Pharmacokinetic/Pharmacodynamic (PK/PD) Model for the Safety of Tigecycline (T) in Patients with Complicated Intra-Abdominal Infections (cIAI)

Authors: Meagher AK, Liolios K, Passarell JA, Cirincione B, Van Wart S, Fiedler-Kelly J, Babinchak T, Ellis-Grosse EJ
Conference: ECCMID
Keywords: anti-infectives, antibiotic, antiemetic, cIAI, cSSSI, female, hosmer-lemeshow, infectious disease, Kaplan-Meier, logistic regression, nausea, region, ROC, SAS, vomiting, weight
Division: Clinical Pharmacology and Pharmacometrics (CPP)

Abstract

Purpose: Nausea (N) and vomiting (V) have been reported with tigecycline, a new glycylcycline with expanded broad spectrum activity. 140 exposure-response relationships and patient covariates predictive of the first N and V occurrence were evaluated in patients with complicated intra-abdominal infections (cIAI).

Methods: Data from three cIAI studies (one Phase 2 and two Phase 3), receiving 100mg loading dose and 50mg every 12 hours, were pooled for analysis. N and V (classified as definitely, possibly, or probably related) reported from the start of infusion until 24 hours after the last dose were included. Individual exposure measures [AUCSS(0-12) and Cmax] were calculated using a previously developed population PK model. Logistic regression was used to evaluate predictors of first N and V occurrence. Covariates included age, weight, sex, region of treatment, and baseline N and V.

Results: The dataset included 928 patients (218 with PK). Mean (SD) age and weight were 46 (18) years and 73 (16) kg. 64% of patients were men and 24%, 37%, and 18% were enrolled in North America, Europe, and Latin America, respectively. Baseline nausea or vomiting was reported in 47% and 35%. Overall, N and V occurred in 18% and 13% of patients receiving tigecycline, however most (62%; 67%) of first N and V events were mild in nature. Women had more N and V (23%; 17%) than men (15%; 11%). N and V were lower in Europe (10%; 6%) than in other regions. AUCSS(0-12) and Cmax were not predictive. The final nausea model included weight, sex, region, baseline nausea, and the interaction of weight/region as predictors of the first nausea occurrence (p=0.671, 0.0006, 0.205, 0.033, & 0.023, respectively). The final vomiting model included weight, sex, region, & 4 interactions (weight/sex, weight/region, sex/region, & weight/sex/region) as predictors of the first vomiting occurrence (p=0.054, 0.819, 0.083, 0.815, 0.02, 0.005, & 0.01, respectively).

Conclusions: AUCSS(0-12) and Cmax were not predictors of nausea and vomiting events for tigecycline. The final nausea model would
predict: nausea to be less in men, Europeans, and in the absence of baseline nausea. The final vomiting model would predict: heavier men, from all regions except Latin America, and heavier women have less vomiting.

European Congress of Clinical Microbiology and Infectious Disease (ECCMID), Nice, France, April 2006

By AK Meagher, K Liolios, Julie A Passarell, BB Cirincione, SA Van Wart, J Fiedler-Kelly, T Babinchak, and EJ Ellis-Grosse