Patients and Methods

This was a randomized, double-blind study of the efficacy and tolerability of ceftobiprole monotherapy versus combination therapy with vancomycin plus ceftazidime for patients with complicated skin and skin-structure infections (http://www.clinicaltrials.gov; no. NCT00210899). This international study was conducted at 129 sites from September 2005 through October 2006. Written consent for participation in the study was obtained from all patients, and the study was conducted in accordance with the Declaration of Helsinki.

Study population. Eligible patients were ⩾18 years of age and had received a diagnosis of a complicated skin or skin-structure infection. Complicated skin or skin-structure infection was defined as infection involving subcutaneous tissues or requiring significant surgical intervention that required intravenous therapy, along with ⩾1 of the following characteristics: (1) onset of infection within 30 days after surgery or trauma (including partial thickness burns over <10% of the body surface), with purulent drainage or ⩾3 symptoms, including rectal temperature >38°C, swelling, erythema of ⩾10 mm, pain, or tenderness; (2) onset of abscess (without open wound) in the 7 days before enrollment, with purulent drainage or aspirate and evidence of loculated fluid requiring intervention within 48 h after study enrollment and with erythema and/or induration of ⩾20 mm in diameter or with tenderness; (3) onset of cellulitis in the 7 days before enrollment, with advancing edema, erythema, or induration and 1 symptom, including fever or reported fever in the 3 days before enrollment, WBC count ⩾10×10⁹ cells/L or ⩾10% bands, or associated lymphangitis and adenopathy; and (4) in diabetic patients, a foot infection consisting of inframalleolar full-skin-thickness ulcer, cellulitis, myositis, or tendonitis with ⩾3 symptoms, including swelling, erythema, tenderness, or increased skin temperature.

Inclusion and exclusion criteria were similar to those published elsewhere [16]. Patients with foreign body infection, osteomyelitis, critical limb ischemia, or septic arthritis were excluded from this trial.

Study design. Patients were stratified by infection type, and the percentage of patients with cellulitis was limited to 20%. Diagnosis was classified according to a hierarchy that placed diabetic foot infection first, followed by abscess or wound infection. Subjects were randomized via a central interactive voice response system in a 2:1 ratio (ceftobiprole group to comparator group).

Patients in the ceftobiprole group received 500 mg for 120 min every 8 h. To ensure that the study was double blind, these patients also received placebo in a manner that matched the manner in which the vancomycin regimen was administered in the comparator arm. In the vancomycin plus ceftazidime group, the starting dose was 1000 mg of vancomycin infused over 60 min every 12 h plus 1000 mg of ceftazidime infused over 120 min every 8 h. Dosages of vancomycin were adjusted according to local prescribing information. The double-blind trial design was maintained by assignment of an independent unblinded pharmacist at each site, who prepared placebo and study drug infusions and used colored sleeves to blind the appearance of infusions to patients and blinded study staff. An unblinded site monitor who operated independently of the blinded site monitor and all blinded activities further monitored this activity. Study drug was given for 7–14 days. At the investigating physician's discretion, empirical treatment with metronidazole to provide activity against anaerobic bacteria was permitted for 48 h, pending the results of culture.

Clinical and microbiological assessments. Clinical and microbiological outcomes were assessed at a test-of-cure visit (TOC) that was performed 7–14 days after the end of therapy and were categorized as shown in table 1. The microbiological assessments included pathogen identification confirmed by coagulase production and API testing (API; bioMérieux) and by susceptibility testing, according to Clinical and Laboratory Standards Institute methodology [18, 19]. This testing was conducted at a single, central laboratory. MRSA was identified by oxacillin MIC (⩾4 µg/mL for S. aureus) or by mecA gene testing using PCR. The presence of Panton-Valentine leukocidin or mecA was assessed by multiplex analysis, using a method similar to that described by McClure et al. [20]. Cultures for anaerobic organisms were performed if anaerobic pathogens were sus-pected.

Safety and tolerability. Safety and tolerability were evaluated by assessment of adverse events, laboratory values, and vital signs and by physical examinations. Other prespecified assessments included analysis of rates of nausea, vomiting, dysgeusia, infusion site reactions, renal-related adverse events, and hypersensitivity reactions.

Analysis populations. The intent-to-treat (ITT) population included all randomized patients. The clinically evaluable (CE) population included patients who had protocol-defined complicated skin and skin-structure infections and completed the TOC visit 6–17 days after completing ⩾80% of the study drug course. Patients who received study drug for <48 h, who received a systemic antibiotic to treat an infection other than the primary complicated skin and skin-structure infection, who were considered to be cured after receiving study drug for <5 days, whose pathogens isolated at baseline were resistant to both ceftobiprole and the comparator, or who were missing their TOC assessment were not clinically evaluable. The microbiologically evaluable population consisted of patients from the CE population who had a pathogen isolated at baseline. The safety analysis population included all randomized patients who received at least 1 dose of study medication.

Statistics. The primary objective was to determine whether ceftobiprole was noninferior to vancomycin plus ceftazidime in improving the clinical cure rate at TOC. The primary analysis of clinical cure rate was assessed in both the CE and ITT populations. A total of 816 randomized patients, assuming a nonevaluable rate of ∼30%, was needed to allow the null hypothesis of inferiority of ceftobiprole of 10% to be rejected with a power of 80% using a 2-sided α of 5%. The secondary efficacy end point, microbiological eradication rate at the TOC visit, was also analyzed, and the noninferiority of ceftobiprole treatment, compared with vancomycin plus ceftazidime treatment, was concluded if the lower limit of the 95% CI was −10% or more. The 95% CIs for the differences of 2 proportions were calculated using the exact confidence intervals and were constructed on the basis of the inverted 2-sided test [21].

Results

Patient disposition. Of 828 patients enrolled in the study, 547 (66%) received ceftobiprole (table 2). Ninety-seven of the 129 sites enrolled at least 1 patient. Sites in North America, Europe, and South America, Asia, and Africa enrolled 276 patients (33%), 441 patients (53%), and 111 patients (13%), respectively. The proportion of patients completing the trial was 92% in the ceftobiprole arm and 90% in the comparator arm. The most common reason for discontinuation was loss to follow-up, which occurred for 3% of patients in both treatment groups.

Demographic and baseline characteristics. Treatment groups in the ITT population were balanced with respect to demographic and baseline characteristics (table 3). Approximately 25% of patients (138 [25%] of 547 patients in the ceftobiprole group and 73 [26%] of 281 patients in the comparator group) were aged ⩾65 years. Infections involved either the fascial plane or muscle in 36% of patients, and surgical debridement (including incision and drainage) was performed for 39% of patients as part of their initial therapy. The mean duration of treatment in the CE population was similar in ceftobiprole- and comparator-treated patients (9.0 days and 9.1 days, respectively); 411 (85%) of 485 ceftobiprole-treated patients and 199 (82%) of 244 comparator-treated patients received therapy for 5–11 days. Thirty-nine patients (22 [4.0%] of 547 ceftobiprole-treated patients and 17 [6.0%] of 281 comparator-treated patients) in the ITT population received metronidazole.

Bacteria identified as causes of infection were similar in both treatment groups. The list of pathogens identified in ⩾10 subjects in either treatment arm is shown in table 4. Among the ITT population, 53% of patients had infection due only to gram-positive pathogens, 16% had infection due only to gram-negative pathogens, and 11% had mixed infections. Among the ME population, 64% of pathogens isolated were S. aureus. Approximately one-third of all S. aureus isolates (123 [32.8%] of 375 isolates) were methicillin resistant. The next most frequently isolated pathogens were Escherichia coli (10.7% of pathogens) and P. aeruginosa (6.6%). The most common causes of infection in the microbiologically evaluable ITT population (patients with diabetes who had foot infections and patients with other complicated skin and skin-structure infections) are shown in table 5.

Clinical outcomes. At the TOC visit, among the CE population, clinical cure occurred for 439 (90.5%) of 485 ceftobiprole-treated patients and for 220 (90.2%) of 244 comparator-treated patients (95% CI, −4.2% to 4.9%). Cure rates in the ITT population at the TOC visit were also similar for ceftobiprole-treated (448 [81.9%] of 547 patients) and comparator-treated patients (227 [80.8%] of 281 patient; 95% CI, −4.5% to 6.7%). Differences in cure rates in the CE population by region were not statistically significant (cure rates in North America, 132 [84%] of 157 in the ceftobiprole group and 66 [93%] of 71 patients in the comparator group; cure rates in South America, Africa, and Asia, 53 [88%] of 60 patients in the ceftobiprole group and 26 [79%] of 33 patients in the comparator group; cure rates in Europe, 254 [95%] of 268 patients in the ceftobiprole group and 128 [91%] of 140 patients in the comparator group).

The cure rates exceeded 90% among ceftobiprole-treated patients with abscess, cellulitis, or wound infection and among patients who had surgical debridement of their infection performed within 48 h after study enrollment. These rates were comparable to those observed among comparator-treated patients (table 6). The majority of patients in this trial did not have additional surgical debridement of their infection performed after study enrollment, and there were no apparent differences in cure rate when patients who required this adjunctive therapy were compared with those who did not. Among subjects with infections involving fascia or muscle, 43 (22.3%) of 193 ceftobiprole-treated patients and 20 (19.4%) of 103 comparator-treated patients underwent a surgical procedure ⩾48 h after study enrollment; 20 (46.5%) of these 43 ceftobiprole-treated patients and 11 (55%) of these 20 comparator-treated patients were considered not to have been cured by study drug because of this need for surgical therapy.

Cure rates among subjects suspected of having the most severe infections are shown in table 6. These subjects were identified using the following criteria: (1) baseline C-reactive protein level >50 mg/L [22], (2) involvement of fascia or muscle at baseline, (3) systemic inflammatory response syndrome [23], and (4) infection due to Panton-Valentine leukocidin-positive MRSA [24]. The cure rate of complicated skin and skin-structure infections associated with bacteremia was 84.6% (11 of 13 patients) among ceftobiprole-treated patients and 62.5% (5 of 8 patients) among comparator-treated patients. The difference in these rates was not statistically significant (difference, 22.1%; 95% CI, 16.7%–61.0%).

Clinical cure rates among the ME population were similar for both gram-positive and gram-negative bacterial infections in the ceftobiprole-treated and comparator-treated patients (table 7). For patients with S. aureus, MRSA, and methicillin-susceptible S. aureus infections, cure rates at the TOC visit among ceftobiprole-treated patients and among comparator-treated patients were comparable (table 8). Among patients with infections involving gram-negative bacteria, when the experience involved at least 10 patients, the cure rates were similar between arms and were at least 75%. Among the 30 ceftobiprole-treated patients from whom P. aeruginosa was isolated at baseline, 26 (86.7%) were clinically cured. Among the 12 ceftobiprole-treated patients from whom P. aeruginosa was isolated as the sole pathogen at baseline, 9 (75%) were assessed as clinically cured at the TOC visit. Six of these 9 cured patients had isolates with MICs ⩽4 µg/mL, and all 3 of the patients who were not cured had isolates with MICs ⩾8 µg/mL.

Microbiological outcomes. In the ME population, eradication rates at the TOC visit were similar among the ceftobiprole arm (88% [344 of 391 patients]) and among the comparator arm (89% [177 of 199 patients]; 95% CI, −6.4% to 4.5%). Similar eradication rates were observed among ceftobiprole-treated and comparator-treated patients infected with either gram-positive or gram-negative bacteria (table 7). For each infection type, microbiological eradication rates were similar among ceftobiprole-treated and comparator-treated patients (microbiological eradication rate, 79.5% vs. 78.0% for diabetic foot infection; 91.6% vs. 93.2% for wound; 91.8% vs. 95.8% for abscess; and 90.3% vs. 88.0% for cellulitis).

Safety and tolerability. Approximately one-half of the patients in each arm experienced ⩾1 adverse event (table 9), of which the majority were mild and not considered to be treatment related. Five of the following specific adverse events occurred in >5% of ceftobiprole-treated patients: nausea, infusion site reaction, diarrhea, headache, and vomiting (table 9). The frequency of adverse events among ceftobiprole-treated patients was not different from that among comparator-treated patients, except for hypersensitivity-related adverse events, which occurred more frequently among the subjects treated with vancomycin plus ceftazidime (95% CI, −9.3% to −1.2%).

Adverse events leading to discontinuation of study treatment were experienced by 26 patients (5%) in the ceftobiprole arm and by 16 patients (6%) in the comparator arm. No individual adverse event leading to treatment discontinuation occurred in >1% of patients in either arm. One patient (0.2%) in the ceftobiprole group discontinued study treatment because of vomiting, and 2 patients (0.7%) in the comparator group discontinued treatment because of nausea. Severe adverse events were reported by 39 ceftobiprole-treated patients (7%) and 24 comparator-treated patients (9%). All specific severe adverse events occurred in ⩽1% of patients. Four deaths occurred during the study period, 3 of which occurred in ceftobiprole-treated patients. Causes of death for the ceftobiprole-treated patients were cardiac arrest, acute cardiovascular and pulmonary insufficiency 12 days after study treatment was discontinued, sepsis associated with a vascular catheter-associated infection, and multiorgan failure 25 days after study treatment was discontinued. One patient in the comparator arm died of cardiorespiratory insufficiency. None of these deaths were considered to be related to study treatment.

Discussion

In this trial, similar cure rates were observed among ceftobiprole-treated patients and among vancomycin plus ceftazidime-treated patients with a range of complicated skin and skin-structure infections caused by either gram-positive or gram-negative bacteria. This experience included the study of patients with diabetic foot infections, abscess, cellulitis, or wound infection and patients with infections caused by a broad range of pathogens, including MRSA and P. aeruginosa.

Because complicated skin and skin-structure infections can represent a spectrum of disease that may differ in response to therapy, subgroup analyses were conducted to assess the clinical outcomes in each of the treatment arms. As expected, patients with diabetes who had foot infections had the lowest cure rates of the categories analyzed. In this subgroup, the cure rate observed among ceftobiprole-treated patients (86.2%) was comparable to that observed among patients treated with vancomycin plus ceftazidime (81.8%). Additional analyses of outcomes in patients with severe infection, as suggested by marked elevation in C-reactive protein level, involvement of fascia or muscle, sepsis, or infection due to Panton-Valentine leukocidin-positive MRSA, further underscore that ceftobiprole monotherapy was as effective as the combination comparator regimen. These results support the conclusion that ceftobiprole can be used to treat the most severe complicated skin and skin-structure infections, with outcomes that are likely to be similar to those achieved with a combination regimen (specifically vancomycin and ceftazidime) that provides activity against the broad spectrum of pathogens.

Gram-positive bacteria and, specifically, S. aureus are major causes of complicated skin and skin-structure infections, and this trial further demonstrates the predominant role that these pathogens play. With the exception of S. aureus, which was isolated at the site of infection in nearly two-thirds of patients enrolled in the study, no single bacterial species was identified as causing infection in >11% of patients. Nevertheless, the trial experience did provide additional information regarding the efficacy of ceftobiprole, compared with ceftazidime, as treatment for infections due to gram-negative bacteria, specifically E. coli and P. aeruginosa. Consistent with the similarity of in vitro activity of ceftobiprole and ceftazidime against Enterobacteriaceae and P. aeruginosa, ceftobiprole treatment was associated with clinical and microbiological cures comparable to those achieved with ceftazidime treatment, suggesting that the spectrum of clinically relevant antibacterial activity of ceftobiprole extends to gram-negative bacteria.

The emerged importance of MRSA as a cause of complicated skin and skin-structure infections has been further confirmed by this trial. Approximately one-third of S. aureus isolates were methicillin resistant. In both treatment arms, cure rates among patients with MRSA infection were slightly lower that those among patients with methicillin-susceptible S. aureus infection. The experience with Panton-Valentine leukocidin-positive MRSA infection in this trial was substantial: 57 (46%) of 123 clinically evaluable patients with MRSA infection had Panton-Valentine leukocidin-positive isolates. In patients with Panton-Valentine leukocidin-positive MRSA infection, the cure rate among ceftobiprole-treated patients was 92.1%, compared with a cure rate of 84.2% among comparator-treated patients.

Consistent with experience with other β-lactams (specifically cephalosporins) and with findings from the ongoing ceftobiprole development program, ceftobiprole was well tolerated in this study. The incidence of serious adverse events was similar in both treatment groups. In addition, the incidence of common adverse events was comparable between treatment groups, with the exception of hypersensitivity reactions, which occurred more frequently in patients treated with vancomycin plus ceftazidime. Although nausea and vomiting were reported to occur more frequently in ceftobiprole-treated subjects than in comparator-treated subjects in a previous trial using a 60-min infusion [16], the frequency of these symptoms in this trial, which used a 120-min infusion, was similar in both groups.

In an era in which MRSA has assumed an important role in causing complicated skin and skin-structure infections in both hospital- and community-acquired disease, providing initial therapy for the most severely ill patients with an agent that has a spectrum of antibacterial activity that includes MRSA is emerging as the standard of care. The results of this trial support ceftobiprole as an effective anti-MRSA agent that also has activity against important gram-negative bacteria. Therefore, ceftobiprole has promise of meeting the therapeutic challenge of being a well-tolerated, single agent that has sufficient antibacterial activity to treat serious complicated skin and skin-structure infections, including diabetic foot infections.