diabetes = 2990 Walters et al.23 Cohort, 10 UK + + Absent light touch + Absent 0 (Logistic general practices + Impaired pain pulses 0 20:30 regression) type 1 and 2 Perception Doppler 7 patients = 1077 0 VPT Litzelman RCT Type 2 0 + Monofilament 0 0 + Exclusion et al.49 patients = 352 criteria (GEE) Kumar et al.24 Cross-sectional UK + + NDS + 0 0 + Char Count= 0
(Logistic general practices regression) type 2 patients = 811 Carrington Cohort, single UK 0 + Motor neuropathy 0 Exclusion 0 0 Exclusion et al.39 (Cox diabetes clinic VPT 0 Pressure criteria criteria regression attendees sensation 0 Thermal analysis) Non-diabetes = 22 (All were in the Type 1 = 83 Type model 2 = 86 simultaneously)
(Continued )
JWBK089-01
Table 1.4 (Continued )
Neuropathy (monofilament, Low AAI, reflex, vibration or TcpO2 JWBK089-Boulton
Study (type Study design, Long DM neurological or absent High History of analysis) diabetes type duration summary score) pulses HbA1c Deformity Smoking of ulcer Amputation
Abbott et al.25 RCT, patients with 0 0 Monofilament Exclusion Exclusion Exclusion (Cox VPT ≥ 25 V + VPT + Reflex criteria criteria criteria regression (United States, April 21, 2006
analysis) United Kingdom, Canada) Type 1 = 255 Type 2 = 780 20:30 Boyko et al.30 Cohort, veterans 0 + Monofilament + AAI 0 + Charcot 0 + + (Cox Type 1 = 48 Type +TcpO2 8 regression 2 = 701 analysis) Kastenbauer Cohort Type 2 = 187 0 0 Monofilament Exclusion 0 0 0 Exclusion Exclusion et al.37 criteria criteria criteria Char Count= 0 + VPT (Logistic regression) Abbott et al.27 Cohort, United 0 0 VPT + + 0 + (Cox Kingdom + Monofilament regression registered diabetes + NDS analysis) patients from six + Reflex UK health districts Type 1 or 2 = 6613
Blank, not studied; +, statistically significant finding; 0, no statistically significant finding; AAI, ankle, arm index; DM, diabetes mellitus; HbA1c, haemoglobin A1c ; RCT, randomised controlled trial; TcpO2 , transcutaneous oxygen tension; VPT, vibration perception threshold; NDS, neuropathy disability score.
JWBK089-01 JWBK089-Boulton April 21, 2006 20:30 Char Count= 0
RISK FACTORS FOR DIABETIC FOOT ULCERS 9
both feet (at the dorsalis pedis or posterior tibial arteries) predicted increased foot ulcer risk, after controlling for neuropathy measures, history of prior foot ulcer and foot deformity.27 The proportion of foot ulcers with both neuropathy and ischaemia is fairly consistent inter- nationally.
In the United Kingdom, Kumar et al. found neuropathy alone in 46% of those with a history of foot ulcer, ischaemia alone in 12%, neuropathy and ischaemia (neuroischaemia) in 30% and neither in 12%.24 Similarly, Walters et al., also working in the United Kingdom, and Nyamu et al. in a carefully conducted clinic-based study in Kenya found that the greatest pro- portion of foot ulcers were neuropathic, followed by neuroischaemic and lastly by ischaemic ulcers, and that ischaemia was present in about half the ulcers overall.23,40 On the other hand, the proportion of foot ulcers that are ischaemic is less in some lesser developed countries, compared to more developed countries.
Morbach et al., in a study comparing foot ulcers clas- sified by the Wagner system across several countries, found that peripheral vascular disease was present in 48% of foot ulcers in Germany, but in only 11% of ulcers in Tanzania and 10% in India.51 The risk associated with a prior history of ulcers was assessed in five studies.
Boyko et al., Litzelman et al. and Abbott et al. reported that a prior history of foot ulcers significantly increased the likelihood of a subsequent ulcer.27,30,49 Kumar et al. reported a relationship between prior amputation and subsequent ulcer.24 Rith-Najarian and colleagues (though they did not use a multivariate approach and hence their study is not shown in Table 1.4) found an incidence rate of foot ulcers of 6/1000 person-years at risk among people with diabetes who had no history of prior foot ulcer, intact foot sensation and no foot deformity, whereas the rate was 330/1000 person-years at risk if all three of these criteria were present.22 Health care access and availability of diabetes education have been reported to influence development of foot ulcers.
In a randomised trial conducted by Litzelman et al. in a county hospital population, diabetes patients were randomised to education, behavioural contracts and reminders, while concurrently their providers received special education and chart prompts.
The control population in this study received usual care and education.
After 1 year, patients in the intervention group reported more appropriate foot self-care behaviours, including inspection of feet and shoes, washing of feet and drying between toes.
Not all desirable behaviours were adopted.
There was no significant difference between patient groups in testing of bath water temperature and reporting of foot problems.
Patients in the intervention group developed fewer serious foot lesions including ulcers than did those in the control group.49 Among the five studies reported in Table 1.4 that included glycosylated haemoglobin (HbA1c , indicating medium-term glycaemic control) in their analyses,16,30,37,39,49 only HbA1c or blood glucose levels was positively associated and foot ulcers: Moss et al. found a statistically significant association between increasingly poor HbA1c and subsequent foot ulcers in their cohort study, with an odds ratio of 1.6 (95% CI l.3, 2.0) for every 2% deterioration in this measure.16 The relationship between smoking and foot ulcers was assessed in six studies reported in Table 1.4;23,24,27,30,37,39 however, it was only of borderline significance in the younger population in the Wisconsin study.16 Moss and colleagues found that current smokers younger than 30 years were more likely to ulcerate, with odds ratio 2.3 (95% CI 1.0, 5.6).16 Kastenbauer and colleagues found that daily intake of alcohol also increased ulcer risk.37 The cohort study by Boyko et al. also identified higher body weight, insulin use and history of poor vision as three additional independent predictors of foot ulcer.30 Four of the studies reported in Table 1.4 address the relationship between deformity and subsequent foot ulcer.27,30,37,49 The study by Boyko et al. found an independent association between Charcot deformity and foot ulcer, but other foot deformities were not independent ulcer