The effectiveness of footwear and offloading interventions to prevent and heal foot ulcers and reduce plantar pressure in diabetes: a systematic review

Authors


Abstract

Background

Footwear and offloading techniques are commonly used in clinical practice for the prevention and treatment of foot ulcers in diabetes, but the evidence base to support this use is not well known. The goal of this review was to systematically assess the literature and to determine the available evidence on the use of footwear and offloading interventions for ulcer prevention, ulcer treatment, and plantar pressure reduction in the diabetic foot.

Methods

A search was made for reports on the effectiveness of footwear and offloading interventions in preventing or healing foot ulcers or reducing plantar foot pressure in diabetic patients published prior to May 2006. Both controlled and uncontrolled studies were included. Assessment of the methodological quality of studies and data extraction was independently performed by two reviewers. Interventions were assigned into four subcategories: casting, footwear, surgical offloading and other offloading techniques.

Results

Of 1651 articles identified in the baseline search, 21 controlled studies were selected for grading following full text review. Another 108 uncontrolled and cross-sectional studies were examined. The evidence to support the use of footwear and surgical interventions for the prevention of ulceration is meagre. Evidence was found to support the use of total contact casts and other non-removable modalities for treatment of neuropathic plantar ulcers. More studies are needed to support the use of surgical offloading techniques for ulcer healing. Plantar pressure reduction can be achieved by several modalities including casts, walkers, and therapeutic footwear, but the diversity in methods and materials used limits the comparison of study results.

Conclusions

This systematic review provides support for the use of non-removable devices for healing plantar foot ulcers. Furthermore, more high-quality studies are urgently needed to confirm the promising effects found in both controlled and uncontrolled studies of footwear and offloading interventions designed to prevent ulcers, heal ulcers, or reduce plantar pressure. Copyright © 2008 John Wiley & Sons, Ltd.

Introduction

There is a long clinical tradition in the use of footwear and offloading techniques for the prevention and healing of plantar foot ulcers in the diabetic patient. However, the research literature in this area has lagged behind clinical practice, often leading the clinician to recommend interventions based on opinion and past experience rather than published evidence. A recent systematic review pointed out, for example, that evidence for the use of therapeutic footwear to prevent the recurrence of plantar ulcers was not yet strong 1. The purpose of the present review was to systematically assess the refereed literature and to determine the available evidence on footwear and offloading interventions, which are used in clinical practice to prevent and treat ulceration or to reduce plantar pressure in the diabetic foot.

Materials and methods

The population of interest for this systematic review was patients with type 1 or type 2 diabetes mellitus, and the clinical problem addressed was plantar foot ulceration. Main outcomes were ulcer prevention, ulcer healing, and the reduction of plantar foot pressure. The interventions considered were four groups of techniques commonly used throughout the world in clinical practice:

  • 1.Casting techniques: (a) total contact cast (TCC) and (b) cast shoes
  • 2.Footwear related techniques: (a) shoes, (b) insoles, (c) in-shoe orthoses, and (d) socks
  • 3.Surgical offloading techniques: (a) Achilles tendon lengthening (ATL), (b) silicone injections/tissue augmentation, (c) callus debridement, (d) metatarsal head resection, (e) osteotomy/arthroplasty/ostec- tomy/exostectomy, and (f) external fixation.
  • 4.Other offloading techniques: (a) bed rest, (b) crutches/ canes/wheelchairs, (c) bracing (patella tendon bearing, ankle-foot orthoses), (d) walkers, (e) offloading dressings, (f) felted foam/padding, and (g) plugs.

Studies on healthy subjects or patients with other diseases were not considered. The search, performed on 1 May 2006, covered references in all languages and was not limited by date. Only original research qualified for inclusion. Study designs that were included were randomized controlled trials (RCTs), case-control studies, cohort studies, control before-and-after (CBA) studies, interrupted time series (ITS), prospective and retrospective uncontrolled studies, cross-sectional studies, case series, and case studies. Tracking of references in included articles was not performed. Search strategies, shown in Appendix A, which included search terms on patient group, study design, outcome, and intervention, were defined using the following databases: MEDLINE, Excerpta Medica Database (EMBASE), Cumulative Index to Nursing & Allied Health Literature (CINAHL)®, Cochrane database of systematic reviews, Cochrane central controlled trials register, Database of Abstracts of Reviews of Effects (DARE), National Health Service Economic Evaluation Database (EED), and Health Technology Assessment (HTA) Database.

In each of the four intervention groups, one reviewer assessed references by title and abstract to determine possible eligibility on three criteria: patient group, outcome, and intervention. Full-paper copies of identified articles were retrieved, and each was assessed for eligibility by two independent reviewers on four criteria: patient group, study design, outcome, and intervention. All reviewers were authors of this systematic review. Findings were discussed between co-reviewers, and a final decision regarding inclusion and exclusion was made. Each included article describing a controlled study with a RCT, CBA, cohort, case-control, or ITS study design was assessed for methodological quality by two independent reviewers using standard scoring lists specific for each study design based on scoring lists developed by the Dutch Cochrane Centre (www.cochrane.nl). Reviewers did not participate in any discussion of articles of which they were authors at any stage of the process to prevent conflict of interest. Findings were discussed between co-reviewers and a final decision was made on methodological quality. Equal weighting was applied to each validity criterion. Quality items were rated as ‘done’, ‘not done’, or ‘not reported’ and only those rated as ‘done’ contributed to the total quality score. Evidence grading of each article was based on study design and the total score for methodological quality using the Scottish Intercollegiate Guidelines Network (SIGN) instrument: level 1 for RCTs and level 2 for case-control, cohort, CBA or ITS studies. Possibility for bias was scored for each study as + + (high quality with low risk of bias), + (well conducted with low risk of bias), or − (low quality with high risk of bias). Extracted data were summarized in evidence tables and described on a study-by-study narrative basis. Evidence tables were thoroughly discussed by the working group. The included studies at level 3 evidence, i.e., prospective and retrospective uncontrolled studies, cross-sectional studies, case series, and case studies, were also assessed and summarized but not included in the evidence tables. Finally, conclusions were drawn for each subcategory of interventions based on the strength of the available evidence.

Results

A total of 1651 articles were identified in the database search, and 212 of these articles were considered potentially eligible after review of the title and abstract. Twenty-one level 1 and level 2 studies were selected for methodological assessment following full text review and 108 level 3 studies were selected for additional description. Detailed results and study ratings of the 21 controlled studies are contained in an evidence table shown in Appendix B.

Ulcer prevention

Footwear

No eligible studies related to the primary prevention of ulceration (i.e. first ulcer) in diabetes were found. Several studies have assessed the effects of footwear on ulcer recurrence (i.e. secondary prevention). One large RCT found no significant difference in the number of ulcer occurrences or the proportion of people with ulcers over a 2-year period between therapeutic shoes and control shoes 2. However, the methodological quality of this study has been challenged 3, 4. A small RCT found a significant difference in the proportion of people with ulcers over a 1-year period between two groups, one in which subjects had worn therapeutic shoes and the other in which subjects continued to use their own shoes 5. Plantar pressures and shoe use (treatment adherence) were not measured in either of these studies. Another large RCT investigated the recommendation for footwear within the context of a multi-faceted educational intervention 6. Owing to methodological flaws, no relevant conclusions could be drawn from this study. A CBA study compared clinical treatment and education including therapeutic footwear with treatment at a different clinic that did not include therapeutic footwear. A significant difference in ulcer occurrence in favour of the therapeutic footwear intervention was found in 2 years 7. A cohort study found that the use of therapeutic sandals resulted in a significantly lower ulcer occurrence at 9 months compared to a control group wearing sandals with a hard leather board insole. However, selection bias could not be excluded in this study 8. Finally, one prospective uncontrolled study provided some support that footwear can be effective in ulcer prevention 9.

Surgical offloading

One RCT 10 compared treatment with ATL in addition to TCC with TCC alone in patients affected by forefoot neuropathic plantar ulceration. The percentage of plantar ulcer recurrences was significantly lower in the ATL group than in the TCC alone group at follow-up periods of 7 months (15% vs 59%, respectively) and 2 years (38% vs 81%, respectively). Another RCT studied the combined effect of surgical excision, debridement, removal of bone segments underlying the lesion, and surgical closure compared with conservative offloading treatment on plantar forefoot ulcer recurrence 11. The results showed a significantly lower ulcer recurrence rate after 6 months in the surgical group. Two cohort studies by the same research group showed positive effects of metatarso-phalageal joint arthroplasty of the great toe and metatarsal head resection on ulcer recurrence rate at 6 months (5%) when compared to TCC treatment (28%–35%) 12, 13.

Casting and other offloading techniques

Casting techniques were not considered in the context of ulcer prevention. For offloading interventions other than footwear or surgical offloading, no relevant studies were identified.

Ulcer treatment

Casting

Several level 1 studies have assessed the effect of TCCs on healing of plantar foot ulcers. In one RCT, the TCC was shown to heal a higher percentage of plantar ulcers at a faster rate when compared with standard treatment (i.e. accommodative footwear) 14. In another RCT, treatment with a windowed TCC showed significantly higher healing proportions than treatment with therapeutic footwear after a short follow-up time of 30 days 15. In a medium-sized RCT of several commonly used devices for ulcer treatment, the TCC was found to be superior to a removable walker and a half-shoe in the proportion of healed ulcers in 12 weeks and in time to healing 16. However, data for the removable walker and half-shoe were pooled for the primary outcome, which limits the conclusions that can be drawn. In this study, patients wearing TCCs were much less active than patients wearing half-shoes. A low-quality CBA study with undefined study duration found healing proportions of 81% and 70% for TCC and half-shoes, respectively (significance not reported) 17. Adherence to treatment was much larger in the TCC (98%) than in the half–shoe group (10%), which is surprising in view of the small difference in healing proportion. In a low-quality cohort study, healing proportions of neuropathic ulcers were 83% in a windowed TCC and 25% in two types of offloading shoes 18. The other evidence largely describes clinical outcomes from predominantly prospective and retrospective epidemiological studies of various offloading techniques 19–38. Between 73% and 100% of wounds were reported to have healed in studies of TCCs, with healing times ranging between 30 and 63 days 23, 29, 39. Fiberglass cast shoes were found to heal 91% of subjects in an average of 34 days 32. Studies of Scotch-cast boots, windowed casts, and custom splints have reported between 70% and 81% of subjects healed between 42 and 300 days 23, 40, 41.

Footwear

In controlled studies, the use of footwear in the context of ulcer healing is most often reported as a comparative control condition for other interventions such as casting. The results of these studies were discussed in the previous sub-chapter. Within the limitations of their designs, the results from level 3 studies suggest that molded and ethylene vinyl acetate (EVA) boots can play a role in ulcer healing 42, 43, that various types of shoes and offloading insoles can, under some circumstances, result in healing 44–47, and that half-shoes can be effective in the treatment of plantar ulcers 23, 48, 49.

Surgical offloading

Many of the studies on offloading of neuropathic plantar ulcers through surgical intervention have been discussed above in the context of ulcer prevention. One RCT comparing ATL in addition to TCC with TCC alone in patients with forefoot ulceration found 100% healing in 41 days and 88% healing in 58 days in these two study groups, respectively (not significant) 10. Another RCT showed higher healing percentages through a combination of surgical excision, debridement, removal of bone segments underlying the lesion, and surgical closure when compared with conservative offloading treatment: 95% vs 79% (p < 0.05), although there is some question regarding the statistical analysis used 11. Time to healing was significantly shorter in the intervention group (47 vs 129 days (p < 0.001)). Two level 2 studies on the efficacy and safety of metatarsal-phalangeal joint arthroplasty in treating plantar sub-interphalangeal joint ulcers were identified. A case-control study 36 and a cohort study 12 of arthroplasty in addition to TCC both showed no change in proportion of ulcers healed compared to TCC alone, but a significantly faster time to healing in the arthroplasty group. In cases of recalcitrant plantar ulcers, fifth metatarsal head resection was found to be as effective as offloading treatment (both 100% healed in 8.7 weeks), with shorter healing times in the surgical group (maximum 5.8 vs 8.7 weeks) 13. The majority of uncontrolled studies focused on descriptions of surgical procedures and case reports/series of outcomes 50–79. Carefully selected surgical procedures such as arthroplasties of affected joints with limited joint mobility in the forefoot may have value in promoting wound healing 51. Furthermore, exostectomy procedures have been reported to be relatively reliable methods of treatment for rigid, prominent deformities secondary to Charcot's neuro-osteoarthropathy 58, 59.

Other offloading interventions

Studies on the use of bed rest, crutches, canes, wheelchairs, offloading dressings, and plugs on ulcer healing were not located in the literature search. Only the effects of prefabricated removable walkers, bracing, and felted foam have been studied using controlled designs. One RCT on removable walkers 16 has been discussed above in the section on ‘casting’. In two other RCTs the removable walker was made non-removable, and this approach resulted in significantly increased proportions of healed ulcers compared with the standard removable walker 80 and similar healing rates compared with the TCC 81. A methodologically flawed RCT examined the effect of felted foam in post-operative shoes and found significantly shorter time to heal compared to a pressure relief half-shoe 82. Healing proportions were not reported. The use of a walking brace (plantar healing orthosis) in a small CBA study of poor quality did not lead to higher healing proportions than topical ulcer care, but the study may have been underpowered 83. In a retrospective analysis, accommodative dressing (felted foam) worn in a surgical shoe, a healing shoe, or a walking splint was found to be as effective as a TCC in both healing proportion and time to healing 23. Finally, two case studies of patients with recalcitrant plantar ulceration showed that continued treatment with a Cam walker with plastazote arch filler 84 or an ankle-foot orthosis 85 resulted in healing within 4 months and 12 weeks, respectively.

Plantar pressure reduction

Casting

No controlled studies on the effectiveness of casting devices in reducing plantar pressure in the diabetic foot were found. However, five uncontrolled studies on this topic were identified 86–90. TCCs appear to offload the affected extremity effectively. An unreplicated study suggested no substantial difference between standard fracture casts and TCCs in the ability to offload 90.

Footwear

There are no RCTs that report on the effect of footwear in terms of plantar pressure reduction. One cohort study reported reductions in plantar load of 10%–19% in therapeutic sandals compared with sandals with a hard leather board insole, but there is doubt about the accuracy of the pressure results in this study 8. A total of 25 predominantly cross-sectional studies that report on plantar pressure reduction for various footwear interventions were included in this review. These include a variety of prefabricated and custom-made therapeutic footwear designs, hosiery (socks) and running shoes. Shoes with a rocker-bottom outsole were found to be effective in reducing forefoot peak pressures 91. Footwear with some sort of molded insole also provided significant reductions in plantar pressure 88, 89, 92–102. Shoes with flat insoles were less effective 99. The results for (padded) hosiery are not conclusive.

Surgical offloading

Surgical approaches to reduce plantar pressure have not been widely investigated. One RCT compared the pressure–reducing effect of regular liquid silicone injections and saline injections under callused metatarsal heads and found significantly reduced peak plantar pressures in the silicon group at 12 months, but not at 24 months follow-up 103. In a subset analysis of a larger RCT 10, peak plantar pressures were evaluated in patients subjected to ATL + TCC treatment or to TCC alone 104. The study showed significant reductions of forefoot plantar pressure post-ATL compared to pre-ATL, which was not sustained at 8 months follow-up. Several uncontrolled studies suggest that ATL and metatarsal head resections reduce pressure in the forefoot 53, 105. Metatarsal head resections, however, may increase plantar pressure in other areas of the foot 106, 107. Furthermore, regular callus removal has been reported to have a beneficial effect on the reduction of plantar pressure 108, 109.

Other offloading interventions

Several well-conducted cross-sectional studies have investigated the effect of a variety of offloading modalities on plantar pressures at sites of previous ulceration. Removable walkers were shown to be as effective as TCCs and more effective than forefoot offloading shoes and extra–depth shoes in reducing forefoot peak pressures 87, 88. Forefoot offloading shoes were found to be more effective than accommodative felt and foam dressings (worn in a post-operative shoe) or post-operative shoes alone 87. The heel region was found to be best off-loaded with a TCC, followed by removable walkers, which, in turn, were more effective than depth-inlay shoes 86. Polyurethane foam sheets applied at forefoot ulcer locations and tested inside the patients' own shoes showed a significant reduction in peak pressures compared with wearing standard shoes 110.

Discussion

This systematic review included all levels of study designs, ranging from RCTs to case reports, since it was known at the outset of this project that evidence from controlled prospective studies would be scarce. Only 21 prospective controlled studies were finally selected for review, together with 108 uncontrolled and cross-sectional studies, which were additionally described. Most prospective controlled studies have been published on plantar ulcer treatment, reflecting the central role that offloading plays in this context. Furthermore, although the goal of most footwear and offloading devices is to reduce pressure at ulcer sites or at-risk locations, it is surprising to see how many studies actually fail to measure plantar pressure to determine efficacy of the intervention used. Also the measurement of adherence to footwear or offloading treatment and daily activity levels, which most authorities consider important in the development of foot ulcers, is missing from most publications. As the grading in the evidence table (Appendix B) shows, the methodological quality of most studies was generally rather poor, with only six of 21 publications being ‘well conducted’ studies, and none being of ‘high quality’. Therefore clinicians should view the results from these studies with some caution.

Ulcer prevention

The literature shows that ulcer recurrence remains a significant problem in diabetes. Various authors have reported between 27% and 100% recurrence rates at 1 and 4 years, respectively 5, 49. Although these numbers clearly indicate that research needs to emphasize the prevention of ulcer recurrence, very few studies of sufficient methodological quality are available in this regard. No experimental studies exist on the role of footwear and offloading in primary ulcer prevention. There are indications that therapeutic shoes may be effective in secondary prevention compared to standard footwear, although one RCT has found no effect. These contrasting results are likely to be the result of the wide diversity of interventions and control conditions investigated in these studies as well as the lack of knowledge of unloading efficacy, which is rarely measured in advance of prescription. This lack of standardization complicates the comparison of studies in the literature and prevents the articulation of useful directions for the optimum design of footwear to prevent ulcer recurrence. The development and documentation of standard procedures in the future is therefore highly recommended.

Lengthening of the Achilles tendon, surgical excision of bony prominences, joint arthroplasty, and metatarsal head resection all appear to reduce the risk of ulcer recurrence in selected diabetic patients with forefoot neuropathic plantar ulceration when compared to conservative offloading treatment. However, the disadvantages and potential complications of surgical interventions should always be taken into consideration. For ATL, for example, the influence on locomotion and other functional tasks and the risk of heel ulceration should be considered 111, 112. Clearly, more studies are needed before a definitive statement can be made about the safety and efficacy of preventive surgery.

Research on offloading interventions other than footwear and surgery is missing in the literature, and this gap must be filled in order to build up an evidence base to properly assess clinical effectiveness and the use of these interventions.

Ulcer treatment

There is a fairly strong evidence base showing that TCCs heal a higher proportion of neuropathic plantar ulcers at a faster rate than other, mainly removable, offloading modalities. Interestingly, however, one RCT 16 also showed that patients wearing a TCC were much less active than patients wearing half-shoes, a result that may have contributed to improved healing in TCC and shows that some form of activity measurement should ideally be part of any study comparing offloading modalities for ulcer healing. Alternative casting devices such as cast shoes and cast boots may heal the same percentage of neuropathic plantar foot ulcers and at a similar rate as TCC, although this will require confirmation in prospective controlled studies before these devices can become widely accepted as first-choice offloading treatment modalities.

On the basis of the available evidence, therapeutic footwear does not appear suitable for ulcer treatment since other offloading modalities such as TCC are more effective. Two RCTs showed that removable walkers that are made non-removable are more effective than a standard removable walker and as effective as a TCC in healing diabetic neuropathic plantar foot ulcers. Although adherence to offloading treatment was not specifically measured in these studies, the positive effect of making these walkers non-removable points to the importance of the constant use of devices intended to promote wound healing. The evidence base for the use of felted foam for the treatment of neuropathic plantar forefoot ulcers is not strong.

Lengthening the Achilles tendon seems to have limited additional value compared to treatment with TCC alone in healing plantar forefoot ulcers.

The evidence also indicates that other surgical options, such as arthoplasty, metatarsal head resection, surgical excision of bony prominences, and exostectomy, do not improve the proportion of healed ulcers; only the time to heal is slightly shorter. Therefore, more studies are needed to better define the role of surgical approaches compared to conservative treatment. The evidence base for the use of other offloading modalities such as bed rest, crutches, canes, wheelchairs, bracing, and plugs is small or non-existent.

Plantar pressure reduction

Several cross-sectional and uncontrolled studies show that TCCs and removable walkers are very effective in reducing plantar pressure at sites of ulceration and high peak pressure in the diabetic neuropathic foot. This effective degree of offloading is likely an important aspect of plantar foot ulcer healing using these devices, together with the non-removable nature of the intervention. Half shoes and forefoot offloading shoes also reduce peak pressures significantly but not as much as TCC and removable walkers, which likely contributes to the lesser clinical effectiveness of these devices. Within this context, it is noted that the threshold for offloading (expressed as a peak pressure level or pressure-time integral) which is required to adequately heal neuropathic foot ulcers in diabetic patients is currently unknown.

The mechanical effect of special footwear is thought to rely on plantar pressure reduction over the at-risk or previously ulcerated area and a transfer of load to other regions. Significant reductions in plantar pressure can be achieved with footwear with a rocker bottom outsole and footwear that includes a molded insole when compared with standard footwear. Similarly, polyurethane foam sheets worn in shoes might be effective in reducing forefoot plantar pressures when compared with wearing shoes alone. Again, the wide diversity of methods and materials used in footwear studies prevents firm conclusions on this topic. Clearly, more comprehensive analyses of different interventions tested in the same group of patients using the same pressure measuring device are required.

The findings on the use of silicone injections for the purpose of pressure reduction suggest that these injections should be considered only as a temporary method for reducing plantar pressure (i.e. not more than 12 months). The efficacy of long-term use of this approach is still in question. Similarly, data on the effect of ATL suggests that this procedure reduces pressure in the immediate post-operative period, but this effect may not persist. The adverse biomechanical effects of other surgical procedures should also be considered, as they may reduce pressure in the immediate post-operative period, but may cause unintended increases in pressure in other areas of the foot. Two studies showed that regular callus removal has a beneficial effect on plantar pressure in the neuropathic diabetic foot. However, despite its widespread use in clinical practice throughout the world, quite surprisingly, the effectiveness of callus removal in preventing ulceration has, to date, not been examined in a prospective controlled study design.

Other considerations

Several issues related to the findings and conclusions in this systematic review should be considered. First, we acknowledge the difficulties inherent in the design of trials involving surgical procedures. Regional variations in equipment, technique, and surgical practice make randomized trials more challenging in surgery than in other interventions in this area. For this reason, we accept that the key factors of pressure reduction, wound healing and/or prevention of recurrence may be suitable endpoints for other types of investigations in this area. Furthermore, we also urge individual centers not to abandon publication of technique-driven studies. Second, the effectiveness of footwear and offloading in ulcer healing and prevention is always likely to be complicated by patient adherence to treatment. It is self-evident that even the most effective device will not be successful if it is not worn. Studies have reported very low percentages between 25% and 28% of the total ambulatory time in the use of prescribed footwear and offloading devices by diabetic patients 113, 114. While non-removable devices can overcome this problem during ulcer healing, there is clearly still much to learn regarding ways to encourage patients to adhere to footwear treatment. The provision of more attractive footwear may help in this regard 115. Third, even though the costs of footwear and offloading devices are often considered to be substantial, it is important that such costs are viewed in relation to the total cost of care for a foot ulcer and the inherent increased risk of amputation. Although the direct costs of a foot ulcer have been estimated to be substantial, the question of cost effectiveness of the various treatments still remains to be determined 116. Fourth, a major obstacle to comparison of studies is the lack of standardization in terminology, prescription, manufacture, and material properties of footwear and offloading devices. Finally, our choice to assess only studies, which have diabetic patients as subject population has prevented drawing conclusions on interventions that prove to be effective in studies on other patient groups or healthy people. While similar results have been observed clinically in diabetic patients, definitive studies in this patient group are required before recommendations for the use of these interventions can be made.

Conclusions

This systematic review of the literature shows that the evidence base to support the use of footwear and offloading in the prevention and treatment of diabetic foot ulcers is small and in some areas non-existent. Generally, the methodological quality of the studies assessed was rather poor. The best support from the evidence gathered is for the use of non-removable devices such as TCC and non-removable walkers in the treatment of neuropathic plantar foot ulcers. More high-quality controlled studies are urgently needed to confirm the promising effects of footwear and offloading interventions designed to prevent ulcers, heal ulcers, or reduce plantar pressure that have been demonstrated in several controlled and many uncontrolled studies reviewed here. Only with such evidence from robust well-controlled studies, which incorporate measures of offloading efficacy, will these modalities be widely accepted and implemented in diabetic foot practice throughout the world.

Acknowledgements

The authors gratefully acknowledge the help of Catherine Griffis in obtaining full paper copies of original articles.

Conflict of interest

P. R. Cavanagh owns stock in DIApedia and is an inventor on US patents 6,610,897 6,720,470 and 7,206,718 which elucidate a load relieving dressing and a method of insole manufacture for offloading diabetic feet. DIApedia receives royalties from a licensing agreement with Acor Orthopaedic. P.R. Cavanagh has received honoraria from Merck, Eli Lilly, and ConvaTec and he is a recipient of grants from the National Institutes of Health. None of the other authors have conflicts of interest.

APPENDIX A

Literature search strings for each database

Medline database (1950–1 May 2006)
General

(((“Diabetes mellitus”[MeSH]) OR (diabetes mellitus) OR (diabetes) OR (diabetic)) AND ((“epidemiologic study characteristics”[MeSH]) OR (“Case Reports”[Publication Type]) OR (clinical trial*) OR (case-control stud*) OR (cohort stud*) OR (comparative stud*) OR (cross-sectional stud*) OR (case series) OR (case stud*)) AND ((“Foot ulcer”[MeSH]) OR (ulcer*) OR (“Wound healing”[MeSH]) OR (“Diabetic Foot”[MeSH]) OR (diabetic foot) OR (“Pressure”[MeSH]) OR (pressure*) OR (“Weight-bearing”[MeSH]) OR (“Biomechanics”[MeSH]) OR (biomech*) OR (“Stress”[MeSH]) OR (stress*) OR (offload*) OR (off-load*)))

For each intervention subcategory, this general search string was completed as follows:

Casting

AND ((“Casts, Surgical”[MeSH]) OR (cast*) OR (casting*))

Footwear

AND ((Footwear) OR (“Shoes”[MeSH]) OR (shoe*) OR (insole*) OR (“Orthotic Devices”[MeSH]) OR (orthoses) OR (orthosis) OR (orthotic*) OR (sock*) OR (hosiery))

Surgical offloading

AND ((surgic*) OR (surger*) OR (“Surgical Procedures, Operative”[MeSH]) OR (“Surgery”[Sub-heading]) OR (“Diabetic Foot/surgery”[MeSH]) OR (“Achilles Tendon/surgery”[MeSH]) OR (“Equinus Deformity/surgery” [MeSH])) AND ((“Achilles Tendon”[MeSH]) OR (Achilles*) OR (silicon*) OR (callus*) or (callosit*) OR (“Debride- ment” [MeSH]) OR (debrid*) OR (resect*) OR (metatar- sal*) OR (augment*) OR (“Osteotomy”[MeSH]) OR (osteotom*) OR (“Arthroplasty”[MeSH]) OR (Arthro- plast*) OR (Ostectom*) OR (Exostectom*) OR (“External Fixators”[MeSH]) OR (External fixat*))

Other offloading interventions

AND ((“Bed rest”[MeSH]) OR (bed rest) OR (“Canes” [MeSH]) OR (cane*) OR (“Crutches”[MeSH]) OR (crutch*) OR (“Wheelchairs”[MeSH]) OR (wheelchair) OR (brace*) OR (bracing*) OR (“Walkers”[MeSH]) OR (walker*) OR (prefab*) OR (“Bandages”[MeSH]) OR (offloading dressing*) OR (off-loading dressing*) OR (foam*) OR (felt*) OR (padding*) OR (plug*))

EMBASE database (1980–1 May 2006)

General

(((exp diabetes mellitus/) OR (diabetes mellitus) OR (diabetes) OR (diabetic)) AND ((clinical study/or case control study/or case report/or case study/or longitudinal study/or major clinical study/or prospective study/or clinical trial/or randomized controlled trial/) OR (cohort analysis/) OR (clinical trial$ ) OR (case-control stud$ ) OR (cohort stud$ ) OR (comparative stud$ ) OR (cross-sectional stud$ ) OR (case series) OR (case stud$ )) AND ((Foot ulcer/) OR (ulcer/) OR (ulcer$ ) OR (ulcer healing/) OR (Diabetic Foot/) OR (diabetic foot) OR (Pressure/) OR (pressure$ ) OR (Weight-bearing/) OR (biomechanics/) OR (biomech$ ) OR (mechanical stress/) OR (stress$ ) OR (offload$ ) OR (off-load$ )))

For each intervention subcategory, this general search string was completed as follows:

Casting

AND ((plaster cast/) OR (cast$ ) OR (casting$ ))

Footwear

AND ((Footwear) OR (shoe/) OR (shoe$ ) OR (insole$ ) OR (orthosis/) OR (orthoses) OR (orthosis) OR (orthotic$ ) OR (sock$ ) OR (hosiery))

Surgical offloading

AND ((Achilles Tendon$ ) OR (Achilles$ ) OR (silicon$ ) OR (callus/) OR (callus$ ) or (callosit$ ) OR (Debridement/) OR (debrid$ ) OR (resect$ ) OR (metatarsal$ ) OR (augment$ ) OR (metatarsal osteotomy/) OR (osteotom$ ) OR (Arthroplasty/) OR (Arthroplast$ ) OR (Ostectom$ ) OR (Exostectom$ ) OR (External Fixator$ ) OR (External fixat$ ))

Other offloading interventions

AND ((bed rest/) OR (bed rest) OR (walking aid/) OR (cane$ ) OR (crutch$ ) OR (Wheelchair/) OR (wheelchair) OR (brace/) OR (brace$ ) OR (bracing$ ) OR (walker$ ) OR (prefab$ ) OR (Bandage/) OR (offloading dressing$ ) OR (off-loading dressing$ ) OR (foam/) OR (foam$ ) OR (felt$ ) OR (padding$ ) OR (plug$ ))

CINAHL database (1982–1 May 2006)

General

(((exp diabetes mellitus/) OR (diabetes mellitus) OR (diabetes) OR (diabetic)) AND ((exp experimental studies/) OR (exp nonexperimental studies/) OR (exp quasi-experimental studies/) OR (Case Studies/) OR (Comparative studies/) OR (clinical trial$ ) OR (case-control stud$ ) OR (cohort stud$ ) OR (comparative stud$ ) OR (cross-sectional stud$ ) OR (case series) OR (case stud$ )) AND ((Foot ulcer/) OR (ulcer/) OR (ulcer$ ) OR (Diabetic Foot/) OR (diabetic foot) OR (Pressure/) OR (pressure$ ) OR (Weight-bearing/) OR (exp biomechanics/) OR (biomech$ ) OR (stress$ ) OR (offload$ ) OR (off-load$ )))

For each intervention subcategory, this general search string was completed as follows:

Casting

AND ((Casts/) OR (cast$ ) OR (casting$ ))

Footwear

AND ((Footwear) OR (Orthopedic footwear/) OR (shoes/) OR (shoe$ ) OR (insole$ ) OR (exp orthoses/) OR (orthoses) OR (orthosis) OR (orthotic$ ) OR (sock$ ) OR (hosiery))

Surgical offloading

AND ((surgic$ ) OR (surger$ ) OR (exp surgery, operative/)) AND ((Achilles Tendon/) OR (Achilles$ ) OR (silicon$ ) OR (callus/) OR (callus$ ) or (callosit$ ) OR (Debridement/) OR (debrid$ ) OR (resect$ ) OR (metatarsal$ ) OR (augment$ ) OR (Osteotomy/) OR (osteotom$ ) OR (Arthroplasty/) OR (Arthroplast$ ) OR (Ostectom$ ) OR (Exostectom$ ) OR (External fixat$ ))

Other offloading interventions

AND ((bed rest/) OR (bed rest) OR (Canes/) OR (cane$ ) OR (Crutches/) OR (crutch$ ) OR (Wheelchairs/) OR (wheelchair) OR (brace$ ) OR (bracing$ ) OR (Walkers/) OR (walker$ ) OR (prefab$ ) OR (offloading dressing$ ) OR (off-loading dressing$ ) OR (foam$ ) OR (felt$ ) OR (padding$ ) OR (plug$ ))

Cochrane database (systematic reviews, central controlled trials register, DARE, NHS EED, and HTA) (1800–1 May 2007)

General

(((Diabetes mellitus[MeSH]) OR (diabetes mellitus) OR (diabetes) OR (diabetic)AND (epidemiologic study characteristics[MeSH]) OR (Case Reports[Publication Type]) OR (clinical trial*) OR (case-control stud*) OR (cohort stud*) OR (comparative stud*) OR (cross-sectional stud*) OR (case series) OR (case stud*) AND (Foot ulcer[MeSH]) OR (Wound healing[MeSH]) OR (Diabetic Foot[MeSH]) OR (Pressure[MeSH]) OR (Weight-bearing[MeSH]) OR (Biomechanics[MeSH]) OR (Stress[MeSH]) OR (ulcer*) OR (diabetic foot) OR (pressure*) OR (biomech*) OR (stress*) OR (offload*) OR (off-load*)

For each intervention subcategory, this general search string was completed as follows:

Casting

AND (Casts, Surgical[MeSH]) OR (cast*) OR (casting*)

Footwear

AND (Shoes[MeSH]) OR (Orthotic Devices[MeSH]) (Footwear) OR (shoe*) OR (insole*) OR (orthoses) OR (orthosis) OR (orthotic*) OR (sock*) OR (hosiery)

Surgical offloading

AND (Surgical Procedures, Operative[MeSH]) OR (Diabetic Foot/surgery[MeSH]) OR (Achilles Tendon/surgery [MeSH]) OR (Equinus Deformity/surgery[MeSH]) AND (Surgery[Subheading]) OR (surgic*) OR (surger*) (Achilles Tendon[MeSH]) OR (Debridement[MeSH]) OR (Osteotomy[MeSH]) OR (Arthroplasty[MeSH]) OR (External Fixators[MeSH]) OR (achilles*) OR (silicon*) OR (callus*) or (callosit*) OR (debrid*) OR (resect*) OR (metatarsal*) OR (augment*) OR (osteotom*) OR (arthroplast*) OR (ostectom*) OR (exostectom*) OR (external fixat*)

Other offloading interventions

AND ((Bed rest[MeSH]) OR (Canes[MeSH]) OR (Crutches [MeSH]) OR (Wheelchairs[MeSH]) OR (Walkers[MeSH]) OR (Bandages[MeSH]) (bed rest) OR (cane*) OR (crutch*) OR (wheelchair) OR (brace*) OR (bracing*) OR (walker*) OR (prefab*) OR (offloading dressing*) OR (off-loading dressing*) OR (foam*) OR (felt) OR (padding*) OR (plug*)

APPENDIX B

Evidence tables
ReferenceStudy design + qualityStudy population and characteristicsIntervention (I) and control (C) conditionsOutcome categoryResults on primary/secondary outcomes + statisticSIGN Level of evidenceComments/weaknesses
  • *

    Texas, University of Texas Classification of Diabetic Foot Wounds

Agas et al., 2006 (18)CohortPooled groups (hydrogel+ platelet-derived growth factor (PDGF)):I: Windowed TCCUlcer treatmentHealing proportions (PDGF groups):2Data extracted from 3 different patient groups
 Quality: 1/8Patients: I : 20, C : 26C: Offloading shoes (two different types) I : 15/18 (83%)  
  Wounds: I : 21, C : 29  C : 6/24 (25%) PDGF was part of the treatment in both I and C groups
     P = 0.002  
  PDGF groups:  Duration of therapy (pooled groups): Healing proportions reported only in the PDGF groups
  Patients: I : 17, C : 21  I: mean 59 days (SD: 41)  
  Wounds: I : 18, C : 24  C: mean 110 days (SD: 75)  
     P < 0.008  
  Non-ischemic, non-infected full-thickness ulcers (Wagner II and III) between 1 and 16 cm2 on any part of the foot    Some subjects were exposed to both treatments (hydrogel + PDGF)
       No information on adherence in control group
  Study duration: healing up to 1 year    Plantar pressure not measured
  Lost to study: not reported    Presence of peripheral neuropathy and foot deformities not reported
       No data on complications
Armstrong et al., 2001 (16)RCTPatients: I : 25, C1 : 25, C2 : 25I:TCCUlcer treatmentHealing proportions:1−Selective loss with dropout percentage larger than 20% in TCC
 Quality: 3/9Superficial non-infected, non-ischemic neuropathic plantar foot ulcers (grade 1A, Texas*)C1:Removable walker (RW) I : 17/19  
   C2:Half shoe C1 : 13/20  
     C2 : 14/24 Pooling of data from the RW and Half-shoe prevents conclusions on individual devices
     Ivs.C1 + C2:  
     P = 0.026  
     OR = 5.4 (95%CI = 1.1–26.1)  
       No information on adherence in control groups
  Study duration: 12 weeks or healing     
  Lost to study: 12 (I : 6, C1 : 5, C2 : 1)  Average time to healing (days):  
     I : 33.5 Plantar pressure not measured
     C1 : 50.4  
     C2 : 61.0  
     IvsC1: P = 0.07, ns)  
     IvsC2: P = 0.005  
 Average number of daily steps:  
     I : 600  
     C1 : 768  
     C2 : 1462  
     IvsC2: P = 0.04  
Armstrong et al., 2003 (12)CohortPatients: I : 21, C : 20I: Metatarsal-phalangeal joint arthroplastyUlcer treatmentHealing proportions and time to healing:2−No information on adherence to wearing post-healing footwear
 Quality: 3/8Neuropathic non-infected plantar ulcers (grade 1A + 2A, Texas*)C: TCCUlcer recurrenceI: 21/21, all within 17 days  
  Study duration: until healing (67 days)Post-healing: comfort shoes or prescription depth inlay shoesComplications (infection, amputation)C : 20/20, all within 67 days  
  Follow-up: 6 months  P = 0.0001 (days)  
  Lost to follow-up: not reported  Recurrence:  
     I: 4.8%, C: 35%  
     P = 0.02,  
     OR 7,6 (95%CI: 1.1–261.7)  
     No differences in infection or amputation rate between groups  
Armstrong et al., 2005 (13)CohortPatients: I : 22, C : 18I: Fifth metatarsal head resectionUlcer treatmentHealing proportions and time to healing:2−No description about the type of offloading in control group
 Quality: 2/8Neuropathic non infected plantar ulcers (grade 1A + 2A, Texas*)C: “Aggressive” offloadingUlcer recurrenceI : 22/22, all within 5.8 weeks No information on adherence to wearing post-healing footwear
  Study duration: until healing (8.7 weeks)Post-healing: C : 18/18, all within 8.7 weeks  
  Follow-up: 6 monthsDepth-inlay shoes with multidurometer, multilaminar inlays P = 0.02 (time to healing)  
  Lost to follow-up: not reported  Ulcer recurrence:  
     I:4,5%, C:27,8%  
     P = 0.04  
     OR = 8.3 (95%CI: 1.1–67.9)  
Armstrong et al., 2005 (80)RCTPatients: I : 23, C : 27I: Non-removable walkerUlcer treatmentHealing percentages:1 +No information on adherence
 Quality: 6/9Superficial non-infected, non-ischemic neuropathic plantar foot ulcers (grade 1A, Texas*)C: Removable walker I:19/23 Plantar pressure not measured
     C:14/27  
     P = 0.02  
     OR = 1.8 (95%CI = 1.1–2.9)  
  Study duration: 12 weeks or healing     
  Lost to study: 4     
Black, 1990 (83)CBAPatients: I : 10, C : 10I: Walking brace (plantar healing orthosis) and topical ulcer careUlcer treatmentSurface area reduction:2−Poorly described study
 Quality: 2/8Non-ischemic neuropathic plantar foot ulcers  No significant difference between I and C groups No information on adherence Plantar pressure not measured
  Study duration: 90 daysC: Topical ulcer care alone   Plantar pressure not measured
  Lost to study: not reported     
Caravaggi et. al., 2000 (15)RCTPatients: I : 26, C : 24I: TCCUlcer treatmentHealing proportions:1−No exact values on ulcer size
     I: 13/26  
     C: 5/24 Short study duration
 Quality: 4/9Non-infected, non-ischemic neuropathic plantar foot ulcersC: Therapeutic shoe (rocker-bottom, extra-depth + 8-mm plastazote insoles) P = 0.03  
  Study duration: 30 days    No information on adherence in control group
  Lost to study: not-reported    Plantar pressure not measured
Dargis et al., 1999 (7)CBAPatients: I : 56; C : 89I: Clinic treatment and education Extra-depth shoes includedUlcer preventionUlcers in 2 years:2−Plantar pressure was not measured
 Quality: 2/7Study duration: 24 months  I: 30.4%  
     C: 58.4%  
     P < 0.001  
  Lost to study: I : 8, C : 13    No information on adherence
  All patients had a history of ulceration.C: Clinic treatment and education    
  Neuropathic and Vascular status not different between groups     
Ha Van et al., 2003 (17)CBAPatients: I : 42, C : 51I: TCCUlcer treatmentHealing proportions :2−Very long study duration (approximately 600–900 days)
     I: 81%  
     C: 70%  
     P = 0.017  
 Quality: 1/7Non-infected, non-ischemic neuropathic plantar foot ulcers (grade 1A, Texas*)C: Half shoe or heel-relief shoe Hazard ratio:  
     1.68 (CI :1.04–2.70) Plantar pressure not measured
  Study duration: until healing/complications  Average time to heal (days): P-value for healing proportions appears non-significant when recalculated using Fisher exact test
     I: 69 (SD 35)  
     C: 134 (SD 133)  
     No statistic  
  Lost to study: not-reported  Adherence: I:98%, C:10% Hazard ratio is normally limited to survival analysis
     P = 0.001  
Katz et al., 2005 (81)RCTPatients: I : 21, C : 20I: Non-removable walkerUlcer treatmentHealing rates:1 +Reported power of 95% for detecting a 5% difference in outcome seems unrealistic with number of patients
     I: 17/21 (80%)  
     C: 15/20 (74%)  
     P = 0.65  
 Quality: 5/9Non-infected, non-ischemic neuropathic plantar foot ulcers (grade 1A or 2A, Texas*)C: TCC   Study seems underpowered for equivalence
  Study duration: 12 weeks or healing    Plantar pressure not measured
  Lost to study: 7 (I : 4, C : 3)     
Lin et al., 2000 (36)Case- ControlPatients I : 14, C : 14I: First metatarsal-phalangeal joint arthroplasty + TCCUlcer treatmentHealing proportions:2−No inferential statistical analysis
     I: 14/14, all in 23 days  
     C: 14/14, all in 47 days  
 Quality: 2/8Neuropathic non- infected plantar inter-phalangeal joint ulcersC: TCC No statistic  
  Study duration: mean 26 weeks     
  Lost to study: not reported     
Litzelman et al., 1997 (6)RCTPatients: 395 were enrolled and randomized.I: Multifaceted educational intervention including recommendation for footwear where appropriateUlcer preventionIn total 63 ulcers in 53 patients at 1 year1−Poorly described study design: no information on assignment to treatment or control.
 Quality: 2/9All patients had type 2 diabetes. No details about neuropathy or foot deformities provided.  Recommendation for special shoes predicted ulcers at 1 year (OR = 2.19) Study patients did not receive special footwear. Footwear was variable and uncontrolled.
  Study duration: 1 year Purchase of new shoes prevented ulcers at 1 year (OR 0.6)  Patients were poor and non-compliant(mean HBA1c=10.2)
  Lost to studyC: Standard care and observation   Data collected before Medicare Shoe Bill
       Only one patient reported possessing and wearing the special footwear
       No information on outcome in control group
 Plantar pressure not measured No information on adherence
Maluf et al., 2004 (104)RCTPatients: I : 14, C : 14I: TCC + TALBarefoot plantar pressure reductionForefoot peak pressure (N/cm2):1 +Subset analysis of Mueller et al., 2003
 Quality: 5/9Neuropathic non ischemic plantar ulcer (Wagner 2)Charcot feet excludedC: TCC I: 89.2, 64.7, and 88.8 No statistical analysis between groups at different stages (only one overall p value from analysis of variance)
     C: 82.6, 86.5, and 97.0  
  Study duration: mean 8 (SD 2) months     
  Lost to study: 0  Rearfoot peak pressure (N/cm2):  
     I: 52.2, 70.0, and 70.6  
     C: 51.8, 49.0, and 55.8  
     At pre, 3-week post, and 8-month post treatment, respectively  
     Group comparisons:  
     Forefoot: P = 0.005  
     Rearfoot: P = 0.018  
Mueller, et al., 1989 (14)RCTPatients: I: 21, C:19I: TCCUlcer treatmentHealing proportion :1−Wounds that did not reduce in size by six weeks considered “not healed”
     I:19/21  
     C:6/19  
     P < 0.05  
 Quality: 2/9Non-infected neuropathic plantar foot ulcers (Wagner 1 and 2)C: Accommodative footwear (i.e. healing sandal + extra depth shoe with a plastazote insole)   No study duration provided
  Study duration: not-reported    No information on adherence in control group
  Lost to study: not reported  Healing time (days): Plantar pressure not measured
     I: 42 (SD 29)  
     C: 65 (SD 29)  
     P < 0.05  
Mueller et al., 2003 (10)RCTPatients: I:30, C:33I : TCC + TALUlcer treatmentHealing proportions:1 +Confusing primary outcome
     I:30/30 (100%)  
     C:29/33 (88%)  
     P = 0.12  
 Quality: 5/9Recurrent neuropathic plantar forefoot ulcersLost to study: 1C : TCCUlcer recurrence (prevention)  Powered on 7-month recurrence
  Study duration: until healing Lost to study : 1Post-healing: Prescription shoes and insoles Ulcer recurrence at 7 months: Plantar pressure not measured (for ulcer recurrence)
     I:4/27 (15%)  
     C:16/27 (59%)  
  Follow-up 1: 7months  P = 0.001  
  Lost to follow-up: 8  RR = 4.0 (95%CI = 1.8–8.9) No information on adherence to wearing post-healing footwear
  Follow-up 2:2.1 (SD 0.7) years Lost to follw up:0  Ulcer recurrence at 2.1 years:  
     I:10/26 (38%)  
     C:21/26 (81%)  
     P = 0.002  
     RR = 2.1 (95%CI = 1.7–9.6)  
Piaggesi et al., 1998 (11)RCTPatients: I:21, C:20I: Surgical excision of ulcer, debridement of bone and wound closureUlcer treatmentHealing proportions:1 +P-value on proportion of healing and infection appear non-significant when recalculated
     I: 21/22  
     C: 19/24  
     P < 0.05  
 Quality: 5/9Neuropathic non-ischemic, non infected plantar ulcersFootwear: fitted shoes with orthosis + crutchesUlcer recurrence (prevention)Time to healing (mean) : I:47 days Same follow-up period of 6 months for both healing and recurrence
     C:129 days  
     P < 0.001  
  Study duration: 6 monthsC: Standard care, including saline gauze dressing and offloading (fitted shoes with custom made orthosis + crutches)Infection (complication)Ulcer recurrence: No information on adherence to wearing fitted shoes and post-healing footwear
     I: 3/21  
     C: 8/19  
     P < 0.01  
  Lost to study: 0Post-healing footwear: Molded shoes with orthosis    
  Follow up: 6 months     
  Lost to follow-up: 0  Proportion infections:  
     I:3/24  
     C:1/22  
     P < 0.05  
Reiber et al., 2002 (2)RCTPatients: I1 : 121, I2: 119; C:160I1: Therapeutic shoe corkUlcer prevention (recurrence)Number of ulcers in 2 years:1−Ulcer badly defined
 Quality: 3/9Study duration: 24 monthsI2: Therapeutic shoe I1: 26 Plantar pressure was not measured
  Lost to study: 66 (I1: 17, I2: 23, C: 26)Polyurethane I2: 31 Many patients may not be neuropathic
  All patients had history of ulcerationC: Control shoe C: 38 Multiple ulcers could have occurred in one person
  Foot deformities requiring custom footwear were  Not significant  
     Number of patients with an ulcer in 2 years:  
     I1: 18/121 (14.9%)  
     I2: 17/119 (14.3%)  
     C: 27/160 (16.9%)  
  excluded  Not significant. 37% crossover to other footwear from control group
  58% had loss of protective sensation (10g monofilament)    No information on adherence
       Low reulceration rates in all study groups
Uccioli et al., 1995 (5)RCTPatients: I:33; C:36I: Therapeutic shoeUlcer preventionPercentage ulcers in 1 year:1-Plantar pressure not measured
 Quality: 3/9Study duration: 12 monthsC: Own shoe I: 27.7%  
  Lost to study: not reported  C: 58.3%  
  All patients had history of ulceration or were at risk of ulceration (not defined)  P = 0.009  
  Neuropathic and vascular status not different between groups  Shoe use: frequent and continuous in both groups (qualitative assessment)  
van Schie et al., 2000 (103)RCTPatients: I:14, C:14I: 6 liquid silicon injections at 2-week intervals beneath debrided callus at site of highest pressurePlantar pressure reduction (forefoot)Forefoot peak pressure:1 +Selective loss to follow-up after 2 years
 Quality: 7/9Study duration: 3, 6, and 12 monthsC: 6 saline injections at same interval and locationTissue thicknessI: − 232, − 170, − 165, and − 24 kPa No statistical comparison between study groups, only comparison with baseline
  Lost to study: 0  C: − 50, − 5, + 77, and + 37 kPa  
  Follow-up: 1 year  At 3, 6, 12, and 24 months compared wit baseline  
  Lost to follow-up: 12 (I: 3, C: 9)  P< 0.05 at 12-months, non-significant at 24 months in intervention group  
  Neuropathy and callus present in the forefoot  Tissue thickness:  
     I: + 1.8, + 2.0, + 1.3, and + 1.1 mm  
     C: + 0.1, + 0.2, + 0.3, and − 0.1 mm, at each interval  
     P < 0.05 at 12 months and 24 months in intervention group  
Viswanathan et al., 2004 (8)CohortPatients: I1 : 100; I2 : 59; I3 : 32; C:50I1: MCR insoles in customized sandalsUlcer preventionUlcers in 9 months:2−Confounding factors clearly present (poverty, foot deformity)
 Quality: 2/8Study duration: 9 monthsI2: Polyurethane insoles in customized sandalsPlantar pressure reductionI1: 4% Inconsistencies between body of text and Figure 1.
  Lost to study: not reportedI3: Molded insoles in customized sandals I2: 4% Neuropathy results were not reported
  All patients had a history of ulcerationC: Usual sandals with hard leather board insole I3: 3% Peak pressure appears to be an order of magnitude too low
  Group I3 had foot deformities (not defined)  C: 33% Possible change in calibration of pressure insoles over a 9 month period is not addressed
     P < 0.001 (Ivs.C)  
     Change in plantar pressure:  
     I1: − 57.4%  
     I2: − 61.96%  
     I3: − 58.02%  
     C: + 39.38%  
     P < 0.01 (9 months vs. baseline)  
Zimny et al., 2003 (82)RCTPatients: I:24, C:30I: Felted foam in post-operative shoeUlcer treatmentHealing time (days):1−No data on number of ulcers healed
 Quality: 3/9Non-infected, non-ischemic neuropathic plantar forefoot ulcers (Wagner 1 or 2)C: Pressure relief half shoe I:75(CI:67–84) Survival curve not shown in article
  Study duration: 10 weeks  C:85(CI:79–92) No information on adherence to footwear use
  Lost to study: not reported  P = 0.03 Plantar pressure not measured
     Ulcer radius reduction per week (mm):  
     I:0.48 (CI:0.42–0.56), C:0.39 (CI:0.35–0.42)  
     P = 0.005  
     Survival analysis  
     P = 0.06 (not significant) between groups  

Ancillary