NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

School of Health and Related Research (ScHARR), University of Sheffield. Clinical Guidelines for Type 2 Diabetes: Prevention and Management of Foot Problems [Internet]. Sheffield (UK): University of Sheffield; 2003. (NICE Clinical Guidelines, No. 10.)

  • This publication is provided for historical reference only and the information may be out of date.

This publication is provided for historical reference only and the information may be out of date.

Cover of Clinical Guidelines for Type 2 Diabetes

Clinical Guidelines for Type 2 Diabetes: Prevention and Management of Foot Problems [Internet].

Show details

5Foot examination and risk classification

Foot examination and monitoring

Recommendations

Regular (at least annual) visual inspection of patients’ feet, assessment of foot sensation, and palpation of foot pulses by trained personnel is important for the detection of risk factors for ulceration. (A)

Examination of patients’ feet should include:

testing of foot sensation using a 10 g monofilament or vibration (using biothesiometer or calibrated tuning fork) (A)

palpation of foot pulses (A)

inspection for any foot deformity

inspection of footwear (A)

Monofilaments should not be used to test more than ten patients in one session and should be left for at least 24 hours to ‘recover’ (buckling strength) between sessions. (C)

Classify foot risk as:

low current risk (normal sensation, palpable pulses)

at increased risk (neuropathy or absent pulses or other risk factor)

at high risk (neuropathy or absent pulses + deformity or skin changes or previous ulcer)

ulcerated foot (C)

Self-monitoring and inspection of feet by people with diabetes should be encouraged. (D)

Evidence statements

Evaluation of skin, soft tissue, musculoskeletal, vascular, and neurological condition on an annual basis is important for the detection of feet at raised risk of ulceration. (Ib)

Both vibration perception threshold measurement using a biothesiometer and sensation threshold measurement using a 10 gram monofilament accurately predict neuropathic patients at raised risk of ulceration. The 10 gram monofilament is convenient and easy to use. (III

Longevity and recovery testing suggests that each monofilament will survive usage on approximately 10 patients before needing a recovery time of 24 hours (to restore buckling strength) before further use. (III)

Identification of neuropathy based on insensitivity to a 10 gram monofilament is convenient and appears cost-effective. (III)

Comparisons of different methods of measuring neuropathy reveal no dominant technique. (III)

Self-monitoring and inspection of feet by people with diabetes, when it can be done, is a useful practice to help identify any problems as early as possible. (IV)

Introduction

Fundamental foot care in people with diabetes involves adequate monitoring and the opportunity to reinforce messages of self-care (Boulton et al, 1998). Many people with diabetes are unable to perform this monitoring because of poor eyesight and reduced mobility, making it difficult to inspect their feet (Thomson et al, 1992; Masson et al, 1989), thus regular contact between professionals and patients is important (Edmonds et al, 1996).

Regular foot inspection by a health care professional should include visual inspection of the legs, dorsal, plantar and posterior surfaces of the foot and between the toes. Vascular examinations involve palpation of the pulses in the lower extremities and inspection of the feet and legs for evidence of ischaemic changes. Musculoskeletal evaluation includes foot and ankle range of motion inspection for bone abnormalities and analysis of gait and stance. Neurological examinations may include tests for vibration by a biothesiometer, tuning fork (only calibrated tuning forks should be used), or cutaneous pressure sensation using a 10 gram monofilament. The ability to identify feet at risk of ulceration has been demonstrated prospectively in studies.

Evidence

Pham et al (2000) conducted a prospective study of 248 patients from three large foot centres in the USA. As well as a complete history and physical examination patients were assessed for abnormalities associated with neuropathy, peripheral vascular disease and foot deformity. Neuropathic symptoms were assessed with a modified Neuropathy Symptom Score and its severity with the Neuropathy Disability Score. Other measurements included the vibration perception threshold, cutaneous perception using Semmes Weinstein monofilaments, plantar foot pressure, joint mobility, and peripheral vascular disease. During a 30 month follow-up, 73 patients (29%) developed foot ulcers, on 95 feet (19%). Multivariate logistic regression analysis for risk of ulceration identified high neuropathy disability scores (5), vibration perception thresholds (≥25V), Semmes-Weinstein monofilaments (≥5.07) and foot pressures (≥6 kg/cm2). Sensitivity and specificity for these factors, either singly or in pairs, ranged from 58–99% and 28–78% respectively with the usual trade-off between sensitivity and specificity. The most sensitive variable (99%) was a high neuropathy disability score and/or Semmes Weinstein monofilament, whilst the most specific variable (78%) was a high neuropathy disability score and/or foot pressure. Sex, race, duration of diabetes, history of previous foot ulceration, and palpable pulses were not significant predictors.

Pacaud et al (1999) looked for clinical indicators of neuropathy in a prospective study of 160 Canadian outpatients, 46 with Type 1 and 113 with Type 2 diabetes, attending a diabetes clinic. Patients completed a questionnaire that asked about diabetes duration, history of foot examinations and problems, and other related conditions (retinopathy and nephropathy) and their follow-up,. Patients also underwent bilateral foot examinations to measure their vibration perception threshold (Vibraton II) and any neuropathy, using Semmes-Weinstein monofilaments. Patients had a mean age of 57 years but Type 1 patients were younger on average at 43 years than Type 2 at 62 years. Duration of diabetes was 16 years on average in the total group, but 22 years in Type 1, compared with only 13 years in Type 2 patients. Moderate to severe neuropathy, as assessed by vibration perception threshold (VPT) was present in 56.4% of the patients studied, with more Type 2 (63.4%) than Type 1 (38.7%) (p<0.05) patients affected. For the Type 2 patients, 43% had moderate to severe neuropathy within the first five years from diagnosis, and the proportion increased with longer duration of the disease. The VPT results correlated well (r=0.86, p<0.05) with the monofilament results, and using the VPT as a standard, the monofilament technique was 97% sensitive and 89% specific for neuropathy. However presence of symptoms was a poor indicator of neuropathy, as neuropathy was present in 44% of patients who had no symptoms. Presence of retinopathy was also neither very sensitive (45%) nor specific (61%) as a test for neuropathy. Foot examination was also not be a good identifier of neuropathy as the authors found that 40% of the patients had never had their feet examined, whilst only 43% of Type 1 and 37% of Type 2 patients had in the preceding year. There was no difference in presence of moderate or severe neuropathy between those who had been examined (53.3%) and those who had not (60.3%) (p>0.05).

A prospective study in a Native American Indian population, categorised patients by visual deformity (hallux, varus or valgus, claw and hammer toes, bony prominence or Charcot osteoarthropathy), by using a 10g monofilament, and by patient history of ulceration or amputation (Rith-Najarian, 1992). Sensation status was determined by applying the filament to 8 points on the plantar surface of each foot of blindfolded patients. Patients who failed to sense the filament at one or more locations were retested twice before being classified as insensate. Patients were followed for 32 months or until first ulcer or amputation if sooner. Combining categories 1–3 provides a test sensitivity of 93% and specificity of 86% for predicting ulceration, with Likelihood Ratios, LR (positive test) = 5.2 (95%CI: 4.0 to 6.7); LR (negative test) = 0.12 (95%CI: 0.05 to 0.27). An ulcer was about 5 times as likely to occur in a patient with a history of disease or lack of sensation than in a patient without these factors.

Categorypatientsdiedulcersamputation*
0: Sensate, no history of disease*26674 (1.5%)0 (0%)
1: Insensate3017 (23.3%)0 (0%)
2: Insensate with deformity1636 (37.5%)1 (6.3%)
3: History of disease*46824 (52.2%)13 (28.3%)
*

In lower extremity

Paisley et al (2002) tested the Neuropen (Owen-Mumford Ltd), a device to assess both pain and pressure perception against current clinical tests, the neuropathy disability score and vibration perception threshold, to evaluate peripheral nerve function in 124 patients with Type 1 or Type 2 diabetes attending either a diabetes centre or a high risk diabetes foot clinic, in Manchester, England. The Neuropen has an interchangeable 10g monofilament for cutaneous pressure assessment and a calibrated tip for assessing pain sensation, both contained within a plastic pen-like hand held device. The sensitivity and specificity of the Neuropen to identify patients with moderate to severe neuropathy were calculated against neuropathy disability score abnormal cut-off values defined as 6/10, and for a vibration perception threshold of >25V. Sensitivity ranged from 82–92% in detecting abnormal neuropathy disability scores and 82–96% in detecting abnormal vibration perception thresholds. Specificities were 31–68% and 32–71% respectively. The best sensitivities could be obtained with either the neurotip or the monofilaments but the highest specificities were obtained with a combination of both the neurotip and the monofilaments. The Neuropen therefore may provide an inexpensive alternative screening method for identifying those patients with diabetes and moderate to severe neuropathy. In this study the mean age (SD) of the patients was 55.4 (13.7), 84 were male, 40 female, 34 had Type 1 and 80 Type 2 diabetes, and duration of diabetes was, on average, 10 years (range 0.5–43 years), and 18% had a history of foot ulcers. The authors provided no details of ethnicity of the population studied.

A study, in an outpatient clinic in Liverpool, examined the reproducibility of screening using a monofilament, biothesiometer and palpitation of pedal pulses (Klenerman et al, 1996). The investigators reported that only the monofilament gave adequately reproducible results (over 85%) for measurements repeated after 2 weeks.

Booth et al (2000) investigated the accuracy of 10g monofilaments. They tested 4 types of monofilament, a total of 160, with results for 158 being reported. Each monofilament was subjected to 10 mechanical bucklings of 10 mm while the load cell detected the maximum buckling force. They also investigated longevity of the monofilaments. They found that most monofilaments produced by Owen Mumford and Bailey Instruments were buckling within 0.5 g of 10 g of buckling force, with the remaining monofilaments from these batches falling within 1.0 g of deviation (mean buckling forces were 10.1 ± 0.4 g for the Owen Mumford filaments and 9.7 ± 0.4 g for the Bailey Instruments monofilaments). Significantly fewer (76%) of the Semmes-Weinstein monofilaments manufactured by North Coast Medical buckled within 1.0 g of 10 g of force (p<0.001). These filaments generally exhibited a negative deviation with most buckling forces falling to <10 g (8.6 ± 0.3 g). A similar pattern emerged with Timesco/Sensory Testing Systems monofilaments (8.1 ± 0.5 g). In terms of longevity, performed on Bailey Instruments and Owen Mumford monofilaments, most monofilaments remained within 10% of 10 g of buckling force after 100 continuous compressions, but by 200 compressions, only 50% of monofilaments were within this range. They indicated that their findings implied that centres using North Coast Medical or Timesco/Sensory Testing Systems monofilaments may be producing far too many false-positive results (i.e., they may be categorising patients as insensate who are in fact sensate).

All monofilaments (except the Timesco/Sensory Testing Systems monofilaments) have the Conformity European (CE) mark awarded by the Medical Devices Agency. They concluded that longevity and recovery testing suggests that each monofilament will survive usage on ~10 patients before needing a recovery time of 24 h before further use.

Kumar and colleagues (1991) commented that filaments were easy to use, light (150g) and cheap (£12/set) when compared to a biothesiometer weighing 2.5kg, requiring a power source and costing £400. Considering the findings of the available prospective studies and relative performance in head-to-head studies with surrogate endpoints it is likely that monofilaments provide a portable and cost-effective alternative in first-line monitoring for neuropathy.

A prospective study, in a diabetes centre and foot clinic in Manchester, used a biothesiometer and categorised patients, without previous ulceration or significant ischaemia, according to their vibration perception threshold (VPT) at enrolment (Young et al, 1994). The endpoint was the first ulcer regardless of cause, and results at the end of the study are tabulated below, after (up to) 4 years follow-up. Using a threshold of >25V, these results indicate a test sensitivity of 83%, and a specificity of 62%, with Likelihood Ratios, LR (positive test) = 2.2 (95%CI: 1.8 to 2.5); LR (negative test) = 0.27 (95%CI: 0.14 to 0.48)

VPT (Voltage)patientsulcersincidence
<15V20963.0%
16–24V5823.4%
>25V2024019.8%*
*

p<0.01, group 3 vs. group 1

Abbott et al (1998) report the incidence of foot ulcer after 1 year in a prospective study of 1033 patients with established neuropathy (VPT25 V on at least 1 foot and VPT30V on both feet). Patients were originally enrolled in a randomised control trial in which active treatment showed no benefit, and so both treatment and placebo groups are analysed together. For each 1 unit increase in VPT at study baseline, the hazard of first ulcer increased by 5.6%. Each unit increase in the muscle and reflex components of the Michigan diabetic polyneuropathy (DPN) score increased the hazard of first ulcer by 5.0%.

A number of other techniques are available for assessing neuropathic deficit: e.g. the tactile circumferential discriminator (TCD) (Vileikyte et al, 1997); the graduated tuning fork (Thivolet et al, 1990; Liniger et al, 1990b); thermal discrimination devices and others (Liniger et al, 1990a). However these have not been prospectively evaluated but (generally) compared with other techniques for their ability to detect existing ulcers. A number of such studies have directly compared biothesiometers and monofilaments but not found either to be superior (Kumar, 1991; Vileikyte et al, 1997).

Screening to detect those at risk of amputation

Mayfield et al (2000) looked at the effectiveness of foot examinations as one method for decreasing the risk of foot amputation. Their patients were Pima Indians with diabetes enrolled in a diabetes programme or receiving care in a general medical clinic. Clinicians in all health care settings were encouraged to conduct foot examinations on all patients with diabetes and delivery of these recommendations was monitored through an annual chart audit. Two hundred and forty four patients were recruited to the study, 61 who had had a previous nontraumatic amputation of a lower extremity between 1985 and 1992, and 183 with no amputation by 1992. From medical records the pivotal event leading to the amputation was identified and records were then examined for a 36 month period retrospective to that pivotal event for: the number of clinic visits; number and type of foot examinations; number of missed appointments; management and delivery of foot care; and other health conditions and diabetes complications. Once identified, the pivotal event was assigned to three randomly selected patient controls for every patient case, and their records were examined in a similar method. Over 1100 foot examinations were performed in a 36 month period. However, when the independent effect of foot examination on the risk of amputation was examined in a logistic model, that controlled for differences in demographics, diabetes severity, and foot risk conditions, receiving one or more foot care examinations during a three year period did not provide significant benefit (odds ratio 0.55, 95% confidence interval 0.17, 1.7, p=0.31). Nonadherence with foot care recommendations was also not associated with risk of amputation (OR 1.9, 95% CI 0.88, 4.3, p=0.10).

Comment

No direct evidence was found to identify the optimal content or frequency of visual inspections and examinations, but regular surveillance at a minimum frequency of once annually was held by the guideline development group to be good clinical practice where no complications have previously been found. The benefits of monitoring people with diabetes without foot complications arise from the ability to detect feet at raised risk and the reduced morbidity achieved by aggressive intervention. Hence primary and secondary care should work together to identify a package of care for patients at raised risk of ulceration.

Adequate glycaemic control has been demonstrated to reduce the incidence and progression of microvascular disease in Type 1 (insulin dependent diabetes) patients (Diabetes Control and Complications Trial Research Group, 1993; Reichard et al, 1993). Direct evidence that improved glucose control will equally benefit Type 2 diabetes patients (for whom macrovascular, rather than microvascular, complications are of greater relative importance) is inadequate. A review of available epidemiological and trial evidence suggested a trend towards greater neuropathy with poor glycaemic control but the evidence is far from conclusive (Gaster et al, 1998). The UK Prospective Diabetes Study, which reported in 1998, discussed the impact of differing levels of glycaemic and blood pressure control in terms of macro and micro vascular disease but did not report specific outcomes concerning diabetic foot ulcers.

Risk factors

Evidence statements

A range of risk factors has been identified to indicate an increased risk of ulceration. (III)

Modifiable risk factors include peripheral vascular disease, neuropathy, foot deformities, plantar callus, smoking. (III)

Introduction

Epidemiological and clinical risk factors for ulceration have been extensively but inconclusively researched.

Evidence

Reported markers of increased risk are old age, duration of diabetes, neuropathy, peripheral vascular disease, renal disease, foot deformities, plantar callus, previous ulceration or amputation, poor vision, poor footwear, cigarette smoking, social deprivation and social isolation (Pecoraro et al, 1990; Walters et al, 1992; Chaturvedi et al, 1996; Kumar et al, 1994; Caddick et al, 1994; Murray et al, 1996; de Sonnaville et al, 1997; Litzelman et al, 1997a, b; Abbott et al, 1998). Studies using different methods and patient data and including selections of risk factors come to different conclusions about their relative importance in predicting complications. Hence, deciding at what level, and in what combinations these manifestations become clinically important is hampered by the lack of availability of good intervention studies taking a range of thresholds. However, these risk factors are all easily observable by trained health professionals.

Risk factors for diabetic foot ulceration

(see Table 1)

Table 1. Risk factors.

Table 1

Risk factors.

Four papers were identified in which risk factors associated with development of foot ulcers were examined (Sriussadaporn et al 1997, Frykberg et al 1998, Boyko et al 1999, and Kastenbauer et al 2001). The study designs were case control (Sriussadaporn et al 1997) and cohort (Frykberg et al 1998, Boyko et al 1999, Kastenbauer et al 2001), although Frykberg et al only report baseline characteristics of subjects. Two of the populations studied had Type 2 diabetes, the other two involved both Type 1 and Type 2 (Frykberg et al 1998, Boykyo 1999). In Frykberg et al, of the 251 North American patients studied, (126 men, 125 women) of either Caucasian, black or Hispanic race, 99 patients had a current or prior history of ulceration, with 33 with active ulcers. A multivariate logistic regression analysis for risk of ulceration identified three significant factors, having a vibration perception threshold 25V (OR 4.4, 95% CI 2.58, 7.54), having a Semmes-Weinstein monofilament test ≥5.07 (OR 4.1, 1.89,8.87), and a maximum peak plantar pressure ≥6kg/cm2 (2.1, 11.32, 3.39), after controlling for age, sex, diabetes duration, and race. The case control study, which took place in Thailand, compared 55 patients with full thickness ulcers mid calf or below with 110 patients without ulcers, matched for age. All patients with a previous history of foot ulcer were excluded. In a multiple logistic regression analysis, the risk of developing a foot ulcer was associated with peripheral nerve status – absence of short-latency somatosensory evoked potentials (odds ratio 1.67, 95% CI 0.31, −8.97, p<0.001), visual acuity (OR=0.233 per unit decrease in decimal visual acuity, 95% CI 0.005, 0.39, p<0.005) and fasting plasma glucose level (OR=1.01 per mmol/l increase, 95% CI 1.00, 1.02, p<0.005), after controlling for HbA1c, urea, creatinine, diabetes knowledge, and foot care.

Of the two cohort studies, Kastenbauer et al followed up 187 Austrian patients (excluding 25 withdrawals from the study) without a history of foot ulceration once a year for up to 54 months (mean follow up 3.6 years). Ten patients developed 18 neuropathic foot ulcers during the study period, producing an annual incidence of first foot ulceration of 1.6% (95% CI 0.7, 2.6) and an incidence density of 21.7 per 1000 person years (95% CI 8.2, 35.1). In a multiple stepwise Cox Proportional Hazards regression analysis, the predictors of foot ulceration were having an elevated vibration perception threshold (relative risk 25.4, 95% CI 3.1, 205, p=0.0024), an elevated mean plantar pressure (RR 6.3, 95% CI 1.2, 32.7, p=0.0291), a daily intake of alcohol (RR 5.1, 95% CI 1.1, 24.0, p=0.0404), mediasclerosis (RR 0.07, 95% CI 0.01, 0.6, p=0.0174). The other non-significant covariates in the model were age, diabetes duration, weight, oral antidiabetic therapy, insulin use, history of angiography, flatfoot deformity, hallux valgus, oxford shoes, varicosis, dry skin, skeletal abnormalities, HbA1c, triglycerides, stage or peroneal nerve conduction velocity, and diastolic blood pressure. The Seattle Diabetic Foot Study provided the results for the second cohort study (Boyko et al 1999). In this large study, 749 veterans with diabetes with 1483 lower limbs and no current foot ulcer, were followed up at 12 – 18 month intervals with a foot examination, and quarterly by postal contact to check for the presence of foot ulcers. Over a mean follow up period of 3.7 years, 162 ulcers developed, 3.0/100 person years. Using a Cox proportional hazards stepwise model, nine variables were found to be risk factors for foot ulcer: sensory neuropathy by 10g monofilament (RR 2.17, 95% CI 1.52, 3.08, p<0.001), history of foot ulcer (1.63: 1.17, 2.26, p=0.004), history of amputations (2.81: 1.84, 4.29, p<0.001), insulin use (1.59: 1.14, 2.22, p=0.006), dorsal foot transcutaneous oxygen tension (mm Hg) (0.80: 0.69, 0.93, p=0.004), weight (1.23; 1.06, 1.43, p=0.006), ratio of ankle systolic pressure to brachial systolic pressure (0.83; 0.73, 0.96, p=0.011), Charcot deformity (3.49; 1.22, 9.92, p=0.019), vision <20/40 (1.93; 1.42, 2.63, p<0.001). Other non-significant variables excluded in the multivariate model include height, diabetes duration, Type 2 diabetes, random glucose, HbA1c, erythrocyte sedimentation rate, serum creatinine, claudication, peripheral vascular disease, history of vascular bypass surgery, changes in heart rate with timed breathing, no hallux vibration sensation, no Achilles tendon reflex, foot numbness, foot pain, special footwear, hallux limitus, hammer claw/toe ulcer history, hallux joint mobility, extensor digitorum brevis test result, chronic lower limb oedema.

In a fifth paper the authors (Margolis et al 2000) conducted a meta-analysis of (white) patient data from the standard care groups in five randomised controlled trials, not all of which had been published, to identify risk factors for delayed healing of neuropathic diabetic foot ulcers. Patients were receiving off loading, debridement, or moist wound dressing treatments. Factors significantly associated with healing within 12 weeks or 20 weeks included an ulcer duration of less than six months, an ulcer of small size, and at twelve weeks, but not 20 weeks.

Risk factors for lower limb complications in people with diabetes

(see Table 1)

El Shazly et al (1998) conducted a case control study of 348 people with either type 1 or Type 2 diabetes and major complications of the lower extremities, defined as having a foot ulcer, claudication, gangrene and/or ischaemic rest pain lasting 15 days or more, bypass or angioplasty for peripheral vasculopathy, or amputation within previous 12 months and 1050 controls (no complications of the lower extremities). Multivariate logistic regression analysis identified a number of significant risk factors for developing lower limb complications. These included being 50–70 years of age, male, not married, having Type 1 diabetes or being Type 2 on insulin treatment, having cardio or cerebrovascular disease, diabetic neuropathy, abnormal HbA1c, needing help to reach a health care facility, no regular follow-up visits, and no educational intervention. Non significant factors included level of education, employment status, presence of diabetic retinopathy, nephropathy, or hypertension, co-morbid conditions, duration of diabetes, smoking and alcohol consumption.

Risk factors for lower limb amputations in patients with diabetes

(see Table 1)

Four papers were identified that examined the risk factors for lower limb amputation, all were cohort studies, three prospective (Lehto et al 1996, Adler et al 1999, Hamalainen et al 1999) and one retrospective (Gurlek et al 1998). In three studies the populations involved Type 1 and Type 2 patients, whilst the patients in the fourth study were Type 2 only (Lehto et al 1996). In the smallest of the four studies, medical records were retrospectively examined for 147 consecutive patients (97 men, 50 women, 11 Type 1, 136 Type 2) initially referred to a medical centre in Turkey between 1992 and 1996 (Gurlek et al 1998). Fifty four patients (36.7%) had had an amputation. A logistic regression analysis to identify significant risk factors associated with amputation identified peripheral vascular disease (odds ratio 4.0, 95% CI 1.17, 13.4, p=0.03), presence of osteomyelitis (3.73; 1.08, 12.6, p=0.04), and presence of grade 4 or 5 gangrenous lesions (30.8; 7.39, 121.5, p<0.0001). Variables that were not statistically significant included age, sex, duration of diabetes, smoking history, hypertension, nephropathy, or retinopathy. The primary indication for amputation was the presence of gangrene in 66.6% of patients. Of the three prospective studies, two were based in Finland (Lehto et al 1996, Hamalainen et al 1999)and the in the third, patients were participants on the Seattle Diabetic Foot Study in the USA (Adler et al 1999). In this latter study, 776 US veterans with diabetes (98.2% male, 51 Type 1, 725 Type 2) were followed up for 0 – 5.8 years, (median 3.3 years). Thirty patients underwent amputations in this cohort and three multivariate models, incorporating different measures of peripheral vascular disease: palpation of the posterior tibialis and dorsalis pedis pulses; partial pressure of oxygen at the skin surface; and with the ankle arm index, were developed to identify risk factors associated with amputation. Former amputation, lower extremity ulcers, peripheral vascular disease, and treatment with insulin were significant factors after controlling for duration of diabetes in all three models. In addition, peripheral sensory neuropathy was also significant when the partial pressure of oxygen at the skin surface was used to measure peripheral vascular disease. The age-adjusted incidence rate for amputation in the men only, standardised to the 1991 US male diabetic population was 11.3 per 1000 patient years.

Patients in both of the Finnish studies were selected from the central register of diabetes patients receiving drug reimbursement. Hamalainen et al identified recruited their age stratified sample of 728 patients in 1987 and invited them for a follow-up examination seven years later. In Lehto et al, 1044 patients with Type 2 diabetes were selected in 1983 if they were aged between 45 and 64 years and lived in districts served by two hospitals, one in east Finland, the other in west. These patients were also followed up seven years later in 1990 with a postal questionnaire. Fifty eight of the 1044 had undergone an amputation in Lehto et al’s cohort and 25 of the 728 followed up by Hamalainen et al. The incidence of amputation was similar in both sexes, 5.6% in men and 5.3% in women (Lehto et al 1996) and no differences between those treated with insulin or with diet or oral hypoglycaemic drugs, after adjustment for age, sex and severity of diabetes. Cox regression analysis (Lehto et al) and logistic regression (Hamalainen et al) were used to determine the risk factors for amputation. Considering that the patients in these two studies were identified from the same register and live in the same country, little overlap and some discrepancies were seen in the variables identified in the two studies. Significant factors identified by Lehto et al included having retinopathy (RR 3.6; 2.2, 6/1, p<0.001), urinary protein (1.8; 1.1, 3.2, p=0.003), total cholesterol values >6.2 mmol/l (1.8; 1.1, 1.6, p=0.047), fasting plasma glucose >13.4 mmol/l (2.5; 1.5, 4.3, p<0.001), HbA1c >10.7% (2.4; 1.4, 4.0, p=0.001), duration of diabetes > 9 years (2.2; 1.3, 3.6, p=0.004), absence of two or more peripheral artery pulses (3.9; 2.3, 6.8, p<0.001), femoral artery bruit on auscultation (2.1; 1.1, 4.0, p=0.022), bilateral absence of Achilles tendon reflexes (4.3; 2.5, 7.3, p<0.001) and bilateral absence of vibration sense (2.7; 1.6, 4.7, p<0.001). An HbA1c > 9.8% and a duration of diabetes >7 years both increased the risk for amputation independently of each other (p<0.01). Other non-predictive variables included body mass index, previous myocardial infarction, hypertension, smoking, triglyceride levels, and HDL cholesterol levels. Hamalainen et al found that an abnormal vibration perception threshold (odds ratio 14.5; 95% CI 3.6, 57.8, p=0.0001), ankle/brachial pressure index (8.2; 2.8, 24.0, p=0.0001), retinopathy (6.1; 1.9, 19.6, p=0.0024), visual handicap (4.9; 1.4, 17.4, p=0.0129) and being male (3.3; 1.0, 10.8, p=−.0431) were significant risk factors, whilst age at onset of diabetes, duration of diabetes, history of cardiac failure, claudication, absence of at least one peripheral pulse, and elevated serum creatinine.

The one published trial of multidisciplinary intervention on the basis of risk (McCabe et al 1998, reported in the next section) took a very conservative threshold, identifying only patients with gross risk factors, and demonstrated the value of a protection programme. A number of studies have looked specifically at the role of footwear in reducing lesions in at-risk patients but require further research. No studies specifically addressing education in patients at raised risk (broadly defined) were uncovered (but see Education for patients with foot ulcers).

Surveys indicate sub-optimal supervision of elderly patients in hospital, residential care, and general practice (Fletcher et al, 1996; Wilkes et al, 1980; Benbow et al, 1997; Neil et al, 1989; Dornan et al, 1992). Older patients contain the greatest proportion of patients at raised risk of ulceration and other complications.

Comment

The guideline development group felt it was important that specific consideration should be given to organising care to ensure adequate supervision of these patients with risk factors although there is no available evidence concerning the most appropriate process.

Copyright © 2003, School of Health and Related Research (ScHARR), University of Sheffield.
Bookshelf ID: NBK51719

Views

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...