MANUSCRIPT CITATION

Shankaran S, Laptook AR, Pappas A, McDonald SA, Das A, Tyson JE, Poindexter BB, Schibler K, Bell EF, Heyne RJ, Pedroza C, Bara R, Van Meurs KP, Grisby C, Petrie Huitema CM, Garg M, Ehrenkranz RA, Shepherd EG, Chalak LF, Hamrick SEG, Khan AM, Reynolds AM, Laughon MM, Truog WE, Dysart KC, Carlo WA, Walsh WC, Watterberg KL, Higgins RD for the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD) Neonatal Research Network. Effect of depth and duration of cooling on deaths in the NICU among neonates with hypoxic ischemic encephalopathy. JAMA 2014; 312: 2629-2639. PMID: 25536254

REVIEWED BY

Dmitry Dukhovny
Assistant Professor of Pediatrics
Oregon Health & Science University

John AF Zupancic
Associate Professor of Pediatrics
Harvard Medical School

TYPE OF INVESTIGATION

Treatment

QUESTION

In infants born at greater than or equal to 36 weeks with moderate or severe hypoxic ischemic encephalopathy undergoing therapeutic hypothermia, does longer duration of cooling (120 hours vs. 72 hours), deeper cooling (32.0 C vs. 33.5 C), or both, improve death or disability at 18-22 months of age, as compared to current standard of care (33.5 C for 72 hours)?

METHODS

• Design: Randomized Control Trial, 2×2 Factorial Design
• Allocation: Concealed
• Blinding: Unblinded (patients, clinicians), unclear blinding (data collectors, analysts). Blinding would be likely impossible given the nature of the intervention.
• Follow-up period: Study protocol specified follow-up through 18 to 22 months of age. The current report is an interim analysis through hospital discharge, presented after enrollment was halted by the Data Safety and Monitoring Committee (DSMC) because of concerns regarding safety and futility.
• Setting: 18 US Centers in the NICHD Neonatal Research Network.
• Patients: 364 patients were randomized prior to trial closure by the DSMC, representing 50% of the projected sample size of n=726. Average time at randomization was 5 hours of life, mean cord pH was 6.9, 61.3% white, 35.8% inborn.
  • Inclusion criteria: Infants >36 weeks GA admitted to the NICU within 6 hours of birth, were screened for eligibility if they had poor respiratory effort at birth, a need for resuscitation, or a diagnosis of encephalopathy. Physiological eligibility criteria included a pH of <7.0 or a base deficit of > 16 mmol/L in any blood in first hour of life (including umbilical cord blood). If pH was 7.01-7.15, a base deficit was between 10 mmol/L and 15.9 mmol/L, or no blood gas was available, infants had to have had an acute perinatal event (e.g. late or variable decelerations, cord prolapse, cord rupture, uterine rupture, maternal trauma, hemorrhage, or acute cardiorespiratory arrest) and either a 10-minute Apgar score of 5 or less or assisted ventilation initiated at birth and continued for at least 10 minutes. If these criteria were present, investigators had to ensure the infant was moderately or severely affected according to a standardized neurological assessment. This assessment was used by this trial group previously.¹ This exam aimed to define the level of encephalopathy: from moderate to severe in at least 3 of 6 categories: (1) level of consciousness, (2) spontaneous activity (3) posture (4) tone (5) primitive reflexes and (6) autonomic nervous system. However the level of consciousness was the main determining feature.
  • Exclusion criteria: birth weight <1800 grams, core temperature <32.5°C for two hours, major congenital anomaly, unable to randomize within six hours of birth, or moribund status with no plans for further intensive care.
• Intervention: Patients were randomized to cooling at 33.5°C or 32.0°C and for 72 hours or 120 hours duration in a 2×2 factorial design, (making four groups), within the first six hours of life.
• Outcomes: The primary outcome data of this study was death or disability (moderate or severe) at 18 to 22 months. This outcome will be reported when data collection is complete and is not included in this manuscript. This current manuscript reports both adverse events (most predefined) and secondary outcomes. Adverse events included cardiac arrhythmia, persistent acidosis, thrombosis, major bleeding, alteration of skin integrity and death. Other predefined secondary outcomes were reported, including: hypotension, blood/platelet transfusions, persistent pulmonary hypertension of the newborn [including treatment with inhaled nitric oxide and ECMO (ECMO was a post-hoc adverse event outcome identified at the DSMC review)], oliguria or anuria, hepatic dysfuction, culture proven infection, disseminated intravascular coaguopathy, hypoglycemia, hypocalcemia, length of hospital stay (LOS), oxygen/ventilation days, gastric fundal plication, gastrostomy, tracheostomy, home therapy with ventilation/oxygen/gavage/gastrostomy tube/anticonvulsant medication.
• Analysis and Sample Size: The authors assumed no statistical interaction between depth and duration of cooling. Primary comparison were the 2×2 marginals, firstly of target temperatures (33.5°C vs 32.0°C), and secondly duration (72 hours vs 120 hours). The primary outcome baseline assumptions for death or disability of 45% at 18 month of age for the 33.5°C/72 hour group was based on prior work.¹ It was proposed to reduce the primary outcome event rates to 37.5% in the first marginal (of depth of cooling), and to 27.5% in the second marginal (duration). Enrolling 726 neonates (363 per marginal group comparison) including a 5% loss to follow-up, would have an 80% power and a two-tailed 0.05 Type I error. Safety outcomes were compared between 72 hours and 120 hours using Poisson regression in a GLE model. Treatment interactions between depth and duration of cooling were also assessed. All p values were 2-sided but not corrected for number of comparisons. Finally, interim stopping rules by Pocock bounds had been calculated, and the DSMC requested futility analyses calculated by conditional power estimates given actual data.
• Patient follow-up (% included in analysis): All infants randomized were included in this intention to treat analysis.

MAIN RESULTS

Death in the NICU was as follows for the four subgroups:

When adjusted for degree of encephalopathy and intracenter correlations, the risk ratio for the 120 hours vs. 72 hours was 1.37 (95% CI, 0.92-2.04), and for 32.0°C vs 33.5°C was 1.24 (95% CI, 0.69-2.25).

The predefined safety outcomes were different for only the frequency of major bleeding (3% in the 72 hour group vs 1% in the 120 hour group, p=0.02). There were other notable differences between groups with respect to some of the predefined secondary outcomes, including LOS (21.6 days (SD=15.3 days) in the 72 hour vs. 26.4 days (SD=34.1 days) in the 120 hour group, p=0.002), arrhythmias (1% in the 72 hour vs. 7% in the 120 hour group, p=0.02). Unexpected safety outcomes that were found primarily related to pulmonary hypertension, such that inhaled nitric oxide use increased (24% in 33.5°C vs 34% in the 32.0°C group, p = 0.03), as did ECMO use (4% in 33.5°C vs 9% in the 32.0°C group, p = 0.005).

CONCLUSION

Longer (120 hrs vs. 72 hrs) and deeper (32 vs. 33.5) cooling, either alone or combined, did not improve in-hospital mortality for infants with HIE, and may be associated with adverse consequences.

Visit Acta to access a pdf copy of this EBNEO commentary!

COMMENTARY

On the basis of 11 trials of both whole body and selective approaches, therapeutic hypothermia initiated within six hours of birth and continued at 33.5 degrees for 72 hours, has become the standard of care for neonates with hypoxic ischemic encephalopathy. Despite the relatively low number needed to treat (4-12, depending on the trial and the outcome), this therapy appears most efficacious for mild to moderate encephalopathy and not severe, and rates of impairment remain high even in the treated population.^2-4

In this study, Shankaran et al ⁵ tested the hypothesis that longer duration and depth of cooling might confer further benefit for neonates with hypoxic ischemic encephalopathy. At approximately half of the expected enrollment, the DSMC closed the trial based on futility analysis and emerging safety concerns for increased mortality. Specifically, at that point, there was a non-significant increase in NICU mortality (adjusted risk ratios for the 120 hours vs. 72 hours was 1.37 (95%CI, 0.92-2.04), and for 32.0°C vs 33.5°C was 1.24 (95% CI, 0.69-2.25). Yet there were also statistically significant increases in the need for ECMO for 32.0°C vs 33.5°C group (number needed to harm (NNH) =20, p=0.005), inhaled nitric oxide for 32.0°C vs 33.5°C group (NNH=10, p=0.03), arrhythmia for 120 hours vs. 72 hours (NNH=17, p=0.02). Of note, both ECMO and nitric oxide were not part of the predefined DSMC criteria for stoppage. A futility analysis showed that the probability that longer or deeper cooling would be superior was estimated to be 2%. In addition, resource utilization was higher in the more aggressively treated group, with LOS at 26.4 days versus 21.6 days, more than could be accounted for by just 2 additional days of cooling.

It is important to note that the trial was halted despite not having reached the pre-specified safety stopping bounds by the investigators.⁵ Prematurely stopped trials are known to overestimate effect size,⁶ and at least some of the many variables used to support the decision were not on the list of pre-specified secondary outcomes. Indeed, as an example of the confusion that may result from exploratory comparisons, major bleeding was higher in the 72-hour group vs. 120 hour group (NNH = 50, p=0.04). We are left, therefore, with a cautionary tale that “more may not be better, and may be worse”. However a neonatology community keen to aggressively expand the indications for cooling, may argue that we do not have the needed precision to convince.^7,8 The dilemma for the trialists and the DSMC here was obvious.⁹ The increased need for ECMO seen in neonates randomized to cooling to 32.0°C was compelling, and it would have taken a very brave DSMC to ignore the accumulating signal. They may still face criticism for having “jumped the gun,” but most clinicians will likely be reassured by their vigilance.

With these caveats, however, it appears reasonable to suggest that there is no reason to pursue deeper or longer duration of therapeutic hypothermia in this population. Since the pre-stated and clinically most important outcome is composite death and disability, the results of the 18 to 22 month follow up will be a critical addition to the interpretation of the study. For now, the more important lesson is arguably that any expansion of therapy beyond that in high-quality extant studies must be subject to safety and efficacy testing in randomized trials. If possible, future studies should be continued to conclusion, subject to the planned stopping rules.

REFERENCES

1. Shankaran S, Laptook AR, Ehrenkranz RA, et al. Whole-body hypothermia for neonates with hypoxic-ischemic encephalopathy. N Engl J Med 2005;353:1574-84.
2. Azzopardi D, Strohm B, Marlow N, et al. Effects of hypothermia for perinatal asphyxia on childhood outcomes. N Engl J Med 2014;371:140-9.
3. Jacobs SE, Berg M, Hunt R, Tarnow-Mordi WO, Inder TE, Davis PG. Cooling for newborns with hypoxic ischaemic encephalopathy. Cochrane Database Syst Rev 2013;1:CD003311.
4. Shankaran S, Pappas A, McDonald SA, et al. Childhood outcomes after hypothermia for neonatal encephalopathy. N Engl J Med 2012;366:2085-92.
5. Shankaran S, Laptook AR, Pappas A, et al. Effect of depth and duration of cooling on deaths in the NICU among neonates with hypoxic ischemic encephalopathy: a randomized clinical trial. JAMA 2014;312:2629-39.
6. Bassler D, Briel M, Montori VM, et al. Stopping randomized trials early for benefit and estimation of treatment effects: systematic review and meta-regression analysis. JAMA 2010;303:1180-7.
7. Austin T, Shanmugalingam S, Clarke P. To cool or not to cool? Hypothermia treatment outside trial criteria. Arch Dis Child Fetal Neonatal Ed 2013;98:F451-3.
8. Thoresen M. Who should we cool after perinatal asphyxia? Semin Fetal Neonatal Med 2015;20:66-71.
9. Roberts RS. Early closure of the Watterberg trial. Pediatrics 2004;114:1670-1.