The duration of hospitalisation (length of stay) in patients hospitalised with complicated skin and skin structure infections: identifying clinical and microbiologic risk factors in a comparison of tigecycline with vancomycin/aztreonam

Abstract number: 1134_02_380

Mallick R., Solomon S.


Despite therapeutic advances in the treatment of complicated skin and skin structure infections (cSSSI), prolonged hospital stays are a major cost driver and highlight the need for expanded broad-spectrum therapies that enable faster cure and discharge.


We pooled data from 2 double-blind, randomized, multinational clinical studies in patients with cSSSI. Patients were randomly assigned to receive tigecycline (100-mg dose, then 50 mg twice daily [BID]) or vancomycin with aztreonam (1 g vancomycin + 2 g aztreonam BID) administered intravenously for 5 to 14 days. Formal test of cure was assessed at least 12 days after antibiotic termination (which was based on clinical assessments). Length of stay (LOS) reflects comorbidities and disease severity as well as therapy duration. We used a Cox proportional hazard (CPH) model adjusting for potential clinical and microbiologic risk factors, to evaluate LOS.


Among the 1116 modified intent-to-treat patients, the most common infection diagnoses at baseline were cellulitis (58.9%) and major abscess (27.9%). The most common underlying etiologies were spontaneous infection (52.3%) and trauma (27.6%). About 20.2% of patients had diabetes; 6.9% had peripheral vascular disease. Staphylococcus aureus was the most common causative pathogen (about 50% of patients) among 540 patients with confirmed microbiology and complete hospitalization data. However, about 18% of patients had a gram-negative causative pathogen, primarily Escherichia coli. About half the patients were polymicrobial. Overall, there were no differences between the treatment groups in LOS (p = 0.794). The estimated CPH model identified diabetes, trauma, a gram-negative causative pathogen, absence of cure, use of concomitant medications, infected ulcer, ICU and non-US hospital setting for treatment initiation as risk factors for significantly higher LOS. Adjusting for all of these factors, treatment with tigecycline was associated with faster discharge (hazard ratio = 1.22; p = 0.019 [~1 day shorter LOS]). This difference was almost entirely attributable to the subgroup of patients in which the causative pathogen was gram-negative.


Diabetes, gram-negative pathogen, and infected ulcers are risk factors for prolonged LOS. Tigecycline, the first glycylcycline antibiotic, was associated with shorter LOS in those patients with a primary gram-negative pathogen, consistent with its expanded broad-spectrum activity.

Session Details

Date: 01/08/2007
Time: 00:00-00:00
Session name: XXIst ISTH Congress
Location: Oxford, UK
Presentation type:
Back to top