MANUSCRIPT CITATION

HIP Trial Investigators; Blakely ML, Krzyzaniak A, Dassinger MS, Pedroza C, Weitkamp J, Gosain A, et al. Effect of Early vs Late Inguinal Hernia Repair on Serious Adverse Event Rates in Preterm Infants: A Randomized Clinical Trial. JAMA 2024;331(12):1035-44. PMID 38530261.

REVIEWED BY

Puneet Sharma, MD
Clinical Fellow
Division of Newborn Medicine, Boston Children’s Hospital
Department of Pediatrics, Harvard Medical School
puneet.sharma@childrens.harvard.edu

Brian King, MD
Neonatologist
Department of Neonatology, Beth Israel Deaconess Medical Center
Department of Pediatrics, Harvard Medical School
bking6@bidmc.harvard.edu

TYPE OF INVESTIGATION

Treatment

QUESTION

(P) Among preterm infants born at less than 37 weeks’ gestation diagnosed with inguinal hernia during initial hospitalization, (I/C) does inguinal hernia repair before NICU discharge when compared to repair after NICU discharge and 55 weeks’ postmenstrual age (O) lead to more serious adverse events (T) in a 10-month observation period after randomization?

METHODS

• Design: Multicenter, randomized clinical trial
• Allocation: The infants were randomly assigned when approximately 2 weeks away from anticipated NICU discharge in a 1:1 ratio to either early or late repair.Randomization was performed continuously in permutated blocks ranging from 2 to 4 in size and stratified by center and gestational age.  The specific treatment group allocation was concealed until the infants were randomized.
• Blinding: Given the timing of repair as the primary intervention, blinding the clinicians to the study intervention was not feasible.However, the primary outcome was determined by a committee blinded to which intervention the patient received.
• Follow-up period: 10 months after initial randomization, follow-up at 2 years corrected age is ongoing.
• Setting: 39 medical centers in the United States
• Patients:
  • Inclusion criteria: preterm infants born at less than 37 weeks’ gestation who were diagnosed with inguinal hernia by a pediatric surgeon during initial hospitalization to participating center between September 2013 and April 2021
  • Exclusion criteria: clinical factors that exclude early inguinal hernia repair, inguinal hernia repair planned as secondary procedure to another operative procedure, known major congenital or chromosomal anomaly, family unable to return for follow-up and late inguinal hernia repair.
• Intervention: Early inguinal hernia repair, which was defined as repair promptly after randomization and prior to NICU discharge, versus late inguinal hernia repair, which was defined as repair performed when the infant was older than 55 weeks postmenstrual age.
• Outcomes:
  • Primary outcome: the proportion of infants with at least one serious adverse event during 10-month observation period.Serious adverse events were categorized into pulmonary events, cardiac events, surgical events, events related to hernia, or death.
  • Secondary outcome: number of days in hospital during 10-month observation period, including neonatal intensive unit stay after randomization.
• Analysis and Sample Size:
  • The authors made a priori hypothesis that the early repair group would have a 10% higher rate of severe adverse events than the late repair group.To have 80% power to detect a difference of this size, a sample size of 586 infants was required.  The authors assumed a 5% loss rate, so they had planned to enroll 615 infants.
  • For this study, an intention-to-treat analysis was performed.Frequentist and Bayesian analyses were both prespecified in the trial protocol.
  • A logistic mixed-effects model was used to analyze the primary outcome, and a negative binomial mixed model was used for the secondary outcome.Gestational age was included as a covariate and study site as a random intercept.
  • All Bayesian analyses used neutral priors, with a prior centered at an odds ratio of 1.0 with a 95% credible interval (CrI) of 0.33 to 3.0. Prespecified secondary analyses included stratification by gestational age categories, diagnosis of bronchopulmonary dysplasia (BPD), and open vs laparoscopic approach.
  • A formal interim analysis was performed at 50% of anticipated enrollment which found a 97% probability of decreased rate of severe adverse events in the late repair group.This exceeded the stopping threshold of 95% and trial enrollment was stopped.
• Patient follow-up: A total of 1514 infants were assessed for eligibility of which 338 were randomized.A total of 18 infants (9 in each arm) did not undergo treatment as randomized, and a total of 12 infants (4 in early repair group and 8 in late repair group) were lost to follow-up.  As such, 308 of the 338 randomized (91%) were included in primary analysis.
• Funding: This trial was funded by the Eunice Kennedy Shriver National Institute of Child Health and Human Development, U01 HD076733 (awarded to Dr. Blakely).

MAIN RESULTS

The study population was predominantly male (86%), and the median birthweight was 820 g (IQR, 640-1040 g).  There were no significant differences between the two repair groups at enrollment other than the early repair group had slightly more apnea requiring intervention (85% v 79%).  In the early repair group, the median post menstrual age at repair was 41 weeks (IQR, 39-44 weeks) and median weight was 3.1 kg (IQR, 2.5-3.6 kg).  In the late repair group, the median post menstrual age at repair was 57 weeks (IQR, 52-61 weeks) and median weight was 5.9 kg (IQR, 4.7-6.9 kg).  A greater percentage of procedures were laparoscopic in the early repair group than late repair (43% v 36%), and the duration of repair was shorter in the early repair group (57 min v 70 min).

For the primary outcome, 44 of the 159 infants (28%) in the early repair group had at least 1 serious adverse event compared to 27 of the 149 (18%) in the late repair group (RR 0.68,95% CrI, 0.45 to 1.01). The posterior probability of benefit for the primary outcome was 97%.  The relative risk was 0.61 (95% CrI, 0.39 to 0.94) in infants younger than 28 weeks’ gestation, and 0.92 (95% CrI, 0.47 to 1.75) in infants older than 28 weeks’ gestation.  The relative risk was 0.50 (95% CrI, 0.27 to 0.87) in infants with BPD, and 0.85 (95% CrI, 0.51 to 1.37) in infants without BPD.  For the secondary outcome, the median number of days in the hospital was 19 (IQR, 9.8 to 35 days) for the early repair group and 16 (IQR, 7 to 38 days) for the late repair group (RR 0.91, 95% CrI, 0.74 to 1.11), with an 82% posterior probability of benefit.

CONCLUSION

In preterm infants with inguinal hernia, the late repair strategy led to fewer infants with serious adverse events.  A greater effect was noted in infants born at less than 28 weeks’ gestation or with BPD.  Furthermore, there was also a high probability that later repair was associated with fewer overall hospital days.  Based on these findings, the authors suggest that inguinal hernia repair after NICU discharge may be preferred.

COMMENTARY

Optimal timing of inguinal hernia repair is uncertain due to the competing concerns around risk of anesthesia in neonates and risk of hernia-related complications such as incarceration (1). This study is a well-designed, multicentered trial that demonstrates short term benefits to performing inguinal hernia repair after NICU discharge.  Furthermore, the use of Bayesian analyses allows for more pragmatic and clinically oriented interpretation of the outcomes.

 

Late repair led to fewer serious adverse events and fewer days in the hospital, with a greater effect in vulnerable populations like extremely preterm infants and those with BPD.  Of note, there were more laparoscopic procedures in the early repair group, which are associated with shorter operation time and less complications than laparotomy (2).  However, the combination of better outcomes for patients and potentially less healthcare utilization (and costs), makes late repair an appealing choice. Furthermore, this intervention is very straightforward to implement.

 

However, the broad definition of serious adverse event that is used in the composite primary outcome highlights important limitations when interpreting the results. For example, the most common adverse event in both groups was apnea requiring intervention, which could have a very broad range of clinical significance.  Furthermore, cardiac arrest, of which there were three in the late repair group compared to zero in the early, is weighted equally to apnea. Similarly, more infants in the late repair group had incarcerated hernia, but this was broadly classified and included both surgical repairs and bedside reductions.

 

A major limitation with composite outcomes is that they can provide a false equivalency between included outcomes, making it difficult to assess an intervention’s effect on individual outcomes. This limitation is well documented, and yet many trials in neonatology continue to use them due to their benefits when powering trials (3).  Alternate statistical approaches exist to address some of these limitations, one of which is the desirability of outcome ranking (DOOR). DOOR uses a patient-centric approach to rank outcomes from most to least desirable and was recently used in a post-hoc analysis of a neonatal trial (4).  Interpretation of this study’s findings might change using the DOOR approach. For example, more patients in the late repair group did not undergo surgery, often due to spontaneous closure. This outcome was not emphasized by the authors but may be a desirable outcome ranked highly by families.  Similarly, the outcomes of apnea and incarceration would likely not be ranked equivalently using the DOOR approach.

 

Another important factor to consider with late repair is the additional burden it can place on families. Families that were deemed unable to follow up were excluded, but this is an important cohort.  Families may prefer repair during admission as both a financial decision and convenience.  While the late repair group had fewer hospital days, a readmission was required. Greater financial burden may be placed on families with readmission, both due to increased cost-sharing and out-of-pocket costs (5). If many of the adverse events in the early repair group are short term and mild, some families may still elect for early repair depending on their values and preferences.  Differences in long term outcomes, which are being studied, may also alter the potential tradeoffs.  Studies have been conducted to clarify what outcomes are most important to parents of children with bronchopulmonary dysplasia (6).  A similar study in this patient population would undoubtedly provide important context on the timing of repair.

REFERENCES

1. Wang KS; Committee on Fetus and Newborn, American Academy of Pediatrics; Section on Surgery, American Academy of Pediatrics. Assessment and management of inguinal hernia in infants. Pediatrics. 2012;130(4):768-773. PMID: 23008462.

2. Davies DA, Rideout DA, Clarke SA. The International Pediatric Endosurgery Group Evidence-Based Guideline on Minimal Access Approaches to the Operative Management of Inguinal Hernia in Children. J Laparoendosc Adv Surg Tech A. 2020;30(2):221-227. PMID: 28140751.

3. Cordoba G, Schwartz L, Woloshin S, Bae H, Gøtzsche PC. Definition, reporting, and interpretation of composite outcomes in clinical trials: systematic review. BMJ. 2010;341:c3920. PMID: 20719825.

4. Katheria AC, El Ghormli L, Rice MM, Dorner RA, Grobman WA, Evans SR. Application of desirability of outcome ranking to the milking in non-vigorous infants trial. Early Hum Dev. 2024;189:105928. PMID: 38211436.

5. King BC, Mowitz ME, Zupancic JAF. The financial burden on families of infants requiring neonatal intensive care. Semin Perinatol. 2021;45(3):151394. PMID: 33581862.

6. Callahan KP, Kielt MJ, Feudtner C et al. Ranking future outcomes most important to parents of children with bronchopulmonary dysplasia. J Pediatr. 2023;259:113455. PMID: 37172804.