McKinlay CJ, Alsweiler JM, Anstice NS, Burakevych N, Chakraborty A, Chase JG, et al. Association of neonatal glycemia with neurodevelopmental outcomes at 4.5 Years. JAMA Pediatrics 2017; 171(10): 972-983.
Joshua Clive Anchan MD1, Nicholas Ryan Carr DO1, Kaashif Aqeeb Ahmad MD2,3
1Department of Pediatrics, Division of Newborn Medicine, Brooke Army Medical Center, Fort Sam Houston, Texas
2Pediatrix Medical Group of San Antonio, San Antonio, Texas
3Department of Pediatrics, Baylor College of Medicine, San Antonio, Texas
Joshua Anchan, MD
Brook Army Medical Center
3551 Roger Brooke Drive
Fort Sam Houston, TX 78234
KEYWORDS: neonatal hypoglycemia, neurosensory impairment, executive dysfunction, visual motor impairment
TYPE OF INVESTIGATION
Prognosis; observational analysis of a prospective cohort study
In infants (≥32 weeks gestational age) with hypoglycemia risk factors, what effect does the severity and frequency of hypoglycemia in the first 7 days of life have on neurodevelopmental outcomes at 4.5 years of age?
Design: Single center, prospective cohort study
Allocation: This study did not involve allocation. Infants were grouped based on the occurrence of hypoglycemia defined as follows: hypoglycemic episode (consecutive blood glucoses <47mg/dl), severe neonatal hypoglycemic episode (consecutive blood glucose values <36mg/dl), recurrent hypoglycemic episodes (≥3 episodes of hypoglycemia), or interstitial (measured by a subcutaneous sensor) hypoglycemic episode (glucose <47mg/dl for at least 10 minutes).
Blinding: Clinicians caring for the infants in the neonatal period were masked to neonatal interstitial glucose concentrations. Developmental outcome assessors at 4.5 years of age were masked to neonatal glycemic status.
Follow-up period: Neurodevelopmental outcomes were assessed at 2 years corrected age (previously reported) (1), and at 4.5 years of age (+/- 2mo).
Setting: Waikato Hospital, a tertiary care center in Hamilton, New Zealand, between December 2006 and November 2010.
Patients: Infants ³ 32 weeks' gestation with at least one of the following risk factors for hypoglycemia: infant of a diabetic mother, £ 37 weeks' estimated gestational age, small for gestational age (birth weight £ 10th percentile or £ 2500g), large for gestational age (³ 90th percentile or ³ 4500g), or acute illness.
Intervention: Hypoglycemia was measured by glucose oxidase method using whole blood. Screening for hypoglycemia occurred at 1-2 hours after birth, every 3-4 hours for the first 24 hours, then every 6-8 hours for the following 24 hours. Screening for study purposes continued for up to 7 days or until there was no ongoing clinical concern for hypoglycemia, whichever occurred first. Masked interstitial continuous glucose monitoring (CGM) was also performed for up to 7 days. Interventions for hypoglycemia included additional feeding, buccal dextrose gel, and intravenous dextrose with goal of blood glucose greater than 47mg/dl.
Primary outcome: Presence of neurosensory impairment at 4.5 years of age defined as any of the following: visual impairment, deafness, cerebral palsy, full-scale IQ or visual motor integration score >1 standard deviation below the mean, motion coherence threshold or executive function score >1.5 standard deviation from mean, or Movement Assessment Battery for Children-2 (MABC-2) score <15th percentile.
Pre-specified secondary outcomes: Individual components of the primary outcome along with proportion of children with neurodevelopmental disorders, history of afebrile seizures, sensorimotor impairment, or emotional-behavioral difficulty based on diagnostic criteria or validated tools specified by the study authors.
Analysis and Sample Size / Patient follow-up %: By power analysis, 470 total infants would be required to detect a 48% relative increase in neurosensory impairment due to neonatal hypoglycemia with 80% power. This assumed a baseline neurosensory impairment rate of 25% and that 50% of enrolled infants would develop hypoglycemia. The reason for using 25% as the baseline neurosensory impairment rate is not discussed. 614 infants were recruited. 604 (98%) were eligible for follow-up at 4.5 years (3 died, 7 withdrew). 127 patients (21%) were lost to follow-up, leaving 477 infants that were followed up. Primary outcomes were analyzed for 473 infants (77%); 59% had at least one episode of hypoglycemia.
|Hypoglycemia Severity||Neurosensory Impairment||||Low visual-motor score||||Low executive function score|||
|Normoglycemia||38.5%||Adjusted Risk Ratio (95% CI): 0.96 (0.77-1.21)||1.6%||Adjusted Risk Ratio (95% CI): 3.67 (1.15-11.69)||4.7%||Adjusted Risk Ratio (95% CI): 2.32 (1.17-4.59)|
- Mild Hypoglycemia
- Severe hypoglycemia
- 1-2 episodes
- ≥ 3 episodes
Table 1. Summary of significant study results – Incidence of neurosensory impairment, low visual-motor scores, and low executive function scores in children with history of neonatal normoglycemia and hypoglycemia
The authors concluded there was no association between the presence of neonatal hypoglycemia and neurosensory impairment at 4.5 years of age (37.4% in children with history of neonatal hypoglycemia versus 38.5% in children with no history of neonatal hypoglycemia, relative risk 0.96 [CI 0.77-1.21]). The risk of neurosensory impairment was also not associated with the number of hypoglycemic episodes or minimum blood glucose concentration during episodes. However, the lowest quintile for maximum blood glucose in the first 12 hours of life was associated with a higher relative risk of neurosensory impairment.
Among secondary outcomes, children with exposure to neonatal hypoglycemia did have a statistically significant increased risk of low executive function score (10.6% versus 4.7%, p=0.02) and low visual motor integration score (4.7% versus 1.6%, p=0.04), with the greatest risk for those exposed to severe hypoglycemia. Hypoglycemia detected with interstitial monitoring only was associated with a higher risk of low executive function score (12.5% versus 3.5%, RR 4, CI 1.23-12.85). Recurrent hypoglycemia was associated with higher risk of low visual motor scores.
Neonatal hypoglycemia was not associated with an increased risk of neurosensory impairment at 4.5 years of age but it was associated with an increased risk of executive dysfunction and visual motor impairment.
Neonatal hypoglycemia is potentially one of the most common preventable causes of neurodevelopmental impairment. There is debate over the degree to which hypoglycemia causes neurodevelopmental delays. There is no consensus for the glucose level at which to initiate interventions. This study was adequately powered to answer the primary question and did not find neonatal hypoglycemia to be associated with an increased risk of neurosensory impairment. However, the contemporary generalizability of the study is impacted by the methods used to define and screen for hypoglycemia. In 2011, the American Academy of Pediatrics suggested a pre-feed glucose of ≤25mg/dl from 0-4 hours of life and ≤35mg/dl from 4-24 hours of life was an appropriate intervention threshold for an asymptomatic term or late-preterm infant (2). The initial phase of this study was performed prior to the release of these guidelines and defined hypoglycemia as glucose <47mg/dl without reference to timing of the measurement in relation to feeding or if the infant was symptomatic. Categorization of infants with clinically insignificant low glucose values as hypoglycemic could reduce the incidence of neurosensory impairment in this group, although this is difficult to extrapolate given that in this study the overall rate of neurosensory impairment was similar with or without hypoglycemia
Additionally, the prevalence of neurosensory impairment in normoglycemic controls was 38.5%, higher than the expected baseline prevalence of 25% used in the power analysis. There is no clear reason for the high baseline neurosensory impairment rate even when infants who were lost to follow-up and inclusion of preterm infants were considered which further reduces the generalizability of the study.
The study did find that neonatal hypoglycemia was associated with reduced executive and visual motor function in a dose-response relationship with severe hypoglycemia and recurrent hypoglycemia. The authors suggest dysfunction in executive and visual motor function could result in behavioral and emotional difficulties at school age, potentially underlying findings by Kaiser et al of an association between neonatal hypoglycemia and lower fourth-grade test proficiency (3). Drawing such an association however would require further follow-up through school age to define the effect of neonatal hypoglycemia on long-term neurocognitive outcomes. The study also highlights the complexity of using hypoglycemia screening to prevent adverse neurodevelopmental outcomes. The authors note that in their post hoc analysis patients who developed neurosensory impairment had a steeper rise in interstitial glucose concentration after hypoglycemia. This suggests that while hypoglycemia can affect neurodevelopment, rapid resolution of hypoglycemia can also have adverse effects. This is concerning given the common utilization of intravenous dextrose boluses to correct hypoglycemia. The neurodevelopmental impairments associated with transient neonatal hypoglycemia may not be as easily preventable as previously thought.
Overall, study strengths include the relatively large sample size, adequate power for the primary outcome, and follow-up through 4.5 years with an extensive battery of neurodevelopmental testing. While there was no significant association between neonatal hypoglycemia and overall neurosensory impairment, the study did find a higher risk of impairment in executive function and visual motor skills. The clinical significance of these results is less clear and a defined management strategy to prevent the identified outcomes has not been determined. Future research should focus on further delineation of clinical significance of neonatal hypoglycemia and optimal approach to avoid adverse outcomes.
EBM LESSON: Sensitivity Analysis
Observational studies assessing for causal associations are prone to bias from potential confounding (4). Measured confounders can be accounted for through study design. For example, in this study socioeconomic status, sex, and primary risk factor for hypoglycemia were known confounders that were adjusted for in the analysis of primary and secondary outcomes. The potential impact of unmeasured confounders can be estimated through sensitivity analysis (5).
The article references a sensitivity analysis that with exclusion of children born at 32 to 34 weeks, the risk of low executive function and visual motor integration scores at 4.5 years of age was not significantly altered. Sensitivity analysis involves examining how results are affected by changes in various factors including study assumptions, methods, criteria, and compliance (6). Sensitivity analysis procedures will depend on the factor assessed (6). Subgroup analysis represents a type of sensitivity analysis performed to determine whether study results are different across groups.
Ideally, factors that will undergo sensitivity analysis should be identified a priori rather than post hoc to reduce the introduction of bias into which factors are assessed (6). In this study, post hoc subgroup analysis revealed that some infants developed neurosensory impairment between 2 years and 4.5 years while others had a stable (normal or impaired) neurosensory status. Because preterm infants did not have a neurodevelopmental evaluation at 2 years, a sensitivity analysis was performed to determine if exclusion of these children impacted the overall outcome. In this circumstance, it did not. If findings remain consistent after performing sensitivity analysis, the results are considered more "robust" and it further strengthens the study conclusions (6).
Acknowledgement: The Journal Club is a collaboration between the American Academy of Pediatrics - Section of Neonatal Perinatal Medicine and the International Society for Evidence-Based Neonatology (EBNEO.org).
McKinlay CJ, Alsweiler JM, Ansell JM, Anstice NS, Chase JG, Gamble GD, the CHYLD Study Group. Neonatal glycemia and neurodevelopmental outcomes at 2 years. N Engl J Med. 2015;373(16):1507-1518.
Committee on Fetus and Newborn., Adamkin DH. Postnatal glucose homeostasis in late-preterm and term infants. Pediatrics. 2011 Mar;127(3):575-9.
Kaiser JR, Bai S, Gibson N, et al. Association between transient newborn hypoglycemia and fourth-grade achievement test proficiency: a population-based study. JAMA Pediatr. 2015;169 (10):913-921.
Groenwold RH, Nelson DB, Nichol KL, Hoes AW, Hak E. Sensitivity analyses to estimate the potential impact of unmeasured confounding in causal research. Int J Epidemiol. 2010 Feb;39(1):107-17.
Ding P, VanderWeele TJ. Sensitivity Analysis Without Assumptions. Epidemiology. 2016 May;27(3):368-77.
Thabane L, Mbuagbaw L, Zhang S, Samaan Z, Marcucci M, Ye C, et al. A tutorial on sensitivity analyses in clinical trials: the what, why, when and how. BMC Med Res Methodol. 2013 Jul 16;13:92.