The optimum duration of intravenous antibiotic therapy for uncomplicated neonatal bacterial septicemia is not known. Pediatricians administer anywhere between 7 to 14 days of antibiotics, but these practices are not evidence based. C reactive protein (CRP) guided antibiotic duration is based on limited data and serial quantitative CRP is both cumbersome and not universally available. If it could be demonstrated that a 7-day course of antibiotics is not inferior to a 14-day course of antibiotics in terms of relapse rates of infection, then a 7 day course of antibiotics could be uniformly adopted, resulting in economic savings, shorter duration of hospitalization, less chances of hospital acquired infections, less chances of antibiotic induced adverse events and less antibiotic resistance. To test this hypothesis, a randomized, active-controlled, multi-centric, non-inferiority trial to compare the efficacy of a 7-day course of intravenous antibiotics with a 14-day course has been planned. Subjects weighing more than 1000 g at birth with suspected sepsis will be enrolled and observed for a 7-day period to see if they meet eligibility criteria for randomization. Subjects will be randomized on the 7th day of antibiotics, if the initial blood culture grows a non-Staphylococcus aureus bacterial organism, if they have no meningitis, osteomyelitis, septic arthritis or deep seated abscess and if the sepsis goes into clinical remission by the 5th day and remains in remission up to the 7th day of sensitive antibiotics. Subjects in the 14-day group will receive 7 more days of antibiotics after randomization, whereas those in the 7-day group will receive no further antibiotics after randomization. Subjects will be followed up for a 35-day period after randomization. The key outcome will be treatment failure as measured by "definite or probable relapse" within a 21-day period after completion of antibiotic therapy. Secondary outcomes will include definite relapse within 21 and 28 days after antibiotic completion and within 28 and 35 days after randomization; and probable relapse at similar time points. Other secondary outcomes will include secondary infections and adverse events. A total sample size of 700 (350 in each arm) will be required to detect a non-inferiority margin of 7%, assumed event rate of 10%, with 90% power, one-sided alpha error of 5% and loss to follow-up of approximately 10%. Data safety monitoring board will monitor serious adverse events in the trial and will perform at 1/3rd and 2/3rd of expected recruitment. At the time of interim analysis, the DSMB will revisit the sample size of the study. O'Brien Fleming's stopping criteria will be used for the primary outcome while Pocock's stopping rule will be used for the serious adverse events.
