Incidence and mortality trends in congenital diaphragmatic hernia in the United States

Data source

First, we performed a retrospective, repeated cross-sectional analysis of the Centers for Disease Control and Prevention’s Wide-Ranging online Data for Epidemiologic Research (CDC WONDER) Natality dataset [11]. The natality dataset reports statistics for live births in the United States to US residents based on the 2003 revised birth certificate. All states in the US have been using the 2003 revised birth certificate since 2015. The data pertains to all live births (stillbirths excluded) and are available by a variety of demographic characteristics, such as the mother’s race, and mother’s age, and health and medical items, such as tobacco use, and method of delivery. The natality dataset is derived from birth certificates, and 99 % of all births in the US are registered with the CDC [12].

Two worksheets were developed to facilitate the accurate collection of data for the completion of the revised 2003 standard birth certificate: the “Mother’s Worksheet” (available at https://www.cdc.gov/nchs/data/dvs/moms-worksheet-2016.pdf) and the “Facility Worksheet” (available at https://www.cdc.gov/nchs/data/dvs/facility-worksheet-2016.pdf). In the Mother’s Worksheet, data are directly obtained from the mother and include items such as race, Hispanic origin, and educational attainment. For the Facility Worksheet, data are obtained directly from the medical records of the mother and infant for items such as the date of the first prenatal care visit, pregnancy risk factors, and method of delivery [12]. On the Facility Worksheet, twelve congenital anomalies are separately identified in a checkbox format: 1) anencephaly; 2) meningomyelocele/spina bifida; 3) cyanotic congenital heart disease; 4) congenital diaphragmatic hernia; 5) omphalocele; 6) gastroschisis; 7) limb reduction defect; 8) cleft lip with or without cleft palate; 9) cleft palate alone; 10) Down syndrome; 11) suspected chromosomal disorder; and 12) hypospadias. This item allows for the reporting of more than one anomaly and includes a choice of “None of the above”. If the item is not completed (i.e., none of the boxes are checked), it is classified as “Not stated”. It is recommended that this information be collected directly from the medical record using the Facility Worksheet.

To examine the CDH-related infant mortality rate (CDH-IMR), we queried the linked birth/infant death dataset within CDC WONDER from 2007 to 2022 for CDH-IMR through 1 year of age. This database includes population-wide linked birth/infant death records, and more than 99 % of infant death records are linked to their corresponding birth certificates [13]. The purpose of the linkage is to use the many additional variables available from the birth certificate to conduct more detailed analyses of infant mortality patterns. The linked files include information from the birth certificate such as age, race, and Hispanic origin of the parents, birth weight, period of gestation, plurality, prenatal care usage, maternal education, live birth order, marital status, and maternal smoking, linked to information from the death certificate such as age at death and underlying and multiple cause of death [14]. The linked birth/infant death dataset has been used previously to examine the trends in cause-specific infant mortality in the US [15], 16]. This cross-sectional study did not require Institutional Review Board approval or patient informed consent because it used publicly available de-identified data per Common Rule 45 CFR § 46.

Study population, exposures, and outcomes

We queried the 2016–2023 natality dataset for live births with CDH (ICD-10 code Q79.0) for each year [17]. Next, we queried the linked birth/infant death records for infant deaths with CDH as the underlying cause of death. The World Health Organization defines the underlying cause of death as the disease or injury that initiates a sequence of events that leads directly to death. The exposure was the year of birth and the year of death, and the outcomes were the changes over time in the incidence of CDH and CDH-IMR. The incidence of CDH and CDH-IMR were further stratified by the mother’s bridged race, gestational age, birth weight, geographic region, and sex. We used race data recorded on birth and death certificates, as captured in CDC WONDER. Race categories included Black or African American, White, Asian, American Indian or Alaskan Native, more than one race, Native Hawaiian, and others. Data on the mother’s bridged race was available up to 2019. Except for non-Hispanic Black (NHB), non-Hispanic White (NHW), and Hispanic ethnicity, we did not study the other racial/ethnic categories due to suppression constraints (death counts fewer than 20) for most of the years of the study. Urbanization level was categorized into metropolitan and nonmetropolitan regions. Metropolitan regions included large central metropolitan regions, large fringe metropolitan regions, medium metropolitan regions, and small metropolitan regions. Nonmetropolitan regions included micropolitan regions and noncore nonmetropolitan regions. The incidence of CDH was calculated and expressed as per 10,000 live births. The CDH-IMR was calculated and expressed as per 100,000 live births.

Statistical analysis

We tabulated baseline demographic and perinatal characteristics of infants with CDH with frequencies and proportions. The Shapiro-Wilk and Kolmogorov-Smirnov tests were used to evaluate whether continuous data were normally distributed. Continuous data with normal distribution were presented as means (SD) and were analyzed using the t-test. Continuous data with nonnormal distribution were presented as median with interquartile range (25th-75th quartile) and were analyzed using the Mann-Whitney U test for two groups. Differences among racial groups or geographic regions were ascertained using analysis of variance (ANOVA) with posthoc Tukey’s Honestly Significant Difference (HSD) test. p-value <0.05 defined statistical significance.

We used Joinpoint regression version 5.0.2 (National Cancer Institute, Bethesda, Maryland, USA) to examine trends in the incidence of CDH (2016 through 2023) and CDH-IMR (2007 through 2022) [18]. Joinpoint is statistical software for the analysis of trends using joinpoint models, where several different lines are connected at the inflection points or “joinpoints.” The software takes trend data and fits the simplest joinpoint model that the data allow, starting with the minimum number of joinpoints (for example, zero joinpoints is a straight line), and tests whether more joinpoints are statistically significant and must be added to the model, up to the maximum number of joinpoints the data allow. The models may incorporate estimated variation for each point (e.g., when the responses are age-adjusted rates) or use a Poisson model of variation. In addition, the models may also be linear on the log of the response (e.g., for calculating annual percentage rate change). Trends were reported using average annual percentage change (AAPC) with 95 % confidence intervals (CI). The trend was considered significant if the 95 % CI did not include zero. Because the study objective was to examine the temporal trends (rather than the causal relationship between calendar years and outcomes), we did not adjust for covariates [19]. Further, to avoid any trends and racial differences from being masked, we did not perform an adjusted analysis [7]. The analysis for this manuscript was completed between 1st to 7th February 2025.

Incidence of CDH

Among 29, 880, 509 live births between 2016 and 2023, 3,797 had the presence of CDH (1.3 per 10,000) indicated on their birth certificates. Of these, 57.2 % were male, 59.3 % were Non-Hispanic White (NHW), 11 % were Non-Hispanic Black (NHB), and 82.1 % were in metropolitan areas. Furthermore, the average gestational age and birthweight were 37 weeks and 2,852 gm, respectively, and 33.4 % were transferred within 24 h after birth (Table 1).

As shown in Table 2, the incidence of CDH was more likely to be significantly higher in preterm and low birth weight infants, males, rural areas, NHW infants, and in the Midwest census region. There was no significant change in the CDH incidence during the study period (remained stable at 1.2 to 1.3 per 10,000; AAPC 0.93 %; CI: −0.1, 2.0).

Infant mortality rate attributable to CDH

From 2007 through 2022, 3,564 infants out of a total livebirth population of 62.8 million had CDH listed as the underlying cause of death through 1 year of age (overall CDH-IMR of 5.66 ± 0.49 per 100,000). Of all the deaths attributed to CDH, 7.4 , 27.6, 18.95, 24.45, and 21.7 % died within<1 h, 1–23 h, 1–6 days, 7–27 days, and 28–364 days of life, respectively. Thus, the majority of deaths from CDH occurred within the first 7 days of life (53.9 %). The CDH-IMR declined significantly from 6.32 in 2007 to 4.72 in 2022 (APC: −1.5 %; CI: −2.2 to −0.8 %) [Figure 1].

Figure 1:

Changes in the CDH-IMR over time in the United States from 2007 through 2022.

As shown in Table 3, the CDH-IMR was significantly higher in males (6.11 ± 0.77 vs. 5.19 ± 0.41 in females; p<0.001), but there was no significant difference between NHW and NHB (6.29 ± 0.76 vs. 6.05 ± 0.66; p=0.79). However, the CDH-IMR was significantly lower in Hispanic infants (5.40 ± 0.52) when compared to NHB infants (p=0.009) but not significantly lower than NHW infants (p=0.08). The CDH-IMR did not change in NHB infants (APC 0.03 %; CI: −2.9 to 3.7 %), males (AAPC: - 0.3 %; CI: −1.3, 0.7), and females (AAPC: −0.3 %; CI: −1.3, 0.7). For NHW, the CDH-IMR decreased significantly from 6.54 to 4.61 per 100,000 live births from 2007 to 2019 (APC: −1.8 %; CI: −3.2 to −0.4 %).