Clinical response to antibiotic regimens in lower limb cellulitis: a systematic review

There is variation in the treatment of lower limb cellulitis (LLC) with no agreement on the most effective antibiotic regimen. Many patients with cellulitis fail to respond to first‐line antibiotics. This can negatively affect patient care and result in unnecessary hospital admissions. The aim of this systematic review was to determine the clinical response and safety of antibiotic regimens for the management of LLC. A systematic review for randomized controlled trials (RCTs) was conducted using OVID MEDLINE, Ovid Embase and Cochrane Central Register of Controlled Trials in January 2019. Outcomes of interest included the clinical response to antibiotic regimens (type, dose, route, duration) and the safety of antibiotics in LLC. Trial quality was identified using the Cochrane Risk of Bias tool. Four RCTs were included. All included studies showed no significant differences between the clinical response to different antibiotic type, administration route, treatment duration or dose. LLC may be overtreated and shorter courses of oral antibiotics, possibly with lower doses, may be more suitable. There is a lack of published data on the clinical response and safety of antibiotics in LLC. Three studies were high risk for bias overall. Further high‐quality studies may help determine whether less intensive antibiotic regimens can effectively treat LLC.


Introduction
During 2017, 88 664 National Health Service (NHS) patients were admitted to hospitals in the UK with cellulitis, receiving inpatient treatment costing £226m. 1,2 Without effective treatment, cellulitis may cause sepsis and recurrent disease. 3 Given the prevalence and consequences of cellulitis, it is essential to effectively manage this condition.
The antibiotic choice to treat cellulitis is influenced by hospital guidelines, causative bacteria and clinical experience. 4,5 One review identified 25 different antibiotic regimens across 5 emergency departments. 6 A study of over 100 hospitals showed that 16.6% of patients with acute cellulitis experienced initial antibiotic failure. 7 A previous review highlighted a lack of high-quality studies and was unable to define best treatment for cellulitis. 4 However, it was conducted 9 years ago and did not focus on the lower limb, which is affected by cellulitis in 66% of cases. 8 Lower limb cellulitis (LLC) may behave differently from cellulitis at other sites, owing to differences in circulation and flora.
The 2017 Cellulitis Priority Setting Partnership (PSP), emphasized concern from patients and clinicians for more conclusive treatment guidelines. 9 This review aimed to better define how best to treat LLC.

Reporting
This review was reported in line with the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) statement. The protocol was registered with PROSPERO in January 2019 (CRD42019116416).

Outcome
The primary outcome was the clinical response to the antibiotic regimen (type, dose, route, duration). Clinical response was operationalized using validated outcome measures, 10,11 namely, symptom response, laboratory markers, therapeutic failure and quality of life. One of the authors (PS) is a patient expert, and ensured the review was patient-focused. The secondary outcome was the frequency of adverse events.

Study types
Studies included randomized conrolled trials (RCTs) of patients with LLC, in which the clinical response to antibiotics was evaluated. No restrictions on study date or patient demographics were applied. The exclusion criteria included: manuscript not in English; use of prophylactic antibiotics; and presence of eosinophilic cellulitis. Full eligibility and exclusion criteria are given in Table S1 online.

Search strategy
The search (Table S2) was developed with an information specialist (DG) and cellulitis expert (NJL). Ovid MEDLINE, Ovid Embase and the Cochrane Central Register were searched on 5 January 2019 using key terms (Table 1). Grey literature was identified using National Institute for Health and Care Excellence evidence and Google Scholar. The World Health Organization (WHO) International Clinical Trials Registry Platform was screened for unpublished studies. Deduplication was performed using EndNote X9 (https:// endnote.com). The reference lists of included studies were scanned for eligible studies.

Analysis
Eligibility screening, data extraction and quality assessment was performed independently by KM and SS. Disagreements were resolved through a third reviewer (NJL). Data was entered into Revman 5.3 (Cochrane Library). Studies were described in narrative synthesis. Methodological quality was assessed using a modified version of the Cochrane risk of bias (RoB) tool. 12 For each domain a study was assessed as 'low risk' if all signalling questions were 'yes', 'high risk' if at least one signalling question was 'no' or 'unclear' if not reported. If a study was 'high risk' in any domain, it was considered 'high risk of bias' overall. If a study was 'unclear risk' in multiple domains, it was considered to have 'some concerns of bias' overall. 12

Results
The PRISMA diagram shows the search result ( Fig. 1). Only four studies were eligible ( Table 2). [13][14][15][16] In total, 529 patients were included across 13 countries. Two studies were based in hospitals. 13,15 No studies reported who had made the diagnosis or the duration of LLC. Two studies defined LLC diagnosis. 15,16 Two studies specified LLC severity and excluded patients with mild disease. 14,16 No studies compared the same antibiotic regimen, hence I 2 and meta-analysis was inappropriate.
Leman and Mukherjee showed that the addition of intravenous (IV) benzylpenicillin to IV flucloxacillin did not result in a more rapid clinical response (Tables 3 and 4). 13 Peterson et al. demonstrated that 2 of 8 patients (25%) on ciprofloxacin 1500 mg and 6 of 8 (75%) on 2000 mg achieved a long-term satisfactory response [relative risk (RR) = 0.33, 95% CI 0.09-1.18]. 16 This raises the possibility that in some patients lower doses of ciprofloxacin could potentially be used with no reduction in clinical response. 16 Joseph et al. 14 performed a pooled analysis of the ESTABLISH 1 and 2 trials. 17,18 In the tedizolid group 114 of 162 patients (70.4%) achieved early clinical response compared with 115 of 158 (72.8%) in the  14 Tedizolid offered a short and well-tolerated treatment. 14 Zeglaoui et al. revealed that 11 of 55 patients (20%) on IV benzylpenicillin and 8 of 57 (14%) on intramuscular (IM) bipenicillin had failure of treatment at Day 10 based on their symptoms (RR = 1.43, 95% CI 0.62-3.27). 15 Mean time to recovery was 6.3 days in patients on IV benzylpenicillin compared with 6.5 days in patients on IM bipenicillin (t-test, P = 0.75). 15 This suggested that IM bipenicillin could be a possible alternative to IV benzylpenicillin. 15 Zeglaoui et al. also found that 9.1% (4/44) on IV benzylpenicillin and 7% (3/44) on IM bipenicillin experienced local skin complications such as cutaneous necrosis (RR = 1.33, 95% CI 0.32-5.61). 15 The authors reported that 14/44 (25.5%) patients on IV benzylpenicillin experienced venitis, whereas no patients on IM bipenicillin did (RR = 29.00, 95% CI 1.78-471.58). 15 By contrast, Leman and Mukherjee identified no adverse effects for benzylpenicillin. 13 RoB was high for three of these studies, 13,15,16 and there were some concerns about RoB in the fourth 14 ( Fig. 2). One study did not provide the baseline characteristics. 16 Leman and Mukherjee reported that 18 of 99 participants (18%) withdrew from the study, and identified deviation from standard practice due to the early discharge of participants. 13 The study of Peterson et al. was underpowered to show a significant difference. 16 Two studies used subjective outcome measures. 15,16

Discussion
All included studies showed no significant differences between groups related to antibiotic types, administration routes (IV and alternative routes), durations of treatment (long vs. short) or dosages (high vs. low). This suggested that LLC may be overtreated in some patients. Another review, which included sites other than the lower limb also showed oral route, lower dosages and shorter durations of antibiotics to be equally effective. 19 The first surveillance data by the WHO on antimicrobial resistance, published in 2018, revealed that penicillin resistance ranged from 0 to 51% globally. 20 The first-line antibiotic in the UK is flucloxacillin, 5 but this is ineffective against methicillin-resistant Staphylococcus aureus (MRSA) and its broad-spectrum action potentiates resistance. 21   complicated by allergic reactions and Clostridium difficile. 23 With the increase in rates of resistance and adverse events, it is important to identify new antibiotics. Fear of MRSA and poor outcomes may have driven overtreatment in the cases reported. 24 A previous review revealed 14% of cellulitis admissions were overtreated. 25 The main limitation of this review was the number and quality of included studies. Although an abundance of cellulitis literature existed, few focused on the lower limb. As this is the first systematic review focused on the antibiotic treatment of LLC, the less stringent eligibility criteria provided a comprehensive scope of the current literature.
This review identified a profound gap in the literature regarding high-quality studies to identify best treatment of LLC. Future studies could explore whether we can effectively treat mild LLC with less intensive antibiotic regimens. Such regimens would save health services money, reduce treatment complications and contribute to guidelines based on patient characteristics.

Conclusion
The evidence for the clinical response to antibiotics for LLC is limited; however, there is low-quality evidence to support the possible use of shorter courses of lower doses of oral antibiotics in some patients. Further trials comparing lower-intensity antibiotic regimens are warranted.

Acknowledgement
We thank Dr L. Hooper for help with the design of the study, Mr J. Lossaso for helping to identify papers from the search strategy and Professor K. Thomas for helpful advice regarding the research topic.

Learning points
• Cellulitis is a common condition that can have a negative impact on patients; however, the clinical response of different antibiotic treatment regimens remains unclear.
• There is variation in the treatment of LLC, with many patients failing to respond to first-line antibiotics, and an increase in antibiotic resistance.
• There is weak evidence that LLC may be overtreated in some patients, for whom shorter courses of lower-dose, oral antibiotics may be suitable.
Users are encouraged to • Read the article in print or online, paying particular attention to the learning points and any author conflict of interest disclosures. • Reflect on the article. • Register or login online at http://www.wileyhealth learning.com/ced and answer the CPD questions. • Complete the required evaluation component of the activity.
Once the test is passed, you will receive a certificate and the learning activity can be added to your RCP CPD diary as a self-certified entry. This activity will be available for CPD credit for 2 years following its publication date. At that time, it will be reviewed and potentially updated and extended for an additional period.

Supporting Information
Additional Supporting Information may be found in the online version of this article: