Authors: Taylor M. Douglas MD, Taylor Conrad, MD, Ted Segarra, MD, Rithvik Balakrishnan MD, Christianna Sim, MD, MPH
Editors: Kelly Maurelus MD FAAEM, Kami Hu MD FAAEM
Originally published: Common Sense
November/December 2019
Skin and soft tissue infections (SSTIs) result in over two million visits to the emergency department (ED) every year. While this term encompasses infections ranging from erysipelas to necrotizing fasciitis, this article focuses on superficial cellulitis. Due to the lack of good data, there is no true consensus in the medical community regarding standard of care, i.e. whether intravenous (IV) antibiotics are required or what clinical presentations mandate admission. The Infectious Disease Society of America (IDSA) makes recommendations for treatment duration (five days), when to cover methicillin-resistant Staphylococcus aureus (MRSA) [penetrating wounds, intravenous drug use, systemic inflammatory response syndrome (SIRS), nasal colonization, evidence of other MRSA infection] and outpatient management (for those without SIRS, altered mental status, or hemodynamic instability) 1. Most of these guidelines, however, are based on retrospective studies. Studies in the surgical field have attempted to identify grading systems to help guide management without success.2 Without consensus, emergency physicians are left with the following questions on how to best treat and disposition our patients with cellulitis to ensure their infection resolves.
- What is the evidence behind treatment of cellulitis with IV antibiotics and which patients should receive them?
- What risk factors have been identified to predict outpatient failure and the need for inpatient treatment of cellulitis?
Peterson D, McLeod S, Woolfrey K, McRae A. Predictors of failure of empiric outpatient antibiotic therapy in emergency department patients with uncomplicated cellulitis. Acad Emerg Med. 2014;21(5):526-31. doi: 10.1111/acem.12371.
Peterson et al. aimed to identify risk factors that predict failure of initial outpatient antibiotic therapy in patients presenting to the ED with uncomplicated cellulitis. They performed a prospective cohort study conducted at two academic tertiary care center EDs in London, Ontario over an 18-month period, including adults 18 years or older who were diagnosed with cellulitis in the ED. They designed a multivariate, logistic regression model including variables based on the known epidemiology of cellulitis as well as the hypothesized relationship between potential risk factors and treatment failure. Exclusion criteria included hospital admission, current or recent treatment for cellulitis prior to ED presentation, and abscess without concomitant cellulitis. The primary outcome – treatment failure – was defined as a required change in antibiotic therapy or subsequent hospitalization for cellulitis within two weeks after initial presentation.
The study enrolled 598 patients; 52 were excluded and 49 were lost to follow-up. Of the 497 included in the final analysis, 102 (20.5%) had treatment failure but only 21.6% of these required subsequent hospitalization. Failure rates were similar in those treated with oral (38.2%, 95% CI 29.4-47.9%) and IV (40.2%, 95% CI 31.2-49.9%) antibiotics, but lower in those who had received both (21.6%, 95% CI 14.7-30.5%). Five covariates were identified as statistically significant predictors of failure of empiric outpatient antibiotic therapy:
1) fever at triage >38°C (odds ratio [OR] 4.3, 95% CI = 1.6-11.7)
2) chronic leg ulcers (OR 2.5, 95% CI = 1.1-5.2)
3) chronic edema or lymphedema (OR 2.5, 95% CI = 1.5-4.2)
4) prior cellulitis in same area (OR 2.1, 95% CI = 1.3-3.5)
5) cellulitis at a wound site (OR 1.9, 95% CI = 1.2-3.0)
Covariates included in the study found NOT to have a significant association with treatment failure were heart rate at triage > 90 beats/min, diabetes mellitus, smoking and obesity.
Abetz JW, Adams NG, Mitra B. Skin and soft tissue infection management failure in the emergency department observation unit: a systematic review. Emerg Med J. 2018;35(1):56-61. doi: 10.1136/emermed-2016-205950.
The past twenty years have seen a dramatic rise in the number of observation or short-stay units designated for patients who are not necessarily sick enough to require inpatient admission but will require fewer than 24 or 48 hours of observation and management prior to discharge. SSTIs are one of the common diagnoses that warrant placement in an ED observation unit (EDOU). Abetz et al. conducted a systematic review to quantify the rate of and identify patient risk factors for management failure.
The authors searched for manuscripts using the outcome of management failure (inpatient admission, stay >28 hours, or death). Studies were excluded if patients were observed on an outpatient basis, if management failure was defined as ED revisits or re-presentation, or if the studies did not differentiate between SSTIs and other conditions in the observation unit. All of the studies that met eligibility criteria reported management failure rates (15%-38%) higher than generally accepted EDOU failure rates (15%), with higher failure rates in the later studies than in the earlier studies. The most commonly identified risk factors for treatment failure in the separate studies were MRSA infection (OR 4.2, 95% CI 1.4-12.3) or exposure (OR 1.9, 95% CI 1.1-3.4), subjective fever (OR 3.02, 95% CI 1.41-6.43), history of fever (OR 2.3, 95% CI 1.7-3.1), or measured fever with temperature >38°C (OR 2.5, 95% CI 1.1-5.5, and WBC count >15,000 (OR 4.06, 95% CI1.53-10.74).
Limitations of the study include the fact that only one study reported outcome blinding, and the vast majority of the patient population (206,000 patients) came from a single study by Venkatesh et al. The overall level of evidence was low; three of the 10 studies were rated as NHMRC (National Health and Medical Research Council) evidence Level II, and the remaining seven were Level III-C (including the Venkatesh study) or lower. Only two of the ten studies commented on culture-positive MRSA infection, three on the number of purulent SSTIs, three on SSTI anatomic location, and one on antibiotic choice. This contributed to the marked inter-study heterogeneity (I2 = 93.9%, p<0.01) which, combined with the low levels of evidence among the identified studies, limits any definitive conclusions that can be taken from the analysis.
Aboltins CA, Hutchinson AF, Sinnappu RN, et al. Oral versus parenteral antimicrobials for the treatment of cellulitis: a randomized non-inferiority trial. J Antimicrob Chemother. 2015;70(2):581-6.
While some institutions have outpatient IV infusion clinics, these are not common; a big decision point in determining the disposition for patients with cellulitis is the need for IV antibiotics. Perceived severity of disease is often a factor in the decision, along with patient comorbidities or the character of the SSTI, but no consensus exists on which patients are safe to discharge on oral antibiotics. A Cochrane study in 2010 confirmed that data was limited in this area.3
In this prospective, randomized, open-label, non-inferiority single center study out of Australia, patients with cellulitis who had been determined by treating providers to require IV antibiotic therapy were randomized to either IV cefazolin or oral cephalexin. The diagnosis of cellulitis was defined as characteristic skin findings present for < 5 days with pain, temperature ≥37.88ºC, heart rate >90 beats/min, systemic symptoms, or elevated inflammatory markers. Patients were excluded if the cellulitis was mild, if they had complicated or severe disease (bullous disease, severe sepsis, presence of abscesses, necrotizing fasciitis), or if it was associated with trauma, patient immunosuppression, or was periorbital. Patients who had already received antibiotics prior to randomization were included if oral therapy was < 48 hours or IV therapy was < 12 hours. If a patient was allergic to penicillin, clindamycin was used instead. The primary outcome was the duration of time until no advancement of the area of cellulitis, at which point patients on IV therapy were transitioned to oral. Secondary outcomes included failure of treatment (defined as requiring a change in or prolonged course of antibiotics, readmission to the hospital, or abscess drainage), pain, complications, and patient satisfaction.
Of the 47 patients enrolled, one patient in the parenteral treatment arm was lost to follow-up and one patient in the oral arm received clindamycin due to development of a rash. There was no statistically significant difference between IV and oral therapy the primary outcome, suggesting oral treatment was non-inferior to parental therapy. A large portion of patients (43%) had received antibiotic therapy prior to enrollment, but seeing as the cutoff was 48 hours, this likely represents a different cohort than those who received antibiotics prior to ED presentation. The inclusion of patients with signs of systemic illness who were slated to receive IV therapy is a strength of this study. Major weaknesses of this study are its very small size and that the primary outcome was non-progression of disease, rather than treatment failure, which is arguably more important to providers in decision-making. Interestingly, although the study was not powered for treatment failure, it trended towards favoring oral therapy. Lastly, the authors note in the discussion that their community does not have a high rate of community-acquired MRSA, which may affect overall generalizability as many communities in the U.S. do have a high prevalence of MRSA. This study emphasizes the need for larger validation studies that can help guide the use of oral antibiotics for cellulitis, even in those with signs of systemic disease, but does suggest that many patients likely can be safely treated with oral therapy.
Yarbrough PM, Kukhareva PV, Spivak ES, et al. Evidence-based care pathway for cellulitis improves process, clinical, and cost outcomes. J Hosp Med. 2015;10(12):780-6. doi: 10.1002/jhm.2433. Epub 2015 Jul 28.
In order to address the wide variation in antibiotic choices for cellulitis and overuse of broad-spectrum antibiotics and advanced imaging, Yarbrough et al. designed and implemented an EMR-based care pathway for the management of cellulitis with the goal of improving clinical, logistical, and economic outcomes.
The pathway (shown below) was primarily based on guidelines from the IDSA for the management of SSTIs and infections caused by MRSA. It focuses heavily on antibiotic selection, indications for obtaining blood cultures and advanced imaging, and frequently overlooked patient care principles (i.e. elevation of the affected limb). Similar to the 2014 IDSA updates, the pathway made clear distinctions between purulent and non-purulent cellulitis in an attempt to distinguish between staphylococcal and streptococcal infections. It additionally excluded patients with more complicated infections, including those with neutropenia, osteomyelitis, diabetic foot ulcerations, cellulitis of the hand, perineum, or periorbital region, surgical site infections, and human or animal bites. Use of the pathway was not mandatory, but all providers were repeatedly educated regarding the pathophysiology and management of cellulitis, the IDSA guidelines, and the availability of the pathway order set.
The authors designed a single-center, retrospective, observational, pre/post-intervention study at a 500-bed academic center in Salt Lake City, Utah. They included all patients over the age of 18 who were either admitted or placed on observation with a diagnosis of cellulitis over a 2½ year period. Exclusion criteria were discharge from the ED and diagnosis other than cellulitis. The primary outcome was use of a broad spectrum antibiotic (vancomycin, piperacillin/tazobactam, or meropenem). Secondary outcomes included computed tomography (CT) or magnetic resonance imaging (MRI) use, length of stay (LOS), 30-day readmission (excluding visits for diagnoses other than SSTI), and pharmacy, lab, imaging, and total facility costs. A total of 677 visits met inclusion criteria, of which 370 (54.6%) employed the use of the order sets and 307 (45.3%) did not.
The authors noted a significant hospital-wide decrease in the odds of ordering broad spectrum antibiotics (59% decrease, P<0.001) and decrease in pharmacy costs (23% decrease, P=0.002). Overall total facility costs were also significantly decreased (13% decrease, P=0.006), but this decrease was actually driven by the non-intervention group (in which order sets were not used), rather than by the intervention group.
In the subpopulation of patients in whom the order sets were employed, the authors noted statistically significant reductions in the odds of using broad spectrum antibiotics (75% decrease, P<0.001), pharmacy costs (25% decrease, P=0.074), and clinical LOS (13% decrease, P=0.041). However, the absolute differences in LOS were minimal and had notable overlap between groups (2.0 +/- 2.1 days vs 1.7 +/- 1.6 days). Results were similar between those patients who were admitted and those placed on observation.
The authors concluded that implementing a care pathway along with education, pathway-compliant electronic order sets, and audit and feedback can reduce costs and improve the quality of care, without increasing readmission rates. They further hypothesized that the reduction in broad spectrum antibiotic use could potentially reduce rates of Clostridium difficile, but admit that this outcome was not studied in their analysis. Nevertheless, the paper has several limitations in addition to those typically found in retrospective, observational studies. Most notably, there was an ongoing parallel trial during the post-intervention period aimed at reducing laboratory use, and this important confounder unfortunately invalidates any pre-/post-intervention laboratory use comparisons, and likely invalidates the pre-post/intervention total facility costs comparison as well. Overall, this was a fairly well-designed retrospective study, and demonstrates a role for care-pathways to reduce broad spectrum antibiotic use, but its broader applicability and clinical implications are limited by its inherent flaws and confounders.
Conclusion
The studies highlighted above demonstrate the need for further research into the management and disposition of patients with cellulitis. Based on the current guidelines as well as the designs and results of the reviewed studies, consideration of patient-based risk factors such as chronic skin conditions and comorbidities should factor into the treatment of SSTIs. Interestingly, most papers reported cefazolin and ceftriaxone (IV) and cephalexin (PO) as the most common antibiotics used, with clindamycin as a second line agent in patients with allergies. The use of antibiotics that cover MRSA should be made based on local prevalence and susceptibility trends, but are likely unnecessary in simple non-purulent cellulitis. Regardless of treatment choice, the IDSA recommends reassessment for treatment failure 48 to 72 hours after initiating treatment1, so close primary care or consideration of scheduled ED follow-up may play a role in outpatient management.
Now we return to the questions posed in the introduction:
- What is the evidence behind treatment of cellulitis with IV antibiotics and which patients should receive them?
There is no clear evidence indicating which patients should receive IV antibiotics, however even patients demonstrating systemic symptoms may do well with oral therapy. - What risk factors have been identified to predict outpatient failure and the need for inpatient treatment of cellulitis?
Predictors of failed outpatient therapy include fever at triage, chronic leg ulcers, edema or lymphedema, prior cellulitis at the same site, and cellulitis at a wound site, while fever, MRSA infection or exposure, and leukocytosis >15,000 are the most commonly identified factors predicting treatment failure in ED observation units.
References
- Stevens DL, Bisno AL, Chambers HF, Dellinger EP, Goldstein EJC, Gorbach SL, Hirschmann JV, Kaplan SL, Montoya JG, Wade JC. Practice Guidelines for the Diagnosis and Management of Skin and Soft Tissue Infections: 2014 Update by the Infectious Disease Society of America. Clin Inf Dis. 2014;59(2):e10-52.
- Savage SA, Li SW, Utter GH, Cox JA, Wydo SM, Cahill K, Sarani B, Holzmacher J, Duane TM, Gandhi RR, Zielinski MD, Ray-Zack M, Tierney J, Chapin T, Murphy PB, Vogt KN, Schroeppel TJ, Callaghan E, Kobayashi L, Coimbra R, Schuster KM, Gillaspie D, Timsina L, Louis A, Crandall M. The EGS grading scale for skin and soft-tissue infections is predictive of poor outcomes: a multicenter validation study. J Trauma Acute Care Surg. 2019 Apr;86(4):601-608.
- Kilburn SA, Featherstone P, Higgins B et al. Interventions for cellulitis and erysipelas. Cochrane Database Syst Rev. 2010; issue 6