Trial design

This was an open randomized controlled trial of the impact of ambient operating room temperature on the rate of neonatal hypothermia. The study was approved by the University of Texas Southwestern Medical Center Institutional Review Board and was locally funded by the department of obstetrics and gynecology. Women undergoing cesarean delivery by the Parkland Obstetrics Service on the high-risk labor and delivery unit were included.

Cluster randomization was performed with the treatment group assigned on a rotating weekly basis according to a computer-generated random schedule. The preexisting standard for operating room temperature was 20°C (67°F); women were assigned via the schedule to either 20°C or 23°C (73°F), which was the maximum temperature allowable per hospital policy due to an increase in humidity with an increase in temperature. Operating room temperature as displayed on the room thermostat was recorded for each patient, and the accuracy of thermostat readings was audited daily using an independently calibrated infrared thermometer. In addition, a continuous temperature recording device was placed in each operating room for the duration of the study period to be reviewed regularly for any temporary fluctuations in temperature. Providers were not blinded to room temperature, however the randomization schedule was only known to the primary investigator, statistician, and building engineering staff. Both study groups otherwise received identical care according to prevailing institutional practices. This included women undergoing cesarean delivery being covered with warmed gowns and blankets and receiving warmed intravenous fluids. Upper-body forced-air warming devices were placed if general anesthesia was used. For infants >32 weeks’ gestation, resuscitation included use of radiant warming, thorough drying, and placement of prewarmed dry blankets and a knitted cap. In contrast, for infants <32 weeks’ gestation radiant warmers were used with drying, followed by placement of a plastic poncho and cap. Warmed gel mattresses were reserved for those infants <28 weeks’ gestation. The infant, still wrapped in warmed blankets, was placed on the mother’s chest as allowed by neonatal and maternal condition in those infants triaged to the newborn nursery. Transport to the nursery was in an open bassinet. Infants transported to the neonatal intensive care unit were placed in an incubator prewarmed to 38.0°C (100.4°F). A rectal temperature measurement was obtained for every infant in the operating room at approximately 5 minutes of life following initial evaluation and stabilization. Subsequently, axillary temperature was measured on arrival to the admitting nursery, which was the primary outcome.

Outcomes

The primary outcome was neonatal hypothermia on arrival to the admitting nursery, which was defined as <36.5°C (97.5°F) by WHO criteria. Assuming a neonatal hypothermia rate of 45% based on internal hospital data, 788 infants or 394 per treatment group would be required to have 90% power to detect a 25% reduction in hypothermia. Secondary neonatal outcomes included moderate to severe hypothermia (defined as a temperature <36.0°C [96.8°F]), which, assuming a rate of 15% based on internal hospital data, would give us 90% power to detect a 50% reduction. Hyperthermia was defined as temperature ≥38.0°C (100.4°F). Other secondary outcomes included intubation in the operating room, ventilator use in the first 24 hours, diagnosis of respiratory distress, surfactant administration, hypoglycemia requiring treatment, umbilical artery pH <7.0, grade-3 or -4 intraventricular hemorrhage, and culture-proven sepsis.

Maternal outcomes were also examined including hypothermia in the operating room and upon arrival to the postoperative care area. Also studied were hyperthermia, endometritis, and wound infection. Chorioamnionitis was defined clinically, based on the presence of maternal fever, uterine tenderness, and/or maternal or fetal tachycardia prior to delivery. Endometritis was diagnosed if puerperal fever persisted >24 hours postpartum with uterine infection suspected as the source. An electronic survey was distributed to operating physicians to evaluate the impact of changes in ambient operating room temperature on provider experience and performance.

Statistical methods

Data analysis included Pearson χ ² and Fisher exact test for categorical data, Student t test for continuous data, and Mantel-Haenszel test for trend analysis. For inestimable relative risks the logit estimate was used. P values <.05 were considered significant. All statistical analyses were performed using software (SAS 9.3; SAS Institute Inc, Cary, NC). This trial was registered with Clinicaltrials.gov ; the trial number is NCT02436382 .

Trial design

This was an open randomized controlled trial of the impact of ambient operating room temperature on the rate of neonatal hypothermia. The study was approved by the University of Texas Southwestern Medical Center Institutional Review Board and was locally funded by the department of obstetrics and gynecology. Women undergoing cesarean delivery by the Parkland Obstetrics Service on the high-risk labor and delivery unit were included.

Cluster randomization was performed with the treatment group assigned on a rotating weekly basis according to a computer-generated random schedule. The preexisting standard for operating room temperature was 20°C (67°F); women were assigned via the schedule to either 20°C or 23°C (73°F), which was the maximum temperature allowable per hospital policy due to an increase in humidity with an increase in temperature. Operating room temperature as displayed on the room thermostat was recorded for each patient, and the accuracy of thermostat readings was audited daily using an independently calibrated infrared thermometer. In addition, a continuous temperature recording device was placed in each operating room for the duration of the study period to be reviewed regularly for any temporary fluctuations in temperature. Providers were not blinded to room temperature, however the randomization schedule was only known to the primary investigator, statistician, and building engineering staff. Both study groups otherwise received identical care according to prevailing institutional practices. This included women undergoing cesarean delivery being covered with warmed gowns and blankets and receiving warmed intravenous fluids. Upper-body forced-air warming devices were placed if general anesthesia was used. For infants >32 weeks’ gestation, resuscitation included use of radiant warming, thorough drying, and placement of prewarmed dry blankets and a knitted cap. In contrast, for infants <32 weeks’ gestation radiant warmers were used with drying, followed by placement of a plastic poncho and cap. Warmed gel mattresses were reserved for those infants <28 weeks’ gestation. The infant, still wrapped in warmed blankets, was placed on the mother’s chest as allowed by neonatal and maternal condition in those infants triaged to the newborn nursery. Transport to the nursery was in an open bassinet. Infants transported to the neonatal intensive care unit were placed in an incubator prewarmed to 38.0°C (100.4°F). A rectal temperature measurement was obtained for every infant in the operating room at approximately 5 minutes of life following initial evaluation and stabilization. Subsequently, axillary temperature was measured on arrival to the admitting nursery, which was the primary outcome.

Outcomes

The primary outcome was neonatal hypothermia on arrival to the admitting nursery, which was defined as <36.5°C (97.5°F) by WHO criteria. Assuming a neonatal hypothermia rate of 45% based on internal hospital data, 788 infants or 394 per treatment group would be required to have 90% power to detect a 25% reduction in hypothermia. Secondary neonatal outcomes included moderate to severe hypothermia (defined as a temperature <36.0°C [96.8°F]), which, assuming a rate of 15% based on internal hospital data, would give us 90% power to detect a 50% reduction. Hyperthermia was defined as temperature ≥38.0°C (100.4°F). Other secondary outcomes included intubation in the operating room, ventilator use in the first 24 hours, diagnosis of respiratory distress, surfactant administration, hypoglycemia requiring treatment, umbilical artery pH <7.0, grade-3 or -4 intraventricular hemorrhage, and culture-proven sepsis.

Maternal outcomes were also examined including hypothermia in the operating room and upon arrival to the postoperative care area. Also studied were hyperthermia, endometritis, and wound infection. Chorioamnionitis was defined clinically, based on the presence of maternal fever, uterine tenderness, and/or maternal or fetal tachycardia prior to delivery. Endometritis was diagnosed if puerperal fever persisted >24 hours postpartum with uterine infection suspected as the source. An electronic survey was distributed to operating physicians to evaluate the impact of changes in ambient operating room temperature on provider experience and performance.

Statistical methods

Data analysis included Pearson χ ² and Fisher exact test for categorical data, Student t test for continuous data, and Mantel-Haenszel test for trend analysis. For inestimable relative risks the logit estimate was used. P values <.05 were considered significant. All statistical analyses were performed using software (SAS 9.3; SAS Institute Inc, Cary, NC). This trial was registered with Clinicaltrials.gov ; the trial number is NCT02436382 .