Rationale and impact
These sections briefly explain why the committee made the recommendations and how they might affect practice.
All the risk assessment tools reviewed by the committee were able to predict ulcer occurrence with acceptable accuracy. There were no significant differences in classification accuracy (assessed using c‑statistics) between the different risk assessment tools. When considering classification accuracy, sensitivity and specificity together, the PODUS and SIGN systems were the best. PODUS had a higher c‑statistic than SIGN, but it did not report sensitivity or specificity. SIGN had the best overall sensitivity and specificity.
The committee agreed that the most important factor for an assessment tool was the ability to accurately identify people who are at high risk of developing a diabetic foot ulcer. Accurate identification allows people to be referred to appropriate services, where monitoring and preventative treatment can be started. A focus on high sensitivity over high specificity may lead to more false positives, with more people incorrectly receiving increased monitoring and referral to specialist services. However, the committee believe that this is preferable to using a system with lower sensitivity, because an increased risk of ulcer, infection and amputation is much worse than wasted resources from unnecessary monitoring or referrals. Overall, the SIGN system showed the highest sensitivity for both high-risk and combined high- and moderate-risk groups.
The committee considered recommending the PODUS clinical prediction rule because:
evidence from the prospective cohort study was high quality
it has higher classification accuracy than the SIGN system and
it is a short and simple assessment with only 3 items, and it could be completed by primary care professionals who do not have specialist knowledge of diabetic foot care.
Despite the good evidence for the PODUS system, the committee decided not to change the 2015 recommendations, because:
SIGN had good sensitivity and specificity (although this assessment was based on a study with a high risk of bias).
PODUS did not include an assessment of foot deformity in its final model (in an earlier systematic review to identify the factors that most accurately predicted foot ulceration, foot deformity was rejected for being inconsistently defined). Based on their experience and knowledge of established research, the committee believe that this is an important clinical risk factor and disagreed with it being left out of the final PODUS model.
The SIGN system is also relatively simple. It uses the same 3 items as PODUS, but also includes an assessment of foot deformity. The committee think that assessments using SIGN would only take slightly longer than assessments using PODUS, and could also be completed by primary care professionals who do not have specialist knowledge of diabetic foot care.
SIGN is recommended by the 2015 guideline. It is well established in clinical practice, and widely recognised and understood by practitioners. If the committee recommended SIGN, existing processes could be used without issue and there would be no risk of a disruptive change to practice.
If the committee recommended PODUS, staff would need training on how to use the new system. Several free online training courses for primary care professionals would need changing. Primary care electronic patient record systems would also need to be modified.
There is no evidence assessing the use of PODUS in NHS practice. Given the difficulties the NHS and primary care are currently facing, the committee did not want to introduce a potentially expensive and time-consuming change in practice without clear evidence of a significant benefit. They were particularly concerned about the impact because of current low staffing levels and the time staff will need for retraining.
The 2015 guideline recommended a modified version of SIGN that includes a check for renal disease. The committee agreed that this modification is useful and should be retained, because renal disease is a known risk factor for diabetic foot problems.
The recommendations have not changed, so no resource impact is expected.
The evidence showed that 95.5% of people assessed as low risk at their first clinical assessment remained in the low-risk group at their final assessment 8 years later. The ulceration rate in the low-risk group is also very low. Given this evidence, the committee discussed reducing the frequency of foot risk assessments to once every 2 years. However, they were concerned about the impact this may have on patient care.
The annual foot assessment is not just a foot examination and risk assessment. It is also a chance to teach people how to look after their feet, and to emphasise the importance of doing so. Many people with diabetes do not have good foot care routines, or do not have foot care routines at all. They may not know what to do if they have a foot problem, or who to contact. And they may benefit from regular advice about risk factors for foot problems. Reducing the frequency of foot assessments would mean reducing the number of chances to encourage good foot care and direct people to sources of support.
The committee discussed options for providing education and support outside of foot assessments (for example, remote appointments). However, it was not clear how feasible it would be to run these extra appointments in practice. Foot assessments are currently part of the annual diabetes review, so it makes sense to continue to include the foot check and risk assessment in that appointment. There are also Quality and Outcomes Framework (QOF) indicators for annual foot examination and risk classification, which further justify retaining the current system.
Given the risk of reducing access to education and support, the committee agreed to continue recommending annual foot assessments. They agreed that, for the recommendations to change, better evidence would be needed comparing annual and 2‑yearly foot assessments. The committee therefore made recommendations for research on:
The recommendations have not changed, so no resource impact is expected.
The committee agreed that in people with diabetes, all foot wounds are likely to be colonised with bacteria. However, for people with a diabetic foot infection, prompt treatment of the infection is important to prevent complications, including limb-threatening infections.
The committee agreed to retain the recommendation from the 2015 guideline that antibiotics should be started as soon as possible if a diabetic foot infection is suspected. The choice of antibiotic would depend on the severity of infection, although the committee acknowledged that the studies they looked at did not always differentiate between severities. The committee accepted the Infectious Diseases Society of America's definitions of mild, moderate and severe infection, and recommended that this should be taken into account when choosing an antibiotic.
The committee retained the 2015 recommendation that samples should be taken for microbiological testing before, or as close as possible to, the start of antibiotic treatment. This would allow empirical antibiotic treatment to be changed if needed when results are available.
These recommendations are consistent with current practice.
The committee agreed that in their experience, the incidence of diabetic foot infections in children and young people is rare. The mean age of participants in the evidence considered ranged from 54 to 64 years. Based on these factors, the committee included an antibiotic prescribing table for adults, but not for children and young people. They recommended that if a diabetic foot infection is suspected or confirmed in children or young people, specialist advice should be sought regarding antibiotic choice and regimen.
The evidence showed no difference in clinical outcomes for most antibiotics. But the antibiotics used in the studies were not wholly representative of UK practice, with some not being available in the UK and others not widely used. There were no differences in adverse events for many antibiotic comparisons. However, there were differences between some antibiotic classes, with lower rates of adverse effects generally for beta-lactam antibiotics.
The committee agreed that the choice of antibiotic in adults should be based on severity of infection (mild, moderate or severe) and the risk of complications, while minimising adverse effects and antibiotic resistance. This means using narrow-spectrum antibiotics first where possible, and using microbiological results, when available, to guide treatment.
The antibiotics recommended have good activity against many of the pathogens that cause diabetic foot infection, have good penetration for skin and soft tissue infections, and can be used in the different settings where treatment may take place, including ambulatory care. Based on evidence, their experience and resistance data, the committee agreed that flucloxacillin is an effective empirical antibiotic for mild diabetic foot infections (with dosing taking account of a person's body weight and renal function). The committee agreed that flucloxacillin has poor oral bioavailability and in people with diabetes who could have impaired circulation, a higher (off‑label dose) of up to 1 g four times a day may be needed to adequately treat diabetic foot infection.
For adults with a moderate or severe diabetic foot infection, a choice of antibiotics (or combinations of antibiotics) should be available. This enables selection based on individual patient factors, likely pathogens, and guided by microbiological results where available. In moderate and severe infection (which includes osteomyelitis), broader cover is needed because aerobic and anaerobic bacteria may be present. Severe infections can become limb-threatening quickly so antibiotic choices with the broadest spectrum of cover are appropriate; this can be changed to a narrower-spectrum antibiotic based on microbiological results when available, in line with principles of good antimicrobial stewardship. For moderate or severe infection, the committee recommended flucloxacillin at a dose of 1 g four times a day.
Patient preference is also important, particularly for treatment that will involve a hospital stay or be prolonged. Diabetes is a chronic condition and people may have had previous foot infections, with previous courses of antibiotics, that will influence their preferences.
No evidence was identified comparing antibiotic dose, frequency or route of administration. However, the committee acknowledged that a person with a diabetic foot infection may already be on a number of other medications, and this should be taken into account when deciding on dose, frequency and route of administration of an antibiotic.
In line with the NICE guideline on antimicrobial stewardship and Public Health England's Start smart – then focus, the committee agreed that oral antibiotics should be used in preference to intravenous antibiotics where possible. Intravenous antibiotics should only be used for people who are severely ill, unable to tolerate oral treatment, or where oral treatment would not provide adequate coverage or tissue penetration. The use of intravenous antibiotics should be reviewed by 48 hours (taking into account the person's response to treatment and any microbiological results) and switched to oral treatment where possible.
The committee agreed that a shorter course was generally as effective as a longer course for adults with a mild diabetic foot infection, and a 7‑day course was sufficient for most people. However, it agreed that a longer course (up to a further 7 days) may be needed for some people based on a clinical assessment of their symptoms and history. They discussed the limited evidence on antibiotic course length, which compared 6 weeks with 12 weeks in adults with diabetic foot osteomyelitis. The committee agreed that for adults with a moderate or severe diabetic foot infection (which includes osteomyelitis), a 7‑day course would be a minimum, with antibiotic treatment for up to 6 weeks if they have osteomyelitis. When prolonged antibiotic treatment is given, oral options should be used and treatment should be reviewed regularly, taking into account the need for continued antibiotics. The committee discussed antibiotic choices for osteomyelitis and agreed that the empirical choices for moderate and severe diabetic foot infection are also effective empirical choices for osteomyelitis.
The recommendations aim to optimise antibiotic use and reduce antibiotic resistance.
The committee based the recommendation on their experience and safety netting advice from the NICE guideline on antimicrobial stewardship. They agreed that if symptoms worsened rapidly or significantly at any time, or did not improve within 1 to 2 days, people with a diabetic foot infection should be advised to seek medical help.
The recommendation should ensure that appropriate safety netting is in place.
The committee agreed that when microbiological results are available, they should be used to guide antibiotic choice. The committee recognised the complexity around interpreting microbiological results, and agreed that the quality and type of specimen should be taken into account when making decisions around whether to change an antibiotic. The committee also discussed factors that would indicate that a person with a diabetic foot infection would need to be reassessed. These included if an infection was rapidly or significantly worsening or not improving, if other diagnoses were possible, or symptoms suggested a more serious illness or condition.
These recommendations should ensure that appropriate reassessment is in place.
The committee agreed to retain the 2015 recommendation that antibiotics should not be given to prevent diabetic foot infections. No evidence was identified for antibiotic prophylaxis and the committee agreed that antibiotic prophylaxis is not appropriate because of concerns about antimicrobial resistance. People should be advised to seek medical help if symptoms of a diabetic foot infection develop.
This recommendation is consistent with current practice.