A prospective surveillance program was initiated to determine the relative role of
antibiotics containing
N-methylthiotetrazole (NMTT) versus patient risk factors in producing
antibiotic-associated
bleeding. Five hundred forty-six
critically ill patients with
serum albumin 30 g/L or below were evaluated for evidence of a
bleeding event as documented by clinical observation,
hemoglobin changes, and transfusions.
Bleeding events occurred in 16% of patients receiving an
aminoglycoside combination, 10% receiving
antibiotics with the NMTT side chain, and 14.5% receiving
antibiotics not containing NMTT (p greater than 0.05). The
bleeding rate was highest in febrile patients with
cancer (14.5%) and lowest in those with a suspected or documented abdominal
infection (10%) (p = 0.04), but within each patient group there was no difference among the
antibiotics. We conclude that the use of NMTT-containing
antibiotics is not an independent risk factor for
bleeding, but the role of severity of illness may be underappreciated.