Of the 546 patients randomly assigned, 3 did not receive any drug, 274 received tigecycline (50.5%), and 269 (49.5%) received the combination of V/A (Fig. ) and were included in the mITT population. Because 23 patients did not meet the criteria for the severity of infection, 520 patients (tigecycline group, n = 261; combination group, n = 259) were included in the c-mITT population. Eighty-four patients of the c-mITT population did not meet the evaluability criteria, and 436 patients (tigecycline group, n = 223; V/A combination group, n = 213) were included in the CE population. No baseline and/or susceptible isolates were found in 124 CE patients; therefore, 312 patients (tigecycline group, n = 164; V/A combination group, n = 148) were included in the ME population.
Both treatment groups had similar demographic characteristics, clinical diagnoses, causes of infection, and comorbid conditions (Table ). The most common diagnosis in both treatment groups was deep soft tissue infection involving cellulitis. Approximately half of the original infections were spontaneous, about 30% were due to traumatic injury, and 12% resulted from surgeries. In the mITT population, fewer patients in the tigecycline group (n = 13; 4.7%) than in the V/A combination group (n = 27; 10%) received concomitant antibiotics, administered after the start of treatment with the study drug through the last dose of study drug (Fisher's exact test, P = 0.021).
Demographic and baseline characteristics of mITT population
Approximately 80% of the mITT population in each treatment group completed therapy and were included in the CE population. In both groups, the reasons for exclusion from the CE population were broken treatment blind (12 of 274 patients [4.4%] and 17 of 269 patients [6.3%] in the tigecycline and the V/A groups, respectively), failure to meet the inclusion criteria and/or the minimal disease criteria (13 of 274 patients [4.7%] and 11 of 269 patients [4.1%] in the tigecycline and the V/A groups, respectively), no clinical evaluation at test of cure (10 of 274 patients [3.6%] and 8 of 269 patients [3.0%] in the tigecycline and the V/A groups, respectively), the presence of Pseudomonas at the baseline (1 of 274 patients [0.4%] and 3 of 269 patients [1.1%] in the tigecycline and the V/A groups, respectively), and the receipt of more than two prior doses of antibiotic after the sample for the baseline culture was obtained (1 of 274 patients [0.4%] and 2 of 269 patients [0.7%] in the tigecycline and the V/A groups, respectively).
The trial met the predefined statistical criteria for demonstration that the efficacy of the tigecycline montherapy was not inferior to that of combination V/A therapy (Table ). For the CE population, the success rates were 89.7% and 94.4% (difference, −4.7 [95% CI differences, −10.2 to 0.8]) for the tigecycline and the V/A treatment groups, respectively. In the c-mITT population, the cure rates were 84.3% and 86.9% (difference, −2.6 [95% CI differences, −9.0, 3.8]) for the tigecycline and the V/A treatment groups, respectively. In the secondary analyses, the results for the ME population were generally consistent with those for the CE and the c-mITT populations. At the test-of-cure visit, tigecycline was noninferior to V/A and thus may be as effective as the combination of V/A in the ME and microbiologic modified ITT (m-mITT) patients with both monomicrobial and polymicrobial infections (Table ).
Clinical success rates by study population at test-of-cure visit
The clinical cure rates were also equivalent in the two treatment groups when they were compared across the baseline diagnoses of infection (Table ). Although there is little power to determine differences between the dose groups, tigecycline monotherapy was comparable to V/A for the subsets of patients with baseline diagnoses of diabetes, peripheral vascular disease, or bacteremia. A small number of CE patients presented with bacteremia at the baseline (15 patients in the tigecycline arm and 10 patients in the V/A arm). Although the calculations are based on asymptotic properties, which may not be verified in the case of a small sample size, no differences in cure rates were seen between the two treatment groups in the presence or absence of bacteremia. Similarly, no differences in the cure rates were seen between the two treatment groups in the few patients in the CE population with peripheral vascular disease (10 and 12 patients in the tigecycline and the V/A treatment groups, respectively). In the ME population, the microbial eradication rates (documented and presumed) were 84.8% (95% CI, 78.3, 89.9) and 93.2% (95% CI, 87.9, 96.7) for tigecycline and the combination of V/A, respectively, at the test-of-cure visit (difference, −8.5 [95% CI differences, −16.0, −1.0]) (Table ).
Clinical success rates by exploratory subgroupsa at test-of-cure visit
Microbiologic response at the subject level (ME population) at test-of-cure visit
The microbiologic eradication rates (documented and presumed) at the test-of-cure visit for seven selected isolates of clinical interest, as they are the causes of cSSSIs, were comparable in the tigecycline and the V/A groups (Table ). Among patients infected with MRSA and MSSA, the clinical and the microbiologic success rates were 83% and 87%, respectively, for tigecycline-treated patients and 50% and 95%, respectively, for the V/A-treated patients. MIC90 values for tigecycline were uniformly low for the most prevalent isolates, including the MRSA and MSSA isolates, compared with those for V/A (Table ).
Microbiologic response at isolate level: selected baseline isolates at test-of-cure visit (ME population)
MIC range, MIC50s, and MIC90s of selected primary baseline isolates (ME population)
Isolates with a fourfold or greater increase in the MIC from the baseline were evaluated for patients treated with tigecycline. Two patients in the study were identified to be infected with isolates deemed to be indeterminate or resistant to tigecycline, as defined by the provisional breakpoints, based on previous preclinical investigations.
Ribotyping analysis showed that the first patient acquired a second strain of K. pneumoniae (resistant) after the baseline culture grew a susceptible strain. In the second patient, K. pneumoniae grew only in a culture of a sample taken after the baseline culture, and ribotyping analysis was not performed. These results showed that neither of the strains developed documented resistance to tigecycline during therapy.
Data from all patients in the mITT population were used for the analysis of safety. Tigecycline had a safety profile consistent with that seen in phase 1 (23
) and phase 2 (25
) trials. Table shows the frequency and distribution of TEAEs that occurred in at least 3% of the patients. The majority of AEs were considered unrelated to a study medication and were mild to moderate in intensity. Nausea and vomiting were the most commonly reported TEAEs in the tigecycline group, occurring in 25% and 12% of patients, respectively, whereas they occurred in 5% and 2% of patients, respectively, in the V/A combination group. Although nausea and vomiting were usually related to tigecycline, approximately 96% of the incidents were mild to moderate in severity (grade 1 or 2). Significantly more tigecycline patients than patients in the V/A combination group (P
< 0.001) received antiemetic therapy. The most common antiemetic treatment was metoclopramide.
Treatment emergent adverse events that occurred in ≥3% of patients
Serious adverse events were reported for 28 patients during the study: 15 patients in the tigecycline group (5.5%) and 13 patients in the V/A combination group (4.8%). The most frequent SAEs in the tigecycline-treated patients were cellulitis and infection (three patients each; 2.2% combined). The most frequent serious adverse event among V/A combination-treated patients was cellulitis (two patients; 0.7%).
Adverse events were the most frequent primary reason for discontinuation of the study drug (tigecycline, n = 6 [2.2%]; V/A combination, n = 13 [4.8%]). Three tigecycline patients discontinued treatment because of nausea. Seven V/A combination patients but no tigecycline patients (P < 0.007) discontinued treatment because of skin disorders: dermatitis allergic, furunculosis, pruritic rash, pruritis, and rash.
There was one death during the study. A 74-year-old man died from adenocarcinoma, but in the investigators' opinion, the death was not related to the study treatment. The patient received 7 days of treatment with tigecycline and died 1 day after the completion of therapy. Adenocarcinoma was present at the start of the study but was not discovered until surgery was performed.
No hematologic or serum chemistry abnormalities were associated with the use of tigecycline. In the V/A combination group, AST and ALT liver enzyme level increases were the most commonly reported TEAEs, occurring in 5% and 7% of patients, respectively, whereas they occurred in 2% and 2% of patients in the tigecycline group, respectively. Elevations in AST and ALT levels have been reported in <1% of aztreonam-treated patients (20
) but are not described in the vancomycin prescription information (21
). Significantly more patients in the V/A combination group (13.9%) had low lymphocyte levels compared with the number in the tigecycline group (4.4%) (P
< 0.001). A higher percentage of patients in the tigecycline group (15.4%) had a high platelet count compared with the percentage of patients in the V/A combination group (8.9%) (P
= 0.024). All but one patient (in the V/A combination group) with low platelet counts entered the study with preexisting thrombocytopenia.