NB This page is from the 2005 Hyperbook.
Last evidence check September 2006 but no substantial revision since Feb 2005.
Diabetics develop a number of problems with their feet
The causes of these are linked together.
Ulcers usually begin with small abrasions which deepen. However, some ulcers begin with undermined skin which breaks down. Ulceration may progress to expose subcutaneous tissues, tendon, bone or joint. Infection, gangrene or osteomyelitis may supervene. The severity of a diabetic ulcer is usually expressed on the Meggitt-Wagner scale (see diagram). A more comprehensive scale has been developed by Armstrong and colleagues at the University of Texas, which includes risk stratification and expresses tissue breakdown, infection and gangrene separately (Lavery et al 1996). The University of Texas scale has been shown to be reproducible and to correlate with prognosis, especially risk of amputation (Armstrong et al 1998, Treece et al 2004).
Charcot arthropathy probably begins with trauma – about 30-50% of patients have a recognised injury, others probably have subclinical injuries. Injury to the joint is followed by rapid destruction of the joint surfaces and demineralisation. There appears to be osteoclast overactivity, bone vascular shunting and bone breakdown. Some of the metabolic features are similar to those of complex regional pain syndromes and regional osteoporosis, but in these conditions there is no joint destruction. Charcot arthropathy is often said to be painless, but in fact there is usually pain, though this may be less than might be expected from the degree of arthropathy and deformity. The joint destruction and demineralisation may lead to deformity, skin pressure and ulceration. After a few weeks healing begins and after a few months there is usually bony union with joint incongruity and deformity. Progress of Charcot arthropathy is usually expressed using the Eichenholz scale:
Charcot arthropathy usually begins in the tarsometatarsal region, but sometimes it is seen in the midtarsal or ankle joints, or as pathological calcaneal fractures. The distribution is usually expressed by the Brodsky classification. Schon has described a more precise classification, which has been validated, but it has not yet been shown that the additional complexity is clinically useful.
In a community study of 1150 diabetics (Walters et al 1992) the prevalence of active ulceration was 3.3% and of active or past ulceration 7.4%. Forty percent were neuropathic, 24% vascular and 36% mixed. One third of the ulcers were deep and 6% had osteomyelitis. Abbott et al (1998) reported an incidence of new ulcers of 7.2% in one year of a large multi-centre trial of a drug for ulcer prophylaxis.
Apelqvist (1993) followed a cohort of 558 diabetics with foot ulcers for a median of 4 years. Recurrent ulcer rates were 34% at 1 year, 61% at 3 years and 70% at 5 years, while the amputation rate was 6% at one year, 16% at 2 years and 22% at 3 years. Carrington et al (2002) followed a cohort of 169 diabetics (some with previous ulcers). After 6 years, 37% had at least one new ulcer and 11% had an amputation.
Current or previous ulcers (Meijer et al 2001) and Charcot arthropathy (Pinzur et al 2003) seriously reduce mobility and quality of life).
The following increase the risk of major complications in the diabetic foot:
The International Working Group on the Diabetic Foot produced a risk stratification, which was studied in a prospective cohort of 225 patients followed for 3 years (Peters et al 2001).
Group |
Criteria |
Ulceration rate (%) |
Amputation rate (%) |
0 |
no neuropathy |
5 |
0 |
1 |
neuropathy, no deformity or vascular disease |
14 |
0 |
2 |
neuropathy and deformity or vascular disease |
19 |
3 |
3 |
previous ulcer or amputation |
55 |
21 |
Prevention of complications through a community based screening programme to identify patients at risk of ulceration is the cornerstone of a diabetic foot management programme. All patients are taught about the risks and about basic care of their feet. The Cochrane database systematic review by Valk et al (2001) concluded that education may reduce ulceration and amputations, although the effect was greatest in high-risk patients and in the short term (one-year follow-up). Podiatric checking and care are provided as indicated by risk category. Patients at high risk of ulceration also have accommodative shoes and pressure-relieving insoles provided.
Screening, education and good metabolic care can reduce the incidence of neuropathy and ulceration. In patients at high risk, appropriate shoewear and insoles can delay ulceration (and re-ulceration after ulcer healing)(McLaughlin et al 2004). This group of patients are at high risk of death from cardiovascular and renal disease, and treatment of hypertension and hyperlipidaemia is important.
Prophylactic correction of hammertoes to prevent ulceration in 31 neuropathic diabetics was described by Armstrong et al (1996). 2/14 with previous ulceration developed wound infections and 1/14 a recurrent ulcer. No ulcers or infections occurred in the other 17. Without a RCT standardised for risk it is difficult to interpret this study.
Decompression of the tarsal tunnel in diabetics with neuropathy and ulceration is effective in healing ulcers and preventing future ulceration (Dellon et al 2004)
The key to healing diabetic foot ulcers is relief of pressure. This is classically achieved with a total contact plaster cast (TCC), and the efficacy of TCC was established by an RCT of TCC versus dressings and nonweightbearing reported by Mueller et al (1989). Some evidence suggests that pressure transfer with a TCC is to the shank rather than elsewhere in the foot. However, the TCC is time-consuming to apply and maintain. Alternative approaches include shoe casts such as the Scotchcast boot (Burden et al 1983, ), MABAL shoe or walker boots. Comparative trials suggest healing rates are similar with all these off-loading methods (Birke et al 2002), but the methodological quality is low. Removable cast walker boots have been popular in recent years, but Armstrong (2003) found that patients tend to remove the boot and walk without it. By wrapping casting tape round a walker boot to make it non-removable, Armstrong achieved a healing rate comparable with total contact casting.
Patients with ulcers were traditionally advised to avoid weightbearing, and were often admitted to hospital for bed rest. Elevation will help oedema which may inhibit wound healing. However, provided the ulcer itself is protected from pressure, walking may actually improve ulcer healing.
Local wound care is also important. Removal of slough and callus allows better healing. The Cochrane systematic review (Smith et al 2002) found that the available trials favour the use of hydrogel dressings.
Growth factors have been used to improve ulcer healing, but it can be difficult, in trials, to distinguish the effect of the growth factor from ancillary treatment such as debridement and total contact casting.
A tight Achilles tendon is often associated with ulceration and resistance to healing. Mueller et al (2003) found that percutaneous Achilles tendon lengthening was associated with increased ulcer healing compared with a control who had standard treatment. At 2 year follow-up the rate of recurrent ulceration was 38% in the treatment group and 81% in the control group. Achilles lengthening leads to short-term plantar flexor weakness, but this recovers over the subsequent 6-12 months (Maluf et al 2004).
Metatarsal osteotomies (Fleischli et al 1999), Keller arthroplasty (Lin et al 2000) and excision interphalangeal arthroplasty may also be useful in relieving pressure on ulcers and allowing them to heal.
Healing will occur in the majority of ulcers within 2-3 months with treatment as outlined above. The probability of successful healing (Margolis et al 2003) is predicted by:
Whatever local methods are used to heal an ulcer, it is important to optimise diabetic control, treat other complications such as renal failure and optimise arterial inflow to the limb if there is an ischaemic component to the ulcer.
Once the ulcer is healed, total contact insoles need to be provided by the orthotists (Bus et al 2004), with shoes large enough to accommodate the foot and the insole without rubbing. The orthotist will review the patient regularly and repair and replace shoes and insoles as required. Protective shoewear and follow-up can prevent re-ulceration, with over 50% of patient who survived 10 years still ulcer-free (McLaughlin et al 2004).
Infection in an ulcer can be difficult to treat because of:
The usual signs of swelling, rubor, heat and discharge may be obvious, but in an ischaemic foot with deficient neutrophil function the diagnosis may be difficult. If a probe inserted into the ulcer goes directly down to bone, osteomyelitis is probably present. Ultimately, surgical exploration and biopsy may be necessary to make the diagnosis.
All standard imaging modalities can play a part in diagnosing infection, and normal pathways should be followed. There is a significant false negative rate. In the presence of Charcot arthropathy all imaging is particularly difficult to interpret and labelled antibody fragments may be the most accurate method.
An infected foot should be elevated and broad-spectrum antibiotics administered intravenously. Each unit needs to develop their own antibiotic protocol with the microbiologist according to local organisms and sensitivities. Improvement of the vascular inflow may improve the chance of limb salvage.
Negative-pressure dressings may improve healing rates over conventional dressings, although the evidence is not yet strong (Evans et al 2001, Eginton et al 2003). The technique is useful in wounds which have failed other attempts at healing (Clare et al 2002).
Established deep infection may require surgical debridement and this may include amputation at various levels:
Distal amputations have a good healing rate overall and can maintain mobility in appropriate shoes and orthoses. However, losing one toe is a major risk factor for further amputation (Murdoch et al 1997). Loss of the great toe significantly increases pressure under the remaining forefoot (Armstrong and Lavery 1998) and there is a high incidence of subsequent lesser toe deformity (Quebedeaux et al 1996). Midfoot amputations are probably best when a good plantar flap can be used for durable closure without tension. Midtarsal (Chopart) amputations require motor balancing by transfer of all available dorsiflexors into the talar neck to balance the Achilles tendon and prevent equinus. Hindfoot amputations have the advantage that they can usually be walked on a short distance without a prosthesis (useful for going to the toilet in the middle of the night) (Pinzur et al 2003). However, prosthetic fitting for outdoor mobility is more difficult than a below-knee stump. More than at most sites, the decision to do a hindfoot or BK amputation should be taken in close collaboration with the prosthetist.
Peters et al (2001) found that Sickness Impact Profile scores in diabetics with forefoot or midfoot amputations were no different from those without amputations, but diabetics with trans-tibial amputations scored significantly worse. It is worth trying to amputate distally.
Our own policy is to do forefoot procedures, up to the level of the Lisfranc joint, where possible. Wounds are left open and the patient mobilises in a Scotchcast boot, TCC or walker as soon as the wound has settled. Patients with wounds which cannot be closed at the Lisfranc joint or distal would usually be offered a proximal amputation, trans-tibial if possible.
A few resistant ulcers around the hindfoot can be closed by excision and grafting or flap closure in collaboration with a plastic surgeon (Attinger et al 2002, Musharrafieh et al 2003). However, ischaemia and infection limit what is feasible.
The principles are:
Surgery may be required in two main situations: