Thermal Effect of a Woolen Cap in
Low Birthweight Infants during Kangaroo Care

There were no dropouts after enrollment. In the intervention group, all babies wore the cap at the time of temperature measurement. Median length of stay in KMC ward was 4 days (IQR 2-9) in CAP group and 5 days (IQR 3-9) in NOCAP group (p=0.38). All participants were breastfed during the study period.
Settings
Overall, median KMC room temperature was 27.9°C (IQR 25.4-29.4 °C) . Median adherence to SSC was 48%, by site: Aber 16%, Beira 41%, Wolisso 69% (p<0.0001).
Primary outcome
At KMC admission, median infant temperature was 36.6°C (IQR 36.1-37.2 °C). A total number of 5064 temperature recordings were collected. Time spent with temperature in predefined ranges is shown in TABLE 2. Multivariable analysis estimated a rate ratio of 0.92 (95% C.I. 0.84 to 1.00; p=0.06) for the effect of the cap vs. no-cap on time spent in normal temperature range.
In addition, multivariable analysis showed that cap had no statistically significant effect on time spent with temperature below 36.5°C (rate ratio 1.06; 95% C.I. 0.92 to 1.23; p=0.40) or time spent with temperature above 37.5°C (rate ratio 0.90; 95% C.I. 0.67 to 1.21; p=0.48) . Twenty-nine infants (13 in CAP and 16 in NOCAP arms) had episodes of severe hypothermia (<35°C) and 11 (4 in CAP and 7 in NOCAP arms) had episodes of severe hyperthermia (>39°C).
Secondary outcomes
At multivariable analysis, the cap had no statistically significant effect on respiratory problems (odds ratio 0.88; 95% C.I. 0.49 to 1.58; p=0.67), sepsis (odds ratio 1.48; 95% C.I. 0.82 to 2.69; p=0.19) or in-hospital growth (β 3.00; 95% C.I. -0.70 to 6.70; p=0.11). Nineteen patients died (6.3%) before discharge (9 in CAP and 10 in NOCAP arms).
Sub-analysis
In the 72 infant-mother couples with adherence to SSC ≥ 70%, multivariable analysis estimated a rate ratio of 0.90 (95% C.I. 0.74 to 1.11; p=0.34) for the effect of the cap vs. no-cap on time spent in normal temperature

range.
In the 228 infant-mother couples with adherence to SSC <70%, multivariable analysis estimated a rate ratio of 0.94 (95% C.I. 0.85 to 1.03; p=0.18) for the effect of the cap vs. no-cap on time spent in normal.

Discussion

In this study, the use of the cap was safe but did not provide any advantages in maintaining LBWI in normal thermal range during KMC in African low-resource settings. WHO guidelines recommend the use of a cap/hat as additional thermal protection during KMC [18].
In fact, the surface area of the head represents around 21% of the total body surface area [28], thus accounting for a relevant amount of heat dispersion.[19] Previous studies showed that the use of a cap could contribute to reducing heat loss immediately after birth [19, 29], but its effect during KMC in the postnatal period remains to be established.
The reason for the present study was that during KMC, that was supposedly standard of care, the practice of using a cap had not been fully implemented in the trial hospitals.
In the present study, infants spent only half of the time with the temperature in normal range and the cap did not provide any significant advantages in thermal control.
These findings underline the limited capacity for thermoregulation during the first weeks following birth [14], even with the application of thermal protection strategies and suggest that the thermal effect of the cap is limited during stay in a KMC ward.
The complexity of the scenario (different hospitals, very warm KMC rooms, variable adherence to SCC) might have jeopardized the observation of the effect of the cap. However, the study design took into account for all potential confounders (environmental and maternal temperature, adherence to SSC during temperature recording and participating hospital) that could

Rivista Italiana on line "LA CARE" Volume 11, Numero 1-2, anno 2018