Association between history of hookah use and symptoms of postpartum depression: A population‐based study

Although several biologic, psychosocial, and behavioral factors have been linked to postpartum depressive symptoms, studies examining the association between non‐cigarette tobacco products and symptoms of postpartum depression are currently lacking. This study examined the association between hookah use and postpartum depressive symptoms.


| INTRODUC TI ON
Postpartum depression (PPD) is one of the most common complications of childbirth. 1 A recent meta-analysis showed that the prevalence of PPD globally is 17.7%. 2PPD typically occurs 4-6 weeks after childbirth and can have adverse consequences for mothers, infants, partners, and family members if untreated. 3Women who suffer from PPD may experience poor psychological well-being and quality of life. 3PPD can create an unfavorable environment that hampers child development and growth due to reduced maternalinfant bonding. 4veral biologic, psychosocial, and behavioral factors have been linked to PPD.A recent review identified risk factors for PPD, including high life stress, lack of social support, current or past abuse, prenatal depression, and marital or partner dissatisfaction. 5Other studies have identified substance use and smoking, 6 stressful life events, 7 and a combination of multiple genetic and environmental factors contributing to the development of PPD. 8 Recent studies have shown that prepregnancy, 9 prenatal, 10,11 and postpartum smoking is associated with PPD.Other health outcomes associated with smoking during pregnancy include an increase in the risk for deep vein thrombosis, cardiovascular events, pulmonary complications, and gastrointestinal ulcers, among many other negative health outcomes for pregnant women. 12 the USA, hookah use has increased, particularly among young adults.Data from the Tobacco Products and Risk Perceptions Surveys (2014-2015) showed that 15.8% of US adults have smoked using a hookah. 13Among women of childbearing age who participated in the 2013-2014 Population Assessment of Tobacco and Health national survey, 6.5% reported using a hookah, whereas only 2.5% of the 388 pregnant women who participated in the survey smoked using a hookah during pregnancy. 14Interestingly, hookah quit rates were significantly higher among pregnant women, with 98.3% of them quitting by the next round of surveys administered in 2014-2015.It is increasingly concerning that most hookah users perceive hookahs to be less harmful to their health than cigarettes. 15e study has shown that a hookah delivers similar toxins and causes the same subjective effects as smoking cigarettes, although the physiological effects and amount of smoke inhaled are higher in magnitude than cigarette smoking. 16Although a hookah produces heavy metals and higher amounts of carbon monoxide 17 than a regular cigarette, pregnant women perceive hookah smoking as less harmful than cigarette smoking. 18spite well-established evidence on adverse maternal and infant health outcomes associated with tobacco use, few studies have investigated pregnancy-related health outcomes related to hookah use. 19These studies have found positive associations between hookah use and infant mortality, 20 preterm birth, 21 and low birth weight. 21Another study has demonstrated significant associations between hookah use and early weaning from breastfeeding. 22To our knowledge, no prior studies have examined the association between hookah use and maternal mental health, including PPD.We examined the association between a history of hookah use and PPD symptoms using population-based data.We hypothesized that hookah use would be associated with a higher likelihood of having PPD symptoms.

| Study design and population
This cross-sectional study used data from the most recent phase of the Pregnancy Risk Assessment and Monitoring System (PRAMS) 2016-2020. 23In collaboration with state health departments, the US Centers for Disease Control and Prevention (CDC) conducts PRAMS to better understand maternal behaviors, attitudes, and experiences before, during, and shortly after pregnancy.A detailed description of the PRAMS study design and methods are available elsewhere. 24In brief, the PRAMS survey contains data from participants randomly selected from state birth registries in 51 PRAMS sites covering 47 states, with each participating state contributing between 1000 and 3000 participants. 24To improve representativeness, samples are stratified by specific characteristics to ensure adequate data for smaller, high-risk subgroups.There were no significant differences in hookah use in the last 2 years.However, participants with missing data were more likely to report symptoms of PPD (Table S1).

| Outcome of interest: symptoms of PPD
We used the symptoms of the PPD indicator variable in the PRAMS dataset to determine the symptoms of PPD.In PRAMS, this variable is computed based on two questions that PRAMS has widely used to assess the presence of symptoms of PPD status: "Since your new baby was born, how often have you felt down, depressed, or hopeless?"and "Since your new baby was born, how often have you had little interest or little pleasure in doing things you usually enjoyed?"Participants had options to select one of the following responses: "Always", "Often", "Sometimes", "Rarely", or "Never".Participants that answered "Always" or "Often" for either question were considered as having symptoms of PPD.Consistent with the PRAMS coding schemes, we dichotomized this variable as yes/no.

| Independent variable: Hookah use in the past 2 years
Participants were first told the following definitions before being asked about their hookah use in the past 2 years: "A hookah is a water pipe used to smoke tobacco, and it is not the same as an ehookah or hookah pen".Participants were then asked: "Have you used a hookah in the past 2 years?"This variable was coded as a dichotomous variable (yes/no).

| Covariates
Covariate selection was based on a review of existing literature.

| Statistical analysis
Frequencies and percentages were used to describe the study sample.Differences in the proportions of symptoms of PPD by hookah use in the past 2 years and each covariate were assessed using Rao-Scott χ 2 tests.Multivariable logistic regression analysis was then used to determine the independent association between hookah use in the past 2 years and symptoms of PPD.Variables with P less than 0.05 were retained in the multivariable regression.The potential presence of collinearity was checked using the variance inflation factor greater than 4, and no collinearity was detected.Odds ratios and 95% confidence intervals for the odds ratios were reported.The regression model included pairwise interactions between hookah use in the last 2 years and covariates.Two-sided P-values less than 0.05 were considered statistically significant.Complex survey design elements of PRAMS were applied in all analyses as recommended by the CDC PRAMS.All analyses were performed using SAS 9.4 (SAS Institute Inc., Cary, NC, USA).

| RE SULTS
Of the 106 894 participants included in the study sample, a greater proportion of participants were between the ages of 25 and 34 years old (65 094; 61.1%), were married (69 916; 67.8%), were non-Hispanic white (53 702; 60.9%), had greater than a high school level of education (74 549; 70.1%), and did not receive WIC food assistance during pregnancy (71 968; 70.6%).Approximately 9643 (8.2%) participants had symptoms of PPD and 4034 (4.1%) women reported using a hookah in the last 2 years.Additional characteristics of the study sample are shown in Table 1.
In the bivariate analysis, compared with those without PPD symptoms, participants with PPD symptoms were more likely to be hookah users (5.5% vs 4.0%, P < 0.001).Additionally, participants reporting PPD symptoms were more likely to be younger, unmarried, identify as non-Hispanic black, or other race/ethnicity, have a high school education or less than a high school education, have Medicaid or no health insurance, received WIC food assistance during their pregnancy, were underweight or obese, had gestational weight gain below the IOM guideline, had inadequate prenatal care, had an unintended (mistimed, unwanted, unsure) pregnancy, had a cesarean delivery, had no postnatal check-up, were postnatal smokers, never had alcohol before pregnancy, experienced abuse during pregnancy by a husband or partner, did not breastfeed, had a low birth weight infant and reported infantmother room-sharing (Table 1).
Table 2 presents factors associated with symptoms of PPD.
After adjusting for covariates, the odds of having PPD symptoms were 20% higher among participants who reported hookah use in the last 2 years compared with non-users (adjusted odds ratio

| DISCUSS ION
This cross-sectional study found that after adjusting for covariates, the odds of having symptoms of PPD were higher among women who used a hookah in the past 2 years.Previous studies have found a relationship between perinatal smoking and PPD symptoms, [26][27][28][29] but those studies did not exclusively focus on hookah smoking and PPD symptoms.Our study also found that pregnant women who were young (<35 years), self-reported non-Hispanic black or Hispanic, belonged to low socioeconomic strata, had endured intimate partner violence, and had adverse birth outcomes, such as an infant with a low birth weight, had higher odds of having PPD symptoms.
Many studies have found an association between smoking and depression, but the causality as well as the direction of causality have not been established. 30Some studies found that smoking could elevate mood due to nicotine's anxiolytic effects and helping in the adaptation to stressful conditions.However, the brief period of mood elevation is offset by large diurnal variations in smokers compared with non-smokers. 31r study also found that PPD symptoms were most likely to occur among mothers from historically marginalized racial/ethnic groups, participants who identify as non-Hispanic black and Hispanic, and other race/ethnicity.This finding is supported by previous research findings where smoking was associated with PPD symptoms among non-Hispanic black and other race/ethnicity. 32previous study using PRAMS data found that prenatal smoking has a 1.6 times higher odds of PPD symptoms among mothers who experience poverty or low socioeconomic status and received WIC during pregnancy.32 Our findings underscore an important birth equity issue.
Having an infant with a low birth weight was positively associated with PPD symptoms and similar findings were reported in a previous study. 33Although a healthy newborn sparks a joyous occasion, a newborn in need of medical attention can lead the mother to PPD symptoms. 34Room-sharing with an infant also increased the odds of having PPD symptoms in mothers.Bed-sharing with an infant might disrupt the mother's sleep and thereby increase the likelihood of having depressive symptoms.
Having an unintended pregnancy was also linked with a higher likelihood of having PPD symptoms.Previous research also found a similar association; the unintentional pregnancy is further exacerbated by other factors, including poverty and experiencing intimate partner violence, 35 which were also found to have a significant association with having PPD symptoms in our study.
Therefore, states that pass laws restricting abortion must, at the very least, consider how unwanted pregnancies affect those who are at risk of having PPD symptoms and establish measures to protect them. 36r study had several limitations.Due to the cross-sectional nature of the study, recall bias could not be avoided and causality cannot be determined.Stigma related to mental health and social desirability may have introduced measurement error and misclassification bias.
Additionally, in the current study, PPD symptoms were not clinically diagnosed and were self-reported using a structured questionnaire, which may have introduced misclassification bias.However, the questions have widely been developed and used by the CDC and were previously pretested and closely resemble validated clinical screening tools. 37Future studies should therefore employ the American Psychiatric Association's Diagnostic and Statistical Manual of Mental Disorders (DSM-V) definition of PPD as the onset of a major depressive episode during pregnancy or within 4 weeks of delivery. 38Additionally, the association between a history of hookah use and PPD symptoms may be dose dependent.However, data on the frequency of hookah use in the past 2 years were not captured in the PRAMS data.Finally, To the best of our knowledge, this is the first study to examine the association between hookah use and PPD symptoms in a US-based cohort.We used population-based data pooled across 5 years, which ensured a large sample size of 106 894.Therefore, our findings can be considered generalizable.The study also included only women with no prior history of prepregnancy or prenatal depression, which ruled out the effect of a prior history of depressive symptoms.The availability of several sociodemographic and behavioral variables in PRAMS also allowed us to adjust for possible confounders, strengthening the study.
Although the harmful effects of traditional cigarettes have been well-established, the hookah has been regarded as less addictive and easy to quit, especially among young adults. 39Therefore, hookah use has been found even among non-smokers during pregnancy. 40Pregnant women also preferred the flavored hookah products; 41 the flavor may increase the appeal to help with nausea during pregnancy and reduce the harm perception of hookah smoking. 42In this context, a targeted approach should be taken to build awareness of the harmfulness of hookah smoking during pregnancy.Pregnancy has been regarded as a "teachable moment" 43 and smoking cessation is hard, even during pregnancy.
Eligible participants are recruited primarily by mail, and multiple attempts, including telephone calls, are made to contact women in the sample.Between 2016 and 2020, 206 080 women participated in the survey.The PRAMS study protocol was reviewed and approved by the CDC Institutional Review Board and the institutional review boards of the local health departments participating in the state PRAMS projects.An informed consent document is included within each survey packet explaining the participant's rights, and consent is implied if the survey is completed and returned.For interviews completed by telephone, the informed consent document is read verbally on the telephone, and verbal consent is obtained from the participant prior to proceeding with the survey.De-identified PRAMS data are available to researchers upon request from the CDC.The Office of Research Compliance at Ohio University has reviewed the protocol for the current study and determined that it does not meet the government definition of human subject research, in accordance with 45 CFR 46.102 (l) and (f).The current study was restricted to women who had given birth to a singleton infant who was alive at the time of the survey and had no prior history of prepregnancy or prenatal depression (n = 155 480).Participants with missing data on PPD symptoms (n = 3962), hookah use in the past 2 years (n = 2366), and other covariates adjusted in the multivariable model (n = 42 258) were excluded from the study.The final study sample consisted of 106 894 participants.To examine the potential impact of missing data on the interpretation of study findings, we compared the characteristics of participants with and without complete data.

1. 20 ;
95% confidence interval 1.03-1.40;P = 0.022).Other factors significantly associated with higher odds of PPD symptoms include being younger than 35 years old, being non-Hispanic black, TA B L E 1 Descriptive statistics of the study sample (n = 106 894).

Overall Symptoms of postpartum depression n (wt%) No Yes P
TA B L E 1 (Continued)Hispanic or other race/ethnicity, having a high school education, receiving WIC assistance during pregnancy, having an unintended pregnancy, having a cesarean delivery, not receiving a postnatal care check-up, being abused during pregnancy by a husband or partner, having a low birth weight infant, and reporting infantmother room-sharing.
Institute of Medicine; WIC, Special Supplemental Nutrition Program for Women, Infants, and Children; wt%, weighted percent.a Calculated as weight in kilograms divided by the square of height in meters.Adjusted odds of symptoms of postpartum depression by hookah use in the last 2 years (n = 106 894).
TA B L E 1 (Continued) TA B L E 2

33 Adjusted odds ratio (95% confidence interval) P
Institute of Medicine; WIC, Special Supplemental Nutrition Program for Women, Infants, and Children.a Calculated as weight in kilograms divided by the square of height in meters.
TA B L E 2 (Continued)