Maternal hypertensive disorders and subtypes of hypospadias: A Dutch case‐control study

Abstract Background Hypospadias is a frequently occurring congenital anomaly in male infants, in which the opening of the urethra is located along the ventral side of the penis. Although various studies attempted to identify its causes, the aetiology of the majority of hypospadias cases remains poorly understood. Maternal hypertensive disorders are believed to be associated with hypospadias, but the results of previous studies are not consistent, especially for subtypes of hypospadias. Objectives To investigate the associations between maternal hypertensive disorders, stratified by pharmacological treatment, and the occurrence of hypospadias divided into subtypes in a large population‐based case‐control study. Methods We included 887 hypospadias cases and 1005 male controls from the AGORA data‐ and biobank. Cases and controls were born in the periods 1975‐2016 and 1990‐2011, respectively. All data were collected in the period 2004‐2018. Maternal questionnaires were used to obtain information on hypertensive disorders during pregnancy, antihypertensive medication treatment, and potential confounders. Adjusted odds ratios (aORs) with 95% confidence intervals (CIs) for the associations between hypertensive disorders and hypospadias were estimated using logistic regression. Results Hypertensive disorders were reported by 15.3% of the women in this study. Maternal hypertensive disorders in general, chronic hypertension, and gestational hypertension were not associated with hypospadias or its subtypes. Preeclampsia was associated with posterior hypospadias (aOR 3.09, 95% CI 1.49, 6.43), whether it was untreated (aOR 2.81, 95% CI 1.24, 6.38) or pharmacologically treated preeclampsia (aOR 4.96, 95% CI 1.08, 22.80). Conclusions Our findings indicate that preeclampsia is associated with posterior hypospadias, irrespective of pharmacological treatment. This result supports the hypothesis of aetiological heterogeneity among the subtypes of hypospadias, with pregnancy‐related risk factors being associated with the more severe types of hypospadias.


| BACKG ROU N D
Hypospadias is one of the most frequently occurring congenital anomalies in male infants. 1,2 Boys diagnosed with hypospadias have a displaced opening of the urethra on the ventral side of the penis, due to incomplete fusion of the urethral folds in the third and fourth months of gestation. [3][4][5] Based on the location of the meatus, subtypes of hypospadias can be defined as anterior hypospadias (glandular, coronal, and sub-coronal), middle hypospadias (distal penile, midshaft, and proximal penile), and posterior hypospadias (penoscrotal, scrotal, and perineal), 6-8 similar but not equivalent to first, second, and third-degree hypospadias and other classifications that are being used. 9 This birth defect is usually surgically repaired in the first 2 years after birth, but may still cause complications later in life, including social and sexual problems. 10,11 Although hypospadias is associated with specific genetic mutations in a small percentage of cases, the aetiology of the majority of hypospadias cases is still unknown. 7,12 Hypertensive disorders affect approximately 5% to 10% of pregnancies. 13 They are divided into four mutually exclusive groups: chronic hypertension, gestational hypertension, preeclampsia, and preeclampsia superimposed on chronic hypertension. 14 Various studies suggested possible associations between hypospadias or its subtypes and maternal hypertensive disorders and/or the use of antihypertensive medication during pregnancy. 6,8,[15][16][17][18][19][20][21][22][23][24][25][26] Some of these studies reported that hypertensive disorders as a group were associated with hypospadias. 18,21,22,25 The majority of previous studies, however, divided the hypertensive disorders into subgroups and reported that chronic hypertension, 26 gestational hypertension, 6,16,19,20,25 and/or preeclampsia 6,8,15,17,20,24,26 were associated with hypospadias. Other studies reported that neither chronic 15 nor gestational hypertension 26 was associated with hypospadias. A few studies investigated these associations in more detail by looking at antihypertensive medication use during pregnancy and subtypes of hypospadias. Untreated hypertension and late treated hypertension were found to be associated with second-and third-degree hypospadias in three studies, 18,21,22 while these subtypes of hypospadias were associated with untreated gestational hypertension and untreated preeclampsia in another. 23 In the latter study, early treated chronic hypertension was associated with first-degree hypospadias. 23 Unrelated to treatment, preeclampsia was found to be associated strongly with posterior hypospadias and only weakly with anterior and middle hypospadias. 6 As stated above, previous studies showed rather consistent results, but suggest that associations with hypospadias are dependent on the type of hypertensive disorder, maternal exposure to antihypertensive medication during pregnancy, and subtype of hypospadias. So far, however, only one study reported on a combination of all three of these factors. 23 Therefore, we explored the associations between the separate maternal hypertensive disorders, stratified by pharmacological treatment, and the occurrence of hypospadias including its subtypes in a large population-based case-control study.

| Case-control selection
For this case-control study, we used data from the AGORA (Aetiologic research into Genetic and Occupational/environmental Risk factors for Anomalies in children) data-and biobank at the Radboud university medical center in the Netherlands. The continuous data collection for AGORA started in December 2004.
During the first hospital visit, parents of children diagnosed with or treated for a birth defect are asked to participate in the AGORA data-and biobank. For specific birth defects, such as hypospadias, hospital records were used to also identify older cases for retrospective data collection. Population-based controls were

What's already known
Various studies found that maternal hypertensive disorders separately or as a group were associated with hypospadias.
A few studies explored these associations further by looking at antihypertensive medication use during pregnancy and/or the subtypes of hypospadias. Previous studies suggest that associations with hypospadias are dependent on the type of hypertensive disorder, maternal exposure to antihypertensive medication during pregnancy, and the subtypes of hypospadias.

What this study adds
This study explored the associations between maternal hypertensive disorders and hypospadias in detail, by stratifying both exposures and outcomes. The results indicate that preeclampsia is associated with posterior hypospadias. recruited through random sampling from 39 municipal registries throughout the Netherlands in 2011. The participation rates were 34% among control parents and >65% among case parents. Data have been collected using parental questionnaires, by reviewing medical records, and by collecting blood or saliva from children and parents for DNA extraction. The data collection methods for AGORA were described in detail elsewhere. 27  The case population consisted for almost 60% of cases included in two earlier studies of our group using a different control group. 6,8

| Exposures
All exposure information was collected using the AGORA questionnaire, among others containing questions on chronic illnesses, pregnancy-related complications, and medication use in the three months before conception and during pregnancy, divided into gestational weeks or months (months 1-2, months 3-4, and months 5-9).

| Outcomes
The medical records including surgery reports of the cases were reviewed for details on location and severity of the disorder.
Hypospadias was classified into three subgroups depending on the location of the meatus: anterior (glandular, coronal, and sub-coronal), middle (distal penile, midshaft, and proximal penile), and posterior (penoscrotal, scrotal, and perineal) hypospadias. Our primary outcome was any type of hypospadias, irrespective of additional birth defects, while the three subtypes of hypospadias were used as secondary outcomes.

| Statistical analysis
The maternal characteristics of cases and controls were listed as counts and percentages. Unadjusted odds ratios (OR) with 95% confidence intervals (CI) were estimated for the associations between the four primary exposures (maternal hypertensive disorders), as well as the hypertensive disorders stratified by exposure to antihypertensive medication, and hypospadias including its subtypes.
Subsequently, we estimated adjusted odds ratios (aOR) using multivariable logistic regression analyses. Only variables that changed the effect estimate of the exposure of interest by more than 10% upon removal were retained as confounders in the final model. In case of fewer than three exposed cases or controls, only unadjusted ORs with 95% Fisher's exact 95% CI were estimated.

| Missing data
The numbers of missing data for all maternal characteristics are listed in Table 1. As almost all maternal characteristics and the exposure and outcome variables had limited numbers of missing data (<5%), no imputation was performed.

| Sensitivity analyses
Three sensitivity analyses were performed using the analysis strategy outlined above. In the first analysis, we excluded cases and controls who had a first-degree relative with hypospadias to minimise information bias due to familiarity with the disorder. Secondly, we restricted the study population to cases and controls born in the period 1990-2011 with a known time lag between birth and completion of the questionnaire, in order to provide insight into the potential influence of the disparity in birth years between cases and controls. By stratifying the time lag into ≤5 years and >5 years, we tried to identify the role of misclassification due to memory effects.
The third sensitivity analysis comprised the exclusion of cases and controls with a preterm birth to minimise the effect of early-onset preeclampsia (<34 weeks of gestation), which usually results in preterm delivery.
The data were analysed using SPSS Statistics version 25 (IBM Corp).

| Ethics approval
The initial data collection protocol for the AGORA data-and biobank and several updates were approved by the Regional Committee on Research Involving Human Subjects Arnhem-Nijmegen.

| RE SULTS
We extracted 960 cases and 1050 male controls from the AGORA data-and biobank. Due to multiple gestations, 71 cases and 42 controls were excluded. Furthermore, we excluded two cases and three controls because of pre-existing diabetes of the mother. As a result, 887 cases and 1005 controls were included in the analyses. Our patient population contained 366 new cases as well as 161 and 360 cases used in two previous studies. 6,8 Table 1 summarises the maternal characteristics of the cases and controls included. Patient mothers were younger at delivery and more likely to be primiparous and overweight or obese compared to control mothers. Mothers of controls slightly more often used oral contraceptives and folic acid supplements in the periconceptional period. The proportions of women who drank alcoholic beverages or smoked in that period were similar in the two groups. For cases, a family history of hypospadias was reported much more often than for control children.
In total, 150 (16.9%) patient mothers and 139 (13.8%) control mothers were diagnosed with one of the four hypertensive disorders of pregnancy. Table 2 shows the frequencies of the specific hypertensive disorders and the use of antihypertensive medication  Unadjusted and adjusted ORs for the associations between hypospadias and the four hypertensive disorders, stratified by antihypertensive medication exposure, are also provided in Table 2.
All aORs were adjusted for parity, which proved to be the only true confounder in this study. Neither chronic nor gestational hypertension was associated with hypospadias, regardless of pharmacological treatment. Preeclampsia seemed to be weakly associated with hypo-   1.54, 10.05)) were somewhat higher than in Table 3, while the aOR for pharmacologically treated preeclampsia could not be estimates due to small numbers.

| Principal findings
In this population-based case-control study, we explored the associations between maternal hypertensive disorders and hypospadias in detail, by stratifying both exposures and outcomes. We observed clear associations between preeclampsia and posterior hypospadias for both treated and untreated preeclampsia. No other meaningful associations were found between maternal hypertension in general or specific hypertensive disorders and hypospadias or its subtypes.

| Strengths of the study
A major strength of this study was the ability to study stratified exposures and outcomes by subtypes due to the use of a large and well-established population-based data collection. The detailed clinical data available in the AGORA data-and biobank, collected from TA B L E 3 Associations between the three subtypes of hypospadias and maternal hypertensive disorders stratified by treatment with antihypertensive medication during pregnancy medical files including surgical reports, assured proper diagnoses of hypospadias with clear differentiation between the hypospadias subtypes. The extensive questionnaires used during data collection enabled us to assess the presence or absence of the distinctive hypertensive disorders and the treatment with antihypertensive medication for each mother. In a validation study of the AGORA questionnaire used for prescription medication, the sensitivity for antihypertensive medication proved to be 0.83 while the specificity was 1.00. 28 In addition, the questionnaire data were used to gather information on potential confounders, which are usually not available from medical records. The sensitivity analyses performed did not point towards large amounts of selection bias or information bias in the results from the primary analyses, despite the low participation rate among control parents.

| Limitations of the data
The self-reported nature of the exposure and confounder informa-  29 Therefore, some misclassification of exposure seems likely but is presumably nondifferential, since case mothers would not have associated these hypertensive disorders with hypospadias in their sons. The same may apply to the lifestyle factors included in the study, as these were quite equally distributed among cases and controls and did not play a role as confounders. If differential recall would have occurred in any of these factors, however, confounding could have been obscured. Still, we believe our main results to be valid, especially since preeclampsia is a clearly defined condition that is not easily misclassified, and neither is the confounder parity.
The prevalence rates of preeclampsia in this study were also in the same order of magnitude as in two very large studies. 23,26 Due to small numbers and lack of information, we were not able to stratify preeclampsia by severity, estimate the associations between preeclampsia superimposed on chronic hypertension and hypospadias, use the information on timing and type of medication, or fully examine the potential confounders in the stratified analyses. For some of the same reasons, we could not estimate the associations between chronic hypertension and specific subtypes of hypospadias.

| Interpretation
In the study by Brouwers et al, 6 25 Several others reported associations between gestational hypertension and hypospadias in general, 6,16,19,20,23,25 whereas our study and the cohort study of Arendt et al in Denmark that also used mutually exclusive groups of hypertensive disorders, showed no associations. 26 Similar to two previous studies, 16,25 we did not find associations between chronic hypertension and hypospadias either, but two other larger studies did report this association. 23,26 Preeclampsia was clearly associated with hypospadias in general in several studies, 6,8,15,17,20,23,24,26 whereas we observed only a weak association between preeclampsia and hypospadias as a group. These variations in findings may partly be due to differences in the definitions and demarcations of the hypertensive disorders, in the mix of subtypes of hypospadias, and in treatment regimens among the various study populations. Timing of onset of preeclampsia may also play a role, as our results point towards an increased risk of hypospadias associated with early-onset and probably more severe preeclampsia.
The results of this study cannot be explained by a contributing role for diabetes as was recently found in a study on several other birth defects. 30 Women with pre-existing diabetes were excluded from the study, and only a limited number of case mothers (n = 4) and control mothers (n = 5) reported gestational diabetes in the aetiologically relevant period. Other birth defects (cryptorchidism and kidney defects) were only observed in two of the 12 cases with posterior hypospadias and a mother with preeclampsia. Our results for preeclampsia underline the hypothesis of aetiological heterogeneity among the subtypes of hypospadias that was suggested in our previous studies, in which primiparity, multiple pregnancy, preterm birth, low birthweight, and being small for gestational age were also associated with posterior hypospadias. As the male external genitalia develop between the 8th and 14th weeks of gestation, however, the development of hypospadias cannot be influenced by factors that occur later in pregnancy, such as preeclampsia and treatment with antihypertensive medication. Therefore, a shared risk factor, that is placental dysfunction in early pregnancy, may be the explanation for the associations between preeclampsia and the other above-mentioned factors and posterior hypospadias. 6,8 During development of the male genitalia, the placental production of human chorionic gonadotropin (HCG) is crucial, as HCG is responsible for regulating the production of testosterone in the Leydig cells. [31][32][33][34] Insufficient production of HCG due to placental dysfunction could lead to low testosterone levels resulting in abnormal development of the male external genitalia.

| CON CLUS IONS
We observed that both pharmacologically treated and untreated preeclampsia were associated with posterior hypospadias, with a higher risk estimate for treated early-onset preeclampsia. Our results underscore the hypothesis of aetiological heterogeneity among the subtypes of hypospadias, in which severe forms of hypospadias are associated with pregnancy-related risk factors. Our findings also suggest that preeclampsia and posterior hypospadias may have a shared risk factor in placental insufficiency.

ACK N OWLED G EM ENTS
We would like to thank all participants in the AGORA data-and biobank included in this study, as well as the staff members and nurse practitioners of the Department of Urology at the Radboud university medical center for their collaboration in the data collection.