Preventing Foot Ulcers in Patients With Diabetes

Context Among persons diagnosed as having diabetes mellitus, the prevalence of foot ulcers is 4% to 10%, the annual population-based incidence is 1.0% to 4.1%, and the lifetime incidence may be as high as 25%. These ulcers frequently become infected, cause great morbidity, engender considerable financial costs, and are the usual first step to lower extremity amputation.

Objective To systematically review the evidence on the efficacy of methods advocated for preventing diabetic foot ulcers in the primary care setting.

Data Sources, Study Selection, and Data Extraction The EBSCO, MEDLINE, and the National Guideline Clearinghouse databases were searched for articles published between January 1980 and April 2004 using database-specific keywords. Bibliographies of retrieved articles were also searched, along with the Cochrane Library and relevant Web sites. We reviewed the retrieved literature for pertinent information, paying particular attention to prospective cohort studies and randomized clinical trials.

Data Synthesis Prevention of diabetic foot ulcers begins with screening for loss of protective sensation, which is best accomplished in the primary care setting with a brief history and the Semmes-Weinstein monofilament. Specialist clinics may quantify neuropathy with biothesiometry, measure plantar foot pressure, and assess lower extremity vascular status with Doppler ultrasound and ankle-brachial blood pressure indices. These measurements, in conjunction with other findings from the history and physical examination, enable clinicians to stratify patients based on risk and to determine the type of intervention. Educating patients about proper foot care and periodic foot examinations are effective interventions to prevent ulceration. Other possibly effective clinical interventions include optimizing glycemic control, smoking cessation, intensive podiatric care, debridement of calluses, and certain types of prophylactic foot surgery. The value of various types of prescription footwear for ulcer prevention is not clear.

Conclusions Substantial evidence supports screening all patients with diabetes to identify those at risk for foot ulceration. These patients might benefit from certain prophylactic interventions, including patient education, prescription footwear, intensive podiatric care, and evaluation for surgical interventions.

Among persons diagnosed as having diabetes mellitus, the lifetime risk of developing a foot ulcer is estimated to be 15%.¹ Based on recent studies, the annual population-based incidence ranges from 1.0% to 4.1%² and the prevalence ranges from 4% to 10%, which suggests that the lifetime incidence may be as high as 25%.^3,4Quiz Ref IDLower extremity disease, including peripheral arterial disease, peripheral neuropathy, foot ulceration, or lower extremity amputation, is twice as common in diabetic persons compared with nondiabetic persons and it affects 30% of diabetic persons who are older than 40 years.⁵ Foot ulcers cause substantial emotional, physical, productivity, and financial losses.^6- 9 The estimated costs of treating a diabetic foot ulcer were $28 000 in a 1999 US study,¹⁰ and $18 000 (with no amputation) and $34 000 (with amputation) in a 2000 Swedish study.¹¹

The most costly and feared consequence of a foot ulcer is limb amputation, which occurs 10 to 30 times more often in diabetic persons than in the general population.^12,13 Diabetes underlies up to 8 of 10 nontraumatic amputations, of which 85% follow a foot ulcer.^1,3,14 The age-adjusted annual incidence for nontraumatic lower limb amputations in diabetic persons ranges from 2.1 to 13.7 per 1000 persons.² Mortality following amputation ranges from 13% to 40% at 1 year, 35% to 65% at 3 years, and 39% to 80% at 5 years—worse than for most malignancies.²

In light of the enormous disease burden of diabetic foot ulcers, it is crucial to know if they are preventable. This review summarizes and critically evaluates evidence on the efficacy of identifying diabetic persons at high risk for foot ulcers and of interventions designed to prevent them.

Assisted by a medical librarian, we conducted a systematic literature search using the EBSCO (EBSCO Information Services, Birmingham, Ala), MEDLINE, and the National Guideline Clearinghouse databases for articles published between January 1980 and April 2004 and used the following phrases: diabetes or diabetic, foot ulcer or infection, and prevention or preventing. The EBSCO database includes the American Medical Association Collection, Comprehensive Biomedical Reference Collection, Cumulative Index to Nursing and Allied Health Literature, Cochrane Database of Systematic Reviews, Cochrane Controlled Trials Register, Database of Abstracts of Reviews of Effectiveness, Health Business Fulltext Elite, International Pharmaceutical Abstracts, Comprehensive Nursing and Allied Health Collection, and the American Medical Association’s Archive. We also searched (1) the bibliography of each identified article; (2) the National Guideline Clearinghouse Web site (http://www.guidelines.gov); (3) an extensive printed diabetic foot reference collection¹⁵; and (4) several Web sites specializing in issues related to the diabetic foot.

This search identified 165 articles that addressed preventing diabetic foot ulcers, including 22 randomized controlled studies, most of which measured changes in the rates of foot ulceration and amputations related to various interventions. For topics on which there were only a few randomized controlled studies, we focused on selected case-control and cohort studies.

Causative Factors.Quiz Ref IDThe causal pathways leading to foot ulceration include several component causes, the most important of which is peripheral neuropathy.¹⁶ This is present to some degree in more than 50% of diabetic persons older than 60 years.¹⁷ Peripheral neuropathy must usually be profound before leading to loss of protective sensation; the consequent vulnerability to physical and thermal trauma increases the risk of foot ulceration 7-fold.^18,19 A second causative factor in foot ulceration is excessive plantar pressure.²⁰ This is related to both limited joint mobility (at the ankle, subtalar, and first metatarsophalangeal joints) and to foot deformities.^21- 23 In one study of patients with peripheral neuropathy, 28% with high plantar pressure developed a foot ulcer during a 2.5-year follow-up compared with none with normal pressure.²⁴ A third component cause is trauma, especially when repetitive. Among 669 persons with a foot ulcer, 21% were attributed to rubbing from footwear, 11% were linked to injuries (mostly falls), 4% to cellulitis complicating tinea pedis, and 4% to self-inflicted trauma (eg, cutting toenails).²⁵ Persons who had a previous foot ulceration could withstand fewer cycles of stress to their feet before an ulcer recurred.²⁶

Contributory Factors. Once a foot ulcer develops, several factors may contribute to adverse outcomes. The most important is atherosclerotic peripheral vascular disease, which is twice as common in persons with diabetes as in persons without diabetes⁵ and particularly affects the femoropopliteal and smaller vessels below the knee, while frequently sparing the pedal vessels.²⁷ Diabetes is also associated with several intrinsic wound-healing disturbances, including impaired collagen cross-linking and matrix metalloproteinase function,^27,28 and immunologic perturbations, especially in polymorphonuclear leukocyte function.^29,30 Furthermore, persons with diabetes have a higher rate of onychomycosis and toe-web tinea infections that can lead to skin disruption.^31- 34

Having a foot ulcer dramatically worsens physical, psychological, and social quality of life.^6- 8,35,36 The obesity and poor vision that are associated with diabetes may also impair self-care. Optimal prevention (and treatment) outcomes require both a motivated patient and an effective medical care system.

Quiz Ref IDPreventing foot complications begins with identifying those at risk. Primary care clinicians should inquire about factors known to be associated with foot ulcers, namely, previous foot ulceration (relative risk [RR], 1.6; 95% confidence interval [CI], 1.2-2.3; P = .004),³⁷ prior lower extremity amputation (RR, 2.8; 95% CI, 1.8-4.3; P<.001),³⁷ long duration (>10 years) of having diabetes (odds ratio [OR], 3.0; P<.04),³⁸ poor glycemic control (glycosylated hemoglobin >9%; OR, 3.2; P<.03),³⁸ and impaired vision (acuity <20/40; RR, 1.9; 95% CI, 1.4-2.6; P<.001).³⁷ Clinicians should also examine the feet for structural abnormalities (eg, calluses, hammer or claw toes, flat feet, bunions), reduced joint mobility, dry or fissured skin, tinea, or onychomycosis,^39,40 and also inspect footwear to ensure proper fit.

Screening for Loss of Protective Sensation. Nerve conduction studies are generally considered the criterion standard for diagnosing peripheral neuropathy. They are less useful in screening for loss of protective sensation (ie, degree of neuropathy beyond which the patient has a measurably increased risk for diabetic foot ulceration),⁴¹ and are not widely available.

Monofilament. The most frequently used instrument for detecting neuropathy is the nylon Semmes-Weinstein monofilament.⁴² Inability to perceive the 10g of force a 5.07 monofilament applies is associated with clinically significant large-fiber neuropathy (Figure). In 3 prospective studies, the Semmes-Weinstein monofilament identified persons at increased risk of foot ulceration with a sensitivity of 66% to 91%, a specificity of 34% to 86%, a positive predictive value of 18% to 39%, and a negative predictive value of 94% to 95%.^37,45,46 Certain brands of monofilaments are more accurate than others⁴⁷ and they should not be used on more than 10 patients without a recovery period of 24 hours.^42,47

While authorities recommend testing 8 to 10 anatomic sites, testing just 4 plantar sites on the forefoot (great toe and base of first, third, and fifth metatarsals) identifies 90% of patients with an insensate site.⁴⁸ Most consider a lack of perception at any site(s) to be abnormal, but as the threshold for an abnormal test is raised from 1 to 4 insensate sites, the sensitivity remains higher than 90% while the specificity improves from 60% to 80%.⁴⁴ Asking the patient to say “yes” or “no” when asked if he/she believes the Semmes-Weinstein monofilament is being applied is equally accurate and quicker than the “forced-choice” method (asking the patient to correctly identify whether it was at time “A” or “B” that the monofilament was applied).⁴⁹

Biothesiometer. A biothesiometer (Xilas Medical, San Antonio, Tex) is a handheld device that assesses vibration-perception threshold.⁵⁰ A rubber tactor is applied to the distal aspect of the toe and the amplitude is increased to a maximum of 100 V (converted from microns).^41,46 In one study, a vibration-perception threshold of more than 25 V had a sensitivity of 83%, a specificity of 63%, a positive likelihood ratio of 2.2 (95% CI, 1.8-2.5), and a negative likelihood ratio of 0.27 (95% CI, 0.14-0.48) for predicting a foot ulceration over 4 years.^19,51 A case-control study with 255 diabetic persons found that having either abnormal Semmes-Weinstein monofilament perception or a vibration-perception threshold of more than 25 V predicted foot ulceration with a sensitivity of 100% and a specificity of 77%.⁴³

Tuning Fork. The tuning fork provides an easy and inexpensive test of vibratory sensation. With a conventional fork, an abnormal response occurs when the patient loses vibratory sensation while the examiner still perceives it.³⁷ With a graduated (Rydel-Seiffer) fork (Gebrueder Martin, Tuttlingen, Germany), persons indicate first loss of vibration at the plantar hallux as the intersection of 2 virtual triangles moves on a scale exponentially from 0 to 8 in a mean (SD) of 39.8 (1) seconds.⁵² This test correlates more strongly with biothesiometer results (r, –0.90; P<.001)⁵³ than the conventional tuning fork,⁵⁴ but the latter predicted foot ulceration in 2 studies.^37,55 Tuning fork results are less predictive of ulceration than results from using the monofilament.³⁷

Screening for Patients With Elevated Plantar Pressure. Devices identifying high plantar pressure include mats to measure barefoot plantar load distribution and transducers distributed in a removable shoe insole to measure pressure inside footwear.⁵⁶ The numerical values generated are often device-specific and cannot easily be compared. There is no generally accepted plantar pressure level associated with an increased risk of diabetic foot ulceration. In case-control studies using the EMED pressure platform system (Novell, Minneapolis, Minn), a peak barefoot dynamic pressure of 70 N/cm² had a sensitivity of 70.0% and a specificity of 65.1%, while a cutoff of 87.5 N/cm² had a sensitivity of 64%, a specificity of 46%, a positive predictive value of 17%, and a negative predictive value of 90% (Table 1).^57,58

Screening for Peripheral Vascular Disease.Quiz Ref IDPeripheral vascular disease is most easily detected by the ankle-brachial index (ABI), which is the ratio of systolic blood pressure in the ankle to that in the brachial artery. An ABI of 0.90 or less suggests peripheral vascular disease, while higher than 1.1 may represent a falsely elevated pressure caused by medial arterial calcinosis.⁵⁹ This test is easily performed, objective, and reproducible.⁵⁹ One large study found that the ABI was strongly related to the risk of foot ulceration (0.3 higher ABI is associated with an RR of 0.83; 95% CI, 0.73-0.96; P = .01).³⁷

Arterial oxygen supply can also be measured by transcutaneous oximetry.⁵⁹ A transcutaneous oxygen tension higher than 30 mm Hg correlates with a high likelihood of wound healing.⁵⁹ Transcutaneous oxygen tension is also inversely associated with the risk of foot ulceration (15 mm Hg higher dorsal foot transcutaneous oxygen tension is associated with an RR of 0.80; 95% CI, 0.69-0.93; P = .004).³⁷ Because accurately measuring transcutaneous oxygen tension requires expensive equipment and a trained technician, it is not routinely used.

Patient Education. Most patient education studies emphasize foot care, but have been short-term and have measured changes in behavior and cognition rather than the incidence of relevant clinical outcomes such as ulceration. Patient education formats have included lectures, hands-on workshops, skills exercises, behavioral modification programs, and telephone reminders (Table 2).

Two recent reviews concluded that patient education improves short-term knowledge and may modestly reduce risk of foot ulcerations and amputations.^51,67 Larger randomized clinical trials are needed to assess which patient education formats are the most effective, how often periodic reinforcement is required, and the long-term effectiveness of various programs.

Physician Education. Health care organizations have used various strategies to improve clinicians’ performance with patient education.^68,69 In one strategy, a computerized registry reminded physicians to enter the patient’s risk status for lower extremity amputation. After 28 months, the percentage of patients who had received foot screening and risk assessment increased from 15% to 76%.⁶⁸ Project LEAP (Lower-Extremity Amputation Prevention), developed by the US Department of Health and Human Services, is a 1-day workshop on diabetes foot care. When given to 560 clinicians from 85 organizations, it improved the rate of documenting foot care education from a baseline of 38% to 62% after 9 months.⁷⁰ More importantly, appropriate foot care self-management increased from 32% to 48%, and there was a trend toward reduced lower extremity amputations.⁷⁰

Another approach is implementing foot care clinical practice guidelines. An Indian Health Service diabetes program observed 669 patients during a standard care period (1986-1989) with routine foot screening; a public health period (1990-1993) with an annual foot examination and initial risk stratification to give those at high-risk special interventions; and a staged diabetes management period (1994-1996) during which clinicians used clinical practice guidelines.⁷¹ The average lower-extremity amputation incidence per 1000 diabetic person-years was 29 during the standard care period, 21 during public health, and 15 during staged management. The overall reduction in lower extremity amputation was 48% (P = .02), and the incidence of first amputation decreased from 21 per 1000 to 6 per 1000 from the first to the third period (P<.001).⁷¹

Clinical Practice Guidelines on the Diabetic Foot.Quiz Ref IDPublished guidelines^72- 77 uniformly recommend that all diabetic persons have an annual foot examination that includes assessing for anatomic deformities, skin breaks, nail disorders, loss of protective sensation, diminished arterial supply, and improper footwear. The clinician should then assign the patient to a risk category by using any of several systems. The recommended interventions for various risk groups differ slightly among the guidelines, but persons at higher risk for foot ulceration should have more frequent foot examinations.^72- 77 (Table 3)

Optimizing Glycemic Control. The Diabetes Complications and Control Trial reported a 57% reduction in the incidence of clinical neuropathy in patients managed with intensive compared with conventional glycemic treatment.⁷⁸ In the United Kingdom Prospective Diabetes Study, a 1% mean reduction in hemoglobin A_1c was associated with a 25% reduction in microvascular complications, including neuropathy. There was also a nonsignificant reduction in amputations (by 36%) in the intensive compared with the conventional treatment group.⁷⁹

Smoking Cessation. Some but not all studies have found a direct causal association between tobacco use and foot ulceration or amputation.³⁷ A case-control study of diabetic persons in the United Kingdom found the lower risk of leg amputation in those of South Asian compared with European ancestry (OR, 0.26; 95% CI, 0.11-0.65; P = .004) partly attributable to their lower rates of smoking (31% vs 57%; P = .03).⁸⁰ Similarly, a cross-sectional study of 1142 patients with type 2 diabetes in Jordan found smoking to be a strong predictor of amputation.⁸¹

Foot Examination by a Clinician. Foot examinations did not significantly reduce amputations among 244 diabetic patients in 1 case-control study (OR, 0.55; 95% CI, 0.2-1.7; P = .31).⁸² These results may reflect the study’s limited sample size, high rates of foot examination in both case and control patients, different degree of risk between the groups, as well as the unusually high rates of diabetes and amputations among the Pima Indian population studied.⁸³ Another randomized study of diabetic persons (N = 91) with a previous foot ulceration found a significantly reduced risk for ulceration recurrence (RR, 0.52; 95% CI, 0.29-0.93; P  = .03) at 1 year for those who received routine podiatric care.⁸⁴ Thus, screening foot examinations are unlikely to reduce the incidence of foot complications unless they eventuate in appropriate specialist referrals (eg, for intensive podiatric care and customized footwear; Table 4).

Custom Footwear and Orthotics. Prescription shoes for high-risk patients should reduce areas of high plantar pressure and friction and accommodate foot deformities (eg, with a deep, wide toe box and ample padding).⁸⁵ Similarly, shoe inserts should cushion the plantar surface and redistribute pressure over a greater surface area.⁸⁵ Clinical data supporting the benefits of prescription footwear are surprisingly meager. In the largest of several studies, 400 persons with a history of a foot ulcer (but without a severe deformity) were randomized to receive extradeep, extrawide therapeutic shoes with customized neoprene-covered cork inserts; therapeutic shoes with nylon-covered polyurethane inserts; or instructed to wear usual footwear.⁸⁶ Persons assigned to therapeutic shoes had a similar incidence of foot ulcer recurrence as controls.⁸⁶ These surprising findings may have resulted from excluding patients with severe foot deformities, a person’s low baseline prevalence (58%) of “foot insensitivity,”⁸⁷ and defining a foot ulcer as existing for 30 days or longer. This and other studies suggest that patients at low risk for foot complications may safely wear well-fitting, good-quality over-the-counter athletic or walking shoes, whereas those with neuropathy and foot deformities may benefit from custom shoes (Table 5). Larger randomized studies should explore which type of therapeutic footwear (including stockings) may best reduce ulceration in patients with neuropathy and deformities and whether patients with only neuropathy require prescription footwear.

Debridement of Calluses. Calluses (hyperkeratotic lesions caused by pressure) further increase pressure, which is a component cause of ulceration. Because debriding hyperkeratoses can reduce peak plantar pressure by 26%,⁹¹ this should be routinely provided by trained personnel. Wearing proper footwear may not only prevent but also reduce development of calluses. Among 78 diabetic persons, the mean size of plantar calluses decreased in direct proportion with the amount of time spent wearing running shoes.⁹² Similarly, among high-risk persons, those who visited podiatrists most frequently (every 3-4 weeks) had the lowest mean plantar pressure before and after callus removal.⁹³ The optimal frequency of podiatric evaluation and management is uncertain.

Foot Specialist and Multidisciplinary Team Care. A few studies have assessed the role of foot specialist care as the main intervention in preventing diabetic foot ulcers.^84,94 Among 91 diabetic persons with a healed foot ulcer, there were 20 ulcer recurrences in those randomized to podiatric care and 32 in the control group after a median follow-up of 386 days (RR, 0.52; 95% CI, 0.30-0.93; P = .03).⁸⁴ In another trial of diabetic persons with neuropathy, 235 were randomized to receive podiatric care at least twice a year and 263 to receive no podiatric treatment.⁹⁵ During the study period (≤3 years), there was no difference in the incidence of foot ulcers, but the podiatric care group had fewer deep ulcers (6 vs 12), infected ulcers (1 vs 10; P<.01), and hospital admission days (24 vs 346; P<.01).⁹⁵

Other studies have used multidisciplinary (eg, podiatrists, internists, surgeons, nurses, dieticians, social workers) care teams. In one study, 341 diabetic persons were examined to categorize baseline risk,⁹⁶ initiate appropriate education and interventions, and schedule follow-up foot examinations and podiatric care.⁹⁷ After 3 years, the incidence of lower-extremity amputation was only 1.1 per 1000 persons per year. Among high-risk persons, those who missed more than 50% of their appointments with the team were 54 times more likely to develop an ulcer and 20 times more likely to require an amputation than those who kept most appointments.⁹⁷

Prophylactic Foot Surgeries. A dramatically increased interest in reconstructive surgery has occurred in the past 2 decades.^72,98- 113 One proposed classification system divides nonvascular foot surgery into elective (to alleviate pain), prophylactic (to reduce risk of ulceration), curative (to heal an open wound), and emergent (to help control a limb-threatening infection).¹¹⁴ Only a few small studies have reported long-term outcomes for prophylactic procedures, generally aimed at correcting deformities that increase plantar pressure (Table 6). For example, a short Achilles tendon leads to increased pull on the calcaneus, elevated plantar-flexory movement about the ankle, and subsequent elevated forefoot plantar pressure; this may be improved by tendon lengthening. Preventing foot ulcers in patients with Charcot arthropathy usually requires an expert pedorthist and potentially a foot surgeon. In this condition, some advocate surgical options including removal of osseous prominences and reconstruction of the deformed foot or ankle, but controlled trials are lacking.^103,120

Revascularization Surgery. Vascular surgeons have developed techniques (eg, bypass grafts from femoral to pedal arteries and peripheral angioplasty) to improve blood flow to an ischemic foot. While these procedures help heal ischemic ulcers, no prospective study shows that they reduce foot ulceration.¹²¹ The reported effect of revascularization procedures on the incidence and site of amputations varies, but most recent studies suggest benefits.^122- 124

Cost-Effectiveness. A recent cost of illness model, based on published data about diabetic complications and the value of health resources from numerous sources found that the mean annual cost of treatment in 2001 was $9306 for an uninfected diabetic foot ulcer, $24 582 for an infected foot ulcer, and $45 579 for a foot ulcer with osteomyelitis.¹²⁵ Another review compiled cost data from 1990 to 1997 from 7 studies—4 conducted in the United States and 3 in other countries.¹²⁶ After inflation adjustment and currency conversion, the cost of treating foot ulcers not requiring amputation ranged from $993 to $17 519, and approached $30 724 in 1 study that spanned 2 years after diagnosis.

A few groups have modeled cost-utility analyses for strategies to prevent foot ulcers. A Markov model from Sweden of intensive prevention (patient education, use of appropriate footwear, and access to therapeutic foot care) for high-risk patients was cost-effective if the incidence of foot ulcers and lower extremity amputations was reduced by 25%.¹²⁷ A similar model for patients with newly diagnosed type 2 diabetes found that implementing a guideline-based foot program that included intensive glycemic control, regular foot examinations, risk stratification, patient education, clinician education, and multidisciplinary foot care increased life expectancy and quality-adjusted life-years and reduced the incidence of foot complications.¹²⁸ The cost of achieving a 10% reduction in the incidence of foot lesions was less than $25 000 per quality-adjusted life-year gained.¹²⁸

Diabetes confers a dramatically increased risk of foot ulceration, but available evidence suggests that this risk may be reduced to some degree by appropriate screening and intervention measures. Clinicians should screen all patients with diabetes to identify those at risk for foot ulceration. This includes reviewing relevant past history, identifying any current foot deformities, and especially assessing for loss of protective sensation with a monofilament. Other helpful screening methods include assessing for peripheral vascular disease by measuring ABIs, ensuring that the patient is wearing appropriate footwear, and checking for high plantar pressure when possible.

Screening allows the clinician to assign the patient to a risk category that dictates both the type and frequency of foot interventions needed. Effective interventions include patient (and clinician) education. Possibly effective interventions include optimizing glycemic control, smoking cessation, intensive podiatric care, and debridement of calluses. The value of prescription footwear for ulcer prevention is unclear. In selected cases, evaluation for surgical procedures may be indicated. Each of these interventions, when used appropriately, may reduce the risk of foot ulceration and its devastating consequences.

Corresponding Author: Nalini Singh, MD, VA Puget Sound Healthcare System, Mailcode: S-111-ENDO, 1660 S Columbian Way, Seattle, WA 98108 (Nalini.Singh2@med.va.gov).

Author Contributions: All of the authors had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis except for the few cases mentioned in the tables.

Study concept and design: Singh, Armstrong, Lipsky.

Acquisition of data: Singh, Lipsky.

Analysis and interpretation of data: Singh, Lipsky.

Drafting of the manuscript: Singh, Armstrong, Lipsky.

Critical revision of the manuscript for important intellectual content: Singh, Armstrong, Lipsky.

Statistical analysis: Singh.

Administrative, technical, or material support: Singh, Armstrong, Lipsky.

Study supervision: Lipsky.

Role of the Sponsor: There was no sponsor for this study and no agency or company reviewed the manuscript.

Acknowledgment: We thank VA Puget Sound Healthcare System employees Ted Hamilton, MLIS, for his invaluable assistance with the literature searches, and Christopher Pacheco for providing the initial version of the monofilament figure. We also thank Edward J. Boyko, MD, MPH, for his time and expertise in calculating measures of effect in the tables.