Infantile hypertrophic pyloric stenosis (IHPS) is characterized by muscular hypertrophy of the pyloric sphincter, causing obstruction of the gastric outlet and projectile vomiting in the newborn. This defect usually presents at 2–12 weeks after birth, often in a previously healthy infant, with peak onset at week 5 (Jobson & Hall, 2016). IHPS is one of the most frequently treated surgical conditions in children (El-Gohary, Abdelhafeez, Paton, Gosain, & Murphy, 2018). If left untreated, IHPS can lead to dehydration, weight loss, metabolic alkalosis and, in severe cases, death (Spicer, 1982). Pyloromyotomy is the standard method of treatment for IHPS and generally has a good prognosis and low fatality rate (<1%) (reviewed in MacMahon, 2006; Tigges and Bigham, 2012).

A previous review has suggested an increased predisposition for IHPS in populations living in temperate regions of North America and Western Europe compared to those in tropical countries (Spicer, 1982). Reported prevalence estimates (per 10,000 live births) for IHPS range from 17 to 50 in the United States (US) (Applegate & Druschel, 1995; Markel, Proctor, Ying, & Winchester, 2015; Schechter, Torfs, & Bateson, 1997; St Louis et al., 2017) and Western Europe (de Laffolie, Turial, Heckmann, Zimmer, & Schier, 2012; Hedback, Abrahamsson, Husberg, Granholm, & Oden, 2001; O’Donoghue et al., 1993; Pedersen et al., 2008; Sommerfield et al., 2008; Sule, Stone, & Gilmour, 2001). Some populations in Western Europe have reported declines in IHPS prevalence across different birth periods from the 1980s through the 2000s, although others have not (de Laffolie et al., 2012; O’Donoghue et al., 1993; Pedersen et al., 2008; Sommerfield et al., 2008). A US study reported a decline in prevalence in New York State from 1983–1990 (Applegate & Druschel, 1995) and a more recent study using data from several US population-based birth defect surveillance programs showed rather stable prevalence from 1999–2007 (St Louis et al., 2017).

The recurrence of IHPS in families suggests a genetic component for this defect (reviewed in MacMahon, 2006), and analyses of infant and parental characteristics suggest a role for non-inherited factors. Specifically, studies of infant characteristics report a fourfold or higher male excess of IHPS (Applegate & Druschel, 1995; Hedback et al., 2001; Krogh et al., 2012; Lammer & Edmonds, 1987; Markel et al., 2015; Schechter et al., 1997; To, Wajja, Wales, & Langer, 2005; Vermes, Laszlo, Czeizel, & Acs, 2016; Wang, Waller, Hwang, Taylor, & Canfield, 2008) and largely positive associations with preterm birth (Krogh et al., 2012; Schechter et al., 1997; Stark, Rogers, Eberly, & Nylund, 2015; Svenningsson, Svensson, Akre, & Nordenskjold, 2014; Wang et al., 2008), but inconclusive findings for birth weight (Applegate & Druschel, 1995; Lammer & Edmonds, 1987; Schechter et al., 1997; Wang et al., 2008), multiple births (Applegate & Druschel, 1995; Markel et al., 2015; Rider, Stevenson, Rinsky, & Feldkamp, 2013; Schechter et al., 1997; Stark et al., 2015), and month or season of birth (Dodge, 1975; Kwok & Avery, 1967; Lammer & Edmonds, 1987; Schechter et al., 1997; Zamakhshary et al., 2011).

Studies of parental characteristics report varying associations with maternal (Applegate & Druschel, 1995; Krogh et al., 2012; Markel et al., 2015; Pedersen et al., 2008; Schechter et al., 1997; Svenningsson et al., 2014; Wang et al., 2008) and paternal (Archer, Langlois, Suarez, Brender, & Shanmugam, 2007; Grewal, Carmichael, Yang, & Shaw, 2012; McIntosh, Olshan, & Baird, 1995) age at delivery. Similarly, varying associations have been reported with maternal race/ethnicity with the highest risks observed among non-Hispanic White mothers and lower risks for non-Hispanic Blacks, Hispanics (Applegate & Druschel, 1995; Lammer & Edmonds, 1987; Markel et al., 2015; Schechter et al., 1997; Wang et al., 2008), and Asians (Markel et al., 2015; Schechter et al., 1997; Wang et al., 2008). Conversely, more consistent associations have been reported for maternal education at delivery with decreasing prevalence observed with increasing education (Applegate & Druschel, 1995; Markel et al., 2015; To et al., 2005; Wang et al., 2008) and for maternal parity with decreasing prevalence observed with increasing parity (Applegate & Druschel, 1995; Dodge, 1975; Krogh et al., 2012; Schechter et al., 1997; Svenningsson et al., 2014; Wang et al., 2008). Few previous studies of infant or parental characteristics examined associations stratified by IHPS phenotype (isolated vs. multiple birth defects) or by infant sex.

The differing temporal reports of IHPS prevalence and somewhat inconsistent findings from previous descriptive studies on infant and parental characteristics suggest the need for a large study among a racially/ethnically diverse population by pooling population-based data on cases enumerated using systematic surveillance methods and comparing to population-based data on live births. To accomplish this, we used retrospective data from 11 US population-based birth defect surveillance programs to conduct a comprehensive investigation of the prevalence and descriptive epidemiology of IHPS. Our findings offer an improved understanding of the prevalence and descriptive characteristics of this defect in a large, well-described US population.

Our multistate, population-based retrospective study pooled data from 11 US birth defect surveillance programs to estimate the prevalence of IHPS from 1999–2010. Prevalence for all cases during this birth period was 20.09 per 10,000 live births and ranged from 5.52 in HI to 33.28 in OK. The difference in prevalence observed across surveillance programs may reflect the population demographics in these programs. As examples, the HI surveillance program monitors a higher proportion of Asian residents than the other programs included in our analyses, whereas the OK surveillance program monitors a high proportion of non-Hispanic Whites. IHPS prevalence estimates among offspring of Asian parents have been observed to be considerably lower than those among offspring of other parental races/ethnicities (Schechter et al., 1997; Tiao, Tsai, Kuo, & Yang, 2011; Wang et al., 2008). In comparing estimates by birth year, we observed that prevalence was relatively stable through 2002, followed by statistically significant increases from 2003–2006 and then significant decreases from 2007–2010. The cPRs estimated for our descriptive analysis using data from a subset of the birth period (1999–2005) and from surveillance programs with complete information on all selected characteristics suggested a higher prevalence of IHPS among males, preterm births, multiple births, and first-born children, as well as infants born to mothers or fathers <20 years of age at delivery or whose mothers had less than a high school education at delivery. These findings tended to persist in adjusted analyses, although estimates for parental age <20 years were attenuated. Findings restricted to isolated cases and those stratified by sex were similar to findings for all cases.

Our IHPS prevalence estimates differ from another US study which reported an overall prevalence (per 10,000 live births) of 15.8 (95% CI = 15.61, 16.04) from 1999–2007, but no significant change in prevalence during this birth period (St Louis et al., 2017). Although our study and that of St Louis et al. (2017) included mostly overlapping time frames and data from multiple surveillance programs in the US, only six programs (AZ, CO, FL, GA, NY, and TX) overlapped. Differences in the racial/ethnic distributions of cases and live births included between the two studies may have contributed to the variable findings for IHPS prevalence.

Our observation of a male predominance for IHPS was consistent with previous population-based studies from the US (Applegate & Druschel, 1995; Lammer & Edmonds, 1987; Schechter et al., 1997) and European countries (Hedback et al., 2001; Krogh et al., 2012; Markel et al., 2015; Vermes et al., 2016). The inverse association we observed for birth weight <2,500 g was not directly comparable with previous studies because of differences in study methods. Conversely, the higher PRs we observed for preterm infants were consistent with previous studies that examined all IHPS cases (Krogh et al., 2012; Schechter et al., 1997) and isolated IHPS cases (Schechter et al., 1997). Even so, the seemingly paradoxical findings we observed for low birth weight and preterm birth were observed when grouping birth weight and gestational age into meaningful categories considering small and large for gestational age. Additionally, our null findings for season of birth contrast with similar previous studies that reported seasonal variation in IHPS (Kwok & Avery, 1967; Zamakhshary et al., 2011), and our finding of increased prevalence among multiple births was consistent with two other recent studies (Rider et al., 2013; Stark et al., 2015). The results of a Texas study which used data that overlapped with those used in our study generally reported similar associations for infant sex and gestational age, although the Texas study did not observe the lower prevalence in IHPS with low birth weight observed in our study, but this may be because of differences in low birth weight categorizations used between studies (Wang et al., 2008).

With regard to parental characteristics, we observed increased IHPS prevalence among infants born to mothers <20 years of age at delivery, which supported some previous studies (Markel et al., 2015; Pedersen et al., 2008). Similarly, our observation of increased prevalence among nulliparous mothers, supported some previous studies (Applegate & Druschel, 1995; Schechter et al., 1997) as did our observation of a decreased prevalence among infants born to mothers ≥35 years of age at delivery (Applegate & Druschel, 1995; Pedersen et al., 2008; Schechter et al., 1997). Additionally, the decreased prevalence observed for mothers with race/ethnicity other than non-Hispanic White was similar to those reported in three population-based studies (Applegate & Druschel, 1995; Lammer & Edmonds, 1987; Schechter et al., 1997). The association with maternal education at delivery in our study was consistent with three previous studies (Applegate & Druschel, 1995; Markel et al., 2015; To et al., 2005). The study by Wang et al. (2008) that used data which overlapped with our study tended to report similar associations between maternal characteristics (i.e., age, race/ethnicity, education, and parity) and IHPS.

Our observation of increased IHPS prevalence associated with younger paternal age was consistent with the findings of one previous study, although the results may not be directly comparable because of differences in paternal age categorization (Grewal et al., 2012). Associations with paternal age at delivery reported in a Texas study with data that overlapped those used in our study were similar for fathers ≥35 years of age, but not for those <20 years of age at delivery (Archer et al., 2007). Additionally, the decreased prevalence we observed among infants of fathers with a race/ethnicity other than non-Hispanic White is the first examination of IHPS prevalence by paternal race/ethnicity.

Our study has several strengths. It represents one of the largest population-based studies of IHPS to date in the US, examining approximately 35% of US deliveries during the birth period included in the study. Our study also provides the most recent and longest temporal examination of IHPS prevalence estimates in the US. Along with these strengths, our study included a racially/ethnically diverse population, improving our ability to generalize findings to the US population. Other strengths include use of active case finding approaches at most of the participating surveillance programs; systematic approaches for record abstraction, including data on co-occurring birth defects; and population-based birth data from the corresponding jurisdictions for each surveillance program.

A limitation of our study was that data obtained from birth certificates may not always be reliable because of misclassification of information for gestational age (Barradas et al., 2014; Dietz et al., 2014) and race/ethnicity (Mason, Nam, & Kim, 2014), and inconsistencies that may exist in how information is collected and reported on birth certificates by different states. Another limitation was the lack of IHPS data for the entire study period from three of the 11 participating surveillance programs. Also, we lacked data on selected infant and parental characteristics from some surveillance programs, as well as data on maternal behavioral and medication exposures that may contribute to IHPS, including maternal cigarette smoking during pregnancy (Krogh et al., 2012; Leite, Albieri, Kjaer, & Jensen, 2014; Markel et al., 2015; Sorensen, Norgard, Pedersen, Larsen, & Johnsen, 2002; Svenningsson et al., 2014) and maternal use of medications during pregnancy, such as any macrolide (Ludvigsson, Lundholm, Ortqvist, & Almqvist, 2016; Lund et al., 2014), erythromycin (Cooper et al., 2002; Louik, Werler, & Mitchell, 2002), azithromycin (Eberly, Eide, Thompson, & Nylund, 2015), decongestants (Yau, Mitchell, Lin, Werler, & Hernandez-Diaz, 2013), and bendectin (Eskenazi & Bracken, 1982). Nonetheless, more than 99% of data were available for each selected infant or parental characteristic, except maternal race/ethnicity, parity, and paternal race/ethnicity, for which only 80%, 75%, and 67% of data, respectively, were available.

Using multistate population-based data covering one of the largest sample sizes and longest temporal period examined for IHPS in the US, our findings suggest that the prevalence of this birth defect appears to have decreased in the US during 2007–2010. Our descriptive analyses supported several previously reported findings, contributing to an increased understanding of selected infant and parental antecedents of IHPS. The suggestion of a decrease in IHPS prevalence warrants its continued monitoring to see if this pattern continues and to examine additional risk factors for IHPS to help explain the changing prevalence of this defect.