Materials and Methods
Setting and Study Population
The NICU at Parkland Memorial Hospital (PMH, Dallas, TX) is a Level-IIIC, 90-bed, predominantly inborn unit that admits approximately 1400 infants annually. Staffing is provided by 22 neonatologists from the University of Texas Southwestern Medical Center, along with neonatology fellows, neonatal nurse practitioners and pediatric residents. Per protocol, all infants <35 weeks gestation or <2100 g birth weight are admitted to the NICU. Infants born to mothers with prepregnancy diabetes of any class are admitted for 4 hours of observation and glucose monitoring. Infants with surgical conditions are cared for in the NICU before and after surgery; infants with critical heart disease requiring surgical intervention are transferred to Children's Medical Center Dallas.
Empiric therapy for early-onset sepsis (<72 hours of age) is ampicillin and gentamicin; oxacillin and gentamicin are initiated for suspected late-onset sepsis (>72 hours of age). Since 1997, there has been a vancomycin reduction protocol in the PMH NICU whereby oxacillin is the preferred empiric agent and vancomycin is initiated only if the infant is colonized with methicillin-resistant Staphylococcus aureus or the cultures of normal sterile body sites yield bacteria susceptible only to vancomycin.[16,17] All infants receive a minimum of 2 blood cultures when evaluated for sepsis, and coagulase-negative Staphylococcus is considered a pathogen only when recovered in ≥2 cultures drawn within 2 days of each other. Third-generation cephalosporin agents are reserved for the treatment of Gram-negative meningitis, while piperacillin–tazobactam is the preferred second agent for complicated Gram-negative infections that do not involve the central nervous system. Decisions regarding specific antimicrobial agents and duration for pneumonia, necrotizing enterocolitis and other clinical scenarios are at the discretion of the attending neonatologist, with pediatric infectious disease consultation available.
The 14-month surveillance ran from October 3, 2011 to November 30, 2012 and consisted of 2 periods, 1 prospective and 1 retrospective. We performed prospective surveillance of all antibiotics provided to every infant who was admitted to the PMH NICU from March 1, 2012 to November 30, 2012 (9 months). Each infant's electronic medical record (EMR) was reviewed; evaluation and analysis of all antibiotic use was continued until discharge. Specifically, the following information was obtained for each antibiotic course: the agent(s) prescribed, including dose and frequency, as well as the reason(s) for its initiation, discontinuation and duration. Pertinent clinical, laboratory, microbiologic and outcome data also were collected. For infants ≤32 weeks gestation, the Clinical Risk Index in Babies-II score was calculated as a marker for severity of illness. Antibiotics initiated after the infant had not received any for >48 hours were considered a new course of therapy. Antibiotics used for surgical prophylaxis always were considered to be a new course. The neonatology faculty and fellows were aware of the study. To control for the observer (Hawthorne) effect, we also performed a retrospective analysis of all antibiotic use from October 3, 2011 to February 29, 2012 (5 months). The start date was selected because on October 2, 2011, the "K" Nursery—a Level-II area staffed by the newborn nursery providers—was closed, and infants who would have been admitted to this nursery subsequently were cared for in the NICU.
At PMH, a combined maternal and neonatal chemoprophylaxis protocol is used for prevention of early-onset group B streptococcal infection. Antibiotic prophylaxis is provided to all women at delivery who have identifiable risk factors, while newborns without clinical signs of infection receive a single intramuscular dose of aqueous penicillin G within 1 hour of birth.[21–24]
Antibiotic use was calculated using 2 methods, length of therapy (LOT) and days of therapy (DOT). LOT refers to the number of calendar days that an infant received antibiotics, regardless of the number of antibiotics provided.[26,27] DOT is an aggregate sum of the days of exposure accounting for each antibiotic, and is determined by multiplying the number of antibiotic doses by the dosing interval, then dividing by 24 hours. For example, 6 doses of ampicillin given every 8 hours and 2 doses of gentamicin given every 24 hours equal 2 LOT (2 DOT) and 4 DOT [(6 ampicillin doses × 8 hours/dose ÷ 24 hours) + (2 gentamicin doses × 24 hours/dose ÷ 24 hours) = 4]. Four doses of ampicillin given every 12 hours equal 2 LOT (2 DOT) and 2 DOT (4 ampicillin doses × 12 hours/dose ÷ 24 hours = 2). Intramuscular penicillin for GBS prophylaxis counted for 0.5 DOT (1 dose × 12 hours/dose ÷ 24 hours = 0.5) and was assigned a LOT of 1 day. Antibiotics with dosing intervals of >24 hours due to renal dysfunction or extreme prematurity had the intervening days included [ie, 2 doses of gentamicin given every 48 hours would equal 4 LOT (4 DOT) and 4 DOT (2 gentamicin doses × 48 hours/dose ÷ 24 hours = 4)].
Descriptive analyses, including mean and standard deviation or median and interquartile range as appropriate, were performed to describe the overall and condition-specific prevalence of antibiotic prescribing. Student t, Wilcoxon rank-sum and χ2 tests were used, where appropriate, to compare findings from the prospective and retrospective periods. All tests were 2-tailed, and a P value <0.05 was considered significant. The study was approved by the Institutional Review Board of the University of Texas Southwestern Medical Center.
Pediatr Infect Dis J. 2015;34(3):267-272. © 2015 Lippincott Williams & Wilkins