BACKGROUND

Hypothermia in neonates is a frequent complaint in the pediatric emergency department (ED). The differential diagnosis is broad and comprises potential life-threatening conditions, including infectious causes. The International Pediatric Sepsis Consensus defines hypothermia as body temperature less than 36 degrees Celsius.1 Clear evidence exists to guide the evaluation of febrile neonates,2 but evidence guiding the evaluation of hypothermic neonates is lacking. Previous work in this area has looked at infants less than 60 days old with varying temperature cutoffs.1,3–6

To date, there are limited infant hypothermia studies, and no direct studies assessing the incidence of serious bacterial infection (SBI) and herpes simplex virus (HSV) infection in neonates (<30 days) with a body temperature less than 36 degrees Celsius presenting to the ED.

The purpose of this retrospective review is to study neonates presenting with hypothermia and assess the associated risk for serious infections.

METHODS

We conducted a single-center retrospective review of patients seen from 1/1/2015 through 12/01/2020 with a presenting complaint of hypothermia in the emergency department of an urban, free standing, tertiary care children’s hospital. This activity was approved by the University of Louisville Institutional Review Board.

Inclusion criteria: Patients under 30 days of life at time of ED presentation, no prior hospitalizations, born at term (37+3 weeks and greater), initial temperatures less than 36 degrees Celsius documented, either in the ED or in a pediatric office affiliated with our healthcare system and therefore using the same electronic medical record, and patients admitted with chief complaint or ICD 9 or 10 diagnosis involving hypothermia of newborn. Diagnoses are displayed in Table 1.

Table 1
Table 1.Associated Hypothermia Diagnoses and ICD codes

Data gathered included sex, age at time of referral, culture results, laboratory results, initial vital signs, ED disposition, days as inpatient, method of referral (home vs pediatrician vs incidental finding in the ED), month presenting, birth history, medical history, temperature that initiated the evaluation, and discharge diagnoses. For subjects discharged from the ED, the EMR was reviewed for evidence of re-presentation within the healthcare system within the subsequent 30 days.

Serious infectious etiologies were defined a priori as: positive blood, urine, or cerebrospinal culture for known pathogens and excluding common contaminants, listed by the World Health Organization in 2017. Surface cultures of HSV included conjunctival, nasopharynx, mouth, and rectal swabs.

Statistical analysis: We used descriptive statistics to describe the demographic characteristics of the study population. We calculated the proportion of subjects with a serious infection.

RESULTS

A total of 144 subjects seen in the ED during the five-year study period fulfilled inclusion criteria (Figure 1). Demographic characteristics of the study population are summarized in Table 2. There were 138 subjects admitted to the hospital and six who were discharged home from the ED. None of those discharged home returned within 30 days based on electronic medical record review.

Figure 1
Figure 1.Subject Recruitment
Table 2.Demographic Characteristics of the Study Population
Sex
Male 78 (54%)
Female 55 (46%)
Age
0-7 days 122 (84.7%)
8-14 days 10 (6.9%)
15-21 days 3 (2.1%)
22-30 days 9 (6.3%)

Blood and urine cultures were obtained on 138 of subjects, all of whom were admitted to the hospital for further observation (Figure 1). CSF studies were performed on 130 subjects. Of the 138 subjects admitted, duration of hospitalization ranged from 2 to 19 days. There were no deaths in the study cohort. Of the 130 CSF cultures, none were positive for bacterial pathogens (95% confidence interval (CI) 0-2.3%). Additionally, one of the CSF pathogen panels was positive for enterovirus (95% CI 0-4.2%). Of the 138 blood cultures one was positive, a presumed contaminant (Staphylococcus epidermidis) (true positivity rate 95%, CI 0-2.2%). Of the 138 urine cultures, one was positive for (Group B streptococcus) (1% 95%, CI 0-4%) (Table 3).

Table 3
Table 3.Test Results

There were 132 subjects who had surface cultures performed. Of these, two were positive for HSV without any associated serum or CSF HSV findings (95% CI 0.2-5%). Those subjects were referred for hypothermia but were ultimately normothermic in the ED and had unremarkable hospital courses. In total, there were four serious infections (3%, 95% CI 1-7%).

DISCUSSION

Neonates are a difficult population in which to decipher life threatening illnesses. In addition to infectious causes, physiologic factors could contribute to temperature instability such as central nervous system development and body weight.7 These factors also make this population more susceptible to temperature dysregulation secondary to the ambient environment.

There have been several prior studies on neonatal hypothermia. The complexity of analysis lies in the differences in age and temperature cutoffs. Most of the studies were in those subjects 60 days and under, while two were isolated to the first month of life.4,6

On review of infant hypothermia studies, one study was a single center retrospective review which included 360 subjects with an age cutoff of 60 days, finding a SBI risk of 2.8%.3 A second study was a multicenter cross sectional cohort study which included 3565 patients. The study used a cutoff of 30 days, finding a SBI risk of 8%.4 Of note, there were no specified temperature cutoffs in either study. In a single center retrospective review with 4797 subjects, using an age cutoff of 60 days, a third study found a SBI rate of 2.6%.5 A fourth study was a single center retrospective review with 68 subjects, an age cutoff of 28 days and a SBI rate of 5.8%.6

In our study, two subjects required intervention. The subject with enterovirus meningitis presented with apneic spells and required endotracheal intubation. One of the subjects with HSV surface cultures presented with hypoxia and required nasal cannula oxygen. In summary, well appearing infants on ED presentation did not require intervention throughout their hospital encounters.

Looking at strictly the neonates, we had used 30 days for a cut off. If the age was set at 28 days for this study, we would not have discounted any serious infections. The serious infections in our study were seen in those less than 14 days of age. That age group includes non-infectious etiologies, which were documented but not analyzed in this study (Supplemental Table 1).

There were several limitations of our study. We searched for ICD 9 and 10 codes as well as chief complaints for hypothermia. There is certainly the possibility that we failed to pick up all who would have fulfilled inclusion criteria over the selection period. This was a single center study with a relatively small sample size. While our healthcare system has greater than 95% of the pediatric market share in our area, it is possible that some of the subjects who were discharged subsequently presented outside of our system.

Our findings inform the need for larger, multicenter studies. Our next steps are to enroll in a multicenter study to assess etiologies of hypothermic infants. Fortunately, larger studies are currently being designed to establish treatment guidelines for hypothermic infants like those for fever.

CONCLUSION

In neonates who present to the ED for temperatures <36 degrees Celsius, serious infections are extremely rare. Non-infectious etiologies and complications were also rare in well-appearing neonates presenting to the ED.

ACKNOWLEDGMENTS

Thank you to Dr. Samantha Lucrezia for her assistance in review of the data.