Lower versus Traditional Treatment Threshold for Neonatal Hypoglycemia: No Difference Between the groups at 18 months of Age, Awaiting Long-term Outcomes

April 16, 2020


van Kempen, Anne A.M.W. et al. “Lower Versus Traditional Treatment Threshold for Neonatal Hypoglycemia.” The New England Journal of Medicine 382.6 (2020): 534–544. Web. PMID 32023373


Rajesh Pandey
Assistant Professor
McGovern Medical School, University of Texas Health Sciences Center at Houston, Houston, TX, United States

Lisa M. Scheid
Assistant Professor
McGovern Medical School, University of Texas Health Science Center at Houston, Houston, TX, United States




Do asymptomatic infants ≥35 weeks, with birth weight ≥ 2000 grams who are late preterm (35-37 weeks’ GA) or SGA, LGA and infant of diabetic mothers’ with moderate hypoglycemia (P), who are treated for plasma glucose level < 36 mg/dl (I) have lower BSID III scores at 18 months (O) compared to infants who are treated for plasma glucose level < 47 mg/dl (C)?


  • Design: Randomized, controlled, non-inferiority trial.
    • Randomization was centrally controlled and web-based, using a dedicated, secure website, randomization was stratified by study center, 1:1 ratio between the low and high threshold groups
    • The randomization sequence was generated by using GCP and U.S. Food and Drug Administration compliant ALEA® randomization software (www.formsvision.com; ALEA Clinical/FormsVision, Abcoude, the Netherlands). Randomization was centrally controlled and web-based, using a dedicated, password-protected, SSL-encrypted website.
  • Allocation: The authors did not report or provide us with a description of an allocation concealment approach that allowed for classification as concealed or not concealed.
  • Blinding: Clinical team was not blinded, BSID examiners were blinded
  • Follow-up period: 18 months
  • Setting: 17 academic hospitals in the Netherlands
  • Patients: Also include inclusion/exclusion criteria
    • Inclusion Criteria:
      • Preterm infants born at 35 0/7–36 6/7 weeks’ gestation
      • Full-term small-for-gestational-age infants, birth weight below the 10th percentile
      • Full-term large-for-gestational-age infants, birthweight above the 90th percentile
      • Infants of mothers with diabetes treated with insulin, and newborns with a birthweight > 90 percentile if their mothers with diabetes were treated with diet only
    • Exclusion Criteria:
      • Gestational age ≤34 6/7 weeks or birthweight
      • Serious comorbidity, such as severe perinatal asphyxia; conditions requiring the support of vital functions; major congenital malformations or syndrome
      • The requirement of intravenous glucose before randomization for clinical reasons or initial severe hypoglycemia
      • (Suspicion of) hyperinsulinemia or inborn error of metabolism
      • No parental informed consent
  • Intervention:
    • The researchers aimed was to maintain the infants’ glucose concentration at 36 mg per deciliter or greater in the lower-threshold group and at 47 mg per deciliter or greater in the traditional threshold group. If an infant’s glucose concentration fell below the pre-specified threshold, oral feeding, tube feeding, or intravenous glucose administration was started , and carbohydrate intake was increased by 1 to 2 mg per kilogram of body weight per minute. The glucose concentration was checked 1 hour after each increase, and, if necessary, the infant’s carbohydrate intake was further adjusted. Treatment interventions were similar in the two groups. Criteria to stop the screening were at least two consecutive glucose concentrations ≥47 mg/dl without the need for tube feeding or intravenous glucose administration.
  • Outcomes:
    • Primary outcome: Bayley Scales of Infant and Toddler Development, third edition (BSID 3) at 18 months on lower threshold group will be within 7.5 points of the BSID on the traditional group.
    • Secondary outcomes:
      • Related to burden (the number of glucose measurements; treatment of hypoglycemia with supplemental oral feeding, tube feeding, or intravenous glucose; and duration of breastfeeding)
      • Related to the efficacy of the intervention (glucose concentrations and episodes of hypoglycemia after randomization)
      • Use of health care resources (hospital stay and health care costs within 18 months after randomization).
  • Analysis and Sample Size:  The study was designed as a non-inferiority trial to determine whether Bayley’s scores were less than 7.5 points (½ SD) lower after a management strategy using a lower treatment threshold compared with a traditional threshold. Sample sizes were calculated for the four predefined subgroups (preterm, term SGA, term LGA, term IDM) separately. Sample size calculation for noninferiority trials was applied using the absolute difference between the Bayley-III-NL outcomes (normative mean 100 [SD 15]). The non-inferiority margin was set at minus ½ SD (–7.5 points) for the lower threshold group, reflecting a one-month delay in development, which was considered clinically relevant. With a power of 90% and a one-sided alpha of 0.025, 85 newborns were required in each management group. Anticipating 15% withdrawal, we aimed to include 200 newborns (100 in each treatment arm) in each of the four subgroups. Per protocol analysis was done.
  • Patient follow-up: Overall, 82.5 % of infants enrolled in the low threshold, and 86.5 % enrolled in the traditional threshold were included in the analysis.


Total of 348 infants in lower- Threshold group and 341 infants in traditional threshold group were enrolled were comparable for GA (38.6± 0.1 vs 38.7 ± 0.1), birth weight (3316 ± 47 vs.3354 ± 46), males ( 58 % vs. 54 %) and socioeconomic status ( -0.01 ± 0.07 vs. 0.11 ± 0.08).

The main result of the trial, in the table below, did not show the difference between lower and traditional threshold – both among the total trial population as well as subgroups of preterm, SGA, LGA, and IDM.

Table 1: Primary outcomes

Total trial population Lower- Threshold

( N = 348)

Traditional- Threshold             (N = 341) Mean difference
Bayley-III-NL cognitive score 102.9±0.7 102.2±0.7 0.7 (−1.5 to 2.9)
Bayley-III-NL motor score 104.6±0.7 104.9±0.7  −0.3 (−2.4 to 1.8)
Preterm infants
Bayley-III-NL cognitive score 104.3±1.2 103.9±1.2 0.3 (−3.5 to 4.2)
Bayley-III-NL motor score 105.0±1.3 106.5±1.3 −1.5 (−5.7 to 2.6)
Infants small for gestational age
Bayley-III-NL cognitive score 97.5±1.3 98.5±1.3 −1.0 (−5.1 to 3.2)
Bayley-III-NL motor score 102.4±1.3 102.5±1.4 −0.1 (−4.4 to 4.3)
Infants large for gestational age
Bayley-III-NL cognitive score 105.4±1.3 103.7±1.1 1.7 (−2.1 to 5.5)
Bayley-III-NL motor score 106.0±1.2 106.0±0.9 −0.1 (−3.4 to 3.2)
Infants of mothers with diabetes
Bayley-III-NL cognitive score 106.5±1.8 103.1±2.3 3.4 (−3.2 to 10.0)
Bayley-III-NL motor score 105.7±1.5 104.2±2.0 1.4 (−4.1 to 7.0)

Secondary outcomes:  Proportions of infants with hypoglycemia and frequency of hypoglycemia was more common in lower threshold group.

Table 2: Secondary outcomes

Lower Threshold

( N = 348)

Traditional-Threshold (N = 341) Mean Difference

( 95 % CI )

Infants with hypoglycemic episodes 57 % 47 % 9.7 (2.2 to 17.0)
No. of hypoglycemic episodes
2 or 3 21 % 16 % 5.1 (–0.7 to 10.9)
4 or more 9 %  2 % 6.9 (3.5 to 10.5)
Severity of Hypoglycemia
Moderate 47 % 42 % 5.5 (-1.9 to 12.8)
Severe or both moderate and severe 10 % 5 % 4.2 (0.3 to 8.2)

Moderate hypoglycemia was defined was plasma glucose level 36-46 mg per deciliter and severe hypoglycemia was defined as plasma glucose level 35 mg per deciliter or less

Cost of neonatal care {–411 (–888 to 77)}, health care cost during follow up period {–40 (–511 to 436)}, and total healthcare cost {–346 (95% CI –1034 to 386)}; was lower threshold group that did not reach statistical significance. There was no difference in adverse events.


In otherwise healthy newborns with asymptomatic moderate hypoglycemia, a lower glucose treatment threshold (36 mg per deciliter) was non-inferior to a traditional threshold (47 mg per deciliter) for psychomotor development at 18 months.


Neonatal hypoglycemia has long been a medical conundrum. The mere definition and operational threshold with which to offer treatment has long been and continues to be debated. Severe and symptomatic hypoglycemia negatively impacts long-term neurodevelopment. How to balance treatment without risking adverse consequences and without overutilization of health care resources are discussed often. As a result, this research study sets out to challenge the dogma of the required treatment at glucose <47 mg/dl.

One aspect that distinguishes this study from others was its unique clinical design. While most other studies are retrospective, this study was a prospective, randomized, non-inferiority study assessing the current operational threshold for hypoglycemia treatment with a lower threshold. The study allowed for neurodevelopmental outcomes <0.5 standard deviations below the mean, as a threshold for non-inferiority. Given one standard deviation from the mean is considered normal, this was a reasonable threshold. Additionally, the study was able to achieve its goal of 85% neurodevelopmental follow-up at age 18 months, which is impressive for a condition that is usually short-term and often viewed as benign.

There were a few critical aspects of this study. First, it was underpowered to assess infants of diabetic mothers. Second, aside from the treatment threshold, there was no defined protocol for how to treat hypoglycemia, and it was left to provider discretion despite wide practice variations. Lastly and most importantly, follow-up at 18 months of age is likely too soon to detect differences in neurodevelopmental outcomes; therefore, most follow-up studies focus on 24 months of age or higher. The CHYLD studies by McKinlay, et al. found no difference in neurodevelopmental outcomes between neonate with hypoglycemia (defined as glucose <47 mg/dL) and those without at 24 months of age, but did find differences in executive and visual-motor functioning at 4.5 years of age. (1, 2) Several additional studies with longer-term follow-up have shown differences between cohorts with and without neonatal hypoglycemia. For instance, Wickström, et al. (2018), showed evidence of motor and cognitive delay at 2-6 years of age-associated with moderate hypoglycemia (defined as glucose < 40 mg/dL). (3) Kaiser used 4th-grade standardized testing to assess 10-year follow-up and found an association with decreased literacy and mathematics scores with a single incidence of glucose <45 mg/dL, and even further disparity at lower thresholds of <40 mg/dL and <35 mg/dL. (4) A systematic review by Shah in 2019, concluded neonatal hypoglycemia is associated visual-motor impairment and executive dysfunction in early-childhood (2-5 years of age) and neurodevelopmental impairment and low literacy in mid-childhood (6-11 years of age). (5) Given this evidence, it is reasonable to conclude while this study showed no difference at 18-month follow-up that longer-term follow up might show differences. As this study completed in 2011, it would be interesting to see if a 4 to 6-year follow-up is assessed, and if so, what the results show.

In conclusion, while it may be too soon to instill practice changes based on the results of this study, there is potential if longer-term follow-up can show similar results.


  1. McKinlay CJD, Alsweiler JM, Ansell JM, Anstice NS, Chase JG, Gamble GD, et al. Neonatal Glycemia and Neurodevelopmental Outcomes at 2 Years. N Engl J Med. 2015;373(16):1507-1518. doi:10.1056/NEJMoa1504909
  2. McKinlay CJD, Alsweiler JM, Anstice NS, Burakevych N, Chakraborty A, Chase JG, et al. Association of Neonatal Glycemia With Neurodevelopmental Outcomes at 4.5 Years. JAMA Pediatr. 2017;171(10):972-983. doi:10.1001/jamapediatrics.2017.1579
  3. Wickström R, Skiöld B, Petersson G, Stephansson O, Altman M. Moderate neonatal hypoglycemia and adverse neurological development at 2–6 years of age. Eur J Epidemiol. 2018;33(10):1011-1020. doi:10.1007/s10654-018-0425-5
  4. Kaiser JR, Bai S, Gibson N, Holland G, Lin TM, Swearingen CJ, et al. Association Between Transient Newborn Hypoglycemia and Fourth-Grade Achievement Test Proficiency: A Population-Based Study. JAMA Pediatr. 2015;169(10):913-921. doi:10.1001/jamapediatrics.2015.1631
  5. Shah R, Harding J, Brown J, McKinlay C. Neonatal Glycaemia and Neurodevelopmental Outcomes : A Systematic Review and Meta-Analysis. Neonatology. 2019;115:116-126. doi:10.1159/000492859