This month I decided to do a bit of a lit review on a topic that’s been bugging me recently in our department. It’s not a new or landmark clinical trial, but a topic that I think deserves some attention and discussion! In our busy summer department, I’m often surprised at how many patients are brought back for standing orders for IV antibiotics (often Ceftriaxone!) for moderate cellulitis.  There is a lot of variability in who is given IV vs PO antibiotics, and for what indications, so I thought I would look at the evidence. 

TL:DR –  (“too long, didn’t read” for the non-millennials!)

• In multiple (although small) studies, no difference in clinical resolution of cellulitis has been demonstrated between IV and oral antibiotics for simple/uncomplicated cellulitis. 
• The IDSA recommendation that IV antibiotics for nonpurulent cellulitis be reserved for patients who are immunocompromised or have systemic signs of infection, hemodynamic instability, or altered mental status.  
• Even one dose of IV antibiotics given in the ED can increase risk of antibiotic associated diarrhea and C.diff infection (25.7% vs 12.3% for PO Rx only)
• PO availability of most common antibiotics is excellent – Keflex 90-100%, Clinda 90%
  • A RCT comparing resolution of cellulitis with PO Keflex vs IV Ancef + Probenecid showed NO benefit of IV therapy.
  • A previous RCT comparing IV Ceftriaxone to IV Ancef + Probenecid showed no superiority of Ceftriaxone over the Ancef… 
• So… most of the time can we just go with Keflex?! Save the patient the time for the ER visit, the nursing time, the antibiotic use, the diarrhea risk, the risk of IV complications… and ensure appropriate outpatient follow-up in 72h.

Hey, this patient has cellulitis. Should I use Keflex or Ancef?

Study 1: Oral versus parenteral antimicrobials for the treatment of cellulitis: a randomized  non-inferiority trial 

Craig A. Aboltins, Anastasia F. Hutchinson, Rabindra N. Sinnappu, Damian Cresp, Chrissie Risteski, Rajasutharsan Kathirgamanathan, Mark A. Tacey, Herman Chiu, Kwang Lim, Oral versus parenteral antimicrobials for the treatment of cellulitis: a randomized non-inferiority trial. Journal of Antimicrobial Chemotherapy. Volume 70, Issue 2, February 2015, Pages 581–586, https://doi.org/10.1093/jac/dku397

This was a randomized trial done in 2015 in a single centre ED. Patients age 16+ were referred by the ED staff for treatment of cellulitis with IV antibiotics – i.e. staff thought the severity of cellulitis warranted IV therapy or that patient had failed outpatient treatment with PO antimicrobials. Cellulitis was defined by the presence of acute dermal/epidermal inflammation lasting <5 days and associated with pain, fever with a temperature of ≥37.8°C, tachycardia >90 beats/min, systemic symptoms or elevated inflammatory markers (pretty clinically clear definition, and probably includes the more severe end of the spectrum, people I would also automatically assume need IV therapy!) Patients were excluded for complicated cellulitis (defined as severe sepsis, extensive bullous skin changes or abscess formation), necrotizing fasciitis, periorbital cellulitis, cellulitis complicating trauma, immunosuppressed patients, mild cellulitis (defined as limited area and no systemic symptoms), vomiting precluding oral treatment, or prior treatment with oral antimicrobials for >48h or with IV for >12h. Patients were assigned to either Keflex 1g PO QID x 10 days (interestingly a higher dose than our usual 500mg PO QID), or Ancef 2g IV q12h x 10d, with a change to PO indicated when afebrile and area of cellulitis stopped progressing. The primary outcome was time until no advancement of the area of cellulitis. Secondary outcomes were failure of treatment, pain, complications of treatment and satisfaction with care. Only 47 patients were included in the trial, but it showed that PO therapy was non-inferior to IV therapy for uncomplicated cellulitis.

Study 2: Intravenous cefazolin plus oral probenecid versus oral cephalexin for the treatment of skin and soft tissue infections: a double-blind, non-inferiority, randomised controlled trial.

Dalen D, Fry A, Campbell SG, Eppler J, Zed PJ. Intravenous cefazolin plus oral probenecid versus oral cephalexin for the treatment of skin and soft tissue infections: a double-blind, non-inferiority, randomised controlled trial. Emerg Med J. 2018 Aug;35(8):492-498. doi: 10.1136/emermed-2017-207420. Epub 2018 Jun 18. 

This was a prospective multicentre RCT including 206 patients comparing IV Ancef + Probenecid vs PO Keflex for the treatment of uncomplicated cellulitis in two Canadian EDs. Patients age 19+ were included if they presented to the ED with a presumed diagnosis of mild-moderate cellulitis as assessed by the attending physician, were deemed well enough to have outpatient treatment, and could return to the ED daily for assessments. There were a bunch of relevant exclusion criteria: patients were excluded if oral antibiotic therapy was indicated but intravenous antibiotics were deemed to not be required by the emergency physician (ie, mild enough to be given oral therapy and sent home without follow-up in the ED). Patients with known chronic kidney disease (creatinine clearance <30 mL/min), known previous methicillin-resistant Staphylococcus aureus (MRSA) infection, use of antibiotics for >24 hours in the past seven days, wound/abscess requiring operative debridement or incision and drainage (either via clinical exam and/or ultrasound examination), suspected necrotising fasciitis, osteomyelitis or septic arthritis, febrile neutropenia, concomitant documented bacteraemia, two or more signs of systemic sepsis, infections at a site involving prosthetic materials, animal or human bite wound infections, postoperative wound infections, known peripheral vascular disease, superficial thrombophlebitis, pregnant/breast feeding, obesity (BMI>30 ) or a known allergy to study medication. Importantly, local antibiograms at both centres in the study reported MRSA rates <20%, similar to our centre.  Patients were randomized to receive either Keflex 500mg PO QID at home + Placebo pill + IV NS in the ED daily, or Ancef 2g IV + Probenecid 1g PO daily in the ED + PO Placebo QID at home. Primary outcome was treatment failure at 72h. Secondary outcomes were clinical cure at 7 days, admission to hospital and adverse events.  There were no significant differences in any outcomes. Failure rate at 72h was 4.2% (PO) vs 6.1% (IV). Conclusion: Cephalexin at appropriate doses appears to be a safe and effective alternative to outpatient parenteral cefazolin in the treatment of uncomplicated mild-moderate SSTIs who present to the ED.

But what if I really think this patient needs IV therapy? Should I just put them on Ceftriaxone?

Study 3: Once-Daily Intravenous Cefazolin Plus Oral Probenecid Is Equivalent to Once-Daily Intravenous Ceftriaxone Plus Oral Placebo for the Treatment of Moderate-to-Severe Cellulitis in Adults

M. Lindsay Grayson, Malcolm McDonald, Kimberley Gibson, Eugene Athan, Wendy J. Munckhof, Phillip Paull, Fran Chambers. Once-Daily Intravenous Cefazolin Plus Oral Probenecid Is Equivalent to Once-Daily Intravenous Ceftriaxone Plus Oral Placebo for the Treatment of Moderate-to-Severe Cellulitis in Adults. Clinical Infectious Diseases. Volume 34, Issue 11, 1 June 2002, Pages 1440–1448, https://doi.org/10.1086/340056

This Australian study compared Ancef 2g IV + probenecid 1g PO daily to Ceftriaxone 1g IV daily + oral placebo in 132 ED patients treated for moderate-to-severe cellulitis in adults. These patients were slightly sicker than the previous studies mentioned: 41% had fever, 33% had lymphangitis, and 19% had ulceration. More than 50% of patients in each group had previously received ineffective antibiotic therapy. The duration of intravenous therapy was left to the discretion of the managing physician and outcomes were assessed at the end of therapy and at 1 month.  The duration of study therapy was similar for both groups (~6-7 days), and after they completed intravenous therapy, the vast majority of patients completed treatment with 7–10 days of oral antibiotic therapy (generally cephalexin or clindamycin). Cure rates at end of therapy were 86% vs 96% (no statistical difference). Cure rates at 1 month were also similar (82% vs 85%, no statistical difference). Nausea was more common in the Ancef-probenecid group, probably due to the probenecid. Conclusion: The once-daily regimen of cefazolin-probenecid is a cheap, practical, and effective treatment option for moderate-to-severe cellulitis, and it avoids the need to use third-generation cephalosporins in most patients.

What about just giving 1 dose of Ceftriaxone today, just to get them started, and then discharging with Keflex?

This has no evidence behind it. There is no support for the idea that a single dose of IV antibiotics results in better outcomes for any condition (even for pyelo/severe UTI! See this Cochrane study: Pohl A. Modes of administration of antibiotics for symptomatic severe urinary tract infections. Cochrane Database Syst Rev. 2007 Oct 17;2007(4)). It makes no sense to give a broad spectrum antibiotic and then step down to a narrow spectrum one, and in fact this might just increase antibiotic resistance. If we are using IV to PO stepdown strategy, we should be using a antibiotics with equivalent coverage – i.e. IV cipro to PO cipro, or IV Ancef to PO Keflex. However, even one dose of an IV antibiotic in the ED has been shown to increase the risk of antibiotic associated diarrhea and C.difficile infection from 12.3% for PO Abx Rx only to 25.7% for 1 dose IV followed by a PO Rx. (Haran JP, Hayward G, Skinner S. Factors influencing the development of antibiotic associated diarrhea in ED patients discharged home: risk of administering IV antibiotics. The American journal of emergency medicine. 2014; 32(10):1195-9.) 

Furthermore, the first line medications for cellulitis all have excellent bioavailability. Cephalexin is 90-100% bioavailable, clindamycin 90%, and doxycycline >90% when taken with food. (MacGregor 1997 – see table below for more detail). Really, theoretically, the only reason to use IV antibiotics (in the absence of sepsis or underlying immunocompromise) is if the patient isn’t able to tolerate/absorb PO.

What if I think the patient has already failed oral/outpatient therapy for cellulitis?

As Dr. Helmen points out in his post in ACEP Now,  “One reason that IV antibiotics are overused is an incorrect diagnosis of “treatment failure.” All too often, patients with skin and soft tissue infections are deemed to have failed oral antibiotics after fewer than 48 hours of oral antibiotics. They then are needlessly switched to IV antibiotics. There is no evidence to support this practice. Treatment failure of simple cellulitis should only be entertained after a 48- to 72-hour trial of oral antibiotics. Even in many of these cases, switching classes of oral antibiotics is sufficient. IV antibiotics are not the automatic answer to ‘treatment failures.’”

Good reasons to admit patients to hospital might be more for wound care, or ensuring compliance with leg elevation, rather than simply the need for IV therapy. 

OK, but still, PO antibiotics can’t work for everyone, we see so many cases bounce back. What patient factors might make me choose IV therapy?

Yadav K, Suh KN, Eagles D, et al. Predictors of oral antibiotic treatment failure for nonpurulent skin and soft tissue infections in the emergency department. Acad Emerg Med. 2019;26(1):51-59.

A 2019 retrospective chart review of 500 patients published in Academic Emergency Medicine looked at predictors of failure of PO antibiotics for nonpurulent SSTIs in adults. Independent predictors of oral antibiotic treatment failure (defined as hospitalization, change in class of oral antibiotic, or switch to IV therapy after 48 hours of oral therapy) included tachypnea at triage, the presence of chronic ulcers, history of MRSA colonization or infection, previous recent cellulitis (in the past year), chronic kidney disease, and diabetes.

What about guidelines?

I wish there were clearer guidelines about all of this. The 2014 Infectious Diseases Society of America does recommend that IV antibiotics for nonpurulent cellulitis be reserved for patients who are immunocompromised or have systemic signs of infection, hemodynamic instability, or altered mental status. “Systemic signs of infection” to me would mean fever, but remember that patients with fever were included in the studies above that showed no benefit of Ancef over Keflex.  While the studies on the subject are small, and guidelines are based on expert consensus, at this time it seems like perhaps the burden of proof should be laid on the IV therapy with current standard of care being PO antibiotics?!  Will this change your current practice? Who do you think we could safely try PO therapy for? The young healthy female with 24h lymphangitis? The non-diabetic with cellulitis and fever, but otherwise systemically well? If reassessing someone at 72h with no improvement in their forearm cellulitis, is looking for a foreign body or switching classes of antibiotic (i.e. to Septra for MRSA coverage) perhaps more appropriate than automatically giving them Ceftriaxone? I am interested in your thoughts!

Resources and further reading if you want a deep dive:

Ramakrishnan K, Salinas RC, Agudelo Higuita NI. Skin and Soft Tissue Infections. Am Fam Physician. 2015 Sep 15;92(6):474-83. 

Helman, Anton. IV vs. PO: Which Antibiotics Are Better for Common ED Infections? ACEP Now. Jan 21, 2020. https://www.acepnow.com/article/iv-vs-po-which-antibiotics-are-better-for-common-ed-infections/ Accessed July 15, 2021.

Stevens DL, Bisno AL, Chambers HF, et al. Practice guidelines for the diagnosis and management of skin and soft tissue infections: 2014 update by the Infectious Diseases Society of America. Clin Infect Dis. 2014;59(2):e10-52.