Age and Probability of Spontaneous Umbilical Hernia Closure

Abstract

This cohort study of children younger than 6 years uses electronic health records to investigate whether a child’s age is associated with the probability of spontaneous umbilical hernia closure and to refine guidelines for surgical repair.

Umbilical hernias (UHs) are common in childhood, affecting 10% to 30% of newborns in the US annually.¹ Most resolve spontaneously and the safety of observation is well established. Repair prior to age 4 years is associated with higher rates of recurrence and complications, and practice variation including demographic disparities has been reported.^2,3,4,5 An important driver of practice variation is the lack of population-based data characterizing spontaneous closure (SC) rates and conflicting recommendations for repair, ranging between ages 3 and 5 years.⁶ We aimed to characterize SC rates for this age group to refine guidelines for surgical referral and repair.

Methods

This retrospective cohort study of children younger than 6 years was conducted from March 2017 to September 2022 at 68 primary care pediatric practices throughout Massachusetts affiliated with Boston Children’s Hospital (BCH). Electronic health records (EHRs) were screened using automated text searches; manual medical record review confirmed diagnoses through physical examination findings or documentation as an active problem in the assessment and plan. Race and ethnicity data, included in the EHR, were not part of this analysis. Umbilical hernia persistence, SC, and repair were assessed through manual review of subsequent EHR encounters. The institutional review board of BCH approved the study; we followed STROBE.

Spontaneous closure was assigned if repair was not documented and 1 of 2 criteria was met: language indicating absence of a UH, or 2 consecutive well-child visits without UH documentation in the examination or active problem list. Accuracy of the EHR in documenting repair (and to assess misclassification of an undocumented repair as SC) was evaluated by cross-referencing the study cohort with all patients who underwent repair at BCH. Competing risks analysis estimated the cumulative SC incidence by age while accounting for the competing events of surgical repair and censored follow-up. Secondary analyses were performed, stratifying on hernia size (small if described as “small” or ≤1 cm, and large if described as “large” or >1 cm).

Spontaneous closure rates were calculated for persistent hernias at each year of age to estimate the probability of resolution with different observation periods. Analyses were performed with SAS, version 9.4 (SAS Institute Inc). Statistical significance was defined as P = .05 (2-tailed).

Results

Of 167 557 eligible patients, 4486 (2.7%) were diagnosed with UH at a median (IQR) age of 1.6 (1.0-2.3) months. Of these, 2257 (50.3%) were male and 2230 (49.7%) were female. Hernia size categorization was possible for 3423 children (76.3%) (small, 2903 [84.8%]; large, 520 [15.2%]). Spontaneous closure occurred in 88.6% of all children by age 5 years, and rates were associated with hernia size (small, 89.5%; large, 80.1%; P = .002; Figure). The probability of SC in children with a persistent UH at age 3 years was 20.0% by 4 years and increased to 34.8% if observed for a 2-year period through age 5 years. The probability of SC between ages 3 and 5 years was 29.4% for large hernias and 37.0% for small hernias during this 2-year observation period (Table). Documentation was present in the primary care EHR for 43 of 44 repairs (97.7%) performed at BCH.

Figure. Kaplan-Meier Estimates of Spontaneous Umbilical Hernia Closure in Children During the First 5 Years of Life.

Interval rates of spontaneous closure represent probability of closure after further observation for hernias that persist at a given age. The shaded areas represent 95% CIs for persistence rates among children with small and large hernias. Hernias were categorized as small if they were described as “small” or measured 1 cm or less, and large if described as “large” or measured greater than 1 cm.

Table. Probabilities of Future Spontaneous Closure Based on Most Recent Age When an Umbilical Hernia Was Documented and Different Observation Periods.

Most recent age when umbilical hernia documented, y	Interval probability of spontaneous closure with further observation by age observed to, %
	1 y	2 y	3 y	4 y	5 y
All hernias
0 (Birth)	64.3	76.9	82.4	86.0	88.6
1		35.4	50.9	60.7	68.0
2			23.9	39.2	50.4
3				20.0	34.8
4					18.4
Small hernias a
0 (Birth)	64.8	77.5	83.3	86.8	89.5
1		36.1	52.5	62.6	70.1
2			25.7	41.4	53.2
3				21.2	37.0
4					20.1
Large hernias a
0 (Birth)	48.2	62.8	71.9	76.1	80.2
1		28.1	45.7	53.9	61.7
2			24.5	35.9	46.7
3				15.0	29.4
4					16.9

^a Hernias were categorized as small if they were described as “small” or measured 1 cm or less, and large if described as “large” or measured greater than 1 cm.

Discussion

This study represents the largest population-based analysis of spontaneous UH closure in children. Of note, although SC rates were lower for larger hernias, the probability of SC between ages 3 and 5 years was relatively high independent of size. These results provide new insight for primary care clinicians, surgeons, and caregivers that can better inform discussions around the timing of repair.

Limitations of this study include its retrospective design and reliance on EHR documentation, although the results are likely to be generalizable and valid based on the robust study cohort and methods used to minimize misclassification. Despite these limitations, the results of this analysis suggest that repair should be delayed in most children until age 5 years.

Supplement.

Data Sharing Statement