This was a prospective observational study at the tertiary perinatal center of the University Medical Center Hamburg-Eppendorf over a 12-month period (May 2019–2020). Inclusion criteria were preterm neonates with a birth weight < 1,800 g and a gestational age < 33 weeks, who were admitted to the Neonatal Intensive Care Unit (NICU) and treated in an incubator (Caleo®, Draeger, Luebeck, Germany) for at least 14 days. Patients with chromosomal aberrations, congenital malformations, small for gestational age (birth weight < 3rd percentile), and with surgical interventions were excluded from the study. Neonates were eligible for enrollment if parental consent was obtained.

Parenteral nutrition with glucose was started immediately after birth. Electrolytes, amino acids, and lipids were added within the first three days of life. Enteral nutrition with increasing amounts of breast milk or preterm formula was also initiated within the first three days of life. The incubator settings were selected by the nursing staff in accordance with the ward's internal practices (Table 1) to keep the preterm infants’ rectal temperature in the normothermic range (36.5–37.5°C).

Relevant clinical patient data, daily energy intake, and the incubator settings chosen by the caregivers were extracted from the digital patient data management system (ICM®, Draeger, Luebeck, Germany) on days 1, 2, 3, 4, 5, 6, 7, 10, 14, 18, 22, and 28 of life. The thermal equilibrium was assessed using the HeatBalance® program (Draeger Medical Germany, Luebeck, Germany) (13, 14), originally developed to simulate appropriate incubator settings based on physical algorithms. As the calculation was carried out retrospectively and without feedback to the nursing staff, any influence on the caregivers’ actual settings could be ruled out.

The HeatBalance® program accounted for the following parameters: the baby's heat production (Q_neonate, W), the (negative or positive) heat fluxes due to convection (Q_conv, W), radiation (Q_rad, W), evaporation (Q_evap, W), respiration (Q_resp, W), and phototherapy (Q_pt, W), as well as the resulting water loss (ml/kg/h). Heat production was estimated based on the baby's body weight, corrected gestational age, and activity level based on data from the literature (8, 13, 14). Total heat loss (Q_total, W) was defined as the sum of all heat fluxes (Q_total = Q_conv + Q_rad + Q_evap + Q_resp + Q_pt). The heat balance was the difference between heat production and total heat loss (Q_hb = Q_neonate—Q_total). Required independent variables for the program's algorithms included weight (g), gestational age (weeks), postnatal age (days), room temperature (°C), incubator air temperature (°C), incubator relative humidity (%), ventilation (yes/no), phototherapy (yes/no), and clothing (yes/no) (Figure 1). For reasons of comparability with clinical habits, values expressed in watts by the HeatBalance® program were later transformed into calories per unit of time (1 W = 860 cal/h).

The study protocol was approved by the ethics committee of the local Medical Chamber of Hamburg (2020-10105-BO-ff).

Statistical analyses were performed using SPPS Ver. 27 (IBM, NY, USA) and SigmaPlot Ver. 14 (Systat Software, Inc., CA, USA). Data on neonatal demographics were expressed as median and range for continuous variables. Mean values and standard deviations (SD) of caloric intake, absolute and specific weight gain, heat production, and heat balance were plotted against the newborn's postnatal age.

The eligible study population consisted of 32 preterm neonates with a median (range) gestational age and birth weight of 29.1 (25.6–32.9) weeks and 1,213 (720–1,795) g, respectively. Nine consecutive measurements (days 1–14) were carried out in twelve patients, and twelve data points (days 1–28) were recorded in 20 patients. Incubator settings were made according to our institutional regimen, with humidities between 50% and 75% for one week, depending on gestational age, followed by a 5% reduction per week. Temperature settings started between 34 and 37°C depending on birth weight and were regulated by servo (skin control) mode (Table 1).

Body weight initially dropped to a mean (SD) minimum of −7.5 (±6.2)% on day four, followed by a rise to 38.3 (±7.7)% above birth weight on day 28 (Figure 2a).

Consistent with our institutional nutritional regimen, daily caloric intake per weight increased steadily up to a maximum mean of 137 (±9) kcal/kg/d on day 28 of life. The mean heat production estimate was 37.5 (±8.2) kcal/kg/d directly after birth and increased to 61.5 (±9.2) kcal/kg/d on day 28. The mean difference between caloric intake and heat production (energy excess) increased from −7.8 (±9.8) kcal/kg/d on the first day of life to 73.0 (±16.4) kcal/kg/d on day 28. The increase in body weight started once the caloric intake exceeded the heat production rate by roughly 50 kcal/kg/d (Figure 2b).

The growth rate varied to a maximum of 25.7 (±16.6) g/kg/d. The heat balance started at −25.3 (±17.8) kcal/kg/d after birth and reached a temporary steady state by day 10. However, within the following 8 days, it dropped again to a minimum of −9.3 (±11.2) kcal/kg/d, stabilizing at −4.2 (±10.6) kcal/kg/d from day 22 onwards. The variations in growth rate appeared to run parallel to the fluctuations in heat balance (Figures 2c,d).