Reiber GE, Smith DG, Wallace C, Sullivan K, Hayes S, Vath C, Maciejewski ML, Yu O, Heagerty PJ, LeMaster J. Effect of Therapeutic Footwear on Foot Reulceration in Patients With DiabetesA Randomized Controlled Trial. JAMA. 2002;287(19):2552-2558. doi:10.1001/jama.287.19.2552
Author Affiliations: Health Services and Rehabilitation Research and Development, VA Puget Sound Health Care System, Department of Veterans Affairs (Drs Reiber, Wallace, Maciejewski, and Heagerty, Mr Hayes, and Mss Vath and Yu), Departments of Health Services (Drs Reiber and Maciejewski), Epidemiology (Drs Reiber and LeMaster), Orthopaedic Surgery (Dr Smith), Biostatistics (Ms Yu and Dr Heagerty), and Family Medicine (Dr LeMaster), University of Washington, and Joslin Diabetes Center at Swedish Medical Center (Dr Sullivan), Seattle, Wash.
Context Many people with diabetes experience lower-limb ulcers. Footwear has
been implicated as a primary cause of foot ulcers, yet research is limited
on the efficacy of shoe and insert combinations to prevent reulceration.
Objective To determine whether extra-depth and -width therapeutic shoes used with
2 types of inserts reduce reulceration in diabetic individuals with a history
of foot ulcer.
Design, Setting, and Participants Randomized clinical trial of 400 diabetes patients with history of foot
ulcer in 2 Washington State health care organizations who did not require
custom shoes for foot deformity and were enrolled between August 1997 and
December 1998 and followed up for 2 years. Data collected at regular intervals
documented physical, foot, and diabetes characteristics; footwear use; foot
lesions; and ulcers.
Interventions Participants were randomly assigned to receive 3 pairs of therapeutic
shoes and 3 pairs of customized medium-density cork inserts with a neoprene
closed-cell cover (n = 121); to receive 3 pairs of therapeutic shoes and 3
pairs of prefabricated, tapered polyurethane inserts with a brushed nylon
cover (n = 119); or to wear their usual footwear (controls; n = 160).
Main Outcome Measure Foot reulceration, compared among the 3 groups.
Results Two-year cumulative reulceration incidence across the 3 groups was low:
15% in the cork-insert group, 14% in the prefabricated-insert group, and 17%
in controls. In the intent-to-treat analysis, patients assigned to therapeutic
shoes did not have a significantly lower risk of reulceration compared with
controls (risk ratio [RR] for the cork-insert group, 0.88; 95% confidence
interval [CI], 0.51-1.52 and RR the for prefabricated-insert group, 0.85;
95% CI, 0.48-1.48). All ulcer episodes in patients assigned to therapeutic
shoes and 88% wearing nonstudy shoes occurred in patients with foot insensitivity.
Conclusions This study of persons without severe foot deformity does not provide
evidence to support widespread dispensing of therapeutic shoes and inserts
to diabetic patients with a history of foot ulcer. Study shoes and custom
cork or preformed polyurethane inserts conferred no significant ulcer reduction
compared with control footwear. This study suggests that careful attention
to foot care by health care professionals may be more important than therapeutic
footwear but does not negate the possibility that special footwear is beneficial
in persons with diabetes who do not receive such close attention to foot care
by their health care providers or in individuals with severe foot deformities.
Individuals with diabetes incur 67% of all lower-limb amputations in
the United States.1 Approximately half of diabetes-related
amputations have been attributed to poorly fitting footwear initiating a causal
chain leading to foot ulcers and amputations.2- 4
The etiology of diabetic foot ulcers and amputations is well characterized,
but the efficacy of footwear in preventing reulceration has undergone limited
investigation. Edmonds et al3 described follow-up
findings in a tertiary referral population in which 239 ulcer patients were
divided into those with ischemic ulcers (absent foot pulses; n = 91) and neuropathic
ulcers (sensory loss; n = 148). The study described reulceration in 121 patients
in the neuropathic group followed up for an average of 26 months. Among patients
in the neuropathic group who received therapeutic and custom shoes, 26% had
reulceration vs 83% who did not receive this footwear and wore their own footwear.
A randomized clinical trial of a single style of shoe and insert for men and
women was conducted in persons with prior foot ulcer at 2 study sites in Italy.5 After 1 year, reulceration occurred in 28% of the
therapeutic footwear group vs 58% in those wearing their own footwear.
Many questions remain about the efficacy of various shoe and insert
combinations in foot ulcer prevention, even though therapeutic footwear is
a covered Medicare benefit.6 Few studies have
examined specific protective shoe properties.7
We conducted a randomized clinical trial to determine whether provision of
study therapeutic shoes and 2 types of inserts would reduce the incidence
of reulceration in individuals with diabetes. We hypothesized that the group
assigned to therapeutic shoes with custom cork inserts would have the lowest
reulceration rates, followed closely by those assigned to therapeutic shoes
with prefabricated inserts.
Participants were selected from among 21 000 Washington State residents
with diabetes from the VA Puget Sound Health Care System (VA) and Group Health
Cooperative (GHC). Study eligibility criteria were diagnosed diabetes; age
45 to 84 years; men from either the VA or GHC and women from GHC (there were
very few female veterans who could meet the eligibility criteria); history
of a full-thickness foot lesion or foot infection requiring antibiotic treatment;
no foot deformities requiring a custom shoe; shoe size 8 to 12½ for
men and 7 to 10½ for women (sizes were limited to contain manufacturing
costs); ability to walk 1 block and climb 1 flight of stairs per day (to ensure
that the footwear would be worn); and willingness to consent to randomization
and study footwear provisions. Exclusion criteria were prior lower-extremity
amputation of more than 1 digit; presence of either unhealed or healed lesion
in the prior month; requirement of boots, custom shoes, or nontraditional
footwear for daily activities; nonambulatory status; or terminal illness that
would make 2-year survival unlikely. Our footwear could not safely accommodate
persons with severe foot deformities and Charcot feet, so these individuals
also were excluded.
Figure 1 shows that of the
763 patients with diabetes who met the eligibility criteria and were asked
to participate, 400 agreed to be randomized into the study (186 VA men, 123
GHC men, and 91 GHC women). Study participation was declined due to time and
transportation constraints (49%), no interest in research (35%), poor health
(6%), and shoe size outside study range or other reasons (10%). Computer-generated
block randomization was performed according to health care organization and
sex. One hundred twenty-one men and women were randomly assigned to receive
3 pairs of therapeutic shoes and 3 pairs of customized medium-density cork
inserts with a neoprene closed-cell cover; 119 were randomly assigned to receive
3 pairs of therapeutic shoes and 3 pairs of prefabricated, tapered polyurethane
inserts with a brushed nylon cover; and 160 were randomly assigned as controls
to wear their usual footwear. Human subjects approval was granted by the University
of Washington, Seattle, and GHC. Participants provided informed consent and
were enrolled between August 1997 and December 1998 and followed up for 2
years. They continued to receive regular health care and foot care from the
VA or GHC. Participants received no compensation other than the study footwear.
To prevent contamination of the footwear interventions by patient education
or clinical care, no participants received such education or care at the study
The 2 footwear interventions included patient receipt of both study
therapeutic shoes and inserts. Men's and women's dress, casual, and athletic
shoes were designed and manufactured with assistance from Nike Inc (Beaverton,
Ore), Cole-Haan (Livermore, Me), and Lowell Shoe (Hudson, NH) in the most
common shoe sizes specifically for study participants.8,9
Bruce Kilgore of Nike Inc provided our study team with expertise on footwear
materials, design, and manufacturing processes. Our study team designed shoes
and inserts, tested prototypes, and conducted pilot studies of the men's and
women's shoes and inserts prior to the clinical trial. The men's dress, casual,
and athletic shoes were manufactured by Cole-Haan and the 3 women's shoe styles
were manufactured by Lowell Shoe.
The combination shoe last (mold resembling interior shoe shape) provided
extra depth in the toe box and extra width across the metatarsal heads. The
men's and women's outsoles had a 38-in recess that allowed the insert
and foot to drop down into the outsole, providing extra forefoot space without
a "boxy" or orthopedic appearance. The men's outsole for all 3 shoe styles
was a semirockered forefoot made rigid with a lightweight extended composite
shank. Women's shoes were semirockered with a nonextended steel shank and
a uniform, more feminine outsole. Study shoes resembled nontherapeutic shoes,
and materials for uppers were brown, black, and off-white high-quality leather.
A wide range of potential insole materials (eg, polyurethane foams,
viscoelastic polymers) was investigated using evidence of excellent performance.10- 18
Two types of study inserts were designed to fit the interior contour of the
shoes. We chose cork for the customized insert because of its elastic nature
(cork takes very little set or deformation), machinability, and a long history
as an acceptable material for footwear. Custom cork inserts were produced
by scanning the plantar surface of patients' feet using computer-aided design/computer-aided
manufacturing technology. These medium-density, full-length, milled-cork inserts
had a 116-in closed-cell neoprene cover. The prefabricated study insert was
a 58-in noncustomized tapered polyurethane insert with a brushed nylon
cover, designed to be thicker than inserts used in commercial walking shoes.
This insert was not based on the digitized form of the participant's foot.
These 2 medium-density insert materials represented excellent, established
insert technology. A lightweight terry-cloth house slipper (Tru-Stitch Footwear,
Malone, NY) with no internal seams and a textured sole was designed for all
study participants to use to minimize differences in out-of-shoe exposure.
All shoes and inserts for intervention participants were fitted by the same
study pedorthist who manufactured the custom inserts, performed shoe-fitting
adjustments, and replaced footwear based on wear patterns.
Data were collected from patients at baseline and 1 and 2 years on diabetes,
overall health, foot, and functional status. This included peripheral pulses
and foot sensation as measured using a 5.07 Semmes-Weinstein monofilament.
Four visits occurred within 1 month of enrollment to ensure proper footwear
fit in the intervention groups. Thereafter, visits were scheduled every 17
weeks to collect information on physical activity, feet, shoes, inserts, and
the amount of out-of-bed time that participants wore shoes, socks/stockings,
or slippers or went barefoot. Footwear compliance was computed from a physical
activity questionnaire by quantifying the time out of bed that was spent "in
shoes" or "not in shoes."
The main study outcome measure was incidence of a foot ulcer, defined
as a cutaneous erosion extending into or through the dermis to deeper tissue
or other cuts that did not heal within 30 days. The pivotal event (triggering
the sequence of events preceding each lesion or ulcer) was identified by patient
self-report. A footwear-related ulcer was identified from the focused pivotal-event
interviews as patient self-report of a footwear-related incident resulting
in a break in the skin and documentation of either improper break-in of shoes
or a lesion at a site compatible with a footwear-related lesion not healing
within 30 days (verified by photographs and medical records). Ulcer episode
was defined as the interval from ulcer identification to healing. Multiple
ulcers occurring on the same day on the same foot (often the result of minor
trauma) were defined as a single episode. Minor lesions and ulcers were documented,
characterized, and photographed, and patients were referred to their usual
health care practitioners for foot care since foot care was not a study provision.
Health care practitioners were notified in writing of lesions and ulcers.
Patients wearing intervention footwear who experienced foot ulcers resumed
wearing intervention footwear following healing.
Final ulcer classification was determined by a panel of 3 foot care
specialists blinded to study group. They reviewed the descriptions, photographs,
study and medical records, and causal pathway data. They also verified the
assignment of all minor lesions not healing within 30 days, ulcers, and amputations.
Baseline demographic characteristics of participants randomized to each
study group were compared by calculating means and SDs or frequencies and
proportions. Formal comparison was based on analysis of variance for continuous
measurements. An intention-to-treat analysis was performed using all available
follow-up time on all randomized patients to assess the impact of intervention
on the primary end point. Analyses were carried out using SAS software19 and statistical tests were 2-sided, with P<.05 considered statistically significant.
To analyze the binary primary outcome, any vs no ulceration, we used
exact binomial methods to construct confidence intervals (CIs) for the proportion
in each study group. The χ2 test or Fisher exact test was used
to compare across study groups. We also analyzed total ulcers and ulcer episodes.
To compute ulcer rates, we used standard person-time methods and Poisson regression
with robust SEs to allow for possible overdispersion. All comparisons were
between the group assigned to therapeutic shoes with cork inserts and the
control group and between the group assigned to therapeutic shoes with prefabricated
inserts and the control group. In secondary analyses, we excluded person-time
not wearing footwear for medical or personal reasons. Logistic regression
was used to assess the association between covariates and reulceration.
Our power analysis was based on published 2-year reulceration rates
of 72% to 87% for participants using their own footwear and 26% for those
using prescription shoe and insert combinations.3,20
We assumed that 67% of control participants would wear standard footwear and
33% would cross over to some type of custom shoes, insoles, or in-depth shoes.
Over the 2-year study, we anticipated a 20% combined mortality and loss to
follow-up. Therefore, we anticipated 2-year reulceration rates of 30% in the
cork-insert group, 35% in the prefabricated-insert group, and 58% among controls.
Based on these rates, our original sample size was selected to yield 97% power
to detect a difference in the rate for the cork-insert group vs the control
group and to provide 89% power to detect a difference between the prefabricated-insert
group vs the control group.
Baseline participant characteristics are described in Table 1. The average age was 62 years, 23% were female, 78% were
white, and the average education was 14 years. Fifty-eight percent of participants
were insensate to monofilament and 32% had a moderate foot deformity (hallux
limitus, fixed claw toe, or bunion), though significantly fewer patients in
the group assigned to therapeutic shoes with prefabricated inserts had these
foot deformities than among those assigned to therapeutic shoes with cork
inserts or among controls (P = .03).
The time participants were out of bed was divided into "in shoes" and
"not in shoes." Compliance with use of intervention shoes when wearing shoes
was 83% in the group with cork inserts and 86% in the group with prefabricated
inserts. When participants were out of bed and not wearing shoes, they averaged
just under 1 hour per day barefoot, 1 hour per day in stockings, and 3 hours
per day in slippers.
During the 2-year follow-up interval, physicians prescribed custom shoes
for 2.5% of controls and custom inserts for an additional 4.4%. Among controls,
13% purchased therapeutic shoes and an additional 17% purchased over-the-counter
inserts. Over the 2-year follow-up period, 5.5% of study participants died
and 11% withdrew (see Figure 1).
Sixty-two participants had a total of 95 reulcerations occurring in
84 ulcer episodes. Table 2 shows
that the cumulative percentage of participants with foot ulcers during the
study period was 15% with therapeutic shoes and cork inserts, 14% with therapeutic
shoes and prefabricated inserts, and 17% in the control group. There were
no statistically significant differences in ulcers or ulcer episodes between
groups. The risk ratio (RR) for persons with 1 or more reulcerations in the
group with therapeutic shoes and cork inserts compared with those in the control
group was 0.88 (95% CI, 0.51-1.52). The RR comparing the group with prefabricated
inserts with the control group was 0.85 (95% CI, 0.48-1.48). When examining
total ulcer episodes, the RR comparing the group with cork inserts with controls
was 0.86 (95% CI, 0.45-1.63) and the RR comparing the group with prefabricated
inserts with controls was 0.80 (95% CI, 0.41-1.56). In the secondary analysis,
we used logistic regression to control for age, sex, foot deformity, education,
and marital status. The RR for reulceration comparing persons without foot
sensation with those with sensation was 3.68 (95% CI, 1.81-7.49).
Table 3 shows that the 2
pivotal events that contributed most to ulcer episodes were shoes (39%; 10.7%
due to intervention shoes and 28.6% due to nonintervention shoes) and external
trauma (31%). Among the 9 persons who developed ulcers while wearing intervention
shoes, all had foot insensitivity. Among the 24 persons who developed ulcers
while not wearing intervention shoes, 88% had foot insensitivity, and among
the 26 patients whose pivotal event was external trauma, 96% had foot insensitivity.
Amputation was not a study end point because few instances were expected.
Of the 11 amputations that occurred, 6 involved toes and feet and 5 were at
the transtibial or transfemoral level.
The same therapeutic shoes were worn by participants in the 2 intervention
groups, but their inserts differed. Intervention footwear showed no statistically
significant protective effect in ulcer prevention over footwear worn by controls.
Cumulative ulcer incidence in all 3 study groups was low during the 2-year
follow-up. The customized cork inserts with neoprene covers required considerably
more time, equipment, and expense to produce than did the tapered polyurethane
and brushed nylon inserts, which performed similarly but were far less expensive.
Despite low ulcer incidence, footwear-related ulcers occurred. There
are multiple pathways to foot ulcers. In addition to pathways involving footwear,
there are pathways with external trauma when individuals are not wearing shoes.21 Participants' time out of shoes and out of bed averaged
5 hours per day. The slippers we provided to all participants may have influenced
the external trauma pathway. We anticipated no impact of footwear on self-care,
critical ischemia, or decubitus ulcer pathways. Other strategies are needed
to prevent or decrease ulcers from these pathways.
There are other explanations for the low ulcer incidence in our study.
Foot insensitivity was present in 58% of participants at baseline and in 66%
at 2 years. Control participants had a lower but not statistically significant
prevalence of foot insensitivity at baseline. Risk of reulceration was clearly
higher in persons with foot insensitivity. In these patients, the proportions
of ulcers attributed to intervention shoes, nonintervention shoes, and external
trauma were 100%, 88%, and 96% respectively.
Follow-up study visits occurred every 17 weeks, which could have heightened
participant awareness of footwear. In addition, when we identified a lesion
that could progress to an ulcer, human subjects considerations stipulated
that we notify patients and their physicians. Therefore, timely discovery
of minor lesions at clinic visits and physician referral may have kept ulcer
rates low. In addition, our conservative definition of foot ulcer excluded
lesions from external trauma resolving within 30 days. Other investigators
have classified similar lesions as ulcers after shorter intervals.
Our study design assumed and was powered for 33% of control patients
to acquire therapeutic footwear over the 2-year follow-up. After 2 years,
a total of 6.9% of patients received prescriptions and purchased prescribed
shoes and/or custom shoes. An additional 13% of control patients purchased
off-the-shelf therapeutic shoes and 19% purchased over-the-counter inserts.
These crossovers involved a total of 37% of the control population. Footwear
purchases by the control group occurred gradually. Our data suggest that footwear
purchased and worn are not synonymous. However, this footwear crossover may
have attenuated our findings.
Edmonds et al3 reported 26-month reulceration
rates of 26% in patients in therapeutic footwear and 83% in their own footwear.
However, these findings may have been influenced by inclusion of many patients
with severe foot deformities, including uncorrected Charcot feet and prior
toe and ray amputations. Patients with more than 1 toe amputation, Charcot
feet, or a requirement of custom shoes were excluded from our study because
our study shoes could not safely accommodate their custom footwear needs.
Patients in the United Kingdom with very severe foot deformities are routinely
given custom footwear while in the Pacific Northwest, these patients are increasingly
treated with orthopedic foot surgery reconstruction procedures. Differences
also may exist in management of minor toe amputations. In the Pacific Northwest,
when toes 2, 3, and 4 are removed at or proximal to the metatarsophalangeal
joint, leaving the little toe and the great toe, the surgeon often performs
a transverse amputation of all toes or a transmetatarsal amputation rather
than leaving a foot with 1 or 2 toes, which often goes on to deformity and
associated biomechanical difficulties. In the study by Edmonds et al,3 custom footwear, not surgery, was the convention in
treating patients with severe foot deformities.
A study of reulceration was conducted at 2 teaching hospitals in Italy.
In their clinical trial of 69 persons with diabetes at high risk of ulcer,
Uccioli et al5 provided men and women in the
intervention group with 1 style of Velcro-closure shoes plus inserts. They
reported that 1-year ulcer frequency in the intervention group was 28% compared
with 58% in the control group. All participants in our study received 3 pairs
of normal-appearing shoes (dress, casual, and athletic styles). In prestudy
focus groups, these shoes were aesthetically acceptable to our participants.
Over the duration of the study, compliance with our study shoes was 83% to
86%. The Italian study reported intervention footwear compliance as always
frequent or continuous based on participant self-reported compliance on a
4-item scale (infrequent, occasional, frequent, or continuous). Exposure to
minor trauma when not wearing shoes (mean, 5.2 hours per day in our study)
was minimized in our participants by providing specially designed house slippers
with no internal seams to intervention and control participants. The final
comparison between studies is that on an annual basis, the cumulative incidence
of reulceration in all groups in our study was less than one third of that
reported for the intervention group of Uccioli et al.
Accurate classification of foot risk in diabetic populations is difficult.
Based on evidence from population-based and analytic studies, approximately
half of adults with diabetes have 1 or more foot risk conditions (peripheral
neuropathy, lower-extremity arterial disease, or foot deformity).22- 29
Approximately 10% report a prior foot ulcer (similar to our study group),25,26,30 and approximately
5% report a severe foot deformity requiring custom footwear or a prior lower-limb
amputation (these individuals were excluded from our study).29,31
Our study findings can be generalized to those with a prior foot ulcer.
This randomized clinical trial has several limitations. The percentage
of participants with foot insensitivity at baseline was not as high as expected.
Data from this trial and others indicate that participants with foot insensitivity
are at highest risk of ulceration. Lesion identification and referral by the
study staff and follow-up by primary care physicians at the 2 Pacific Northwest
health care organizations may have contributed to our low ulcer rates. We
compared our reulceration rates with recently published diabetic foot ulcer
rates (not necessarily reulceration rates) from other studies conducted in
these 2 organizations and found that their annual incidence was low: 1.9%
at GHC and 3.0% at the VA General Internal Medicine Clinic.29,32
Finally, due to limited resources, we were unable to collect blood for glycated
There are 5 conclusions from this study of persons with prior foot ulcers
but no severe foot deformity. First, while it is a popular belief that therapeutic
shoes and inserts should be dispensed freely to all patients with diabetes,
this study did not provide evidence to support this practice. Second, customized
cork inserts offered no benefit over prefabricated polyurethane inserts, and
control patients wearing their usual footwear fared as well as intervention
patients in therapeutic shoes. Third, this study suggests that careful attention
to foot care by physicians may be more important than therapeutic footwear,
but the study does not negate the possibility that special footwear is beneficial
in persons with diabetes who either do not receive such close attention to
foot care by their physicians or have severe foot deformities. Fourth, strategies
to reduce external trauma unrelated to shoes appear to be indicated, including
safe house slippers. Finally, the highest-risk patients, those who have foot
insensitivity, are also the most vulnerable to complications. Limited health
care resources might be better used to develop a comprehensive approach to
care for these individuals. These patients and their physicians should jointly
explore individualized strategies to decrease the events that give rise to