Skip Navigation

Meta-Analysis of Trials Evaluating Parenteral Antimicrobial Therapy for Skin and Soft Tissue Infections

  1. Rebecca J. McClaine1,
  2. Thomas L. Husted1,
  3. Renee S. Hebbeler-Clark2, and
  4. Joseph S. Solomkin1
  1. 1Departments of Surgery, University of Cincinnati Medical Center, Cincinnati, Ohio
  2. 2Departments of Medicine, University of Cincinnati Medical Center, Cincinnati, Ohio
  1. Reprints or correspondence: Dr Joseph S. Solomkin, Div of Trauma and Critical Care, Dept of Surgery, University of Cincinnati College of Medicine, 231 Albert Sabin Way, ML0558, Cincinnati, OH 45267-0558 (solomkjs{at}ucmail.uc.edu).
 
Next Section

Abstract

Background. Many trials have been carried out to determine the effectiveness of antimicrobial agents in treating skin and soft tissue infections. The results of these studies are often utilized to make determinations about the use of these antimicrobials against other types of infections. Despite the importance of these trials in determining clinical care, we hypothesized that many of these studies failed to include a variety of infections of significant enough severity to effectively draw objective conclusions about antimicrobial efficacy.

Methods. We conducted a modified PubMed search to identify studies of antimicrobial agents in treating soft tissue infections that were published from 1998 through 2008. We then evaluated these trials for specific recommended study criteria, which were based on published US Food and Drug Administration guidelines for the conduct of trials of antimicrobials for soft tissue infection.

Results. Seventeen studies were identified for inclusion in the trial. Upon review, only 30% of trials required both local and systemic signs of infection for inclusion in the trial. One trial stratified results on the basis of operative intervention, less than half reported patient comorbidities, and only 53% provided a specific definition for “cure.”

Conclusions. Our meta-analysis of current trials evaluating antimicrobial therapy for skin and soft tissue infections revealed substantial shortcomings in the design of most of these trials. These data provide evidence for the importance of designing specialist panels to objectively evaluate studies and photographs of included infections to ensure that conclusions drawn from these trials concerning clinical practice are justified.

Many trials have been designed to compare the efficacy of antimicrobial agents in the treatment of skin and soft tissue infections. Complicated skin and skin-structure infections have been the initial indication for regulatory approval for all agents active against methicillin- resistant Staphylococcus aureus (MRSA). Although these study designs often incorporate blinded and randomized assignment and prespecified outcome criteria, their failure to address other clinically relevant variables that independently predict outcome may significantly weaken their validity.

In 1998, the US Food and Drug Administration (FDA) completed an industry guidance document to steer the development of antimicrobial drugs for the treatment of uncomplicated and complicated skin and skin-structure infections [1]. The recommendations from this document, summarized in Table 1, provide definitions of uncomplicated and complicated skin and soft tissue infections, suggest inclusion and exclusion criteria, address specific details concerning study conduct, and provide definitions of outcomes.

Table 1.
View larger version:
Table 1.

Summary of Guidelines from the US Food and Drug Administration for Development of Antimicrobial Drugs for Treatment of Complicated Skin and Soft Tissue Infections

In evaluating the relevance of a study's results to clinical practice, clinicians strive to determine whether the trial's conclusions or recommendations would apply to their patient population. Specifically, information concerning the severity of enrolled patients' illnesses, the anatomy of the infections of the patients enrolled, and the agreement between the trial's recommendations and current practice guidelines all contribute to this evaluative process. For surgeons, information addressing the drainage or debridement procedures performed during the management of such infections is important.

Studies that provide this information as clearly and objectively as possible facilitate clinicians' abilities to apply study conclusions to ongoing patient care. Also, a means of explicitly assuring readers that adequate source control procedures were employed is considered to be of great importance.

With this background, the aim of the current analysis was to evaluate recent clinical trials comparing antimicrobial therapies for skin and soft tissue infections to determine their applicability to current clinical practice. In addition to assessing well-known study design elements, such as randomization and blinding, we specifically examined the studies' efforts to include different types of infections, describe the severity and anatomy of infections included, report patient comorbidities, examine the operative procedures performed, and include clinically relevant pathogens (eg, MRSA).

Previous SectionNext Section

Methods

To identify published studies for inclusion in our analysis, a PubMed search was conducted on 31 December 2008, using the keywords “soft tissue infection clinical trials” or “skin structure infection clinical trials.” Results were limited to randomized controlled trials published within the past 10 years, written in or translated into English, and conducted on humans. Studies demonstrating any of the following criteria were then disqualified: subgroup analyses, overlapping patient study groups (subset reports or combined study reports), those primarily examining length of hospital stay or cost, laboratory or microbiologic analyses, and phase II trials. Additionally, studies conducted in the pediatric population, evaluating topical therapies, evaluating prophylactic antibiotic use, and those including fungal infections, non-soft tissue infections, and uncomplicated infections were also excluded. Of note, studies randomizing patients to receive an oral antibiotic as the study or control drug were eliminated, on the assumption that these included uncomplicated infections.

A second PubMed search was then conducted, using the suggested keywords of the initially identified trials and a strategy described elsewhere [2]. The keywords for this search included “skin infection(s)”, “complicated skin and skin-structure infection(s)”, “skin and skin-structure infection(s)”, “soft tissue infection(s)”, and “Gram-positive infection(s)”. The same exclusion limits were then employed.

After trial selection had been completed, each study was reviewed by 2 independent observers to evaluate specific study characteristics. Results are reported as the percentage of studies with a particular characteristic or as the mean value ( ± standard deviation) for comparing percentages across papers.

Previous SectionNext Section

Results

Trials included in this analysis. The initial PubMed search yielded 132 trials; 118 were then eliminated on the basis of our exclusion criteria, leaving 14 for inclusion in our study. The second PubMed search, using keywords suggested by the results of the first search, yielded 150 publications, many of which had also been included in the original 132 hits. Following application of the exclusion criteria, 3 trials remained, bringing the total number of included trials to 17 (Table 2) [3–19]. Separate evaluations of the papers by 2 independent observers yielded no substantially dissimilar data for any study.

Table 2.
View larger version:
Table 2.

Data on the Studies Included in Our Meta-Analysis

General study characteristics can be seen in Table 3. Overall, 94% of trials were multicenter, 64% provided patient blinding, and the drug administrator and/or evaluator were blinded in 59% of trials. The cure rate for the included trials ranged from 67% to 93% in both investigational and control arms.

Table 3.
View larger version:
Table 3.

Characteristics of the 17 Trials in Our Meta-Analysis

A meta-analysis of the included studies is shown in Figure 1. Of the 17 included studies, only 2 demonstrated significantly higher cure rates with the investigational drug, compared with the control [9, 10], whereas one study demonstrated a significantly higher cure rate with the control arm [11].

Figure 1.
View larger version:
Figure 1.

Meta-analysis of included studies. Seventeen studies were included in our analysis. Of these, only 3 demonstrated significantly higher cure rates for 1 study arm, with 2 favoring the investigational drug and 1 favoring the control agent. CI, confidence interval.

Entry criteria. Trials required various numbers of local and systemic signs for trial inclusion (Figure 2). Of the 17 trials, 13 described specific local or systemic signs to evaluate patients for inclusion. Of these, 9 grouped their suggested local and systemic signs together, and required a certain number of total signs for inclusion, allowing some patients with only local or only systemic signs to be included. As a result, only 2 of these 9 actually required at least 1 local and systemic sign for inclusion. The remaining 4 studies that specifically described signs for inclusion listed local and systemic signs separately. Overall, 29% of trials actually required both specifically described local and systemic signs for inclusion.

Figure 2.
View larger version:
Figure 2.

Local and systemic signs necessary for inclusion of studies. Of the 17 studies included in our meta-analysis, 13 listed specific signs that were necessary for patient inclusion. Of these 13, the inclusion requirements for only 5 trials were such that both local and systemic signs were required for patient inclusion.

Infections included. All 17 studies included information about the types of infections included in their trials (Table 4). Excluded infections included diabetic foot infections (24% of trials), ulcers (31% of trials), and burns (38% of trials). All studies included only “complicated” (not “simple”) abscesses, presumably based upon previously described FDA definitions of these entities, in which abscesses treated with simple incision and drainage alone, furuncles, and folliculitis are considered simple, whereas complicated infections include those involving deeper soft tissue, requiring more extensive surgical intervention, or involving specific anatomic regions [1]. Three studies specifically excluded patients whose infections had undergone incision and drainage alone.

Table 4.
View larger version:
Table 4.

Data on Types of Infections Included in the 17 Trials in Our Meta-Analysis

Numerically, cellulitis represented 33.9% ± 16.0% of included infections in the individual trials and 23.7% of overall infections. Similarly, abscesses were 29.8% ± 12.3% and 21.5% of included and overall infections, respectively. Diabetic foot infections, postoperative or surgical infections, ulcers, and burns represented varying percentages of included infections. In general, the anatomic locations of infections were not described, except in the studies exclusively including diabetic foot infections.

Organisms included. Table 5 lists information concerning the inclusion of specific organisms in trials. Seventy-one percent of trials reported percentages of infection with gram-negative rods; 88% reported percentages of infection with MRSA; and 47% and 59% of the trials reported separate cure rates for gram-negative rod and MRSA infections, respectively. Overall, gram-negative rods represented 26.6% ± 15.0% of isolates; MRSA represented 18.6% ± 17.6%.

Table 5.
View larger version:
Table 5.

Microbiological Data on Specific Organisms Reported in the 17 Trials in Our Meta-Analysis

Compliance with FDA guidelines. Multiple factors indicating compliance with published FDA guidelines were examined (Table 6). Overall, 47% of reports detailed patient comorbidities. Five studies (29%) reported on the number of patients with peripheral vascular disease. No studies reported on immunosuppressed patients, and only 1 recorded those with underlying skin disease. Fifty-nine percent of studies required hospitalization for inclusion in trial. Only 1 trial stratified antimicrobial response data by the need for surgical intervention. Fifty-nine percent of trials stated the percent of patients who underwent surgical intervention; 25%–40% of patients in these studies underwent an operation. The suggestion that >70% of patients be microbiologically evaluable was met in 35% of trials. In 53% of trials, a specific definition of cure was provided.

Table 6.
View larger version:
Table 6.

Data on Adherence to US Food and Drug Administration Guidelines of the 17 Trials in Our Meta-Analysis

Previous SectionNext Section

Discussion

A large number of clinical trials have been performed to evaluate the efficacy of new antimicrobial agents in the treatment of skin and soft tissue infections. These studies have served as the basis for approval of these agents by the FDA for the indications studied. However, noninferiority to approved agents in skin and skin structure trials has led to the use of some of these agents for other indications that may be under discussion or study at the time of the regulatory approval. Our data support the notion that currently employed trial designs are not sufficiently robust to make this approach reliable. As expected, most of the trials included in our meta-analysis were designed to limit bias by blind randomization and by maintenance of blinding of the treatment being provided patients and involved healthcare practitioners [20].

Additionally, these studies incorporated appropriate numbers of both men and women to provide information about the efficacy of the drug in the general patient population. Also, most made efforts to maintain study protocol such that a high percentage of patients were both clinically and microbiologically evaluable. All of these characteristics are mentioned in the FDA recommendations and are generally important in the design of unbiased clinical trials. It should be mentioned that all of these trials were designed to examine the noninferiority of the new agent being tested, compared against standard accepted treatment. As demonstrated in Figure 1, only 2 of the 17 included trials actually demonstrated a significant treatment effect for the investigational drug, compared with the control. One study favored the control group. In 14 of the other studies, there were no differences in cure rate between the treatment groups.

Clinicians judge the applicability of these trials to their patient population by evaluating the types of infection (ie, the level of severity and the anatomic site) included and the objectivity of the data provided. It is not apparent from the data presented in these trials that operation, simple drainage, or even limb immobilization might not be equivalent to similar maneuvers with added antimicrobial therapy for many of the patients enrolled in these trials [21]. Ten trials (59%) did provide information about whether procedures had been performed, and ∼25%–40% of patients in these trials did undergo surgical therapy. The nature of the procedures performed was not detailed, and it is possible that many such procedures were simple incisions.

The trials that provided this demographic information did not stratify their outcomes on the basis of this factor. Information concerning the procedures performed would have provided a better understanding of the anatomic consequences of the infection.

All of the trials included in our meta-analysis provided categorical information about the types of infections included in their studies. Not unexpectedly, the majority of included infections were abscesses and cellulitis, which may be legitimate if these infections are of appropriate systemic severity, are present in patients with significant comorbidities, or are in anatomically difficult areas. However, none of the studies reported specific information that would allow an assessment of the extent of the infections of the patients enrolled or the surgical procedures performed in any detail beyond bedside incision and drainage, or that would provide any assessment of the adequacy of the surgical therapy provided. Given the generally high cure rates reported in these studies, the reader is left with concerns for the need for any therapy and, in the case of failure, whether the failure was due to inadequate source control.

Although 3 trials do specifically exclude abscesses cured by incision and drainage only, we found only 5 trials requiring both local and systemic signs of infection, and 5 trials did not require any evidence of systemic response to the infection. This suggests that, in some trials, the infections of many patients, in the absence of comorbidities, could be described as Eron severity class I, which could be treated in an outpatient setting, possibly with oral therapy at most [22]. In fact, only 59% of the trials that we reviewed required patients to be hospitalized for inclusion in the trial.

Almost no information was provided about the anatomic site of infections in any of the trials. Additionally, only 47% of the trials reported patient comorbidities, and these gave information only about patients with diabetes and peripheral vascular disease, and did not include information on the number of patients with other important comorbidities, such as those who are immunosuppressed, experience chronic renal failure, or have underlying skin conditions; the trials that did report this information did include appropriate numbers of patients with diabetes and peripheral vascular disease.

The microbiologic characteristics of infections included in trials are critical for determining the applicability of the results. To that end, the FDA suggests that 70% of patients enrolled in trials evaluating new antimicrobials be microbiologically evaluable, and provides suggestions about the method by which such samples should be obtained, to avoid culturing colonizing bacteria [1]. A review of the trials included in our meta-analysis reveals that the majority of trials required a microbiologic sample for inclusion in the study, but only a subset of these described the method of obtaining this culture. Overall, in 35% of trials, >70% of patients are microbiologically evaluable by the end of the study. About half of the trials provided information concerning the organisms cultured, including the number of polymicrobial infections and those with gram-negative rods or MRSA; only about half provided some information about the effectiveness of the antimicrobial agent in clinically curing infections caused by specific bacteria, or about microbiological eradication rates per organism. Overall, this limits the ability to extrapolate the effectiveness of the agents being studied to other serious infections with known infective bacteria.

In conclusion, our meta-analysis of current trials evaluating antimicrobial therapy of skin and soft tissue infections reveals substantial and significant shortcomings of the data provided. This in turn limits the applicability of the results to clinical practice. In addition to advocating for the improvement of future trials to address these shortcomings, we recommend the enactment of other strategies, to strengthen the impact of these important trials. We would recommend that specialist panels be convened to review blinded clinical information and pre- and postoperative photography of the treated infections of patients (including those infections not requiring any intervention), to add legitimacy to concerns regarding the need for antimicrobial therapy, the extent of the infection, and the adequacy of any procedure performed.

Previous SectionNext Section

Acknowledgments

Potential conflicts of interest. J.S.S. reports that he is a consultant for Johnson and Johnson, Schering-Plough, and Bayer, that he has received expenses and honoraria for lectures from Merck, and that has received laboratory research support from Pfizer. All other authors: no conflicts.

  • Received September 24, 2009.
  • Accepted December 1, 2009.
Previous Section
 

References

  1. 1.↵
    1. US Food and Drug Administration (FDA)
    . Guidance for industry: uncomplicated and complicated skin and skin structure infections-developing antimicrobial drugs for treatment. 21 July 1998. http://www.fda.gov/downloads/Drugs/GuidanceComplianceRegulatory-Information/Guidances/ucm071185.pdf. Accessed 1 January 2009.
  2. 2.↵
    1. Chalmers I,
    2. Altman DG
    1. Dickersin K,
    2. Scherer R,
    3. Lefebvre C
    . Identifying relevant studies for systematic reviews. In: Chalmers I, Altman DG, editors. Systematic reviews. London, England: BMJ Publishing; 1995. p. 17-36.
  3. 3.↵
    1. Nichols RL,
    2. Graham DR,
    3. Barriere SL,
    4. et al
    . Synercid Skin and Skin Structure Infection Group. Treatment of hospitalized patients with complicated gram-positive skin and skin structure infections: two randomized, multicentre studies of quinupristin/dalfopristin versus cefazolin, oxacillin or vancomycin. J Antimicrob Chemother 1999;44(2):263-273.
    Abstract/FREE Full Text
  4. 4.
    1. Talan DA,
    2. Summanen PH,
    3. Finegold SM
    . Ampicillin/sulbactam and cefoxitin in the treatment of cutaneous and other soft-tissue abscesses in patients with or without histories of injection drug abuse. Clin Infect Dis 2000;31(2):464-471.
    Abstract/FREE Full Text
  5. 5.
    1. Stevens DL,
    2. Smith LG,
    3. Bruss JB,
    4. et al
    . Randomized comparison of Meta-Analysis of Antimicrobial Trials linezolid (PNU-100766) versus oxacillin-dicloxacillin for treatment of complicated skin and soft tissue infections. Antimicrob Agents Chemother 2000;44(12):3408-3413.
    Abstract/FREE Full Text
  6. 6.
    1. Siami G,
    2. Christou N,
    3. Eiseman I,
    4. Tack KJ
    . Clinafloxacin versus piperacillin-tazobactam in treatment of patients with severe skin and soft tissue infections. Antimicrob Agents Chemother 2001;45(2):525-531.
    Abstract/FREE Full Text
  7. 7.
    1. Graham DR,
    2. Lucasti C,
    3. Malafaia O,
    4. et al
    . Ertapenem once daily versus piperacillin-tazobactam 4 times per day for treatment of complicated skin and skin-structure infections in adults: results of a prospective randomized, double-blind multicenter study. Clin Infect Dis 2002;34(11):1460-1468.
    Abstract/FREE Full Text
  8. 8.
    1. Arbeit RD,
    2. Maki D,
    3. Tally FP,
    4. Campanaro E,
    5. Eisenstein BI
    . The safety and efficacy of daptomycin for the treatment of complicated skin and skin-structure infections. Clin Infect Dis 2004;38(12):1673-1681.
    Abstract/FREE Full Text
  9. 9.↵
    1. Lipsky BA,
    2. Itani K,
    3. Norden C,
    4. Linezolid Diabetic Foot Infections Study Group
    . Treating foot infections in diabetic patients: a randomized multicenter open-label trial of linezolid versus ampicillin-sulbactam/amoxicillin-clavulanate. Clin Infect Dis 2004;38(1):17-24.
    Abstract/FREE Full Text
  10. 10.↵
    1. Weigelt J,
    2. Itani K,
    3. Stevens D,
    4. Lau W,
    5. Dryden M,
    6. Knirsch C
    . Linezolid versus vancomycin in treatment of complicated skin and soft tissue infections. Antimicrob Agents Chemother 2005;49(6):2260-2266.
    Abstract/FREE Full Text
  11. 11.↵
    1. Breedt J,
    2. Teras J,
    3. Gardovskis J,
    4. et al
    . Safety and efficacy of tigecycline in treatment of skin and skin structure infections: results of a double-blind phase 3 comparison study with vancomycin-aztreonam. Antimicrob Agents Chemother 2005;49(11):4658-4666.
    Abstract/FREE Full Text
  12. 12.
    1. Jaurequi LE,
    2. Babazadeh S,
    3. Seltzer E,
    4. et al
    . Randomized, double-blind comparison of once-weekly dalbavancin versus twice-daily linezolid therapy for the treatment of complicated skin and skin structure infections. Clin Infect Dis 2005;41(10):1407-1415.
    Abstract/FREE Full Text
  13. 13.
    1. Sacchidanand S,
    2. Penn RL,
    3. Embil JM,
    4. et al
    . Efficacy and safety of tigecycline monotherapy compared with vancomycin plus aztreonam in patients with complicated skin and skin structure infections: results from a phase 3, randomized, double-blind trial. Int J Infect Dis 2005;9(5):251-261.
    CrossRefMedlineWeb of Science
  14. 14.
    1. Fabian TC,
    2. File TM,
    3. Embil JM,
    4. et al
    . Meropenem versus imipenem cilastatin for the treatment of hospitalized patients with complicated skin and skin structure infections: results of a multicenter, randomized, double-blind comparative study. Surg Infect(Larchmt) 2005;6(3):269-282.
    CrossRef
  15. 15.
    1. Giordano P,
    2. Song J,
    3. Pertel P,
    4. Herrington J,
    5. Kowalsky S
    . Sequential intravenous/oral moxifloxacin versus intravenous piperacillin-tazobactam followed by oral amoxicillin-clavulanate for the treatment of complicated skin and skin structure infection. Int J Antimicrob Agents 2005;26(5):357-365.
    CrossRefMedlineWeb of Science
  16. 16.
    1. Lipsky BA,
    2. Armstrong DG,
    3. Citron DM,
    4. Tice AD,
    5. Morgenstern DE,
    6. Abramson MA
    . Ertapenem versus piperacillin/tazobactam for diabetic foot infections (SIDESTEP): prospective, randomized, controlled, double-blinded, multicentre trial. Lancet 2005;366:1695-1703.
    CrossRefMedlineWeb of Science
  17. 17.
    1. Noel GJ,
    2. Strauss RS,
    3. Amsler K,
    4. Heep M,
    5. Pypstra R,
    6. Solomkin JS
    . Results of a double-blind, randomized trial of ceftobiprole treatment of complicated skin and skin structure infections caused by gram-positive bacteria. Antimicrob Agents Chemother 2008;52(1):1-8.
    FREE Full Text
  18. 18.
    1. Noel GJ,
    2. Bush K,
    3. Bagchi P,
    4. Ianus J,
    5. Strauss RS
    . A randomized, double-blind trial comparing ceftobiprole medocaril with vancomycin plus ceftazidime for the treatment of patients with complicated skin and skin- structure infections. Clin Infect Dis 2008;46(5):647-655.
    Abstract/FREE Full Text
  19. 19.↵
    1. Stryjewski ME,
    2. Graham DR,
    3. Wilson SE,
    4. et al
    . Telavancin versus vancomycin for the treatment of complicated skin and skin-structure infections caused by gram-positive organisms. Clin Infect Dis 2008;46:1683-1693.
    Abstract/FREE Full Text
  20. 20.↵
    1. Concato J,
    2. Shah N,
    3. Horwitz RI
    . Randomized, controlled trials, observational studies, and the hierarchy of research designs. N Engl J Med 2000;342(25):1887-1892.
    CrossRefMedlineWeb of Science
  21. 21.↵
    1. Stryjewski ME,
    2. Chambers HF
    . Skin and soft-tissue infections caused by community-acquired methicillin-resistant Staphylococcus aureus. Clin Infect Dis 2008;46(suppl 5):368-377.
    CrossRef
  22. 22.↵
    1. Stevens DL,
    2. Bisno AL,
    3. Chambers HF,
    4. et al
    . Infectious Diseases Society of America.Practice guidelines for the diagnosis and management of skin and soft-tissue infections. Clin Infect Dis 2005;41(10):1373-1406.
    FREE Full Text
| Table of Contents

This Article

  1. Clin Infect Dis. (2010) 50 (8): 1120-1126. doi: 10.1086/651264
  1. AbstractFree
  2. » Full Text (HTML)Free
  3. Full Text (PDF)Free

Classifications

Share

  1. Email this article

Navigate This Article

Current Issue

  1. March 1, 2012 54 (5)
  1. Clinical Infectious Diseases
  1. Alert me to new issues

Most