Guidance

Summary of the evidence

Summary of the evidence

This is a summary of the evidence. For full details, see the evidence review.

Managing cellulitis and erysipelas

  • Cellulitis and erysipelas are infections of the tissues under the skin, which are treated with antibiotics.

  • The main bacteria causing cellulitis and erysipelas are Streptococcus pyogenes and Staphylococcus aureus, but infection can also be caused by Streptococcus pneumoniae, Haemophilus influenza, gram-negative bacilli and anaerobes (NICE clinical knowledge summary on cellulitis).

  • The evidence identified in this guideline was for antibiotics compared with other antibiotics for managing non-surgically acquired cellulitis or erysipelas in adults, young people and children. Most studies did not report the site of infection, but where this was reported, most cases had a lower limb infection or less frequently an upper limb infection. One systematic review excluded a study of facial cellulitis.

Committee discussion on managing cellulitis and erysipelas

  • The committee discussed that to ensure the appropriate treatment of cellulitis and erysipelas, it is important to exclude other causes of skin redness (erythema). This can often be caused by an inflammatory reaction, for example, following an immunisation or an insect bite, or any other non-infective cause such as chronic venous insufficiency, which could be wrongly treated as cellulitis or erysipelas.

  • The committee discussed that in most cases microbiological swabbing of cellulitis or erysipelas yields negative results (particularly if the skin is intact) and in most cases the infecting organism is likely to be either Streptococcus pyogenes or Staphylococcus aureus bacteria. They therefore agreed that swabbing should not be undertaken routinely. However, the committee agreed that where the skin is broken and there is reason to believe a different organism may be involved (for example, if there is a penetrating injury, exposure to water-borne organisms, or infection acquired outside the UK), then a swab may be useful to guide antibiotic treatment.

  • The committee agreed based on experience that to monitor the progression of cellulitis or erysipelas, and help assess the effectiveness of antibiotic treatment, it may be useful to draw around the extent of the infected area using a single-use surgical marker pen before treatment. The committee discussed that a single-use surgical pen should be used because it is designed for this purpose (unlike other pen types that may damage skin or leave permanent marking) and would not risk cross-infection. The committee noted that in people with certain conditions (for example, lymphoedema), drawing around the infected area may be difficult or not possible because the rash may be ill-defined. Additionally, they noted that the extent of redness may be less visible on darker skin tones.

  • The committee agreed that it may take time for antibiotic treatment to take effect, and initially redness or swelling may extend beyond the marked line (if used).

  • The committee agreed that, in line with the NICE guideline on antimicrobial stewardship, prescribers should provide 'safety netting' advice to people with cellulitis or erysipelas about when to seek further help if they become more unwell or have side effects of antibiotic treatment, and also discuss that skin can take some time to return to normal even after a course of effective antibiotics. The committee was aware that the time taken for skin to return to normal appearance is variable. In their experience, it could be a number of weeks. Because no data were available to affirm this, the committee agreed not to specify a timescale.

  • The committee agreed that if a person's symptoms worsen rapidly or significantly at any time they should be reassessed, taking into account other possible diagnoses, the development of serious complications, such as orbital cellulitis, septic arthritis, osteomyelitis, lymphangitis, necrotising fasciitis or sepsis, and the possibility of an uncommon or resistant bacteria. The committee also agreed that reassessment should include managing any underlying condition that may predispose to cellulitis or erysipelas.

  • The committee agreed that taking a swab for microbiological testing should be considered if the skin is broken and this has not been done already. When microbiological results are available, the antibiotic should be reviewed and changed accordingly (for example, if bacteria are found to be resistant) if symptoms are not already improving, using a narrower-spectrum antibiotic if possible.

  • The committee agreed that people with cellulitis or erysipelas should be referred to hospital if they have symptoms or signs suggestive of orbital cellulitis, osteomyelitis, septic arthritis, necrotising fasciitis or sepsis.

  • The committee discussed and agreed that in some cases, the prescriber may need to consider referring or seeking specialist advice on inpatient treatment or locally available options for intravenous treatment at home or in a community setting. These cases include people who are severely unwell, at higher risk of complications, have infection near the eyes or nose (including periorbital cellulitis), could have uncommon pathogens, have lymphangitis, have a spreading infection that is not responding to oral antibiotics, or cannot take oral antibiotics. They discussed that children under 1 year and people who are frail or have underlying disease (such as diabetes or immunosuppression) are at a higher risk of developing complications, as are those who could have an uncommon causative organism, for example, following a penetrating injury, a wound exposed to water (surfers for example), or an infection acquired outside the UK.

Choice of antibiotics

Effectiveness of antibiotics versus other antibiotics in adults

  • There were no differences in the clinical effectiveness of the following antibiotic comparisons in adults with cellulitis or erysipelas:

    • an oral penicillin or cephalosporin compared with an oral macrolide or oral clindamycin (adults and children; Ferreira et al. 2016)

    • oral azithromycin compared with oral cefalexin (Kilburn et al. 2010)

    • oral azithromycin compared with oral erythromycin (Kilburn et al. 2010)

    • intravenous (IV) then oral moxifloxacin compared with IV then oral co‑amoxiclav (Vick-Fragoso et al. 2009)

    • IV or oral linezolid compared with IV vancomycin (Kilburn et al. 2010)

    • IV dalbavancin compared with IV vancomycin (Boucher et al. 2014)

    • IV ampicillin with sulbactam compared with IV cefazolin (Kilburn et al. 2010)

    • IV flucloxacillin compared with IV ceftriaxone (Kilburn et al. 2010)

    • IV moxifloxacin compared with IV piperacillin with tazobactam (Kilburn et al. 2010)

    • IV daptomycin compared with IV penicillin or IV vancomycin (Konychev et al. 2013)

    • IV tigecycline compared with IV ampicillin with sulbactam or IV co‑amoxiclav (Matthews et al. 2012)

    • newer cephalosporins compared with older cephalosporins (Kilburn et al. 2010)

    • IV ceftaroline compared with IV vancomycin plus aztreonam (Frampton 2013)

    • IV daptomycin compared with IV vancomycin (Pertel et al. 2009)

    • IV meropenem compared with IV imipenem with cilastatin (Kilburn et al. 2010).

  • Some differences were seen for some effectiveness outcomes for the following antibiotic comparison in adults with cellulitis or erysipelas:

    • oral macrolides or oral streptogramins (pristinamycin: not available in the UK) improved the number of people who were symptom-free, or had reduced symptoms, at 7 to 14 days' follow‑up compared with a penicillin (Kilburn et al. 2010).

Based on 3 systematic reviews (Frampton 2013, Ferreira et al. 2016 and Kilburn et al. 2010) and 5 randomised controlled trials (RCTs; Boucher et al. 2014, Konychev et al. 2013, Matthews et al. 2012, Pertel et al. 2009 and Vick-Fragoso et al. 2009).

Effectiveness of antibiotics versus other antibiotics in children

  • There was no difference in the clinical effectiveness of the following antibiotic comparison in children with cellulitis or erysipelas:

    • IV linezolid compared with IV vancomycin (Yogev et al. 2003).

Based on 1 RCT (Yogev et al. 2003).

Dual therapy in adults or children

  • There were no differences in the clinical effectiveness of the following antibiotic comparisons in adults or children with cellulitis or erysipelas:

    • oral cefalexin plus oral co‑trimoxazole compared with oral cefalexin alone (Bowen et al. 2017)

    • IV then oral flucloxacillin plus IV then oral benzylpenicillin compared with IV then oral flucloxacillin alone (Kilburn et al. 2010)

    • IV or oral flucloxacillin plus oral clindamycin compared with IV or oral flucloxacillin alone (Brindle et al. 2017)

    • IV ceftazidime plus IV vancomycin compared with IV ceftobiprole alone (Noel et al. 2008).

Based on 2 systematic reviews (Bowen et al. 2017 and Kilburn et al. 2010) and 2 RCTs (Brindle et al. 2017 and Noel et al. 2008).

Safety of antibiotics

  • Antibiotic-associated diarrhoea is estimated to occur in 2 to 25% of people taking antibiotics, depending on the antibiotic used (NICE clinical knowledge summary on diarrhoea – antibiotic associated).

  • About 10% of the general population claim to have a penicillin allergy; this is often because of a skin rash that occurred while taking a course of penicillin as a child. Fewer than 10% of people who think they are allergic to penicillin are truly allergic. See the NICE guideline on drug allergy for more information.

  • People with a history of immediate hypersensitivity to penicillins may also react to cephalosporins and other beta-lactam antibiotics (British national formulary [BNF] information on phenoxymethylpenicillin).

  • Cholestatic jaundice and hepatitis can occur with flucloxacillin up to 2 months after stopping treatment; risk factors are increasing age and use for more than 14 days (BNF information on flucloxacillin).

  • Cholestatic jaundice can occur with co‑amoxiclav, and is more common in people over 65 years and in men; treatment should not usually exceed 14 days (BNF information on co-amoxiclav).

  • Macrolides should be used with caution in people with a predisposition to QT interval prolongation. Nausea, vomiting, abdominal discomfort and diarrhoea are the most common side effects of macrolides. These are less frequent with clarithromycin than with erythromycin (BNF information on erythromycin).

  • Tetracyclines (for example, doxycycline), can deposit in growing bone and teeth (by binding to calcium) causing staining and occasionally dental hypoplasia. They should not be given to pregnant or breastfeeding women, and use in children under 12 years is either contraindicated or cautioned for use only in severe or life-threatening infections where there are no alternatives (BNF information on doxycycline).

  • Clindamycin has been associated with colitis and diarrhoea. Although this can occur with most antibiotics, it is more frequent with clindamycin. Monitoring of liver and renal function is recommended if treatment exceeds 10 days, and in babies (BNF information on clindamycin).

  • Glycopeptide (for example, vancomycin and teicoplanin) doses are based on body weight. Therapeutic drug monitoring and monitoring of various patient parameters (including blood count, urinalysis, auditory function, hepatic function and renal function) is recommended depending on the particular glycopeptide (BNF information on vancomycin).

  • Severe optic neuropathy can occur with linezolid, particularly if used for longer than 28 days. Blood disorders have also been reported and weekly full blood counts are recommended (BNF information on linezolid).

  • See the EMC's summaries of product characteristics for information on contraindications, cautions and adverse effects of individual medicines.

  • Data on adverse events in the included studies were limited because of cellulitis or erysipelas often being a subgroup in larger skin and skin structure infection studies, where adverse event data were presented for the whole study and not cellulitis or erysipelas subgroups.

  • There were no differences in the adverse events of the following antibiotic comparisons in adults or children with cellulitis or erysipelas:

    • oral cefazolin compared with IV ceftriaxone (Kilburn et al. 2010)

    • oral cefalexin plus oral co‑trimoxazole compared with oral cefalexin alone (Bowen et al. 2017)

    • oral cefalexin or oral clindamycin compared with IV cefazolin or IV clindamycin (Aboltins et al. 2015)

    • oral levofloxacin for 5 days compared with 10 days (Kilburn et al. 2010)

    • IV ceftriaxone compared with IV flucloxacillin (Kilburn et al. 2010)

    • IV daptomycin compared with IV vancomycin (Pertel et al. 2009).

  • Some differences were seen for some adverse event outcomes for the following antibiotic comparisons in adults or children with cellulitis or erysipelas:

    • flucloxacillin plus clindamycin was significantly worse for adverse events (most commonly diarrhoea) compared with flucloxacillin alone (Brindle et al. 2017)

    • IV penicillin was significantly worse for adverse events (no details provided) compared with intramuscular penicillin (Kilburn et al. 2010).

Based on 2 systematic reviews (Bowen et al. 2017 and Kilburn et al. 2010) and 3 RCTs (Brindle et al. 2017, Pertel et al. 2009 and Aboltins et al. 2015).

Committee discussion on choice of antibiotics

  • The committee noted that most antibiotics compared with another antibiotic showed no difference in clinical outcomes in adults or children. The committee also noted that dual therapy was no more effective than single antibiotic therapy in adults. Adverse event data were very limited and there were no differences in adverse events between most of the antibiotic comparisons. Given the very limited amount of evidence in children, the committee agreed that antibiotic choice for children can be extrapolated from the choice for adults.

  • The committee agreed based on their experience that choice of antibiotic treatment should be based on the severity of symptoms and the risk of developing complications, while minimising the risk of the development of antibiotic resistance.

  • The committee discussed that in practice, erysipelas can often be difficult to tell apart from cellulitis and recognised that both infections may be caused by Streptococcus pyogenes or Staphylococcus aureus, although there is uncertainty around the evidence for erysipelas, which may be more associated with streptococcus. Therefore, management of both infections, with regard to antibiotic choice, is the same.

  • The committee were aware that severity scoring tools (for example, Eron 2000 and 'Dundee' Koerner and Johnson 2010) have been developed and may be used in practice. However, these have not to date been used in randomised clinical trials. The committee agreed that recommendations for antibiotic treatment should reflect the available evidence and provide guidance on oral and intravenous treatment because this would fit with current severity scoring tools and the risks of developing complications without needing evidence of the effectiveness of such tools.

  • The committee agreed that the evidence for dual therapy (a combination of 2 antibiotics) showed no benefit over monotherapy for treating cellulitis or erysipelas, and dual therapy should not be routinely used given the increased risk of antimicrobial resistance and more adverse effects.

  • The committee agreed based on the evidence, their experience and resistance data that the first-choice oral antibiotic should be flucloxacillin (a relatively narrow-spectrum penicillin). The committee discussed that flucloxacillin has activity against Staphylococcus aureus (because it is not inactivated by penicillinases produced by staphylococci) and Streptococcus pyogenes. They also agreed that this would be the first-choice antibiotic for people with recurrent infection, because the risk of resistance to flucloxacillin is very low. The only exception would be people with a suspected or confirmed meticillin-resistant Staphylococcus aureus (MRSA) infection, but the committee discussed that the likelihood of such a cellulitis or erysipelas infection with MRSA is very low. The committee agreed that flucloxacillin has poor oral bioavailability and in people with cellulitis or erysipelas who could have impaired circulation (such as people with diabetes or venous insufficiency), a higher (off label) dose of up to 1 g four times a day may be needed to adequately treat the infection. The committee were aware that a narrow-spectrum penicillin with a specific antistreptococcal penicillin is sometimes prescribed for cases of cellulitis or erysipelas, because these infections can involve either streptococci or staphylococci, but there is no evidence that dual therapy is more effective than, for example, flucloxacillin alone. Additionally, the committee considered that dual therapy may increase the risk of antimicrobial resistance and adverse effects.

  • The committee agreed that oral macrolides, clarithromycin or erythromycin (in pregnancy), are suitable alternatives to flucloxacillin in people who have penicillin allergy or where flucloxacillin is not a suitable option. Oral macrolide antibiotics were shown to be at least as effective as an oral penicillin in studies and have a similar spectrum of activity to that of a penicillin. There was limited, very low quality, evidence that oral macrolides or oral streptogramins were more effective than a penicillin (oral or IV). However, the committee considered this evidence was limited because oral macrolides and oral streptogramins were analysed together, not as separate classes. Additionally, the oral streptogramin (pristinamycin) and the only oral penicillin (cloxacillin) used in the studies are not licensed in the UK. There was no head-to-head comparison of either oral macrolides or oral streptogramins with flucloxacillin.

  • The committee discussed the MHRA Public Assessment Report on the safety of macrolide antibiotics in pregnancy. This found that the available evidence is insufficient to confirm with certainty whether there is a small increased risk of birth defects or miscarriage when macrolides are taken in early pregnancy. They agreed with the UK Teratology Information Service monograph on the use of macrolides in pregnancy. They decided that there should be an informed discussion of the potential benefits and harms of treatment. Then, after such a discussion, macrolides can be used if there is a compelling clinical need and there are no suitable alternatives with adequate pregnancy safety data. Erythromycin is the preferred choice if a macrolide is needed during pregnancy, for example, if there is true penicillin allergy and the benefits of antibiotic treatment outweigh the harms. This is because there is more documented experience of its use than for other macrolides.

  • The committee also discussed and agreed that doxycycline (an oral tetracycline) may be useful for people over 12 years who have penicillin allergy or if flucloxacillin is unsuitable. Despite a lack of evidence found for its use, doxycycline is commonly used as an alternative to flucloxacillin for cellulitis and erysipelas in UK practice.

  • The committee discussed and agreed based on limited evidence and their experience that for infection near the eyes or nose (the triangle from the bridge of the nose to the corners of the mouth, or immediately around the eyes), the first-choice oral antibiotic should be the broader-spectrum antibiotic, co‑amoxiclav (a penicillin with a beta-lactamase inhibitor). This is because of the risk of a serious intracranial complication in the event of treatment failure, and because co‑amoxiclav provides cover for Haemophilus and anaerobic bacteria. The committee agreed this extended-spectrum antibiotic was needed to prevent treatment failure and reduce complications of infection in people who are at higher risk because of the location and nature of the infection, and the possibility of uncommon pathogens. For people with infection around the eyes or nose, consulting a specialist was recommended because of the particular risk of complications with this infection site.

  • The committee also agreed that co‑amoxiclav could be used as an alternative first-choice antibiotic in children without penicillin allergy if flucloxacillin was unsuitable. However, if flucloxacillin is only not suitable because of poor palatability of the oral solution, the committee agreed that flucloxacillin capsules should be considered because children can often take tablets or capsules if they are supported to do this (see Medicines for Children leaflet on helping your child to swallow tablets).

  • The committee discussed and agreed based on their experience that although routine dual therapy was not recommended, clarithromycin (a macrolide) with metronidazole (an antibiotic with high activity against anaerobic bacteria) is a suitable alternative to co‑amoxiclav in adults with infection near the eyes or nose, if co‑amoxiclav is not suitable or there is penicillin allergy. In children, the committee discussed that anaerobic bacteria are less of a concern and that clarithromycin alone may be sufficient. However, if anaerobes are suspected, the addition of metronidazole was recommended.

  • The committee agreed based on evidence, their experience and resistance data that the first-choice intravenous antibiotic for people unable to take oral antibiotics or who are severely unwell should be the relatively narrow-spectrum penicillin, flucloxacillin. If flucloxacillin is unsuitable, intravenous clarithromycin is recommended, with intravenous co‑amoxiclav an option for infection near the eyes or nose or severe infection.

  • Based on evidence, their experience and resistance data, the committee also agreed to recommend the following alternative antibiotics for people with severe infection:

    • cefuroxime

    • clindamycin

    • ceftriaxone (in adults for ambulatory care only).

      They discussed that other antibiotics may also be appropriate, particularly in ambulatory care for specific people or populations, and prescribers should seek specialist, local advice.

  • In cases of suspected or confirmed MRSA infection, vancomycin, teicoplanin or if these cannot be used, for specialist use, linezolid, were recommended. They also noted that other antibiotics may be appropriate based on microbiological results and specialist advice.

Antibiotic dose frequency

  • There was no difference in the clinical effectiveness of the following antibiotic comparison in adults with cellulitis or erysipelas:

    • oral cefalexin four times a day compared with twice a day, using the same total daily dose.

Based on 1 systematic review (Kilburn et al. 2010).

  • No systematic reviews or randomised controlled trials in children met the inclusion criteria.

Antibiotic course length

  • There were no differences in the clinical effectiveness of the following antibiotic comparisons in adults with cellulitis:

    • oral tedizolid for 6 days compared with 10 days (Hanretty et al. 2018)

    • oral levofloxacin for 5 days compared with 10 days (Kilburn et al. 2010).

Based on 2 systematic reviews (Hanretty et al. 2018 and Kilburn et al. 2010).

  • No systematic reviews or randomised controlled trials in children met the inclusion criteria.

Antibiotic route of administration

  • There were no differences in the clinical effectiveness of the following antibiotic comparisons in adults with cellulitis:

    • oral cefalexin or oral clindamycin compared with IV cefazolin or IV clindamycin (Aboltins et al. 2015)

    • IV benzylpenicillin compared with intramuscular benzylpenicillin (Kilburn et al. 2010).

Based on 1 systematic review (Kilburn et al. 2010) and 1 RCT (Aboltins et al. 2015).

  • No systematic reviews or randomised controlled trials in children met the inclusion criteria.

Committee discussion on antibiotic dose frequency, course length and route of administration

  • The committee acknowledged that there was very limited evidence identified for antibiotic dose frequency.

  • The committee agreed that the shortest course that is likely to be effective should be prescribed to reduce the risk of antimicrobial resistance and minimise the risk of side effects.

  • Based on limited evidence and their experience, the committee agreed that a shorter course of antibiotics was generally as effective as a longer course of antibiotics for cellulitis or erysipelas, and a 5- to 7‑day course was sufficient for most people. However, the committee noted that only 1 RCT (Brindle et al. 2017) used flucloxacillin for 5 days and the authors expressed doubts about the quality of their data for dose and duration. Therefore, the committee discussed that the decision of whether a 5‑day or a 7‑day course was given would be based on clinical judgement of individual cases. The committee discussed that a longer course (up to 14 days in total) may be needed for some people based on a clinical assessment of their symptoms and history. However, skin does take some time to return to normal, even after an effective course of antibiotics, and a full resolution of symptoms at 5 to 7 days would not be expected.

  • Based on limited evidence, the committee agreed that oral antibiotics were as effective as intravenous antibiotics for treating cellulitis and erysipelas.

  • 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, for a total of 5 to 7 days. Again, a longer course (up to 14 days in total) may be needed for some people based on clinical assessment.

Antibiotic prophylaxis for the prevention of recurrent cellulitis and erysipelas

  • Antibiotic prophylaxis (with an intramuscular or oral penicillin, or oral erythromycin) significantly lowered the risk of recurrence of cellulitis or erysipelas compared with no treatment or placebo in a meta-analysis of 5 RCTs in adults (approximately 46 to 70 years) with 1 or 2 previous episodes of cellulitis or erysipelas in the past 3 months to 3 years depending on the RCT.

  • Antibiotic prophylaxis significantly lowered the incidence rate (episodes per person month) compared with no treatment or placebo in a meta-analysis of 4 RCTs in adults with 1 or 2 previous episodes of cellulitis or erysipelas.

  • Antibiotic prophylaxis significantly lowered the risk of an episode (time to next episode) of cellulitis or erysipelas compared with no treatment or placebo in a meta-analysis of 3 RCTs in adults with 1 or 2 previous episodes of cellulitis or erysipelas.

  • Antibiotic prophylaxis was not significantly different to no treatment or placebo for mortality or risk of hospitalisation in adults with 1 or 2 previous episodes of cellulitis or erysipelas.

  • There were no differences in the adverse events of antibiotic prophylaxis compared with no treatment or placebo in adults with cellulitis or erysipelas.

Based on 1 systematic review (Dalal et al. 2017).

  • No systematic reviews or randomised controlled trials in children met the inclusion criteria.

Committee discussion on antibiotic prophylaxis of recurrent cellulitis or erysipelas

  • The committee noted that recurrence is not uncommon in people who have had cellulitis or erysipelas. The committee also noted the limitations of the evidence for antibiotic prophylaxis, which was in adults only and mainly related to lower limb cellulitis. There was variation in the populations in the included studies for the number of previous episodes (1 or 2) and the time periods over which recurrence was defined (up to 3 years).

  • The committee discussed the evidence for prophylactic antibiotics in adults. Overall, antibiotics reduced the risk of cellulitis or erysipelas recurring but did not reduce the risk of hospitalisation or mortality, and the long-term effects on antibiotic resistance are unknown.

  • The committee agreed based on evidence and experience that antibiotic prophylaxis should not be routinely offered to prevent recurrent cellulitis or erysipelas because of the balance of risks and benefits in the overall population.

  • However, they agreed based on evidence and experience that a trial of antibiotic prophylaxis could be considered for a higher-risk population, which the committee defined as adults who have had at least 2 separate episodes of cellulitis or erysipelas in the previous 12 months, which were managed in hospital, or where the care was under specialist advice. The populations in the antibiotic prophylaxis trials were more varied (having 1 or 2 previous episodes over up to 3 years), but the committee wanted to ensure that prophylaxis would only be considered for those at highest risk.

  • The committee agreed that prophylaxis should only be considered following a discussion between a specialist and the person to ensure shared decision making, and should be reviewed every 6 months. Prophylaxis may be appropriate in this higher-risk population because the benefits of prophylaxis may outweigh the risks. However, it is important to ensure that the previous episodes of cellulitis and erysipelas have been correctly diagnosed, any underlying condition (such as oedema, diabetes or venous insufficiency) is being managed optimally, and prophylaxis is reviewed at least every 6 months.

  • The choice of antibiotic was low-dose phenoxymethylpenicillin, which was used in most of the trials in the systematic review, or low-dose erythromycin in penicillin allergy, which was used in 1 trial in the systematic review. The committee discussed that alternative antibiotics may be appropriate with specialist advice, and that choice should be based on recent microbiological results where possible. Based on their experience and resistance data, the committee agreed that using the same antibiotic for treatment and prophylaxis should be avoided.

  • The committee recognised the importance of reviewing antibiotic prophylaxis, and considered that up to every 6 months was reasonable based on possible adverse effects of antibiotics, the risk of resistance with long-term antibiotics, the possible need for any further investigations if recurrence of cellulitis or erysipelas, and to allow time to assess treatment success. People should also know to seek medical help if cellulitis or erysipelas recurs despite taking prophylaxis.

  • To reduce the risk of antimicrobial resistance, the committee agreed that each review should include a discussion around the success of prophylaxis and whether antibiotics should be continued, stopped or changed, taking into account the person's preferences for antibiotic use and the potential risk of antimicrobial resistance with long-term use of antibiotics. If treatment failure occurs and cellulitis or erysipelas recurs, the committee agreed that antibiotic prophylaxis should be stopped or changed to an alternative prophylactic antibiotic once the acute infection has been treated.

  • No recommendation for antibiotic prophylaxis in children was made because there was no evidence in this population.