Neonatal hypothermia is widely recognized as an important contributing factor to neonatal morbidity, especially in low and middle income countries [1, 2] and has been associated with mortality risk in newborns [3] and young infants aged 0-2 months [4]. The World Health Organization (WHO) has included thermal care (including the prevention of neonatal hypothermia) as a component of essential care in newborn among a package of basic interventions recommended universally for all babies [5]. However, in low-resource settings, adequate thermal care of newborns is difficult to achieve and hospital-based studies in South Asia [6–9] and Sub-Saharan Africa [10–13] have demonstrated a high incidence of primary hypothermia, especially in the first 24 h after birth.

Few population-based estimates are available from settings where the majority of babies are born at home, where the risk for hypothermia is presumably higher than in hospital [9, 14–16]. In southern Nepal, measurements of less than the WHO cutoff for any degree of hypothermia (36.5°C) were observed in > 90% of babies during the first 28 days of life [14]. Even less data are available on specific factors associated with an increased risk of hypothermia. The possible risks described have been ecological in nature or limited to hospital-born infants but they do indicate that preterm birth, low birth weight and seasonality (being born during the cold season) are important risk factors for neonatal hypothermia [9, 10, 12, 16–18]. However, one comprehensive review [1] of neonatal hypothermia in low-resources settings noted a need for systematic examinations of the quantitative association between hypothermia and individual, family and community-level risk factors in large population-based cohorts.

Context-specific information on the risk factors for neonatal hypothermia is necessary in order to ensure that community-based neonatal health promotion packages include thermal care interventions for the newborn. Community-based trials of chlorhexidine cleansing interventions in southern Nepal [19, 20] included repeated in-home visits by project workers to collect 213,616 axillary temperatures from more than 23,240 infants, half of whom had moderate to severe hypothermia (< 36.0°C) as defined by WHO [2, 14]. In this manuscript we quantify associations between neonatal hypothermia and the specific characteristics of infants and their carers, newborn care practices and socio-demographic factors.

Of the 23,662 live-born infants in the study area who were eligible for inclusion in the parent trial, 23,257 (98.2%) were met at one or more home visits during the neonatal period; 17 did not have any axillary measured temperatures; 23,240 contributed 213,636 axillary temperatures measures (mean 9.2 measures/infant) to the analytic dataset. There were 3134 measures of less than 35.0°C - equivalent to 1.47/100 measures (95% CI: 1.42, 1.52) from among 2342 infants, giving a cumulative risk estimate of 10.7% (95% CI: 10.3, 11.1).

Babies born outside a clinic or hospital (90.2% of all infants in the sample) were at a higher risk of hypothermia [PR = 1.62 (95% CI: 1.27, 2.06)], as were those born to mothers assisted at birth by non-skilled personnel [PR = 1.47 (95% CI: 1.19, 1.82)] and babies of mothers reporting at a > 30 min interval between the birth and delivery of the placenta. Upper caste, ethnic group and variables indicating improved socioeconomic status provided a greater protection from hypothermia.

Hypothermia was almost 50% more prevalent among babies for whom breastfeeding was delayed beyond 24 h [PR = 1.49 (95% CI: 1.37, 1.62)) and 75% more prevalent among babies for whom the cord was not cut until after the placenta was delivered [PR = 1.74 (95% CI: 1.31, 2.33)]. Risk was marginally higher among babies bathed after birth, an almost universal practice in this setting [24] There was little evidence that the timing of the bath within the first day of life was associated with hypothermia risk, although post-bath oil massage was slightly protective. Babies of mothers who reported warming the room, giving the infant a cap and providing skin-to-skin contact during the first 14 days were at similar risk relative to babies of mothers not reporting these practices.

Hypothermia risk associated with incremental weight differences

Given the strength of the association with infant weight, the risk associated with incremental (50 g) changes in weight was examined (Figure 1). After modelling the risk associated with weight as a continuous measure in a spline Poisson regression model, there was only weak evidence that weight of between 3500-3000 g was associated with hypothermia risk (P = 0.09). However, each 100 g decrement in weight between 3000 g and 2500 g was associated with a 7.4% (95% CI: 4.4%, 10.5%) increase in risk of hypothermia and the association became substantially stronger and of greater magnitude between 2500 g and 2000 g [13.5% (9.7%, 17.5%)] and < 2000 g [31.3% (26.3%, 36.5%)].

In a low resource setting in southern Nepal, where 90% or more of births occur at home, after adjusting for the age at measurement and ambient temperature (the season), the weight of the infant was the most important independent factor related to neonatal hypothermia. Almost 30% of infants had low birth weight (< 2500 g) and those between 2000 g -2499 g and 1500 g -1999 g were at a 1.49 and 4.32 times greater risk of hypothermia, respectively, after adjusting for a multitude of factors. VLBW babies were 11.63 times more likely than normal weight infants to have axillary temperatures < 35.0°C, although the proportion of these babies was very low (146 of 22,748, 0.6%). When examined as a continuous variable, 100 g decrements in weight were associated with substantial increases in risk of hypothermia along the entire spectrum of underweight babies and continued up to 3000 g. Such associations have been documented previously among hospital-born newborns [12, 25, 26].

All newborn babies are prone to heat loss, particularly in the first minutes and hours after birth [1, 2]. Preterm and underweight babies lose heat faster than normal weight or full term babies, due to larger surface to mass ratios [1] and rapid transepidermal loss of water through the compromised skin barrier [27]. Preterm birth and low weight were independently found to be related to hypothermia in these data but the latter appeared to a more important factor; in both cases these infants are more susceptible to hypothermia due to their reduced brown adipose tissue reserves compared to full terms infants with appropriate weight for gestational age. Given that [1] much underweight at birth is due to intra-uterine growth restriction in this setting, [2] the 'normalcy' of small babies in this community relative to early birth (especially very early preterm infants < 34 weeks) and [3] the potential differential capacity to recognize early birth versus small size at birth [28], the intention and timing (and, thus, potential causality) of thermal care practices such oil massage, warming the room and provision of a hat may differ for these babies. Our data on these specific practices are limited by uncertainty as to whether the practice resulted as a therapeutic (reactive) response to the caretaker's perception of her baby being cold or as a preventative measure. Differential intent (more therapeutic versus prophylactic) of these and potentially other non-measured practices between preterm and low birth weight babies might contribute to the observed greater importance of weight as a risk factor.

These weight measures (precise to +/- 2 g) are more objective than gestational age, which was estimated from the date of the last menstrual period reported by mothers at enrollment at approximately 6 months. Thus, there is more scope for misclassification of gestational age, which could result in estimates of the association with hypothermia that are biased downwards [29]. While weight varies over the newborn period, 89.6% of weight measures were recorded within 72 h of birth and hypothermia risk factor models were adjusted for age at every measurement. Limiting the dataset to only infants measured within 72 hours of birth did not substantially change any risk factor estimates or conclusions.

Adj PRs for the lowest birth weight category were five times higher in the warm season than in the cold. This indicates that, while the overall risk of hypothermia is lower during this time relative to the cold months, [14] low birth weight babies are disproportionately at risk and do not receive the protective benefit of the warmer months. Cold season practices, such as warming the room or providing baby with a hat, did not appear to be protective in this setting possibly due to reverse causality (that is, that these are more likely practiced as reactive measure to the caretaker's perception that the baby is cold). However, although caretakers are generally more aware of the importance of keeping the baby warm during the cold months, the absence of such attention during warm periods may differentially impact low birth weight infants, resulting in the substantially increased magnitude of comparative risk estimates among these babies during the warm periods.

The strong association between hypothermia and the sex of the newborn is not easily explained. There may be unmeasured practices that differ according to the sex of the newborn which may change the risk profile [30]. Alternatively, if observed care practices (such as skin-to-skin contact and hat wearing which did not differ by sex) are conducted with differing intent between boys and girls (as being preventive for boys but therapeutic for girls), these sex-related associations could arise even after adjustment.

Delayed breastfeeding was associated with hypothermia after adjustment: early breastfeeding may directly reduce risk through this additional direct contact. In addition, early breastfeeding has been associated with reduced neonatal mortality [21, 31] potentially through a reduced risk of infection and hypothermia.

While kangaroo-mother-care or skin-to-skin contact is recommended as part of a package of targeted care for low birth weight or preterm infants in facilities [32] debate continues as to the benefit of extending this to all infants in community settings [9, 33]. Maternal report of skin-to-skin contact during the first 14 days after birth was not related to the risk of hypothermia. However, this practice is not part of usual care in this community (only 4.4% of infants received this care) and may be a reactive practice undertaken when mothers perceive their baby to be cold. Those receiving this care were more likely to have initiated breastfeeding after 24 h, providing some support for the reactive nature of the practice.

If adequate thermal protection is not begun immediately after birth, the newborn will lose heat rapidly, with the skin temperature losing up to 4°C in the first minutes [34]. Thus, immediate actions, such as wrapping and drying the baby without waiting for the placenta to be delivered, are recommended as part of essential newborn care [2, 5]. In our setting, newborns are often left unattended until after the placenta is delivered. Both delayed placental delivery and the fact that the cutting of the cord occurs after placental delivery were associated with an increased risk of hypothermia. However, as delayed cutting of the cord may improve iron status and reduce the risk of anaemia, [35] early cutting should not be encouraged. Rather, further emphasis should be placed on encouraging attendants to rapidly dry and wrap the newborn, to have skin-to-skin contact with the mother and to avoid leaving the baby unattended during the period between birth and delivery of the placenta.

These data provide an in-depth examination of the risk factors associated with neonatal hypothermia as measured by axillary temperatures collected at home from 23,240 infants. The data highlight the importance of thermal care interventions, especially those targeted towards low birth weight infants. Focused behaviour change messages and skill building on the basic components of thermal care, including early breastfeeding initiation, skin-to-skin contact for all infants and delayed bathing have been promoted in Shivgarh, India and have been resulted in a 54% reduction in neonatal mortality [36]. The relevance and specific importance of these interventions for underweight, preterm and other high risk infants need to be emphasized. Further evaluations of the delivery mechanisms of behavioural change communications which stress the importance of thermal care and skin-to-skin contact within the community is necessary. Estimates of the risk of adverse outcomes such as mortality and morbidity of early exposure to neonatal hypothermia in the community are urgently needed.