Bug Juice Potpourri


In this month's Journal Club, we covered several articles that looked at the use of antibiotics in the Emergency Department.  Does adding Trimethoprim-Sulfamethoxazole to Cephalexin increase the rates of clinical cure in uncomplicated cellulitis? For patients receiving Vancomycin in the ED, how many are appropriately dosed and how many receive a sufficient number of doses to hopefully limit the emergence of resistant bacteria?  Are patients receiving Vancomycin and Piperacillin-Tazobactam really at increased risk of acute kidney injury?

Moran GJ, Krishnadasan A, Mower WR, et al. Effect of Cephalexin Plus Trimethoprim-Sulfamethoxazole vs Cephalexin Alone on Clinical Cure of Uncomplicated Cellulitis. JAMA 2017;317(20):2088–9.

This is a prospective, multicenter randomized control superiority trial. The trial was completed in 5 large emergency departments across the country. Patients twelve years and older presenting with skin and soft tissue infections without MRSA risk factors, including no IV drug use, no wound or foreign body, no underlying skin condition, and no invasive infection at the time of enrollment were included in the study. All of the patients were screened with bedside soft tissue ultrasound and found not to have evidence of abscess or fluid collection on bedside ultrasound at the time of enrollment. Patients were randomized into treatment with a 7 day course of Bactrim 320/1600 twice daily + Keflex 500 mg four times daily combination therapy or treatment with a 7 day course of Keflex 500 mg four times daily + placebo. Treatment regimens were provided in blister packs.

Superiority was defined as a 10% difference between the groups. The primary outcome in the per protocol group was clinical cure defined as absence of clinical failure features. These features of failure included fever, increase in erythema (> 25%), swelling, and tenderness at follow-up on days 3-4; no decrease in erythema, swelling, or tenderness by days 8-10; and more than minimal erythema, swelling, or tenderness at days 14-21. They found no statistically significant difference between the combination therapy vs monotherapy groups in the per-protocol analysis (83.5% vs 85.5%, respectively) or in the modified intention to treat analysis (76.2% vs 69.0%, respectively). There was no significant difference in adverse events rates and secondary outcomes, including surgical drainage procedures, invasive infections, hospitalizations, days of analgesic use, and days missed from normal activities and work/school. Although there were no statistically significant differences between rates of clinical failure in either the per-protocol or intention to treat analyses, the confidence interval in the intention to treat did cross the pre-defined superiority threshold of 10%. This suggests the possibility of superiority with combination therapy that did not reach clinical significance.

Discussion points

  • IDSA guidelines recommends that patients with cellulitis without systemic signs of infection, penetrating trauma, evidence of MRSA elsewhere, or IVD should received antibiotic therapy with a single agent providing coverage for streptococci only without MRSA coverage. However, many providers in the US frequently prescribed regimens that include MRSA coverage as well. This is a practice pattern held by many of the providers in our residency program and physician group, as wall.

  • A strength of this study is that patients were evaluated with bedside ultrasound to confirm absence of abscess prior to initiation of therapy. Whereas most isolated cellulitis are suspected to be caused by streptococci, MRSA is the causative agent in many abscesses. Furthermore, abscesses are unlikely to resolve without incision and drainage. Differentiation of a small, early abscess from isolated cellulitis can be challenging from history and physical exam alone. Many in our group cited the concern of misdiagnosing a small abscess as cellulitis as one of their primary reasons for providing MRSA coverage. Using ultrasound decreases the likelihood of this misdiagnosis and most providers in our group felt comfortable utilizing ultrasound for this clinical indication. However, providers who are not comfortable using bedside ultrasound or in practice environments where ultrasound is not readily available may find these results less applicable to their daily practice, a limitation mentioned in the discussion.

  • There were a large number of patients lost to follow-up in this study and poor adherence to prescribed treatment regimens. Only 58% of patients completed the full course of antibiotics. The per-protocol group, which included patients who took 75% of the prescribed treatment and had a test of cure follow-up, most accurately reflects the clinical response and best determines the effect of adding MRSA coverage. However, the per-protocol analysis may not reflect the effectiveness of course of therapy in the clinical setting. The intention to treat analysis more practically reflect the impact of therapy in the “real world”, where patients are often incompletely adherent to therapy. Because the confidence interval does cross the predefined superiority threshold, further research is necessary to determine whether there may be benefit to adding MRSA coverage in those with imperfect adherence to therapy.

  • Although the tolerability of these medications is well known and in general they are well tolerated, there was lower adherence to therapy in the Keflex + placebo group. This was not seen in a prior study using a similar masking. This may represent a blinding issue. It also suggests that imbalance in evaluable per protocol between the treatment groups was largely due to adherence.

  • Some early data reports spontaneous resolution of cellulitis rates as high as 66%, although is data from the pre-antibiotic era. This may bias toward the null hypothesis and reflect that many cases diagnosed as cellulitis are infect noninfectious conditions.

Take Away

This study did not demonstrate statistically significant rates of clinical cure with bactrim and keflex combination therapy for adult patients without MRSA risk factors with cellulitis, confirmed on bedside ultrasound. Our group’s take away from the paper was that: in patients with isolated cellulitis without MRSA risk factors, we would feel more comfortable prescribing monotherapy with an agent with coverage for streptococci in accordance with the IDSA guidelines.

Mueller K, McCammon C, Skrupky L, Fuller BM. Vancomycin Use in Patients Discharged From the Emergency Department: A Retrospective Observational Cohort Study. JEM 2015;49(1):50–7.


  • Single-center retrospective observational cohort study in an urban, academic, tertiary care center (Barnes Jewish)

Their objective

  • To characterize the indications, dosing, and appropriateness of vancomycin use in patients discharged from the ED

Main outcomes

  • To characterize patients receiving vancomycin prior to discharge home from the ED
  • To identify patients that did not meet indications for appropriate use based on the 2011 Infectious Diseases Society of America guidelines for treating MRSA infections


  • Single-center retrospective observational cohort study
  • Subjects were consecutive adult patients administered Vancomycin in the ED and then discharged home over an 18-month period 
  • Pre-defined list of variables placed in a standard format and collected by the PI via the EMR, which was cross-checked for accuracy prior to analysis 
  • Collected patient demographics, CC, Diagnosis, Dosing of Vanc, other antibiotics given, other antibiotics prescribed, MRSA risk factors, and appropriateness of vanc use 
  • Defined “correct” dose of Vancomycin as at least 15 mg/kg of actual body weight 
  • MRSA risk factors were defined off of recent ED literature as diagnosis of abscess, antibiotic use in the last 30 days, reported insect or spider bite, a personal hx of MRSA, and close contacts with a similar infection 
  • Assessed return to the ED with the same complaint or need for admission with the same complaint, and for resolution of symptoms by subsequent ED visit or PCP visit.  Assessed this for a 12-month period after initial ED visit 


  • 526 patients received Vancomycin over this 18-month period and were discharged home.  The patients in the study were generally health middle-age adults. The majority of patients (75%) had a “MRSA risk factor” - mainly an abscess (n 281 or 53%) 
  • Diagnosis given to patients included SSTI (70%), fever, UTI, laceration, AMS, headache, device complication, meningitis, hypotension, bacteremia, osteomyelitis, pneumonia
  • 19% of all patients grew MRSA (but only 289 or 55% were cultured, of which 195 were wound cultures) 
  • I&D performed on 219 patients (42%), leaving 62 patients (12%) diagnosed with abscess or abscess plus cellulitis never having I&D performed 
  • 73% of patients (n 239) were under-dosed and 68% of patients (n 357) received only one dose 
  • In addition to IV Vanc, 59% of patients (n 313) received some other antibiotic prior to d/c
  • 73% of patients (n 384) were prescribed antibiotics as an outpatient 
  • Only 9% of patients (n 49) met criteria for Vancomycin based on IDSA guidelines 
  • Follow-up data available for 63% of patients (n 332).
    • 139 (42%) came back to ED for same problem within one month. 
    • 42 (12%) required hospital stay for same problem within one month. 
    • 197 (60%) had resolution on follow-up within 12 months

Their take away

  • Vancomycin dosed and used like this is unlikely to have any clinical effect on patient care (based on dynamics/kinetics of drug and ⅓ of patients returning for the same thing)
  • Vancomycin dosed and used like this is likely to lead to resistance 
  • 91% of patients met no IDSA indication for Vancomycin use - it was mainly used for uncomplicated SSTI 


  • It is a retrospective study by design (not entirely sure WHY they got Vancomycin although the fact they were discharged supports likely not indicated)
  • Single-center study at an urban center - maybe less relevant in rural settings 
  • Mainly descriptive data 
  • No control group, so cannot say definitely that these patients did not do better with the Vancomycin 

Our Take-away

  • We agree that patients with uncomplicated SSTI and those that are likely to be discharged home should not receive IV Vancomycin and that using the drug in this way will likely cause antibiotic resistance.
  • We felt this was a great review of the pharmacokinetics and pharmacodynamics of Vancomycin.  For patients with a SSTI with a consultant such as a hand surgeon requesting the patient be given Vancomycin and placed in our ED Observation Unit (where we expect patients to be likely discharged home within 23 hours), we felt this paper is good evidence to discuss with the consultant why Vancomycin would not be the best antibiotic to use. 

Rutter WC, Burgess DR, Talbert JC, Burgess DS. Acute kidney injury in patients treated with vancomycin and piperacillin-tazobactam: A retrospective cohort analysis. J Hosp Med 2017;12(2):77–82.

This is a retrospective cohort analysis of adult patients without renal disease who were hospitalized for at least 48 hours and received vancomycin, pip-tazo, or a combination between September 1, 2010 and August 31, 2014 at the University of Kentucky Medical Center. The primary outcome was incidence of AKI that occurred greater than 48 hours after treatment initiation and less than seven days after discontinuing treatment. Overall, 11,650 patients were included in the analysis. Of those 1,647 (14.1%) developed AKI. 

AKI definition was based on the RIFLE (Risk, Injury, Failure, Loss, End-stage) criteria which uses percent decrease in estimated GFR to stratify the severity of kidney injury. AKI status was defined as meeting any of the RIFLE criteria. The researchers also stratified patients based on hypotension, total days of therapy, and any exposure to other nephrotoxic agents. To control for differences in underlying severity of illness, the researchers stratified based on Charlson Comorbidity Index. 

In total; 5,497 received combination vancomycin and pip-tazo, 3,055 received vancomycin alone, and 3,098 received pip-tazo alone. The combination therapy group had higher Charlson Comorbidity Index scores, were more likely to experience hypotension, and had increased hospital length of stay. AKI occurred in 21% of patients who received combination therapy, 8.3% of patients who received vancomycin monotherapy, and 7.8% of patients who received pip-tazo monotherapy. Patient who received combination therapy had higher rates of risk, injury, and failure as defined by the RIFLE criteria than either monotherapy group. In hospital mortality and transfer to hospice were also higher in the combination therapy group. In subgroup analysis of patients with similar Charlson Comorbidity Indices, patients in the combination therapy group were still more likely to develop AKI than patients in either monotherapy group. 

Discussion Points

  • The patients in these treatment groups were not truly comparable. Patients receiving combination therapy were sicker than patients in the monotherapy groups. The method for stratifying these patients, based on the Charlson Comorbidity Index, is not a great way to stratify these patients, as this is a measure of baseline health and not a measure of severity of acute illness. APACHE or SOFA score may have been a better choice. 
  • These treatments are not equivalent. The alternative to combination vancomycin/pip-tazo is not monotherapy, but rather a different combination therapy such as Vancomycin/Cefepime. This study does not inform our decision between two appropriate treatment options.
  • Although this is a fairly large study population, it is a single-center retrospective study and susceptible to all of the associate biases therein, such as cofounding variables not accounted for. 
  • This study does not speak to patient centered outcomes. AKI as defined by decreased GFR may or may not matter to patients. Patient centered outcomes among others, would be increased length of hospitalization, ICU days, need for CRRT or IHD, or permanent kidney injury. 
  • This study should not necessarily cause us to change our current practice. 


Moran, et al - Kelli Jerrill, MD

Mueller, et al - Shaun Harty, MD

Rutter, et al - Joshua Gauger, MD

Editing, Posting, Podcast Editing - Jeffery Hill, MD MEd