This study draws on data from the 2011–2018 NHANES, a cross-sectional survey intended to evaluate the health and nutritional status of the civilian, non-institutionalized U.S. population. The survey employs a complex, multistage probability sampling design to ensure representativeness of the U.S. population. Given the extensive scope of NHANES data, including detailed urinary biomarker measurements of phthalate metabolites, it provides a unique opportunity to investigate the potential association between phthalate exposure and OAB on a broad scale.

The study cohort comprised adults aged 20 years and above who participated in NHANES between 2011 and 2018 and had complete data on urinary phthalate metabolites and OAB assessments. Based on this, individuals who were pregnant or had a documented history of known diseases affecting bladder function were first excluded. Ultimately, 6,228 participants were included in Models 1 and 2. Then, individuals with missing covariate data were excluded, resulting in a final inclusion of 4,451 participants in Model 3 (Figure 1).

Isotope dilution high-performance liquid chromatography-tandem mass spectrometry (HPLC-MS/MS) was employed for the quantitative analysis of urinary phthalate metabolites. Urine was processed using β-Glucuronidase for the digestion of glucurnide-conjugated metabolites, followed by online solid-phase extraction (SPE) coupled with reverse-phase HPLC-ESI-MS/MS. Detailed experimental methods can be found in the NHANES 2017–2018 Laboratory Methods. Urine samples were collected from participants and analyzed for a range of phthalate metabolites, including high-molecular-weight (HMW) phthalates such as mono(carboxynonyl) phthalate (MCNP), mono(carboxyoctyl) phthalate (MCOP), mono-2-ethyl-5-carboxypentyl phthalate (MECP), mono(3-carboxypropyl) phthalate (MCPP), and monobenzyl phthalate (MBzP). Low-molecular-weight (LMW) phthalates, such as monoethyl phthalate (MEP), monobutyl phthalate (MBP), and mono-isobutyl phthalate (MiBP), were also included. The selection of these metabolites was based on their prevalence in the general population and their potential health implications (17, 18). Urinary creatinine levels were concurrently measured to adjust for urine dilution in the analyses, detailed experimental methods can be found in the NHANES 2017–2018 Laboratory Methods. Quality control procedures involved the use of blanks, calibration standards, and quality control samples with known concentrations of phthalate metabolites. All laboratory analyses were conducted by the Laboratory Sciences Division of the Centers for Disease Control and Prevention (CDC).

The Overactive Bladder Symptom Score (OABSS) is a standardized tool developed to quantify the severity of symptoms in patients with OAB (19). Urinary symptom questionnaire data from the NHANES survey were translated into OABSS scores to determine the presence of OAB in study participants (20, 21) (Table 1). Participants were asked: how often have you experienced urinary leakage? A score of 0 on the OABSS corresponded to “Never,” while “Less than once a month” and “A few times a month” were both scored as 1. “A few times a week” was assigned a score of 2, and “Every day or night” was scored as 3. Nocturia was evaluated in a similar manner, with scores assigned based on the frequency of nocturnal awakenings to urinate. “0 times” corresponded to an OABSS score of 0, “1 time” to a score of 1, and “2 times” to a score of 2. The total OABSS was derived by summing the scores across these categories, providing a quantitative measure of OAB symptom severity for each participant. Individuals with a total score of ≥3 were classified as having OAB (22). This scoring approach provides a standardized evaluation of OAB within the study cohort, thereby enhancing comparability and supporting rigorous statistical analysis. In addition, we performed a sensitivity analysis using an OABSS threshold of ≥5 to assess the robustness of our main findings.

Our analysis included a range of covariates to account for potential confounders. The selection of these covariates was guided by their established or hypothesized relationships with phthalate exposure and OAB (23, 24). The covariates considered in the analysis included age (years), gender (male or female), race/ethnicity, body mass index (BMI, kg/m²), educational attainment, socioeconomic status, smoking status (current, former, never), alcohol consumption (light, moderate, or heavy), hypertension, diabetes, hyperlipidemia, and a history of cardiovascular disease. BMI, hypertension, diabetes, and hyperlipidemia were evaluated using physical examination data, while the other covariates were obtained from self-reported questionnaire data.

All statistical analyses were conducted using Stata 17 and R version 4.3.1. Survey design variables and sample weights were integrated into all analyses, with MEC weights adjusted by a factor of 0.25 to yield final weights for the four survey cycles spanning 2011 to 2018. MEC weight refers to the weight assigned to survey participants based on the Mobile Examination Center data. Descriptive statistics were utilized to summarize the demographic characteristics of the study cohort. Categorical variables were represented using weighted proportions and 95% confidence intervals (95% CI), while continuous variables were presented using weighted means and standard errors (SE). We used SPSS to conduct Pearson chi-squared tests for categorical variables and t-tests for continuous variables to assess group differences. The association between urinary phthalate metabolite concentrations and OAB was assessed using multivariate logistic regression models, with subgroup analyses performed to further investigate the findings.

To investigate the potential linear or non-linear associations between urinary phthalate concentrations and OAB, a restricted cubic spline (RCS) analysis was employed. The RCS method enables flexible modeling of non-linear relationships without requiring the imposition of a specific functional form. The results were subsequently visualized to offer a comprehensive understanding of the dose-response relationship (25).

We applied a Bayesian Kernel Machine Regression (BKMR) model to evaluate the potential interactions among various phthalate metabolites and their collective impact on OAB. The BKMR approach is particularly effective for studies involving multiple exposure mixtures, as it enables the assessment of compound interactions and their combined effects on OAB risk (26).

The baseline characteristics of the study participants are detailed in Table 2. The study included 4,451 participants, with an average age of 46.46 years (SE = 0.33). Compared to participants without OAB, those with OAB were more likely to be older, predominantly female, and have a higher BMI (P < 0.001). Compared to the non-OAB population, the proportion of individuals with hypertension, diabetes, and cardiovascular diseases is higher in the OAB population (P < 0.001), while there is no difference in the proportion of high cholesterol between the two groups (P > 0.05). Furthermore, compared to non-OAB participants, individuals with OAB exhibited elevated average concentrations of specific phthalate metabolites, such as MBzP and MiBP (9.12 vs. 7.94; 15.67 vs. 13.53; P < 0.05), while the concentration of MCOP was found to be lower in participants with OAB (27.29 vs. 37.93; P < 0.001). Compared to the non-OAB population, the proportion of smokers is higher and the proportion of drinkers is lower in the OAB population (P < 0.001).

The association between urinary phthalate concentrations and the risk of OAB was assessed using multivariable logistic regression models (Table 3). After adjusting for potential confounders, we observed that when phthalate concentrations were treated as continuous variables, higher concentrations of HMW phthalate MBzP were associated with an increased risk of OAB. This association was more pronounced in the highest tertile range when phthalate concentrations were treated as categorical variables (OR = 1.401, 95% CI: 1.108, 1.771). Likewise, an increase in LMW phthalate MiBP levels was correlated with a higher OAB risk, especially in the highest tertile (OR = 1.050, 95% CI: 1.045, 1.056). When using the OABSS threshold of ≥5 to define OAB, MBzP, and MiBP still showed an association with OAB in model 3, after adjusting for all covariates, especially in the higher percentiles (OR = 1.509, 95% CI: 1.041, 2.186; OR = 1.555, 95% CI: 1.088, 2.224). The results of this sensitivity analysis are provided in the Supplementary Table 1.

To explore whether the association between phthalate exposure and OAB varies across age, sex, and other covariates, we performed subgroup analyses of MBzP and MiBP (Figure 2). The results indicate that no significant interaction was found across all subgroups (P > 0.05), suggesting that the effects of MBzP and MiBP concentrations on OAB are consistent among these subgroups, and subgroup characteristics did not have a significant moderating effect on the associations between MBzP and OAB, as well as MiBP and OAB.

We conducted a RCS analysis to assess the potential non-linear relationship between urinary phthalate concentrations and OAB risk (Figure 3). The results indicated a positive correlation between MBzP and MiBP with OAB (P for overall < 0.05), with the relationship being linear (P for non-linear > 0.05).

BKMR analysis revealed an overall positive impact of phthalate mixtures on OAB risk, indicating that higher levels of combined exposure are associated with an increased risk of OAB. The results of the BKMR model indicate that MBzP has the highest probability of inclusion (P = 1.00), suggesting that MBzP may be the primary factor contributing to this effect. The joint and individual effects of phthalate metabolites on OAB risk as revealed by the BKMR analysis are presented in Figures 4A, B. Furthermore, the analysis explored potential interactions between different phthalate exposures, but no significant interactions were found (Figure 4C).