ABSTRACT

Diabetic foot ulcers (DFU) are associated with significant impairment of quality of life, increased morbidity and mortality, and are a huge drain on health care resources. In Western countries, the annual incidence of foot ulceration in the diabetic population is around 2%. DFUs develop as a consequence of a combination of factors, most commonly peripheral neuropathy (loss of the gift of pain), peripheral arterial disease (PAD), and some form of unperceived trauma. Recent studies emphasize the very high prevalence of foot ulceration in people with diabetes on dialysis as a consequence of end-stage renal disease. The mortality in this patient group is higher than for most forms of cancer. All patients with diabetes should have an annual screen to identify their foot ulcer risk status: those with any risk factors require specific foot care education as well as regular contact with a health care professional, usually a podiatrist. DFUs should heal if there is an adequate arterial inflow, infection is aggressively managed, and pressure is removed from the wound and its margins. In the management of plantar neuropathic ulcers, offloading is critical and all efforts must be made to enhance patient understanding of the need for offloading. Antibiotic usage should be guided by clinical signs of infection and microbiologic analysis of deep tissue specimens: evidence now exists to show that oral antibiotics are equally efficacious as intravenous in treating most cases of osteomyelitis in the diabetic foot. Most adjunctive therapies have little evidence to support their use although recent trials suggest efficacy for a number of topical therapies including LeucoPatch (3C patch) and sucrose octasulphate; and negative pressure wound therapy has also been shown to be helpful in certain cases. There is currently no indication for hyperbaric oxygen usage, whereas recent studies suggest that topical oxygen therapies help wound healing. Charcot neuroarthropathy (CN) should be easily preventable: most important is to treat any neuropathic patient with a warm swollen foot as having CN until proven otherwise.

INTRODUCTION

At the beginning of the 21^st Century, diabetic foot problems, although eminently preventable, represent one of the commonest causes of hospital inpatient admission in Western countries. In 2005, the International Diabetes Federation realized the global importance of diabetic foot disease and chose to focus their campaign during the whole year on raising awareness with a worldwide campaign to “put feet first” and highlight the common problem of amputation amongst diabetic patients throughout the world. To coincide with World Diabetes Day 2005 (November 14, birth date of Frederick Banting), the Lancet elected to dedicate a whole issue to diabetic foot problems (1).

In this chapter, the global term “diabetic foot” refers to the variety of pathological conditions that might affect the feet in patients with diabetes. Foot ulcers are defined as lesions involving a skin break with loss of epithelium: they can extend into the dermis and deeper layers sometimes involving bone and muscle. Amputation is defined as “the removal of a terminal, non-viable portion of the limb”. The lifetime risk of a person with diabetes developing a foot ulcer (DFU) has been estimated to be as high as 34 % (2).

The suffering of affected individuals and the cost of DFUs are both equally staggering. Those individuals with DFUs usually have other complications of diabetes including nephropathy: data from the UK and the USA confirmed that the outlook for those people with foot complications who are on dialysis is very poor with a high mortality risk (1-3). Data from our group confirm that those people with diabetes who have had an amputation and who are on dialysis have a 75% two-year mortality; the majority of these were of cardiovascular etiology. Data such as these are worse than most malignant diseases, with the possible exception of lung and pancreas. There is therefore an urgent need for preventative strategies to reduce the incidence of foot complications amongst those with diabetes. With respect to costs, in 2008 Rogers et al (4) reported that in the US $18 billion was spent on the care of DFUs and US $11.7 billion on lower extremity amputations. More recently, data from the UK in 2019 suggest that a conservative estimate of the annual cost of diabetic foot problems exceeds UK £900 million which represents approximately 1% of the total budget of the National Health Service (5).

The importance of regular diabetic foot care in very high-risk patients is emphasized by an observational study from Arizona where the State decided to remove routine podiatry from high-risk patients to reduce their health budget. This led to an annual saving of US $351,000 but the cost of this action measured by increased hospitalization, length of stay, and amputations was $16.7 million (6).

This chapter will include a discussion on the epidemiology of foot problems including foot ulceration, amputations, and Charcot neuroarthropathy (CN). The etiopathogenesis will then be described and aspects of management of neuropathic, neuroischemic, and infected DFUs considered. The question of how to address primary and secondary prevention of diabetic foot problems will then be discussed followed by a section on Charcot neuroarthropathy. For more detailed discussion, the reader is referred to review articles on these topics (2,7-9).

EPIDEMIOLOGY OF THE DIABETIC FOOT

The study of the epidemiology of diabetic foot disease has been beset by numerous problems relating to both diagnostic tests used, and population selected. However, there is little doubt that foot complications are common. In the UK, the North West Diabetes Foot Care Study (a community-based study of over 15,000 people) reported that the annual incidence of foot problems amongst the population with diabetes was just under 2% (10), with similar results having been reported from the Netherlands. Similarly, when discussing amputations, the figures vary widely again due to diagnostic criteria as well as regional differences. It must be remembered that many individuals at diagnosis of type 2 diabetes have significant neuropathy: in the United Kingdom Prospective Diabetes Study, for example, 13% of patients at diagnosis had neuropathy of sufficient severity to put them at risk for foot ulceration (7).

With respect to ethnicity, studies from the UK suggest that foot ulcers and amputations appear to be less common in Asians of Indian sub-continental origin and Afro-Caribbean men. In contrast, reports from the USA suggest that amputation rates are more common amongst African-Americans with diabetes than amongst white Americans. Similarly, ulceration is much more common in Hispanic Americans and native Americans than in non-Hispanic whites (2). More recently, reviews have confirmed the importance of healthcare inequities in diabetic foot disease: race, ethnicity, socioeconomic status, and geography are powerful mediators of risk for DFU and lower-extremity amputation (2,11).

ETIOPATHOGENESIS OF DIABETIC FOOT ULCERATION

The foot does not break down spontaneously and in this section, the many warning signs that the feet are at risk of breakdown will be discussed. This was recognized by Elliott Joslin almost 90 years ago when he stated that “diabetic gangrene is not heaven-sent but is earth-borne” (12). It was previously believed that neuropathy, vascular disease, and infection were the main causes of ulceration: it is now recognized that infection occurs as a consequence of ulceration, and is not a cause thereof. There are many contributory factors to foot ulceration, the most important of which are diabetic neuropathy and peripheral arterial disease (PAD). These and other causative factors are listed in table 1.

More common contributory factors shown in bold

Diabetic Neuropathy

Although the association between both somatic and autonomic neuropathy and foot ulceration has been recognized for many years, it is only in the last 20 years that prospective studies have confirmed these assumptions (2,8,10). It has been reported that the risk of developing the first foot ulcer is seven-fold higher in those with moderate to severe sensory loss compared with non-neuropathic diabetic individuals (13). Additionally, poor balance and instability as a consequence of loss of proprioception have been confirmed and are also likely contributory factors not only to foot ulceration, but also to Charcot neuroarthropathy (CN) (2,7,14,15).

Sympathetic autonomic neuropathy in the lower extremity leads to reduced sweating and dry skin that is prone to crack and fissure, and as well, in the absence of PVD, to increased blood flow, arterio-venous shunting, and the warm foot.

As will be discussed later, simple clinical tests may be used to identify the high-risk neuropathic foot (16). Most important in the identification of the high-risk neuropathic foot is good clinical observation and removal of the shoes and socks, with careful inspection of the feet as part of the routine follow up of all patients with diabetes.

Peripheral Arterial Disease (PAD)

A two-center study of causal pathways to foot ulceration reported that peripheral ischemia was a causal component in the pathway to ulceration in 35% of cases (17). In many Western countries, there has been an increase in the percentage of foot ulceration in which ischemia is a contributory factor (18). It is well recognized that patients with diabetes are more prone to distal arterial disease, which may be associated with a poorer outcome.

A detailed discussion of PAD in diabetes is outside the scope of this chapter and readers are directed to reviews on this topic (19, 20). A large follow-up study from Australia has confirmed that the strongest predictors of development of PAD in type 2 diabetes include microvascular complications (particularly macroalbuminuria and photocoagulation for retinopathy (21)).

Other Risk Factors

Of all the risk factors for foot ulceration (table 1), the most important is a past history of ulceration and/or amputation (2). In some series, the annual recurrence rate is up to 50%.

Other Long-term Complications

Those with other late complications particularly nephropathy, have an increased ulcer risk. Visual disturbance as a consequence of retinopathy is a confirmed risk factor; it is easy to understand why this should be. Those patients with sensory loss, particularly large fiber dysfunction, have poor balance and rely on vision as a secondary protective factor. Thus, those who have had for example extensive laser therapy and also have loss of proprioception, are at great risk of foot injury particularly when walking on uneven surfaces and in the hours of darkness.

A strong association between end-stage renal disease and foot ulceration has been emphasized in a number of studies. The temporal association between the start of dialysis treatment and foot ulceration was first confirmed by Game et al (22). A study comprising patients from both the US and the UK subsequently reported a very high prevalence of foot pathology in patients on dialysis, with 46% of patients having past or present foot ulceration and 18% were already amputees (23). The same group later confirmed that being on dialysis is an independent risk factor for foot ulceration in patients with diabetes (3,24). As noted above, preliminary data from the same group suggests that those patients who have already undergone amputation and who are on dialysis have a two-year mortality of up to 75%.

It must also be remembered that patients post-renal transplant or even post-simultaneous pancreas-kidney (SPK) transplant remain at very high risk of developing foot complications. There have been a number of reports of both foot ulceration and Charcot neuroarthropathy occurring in patients post-SPK (25). Theoretically, such subjects are “non-diabetic” but they remain at high risk because they invariably have a dense sensorimotor and autonomic peripheral neuropathy. They should remain under annual review and be coded as ‘diabetes in remission’.

Plantar Callus

Plantar callus forms under weight-bearing areas as a consequence of the dry skin (autonomic neuropathy), insensitivity, and repetitive moderate stress from high foot pressures. Callus itself acts as a foreign body and can cause ulceration in the insensate foot.

Elevated Foot Pressures

Numerous studies have confirmed the contributory role that abnormal plantar pressures play in the pathogenesis of foot ulceration (1, 2, 7). Most studies used sophisticated techniques such as pedobarography to assess foot pressures, but these are not required in day-to-day clinical practice.

Foot Deformities

A combination of motor neuropathy, cheiro-arthropathy, and altered gait pressures is thought to result in the “high-risk” neuropathic foot with clawing of the toes, prominent metatarsal heads, high arch, and small muscle wasting.

Demographics

In Western countries, the male sex has been associated with a 1.6-fold increased risk of foot ulcers (10). There is an increased risk of foot ulceration with increasing age and duration of diabetes.

Psychosocial Factors

There have been a few studies of psychosocial factors in the pathway to foot ulceration and it appears that patients’ behavior is not driven by the abstract designation of being “at risk”; it is driven by patients’ perception of their risk (26) . Thus, if a patient does not believe or understand that a foot ulcer lies on the path from neuropathy to amputation, are they likely to follow educational advice on how to reduce neuropathic ulcers? Moreover, a prospective study has confirmed that depression predicts first, although not recurrent, diabetic foot ulcers (27) .

THE PATHWAY TO FOOT ULCERATION IN DIABETES

As discussed and outlined in Figure 1, the pathway to ulceration is indeed complex and involves an interaction of numerous factors. Whereas none of the factors listed in the last section will alone result in ulceration, it is the interaction and combination of risk factors working together that leads to skin breakdown. In the prospective study of Reiber et al, 63% of all foot ulcers resulted from a combination of neuropathy, deformity, and trauma: in Western countries, the commonest cause of trauma is ill-fitting footwear (17). It must be remembered that as those with neuropathy have reduced sensory input, they will commonly be unable to feel the fit of a shoe until the pressure from the shoe is quite high. Thus, people with neuropathy frequently choose shoes that are too small. All such individuals should be advised to have their feet measured prior to the purchase of any “off the shelf” footwear.

Other simple examples of two risk factors working together in the pathway to ulceration are neuropathy and mechanical trauma (a common scenario is a neuropathic individual with a foreign body in the shoe), neuropathy and thermal trauma (holidays are particularly dangerous), and neuropathy and chemical trauma (such as inappropriate use of over-the-counter chemical corn treatments which should never be used in those with neuropathy).

In summary, whereas neuropathy was present in four out of five cases of new foot ulcers in the Reiber study (17), as noted above, the combination of neuropathy and ischemia is becoming more common in Western countries, and neuro-ischemic ulcers are the commonest type seen in 2023 in diabetic foot clinics.

FOOT ULCERATION

DFUs are common, associated with much morbidity and even mortality but should be eminently preventable. It used to be believed that diabetic foot ulcers were difficult to heal: this is not true: a foot ulcer will heal if it is permitted to do so and this requires attention to three factors-

A That there is adequate arterial inflow to the foot.

B That any infection is appropriately and aggressively managed.

C That all pressure is removed from the wound and its margins.

Figure 1. Pathways to Diabetic Foot Ulceration.

Despite increased knowledge of the pathogenesis and treatment of diabetic foot ulcers in recent years, it is still the third point, offloading the wound, that is poorly adhered to by health care professionals. Many forget that those with a neuropathic or neuroischemic ulcer have “lost the gift of pain”. That pain is a gift which is only realized when it is lost, as first described by Dr Paul Brand when studying leprosy (28). However, before going into more detail on management, it is important to classify wounds appropriately in order to guide therapeutic management.

CLASSIFICATION OF DIABETIC FOOT WOUNDS

Accurate and concise ulcer description and classification systems are required to improve multidisciplinary collaboration and communication, as well as for aiding treatment choices. For many years, the Meggitt-Wagner grading system was regarded as the gold standard. One problem with this system is that the ischemic status of the wound is not included. Thus, a number of new classification systems for diabetic foot wounds have been proposed and validated over the last 20 years. A commonly used system in the United States is the University of Texas Wound Classification System (29). This incorporates the Meggitt-Wagner grades but also enables the practitioner to stage the wound with respect to the presence or absence of infection and/or ischemia (Figure 2). In a comparative prospective study across two Centers, one in the UK and one in the US, the University of Texas Classification System was shown to be superior to the Meggitt-Wagner system at predicting outcomes (30). However, this study also showed that the traditional Meggitt-Wagner system was itself generally accurate in predicting outcomes.

Most recently, the WIFI (Wound Ischemia, Foot Infection) classification was introduced and is the most commonly used today in the USA, particularly in vascular clinics. This was developed and validated as a method to assess three variables – the wound, level of ischemia and the presence and severity of foot infection – to predict the risk of amputation (Figure 3)

Figure 2. The University of Texas Wound Classification System.

Figure 3. WIFI system. Wound, Ischemia, and Foot Infection (WIfI) Classification of Limb Threating diabetic foot disease, tissue loss, ischemia, and infection frequently overlap. However, one is frequently more dominant than the other at different times in the life cycle of an acute-on-chronic event. Here, the amount of tissue loss, ischemia, and foot infection can be ordinally graded to help predict outcome and assist in communicating a plan of action. aA higher score on the WIfI scale is associated with lower extremity amputation and morbidity and can be used to determine the need for revascularization. WIfI scores of 1, 2, 3, and 4 were associated with 1-year amputation rates of 0%, 8%, 11%, and 38%, respectively. Figure from JAMA 2023 Jul 3;330(1):62-75 with permission.

EVALUATION OF THE DIABETIC FOOT ULCER

Clinical evaluation of the foot wound should include a detailed description of the site, size, and depth of the wound. The neuropathic and vascular status of the wound should then be assessed (for details see below). In general, neuropathic ulcers typically occur in the warm but insensate foot, often under pressure bearing areas, and are surrounded by callus. In contrast, ischemic wounds tend to occur in the cool, poorly perfused foot, and are often at lateral fifth metatarsal head regions or the medial first metatarsal head regions. In a predominantly ischemic wound, callus tissue is uncommon. In a neuroischemic wound, the morphology will depend upon the predominance of each of these two pathologies. The correct identification of the degree of ischemia is of the utmost importance when evaluating a wound. If the foot is cool with impalpable pulses, then non-invasive Doppler ultrasound studies are indicated. Conventional methods of assessing tissue perfusion in the peripheral circulation may not be entirely reliable in patients with diabetes. For example, the Ankle Brachial Pressure Index, which is routinely used to screen for PAD in individuals without diabetes, may well be falsely elevated in the those with diabetes because of medial arterial calcification. Toe pressure indices may therefore be more reliable.

Peripheral Arterial Disease

A detailed discussion of vascular procedures is outside the scope of this review, although any person being considered for radiological or surgical procedures will require arteriography. Care must be taken in the use of certain contrast media in patients with chronic renal disease. A detailed discussion of the use of bedside investigations to diagnose PAD in people with diabetes is provided in the recently published guidelines by Fitridge et al (31).

Is Infection Present?

The correct diagnosis of infection in the diabetic foot wound is critical as it is often the combination of untreated infection and PAD that lead to amputation in the diabetic foot. A systematic review that was updated in 2020 still recommend that the diagnosis of infection requiring treatment is a clinical one (32). However, appropriate tissue specimens should be sent to the microbiological laboratory for culture and sensitivity. Superficial swabs are of little use: deep tissue specimens or if osteomyelitis is suspected, bone biopsies are recommended (32) .

A high index of suspicion for the presence of osteomyelitis is essential when assessing the diabetic foot wound. The “probe to bone” (PTB) is often used to diagnose osteomyelitis although there has been much discussion about its accuracy. A systematic review concluded that the PTB test can accurately diagnose osteomyelitis in high-risk patients, and rule out osteomyelitis in low-risk patients (33).

Role of Plain X-Ray in Diagnosing Osteomyelitis

The plain radiograph remains the commonest first radiological investigation of an acutely presenting diabetic foot problem. Despite this, it may be dismissed because of relatively low sensitivity for acute osteomyelitis, with literature over the last 10 years concentrating on CT scanning, MR scanning, and nuclear medicine studies (particularly Gallium Citrate, labelled leucocyte scans and recently PET, PET-CT, SPECT-CT and PET-MR). These latter studies are of limited availability and are expensive, and some carry a high radiation burden. They have their own sensitivity and specificity problems and may not be available in a timely manner. The initial sensitivity of the plain radiograph for acute osteomyelitis is improved by serial studies at one to two-week intervals, during which time therapy for presumed osteomyelitis may be instigated for clinical reasons and whilst awaiting the results of further “high tech” imaging (if still required). The plain radiographic findings could then be considered of high sensitivity and specificity, but with a two-week lag, both for diagnosis and for response to treatment. Appropriate clinical information for the reporting radiologist must include that the patient is diabetic, whether the foot is neuropathic, whether an ulcer is present and if so, its precise anatomical location, and whether it probes to bone. The radiologist should be aware that most sites of acute osteomyelitis in the diabetic foot occur in the floor of an ulcer that probes to bone and that if the foot is neuropathic there may be acute fractures without a history of trauma or acute Charcot neuroarthropathy may be present.

Whilst periosteal reaction is an early feature of osteomyelitis, it is not commonly seen around the small bones of the foot, and if present, is most often seen around metatarsals, and may be due to fracture rather than osteomyelitis.

The hallmark plain radiographic feature of osteomyelitis in the diabetic foot is focal loss of bone density, almost invariably adjacent to the floor of an ulcer. Whilst sometimes described as bone destruction, it is initially bone de-mineralization that causes this appearance, which can reverse on successful treatment, with radiographic re-appearance of the apparently destroyed bone (Figure 4). Obtaining the radiographic view most likely to demonstrate the bone in the floor of an ulcer is therefore an important consideration, often overlooked now that requests are electronic and radiographic views are selected from limited drop-down menus. For example, toe-tip ulcers and ulcers on the dorsum of the inter-phalangeal joints require lateral toe views - best obtained using dental radiographs if available; the inferior surfaces of metatarsal heads are best demonstrated on sesamoid views; the heel requires both lateral and axial views. As a general rule, radiographs tangential to the bone surface at the site of suspected osteomyelitis are ideal, in addition to the standard radiographs of the region. A dedicated team of radiographers familiar with these requirements will improve the relevance and quality of the resultant radiographs.

Plain radiology remains an important investigation in the diagnosis and management of diabetic foot osteomyelitis, but it needs to be of high quality, with appropriate views, and regularly repeated to fulfil its potential.

Figure 4. Acute presentation with an ulcer at the tip of the great toe, probing to bone. The terminal phalangeal tuft does show some irregularity (left panel). B) two weeks later there is marked bone demineralization consistent with osteomyelitis (middle panel). C) After 2 months of treatment there has been partial remineralization of the bone but with an underlying pathological fracture (right panel).

MANAGEMENT OF DIABETIC FOOT ULCERS

The principles of management of different types of foot ulcers will be discussed in brief in this section. The University of Texas Wound Classification System (Figure 2) will be used throughout.

Neuropathic Plantar Ulcers (UT 1A, 1B, 2A, 2B)

As noted above, neuropathic ulcers tend to occur under pressure areas, particularly at the plantar surface of the forefoot. Other recognized sites include the dorsal areas of the toes, particularly the distal inter-phalangeal joint if there is clawing of the toes. In patients with marked deformities such as those caused by Charcot neuroarthropathy, ulcers may occur at other pressure points, particularly in the plantar mid-foot due to, for example, a dropped cuboid bone. When lecturing on the management of neuropathic diabetic foot problems, one is often asked “what can one put on the wound to heal it?”. The answer is invariably that one should be asking “what should one take off the foot to help heal the ulcer?”. Thus, the management of a plantar neuropathic foot ulcer that is not infected is firstly sharp debridement of the ulcer down to bleeding healthy tissue with removal of all callus tissue over the wound and the edge, and secondly, the removal of pressure from the wound while the person is walking. Pain sensation normally protects wounds from further damage causing the non-neuropathic individual to limp. Any subject with a plantar ulcer who walks into the clinic without limping must, by definition, have loss of pain sensation. A neuropathic individual with a plantar ulcer will therefore walk on the ulcer as there is no warning symptom to inform him or her otherwise. Techniques for removing pressure include the use of casts (either removable or irremovable), boots, half shoes, sandals and felted foam dressings. The total contact cast (TCC) is regarded as the gold standard. Studies that randomize patients to an irremovable TCC, a removable cast walker (RCW), or other offloading devices invariably confirm that healing is fastest in the irremovable device (2,7). Although RCWs and irremovable casts (such as the TCC) offload equally well in the gait laboratory, the irremovable device is always associated with more rapid healing in clinical practice. The problem is that those with neuropathic foot ulcers have lost the sensory cue that tells them not to walk on their active ulcer. Studies suggest that individuals are compliant with wearing the offloading RCW during the day, but feel that home is safer and therefore tend to put slippers on, or even walk barefoot at home. A subsequent trial has confirmed that if the RCW is rendered irremovable by wrapping with scotch cast for example, then the outcome is that there is no difference in healing rates between the TCC and the RCW rendered irremovable (34) . Most people with simple neuropathic foot ulcers (UT grades 1A, 2A, 1B, 2B) generally heal in less than three months although of course this does vary with ulcer size. There is no contraindication to casting neuropathic individuals with mild foot infections (UT grades 2A, 2B). It is recommended that after the wound is healed, offloading should continue for a further four weeks to enable the scar tissue to firm up.

Wound dressings are important to keep the ulcer clean, but the placement of a large dressing on a wound may lead the person to a false sense of security by believing that dressing an ulcer is curative. Nothing could be further from the truth in the neuropathic ulcer. Unfortunately, there is little evidence from randomized controlled trials (RCTs) that any dressing is superior to another. Indeed, Jeffcoate et al (35)  randomized people to one of three dressings and could find no difference in outcome according to dressing used: the only difference was in cost. Thus, without an evidence-base, there is no indication to use some of the newer more expensive dressings.

Neuro-ischemic Ulcers

A neuro-ischemic ulcer is one occurring in a foot of a person who has both a neuropathic deficit and impaired arterial inflow: these would be classified UT 1C, 2C in the absence of infection, or 1D, 2D or 3D in the presence of infection. Such individuals warrant full vascular investigation as described above, and referral to the vascular surgery team. The principles of treatment are similar to those for neuropathic ulcers, and it has been confirmed that offloading can safely be used in non-infected neuro-ischemic ulcers under a weight-bearing area. However, antibiotics should be used if there is any suspicion of infection and casting only used with extreme caution in such cases (36). There is now evidence that one dressing, sucrose octasulphate, can improve the healing rates of neuroischemic ulcers in diabetic patients (for further details, see below under adjunctive treatments). With respect to the effectiveness of revascularization of the ulcerated foot in those with neuro-ischemic lesions, results showed that major outcomes following endovascular or open bypass surgery were similar amongst studies (37).

Management of Diabetic Foot Infections

Appropriate wound debridement and offloading together with antibiotics are important in the management of the infected neuropathic foot ulcer, although there are few data from randomized trials to guide the prescriber (32). There is however no evidence that clinically non-infected neuropathic ulcers warrant treatment with antibiotics. With respect to the choice of antibiotic therapy, the reader is directed to the helpful 2012 Infectious Diseases Society of America Clinical Practice Guideline (38).  Commonly used broad-spectrum antibiotics include Clindamycin, Cephalexin, Ciprofloxacin, and the Amoxycillin – Clavulanate potassium. Oral antibiotics usually suffice for mild infections, whereas more severe infections including cellulitis and osteomyelitis require intravenous antibiotic usage initially. Care should also be taken to optimize glycemic control, as hyperglycemia impairs leucocyte function.

The above statements on antibiotics refer to initial treatment: after starting with such broad-spectrum antibiotics, when the results of cultured deep tissue specimens are available, antibiotic therapy should be targeted at the likely primary infective organisms. Finally, with respect to duration of antibiotics, there are no data available from randomized trials to help guide the practitioner. Antibiotics should be continued until clinical signs of infection have resolved, but there is no indication to continue antibiotics beyond this period of time and certainly no indication to continue until the wound has healed. A recent review has identified the challenges facing us due to the increasing threat of multidrug-resistant pathogens (39) .

Osteomyelitis

Diagnosis of osteomyelitis has been discussed above both relating to the PTB test and also the use of plain radiographs. Although the treatment of osteomyelitis has traditionally been surgical, there is increasing evidence from case series and a RCT, that osteomyelitis localized to one or two bones, such as digits, may successfully be treated with antibiotics alone (40, 41). A randomized trial from Spain showed that antibiotics alone were not inferior to localized surgery (41). Again, with respect to duration of antibiotic therapy for osteomyelitis, there is no evidence-base to guide us though a recent trial suggests that six weeks’ antibiotic therapy for non-surgically treated diabetic foot osteomyelitis may be sufficient: traditionally, up to three months has been recommended (42). Lastly, many were surprised to read the results of the OVIVA (Oral Vs Intravenous Antibiotics) study (43) which randomized patients with osteomyelitis to oral vs intravenous antibiotics and showed no superiority of either delivery modality. These observations will certainly challenge the approach to osteomyelitis management in the future. A detailed updated review on infection management has been published by the American Diabetes Association in 2020 (44) .

Adjunctive Treatments

Adjunctive therapies are those which might be considered for complex diabetic foot wounds which fail to heal after 8-12 weeks of standard of care as discussed in the above sections. In recent years, many new such therapies, including skin substitutes, oxygen and other gases, products designed to correct abnormalities of wound biochemistry and cell biology associated with impaired wound healing, applications of cells, bioengineered skin and others, have been proposed to accelerate wound healing in the diabetic foot. Some years ago, an internationally conducted systemic review concluded that there was little published evidence from appropriately designed clinical trials to justify the use of such newer therapies (45).

However, there has been a renaissance in diabetic foot care with many RCTs of new therapies published since 2018 including topical therapies and oxygen-based treatments (46). A number of well-designed RCTs were published in 2018. The first proven therapy for neuro-ischemic ulcers, sucrose octasulfate dressings, was reported in the Explorer study (47). In the active group, 48% of wounds were closed after 20 weeks compared to 30% in the control dressing group (p<0.002). In the same issue of Lancet Diabetes Endocrinology, Game and colleagues reported the positive effect of the Leucopatch (3C Patch) device (a disc containing autologous platelets, leucocytes and fibrin) when applied to the surface of hard-to-heal foot ulcers (48).

Although, as noted above, the International Working Group on the Diabetic Foot (IWGDF) systematic review in 2016 (45) could not support the use of many of the therapies outlined above, this had changed by 2020 when three trials of placenta-derived products were considered (49). Although none was blinded, these were judged to be of low risk of bias as outcomes were assessed in a blinded manner. The first studied a cryo-preserved amniotic membrane allograft (50), the second an umbilical cord product (51), and the third a dehydrated amniotic membrane allograft (52): each showed significantly faster healing in the active treated group versus standard of care. Further details of all these studies referred to above can be found in the most recent American Diabetes Association (ADA) compendium on evidence-based management of complex diabetic foot wounds (53).

Hyperbaric and Topical Oxygen in the Diabetic Foot

HBO has been promoted as an effective treatment in diabetic foot wounds over many years (8). However, early RCTs have been criticized because of the small numbers of patients enrolled, and methodological and reporting inadequacies. A well designed and blinded RCT was conducted in Sweden some years ago suggesting the benefit of HBO in chronic neuro-ischemic infected foot ulcers with no possibility of revascularization (54). More recently, there have been two negative studies including a large retrospective cohort trial (55) and a multi-center Canadian study that showed no benefits of HBO whatsoever in any patient group (56). Thus, at present, the use of HBO in any diabetic foot wound has few data to support its efficacy and the multi-center trial from the Netherlands was also negative (57). The use of HBO in diabetic foot wounds was the topic of a recent debate (58).

There has been increasing interest in the use of topical oxygen-based therapies in wound healing in recent years. Whereas the latest studies of HBO have been negative, there have been interesting developments in the use of devices delivering topical oxygen. There is now evidence that both continuous (59) and cyclical (60) topical wound therapy may improve wound healing rates. A number of more recent studies now support the use of cyclical topical oxygen therapy (TWO2) (53) including some ‘real world’ data (61) and a number of meta-analyses and systematic reviews, the most recent of which has just been published (62). Thus, there is a body of evidence to support the use of TWO2 in the management of hard-to-heal diabetic foot ulcers that fail to respond to standard of care.

Negative Pressure Wound Therapy (NPWT)

The application of NPWT is believed to accelerate healing through reducing edema, removal of exudate, increased perfusion, self-proliferation, and the formation of granulation tissue (63). RCTs have suggested efficacy in rates of wound healing and reduced amputations, with the application of NPWT in both post-surgical and non-surgical chronic non-healing ulcers (64,65). A systematic review confirmed that there was some evidence to support the use of NPWT in post-operative wounds (49).

 ADA Standards of Care and IWDGF Guidelines 2023

The ADA publishes its standards of care and clinical practice guidelines each January in Diabetes Care. In 2023 (66), those adjunctive therapies for foot ulceration recommended and supported by level A evidence (based on large, well-designed randomized controlled trials or well-done meta-analyses of randomized controlled trials) included: negative-pressure wound therapy, placental membranes, bioengineered skin substitutes, several acellular matrices, autologous fibrin and leukocyte platelet patches, and topical oxygen therapy. The IWDGF guidelines are renewed every four years, and in 2023, for adjunctive therapies for foot ulceration, recommended, with variable levels of strength and certainty of evidence, that the following might be considered (67): - regular sharp debridement (strength of recommendation: strong), sucrose-octasulfate impregnated dressings, hyperbaric oxygen in neuro-ischemic or ischemic diabetes-related foot ulcers, topical oxygen therapy, the autologous leucocyte, platelet and fibrin patch (Leucopatch or 3C Patch), placental derived products and Negative Pressure Wound Therapy (only as an adjunct therapy to standard of care for the healing of postsurgical diabetes-related foot wounds).

CHARCOT NEUROARTHROPATHY (CN)

Charcot neuroarthropathy, although uncommon, is a potentially devastating late complication of diabetic neuropathy (68). Although the exact mechanisms resulting in the development of CN remain unclear, much progress has been made in our understanding of the etiopathogenesis of this disorder over the last two decades. CN occurs in a well-perfused foot with both somatic and autonomic neuropathy: the patient presenting with acute CN tends to be slightly younger than is usual for those presenting with foot ulcers. A history of trauma may be present though may be missed because of the severe sensory loss. Although, in its pathogenesis, there are many unanswered questions, improved understanding in recent years of the role of inflammatory pathways might lead to new pharmacologic approaches in the acute phase of the condition. The outcomes in terms of management of CN have been generally poor because of ignorance that leads to delayed diagnosis.

Most important in the management of this condition is recognition of the acute Charcot foot. Any patient with known neuropathy who presents with a warm, swollen foot of unknown causation should be presumed to have acute CN until proven otherwise. Contrary to earlier reports, many patients may present with painful, difficult to describe symptoms in the affected foot despite significant neuropathy.

In its early stages, all investigations may be normal, including the foot x-ray. The role of the radiologist in the diagnosis of acute and chronic CN is discussed in the next section.

Role of Radiologist in in Diagnosing CN

As with acute osteomyelitis (see above), the initial radiographs in acute CN may appear (almost) normal, though it is common for soft tissue swelling to be present and radiographically visible, usually over the dorsum of the foot. It is consequently imperative that both the clinician and the radiologist are aware of the possibility of this condition being present. The first more specific radiographic feature is bone demineralization, usually subchondral or periarticular, around the joint(s) involved by the acute CN process (in contrast to acute osteomyelitis, where it is related to the ulcer location). Focal peri-articular fractures may then develop (Figure 5). If CN is suspected, despite non-diagnostic initial radiographs, then the options are to treat as acute CN (see below) and perform serial radiographs at one-to-two-week intervals until the diagnosis is confirmed or no longer clinically suspected, or treat similarly whilst arranging urgent radiological investigation with a more sensitive test (whilst repeating the radiographs if the further tests are delayed). CT scanning may show small avulsion fractures around midfoot articulations that are invisible on plain radiographs, with minimal increase in the sensitivity and specificity over the plain radiograph, but MR scanning (to include fat suppressed sequences) is better, demonstrating soft tissue edema, bone marrow edema and/or ligamentous disruption. If the MR scan shows no marrow signal abnormality in the foot, acute CN is unlikely. Where the appearances or clinical presentation are complex, with both osteomyelitis and acute CN being suspected, Indium labelled white cell scans and PET/CT have a role, though both can be false positive for osteomyelitis in the presence of acute CN. In infection, MR may demonstrate soft tissue abscesses or sinus tracks that may extend to the (infected) bone surface.

In chronic inactive CN, plain radiographs demonstrate the features of joint distension, destruction, dislocation, disorganization, debris, increased bone density (sclerosis) and deformity. On MR scanning, marrow edema of acute CN is replaced by low signal from sclerosis of the bone. Acute osteomyelitis superimposed on chronic CN produces a mixed picture requiring careful clinical-radiological review.

Diagnosis of acute Charcot neuroarthropathy remains a synthesis of high clinical awareness, clinical findings and radiological findings. The latter should always include serial plain radiography and, where necessary, MR scans.

An overview of imaging in the Charcot foot is available online (69).

Figure 5. Acute Charcot neuroarthropathy. There is widening of the interosseous distance between the medial cuneiform and 2nd metatarsal (arrowheads), indicating disruption of the Lis-Franc ligament and a subtle flake fracture fragment (arrow).

Management of Charcot Neuroarthropathy

The treatment of CN depends upon the stage during which it is diagnosed. The essence of treatment in the acute phase remains non-weight bearing immobilization in a total contact or below-knee cast. Duration of treatment will depend upon response and it is recommended to continue casting until the temperature differential between the active and non-affected foot is down to approximately 1.5°C. As for the foot ulcer, it is recommended that treatment in a cast be continued for up to 4 weeks after the temperature differential has settled. At present, there are no proven medical or pharmacological approaches other than casting that have been shown to improve outcome. The management of advanced CN with bone deformity requiring reconstructive surgery is beyond the scope of this chapter and the reader is referred to a detailed review (70).

PREVENTION OF FIRST AND RECURRENT ULCERS

Prevention will only be successful with the early identification of those patients who have risk factors for foot ulceration. In the 1990s, the concept of the “annual review” was developed, and all those with diabetes should, at whatever stage, be screened for evidence of complications at least annually. The principle aim of such a review is to identify those with early signs of complications and institute appropriate management to prevent progression. The “Comprehensive Diabetic Foot Examination” (CDFE), was developed by a taskforce of the American Diabetes Association (ADA) that was charged with describing what should be included in the annual review for those at risk of foot complications (16). As noted above, the most important aspect of the annual foot review is the removal of shoes and socks with very careful inspection of both feet including between toes. Many neuropathic feet can be identified by this simple clinical observation, looking for features such as small muscle wasting, clawing of the toes, prominence of the metatarsal heads, distended dorsal foot pains (a sign of sympathetic autonomic neuropathy), dry skin, and callus formation. The key components of the diabetic foot annual examination are displayed in table 2.

The ADA Taskforce recommended that for evidence of neuropathy, that the perception of pressure using the 10g monofilament should be used at four sites in each foot (16). An additional test which might include a vibrating 128 Hz tuning fork or others outlined in table 2 should also be used to confirm any abnormality.

For the vascular assessment, foot pulse palpation is most important. Again, as noted above, the ankle brachial index may be falsely elevated in many people with diabetic neuropathy and therefore listening to the Doppler signal may be more helpful as may be a more detailed non-invasive vascular assessment.

More recently, other simple devices for clinical screening have been described. The simplest of all is the “Ipswich Touch Test” developed by Rayman et al in Ipswich, UK. This test simply assesses the ability of the patient to perceive the touch of a finger on the toes (71). The Vibratip which is a battery-operated disposable vibrating stylus can also be used to assess vibration sensation (72), and this has the advantage of using a forced-choice methodology. Both of these tests have been validated in clinical studies (71, 72).

Prevention of Diabetic Foot Ulcers

Surprisingly, there is no evidence from RCTs to confirm the efficacy of preventative foot care education either in the prevention of first foot ulcers or of recurrent foot ulceration (73). This, however, should be interpreted as lack of evidence rather than evidence of no effect. For those patients with no foot ulcer history found to have any of the risk factors listed above or in table 2, they require education in foot self-care and regular podiatric attention.

With respect to secondary prevention, a RCT that looked at the effect of a foot care education program in those with a history of foot ulcers could provide no evidence that such a program of targeted education led to clinical benefit when compared to the usual care (74). It seems likely that those with a history of foot ulcers have such predominant physical abnormalities (e.g., foot deformity, loss of sensation, etc.) that education alone in self-foot care management is insufficient to prevent recurrent ulceration. It may be the combination of foot care education and an intervention that the individual can perform may be more effective. Lavery and colleagues, in studies supported by other RCTs demonstrated in an RCT that patients with a history of neuropathic foot ulcers who were randomized for self-foot temperature monitoring did demonstrate a reduced recurrent ulceration rate. All patients in the active group received foot care education and were provided with a skin thermometer which they used twice a day to check the temperatures of both feet. Those patients who discovered increased unilateral foot temperatures were advised to stop walking and see their health care professional. In the active group there was a highly significant reduction in recurrent foot ulceration (75): however, not all subsequent studies have confirmed this observation (76).

Important in the prevention of foot complications in diabetes is the team approach: members of the team commonly include the diabetes specialists, orthopedic and vascular surgeons, podiatrists, nurse educators, physiotherapists, pedorthists, and others. A study from one district in the UK was able to confirm a 62% reduction in major amputations over an 11-year follow-up period: this decrease occurred after the establishment of such a multi-disciplinary diabetic foot care team (77).

One of the impacts of the recent Covid-19 pandemic has been an explosion in the use of telemedicine and remote monitoring in the care of the diabetic foot (78). A number of studies are currently ongoing looking at “smart technology” in the prevention of recurrent diabetic foot ulcers. These include the use of sensors in socks or shoes to detect pressure change and also various devices to measure differentials in skin temperature: each of these might alert patients in the pre-ulcerative phase with the hope of preventing the actual ulcer from developing. It has now clearly been confirmed that a temperature differential of 2.2C between the feet using remote at home monitoring in patients at high risk of plantar ulceration is a strong predictor of ulcer development (79). Similarly, intelligent pressure sensing insole systems can reduce the incidence of plantar ulcers in those with a past history of ulceration (80). However, face to face consultations remain crucial in the screening for PAD and neuropathy in people with diabetes.

The Foot in Remission

As a recurrence is so common after the healing of neuropathic or neuro-ischemic foot ulcers, it has been suggested that those with a history of foot ulcers should be described as having “a foot in remission” rather than healed. This might better communicate risk of recurrence not only to the patient, but also other healthcare professionals (8). It is hoped that, as in cancers, aggressive treatment during the active disease together with a focus on improving care in “remission” can help to maximize patients’ function and of course improve quality of life (8,26) .

CONCLUSIONS

Although there has been much progress in our understanding of the etiopathogenesis and management of diabetic foot disorders over the last 30 years, much of what we use in clinical practice today still lacks an evidence-base. This is particularly true for example for dressings. The International Working Group on the Diabetic Foot has reported on the details required in the planning and reporting of intervention studies in the prevention and management of diabetic foot lesions (81). Details of the necessary trial design, conduct, and reporting should be taken into account when assessing published studies on interventions in the diabetic foot. Most important of all however in the management of patients with diabetic foot disorders, is to remember that the patient has frequently lost the “gift of pain” that protects most of us from developing significant foot problems but, when absent, can lead to devastating consequences.

REFERENCES

1. Boulton AJM, Vileikyte L, Ragnarson-Tennvall G, Apelqvist J. The global burden of diabetic foot disease.  Lancet 2005; 366: 1719-24.
2. Armstrong DG, Tan TW, Boulton AJM, Bus SA. Diabetic foot ulcers:  a review.  JAMA 2023; 330: 62-75. Doi: 10.1001/JAMA 2023.10578.
3. Lavery LA, Hunt NA, Ndip A, et al. Impact of chronic kidney disease on survival after amputation in individuals with diabetes.  Diabetes Care 2010; 33: 2365-2369.
4. Rogers LC, Lavery LA, Armstrong DG. The right to bear legs.  J Am Podiatr Med Assoc 2008; 98: 166-168.
5. Kerr M, Barron E, Chadwick P, et al. The cost of diabetic foot ulcers and amputations to the National Health Service in England.  Diabetic Med 2019; 36: 995-1002.
6. Skrepnek GH, Mills JL, Armstrong DG. Foot-in-Wallet Disease:  tripped up by “cost-saving” reductions?  Diabetes Care 2014; 37: 196-197.
7. Boulton AJM. The diabetic lower limb – a forty-year journey: from clinical observation to clinical Science. Diabetic Medicine 2019; 36:1539-1549. Doi.org/10.1111/dme.13901.
8. Armstrong DG, Boulton AJM, Bus SA. Diabetic foot ulcers and their recurrence.  New Engl J Med 2017; 376: 2367-2375.
9. Zhang Y, Lazzarini PA, McPhail SM, et al. Global disability burdens of diabetes-related lower-extremity complications in 1990 and 2016.  Diabetes Care 2020; 43: 964-974. Doi: 10.2337/dc19-1614.  
10. Abbott CA, Carrington AL, Ashe H, et al. The North-West Diabetes Foot Care Study:  incidence of, and risk factors for new diabetic foot ulceration in a community-based patient cohort.  Diabetic Med 2002; 19: 377-384.
11. McDermott K, Fang M, Boulton AJM et al. Etiology, Epidemiology and Disparities in the Burden of Diabetic Foot Ulcers.  Diabetes Care 2023; 46: 209-221.  Doi: 10.2337/dci22-0043.
12. Joslin EP. The menace of diabetic gangrene.  N Engl J Med 1934; 211: 16-20.
13. Young MJ, Breddy JL, Veves A, et al. The prediction of diabetic neuropathic foot ulceration using vibration perception thresholds:  a prospective study.  Diabetes Care 1994; 17: 557-560.
14. Brown SJ, Handsaker JC, Bowling FL, et al. Diabetic peripheral neuropathy compromises balance during daily activities.  Diabetes Care 2015; 38: 1116-22.
15. Handsaker JC, Brown SJ, Bowling FL, et al. People with diabetic peripheral neuropathy display a decreased stepping accuracy during walking:  potential implications for risk of tripping.  Diabet Med 2016; 33: 644-649.
16. Boulton AJM, Armstrong DG, Albert SF, et al. Comprehensive foot examination and risk assessment:  a report of the taskforce of the Foot Care Interest Group of the American Diabetes Association, with endorsement by the American Association of Clinical Endocrinologists.  Diabetes Care 2008; 31: 1679-1685.
17. Reiber GE, Vileikyte L, Boyko EJ, et al. Causal pathways for incident lower-extremity ulcers in patients with diabetes from two settings.  Diabetes Care 1999; 22: 157-162.
18. Prompers L, Huijberts M, Apelqvist J, et al. High prevalence of ischemia, infection and serious comorbidity in patients with diabetic foot disease in Europe.  Baseline results from the Eurodiale study.  Diabetologia 2007; 50: 18-25.
19. Hinchliffe RJ, Forsythe RO, Apelqvist J, et al. Guidelines on the diagnosis, prognosis and management of peripheral artery disease in patients with foot ulcers and diabetes.  Diabetes Metab Res Rev 2020; 36 (suppl 1): e3276.doi10.1002/dmrr.3276.
20. Forsythe RO, Hinchliffe RJ. Management of critical limb ischemia in the patient with diabetes.  J Cardiovasc Surg (Torino) 2016; 57: 273-281.
21. Mohammedi K, Woodward M, Hirakawa Y, et al. Microvascular and macrovascular disease and risk for major peripheral arterial disease in patients with type 2 diabetes.  Diabetes Care 2016; 39: 1796-1803.
22. Game FL, Chipchase SY, Hubbard R, et al. Temporal association between the incidence of foot ulceration and the start of dialysis in diabetes mellitus.  Nephrol Dial Transplant 2006; 21: 3207-3210.
23. Ndip A, Lavery LA, Lafontaine J, et al. High levels of foot ulceration and amputation risk in a multiracial cohort of diabetic patients on dialysis therapy.  Diabetes Care 2010; 33: 878-880.
24. Ndip A, Rutter MK, Vileikyte L, et al. Dialysis treatment is an independent risk factor for foot ulceration in patients with diabetes and stage 4 or 5 chronic kidney disease.  Diabetes Care 2010; 33: 1811-1866.
25. Barrado FC, Kuypers DK, Matricali GA. Charcot neuroarthropathy after simultaneous pancreas-kidney transplantation:  risk factors, prevalence and outcome. Clin Transplant 2015; 29: 712-719.
26. Vileikyte L, Pouwer F, Gonzalez JS. Psychosocial research in the diabetic foot:  are we making progress?  Diabetes Metab Res Rev 2020; 36 (suppl 1) e3257.doi10.1002/dmrr3257.
27. Gonzalez JS, Vileikyte L, Ulbrecht JS, et al. Depression predicts first but not recurrent diabetic foot ulcers.  Diabetologia 2010; 53: 2241-2248.
28. Boulton AJM. Diabetic Neuropathy: is pain God’s great gift to mankind?  Semin Vasc Surg 2012; 25: 61-65.
29. Armstrong DG, Lavery LA, Harkless LB. Validation of a diabetic wound classification system.  The contribution of depth, infection, and ischemia to risk of amputation.  Diabetes Care 1998; 21: 855-859.
30. Oyibo S, Jude EB, Tarawneh I, et al. A comparison of two diabetic foot ulcer classification systems:  the Wagner and the University of Texas wound classification system.  Diabetes Care 2001; 24: 84-88.
31. Fitridge R, Chuter V, Mills JL et al.  . The 2023 intersocietal IWGDF, ESVS, SVS guidelines on peripheral arterial disease in patients with diabetes and a foot ulcer.  iwgdfguidelines.org. 2023 
32. Senneville E, Lipsky BA, Abbas ZG, et al. Diagnosis of infection in the foot in diabetes:  a systematic review.  Diabetes Metab Res Rev 2020; 36 (suppl 1): e3281.doi:10.1002/dmrr3281.
33. Lam K, van Asten SA, Nguyen T, et al. Diagnostic accuracy of probe to bone to detect osteomyelitis in the diabetic foot:  a systematic review.  Clin Infect Dis 2016 63: 944-948.
34. Katz IA, Harlan A, Miranda-Palma B, et al. A randomized trial of two irremovable offloading devices in the management of plantar neuropathic diabetic foot ulcers.  Diabetes Care 2005; 28: 555-559.
35. Jeffcoate WJ, Price PE, Phillips CJ, et al. Randomized controlled trial of the use of three dressing preparations in the management of chronic ulceration of the foot in diabetes.  Health Technol Assess 2009; 13: 1-86.
36. Nabuurs-Fransen MH, Sleegers R, Huijberts HS, et al. Total contact casting of the diabetic foot in daily practice:  a prospective follow-up study.  Diabetes Care 2005; 28: 243-247.
37. Hinchliffe RJ, Brownrigg JR, Andros G, et al. Effectiveness of revascularization of the ulcerated foot in patients with diabetes and peripheral artery disease:  a systematic review.  Diabetes Metab Res Rev 2016; 32 (suppl 1): 136-144.
38. Lipsky BA, Berendt AR, Cornia PB, et al. 2012 Infectious Diseases Society of America clinical practice guideline for the diagnosis and treatment of diabetic foot infections.  Clin Infect Dis 2012; 54: e132-173.
39. Ertugrul B, UcKay I, Schoni M, Peter-Riesch B, Lipsky BA. Management of diabetic foot infections in the light of recent literature and new international guidelines.  Expert Rev Anti-Infect Ther 2020; Feb 19 Epub.
40. Game FL, Jeffcoate WJ. Primary non-surgical management of osteomyelitis of the foot   in diabetes.  Diabetologia 2008; 51: 962-967.
41. Lázaro-Martinez JL, Aragón-Sánchez J, Garcia-Morales C. Antibiotics versus conservative surgery for treating diabetic foot osteomyelitis: a randomized comparative trial.  Diabetes Care 2014; 37: 789-795.
42. Tone A, Nguyen S, Devemy F, et al. Six-week versus twelve-week antibiotic therapy for non-surgically treated diabetic foot osteomyelitis: a multicentre open-label controlled randomized study.  Diabetes Care 2015; 38: 302-307.
43. Li HK, Rombach I, Zambellas R, et al. OVIVA Trial collaborators.  Oral Vs IntraVenous Antibiotics for bone and joint infection.  N Engl J Med 2019; 380: 425-436.
44. Boulton AJM, Armstrong DG, Hardman MJ, et al. Diagnosis and management of diabetic foot infections.  Arlington (VA).  American Diabetes Association 2020.
45. Game FL, Apelqvist J, Attinger C, et al. Effectiveness of interventions to enhance healing of chronic ulcers of the foot in diabetes:  a systematic review.  Diabetes Metab Res Rev 2016; 32 (suppl 1):  154-168.
46. Edmonds ME. A renaissance in diabetic foot care:  new evidence-based treatments.  Lancet Diabet Endocrinol 2018; 6: 837-838.
47. Edmonds ME, Lázaro-Martinez JL, Alfayate-Garcia JM, et al. Sucrose octasulfate dressing versus control dressing in patients with neuroischemic diabetic foot ulcers (Explorer):  an international, multicenter, double-blind, randomized controlled trial.  Lancet Diabetes Endocrinol 2018; 6: 186-196.
48. Game F, Jeffcoate W, Tarnow L, et al. LeucoPatch system for the management of hard-to-heal diabetic foot ulcers in the UK, Denmark, and Sweden:  an observer-masked, randomized controlled trial.  Lancet Diabetes Endocrinol 2018; 6; 870-878.
49. Rayman G, Vas P, Dharariya K, et al. International Working Group on the Diabetic Foot (IWGDF).  Guidelines on use of interventions to enhance healing of chronic foot ulcers in diabetes  (IWGDF 2019 update).  Diabetes Metab Res Rev 2020; 36 (suppl 1): 3283.
50. Lavery LA, Fulmer J, Shebetka KA, et al. Grafix Diabetic Foot Ulcer Study Group.  The efficacy and safety of Grafix for the treatment of chronic diabetic foot ulcers:  results of a multi-centre, controlled, randomized, blinded, clinical trial.  Int Wound J 2014; 11: 554-560.
51. Tettelbach W, Cazzell S, Sigal F, et al. A multicentre prospective randomized controlled comparative parallel study of dehydrated human umbilical word (EpiCord) allograft for the treatment of diabetic foot ulcers.  Int Wound J 2019; 16: 122-130.
52. Tettelbach W, Cazzell S, Reyzelman AM, Sigal F, Caporusso JM, Agnew PS. A confirmatory study on the efficacy of dehydrated human amnion/chorion membrane dHACM allograft in the management of diabetic foot ulcers:  a prospective, multicentre, randomized, controlled study of 110 patients from 14 wound clinics.  Int Wound J 2019; 16: 19-29.
53. Boulton AJM, Armstrong DG, Löndahl M, et al. New Evidence-Based Therapies for Complex Diabetic Foot Wounds. Arlington (VA) May 2022: American Diabetes Association. doi: 2337/db2022-02.
54. Löndahl M, Katzman P, Nilsson A, et al. Hyperbaric oxygen therapy facilitates healing of chronic foot ulcers in patients with diabetes.  Diabetes Care 2010; 33: 998-1003.
55. Margolis DJ, Gupta J, Hoffstad O, et al. Lack of effectiveness of hyperbaric oxygen therapy for the treatment of diabetic foot ulcer and the prevention of amputation:  a cohort study.  Diabetes Care 2013; 36: 1961-1966.
56. Fedorko L, Bowen JM, Jones W, et al. Hyperbaric oxygen therapy does not reduce indications for amputation in patients with diabetes with non-healing ulcers of the lower limb:  a prospective, double-blind, randomized controlled clinical trial.  Diabetes Care 2016; 39; 392-399.
57. Santema KTB, Stoekenbroek RM, Koelemay MJW, et al. Hyperbaric oxygen therapy in the treatment of ischemic lower-extremity ulcers in patients with diabetes:  results of the DAMO₂CLES multicenter randomized controlled trial.  Diabetes Care 2018; 41: 112-119.
58. Löndahl M, Boulton AJM. Hyperbaric oxygen:  useless or useful?  A debate.  Diab Metab Res Rev 2020 Mar;36 Suppl 1:e3233. doi: 10.1002/dmrr.3233.
59. Niederauer MQ, Michalek JE, Liu Q, et al. Continuous diffusion of oxygen improves diabetic foot ulcer healing when compared with a placebo control:  a randomized, double-blind, multicenter study.  J Wound Care 2018; 27 (suppl 9): S30-S45.
60. Frykberg RG, Franks PJ, Edmonds ME, et al. A multinational, multicenter, randomized, double-blinded, placebo-controlled trial to evaluate the efficacy of cyclical topical wound oxygen therapy (TWO2) in the treatment of chronic diabetic foot ulcers:  the TWO2 study.  Diabetes Care 2020; 43: 616-624.
61. Yellin JI, Gaebler JA, Zhou FF, et al. Reduced hospitalizations and amputations in patients with diabetic foot ulcers treated with cyclical pressurized topical wound oxygen therapy:  real-world outcomes.  Adv Wound Care (New Rochelle).  Online ahead of print on 6 December 2021 (doi: 10.1089/wound.2021.0118).
62. Carter MJ, Frykberg RG, Oropallo A, et al. Efficacy of topical oxygen therapy in healing chronic diabetic foot ulcers:  systematic review and meta-analysis.  Adv Wound Care 2023, Apr; 12 (4): 177-186. doi: 10.1089/wound.2022.0041.
63. Isaac AL, Armstrong DG. Negative pressure wound therapy and other new therapies for diabetic foot ulceration: the current state of play.  Med Clin North Amer 2013; 97: 899-909.
64. Armstrong DG, Lavery LA. Negative pressure wound therapy after partial diabetic foot amputation:  a multicenter, randomized controlled trial.  Lancet 2005; 366: 1704-1710.
65. Blume PA, Walters J, Payne W, et al. Comparison of negative pressure wound therapy using vacuum-assisted closure with advanced moist wound therapy in the treatment of diabetic foot ulcers:  a multicenter, randomized controlled trial.  Diabetes Care 2008; 31: 631-636.
66. El Sayed NA, Aleppo G, Aroda VR, et al. Retinopathy, Neuropathy and Foot Care:  Standards of Care in Diabetes – 2023.  Diabetes Care 2023; 46 (Suppl 1): S203-S215.  doi: 10.2337/dc23-S012.
67. Chen P, Vilorio NC, Dhatariya K, et al: Guidelines on interventions to enhance healing of foot ulcers in people with diabetes (IWGDF 2023 update).  Diab Metab Res Rev 2023; May 25; 3644 doi: 10.1002/dmrr.3644.
68. Jeffcoate WJ. Charcot foot syndrome.  Diabetic Med 2015; 32: 760-770.
69. Ergen FB, Sanverdi SE, Oznur A. Charcot foot in diabetes and an update on imaging.  Diabetic Foot Ankle 2013; 4: 10.3402/dfa.V4i0.21884.
70. Shen W, Wukich D. Orthopedic surgery and the diabetic Charcot foot.  Med Clin North Am 2013; 97; 873-882.
71. Rayman G, Vas PR, Baker N, et al. The Ipswich Touch Test:  a simple and novel method to identify inpatients with diabetes at risk of foot ulceration.  Diabetes Care 2011; 34: 1517-1518.
72. Bowling FL, Abbott CA, Harris WE. A pocket-sized disposable device for testing the integrity of sensation in the outpatient setting.  Diabet Med 2012; 29: 1550-1552.
73. Dorresteijn JA, Valk GD. Patient education for preventing diabetic foot ulceration.  Diabetes Metab Res Rev 2012; 28 (suppl 1): 101-106.
74. Lincoln NB, Radford KA, Game FL, et al. Education for secondary prevention of foot ulcers in people with diabetes:  a randomized controlled trial.  Diabetologia 2008; 51 1954-1961.
75. Lavery LA, Higgins KR, Lanctot DR, et al. Preventing diabetic foot ulcer recurrence in high-risk patients:  use of temperature monitoring as a self-assessment tool.  Diabetes Care 2007; 30: 14-20.
76. Aan de Stegge WB, van Netten JJ, Bus SA. Does the skin heat up before it breaks down in diabetic foot ulceration?  Diabete Metab Res Rev 2023, 08 Feb 2023.  https://doi.org/10.1002/DMRR.3621.
77. Krishnan S, Nash F, Baker N, et al. Reduction in diabetic amputations over eleven years in a defined UK population:  benefits of multidisciplinary teamwork and continuous prospective audit.  Diabetes Care 2008; 31: 99-101.
78. Boulton AJM. Diabetic Foot Disease during the Covid-19 pandemic.  Medicina 2021; Jan 22; 57 (2): 97. Doi; 10.3390/medicina57020097.
79. Frykberg RG, Gordon IL. Reyzelman AM, et al.  Feasibility and efficacy of a smart mat technology to predict development of diabetic plantar ulcers.  Diabetes Care 2017; 40: 973-980.doi: 10.2337/dc16-2294.
80. Abbott CA, Chatwin KE, Foden P, et al. Innovative intelligent insole system reduces diabetic foot ulcer recurrence at plantar sites:  a prospective, randomized, proof-of-concept study.  Lancet Digit Health 2019; Oct 1 (6): e308-e318.
81. Jeffcoate WJ, Bus SA, Game FL, et al. Reporting standards of studies and papers on the prevention and management of foot ulcers in diabetes:  required details and markers of good quality.  Lancet Diabetes Endocrinol 2016; 4: 781-788.

ABSTRACT

It is recognized that traditional measures of glucose control (such as hemoglobin A1c [A1C]) provide little information regarding the need for day-to-day changes in therapies. While intermittent self-monitored blood glucose (SMBG) provides additional information with which to make treatment decisions, significant barriers to its use exist, such asinconvenience and lack of timely and regular feedback. Furthermore, important information regarding glucose trends may be missed. Continuous glucose monitoring (CGM) has become increasingly reliable and has demonstrated efficacy in terms of improving A1C, reducing hypoglycemia, and improving the time in target glucose range. Incremental progress continues to be made toward a fully functional artificial pancreas, of which CGM will play a vital role. As more and more data are presented to patients and providers, it has become increasingly paramount that the data are organized in a standardized way and that communication of data is streamlined using patients’ mobiledevices where available and within the existing clinic infrastructure. Systems that provide immediate feedback to patients and decision support tools for patients and providers have demonstrated superior outcomes compared to routine SMBG alone. Alternate markers of glucose control may provide complementary information about glucose control and long-term prognosis. This chapter will review the latest evidence for use of professional and personal CGM, mobile glucose monitoring approaches, and biomarkers of glycemic control.

INTRODUCTION

The current technology for monitoring of glucose levels has been well established since the 1980′s. This practice is beneficial to patients with diabetes from both a clinical and an economic standpoint when used optimally. Knowledgeof the glucose levels that are measured can allow a patient to select an appropriate dose of insulin or implement dietary or other lifestyle changes to regulate their glucose levels. Expert groups provide recommendations for glucose targets, including A1C, self-monitored blood glucose (SMBG), and interstitial glucose (1,2). Although targets vary, expert groups recommend individualization based upon risk of hypoglycemia, polypharmacy, comorbidities, and other characteristics that may affect long-term benefit and individual patient characteristics. The ADA has expanded recommendations for assessing overall glucose levels to include the A1C or CGM metrics such as % Time in Range(TIR, the % of time spent 70-180 mg/dl), or the Glucose Management Indicator (GMI), which is an estimate of A1C that is derived from a 14-day CGM report for routine assessment of glucose levels (1).

The landscape of glucose monitoring technologies is expanding and rapidly changing. For a full review of glucose monitoring technologies, the reader is referred to one of many excellent reviews referenced throughout this chapter. Several trends are emerging in glucose monitoring and will be reviewed in more detail in this chapter:

• CGM: This practice is becoming more widely established as evidence supporting its use has accumulated. The data available through CGM can permit significantly more fine-tuned adjustments in insulin dosing and other therapies than spot testing from self-monitoring of blood glucose (SMBG) can CGM technologies forautomatic collection of data have spurred interest in noninvasive glucose monitoring as an additional tool for obtaining information about glucose levels.
• Closed loop control (CLC): Also known as an “artificial” or “bionic” pancreas, this technology links CGM withautomatically controlled insulin The first steps toward CLC are now in use.
• Mobile Technology and Decision Support: In recent years, increasing connectivity between glucose monitoringtechnologies and mobile devices has facilitated ongoing improvements in self-care and communication of data.
• Alternate Markers of Glucose Control: Finally, the use of additional analytes besides glucose is still being established.

This chapter analyzes the technology, benefits, and problems with the use of intermittent SMBG and CGM, mobile technology and decision support, and alternate biomarkers of glycemic control.

CONTINUOUS GLUCOSE MONITORS

CGM measures glucose levels (typically interstitial glucose) continuously and updates the glucose level display every 5 minutes. Most CGMs consist of 1) a monitor to display the information (in some cases, this is the patient’s mobile device), 2) a sensor that is usually inserted into the subcutaneous tissue, and 3) a transmitter that transmits the sensor data to the monitor. Previously, all devices were approved for adjunctive use only due to limitations in accuracy; in this case patients must still perform fingerstick glucose monitoring in order to guide therapy and perform calibrations. However, in 2016, the FDA approved the use of the Dexcom G5 as the first CGM for stand-alone use.Newer technologies have eliminated the requirements for calibration of CGM with a fingerstick glucose. The accuracy of all commercially available CGMs is still the lowest in the hypoglycemic range, which is where the need for sensitivity and specificity is great in terms of serving as an alarm for hypoglycemia.

CGM can provide both retrospective as well as real-time information to detect: 1) hypoglycemic and hyperglycemic excursions; 2) predict impending hypoglycemia; and 3) wide fluctuations in glucose levels, also known as glycemicvariability. 24-hour telephone support is available for all FDA approved CGM devices. Use of CGM can help both thepatient and their medical provider make fine tune adjustments to medication therapy and provide insight to the patient on behavioral changes to achieve glycemic control. Additionally, current efforts to link CGM measurement with automatically controlled insulin delivery, has progressed incrementally toward a fully functional artificial pancreas.Systems can be divided according to their intended use as professional CGM (which is a clinic-owned device and provides either retrospective or real-time glucose data) and personal CGM (which is patient-owned and provides real-time glucose data).

PROFESSIONAL CGM

Professional CGM describes CGM data that are obtained via healthcare provider owned equipment. It does not necessarily provide the glucose results in real time, but downloads the readings after they have been collected, similar to a 24-hour cardiac Holter monitor that provides information about cardiac rhythms after they have occurred. This allows the health care provider to obtain relatively unbiased glucose patterns during typical everyday life. The Endocrine Society recommendations state that professional CGM may be of benefit in adults with diabetes to detectnocturnal hypoglycemia, dawn phenomenon, postprandial hyperglycemia and to assist in management of diabetes therapies (3). Professional CGM is more readily reimbursed than personal CGM, but interpretation of both personal and professional CGM reports by qualified healthcare professionals may be reimbursed on a monthly basis.

Some personal CGM systems can be operated in a blinded fashion in order to provide professional glucose data. These systems will be discussed in more detail later (see “Personal [Real-time] Continuous Glucose Monitoring”). The first device for reading blood glucose levels continuously was a professional CGM that was approved by the FDA inJune 1999. This device was the Continuous Glucose Monitor System (CGMS) manufactured by Medtronic MiniMed (Medtronic Diabetes, Northridge, CA) (4). Since then, newer models have shown improvements in accuracy and patient acceptance. In a meta-analysis of 22 articles, professional CGM resulted in a greater reduction in A1c(-0.28%, 95% CI -0.36% to -0.21%, P < 0.00001) as well as TIR (5.59%, 95% CI 0.12 to 11.06, P = 0.05) compared to usual practice (5).

FreeStyle Libre Pro

The FreeStyle Libre Pro utilizes the same sensor as the Libre personal CGM. The Libre is factory calibrated andtherefore does not require self-monitored blood glucose calibrations. This may be a potential advantage since capillary blood glucose testing is subject to various system and user errors, which in addition to the physiologic lag time between blood and interstitial glucose (which is magnified in the postprandial period) could contribute to CGM error. Itcollects up to 14 days of glucose readings, which are recorded every 15 minutes. The glucose sensor is fully disposable and a single reader is used to activate and scan multiple devices, allowing multiple patients in one office to undergo the procedure simultaneously. Reports are obtained through the LibreView website, which offers a secure cloud-based system, or the FreeStyle Libre desktop reporting software. Reports provide daily patterns, an assessmentof glucose variability and hypoglycemia risk, a daily glucose report, and an overall snapshot report.

The overall MARD (Mean Absolute Relative Difference which is calculated by averaging the absolute values of relative differences between CGM measurement results and corresponding comparison method results) for the FreeStyle Libre is 11.4%, 86.7% of readings were in Zone A of the Consensus Error Grid analysis, and 99.7% of results were in Zones A and B (6). It is important to note that sensor accuracy is lower on day 1 and in thehypoglycemia range (MARD 20.3% for values <72 mg/dl in one study) (7). Accuracy improves and remains steady over the 14-day wear period. The Libre utilizes glucose oxidase in a “direct signaling” approach that is not dependenton oxygen and minimizes interference by other substances, such as acetaminophen, which may falsely elevated readings on other devices.

Dexcom Professional

The Dexcom G6 Pro was approved by the FDA in March 2018 and is available in blinded or unblinded mode depending upon whether the goal is to observe glucose patterns without intervention, to provide immediate feedback to educate and inform patients about their medications and behaviors, or to facilitate decisions about pursuingpersonal CGM. The sensor, transmitter, and receiver are essentially identical to the personal Dexcom G6 system and features expedited startup time and no calibration. The device measures interstitial glucose levels every 5 minutes and is approved for 10 days of use. The device is downloaded using Dexcom CLARITY, a web-based software program that is also used to download and review personal data.

Analysis of Retrospective Data

Data from all CGM devices can be studied retrospectively after downloading (8). It is recommended that diet, activity, symptom, and insulin data are collected during professional CGM to assist with interpretation, either via patient diary,direct entry of events into the device, or use of an accompanying app, depending on the system. Three time periods should be analyzed. These are:

• Overnight: Out-of-target overnight glucose levels can be modified by adjusting the basal insulin dose.
• Pre-prandial Period: Out-of-target pre-prandial glucose levels can be modified by adjusting the previous meal bolus, meal, or exercise pattern.
• Post-prandial period: Out-of-target postprandial glucose levels can be modified by adjusting the immediate meal bolus, meal, or exercise pattern.

In certain special situations, targets may need to be adjusted. Other important elements of a professional CGM analysis are shown in Table 1. An example of a patient who used CGM is presented in Figure 1. The CGMdemonstrated high glucose levels from 6:00 PM to 11:00 PM post-supper and low glucose levels from 12:00 AM to 2AM. Recognition of these patterns allowed appropriately timed treatment interventions.

Ambulatory Glucose Profile

The ambulatory glucose profile (AGP, Figure 2) is a standardized reporting format for glucose data that was developed by an expert panel of diabetes specialists and sponsored by the Helmsley Charitable Trust and is customized for insulin pumps or injection therapy (9). The universal report is intended to simplify and facilitate interpretation of otherwise complex and lengthy reports with varying terminology. It is anticipated that a standardized report would “help clinicians develop expertise in CGM use, enhance quality of care through enhanced pattern recognition, improve practice efficiencies with minimal disruption of workflow, and engage patients, thereby reinforcing consistent use of CGM technology.” A single page report that the medical team can view and file into a patient’s electronic medical record and that can be used as a shared decision-making tool with people with diabetes wasconsidered to be of great value in the report of the 12th International Conference on Advanced Technologies & Treatments for Diabetes (ATTD 2019) (10). The AGP is currently employed by many reporting systems and consists of 3 components:

• Statistical Summary, which utilizes standard metrics and terminology to summarize the number of values,percentage of values, and time in target, above target, and below target, as well as an assessment of glucose variability.
• Modal day report which collapses data from days or weeks to a single day in order to identify patterns by time of day. Data are presented graphically as 5 distribution curves, representing the median, interquartile range, and10^th to 90^th percentiles, on the backdrop of target range.
• Daily View, which facilitates review of within day

Composite Metrics

As a measure of the quality of glycemia, the time in range (TIR), similar to the A1C is limited in its assessment of hypoglycemia. Multiple composite metrics have thus been reported (11). However, the use of multiple metricsincreases complexity and is subject to issues with collinearity. The Glycemia Risk Index (GRI) is a composite metricthat was developed using input from 330 clinical experts who analyzed 14-day tracings from 225 adults with diabetes (12). GRI more heavily weights very high or very low glucose values and correlates with clinician rankings more closely than TIR or time below range (%time < 70 mg/dl, TBR) alone.

Figure 2. Ambulatory Glucose Profile for Insulin Pumps.
Glucose Statistics: Metrics include mean glucose, estimated A1C, glucose ranges, coefficient of variation and standard deviation.
Glucose Profile: Daily glucose profiles are combined to make a one-day (24-hour) picture. Ideally, lines would stay within grey shaded area (target range).
Orange: median (middle) glucose line.
Blue: area between blue lines shows 50% of the glucose values.
Green: 10% of values are above (90% top line) and 10% are below (10% bottom line). Insulin Profile Graph: Shows basal insulin pump settings over a 24-hour period.
Bolus Insulin Graph: Combines all bolus insulin doses into one graph to make a one-day (24-hour) picture. Each box on the graph covers 60 minutes of doses.
Orange: median (middle) dot.
Blue: shaded box shows 50% of the bolus dosages in the hour.
Green: lines above and below the shaded box (whiskers) show how many of the bolus dosages per hour were between 75 - 90% and between 10 - 25%.

PERSONAL REAL-TIME CGM (RT-CGM) OR INTERMITTENTLY SCANNED CGM (IS-CGM)

RT-CGM devices not only display the current glucose every few minutes, but may also alert the patient for impending (projected alert) or actual (threshold alert) hyperglycemia or hypoglycemia or rate of change in glucose. Bycomparison, is-CGM requires patient interaction with the device to obtain readings but may still provide alerts for hypoglycemia or hyperglycemia. While few head to head studies are available, some studies suggest greater reduction in hypoglycemia and improvement in TIR with RT-CGM compared to is-CGM in persons with type 1 diabetes (13,14), even up to 24 months (15).

Over time, accuracy with RT-CGM and is-CGM has improved substantially (16,17,18). In fact, some devices, including the Dexcom and Freestyle Libre are approved for stand-alone use, meaning that under specified conditions, the device may be used to make treatment decisions without confirmatory blood glucose measure. However, the user will still experience a tradeoff between a high alarm sensitivity and specificity for detecting hypoglycemic events, particularly where glucose levels are changing rapidly (Figure 3). Current and recent glucose levels, trend information,and a visual alarm are all presented so that a patient can predict future low or high glucose excursions. Using this information will allow the patient to take actions to spend more time in the euglycemic range and less time in the hypoglycemic or hyperglycemic ranges. This potential decrease in glycemic variability will not necessarily be reflected in an improved A1C value, which reflects mean glycemic levels.

Figure 3. Tradeoffs between emphasis on high sensitivity compared to emphasis on high specificity in a hypoglycemic alarm that is part of a continuous glucose monitor.

Evidence- Type 1 Diabetes

Studies may be divided according to background therapies (insulin pump or injection therapy).

STUDIES UTILIZING EITHER INSULIN PUMP OR INJECTIONS AS BACKGROUND THERAPIES

• The seven-country GuardControl Study was the first randomized controlled trial to ever demonstrate a statisticallysignificant improvement in A1C levels with the use of RT-CGM (19). The Guardian RT was used either continuously or biweekly for three months and both regimens were compared to control treatment which did not include use of CGM. At one month and at three months the continuous users had significantly lower A1C levelsthan the controls. The biweekly users had intermediate improvement which did not reach statistical significance compared to the outcomes in the control group.
• In 2008, the Juvenile Diabetes Research Foundation Continuous Glucose Monitoring Study Group evaluated 322 adults and children with type 1 diabetes (either injection or insulin pump therapy) and A1C 7-10% who wererandomized to either RT-CGM or usual care (20). RT-CGM was associated with a 53% reduction in A1Ccompared to usual care (p<0.001), but was only significant among subjects over age 24 due to lack of consistent use in younger patients. Hypoglycemia was infrequent and was not different between groups.
• In 2011, 120 children and adults with type 1 diabetes on insulin pump or injection therapy and A1C <7.5% wererandomly assigned to RT-CGM (Freestyle Navigator—not available in the US) or masked CGM every other week (21). The time spent in hypoglycemia was reduced over 50% at 26 weeks, and patients spent more time in 70-180 mg/dl range.
• In the IMPACT trial, 241 adults with type 1 diabetes with an A1C less than or equal to 7.5% were randomly assigned to Freestyle flash glucose monitoring (described in more detail under “Overview of Stand-Alone Personal CGM systems”) vs. SMBG. In this group 68% of the patients were treated with multiple daily injections and 32% with CSII. The amount of time spent in hypoglycemia was decreased by nearly 90 minutes per day (P<0.0001) when patients had access to CGM data (22). It must be noted that this technology does not provide real-time alerts for impending hypoglycemia or hyperglycemia and data are accessed via a hand-held device on Ina small study of patients with hypoglycemia unawareness or recent severe hypoglycemia, RT-CGM more effectively reduced the time spent in hypoglycemia compared to flash glucose monitoring (23).
• The CITY study was a randomized study among 153 adolescents and young adults with type 1 diabetes. CGM resulted in a -0.37% greater reduction in A1C compared to usual care (p=0.01) (24). In this study, only 68% of participants used CGM at least 5 days per week in month 6, which is significantly lower than studies reported inadults (25). However, this is more than twice that reported in the pivotal JDRF study of 2008 (20). Moreover, this study utilized an earlier generation CGM which required twice daily calibration; thus, it is possible that newer technologies may support greater persistence with use.
• Among 203 older adults (median age 68) with type 1 diabetes randomized to CGM or usual care, CGM resulted in less hypoglycemia at 26 weeks (estimated treatment difference 27 minutes/day, p<0.001) as well as modest improvement in A1C (estimated treatment difference -0.3%, p<0.001) (26). The improvement in hypoglycemia was sustained over 52 weeks, at which point CGM use was still >90% (27).

STUDIES UTILIZING INSULIN PUMP THERAPY AS BACKGROUND

• In the largest study to date, the STAR3 study, 485 adults and children with A1C 4-9.5% were randomized to sensor-augmented pump therapy (Medtronic Paradigm Revel) or multiple daily injections per day (28). Sensor-augmented pump therapy resulted in better A1C reduction with between-group difference of 0.6%, p<0.001.Hypoglycemia did not differ between groups, but only short-term CGM data were available for comparison and patients with a history of severe hypoglycemia were excluded.

STUDIES UTILIZING INJECTION THERAPY AS BACKGROUND

• In 2016, a 6-month randomized controlled trial, the DIAMOND study, compared RT-CGM (using Dexcom G4system) versus SMBG in 158 patients with type 1 diabetes on multi-dose injection therapy and demonstrated a significantly lower A1C (between group difference 0.6%, p<0.0001), decrease in hypoglycemia (43 minutes vs. 80 minutes per day, p=0.0002) and less glucose variability with RT-CGM compared to SMBG. This study did not address hypoglycemia frequency in the two groups (25).
• The GOLD trial studied 161 patients with type 1 diabetes receiving multiple daily injections with either RT-CGM (Dexcom G4) or standard care in a random order cross-over trial. The mean difference in A1C was 0.43% (p<0.001), favoring RT-CGM. One subject in the CGM group compared to 5 subjects in the standard care group experienced a severe hypoglycemic event. The percentage of time spent in hypoglycemia numerically favored the CGM group but statistical analyses were not presented. There was a significant reduction in standard deviationand MAGE (measures of glucose variability). Overall well-being, diabetes treatment satisfaction, and fear of hypoglycemia improved (29).
• In the FLASH-UK study, 156 participants with type 1 diabetes were randomized to intermittently scanned glucosemonitoring or usual care (30). The intervention group had a significantly greater reduction in HbA1 (adjusted treatment difference -0.5%, p<0.001), higher % TIR, and lower % TBR.
• A randomized controlled trial among 104 adults with type 1 diabetes found that intermittently scanned glucose monitoring improved A1c (estimated treatment difference 0.3% [95% CI, 0.0%-0.6%; P = 0.04) and TIR but not TBR compared to blood glucose monitoring (31).

META-ANALYSES

A Cochrane review and another meta-analysis found modest A1c reductions, particularly among patients who were not using insulin pumps, patients under age 18, and among patients with lower adherence (32). The results were heavily influenced by the STAR3 trial, and the JDRF study did not report a difference between pump users and patients using multiple dose injection therapy. Severe hypoglycemia rates did not differ. However, the quality of most studies was limited due to small sample size, lack of blinding, and lack of sufficient data to compare hypoglycemia rates. Meta-analyses may be hampered by the inclusion of studies with obsolete technology or lack of consideration for the intended use of the device in the study (33,34). In another meta-analysis, studies that specifically enrolled patients at risk for hypoglycemia and used blinded CGM to assess it did show improvement in hypoglycemia (35).

More recently, a meta-analysis of 21 studies published between 2011-2020 encompassing 2149 individuals with type 1 diabetes revealed that CGM led to a significant reduction in A1C by 0.23% (p=0.0005), with larger treatment effect at higher baseline A1C (>8%), and no effect on severe hypoglycemia or DKA (36). However, the meta-analysis did not report CGM derived metrics such as TIR or TBR or clinically significant hypoglycemia. In a 2023 meta-analysis of 22 randomized controlled trials that included participants with type 1 diabetes, there was an overall improvement in A1c,TIR and TBR (37). Reduction in A1C was limited to nonadjunctive devices but all devices resulted in improvement in TIR.

PATIENTS WITH HYPOGLYCEMIC UNAWARENESS

Many older studies specifically excluded patients with a history of severe hypoglycemia or were underpowered todetect significant hypoglycemia. Recent studies have examined the use of RT- CGM in patients with hypoglycemiaunawareness, which is a risk factor for severe hypoglycemia (events requiring outside assistance to treat).

• In the HypoCOMPaSS trial, 96 patients with a history of hypoglycemia unawareness determined by the GOLD Score of at least 4 or more were randomly assigned in a 2x2 factorial design to insulin pump or injection therapy, both with access to a bolus insulin calculator, and either RT-CGM (Medtronic Continuous Glucose Monitoring System) or SMBG. All patients had diabetes education with a goal toward hypoglycemia avoidance (38). Theresults demonstrated a similar reduction in severe hypoglycemia and improvement in hypoglycemia unawareness and fear of hypoglycemia without a significant treatment interaction between insulin or glucose monitoring interventions. Treatment satisfaction was higher with insulin pump compared to injection therapy but similar between RT-CGM and
• The IN CONTROL trial evaluated patients with Type 1 diabetes and hypoglycemia unawareness receiving either injection or insulin pump therapy in a crossover study comparing RT-CTM (Medtronic Paradigm Veo system with a MiniLink transmitter and an Enlite glucose sensor) or SMBG (39). Hypoglycemia was significantly reduced with RT-CGM compared to SMBG (including a 9.8% reduction in events <70 mg/dl and 44% reduction in events <40 mg/dl). Severe hypoglycemic events were significantly reduced but hypoglycemia unawareness was unchanged.
• In a smaller study of 52 adults with type 1 diabetes and problematic hypoglycemia, immediate randomization to CGM was more effective for preventing severe hypoglycemia (39% fewer events, p<0.05) than a dedicated hypoglycemia avoidance education program alone (40). CGM also lead to greater reduction in A1c (treatment difference -0.47%, p<0.05), but impaired awareness was restored in 31% of both groups, supporting the conceptthat CGM assists in earlier recognition and treatment of impending hypoglycemia as opposed to effecting fundamental change in counterregulatory responses.

Differences between studies may be explained by differences in populations and the technologies utilized. In the InCONTROL study, contact with patients was less frequent, sensor use was greater (89 vs. 57% in HypoCOMPaSS) and there were no insulin adjustment protocols. Therefore, more studies are needed to understand the potential role of background therapy, other technologies, and clinical support.

PATIENT REPORTED OUTCOMES

Generic Quality of life scores generally do not improve with RT-CGM but treatment-specific measures, such asdiabetes distress, hypoglycemic confidence, fear of hypoglycemia and to a lesser extent, measures of convenience, efficacy and performance, may be improved (28,41,42).

Evidence- Type 2 Diabetes

In patients with type 2 diabetes, even in patients not on insulin, RT-CGM may act as a motivator and positive influence for patients to improve lifestyle. The change in behavior can potentially lead to better glycemic control and weight loss(43). Moreover, periodic (every 3 months) short- term (14 day) use of real-time CGM may be sufficient to achieve andmaintain clinically relevant improvements in A1c in this population (44).

• In 2012, Vigersky et al. randomized 100 patients with type 2 diabetes on basal insulin and anti-hyperglycemic agents into either a group that used real-time RT-CGM intermittently (2 weeks on, 1 week off) or a group that recorded SMBG four times per day for 12 weeks. At 12 weeks, they found a statistically significantly greater reduction in A1c by 1.0% in the CGM group compared to 5% reduction in the SMBG group. The effect persistedup to the 40-week follow-up, 0.8% and 0.5% reduction in A1c in the RT-CGM versus SMBG group respectively (45).
• In 2017, Beck et al conducted a randomized study to evaluate benefit of RT-CGM use in 158 patients with type 2 diabetes with mean A1C of 8.5% treated using multiple daily injections (46). Over a 24-week period the A1C decreased to 7.7% in the RT-CGM group compared to 8% in the group with usual care (mean difference -0.3%,p=0.022). RT-CGM derived hypoglycemia and quality of life did not differ.
• The Dexcom MOBILE study assessed patients with type 2 diabetes on basal insulin randomly assigned to the Dexcom G6 or usual care for 8 months and reported a significant reduction in A1C, improved TIR andhypoglycemia (47). This was accomplished without an appreciable change in insulin or other medication use, indicating that CGM improves glucose levels by facilitating behavioral changes. Moreover, subsequent discontinuation of CGM for 6 months resulted in loss of about half of the improvement in TIR (48). Moreover, the benefit was similar in older (≥65 years old) vs. younger adults (49).
• In a 10-week study of 101 patients with type 2 diabetes on multiple daily injections of insulin, patients randomized flash glucose monitoring (Freestyle Libre) had greater A1C reduction (-0.82 vs -0.33%, p=0.005), found their treatment to be significantly more flexible and were more likely to recommend it to others (50).
• Among 141 adults with type 2 diabetes treated with insulin or sulfonylurea and recent myocardial infarction, thoserandomized to intermittently scanned glucose monitoring had significantly less TBR (-80 minutes, 95% CI -118, -43 minutes) at 90 days, but marginal difference in A1c or TIR, and the intervention was reported to be cost-effective (51).
• In a randomized trial of 116 adults with type 2 diabetes using non-insulin therapies, intermittently scanned glucose monitoring in combination with diabetes self-management education demonstrated superior A1c reduction at 16 weeks (treatment difference 0.3%, 95% CI 0 to 7%, p=0.048), larger increase in TIR (9.9%, p<0.01), and greatersatisfaction compared to education alone (52).

Real World Outcomes

• In a study of over 29,000 pediatric patients with type 1 diabetes in the Type 1 diabetes Exchange Registry or the German/Austrian DPV Initiative, pediatric CGM use was associated with lower mean A1C regardless of insulindelivery modality (pump or injection) (53).
• In a study of 106 UK hospitals incorporating 16,427 participants, 1241 with repeated TIR data, improvements inTIR were associated with improvement in hypoglycemia unawareness and diabetes related Moreover, TIR>70% was associated with reduced resource utilization (hospital admissions for hypoglycemia or hyperglycemia, paramedic visits, and severe hypoglycemia (54).
• In the Swedish National Diabetes Registry that included 14,372 adults with type 1 diabetes, intermittently scanned glucose monitoring was associated with a small (0.11%, p<0.0001) reduction in A1C after 15-24 months and reduction in severe hypoglycemic episodes (OR 0.79, 95% CI 0.69-0.91) (55).
• Using the French national claims database, a total of 74,011 patients with type 1 or type 2 diabetes initiatedintermittently scanned glucose monitoring and over 98% persisted with the device at 12 months (56). Following initiation of the device, patients had a 39-49% reduction in hospitalizations for acute complications and a 32-40%reduction in diabetes-related Moreover, the reduction in hospitalizations persisted after 2 years (57).
• In Belgium, a study of 1913 adults with type 1 diabetes were studied before and after nationwide reimbursement of intermittently scanned continuous glucose monitoring (58). Following the policy change, treatment satisfaction improved, there was a significant reduction in admissions for acute complications (severe hypoglycemia orketoacidosis), and there were fewer absences from work.
• Among 41,753 patients with insulin requiring diabetes in an integrated health care delivery system, 3806 patientsinitiated CGM, which was associated with a greater reduction in A1C (adjusted treatment difference 0.40%, p<0.001), emergency department or hospitalization for hypoglycemia (adjusted difference -2.7%, p=0.001), areduction in number of outpatient visits and an increase in telephone visits (59). However, there was no difference in hospitalizations for hyperglycemia or ketoacidosis.
• In a Medicare supplemental and commercial claims database study of 2463 patients with type 2 diabetes on multiple injections of insulin/day, intermittently scanned CGM was associated with a reduction in acute diabetes events (HR 0.39, 95% CI 0.30-0.51) and all cause hospitalizations (HR 68, 95% CI 0.59-0.78) at 6 monthscompared to the 6 months prior to initiation (60).

Recommendations

Patients should be adequately informed of the benefits and limitations of this technology, particularly with respect to the role for SMBG. At a minimum, structured education programs encompassing concepts such as carbohydrate counting and active insulin time (insulin on board) should be completed prior to considering RT-CGM, and payers mayrequire that patients demonstrate that they can reliably and consistently perform SMBG (61). Several expert groups have issued guidance in the use of RT-CGM.

• In 2016, the Endocrine Society, co-sponsored by The American Association for Clinical Chemistry, the American Association of Diabetes Educators, and the European Society of Endocrinology, published guidelines for use ofinsulin pumps and CGM. The guidelines recommended RT-CGM in adults with type 1 diabetes and any A1C who are willing and able to use the devices nearly daily. The panel suggested short-term intermittent use for patientswith type 2 diabetes (not requiring prandial insulin) who had an A1C ≥7% and are willing and able to use the device (3).
• The American Diabetes Association (ADA) Standards of Care recommend CGM in adults with type 1 diabetes andthose with hypoglycemia unawareness or frequent hypoglycemia (Table 2a). Among pediatric patients, the ADAnotes that CGM may reduce missed school days with regular usage (1).

A=Clear evidence from well-conducted, generalizable randomized controlled trials that are adequately powered;B=Supportive evidence from well-conducted cohort studies; C= Supportive evidence from poorly controlled or uncontrolled studies; E=expert consensus.

T1D=type 1 diabetes, T2D=type 2 diabetes, SMBG=self-monitored blood glucose

* Level of Evidence:

• High (1) = randomized controlled trial (RCT) or meta-analysis of RCT
• Intermediate (2) = meta-analysis including nonrandomized studies, network meta-analysis, nonrandomizedcontrolled trial, prospective cohort, case control, cross-sectional, hypothesis driven epidemiologic, open label extension, post-hoc analysis
• Weak (3) = discovery/exploratory, economic, consecutive case series, case report, safety/feasibility, high impact basic research
• None (4) = consensus, position, policy, guideline, any highly flawed study, lower impact basic science

BEL=best evidence level

^Grade is based upon evidence level, recommendation qualifiers, subjective factors, and consensus

* Level of Evidence:

• High (1) = randomized controlled trial (RCT) or meta-analysis of RCT
• Intermediate (2) = meta-analysis including nonrandomized studies, network meta-analysis, nonrandomizedcontrolled trial, prospective cohort, case control, cross-sectional, hypothesis driven epidemiologic, open label extension, post-hoc analysis
• Weak (3) = discovery/exploratory, economic, consecutive case series, case report, safety/feasibility, high impact basic research
• None (4) = consensus, position, policy, guideline, any highly flawed study, lower impact basic science

BEL=best evidence level

^Grade is based upon evidence level, recommendation qualifiers, subjective factors, and consensus

The 2021 American Association of Clinical Endocrinologists recommendations for use are summarized in Table 2band 2c. These include all adults and children with type 1 diabetes, especially those with severe hypoglycemia or hypoglycemia unawareness, and all patients with type 2 diabetes on multiple insulin injections, basal insulin, or sulfonylureas who are at risk for hypoglycemia (2).

In 2017, the Advanced Technologies & Treatments for Diabetes (ATTD) Congress organized an international consensus panel, consisting of physicians, researchers, and individuals with diabetes to analyze the existing literature and to provide guidance for utilizing, interpreting, and reporting CGM data (62). This was updated in 2019 (Table 3 and 4). These recommendations are supported by recent data from the DCCT demonstrating that a 10% reduction in time in target glucose range derived from 7-point self-monitored glucose profiles is associated with a 40% reduction in risk of microalbuminuria and a 64% reduction in risk of incident or progressive retinopathy (63).

*Includes time >250, **Includes time <54 mg/dl, ^#Insufficient data

Also, in 2017 the ADA and the European Association for the Study of Diabetes published a joint statement providing recommendations for systematic improvements in clinical use and regulatory handling of CGM devices (64).

Hospital Use

CGM is not currently approved for use in the hospital setting. However, during the COVID-19 pandemic, the FDA announced that it would not object to their use in an effort to support reduction in use of personal protective equipmentand risk of exposures to staff. Thus, there has been increasing interest in their use. Moreover, an increasing number of patients are using these devices in the ambulatory setting and want to continue their use in the hospital. Recent randomized trials support the use of CGM on the hospital wards, where it has been shown to be safe, and may reduce the frequency of hypoglycemia (65,66). In the ICU, there is concern that CGM may be less accurate due to factors such as edema, hypoperfusion, and acidosis but preliminary studies suggest use of CGM in conjunction with periodic point of care (POC) blood glucose (BG) within well-established protocols is safe and may reduce the need for POC BG (67).

In 2020, the Diabetes Technology Society sponsored a panel of experts in inpatient diabetes management to reviewthe evidence for us of CGM in the hospital (68). The panel agreed that CGM had the potential to improve clinicaloutcomes, particularly for patients who are unable to communicate signs or symptoms of hypoglycemia, but use is limited by lack of data demonstrating accuracy (particularly in the hypoglycemic range or in case of diabetic ketoacidosis, poor perfusion, or acetaminophen use) and clinical utility, and a lack of decision support systems, including infrastructure for communicating results to care teams and to the electronic medical record. The panelagreed that patients who are admitted with personal CGM devices should be allowed to continue use of such devices under the condition that they are able to self-manage the devices on their own and are followed by an endocrinologist or experienced practitioner who is specifically trained in their use. In particular, the panel advised implementing institutional policies that recommend continued capillary or blood glucose monitoring, ensuring that CGM data are not used for inpatient insulin dosing (since no CGM device is FDA approved in the inpatient setting), and requiring patients to sign safety waivers which illustrate the potential risks and benefits of continued use. Devices must beremoved for any MR or CT imaging. The panel made the specific recommendations for clinical care including:

• Consider use of CGM to reduce exposures (such as for point of care glucose) and need for personal protective equipment in persons with highly contagious diseases.
• Barring use in the setting of highly contagious disease, CGM values should be confirmed with point of care (POC) glucose prior to making treatment decisions.
• Hospitals should develop implementation plans which include a process map, protocol, provider/staff/patient education and order sets.
• Providers should recognize CGM pattern caused by compression of the device, which can cause a falsely low value.
• Providers should ensure patients are not taking medications or supplements that can interfere with CGM.
• Nurses should be adequately trained on use of CGM, inspect the insertion site every shift, and set expectations that POC values are still necessary to support ongoing use of CGM (typically every 6 hours).
• Hospitals need to develop security protocols, data storage, visualization tools, and integration within the electronic medical record to support the use of CGM.
• Hospitals need to identify CGM values in the electronic medical record to distinguish values from blood glucose values.
• Hospitals need to adopt the Unique Device Identifier (UDI) to track devices in the electronic medical record.

Limitations of Use

It should be emphasized that most prospective randomized controlled trials enroll highly motivated patients. In the real-world setting, there are concerns about limited resources for training, and less motivated patients may be overwhelmed with the additional data, particularly where complex algorithms are required. Nevertheless, in the Type 1 Diabetes Exchange Registry, CGM use increased from 7% in 2010-2012 to 30% in 2016-2018, and rose more than 10-fold in children (69). A1C levels were lower in CGM users compared to nonusers. While CGM use has improved substantially over time, more than half of respondents cited cost or insurance coverage as a significant barrier to use (70). Moreover, disparities in prescribing patterns and implicit bias have been described (71,72,73). Modifiablereasons for avoiding use include the hassle of devices (47%) and aversion to having a device attached to the body (35%). Skin reactions and/or difficulty with adhesion are well known and are an important cause of discontinuation (74). Methods of addressing this barrier such as use of barriers, overlay patches, or topical antihistamines and corticosteroids have been described but additional research is needed (75).

In a multi-national study of 263 patients, persistent sensor use for 12 months was only 30% (76). Improvement in A1C was associated with higher A1C at baseline, older age, and more frequent sensor use. Diabetes related hospital admissions were reduced following the initiation of sensor augmented pump therapy and fear of hypoglycemia improved. In the 6-month follow- up phase of the JDRF-CGM trial, RT-CGM was initiated in the control group in a manner that more closely approximates clinical practice (77). Investigators found a significant reduction in CGM use inall age groups over time. However, increasing sensor use was associated with A1C reduction. It is likely that adherence will improve as technologies improve.

Other limitations include possible interference with acetaminophen, ascorbate, and other active agents in glucose-oxidase based electrochemical sensors. They are also dependent on both the sensitivity and specificity on the enzyme availability on the electrode surface. There are well known delay artifacts due to the time lag between glucose concentration in the interstitial fluid and blood glucose. These time ranges, often between 5 and 10 minutes, are not crucial to analyzing retrospective data, but can be critical when CGM is indicated for real-time decision making (78).

Daily Use

Patients must be aware that sensor readings can deviate from actual blood glucose measurements, particularly during rapid glucose changes such as that which occurs post-meal or during exercise. Calibration, where necessary, should not be performed when trend arrows indicate rapid swings in glucose. While systems are becoming more reliable, patients may need to verify sensor readings before taking action such as meal boluses or treatment of hypoglycemia depending on the device, even if a device is approved for nonadjunctive use.

Alarm thresholds should be set in order to maximize patient compliance, keeping in mind that the sensitivity fordetecting hypoglycemia decreases as the threshold is reduced below 70 mg/dl. Conversely, specificity improves to amuch smaller degree at lower thresholds, and thus false alarms may not be reduced substantially.

Several algorithms have been published that provide specific guidance to patients for responding to trend arrows and alarms and are summarized below. All algorithms are complex and are not integrated within bolus calculators ofexisting insulin pumps. Therefore, they should only be implemented in patients who have demonstrated anunderstanding of CGM technology, including lag times between CGM and BGM, calibration procedures, alerts andtrend arrows, as well as understanding of insulin action time and the risks of insulin stacking. In one small study, trend arrows were accurate approximately 79% of the time outside of mealtime windows (30 minutes before and 120 min after carbohydrate intake) but this dropped to ~60% within mealtime windows (80). Thus, algorithms are not intended for use post-meal. The use of automated insulin delivery systems should increase safety and efficacy and reduce the complexity of the trend arrow approach.

• The algorithm by Jenkins et al. provides tiered recommendations that are based upon the meter glucose and sensor trend arrows (81). In addition, the algorithm advises patients how to review downloads of the data periodically (weekly) and make adjustments. Patients who were randomly assigned to sensor augmented pump with the algorithm had lower A1C and reported better quality of life at 16 weeks compared to patients who did not get the The effect on quality of life persisted at the 32-week follow-up, and was associated with A1Creduction. Importantly, patients who received the algorithm at 16 weeks after initiating sensor augmented pump did not benefit.
• The DirecNet study algorithm (for use with the Navigator system) recommended that patients increase or decreasethe meal + correction bolus by 10-20% based upon the rate of change and provided specific instructions for responding to alarms (82). Algorithm use was high in the first 3 weeks but dropped off by week 13, despite increasing insulin self- adjustments, possibly as patients became more independent over time.
• Subsequent methods recommended adjustment of only the correction insulin dose by the amount needed to cover a glucose level that is incrementally higher or lower than the current glucose, based upon the trend arrow (83,84).
• Klonoff and Kerr proposed a more straightforward correction dose (in 5-unit increments), based upon the trend arrow and the patient’s insulin sensitivity (85).
• A consensus statement facilitated by the Endocrine Society provides expert guidance on the use of trend arrows for making treatment decisions (86). The guidance recommends adjustment of boluses pre-meal and no sooner than 4 hours post-meal in 0.5-unit increments based upon the trend arrow and the patient’s sensitivity. The statement recommends no additional treatment within 2 hours of a previous meal bolus, and correction bolus using the bolus calculator or usual correction dose only in the 4 hours after a meal. Similar expert guidance has been developed for the Freestyle Libre system (87).
• A more recent adaptation of the Endocrine Society guidance incorporated pre-meal glucose levels in addition to theinsulin sensitivity (88) and a small randomized study demonstrated it was more effective than the incorporation of insulin sensitivity alone, particularly among insulin pump patients (89).

Overview of Stand-alone Personal RT-CGM and IS-CGM Systems

The first RT-CGM (Guardian, Medtronic^R) was approved in 2004. Since then, additional models and other devices have entered the market, and accuracy and patient satisfaction have improved. Several personal continuous glucosemonitors have been approved by the US Food and Drug Administration (FDA) for use in the United States or carry CE marking for use in Europe and are currently on the market (Table 5). For a full review of regulatory requirements for glucose monitoring devices, the reader is referred to one of several excellent reviews (90).

GUARDIAN CONNECT

The Guardian Connect utilizes the Medtronic Guardian Sensor 3, the Guardian Connect transmitter, and the GuardianConnect app to transmit data via Bluetooth every 5 minutes to the user’s smart phone or device (initially only available on iOS devices) via the Guardian Connect App on smartphones and via CareLink personal and professional software. A separate receiver is not available with this system. Data can be shared with others remotely, and SMS messages can be sent in times of hypoglycemia. The system is only approved for adjunctive use and at least 2 daily fingerstick calibrations are required.

DEXCOM G6

The Dexcom CGM utilizes a glucose oxidase sensor at the tip of a wire that is implanted in the subcutaneous space. The data are transmitted wirelessly and are displayed on a separate receiver (personal smartphone or device specificreceiver). The Dexcom G6 has a sensor life of 10 days, no longer requires calibrations, and minimizes interference by acetaminophen. G6 is also associated with a smartphone app that allows the patient to log activity, set reminders or alarms, and physically see their glucose levels and trends throughout their time wearing the device. The Dexcom CLARITY Diabetes Management Software organizes and presents the patient’s blood glucose data. Dexcom SHARE app allows users to share data with up to 10 other individuals.

DEXCOM G7

The Dexcom G7 features a 60% smaller size (the size of 3 stacked quarters) vs. the G6, is fully disposable, and has a shorter 30-minute warm up time and a 12-hour grace period to replace completed sensors. The overall MARD was reported to be 8.2% with the abdomen and 9.1% on the arm (17). Consistent with previous studies, accuracy is lower at lower sensor glucose, higher glucose rate of change, and on day one of wear.

FREESTYLE LIBRE 14 Day

The sensor utilizes Wired Enzyme™ technology in which the enzyme and mediator are co- immobilized on the sensor. It offers factory calibration, and therefore nearly eliminates the need for fingerstick monitoring. However, patients are still advised to perform SMBG whenever an alert appears on the reader display (which occurs when the glucose isrising or falling rapidly) or whenever the glucose value does not fit the patient’s symptoms. The reader contains a built-in meter for this purpose. The sensor is FDA approved for 14 days of use. The system is not approved for use in children under age 18, or during pregnancy or in persons requiring hemodialysis. The Libre has minimized the interference by acetaminophen which is present in other devices but interference from other substances such as ascorbic acid or aspirin may be possible. The Libre 14 day differs from other CGM devices in that the system does not alert the user for glucose values surpassing a high or low threshold. In addition, glucose values are not automatically made available to the user but are easily and instantly accessed by scanning the sensor with a handheld reader or theassociated app FreeStyle LibreLink. However, this product may be attractive option for patients who are averse to the hassle imposed by other RT-CGM devices. Glucoses are measured every minute and recorded every 15 minutes. Data can be accessed using the reader or downloaded to LibreView cloud based online management system, or using the FreeStyle Libre desktop software. The MARD is reported by the manufacturer to be 9.7% overall, and as with other CGM devices, less accurate on day 1 of wear and in hypoglycemia range (91).

FREESTYLE LIBRE 2

The FreeStyle Libre 2 system offers real-time alerts for high or low glucose values and improved accuracy, approved for ages 4 years and older (92). However, users must continue to scan the device to obtain glucose readings.Moreover, similar to the Libre 14 day, the sensor memory is only 8 hours and glucose data are lost if the sensor is scanned less frequently.

FREESTYLE LIBRE 3

The FreeStyle Libre 3 is even smaller than other devices (the size of 2 stacked pennies), does not require scanningunlike older models, but does require the use of a compatible smartphone. The bluetooth range is improved from 20 to 33 feet. Accuracy is improved compared to the Libre 2, with an overall MARD of 9.2% in adults and 9.7% in children (16).

EVERSENSE

The Eversense system (Senseonics) is a 90-day implantable sensor that uses fluorescent technology to sendmeasures via a rechargeable transmitter which rests just above the skin to a smartphone app titled Eversense NOW (93). In a pivotal clinical study of 71 patients with type 1 and type 2 diabetes, there were no device-related serious adverse events, and the MARD was 11.1%, with over 99% of samples in clinically acceptable error zones A and B of Clarke Error Grid Analysis (94). One study reported interference with tetracycline and mannitol, but not with acetaminophen or ascorbic acid (95). The Eversense XL CGM system consists of a 180-day implantable sensor thathas been shown to have acceptable safety and accuracy with an overall MARD of 9.1% (96). This option may be particularly useful for patients with privacy concerns, physical disability, needle phobia, allergies or other difficulty with adhesion, or activities or professions that may be barriers to external wear (97).

Sensor Augmented Pumps

To date the largest A1C reductions have been observed when sensors are initiated with insulin pump technology. In the observational (nonrandomized) COMISAIR study, patients initiating CGM (with or without insulin pump) achieved significantly larger reductions in A1C (-1.2%) compared to subjects initiating insulin pump alone (-0.6%) or remaining on injections alone (- 0.3%) (98). There was no difference in outcomes between the DexCom G4 and Enlite sensor.

A reduction in time spent in hypoglycemia was observed only in patients using CGM (8% vs 6%, p<0.001).

STEPS TOWARDS AN ARTIFICIAL PANCREAS

Until recently, RT-CGM technology has operated completely independently of insulin delivery. By combining continuous basal insulin delivery during fasting periods with discrete bolus doses of insulin at mealtimes, insulindelivery can be crafted to mimic the natural pattern of pancreatic insulin release. An artificial pancreas consists of: 1) an automatic and continuous glucose monitor; 2) an implanted continuous insulin delivery system; 3) a control processor to link the insulin delivery rate to the glucose level; and 4) a signal to send the glucose level to the body surface for continuous display onto a monitor. Limitations to full implementation include sensor accuracy and lag time, inadequate onset and offset of currently available rapid acting insulin analogs, meal challenges, and changes in insulinsensitivity due to circadian rhythms, exercise, menstrual cycles, and intercurrent illness (99). However, even incremental advances improve glucose control without increasing the complexity of decision-making on the part of the patient. These include:

• Low glucose (threshold) suspend: the insulin pump suspends when the glucose decreases below a pre-set value.
• Suspend before low: insulin pump suspends when hypoglycemia is
• Hybrid closed loop: insulin delivery increases or decreases based upon the sensor glucose value but meal boluses are still required.
• Closed loop control: fully closed loop delivery without the need for meal boluses
• Dual hormone systems: these are hybrid closed loop or closed loop control systems that utilize glucagon or other peptides (such as amylin) in an effort to more closely mimic the physiology of the endocrine pancreas.

The long-term safety, efficacy, cost, and cost-effectiveness of an artificial pancreas are still largely unknown at this time. However, the urgency of this technology is demonstrated by the #WeAreNotWaiting movement, which has given rise to home-grown, crowd-sourced, patient driven systems that utilize existing devices which are linked by open-source software, such as Open Artificial Pancreas System, and Loop. Recently Tidepool Loop received FDA approval (100). A retrospective observational study of patients with Type 1 diabetes demonstrated lower mean glucose, higher time in target range, and less time in hypoglycemia using Open Artificial Pancreas Systems (OpenAPS) compared to sensor augmented pump use alone (101). In general, open-source systems carry safety concerns, particularly among less tech-savvy patients, in the absence of regulatory approval (102). However, the healthcare provider can providesafety recommendations as well as a back-up plan in case of system failure (103). The reader is referred to one of several reviews as a detailed review is beyond the scope of this chapter (104,105).

Threshold Suspend

Progress is expected toward a fully functional closed loop system in incremental steps. The first step toward a fully automated “artificial pancreas” is the low glucose suspend feature, which is now available. The Medtronic 530Gsystem, containing the Veo insulin pump and Enlite sensor, is the first sensor augmented pump with low threshold suspend and uses the same sensor as the more recent 630G system. The Enlite sensor accuracy is significantly improved over the previous Sof-sensor^R, with a MARD of 13.6% when used with the 530G (106). The Enlite is also one-third of the size of Sof-sensor and the filament is 38% shorter. The Enlite sensor may be worn up to six days. The low threshold suspend SmartGard^TM technology suspends the pump for up to two hours in the event of sensor detected hypoglycemia in which the user does not respond to the alarm. Prior to suspension, a “siren” sounds which is distinct from other high or low alerts, and the suspension can be overridden at any time. The MiniMed Connect mobile accessory sends sensor data to an app on a mobile device where data can be viewed (available only with the 530G system). A study that enrolled 247 patients with type 1 diabetes and documented nocturnal hypoglycemia to sensor-augmented pump with or without a low-glucose threshold-suspend feature demonstrated similar A1C between groups at 3 months but lower frequency of nocturnal hypoglycemia (107). Similar findings were demonstrated in an Australian study of 95 patients, in which the incidence rate ratio for hypoglycemia was 3.6 (95% CI 1.7-7.5, p<0.001) (108). There were no reports of DKA in either study.

Suspend Before Low

The next incremental step in closed loop systems is the suspend before low feature, currently available in theMedtronic 640G (approved only in Europe) and the 670G systems. This feature automatically suspends insulin delivery 30 minutes before a low glucose threshold is predicted and resumes delivery once the glucose recovers, without alerting the patient. In a 6-month randomized study of 154 children and adolescents with type 1 diabetes, the 640G system reduced the time spent in hypoglycemia from 2.6 to 1.5% without causing a change in A1C (109). The t:slim X2 Insulin Pump incorporates Basal IQ technology with predictive low glucose suspend using the Dexcom G5 or G6 sensor. In a randomized cross-over study of 103 participants with type 1 diabetes age 6-72 years of age, predictivelow glucose suspend resulted in a 31% reduction in time spent in hypoglycemia < 70 mg/dl without a change in meanglucose or time in hyperglycemia (110).

Hybrid Closed Loop (HCL)

This step refers to sensor glucose driven automatic adjustment of basal insulin with or without additional auto boluses,and still requires the patient to bolus for meals. In a recent consensus statement, an ideal candidate for automated insulin delivery systems (103):

• Is technically capable of managing a pump, has basic carbohydrate counting skills, and is able to implement a back-up plan (including the use of manual injections).
• Has realistic expectations of system In particular, several situations that are unique to HCL are worth emphasizing:
  • Bolusing: pre-blousing approximately 15 minutes prior to meals is critical to maintain In many systems,delayed boluses not only cause early postmeal hyperglycemia but also precipitate delayed hypoglycemia as the system has already begun to augment insulin delivery in response to hyperglycemia.
  • Exercise management: Similarly, carbohydrate loading prior to exercise while using HCL systems will only stimulate insulin delivery and thus is recommended that users implement other means for management such as setting a higher target, typically with a designated exercise mode, or exiting to manual mode with temporarybasal insulin reduction or temporary suspension of the pump.
  • Hypoglycemia management: HCL users typically need fewer carbohydrates (about half) to manage hypoglycemia since the pump has generally already suspended insulin delivery based upon glucose trends.
• Has adequate support, including diabetes education, insurance coverage, and caregiver or other social support where relevant.
• Has the ability to transmit data to the healthcare
• Is mentally and psychologically able to implement AID

On the other hand, there does not appear to be an ideal threshold A1c for determining candidates for HCL therapy, asthose with lower A1c may benefit by reducing TBR and those with higher A1c benefit from reductions in hyperglycemia without the perceived risk of ketoacidosis traditionally attributed to initiation of insulin pump therapy (111). Thus, less ideal candidates may obtain the greatest benefit in terms of achieving glycemic targets.

HCL demonstrates improvements in a range of glycemic outcomes and may confer psychological benefits as well.Most studies have enrolled patients with type 1 diabetes. More data are needed for special populations including type 2 diabetes, especially those with very high insulin requirements, pregnancy, acute illness, steroid use, renaldysfunction, and persons in assisted living facilities (112).

As with CGM, it is important to evaluate these systems in the real-world setting, where user experience can differ from that of the highly controlled and supportive research environment. Cost and insurance hassles, as well as user wear issues are the most commonly reported barriers to use of any diabetes related device. These barriers contribute todiabetes distress and depressive symptoms which can impede self-management behaviors (70,112). HCL systems improve glucose control but may also introduce additional alarms or alerts which are needed for safety (such as threshold alerts for hyperglycemia or hypoglycemia or HCL mode exits due to insulin delivery exceeding the system’s guardrail) or ongoing functionality such as calibration of the CGM (113). Newer systems have attempted to address many of these barriers through improved algorithms or other features (Table 6). Devices differ considerably with respect to algorithms used for insulin adjustment and a number of other features (Table 6) (114,115,116). There are few head-to-head studies comparing the efficacy and safety of available HCL systems. Details of select systems are presented next.

MEDTRONIC 670G

The first system to gain FDA approval is the Medtronic 670G, which adjusts basal insulin delivery every 5 minuteswhen in auto mode. This system utilizes the Guardian sensor 3, which offers enhanced sensor accuracy, with an overall MARD of 9.64% (117). The system was associated with a reduction in A1C from 7.4 to 6.9% and there were trends in improvement of time in target range and hypoglycemia in a non-randomized study of 124 patients with type 1 diabetes (118). A subsequent randomized trial of 151 adults and children demonstrated a significant reduction in A1c and TBR compared to insulin pump without CGM (119). There are few studies addressing long-term use however. In a 1-year prospective observational study of 84 patients, 28% stopped using auto mode by 3 months, and 33% discontinued by 12 months (120). The most common reasons for discontinuation included sensor issues (62%) and difficulty obtaining supplies (12%), fear of hypoglycemia (12%), and preference for injections (8%) or sports (8%). In astudy of 92 youth, 30% discontinued HCL, typically between 3 and 6 months after initiation, due to issues such asdifficulty with calibrations, alarms, and extra time needed for operation (121).

MEDTRONIC ADVANCED HYBRID CLOSED LOOP (AHCL) SYSTEM (780G)

This HCL system is approved for use in Europe and features substantially reduce frequency of alerts, improved time in auto mode, remote software updating, an adjustable target setting as low as 100 mg/dl and enable users to view data via an app on mobile devices (122). In a single arm study of 157 adolescents and adults the 780G system resulted in nearly 95% time in automated mode with 1.2 exits per week, improved A1C, TIR, and TBR (123). In a randomized study of 82 persons with type 1 diabetes using multiple injections per day and isCGM, AHCL resulted in improvements in A1C (-1·42%, 95% CI -1·74 to -1·10; p<0·0001) and TIR, but no difference in hypoglycemia (124). There was an improvement in treatment satisfaction, fear of hypoglycemia, and similar diabetes quality of life. By comparison in a randomized study of 41 participants with type 1 diabetes who were naïve to both CGM or insulin pump technologies, AHCL also resulted in improvements in TBR (125). In a real-world study of 3211 youth (<age 15 years) and 8874 individuals >age 15 years, ACHL demonstrated >90% treatment persistence over 6 months (126). The ACHL system was reported to have better glucose monitoring treatment satisfaction but similar diabetes distress, technology attitudes, and fear of hypoglycemia compared to the 670G system (127).

TANDOM CONTROL-IQ

This system utilizes a t:slim X2 insulin pump with a calibration-free Dexcom G6 sensor (128). In a 6-month trial of 168 patients (age 14-71) randomized 2:1 to hybrid closed loop vs. sensor augmented pump alone, the % time in target 70-180 mg/dl was increased by 11% more in the hybrid closed loop group compared to sensor augmented pump(p<0.0001), with improvements in hypoglycemia, mean glucose and A1c. Moreover, real-world outcomes among 1435 persons with type 1 diabetes included reduced impact of diabetes on life, improved device related treatment satisfaction, and improved emotional well-being (129).

OMNIPOD 5

The Omnipod 5 is a HCL system that uses the Omnipod DASH platform (130). In a single arm study, the Omnipod 5 demonstrated a reduction in A1C of 0.38%, increase in TIR of 9.3% and decrease in TBR of 1.6% (131).