Risk of bleeding and hypoprothrombinaemia associated with NMTT side chain antibiotics: Using cefoperazone as a test case

A retrospective cohort study was performed to determine the incidence of hy poprothrombinaemia and bleeding in patients receiving cefoperazone, a third -generation cephalosporin that contains an NMTT side chain. 374 patients re ceiving cefoperazone from February 1983 to March 1986 at a teaching hospita l in Philadelphia were compared with 497 patients receiving either ceftizox ime or cefotaxime during the same period, and with 476 patients receiving c eftazidime from April 1985 to December 1987. Adverse events (any bleeding e pisodes, decrease in haemoglobin, prolongation of prothrombin time (PT), an d prolongation of partial thromboplastin times (PTT)) were evaluated, if oc curring during the period from the start of cephalosporin therapy, or the s tart of therapy with one of the two control drugs, for 14 days after the la st date of the first course of therapy were recorded. An increased risk of hypoprothrombinaemia was associated with the use of ce foperazone: the prothrombin time was prolonged by 5 s or more in 12.3% of p atients receiving cefoperazone vs. 5.8% of patients receiving ceftizoxime o r cefotaxime, and vs. 5.8% receiving ceftazidime; the adjusted odds ratios (95% CIs) were 3.6 (1.7-7.4) and 3.8 (1.8-7.8), respectively, and these inc reased at higher doses of cephalosporin. No protection was apparent from th e administration of vitamin K prior to or during the course of cephalospori n, No overall increased risks were observed for bleeding (adjusted odds rat ios (95% CIs) were 1.1 (0.8-1.4) vs. ceftizoxime or cefotaxime, and 0.9 (0. 6-1.2) vs. ceftazidime), decrease in haemoglobin, or increased partial thro mboplastin time. In subgroup analyses, increased risks of bleeding were obs erved with high dose cefoperazone use [2.8 (1.5-5.5) vs. ceftizoxime or cef otaxime, and 2.3 (1.1-4.6) vs. ceftazidime]. Patients receiving NMTT side chain antibiotics should be monitored for hypo prothrombinaemia, but any increase in bleeding is likely to be small, and p rophylactic vitamin K is probably not warranted. Copyright (C) 1999 John Wi ley & Sons, Ltd.