Long-term Consequences of Necrotizing Enterocolitis in the Newborn Period

Rationale

Knowledge of the long-term complications related to necrotizing enterocolitis (NEC) in the neonatal period is essential, when clinicians are counseling parents about the risk for complications following NEC. When deciding upon a follow-up strategy for the infant after discharge, this knowledge is also crucial. If the infant has an increased risk for complications later in life, a closer follow up is important to assure proper handling of complications.

So far, a definite conclusion regarding long-term complications of NEC has not been made. A systematic review including ten studies, have shown a significantly increased risk for neurodevelopmental impairment in NEC children. However, the median follow-up time did not exceed two years and may not reflect the true long-term complications. The ORACLE Children Study investigated neurodevelopmental impairments in children with a history of NEC at 7,5 year follow-up and found no increased risk.

No review has been published regarding growth impairments and so far, available studies have a less than three year follow-up period and are reporting conflicting results.

Gastrointestinal morbidities are even less investigated and reports no association with NEC within three years follow-up. However, the ORACLE Children Study found increased risk of bowel problems at 7,5 years follow-up in children with a history of NEC.

The Oracle Children Study and Roze et al reported behavioural problems in up to 30% of children with a history of NEC at school age, but no statistic significance was found when compared to a control-group.

Further studies investigating the outcome of NEC-children at school age are needed.

Objectives

The investigators plan to conduct a retrospective parental-questionnare study, to determine the long-term complications related to the diagnosis of NEC at school age (5-15 years of age).

Methods

This is a case-control study. Case-group will include all surviving children born in Denmark between 1st of January 2002 and 31st of December 2011 with an international classification of diseases (ICD-10) diagnosis of NEC at discharge. Matching each child with a history of NEC to two children with no history of NEC by gestational age, year of birth and birthweight will form the control group. Matching criteria's have been chosen to balance prematurity and age at follow-up between case and control-group. Parental follow-up questionnaires regarding complications will be sent to all parents of cases and controls.

Clinical data on the children, regarding birth and neonatal history, will be obtained and used in a mixed regression model to adjust for possible confounders.

Primary outcome

Abnormal or borderline 'total difficulties score' in the strenghts-and-difficulties-questionnaire as assessed by parents.

Secondary outcomes

1. Severe cerebral palsy (GMFCS III-IV)
2. Head circumference as measured by parents
3. Height as measured by parents
4. Abdominal symptoms resulting in significant absence from school as reported by parents
5. Diarrhoea within the last six months
6. Constipation within the last six months

Power calculation

The sample size for the case group was estimated to include 200 children with NEC (after a no-response drop out of 20%) and in the control group 300 children without NEC, with a 1:2 match ratio and the assumption that the response rate will be lower in the control group., i.e. 250 parents of children with NEC and 500 parents of control children will be contacted.

SDQ total score Assuming 30% prevalence of abnormal or borderline scores for 'total difficulties' in the strength-and-difficulties-questionnaire in NEC-survivors, compared to an assumed 20% in controls (19), this study will detect the difference with a power of 72% at a 5% significance level.

Severe cerebral palsy (GMFCS score III-V) Assuming a 20% prevalence (9) of severe cerebral palsy in NEC-survivors compared to an assumed 10% in controls (22) this study will detect the difference with a power of 87% at a 5% significance level.

Head circumference The power calculation on our cohort shows that we, at a 5% significance level, can detect the difference in mean head circumference at 0.3 standard-deviations or more, with a power of 90%.

Analysis plan

Primary outcome: a p-value < 0.05 will be considered significant. To determine a difference in SDQ-scores between case- and control group, Chi-squared test or Fischer's Exact test will be used as appropriate.

Secondary outcome: a p-value < 0.01 will be considered significant. Differences in 'height', 'head circumference' and 'absence from school due to abdominal symptoms' will be investigated by t-test or Mann-Whitney's test as appropriate. 'Severe cerebral palsy', 'diarrhoea' and 'constipation' will be analyzed by Chi-squared test or Fischer´s Exact test as appropriate.

At first, an unadjusted analysis will be performed to investigate differences in primary and secondary outcome. Secondly, a mixed regression model will be used to adjust for possible confounders