DIABETICMedicine

DOI: 10.1111/dme.13054

Research: Epidemiology
Association of diabetic foot ulcer and death in a
population-based cohort from the United Kingdom

J. W. Walsh1, O. J. Hoffstad1, M. O. Sullivan2 and D. J. Margolis1

1
Department of Biostatistics and Epidemiology, University of Pennsylvania School of Medicine, Philadelphia, PA, and 2Department of Chemical & Biomolecular
Engineering, University of Delaware, Newark, DE, USA

Accepted 8 December 2015

Abstract
Aims The presence of diabetic foot ulcers is strongly associated with an increased risk of death. In this study, we
investigate whether the effects of diabetes-associated complications can explain the apparent relationship between
diabetic foot ulcers and death.
Methods We analysed data from 414 523 people with diabetes enrolled in practices associated with The Health
Improvement Network in the United Kingdom. Our methods were designed to control for potential confounders in
order to isolate the relationship between diabetic foot ulcers and death. Using proportional hazards models and
the area under the receiver operator curve, we evaluated the effects of diabetic foot ulcers and the covariates on
death.
Results Among the patients, 20 737 developed diabetic foot ulcers; 5.0% of people with new ulcers died within
12 months of their first foot ulcer visit and 42.2% of people with foot ulcers died within 5 years. After controlling for
major known complications of diabetes that might influence mortality, the correlation between diabetic foot ulcers and
death remained strong with a fully adjusted hazard ratio of 2.48 (95% confidence interval: 2.43, 2.54). Geographic
variance existed but was not spatially associated.
Conclusions Diabetic foot ulcers are linked to an increased risk of death. This cannot be explained by other common
risk factors. These results suggest that either there are major unknown risk factors associated with both diabetic foot
ulcers and death, or that diabetic foot ulceration itself is a serious threat, which seems unlikely. A diabetic foot ulcer
should be seen as a major warning sign for mortality, necessitating closer medical follow-up.
Diabet. Med. 33, 1493–1498 (2016)

with a DFU was approximately $35 000 in 2008 [5]. In total,

Introduction
Complications of the foot in individuals with diabetes
include a diverse and severe array of musculoskeletal,
neuropathic and circulatory disorders that impinge upon
the quality of life of a person with diabetes [1,2]. Diabetic
foot ulcers (DFU) can lead to a lower extremity amputation,
which can range in severity from the removal of a toe to the
amputation of the full lower extremity. A DFU is a
particularly debilitating disorder that is often associated
with arterial vasculopathy and/or neuropathy. The preva-
lence of DFU was ~ 28% among US Medicare beneficiaries
with diabetes in 2008, with an incidence of ~ 6% [3,4]. The
overall average cost to Medicare of caring for an individual
Correspondence to: David J. Margolis. E–mail: margo@mail.med.upenn.edu
the care of individuals with diabetes who had foot ulcers cost
Medicare about $25 billion in 2008 [5]. In 2012, in the
United Kingdom (UK), it was estimated that the National
Health Service (NHS) spent between £639 million and
£662 million on foot ulcers and lower extremity amputation,
which was approximately £1 in every £150 spent by the NHS
[6].
A minor traumatic injury to the foot of a person with
diabetes can lead to a chronic wound due to several reasons
including a primary inability of the skin of the foot to heal
(e.g. cutaneous organ failure), neuropathic sensory loss
(which leads to an increased risk of injury and may have a
primary effect on wound repair), the failure of a foot’s
biomechanics, leading to repetitive pressure-induced trauma,
and/or reduced arterial vascular supply to bodily extremities

ª 2015 Diabetes UK 1493

DIABETICMedicine
What’s new?
• This is the largest patient cohort study ever to study the
association between diabetic foot ulcers and risk of
death in adults with diabetes.
• Our findings support the hypothesis that diabetic foot
ulcers are a marker associated with a risk to patients’
longevity independent of other complications of diabetes.
impeding the wound healing process. Even more worrisome
is that the presence of DFU has also been strongly correlated
with an increased risk of death. In the US Medicare
population, 10.7% of individuals with an incident DFU
diagnosed in 2008 died in that year [3]. Death among people
with DFU is often associated with cardiovascular disease and
renal insufficiency, which are themselves major complica-
tions of diabetes [7]. These complications are also common
among people living with DFU, and those who have had an
lower extremity amputation [8–10].
The goal of this study was to better understand the
association of DFU and death. In addition, because many
illnesses like cardiovascular disease and renal failure are
associated with both an increased risk of death and an
increased risk of foot ulcer, we evaluated whether a history
of these illnesses can explain the increased risk of death that
people with foot ulcers face.
Methods
This was a cohort study of people with diabetes from The
Health Improvement Network (THIN) in the UK, which
includes anonymized practice data by general practice
physicians (GPs) from England, Scotland, Wales and North-
ern Ireland. These GPs provide or coordinate all care for
these people. THIN has been used in the past to study
diabetes and chronic wounds. Ethical approval for this study
was obtained from both the Scientific Review Committee in
the UK that oversees THIN and from the Institutional
Review Board at the University of Pennsylvania.
Inclusion criteria
All patients were cared for by a GP whose practice was up to
THIN standards for data collection and contributed infor-
mation to THIN during the years 2003–2012. All patients
had diabetes, at least 1 year of follow-up, and were at least
25 years of age at the time of diagnosis with diabetes. As a
result, this study focuses primarily on those with Type 2 or
adult-onset diabetes. The diabetes diagnosis was based on
previously used coding algorithm(s) [10–13]. To ensure that
all patients had diabetes, they had at least two visits for
diabetes at least 2 months apart. Our primary exposure of
interest was on new onset DFU. This diagnosis was based on
a previously described algorithm [10].
1494
Foot ulcer and death  J. W. Walsh et al.
Additional study variables
The outcome of interest was death. Our risk factors of
interest were selected a priori because they were known to be
associated with the most frequent causes of death in
individuals with diabetes and many are also associated with
an increased risk of DFU. Our risk factor variables included a
history of cardiovascular disease such as a history of
myocardial infarctions, cerebrovascular accident, peripheral
vascular disease–arterial insufficiency, the Charlson Comor-
bidity Index (an index that is used to predict mortality [14]),
and a history of chronic kidney disease as determined by the
eGFR (Table 1) [10,11,14]. The eGFR measurements were
used to group participants by the US National Kidney
Foundation scheme for chronic kidney disease into three
groups based on eGFR of > 60, 30–60 and < 30 ml/min/
1.73 m2, respectively [10]. We also evaluated a history of
malignancy, history of cigarette smoking, gender, HbA1c
levels, categorized as < 42 mmol/ml (< 7%), 42–75 mmol/
ml (7–9%) and > 75 mmol/ml (> 9%), and age over
65 years at the time of diagnosis with diabetes. Finally, in
order to evaluate regional variation within the practices, they
were grouped based on their strategic health authority,
which as a variable was then evaluated both as a risk factor
and as a random effect.
Analysis
All variables were evaluated descriptively with respect to the
full THIN diabetes cohort, just among those with DFU and
among those who had died. We then created explanatory
models to specifically estimate the effect of DFU on the risk
of death using proportional hazards models that allowed the
presence of DFU to time-vary. Time was accrued from the
date of first diagnosis of diabetes until a patient left the
practice, died or until a DFU was diagnosed. In this setting,
when a DFU was diagnosed, time was separately accrued
from the onset of the DFU until the patient left the practice or
died. If data were missing, then for that analysis, the
individual was dropped. In all models, death was our
primary outcome and DFU was our primary exposure. The
risk factor variables were added to the models to determine if
they diminished the association between DFU and mortality
[15]. More specifically, based on our hypothesis, we expected
that when the risk factor variables were added into the
proportional hazard models, the association between DFU
and death would be diminished (i.e. the hazard ratio of DFU
would move towards 1). This method of analysis follows the
conceptual framework of confounding but also of the notion
of a surrogate (i.e. DFU is on the causal pathway to death of
the risk factor variable) developed by Prentice et al., in that a
true surrogate must validate the null hypothesis of no causal
association (i.e. a hazard ratio of 1) between the treatment
and the outcome (in this case, between foot ulcers and death)
[15]. For example, if the causal association between DFU and
ª 2015 Diabetes UK
 Research article DIABETICMedicine

Table 1 Demographic and risk factor variables for the cohort of individuals with diabetes and the subgroups of individuals who had a foot ulcer and
individuals who died presented as per cent or means with SD

Subgroup with foot ulcer Subgroup who died

Full cohort N = 414 523 N = 20 737 N = 77 520

Age > 65 years at diabetes 39.0 49.5* 64.2*

diagnosis (%)
Age at diabetes diagnosis mean (SD) 60.7 (16.7)[60.7 (49.0, 71.0)] 62.4 (17.1)*[64.8 (52.8, 74.9)] 67.9(14.2)*[69.7 (60.0, 77.8)]
[median, (IQR)] (years)
Male (%) 54.8 51.2* 52.7*
HbA1c (%)†
< 42 mmol/ml (< 7%) 49.3 45.3 52.9
42–75 mmol/ml (7–9%) 35.6 35.5 32.7
> 75 mmol/ml (> 9%) 15.1 19.2* 14.4*
Chronic kidney disease (%)
1 88.4 79.4 76.5
2 10.8 18.7 21.2
≥3 0.8 1.9* 2.3*
Charlson Index (SD) 0.8 (1.4) 1.2 (1.6)* 1.6 (1.9)*
Myocardial infarction (%) 10.3 16.1* 19.8*
Cerebrovascular accident (%) 9.2 17.0* 18.3*
Peripheral vascular disease (%) 4.1 15.4* 8.6*
Congestive heart failure (%) 3.4 10.9* 10.8*
Malignancy (%) 12.1 14.2* 24.9*
Cigarette use (%)
Never smoked 42.4 43.2 43.4
Past smoker 41.4 43.3 41.1
Current smoker 16.3 13.5* 15.5*

*P < 0.0001.
†
Sample size for measured HbA1c was 351 347 for the full cohort, 17 878 for the subgroup with a foot ulcer and 66 140 for the subgroup
who died and for HbA1c.

death was actually due to myocardial infarction, which has
been shown to be associated with both, then including
myocardial infarction in a statistical model evaluating the
association between DFU and death should render no
statistical association between DFU and death.
We also reported hazard estimates for each variable with
respect to death (i.e. unadjusted models), for DFU alone with
respect to death, for DFU individually with each covariate
with respect to death (i.e. DFU and myocardial infarction)
and for all covariates together with respect to death (a fully
adjusted model). The fully adjusted hazards models were also
evaluated for their ability to predict death using the area
under the receiver operator curve (AROC). All of our
analyses were conducted using STATA v. 13.1.
Results
From 2003 to 2012, 414 523 people in THIN met the
criteria for inclusion in this study. In the THIN dataset,
20 737 people (5.0%) developed DFU and 77 520 (18.7%)
died. The distributions of the demographic and medical
record risk factor variables are presented in Table 1. Of
those with a new-onset DFU, 8.1% died within 12 months of
the initial GP DFU visit. For comparison, 1.8% of individ-
uals died 12 months after their index visit for diabetes. The
5–year death rate for people with DFU was 42.2%. Death in
the first year after diagnosis of DFU varied by strategic health
authority, with regional rates as low as 2.8% and as high as
12.0% (Fig. 1). The variation by strategic health authority
was statistically significant (P < 0.0001), but it was not
spatially correlated (Moran’s I P-value = 0.80).
Individuals with diabetes and a DFU were three times more
likely to die at any time compared with an individual with
diabetes who did not have a DFU, with DFU yielding an
unadjusted proportional hazards ratio of 3.43 [95% confi-
dence interval (CI) 3.37, 3.50] (Table 2). With respect to
death, only age > 65 years had a larger effect estimate, 5.45
(95% CI 5.37, 5.44) (Table 2). After fully adjusting for all of
the risk factor variables, the hazard ratio of DFU with death
was diminished by ~ 28% to 2.48 (95% CI 2.43, 2.54)
(Table 2). Although this decrease represents a significant
amount of confounding, our expectation was that full
adjustment would have resulted in a hazard ratio of nearly
1. The effect that each risk factor had on the association
between DFU and death is described in Table 2. The risk
factors that had the greatest individual effect on the
association between DFU and death were age > 65 years
and the Charlson Comorbidity Index.
We also evaluated the ability of the risk factors to predict
death in the proportional hazards models. The parameter
that we assessed was the AROC. As a predictor of death,
DFU had an AROC of 0.550 (Table 3). The full model had
an AROC of 0.754. The AROCs for the other covariates are
listed in Table 3. As expected, the best individual predictor
was the Charlson Comorbidity Index with an AROC of
0.694.

ª 2015 Diabetes UK 1495

DIABETICMedicine
FIGURE 1 One-year death rates for people with diabetic foot ulcers by
strategic health authority based on practices in the UK. Key: Percentage
of individuals with a diabetic foot ulcer who died within 12 months of
diagnosis.
We also calculated the hazard ratios for the association
between each risk factor and death just among participants
with a DFU. Compared with the estimates for our full
population, these estimates where generally closer to 1. We
also calculated the hazard ratios for each risk factor with
death in those who did not have a DFU during our period of
observation. Compared with the hazard ratios for our full
population, these hazard ratios were generally higher.
Discussion
Patients with DFUs whose diabetes began after the age of 25
and were treated by GPs in the UK’s THIN network faced
more than triple the risk of death at any given point in time
as compared with people with diabetes without DFU. This
study was conducted with the goal of determining whether
the best-documented illnesses associated with an increased
risk of death in people with diabetes, like cardiovascular
disease, renal failure and kidney disease, fully explain the
higher risk of death that people with DFU face when
compared with their counterparts. Many of these risk factors
are also associated with an increased risk of DFU. More
specifically, although each factor by itself was associated
with death, we did not find that myocardial infarction,
congestive heart failure, peripheral vascular disease–arterial
insufficiency, cerebrovascular accident, chronic kidney dis-
ease or the Charlson Comorbidity Index could explain the
high association between DFU and death. Furthermore,
when placed together in a fully adjusted model, these
1496
Foot ulcer and death  J. W. Walsh et al.
complications of diabetes explained < 30% of the associa-
tion between DFU and death [16].
Many studies have examined factors influencing outcomes
for people with DFU, but few have directly asked why DFU is
associated with poor outcomes. In 1996, Boyko et al.
evaluated a cohort of 745 US Veterans’ Administration
hospital patients with diabetes seeking care in their health
system [17]. They found a more than twofold increased risk
of death in those with DFU compared with those with
diabetes alone [17]. They also were not able to remove this
effect by adjusting for cardiovascular disease and other risk
factors of death associated with DFU [17]. Boyko et al.’s
cohort was made up of US military veterans and was 98%
male [17]. A 2003 study in Liverpool, UK found that DFU
increased risk of death independent of factors associated with
DFU except for age, and reported a 5–year death rate of 44%
[18]. This study had a small sample size of just 185
individuals, and had no DFU-free control group [18]. The
Nord-Trøndelag Health Study in Norway found that over
10 years, 49% of participants with DFU died compared with
35.2% of people with diabetes without DFU, and found that
confounding factors like cardiovascular disease and age
could only partially explain this [19]. A 2014 cohort study in
Portugal of 644 patients from a diabetic foot clinic followed
for 3 years concluded that DFU increased morbidity and
mortality, independent of other complications of diabetes
[20]. Another 2014 study, in the UK, found that people with
DFU died 3 years younger on average than those without
DFU [21]. A 2015 systematic review found a 5-year
mortality rate among people with DFU of ~ 40%, but said
little about DFU’s independence as a risk factor [22]. A 2012
meta-analysis confirmed that people with DFU face an
increased risk of death, and suggested that this risk might
be partially due to the elevated burden of cardiovascular
disease associated with DFU [23]. However, excessive
heterogeneity of the studies combined in this meta-analysis
may have compromised their ability to generate a reliable
meta-estimate. Our current study is the largest conducted to
date and the first using a population-based cohort to examine
the DFU association with death.
This study has several potential limitations. It is possible
that our coding for DFU excluded some people. This form of
bias would have likely biased our results to the null (i.e.,
towards one). It is also possible that an individual might have
developed a DFU and had it resolve before seeking care. It is
therefore important to realize that our study is of people
seeking care for a DFU. We also excluded individuals who
developed diabetes prior to the age of 25 and therefore our
study likely focuses on those with Type 2 or adult-onset
diabetes. Coding for chronic wounds in THIN varies, but
previous authors have validated chronic wound codes [11–
13]. It is also possible that THIN GPs are underdiagnosing
complications of diabetes like cardiovascular disease and
peripheral vascular disease, but regarding potential informa-
tion bias, THIN has been shown to be accurate in the past.
ª 2015 Diabetes UK
 Research article DIABETICMedicine

Table 2 Unadjusted and fully adjusted hazard ratios (HR)

Fully adjusted Just those with an ulcer Just those without an

Full (unadjusted) HR HR (95% CI) (unadjusted) (95% CI) ulcer (unadjusted)
Covariate (95% CI) N = 414 523 N = 351 347 N = 20 737 (95% CI) N = 77 520

Diabetic foot ulcer 3.43 (3.37, 3.50) 2.48 (2.43, 2.54) n/a n/a
Age > 65 years 5.45 (5.37, 5.44) 3.99 (3.91, 4.07) 1.85 (1.79, 1.92) 6.68 (6.56, 6.80)
Chromic kidney disease
Stage 1 Ref Ref Ref Ref
Stage 2 2.29 (2.25, 2.33) 1.32 (1.29, 1.34) 1.65 (1.59, 1.72) 2.28 (2.24, 2.32)
Stage 3 or higher 3.01 (2.87, 3.15) 2.00 (1.90, 2.12) 2.41 (2.18, 2.66) 2.82 (2.68, 2.98)
Charlson index 1.29 (1.29, 1.30) 1.17 (1.16, 1.17) 1.16 (1.15, 1.17) 1.31 (1.30, 1.31)
Myocardial infarction 1.92 (1.88, 1.95) 1.46 (1.43, 1.49) 1.65 (1.58, 1.72) 1.89 (1.85, 1.93)
Cerebrovascular accident 1.98 (1.95, 2.02) 1.33 (1.30, 1.36) 1.55 (1.48, 1.61) 1.96 (1.92, 2.00)
Peripheral arterial disease 1.76 (1.71, 1.80) 1.24 (1.21, 1.28) 1.64 (1.57, 1.72) 1.39 (1.35, 1.43)
Congestive heart failure 3.37 (3.29, 3.44) 1.35 (1.31, 1.39) 2.37 (2.26, 2.49) 3.20 (3.12, 3.29)
Cigarette use
Never smoked Ref Ref Ref Ref
Past smoker 0.96 (0.95, 0.98) 0.91 (0.89, 0.92) 0.79 (0.76, 0.83) 0.98 (0.96, 1.00)
Current smoker 1.06 (1.04, 1.09) 1.34 (1.31, 1.38) 0.87 (0.82, 0.92) 1.11 (1.09, 1.14)
HbA1c
< 42 mmol/ml (< 7%) Ref Ref Ref Ref
42–75 mmol/ml (7–9%) 0.59 (0.58, 0.60) 0.80 (0.79, 0.82) 0.95 (0.91, 0.99) 0.50 (0.49, 0.52)
> 75 mmol/ml (>9%) 0.57 (0.56, 0.58) 0.90 (0.88, 0.92) 0.99 (0.93, 1.04) 0.54 (0.53, 0.55)
Malignancy 2.42 (2.38, 2.46) 1.32 (1.29, 1.35) 1.37 (1.31, 1.43) 2.61 (2.56, 2.65)
Gender (Ref: Female) 0.95 (0.93, 0.96) 1.17 (1.15, 1.19) 0.96 (0.92, 0.99) 0.96 (0.94, 0.97)

Unadjusted hazard ratios are also reported for the subcohorts of individuals who had a diabetic foot ulcer and for those who did not. Ref,
reference category.

THIN medical record. In addition, our adjusted models

Table 3 Hazard ratio for diabetic foot ulcer after adjustment for the
listed covariate. Area under the receiver operator curve (AROC) for included the Charlson Comorbidity Index, which is com-
these two variable models is also listed. As a frame of reference, the monly used to predict mortality. Generalizability is always
hazard ratio for the unadjusted model containing the diabetic foot ulcer a concern in human studies. It is possilbe that our results
covariate is 3.43 (3.37, 3.50) (Table 1) and the area under the curve
(AUC) for the fully adjusted model (as described in Table 1) is 0.754 may not generalize, but it is also important to realize that
our study is very large and, within the UK, THIN is
Covariate Full dataset AROC thought to be representative of the UK general population.
Furthermore, strategic health authorities are no longer the
Diabetic foot ulcer Ref 0.550 administrative unit for the UK healthcare system, so
Age > 65 years 2.55 (2.50, 2.60) 0.672
although this analysis provides information about geo-
Chronic kidney disease 3.22 (3.16, 3.28) 0.603
Charlson Index 3.09 (3.03, 3.15) 0.694 graphical interactions with care in the current study, it
Myocardial infarction 3.36 (3.30, 3.43) 0.593 may not generalize to current administrative groups in the
Cerebrovascular accident 3.31 (3.25, 3.37) 0.590
UK.
Peripheral arterial disease 3.31 (3.25, 3.37) 0.562
Congestive heart failure 3.20 (3.14, 3.26) 0.580 In conclusion, people with DFU are at higher risk of dying
Cigarette use 3.48 (3.41, 3.55) 0.552 at any given time than are those who have diabetes and no
HbA1c 3.57 (3.50, 3.65) 0.575
DFU. The association between DFU and death cannot be
Malignancy 3.32 (3.25, 3.38) 0.615
Gender (Ref: Female) 3.43 (3.37, 3.50) 0.555 explained by other known major complications of diabetes
that are also associated with an increased risk of death. It is
Ref, reference category. very unlikely that a foot ulcer by itself is the cause of the
increased risk of death. It is likely that additional factors
We adjusted for factors that were available in THIN. Not exist. It is also possible that a DFU is a marker of increased
all important factors were available such as the duration of medical frailty necessitating increased healthcare provider
diabetes. An example of this concern may have been vigilance in the care of the patient. An obvious interpretation
demonstrated in that elevated HbA1c compared with those of the results of this study would be for healthcare providers
in the ‘normal range’ decreases the risk of death. However, to carefully evaluate all individuals with DFU in order to
it is important to realize that previous reports have shown assure that their care has been maximized. The presence of a
that tight glycaemic control can be associated with DFU should be viewed as a major warning sign of an
increased mortality. In addition, we used a proportional increased risk of death, necessitating closer medical follow-
hazards approach and, in this statistical technique, time up. It is also important that future studies focus on
was based on the date of first diagnsosis of diabetes in the identifying and evaluating additional measures of medical

ª 2015 Diabetes UK 1497

DIABETICMedicine
illness that are not usually captured during routine care in
medical records databases.
Funding source
None.
Competing interests
None declared.
Acknowledgments
Use of THIN was through agreement with the Clinical and
Translational Research Award from the National Institute of
Health at the University of Pennsylvania.
References
1 Jeffcoate WJ, van Houtum WH. Amputation as a marker of quality
of care. Diabetologia 2004; 47: 2051–2058.
2 Jeffcoate WJ, Chipchase SY, Ince P, Game FL. Assessing the
outcome of management of diabetic foot ulcers using ulcer-relaed
and person-related measures. Diabetes Care 2006; 29: 1784–
1787.
3 Margolis D, Malay DS, Hoffstad OJ, Leonard CE, MaCurdy T, de
Lopez Nava K et al. Incidence of Diabetic Foot Ulcer and Lower
Extremity Amputation Among Medicare Beneficiaries, 2006 to
2008. Rockville, MD: Agency for Healthcare Research and Quality,
2010.
4 Margolis D, Malay DS, Hoffstad OJ, Leonard CE, MaCurdy T, de
Lopez Nava K et al. Prevalence of Diabetes, Diabetic Foot Ulcer,
and Lower Extremity Amputation Among Medicare Beneficiaries,
2006 to 2008. Rockville, MD: Agency for Healthcare Research and
Quality, 2010.
5 Margolis D, Malay DS, Hoffstad OJ, Leonard CE, MaCurdy T, de
Lopez Nava K et al. Economic Burden of Diabetic Foot Ulcers and
Amputations Among Medicare Beneficiaries, 2006 to 2008. Rock-
ville, MD: Agency for Healthcare Research and Quality, 2010.
6 Diabetes UK. Postion statement: Putting feet first: Diabets UK
postion on preventing amputations and improving foot care for
people with diabetes. 2014. Available at https://www.dia-
betes.org.uk/Documents/Position%20statements/Putting%20feet%
20first%20position%20statement%20Nov%202014.pdf Last
accessed 17 July 2015.
7 Boyko EJ. Progress in the estimation of mortality due to diabetes.
Diabetes Care 2005; 28: 2320–2321.
8 Armstrong DG, Kanda VA, Marston W, Mills JL, Boulton AJM.
Mind the gap: disparity between research funding and costs of care
for diabetic foot ulcers. Diabetes Care 2013; 36: 1815–1817.
1498
Foot ulcer and death  J. W. Walsh et al.
9 Boulton AJ, Vileikyte L, Ragnarson-Tennvall G, Apelqvist J. The
global burden of diabetic foot disease. Lancet 2005; 366(9498):
1719–1724.
10 Margolis DJ, Hofstad O, Feldman HI. Association between renal
failure and foot ulcer or lower-extremity amputation in patients
with diabetes. Diabetes Care 2008; 31: 1331–1336.
11 Gelfand JM, Neimann AL, Shin DB, Wang X, Margolis DJ, Troxel
AB et al. Risk of myocardial infarction in patients with psoriasis.
JAMA 2006; 296: 1735–1741.
12 Margolis DJ, Hoffstad O, Strom BL, Margolis DJ, Hoffstad O,
Strom BL. Association between serious ischemic cardiac outcomes
and medications used to treat diabetes. Pharmacoepidem DR S
2008; 17: 753–759.
13 Azfar RS, Seminara NM, Shin DB, Troxel AD, Margolis DJ,
Gelfand JM. Increased risk of diabets mellitus and likelihood of
receiving diabetes mellitus treatment in patients with psoriasis.
Arch Dermatol 2012; 148: 995–1000.
14 Quan H, Li B, Couris CM, Fushimi K, Graham P, Hider P et al.
Updating and validating the Charlson comorbidity index and
scaore for risk adjustment in hospital discharge abstracts using data
from 6 countries. Am J Epidemiol 2011; 173: 676–682.
15 Prentice RL. Surrogate endpoints in clincal trials:definitions and
operational criteria. Stat Med 1989; 8: 431–440.
16 Maldonado G, Greenland S. Simulation study of confounder-
selection strategies. Am J Epidemiol 1993; 138: 923–936.
17 Boyko EJ, Ahroni JH, Smith DG, Davignon D. Increased
mortality associated with diabetic foot ulcer. Diabet Med 1996;
13: 957–972.
18 Moulik PK, Mtonga R, Gill GV. Amputation and mortality in new-
onset diabetic foot ulcers stratified by etiology. Diabetes Care
2003; 26: 491–494.
19 Iversen MM, Tell GS, Riise T, Hanestad BR, Ostbye T, Graue M
et al. History of foot ulcer increases mortality risk among individ-
uals with diabetes. Diabetes Care 2009; 32: 2193–2199.
20 Martins-Mendes D, Monteiro-Soares M, Boyko EJ, Ribeiro M,
Barata P, Lima J et al. The independent contribution of diabetic
foot ulcer on lower extremity amputation and mortality risk. J
Diabetes Complicat 2014; 28: 632–638.
21 Brownrigg JRW, Davey J, Holt PJ, Davis WA, Thompson MM,
Ray KK et al. The association of ulceration of the foot with
cardiovascular and all-cause mortality in patients with diabetes: a
meta-analysis. Diabetologia 2012; 55: 2906–2912.
22 Jupiter DJ, Thorud JC, Cuckley CJ, Shibuya N. The impact of foot
ulceration and amputation on mortality in diabetic patients. From
ulceration to death, a systemic review. Int Wound J 2016; 13: 892–
903.
23 Brownrigg JRW, Griffin M, Hughes CO, Jones KG, Patel N,
Thompson MM et al. Influence of foot ulceration on cause-specific
mortality in patients with diabetes mellitus. J Vasc Surg 2014; 60:
982–986.
ª 2015 Diabetes UK