MANUSCRIPT CITATION

Hintz SR, Vohr BR, Bann CM, Taylor HG, Das A, Gustafson KE, et al.; SUPPORT study group of the Eunice Kennedy Shriver National Institute of Child Health and Human Development Neonatal Research Network. Preterm Neuroimaging and School-Age Cognitive Outcomes. Pediatrics. 2018;142. pii: e20174058. doi: 10.1542/peds.2017-4058. PMID: 29945955.

REVIEWED BY

Dr Nandkishor S Kabra
Surya Hospitals, Mumbai

TYPE OF INVESTIGATION

Prognosis

QUESTION

(P) In preterm infants <28 weeks gestation,  do (I) near-term conventional brain MRI findings  compared with  (C) adverse early and late cranial ultrasound improve the prediction of  (O) school age cognitive outcomes, when assessed between  (T) the age of 6 years 4 months to 7 years 2 months.

METHODS

• Design: Neuroimaging and neurodevelopmental Outcomes (NEURO) study was a prospective multi center cohort study. The participating centers were from NICHD NRN.  It is a secondary study  to the SUPPORT study. This study was approved by institutional review boards of participating centers. The SUPPORT study was prospectively registered with clinicalTrials.gov.
• Allocation: Allocation concealment not applicable.
• Blinding: Neuroimaging cranial ultrasound were evaluated by two central readers independent of each other and were blinded. Brain MRI was evaluated by one central reader who was blinded. There was no clarity regarding blinding of neurodevelopment follow-up assessment at early school age.
• Follow-up period: Subjects were followed to school age ranging between 6 years 4 months to 7 years 2 months.
• Setting: Participating centers from NICHD NRN, USA.
• Patients: A subset of preterm infants <28 weeks gestational age who had participated in SUPPORT cohort. SUPPORT study enrolled infants who were born between 24 weeks 0 days and 27 weeks 6 days of gestation. Not all SUPPORT study centers participated.
• Dates: Infants were enrolled in SUPPORT study from February 2005 through February 2009. The NEURO study was approved and began recruitment after SUPPORT began enrollment. All the participating centers did not launch simultaneously.
• Intervention: Infants underwent early cranial ultrasound (CUS) assessment between 4 to 14 days of age. Late CUS assessment was performed at 35 to 42 weeks of postmenstrual age. Conventional brain MRI was obtained at 35 to 42 weeks of postmentsrual age within 2 weeks of late CUS.
• Adverse findings:
  • A composite adverse finding on early CUS was defined as the presence of grade III or IV intracranial hemorrhage (ICH)‍ or cystic periventricular leukomalacia (cPVL) on either or both sides.
  • A composite adverse finding on late CUS was defined as having cPVL or porencephalic cyst, moderate to severe ventricular enlargement (VE) on either or both sides, or a shunt.
  • For all CUS, assessment of interobserver reliability between central readers revealed κ = 0.75 for the early CUS composite adverse finding and a κ = 0.88 for the late CUS composite adverse finding.
  • Brain MRIs were reported by using a central reader form that included white matter abnormality (WMA)  scoring classification system used to evaluate 5 areas of white matter assessment.‍ Interrater agreement for moderate or severe WMA by using this classification system has been reported to be >95%.‍  Significant cerebellar lesions were defined as lesions that were bilateral, cystic, and/or ≥4 mm in size. Adverse findings on brain MRI were defined as moderate or severe WMA or significant cerebellar lesions.
• Outcomes: Neurodevelopmental follow-up assessment at early school age.
• Primary Outcomes:
  • Prospectively defined outcomes in the study were:
    • Significant cognitive impairment defined as a full-scale IQ (FSIQ) <70.
    • Moderate-to-severe disability defined as an FSIQ <70, CP with a Gross Motor Function Classification System (GMFCS) level ≥2, severe hearing impairment, or severe vision impairment.
• Secondary outcomes:
  • FSIQ <85.
  • Minimal or no disability, which was defined as having all of the following: an FSIQ >85, no CP, and no hearing or vision impairment or impairments that were completely correctable.
  • Severe disability, which was defined as an FSIQ <55, CP with a GMFCS level of 4 or 5, or severe hearing or severe vision impairment.
• Analysis and Sample Size: Formal sample size calculations were not performed. Sample of available subjects for cohort study was utilized. Sensitivity, specificity, positive predictive value and negative predictive values for school-age outcomes for early CUS, late CUS and brain MRI findings were calculated. ROC curves were estimated for various mixed models.
• Patient follow-up: In the SUPPORT study 1316 infants were enrolled. A total of 480 infants had required neuroimaging data in NEURO study.  Out of these 480 children, 17  died before school age assessment and 77 were lost to follow-up.  Therefore, 386/463 (83%) children had school age cognitive outcome data.

MAIN RESULTS

The average birth weight of NEURO study cohort was 861.8 ± 190.1 grams with an average gestational age of 25.9 ± 1 week.  The use of antenatal steriods was 96% (371/386). The baseline demographic characteristics of 77 infants that were lost to follow-up were broadly similar to those who were followed up in the study.

Primary Outcomes:

Sensitivity of all imaging modalities was poor. Specificity of all imaging modalities was moderate to good. PPV of all imaging modalities varied from poor to moderate. NPV of all imaging modalities varied from poor to good. Results are summarized in the following tables.

A late CUS adverse finding was the best amongst the three neuroimaging tools under consideration in regards to the diagnostic test properties of specificity and NPV for the outcomes of FSIQ<70, severe disability and moderate to severe disability.

Secondary Outcomes:

In full regression models, both adverse late CUS findings (odds ratio [OR] 27.9; 95% confidence interval [CI] 6.0–129) and significant cerebellar lesions on MRI (OR 2.71; 95% CI 1.1 – 6.7) remained associated with disability, but only adverse late CUS findings (OR 20.1; 95% CI 3.6 –11) were associated with FSIQ <70. Predictive accuracy of stepwise models did not considerably improve with the addition of neuroimaging.

CONCLUSION

The authors conclude that adverse late CUS findings were most strongly associated with cognitive impairment and disability at school age. Significant cerebellar lesions on MRI were also associated with disability. Near-term conventional MRI did not substantially enhance the prediction of severe early school-age outcomes.

COMMENTARY

Infants born at <28 weeks of gestation are at increased risk of neuro-sensory impairments. Although, neonatologists are generally aware about likely long term outcomes in this group of infants, there is always some degree of uncertainty in precisely predicting long term neurological outcome of an individual infant prior to discharge. Several studies have described utility of neonatal cranial ultrasound and brain MRI in prediction of long-term neurological outcomes (1-5). This study has for the first time revealed that late cranial ultrasound findings were most strongly associated with adverse school age outcomes and near term conventional brain MRI did not significantly enhance prediction.

The manuscript described is a multi center large prospective cohort study of extremely preterm infants that describes the prognostic ability of early and late CUS and brain MRI in predicting early school-age outcomes.   Assessment of early and late CUS and brain MRI was performed by assessors who were blinded. The prospectively defined outcomes in this study were:  [1] significant cognitive impairment defined as a full-scale IQ (FSIQ) <70; and [2] moderate-to-severe disability defined as an FSIQ <70, CP with a Gross Motor Function Classification System (GMFCS) level ≥2, severe hearing impairment, or severe vision impairment.  The other outcomes studied were: [1] FSIQ <85, [2] minimal or no disability (which was defined as having all of the following: an FSIQ >85, no CP, and no hearing or vision impairment or impairments that were completely correctable), [3] severe disability (which was defined as an FSIQ <55, CP with a GMFCS level of 4 or 5, or severe hearing or severe vision impairment).

Out of the 480 children in the NEURO study 17 had died before school age assessment and 77 were lost to follow-up.  Therefore, 386/463 (83%) children had school age cognitive outcome data.  The strengths of the study include large sample size, robust methodology and long term follow up up to school-age. One of the weaknesses of the study was that CUS imaging was performed as per local center’s  protocol rather than standardized protocol. However, it should not affect the overall interpretations from the study.  Conventional brain MRI study had prespecified minimum requirements. It is also important to note that, 14% of those lost-to-follow up received postnatal steroids compared to 7% of those in the study group, p = 0.032. This could potentially change the outcomes.

There is no clarity regarding blinding of neurodevelopmental follow-up assessment at early school age. In full regression models, adverse late CUS findings and significant cerebellar lesions on MRI were associated with poor long term neurological outcomes at school age. However, it is important to note that the 95% confidence interval around these estimates are very wide. A previous publication by the same authors reported that the near-term CUS and MRI abnormalities were associated with adverse 18- to 22-month outcomes, independent of early CUS and other factors, underscoring the relative prognostic value of later neuroimaging (5).

Prognosticating long term neurological outcome in an individual infant with gestational age <28 weeks remains challenging and is dependent on many other factors (such as: gestational age, birth weight, race, gender, multiple gestation, maternal education, use of antenatal steroids, sepsis, cranial ultrasound abnormalities, bronchopulmonary dysplasia, postnatal steroids use, surgery, necrotizing enterocolitis and retinopathy of prematurity). Late CUS findings are consistently associated with cognitive impairment and disability at school age. There still remains the uncertainty regarding the preferred imaging modality of choice or prognostic models that will help in counseling the families of extremely premature infants about the long term neurodevelopmental outcomes. Further research is required to build predictive models for infants born at <28 weeks gestation that are not only sensitive but also specific in predicting long term neurologic outcomes. In addition there is an urgent need  for studying interventions in neonatal intensive care units that attenuate or prevent the neurologic  injury and prevent long term complications. Additional benefits of near-term MRI needs further evaluation in context of the health care utilization prior to recommending its routine use in NICU graduates.

REFERENCES

1. Hintz SR, Barnes PD, Bulas D, Slovis TL, Finer NN, Wrage LA, et al; SUPPORT Study Group of the Eunice Kennedy Shriver National Institute of Child Health and HumannDevelopment Neonatal Research Network. Neuroimaging and neurodevelopmental outcome in extremely preterm infants. Pediatrics 2015; 135: e32-42.
2. Tam EW, Rosenbluth G, Rogers EE, Ferriero DM, Glidden D, Goldstein RB, et.al; Cerebellar hemorrhage on magnetic resonance imaging in preterm newborns associated with abnormal neurologic outcome. J Pediatr 2011; 158: 245-50.
3. Schmidt B, Asztalos EV, Roberts RS, Robertson CM, Sauve RS, Whitfield MF; Trial of Indomethacin Prophylaxis in Preterms (TIPP) Investigators. Impact of bronchopulmonary dysplasia, brain injury, and severe retinopathy on the outcome of extremely low-birth-weight infants at 18 months: results from the trial of indomethacin prophylaxis in preterms. JAMA 2003; 289: 1124-9.
4. Inder TE, Wells SJ, Mogridge NB, Spencer C, Volpe JJ. Defining the nature of the cerebral abnormalities in the premature infant: a qualitative magnetic resonance imaging study. J Pediatr 2003; 143: 171-9.
5. Hintz SR, Barnes PD, Bulas D, Slovis TL, Finer NN, Wrage LA, et.al; SUPPORT Study Group of the Eunice Kennedy Shriver National Institute of Child Health and Human Development Neonatal Research Network. Neuroimaging and neurodevelopmental outcome in extremely preterm infants. Pediatrics 2015; 135: e32-42.