1. Introduction

Polypharmacy has become a major issue in pregnant woman because of the increasing trend of medication use during pregnancy.[1,2] The trend of polypharmacy has progressively increased in the past 4 decades among pregnant women.[2] In the US, the average number of medications used during pregnancy increased from 2.6 to 4.2 prescriptions from 1976 to 2008.[2] Evidence showed that about 97% of pregnant women used at least 1 medication during their pregnancy, and about 30% of pregnant women reported taking at least 5 medications.[1] Moreover, a great proportion of pregnant women also reported taking over-the-counter drugs or herbal medicines during their pregnancy.[3]

Information regarding medication utilization among pregnant women is limited since pregnant women are rarely included in clinical trials during drug development.[4] However, medication use is sometimes mandatory in certain conditions during pregnancy. For example, a pregnant woman may have a preexisting medical condition that requires medication treatment, such as diabetes mellitus (DM), hypertension, mood disorders, epilepsy, asthma, and rheumatic diseases.[5] Furthermore, obstetric complications such as gestational diabetes, preeclampsia, and cholestasis of pregnancy may also need to be managed by medications.[5]

There are several concerns with polypharmacy in pregnant women. First, the risk of drug-drug interactions increases during polypharmacy.[6] Drug-drug interactions affect the efficacy and safety of medication use. In consequence, polypharmacy may create potential teratogenic effects that clinicians are unaware of.[7] Second, medication adherence is harder to manage in polypharmacy.[8]Pregnant women who take multiple drugs at a time have increased chances of skipping doses or doubling dosages.[9] Third, the pharmacokinetics and pharmacodynamic effects of medications may change because of the pregnancy.[10] Therapeutic effects become less predictable during pregnancy; thus, it is even more complicated in pregnant women with polypharmacy.

Previous studies primarily investigated polypharmacy issues in populations other than pregnant women.[11–13] Although some studies showed a trend of polypharmacy in pregnant women, they were either not nationally representative or focused on specific conditions such as psychiatric illnesses, epilepsy, and migraines.[14–17] Therefore, findings from previous studies cannot provide a comprehensive overview of polypharmacy in pregnant women from a US nationally representative perspective.

Given the increased trend of medication use and risks of medication-related complications during pregnancy, it is important to understand the trends and patterns of prescription medication use among pregnant women. The National Health and Nutrition Examination Survey (NHANES) is the most comprehensive survey conducted to consistently collect health and medication-related data in the US population. Our study aimed to use the NHANES to investigate the prevalence and trends of polypharmacy in pregnant women aged 15 to 44 years in the US. Moreover, patterns of prescription medication use and the characteristics of pregnant women with polypharmacy were also demonstrated.

2. Materials and methods 2.1. Data source

We retrieved data from the NHANES database to conduct this cross-sectional study. The NHANES is an ongoing population-based survey designed to provide health and nutrition information of the US noninstitutionalized population.[18] It is a major program conducted by the Centers for Disease Control and Prevention, and is released at 2-year intervals.[18] This large cross-sectional survey contains 5 major domains of data: sociodemographic, dietary, examination, laboratory, and a questionnaire.[18] All of these data were collected through a household interview or medical examination. Household interviews were performed by trained interviewers to collect demographic data and socioeconomic data such as race/ethnicity, household income, marital status, and educational level, and the medical examination was administered by Mobile Examination Centers to collect medical, dental, and physiological measurements, and laboratory tests by highly trained medical personnel.[18] The NHANES protocol was reviewed and approved by the National Center for Health Statistics Research Ethics Review Board.[18]

We used data from 9 NHANES cycles between 1999 and 2016 to conduct this study. The average interviewed response rate of these 9 cycles was 70.1%, and the average examined response rate was 76.1%.[18]

2.2. Study population

We included participants of a childbearing age (15–44 years) who were pregnant and women who were not pregnant. We identified pregnant women based on an answer to the question on the pregnancy status in NHANES questionnaire data.[18]

2.3. Outcome assessments

Polypharmacy was defined as using more than 1 prescription medication during the pregnancy. During the survey, respondents were asked to present medications they were taking. Then, the trained interviewers recorded participants prescription medication use including the active components, dosage form, and route of administration.[18] The utilization of compound medicines made from 2 or more active ingredients was also defined as polypharmacy.

2.4. Covariates

Our study adjusted for covariates, including sociodemographic factors and health-related factors. Sociodemographic factors included age, race/ethnicity, poverty income ratio, educational level, and marital status. Age was divided into 3 groups: <25, 25 to 34, and ≥ 35 years. Our study population consisted of non-Hispanic white, Mexican Americans, non-Hispanic black, other Hispanics, and other races (including multiracial). The poverty income ratio was categorized into 3 groups: <130, 131 to 349, and ≥ 350. In NHANES, the poverty income ratio is obtained by dividing household income by the Department of Health and Human Services poverty guidelines to generate the ratio.[18] Educational level was stratified into 2 groups: some college or higher, and high school graduate or lower. Marital status was divided into 2 groups: not married (single/divorced/widowed/separated) and married or living with a partner. Health-related factors included the self-reported general health condition, hospitalization within the past 1 year, and chronic conditions, including high blood pressure, asthma, arthritis, and DM. All of the covariates were retrieved from NHANES demographic and questionnaire data.

2.5. Statistical analysis

A weighted estimation was used during the analytical procedure to provide a representative number for the noninstitutionalized US civilian population. We calculated the mean prevalence of polypharmacy of each cycle to observe the trend of polypharmacy in pregnant women, and provided the average prevalence of polypharmacy in the NHANES from 1999 to 2016. A join point analysis was used to investigate the significance level of the prevalence trends and identify the year at which significant changes in prevalence occurred during the study period. [19,20] A joinpoint regression model was conducted using the National Cancer Institute’s Join point Regression Program vers. 4.9.0.0.[21] In addition, the annual average percentage change is also presented to estimate the degree of change in the prevalence.

Regarding the univariate analysis, the Rao-Scott Chi-squared test of independence was performed to demonstrate differences in categorical characteristics between pregnant women with and those without polypharmacy. Odds ratios (ORs) with 95% confidence intervals (CIs) are reported in our study. We further executed a multivariable logistics regression model to evaluate associations of demographics and health-related characteristics with polypharmacy in pregnant women. All of these data analyses were conducted using SAS software (vers. 9.4, SAS Institute, Cary, NC). Our study was approved by the Joint Institutional Review Board of Taipei Medical University (TMU-JIRB) (approval no.: N202109058).

3. Results 3.1. Polypharmacy in pregnant women

In total, 1591 participants obtained from the NHANES in 1999 to 2016 were enrolled in our study. The number of the participants represented 3,350,983 pregnant women in the US after weighting. Among them, 247,525 (7.4%) were determined to be experiencing polypharmacy.

Figure 1 shows the proportions of prescription medications used by pregnant and nonpregnant women in the US. Among pregnant women, an estimated 25.0% reported that they had used at least 1 prescription in the past 30 days prior to the interview. About 47% of nonpregnant women reported they had used a prescription in the past 30 days. Furthermore, among pregnant women, 17.7% had taken 1 prescription medication, 3.8% had taken 2 prescription medications, and 3.6% had taken 3 or more different prescription medications during their pregnancy. Among women who were not pregnant, 23.4% had taken 1 prescription medication, 11.2% had taken 2 prescription medications, and 12.3% had taken 3 or more different prescription medications in the past month. Overall, pregnant women were less likely to be experiencing polypharmacy than were nonpregnant women (7.4% vs 23.5%, P < .01).

Figure 1.:

Numbers of prescription medications used by pregnant women and non-pregnant women. * There were 3,350,983 pregnant women and 51,960,165 nonpregnant women identified in this study. ** The number of polypharmacy among pregnant women was 247,525 (7.4%) and among nonpregnant women was 12,197,758 (23.5%). *** Pregnant women were less likely to have polypharmacy compared to nonpregnant women (7.4% vs 23.5%, P < .01).

Table 1 lists the most frequently taken prescription medications among pregnant and nonpregnant women. Levothyroxine (6.7%) was the most commonly taken prescription medication among pregnant women, followed by albuterol (5.4%) and montelukast (5.4%). In women who were not pregnant, the 3 most commonly taken prescription medications were ethinyl estradiol (7.8%), levothyroxine (3.6%), and albuterol (3.0%).

Table 1 - The 10 most frequently used prescription medications by pregnant women and nonpregnant women in the US. Rank Medication Number of women reporting use Percentage of women reporting use (%) Typical indication Pregnant women  1 Levothyroxine 92,687 6.7 Thyroid disorder  2 Albuterol 54,568 5.4 Asthma  2 Montelukast 51,676 5.4 Asthma  4 Amoxicillin 43,803 4.3 Infection  5 Sertraline 40,578 3.4 Depression/anxiety  6 Nitrofurantoin 39,888 3.2 Urinary tract infection  7 Promethazine 39,168 3.0 Vomiting  8 Ondansetron 34,830 2.1 Vomiting  9 Cetirizine 29,908 1.9 Antihistamine  10 Budesonide 26,080 1.7 Asthma Nonpregnant women aged 15–44 yr  1 Ethinyl estradiol 4447,578 7.8 Hormone replacement therapy/contraceptive  2 Levothyroxine 1994,253 3.6 Thyroid disorder  3 Albuterol 1594,144 3.0 Asthma  4 Ibuprofen 1211,873 2.3 Pain/NSAID*  5 Acetaminophen 1093,857 1.8 Pain  6 Sertraline 1060,059 1.7 Depression/anxiety  7 Fluoxetine 998,637 1.7 Depression/anxiety  8 Bupropion 834,183 1.4 Asthma  9 Metformin 794,547 1.5 Diabetes  10 Escitalopram 723,468 1.3 Depression/anxiety

* NSAID, nonsteroidal anti-inflammatory drug.

3.2. Trends of polypharmacy in pregnant women

Figure 2 shows the prevalence of polypharmacy in pregnant women in each survey cycle. Overall, the prevalence of polypharmacy increased from 2.8% (1999–2000) to 10.0% (2015–2016) (P < .01). An increasing trend was found in the join point analysis, but the increase did not reach statistical significance (P > .05). In addition, the Annual average percentage change of polypharmacy in pregnant women for the entire study period from 1999 to 2016 was 14.2% (95% CI: −1.4% to 32.1%) (Table S1, Supplemental Digital Content, https://links.lww.com/MD/J12, Table S2, Supplemental Digital Content, https://links.lww.com/MD/J13).

Figure 2.:

Trends of polypharmacy in pregnant women aged 15–44 years from 1999–2000 to 2015–2016. * The prevalence of polypharmacy significantly increased from 2.8% (1999–2000) to 10.0% (2015–2016) (P < .01). ** The P value was 0.069 for the entire study period (1999–2016) according to the join point regression model. *** The average annual percent change (AAPC) of polypharmacy in pregnant women for the entire study period was 14.2% (95% confidence interval, −1.4% to 32.1%).

3.3. Characteristics of women with polypharmacy

Table 2 demonstrates characteristics of pregnant women with or those without polypharmacy. Race/ethnicity and chronic diseases were 2 variables found to be associated with polypharmacy. Hispanic white pregnant women (P < .05) and pregnant women with asthma (P < .05) or diabetes (P < .01) were more likely to report having polypharmacy.

Table 2 - Characteristics of US pregnant women aged 15–44 years with or those without polypharmacy (N = 3350,983). Variable Estimated population (×1000) Without polypharmacy Polypharmacy P value Predisposing factors, weighted n, ×1000, (%) Race  Mexican American 497 485 (97.6%) 12 (2.4%) <.05  Other Hispanic 199 186 (93.5%) 13 (6.5%)  Non-Hispanic white 1760 1585 (90.1%) 176 (10.0%)  Non-Hispanic black 548 522 (95.2%) 26 (4.8%)  Other races - including multiracial 347 326 (93.9%) 21 (6.1%) Age distribution (yr)  <25 1065 1012 (95.0%) 53 (5.0%) .21  25–34 1647 1511 (91.7%) 136 (8.2%)  ≥35 559 506 (90.5%) 53 (9.5%) Marital status  Married or living with partner 2480 2271 (91.6%) 209 (8.4%) .09  Single/divorced/widowed/separated 777 739 (95.1%) 38 (4.9%) Poverty income ratio  ≥350 870 809 (93.0%) 61 (7.0%) .13  130–349 1122 1056 (94.1%) 65 (5.8%)  <130 1118 1001 (89.5%) 117 (10.5%) Educational level  Some college or higher 1164 1101 (94.6%) 62 (5.3%) .07  High school graduate of lower 2027 1848 (91.2%) 179 (8.8%) Health-related factors, weighted n, ×1000, (%)  Health insurance coverage 1686 1517 (90.0%) 168 (10.0%) .20  General health condition of poor or fair 265 228 (85.9%) 37 (14.1%) .18  Hospitalization in the past year 297 261 (88.1%) 35 (11.9%) .39  Any chronic conditions 919 758 (82.5%) 161 (17.5%) <.01  High blood pressure 306 267 (87.2%) 39 (12.8%) .58  Asthma 541 446 (82.4%) 95 (17.6%) <.05  Arthritis 138 120 (86.4%) 19 (13.6%) .71  Diabetes 48 37 (78.0%) 11 (22.0%) <.01

Table 3 shows the characteristics associated with polypharmacy in pregnant women after adjusting covariates in the multivariable regression model. After adjusting for possible confounders including age, race, marital status, poverty income ratio, educational level, general health condition, health insurance coverage, and chronic conditions, non-Hispanic white pregnant women [OR: 25.64, 95% CI: 2.20–299.18] were more likely to have polypharmacy than other races. In terms of health-related conditions, pregnant women with a poor or fair self-reported general health condition were 5.12-times more likely to be associated with a higher prevalence of polypharmacy compared to pregnant women with a good self-reported general health condition (OR: 5.12, 95% CI: 1.23–21.34). Pregnant women with any of the following chronic conditions including high blood pressure, asthma, arthritis, and DM were 6.91-times more likely to have polypharmacy compared to pregnant women with no chronic conditions (OR: 6.91, 95% CI: 3.08–15.50).

Table 3 - Characteristics associated with polypharmacy among pregnant women: results from the adjusted multivariable regression analysis. Variable Multivariable analysis adjusted odds ratio (95% confidence interval) Race  Mexican American Reference  Other Hispanic 3.88 (0.39–38.20)  Non-Hispanic white 25.64 (2.20–299.18)  Non-Hispanic black 2.52 (0.25–25.94)  Other races - including multiracial 8.18 (0.70–95.59) Age distribution (yr)  <25 Reference  25–34 2.45 (0.69–8.70)  ≥35 1.80 (0.40–8.05) Marital status  Married or living with partner Reference  Single/divorced/widowed/separated 1.11 (0.33–3.73) Poverty income ratio  ≥350 Reference  130–349 2.34 (0.35–15.56)  <130 0.69 (0.27–1.77) Educational level  Some college or higher Reference  High school graduate or lower 0.50 (0.14–1.73)  Health insurance coverage 1.77 (0.30–10.44)  General health condition of poor or fair 5.12 (1.23–21.34)  Hospitalization in the past year 1.45 (0.42–5.01)  Any chronic conditions* 6.91 (3.08–15.50)

* Chronic conditions included high blood pressure, asthma, arthritis, and diabetes.

4. Discussion 4.1. Principle findings

To the best of our knowledge, this is the first cross-sectional study conducted using the NHANES database to investigate prevalence, trends, and characteristics of polypharmacy in pregnant women aged 15 to 44 years in the US. In our study, pregnant women were overall less likely to be experiencing polypharmacy compared to nonpregnant women. The prevalence of polypharmacy among pregnant women increased from 1999 to 2016. We also observed that levothyroxine and albuterol were 2 commonly taken prescriptions in both pregnant and nonpregnant women. As to the characteristics of pregnant women with and those without polypharmacy, pregnant women that were non-Hispanic white or had chronic conditions such as asthma and DM were more likely to experience polypharmacy. Moreover, pregnant women with a general self-rated health condition of poor or fair or who had chronic conditions were more likely to be experiencing polypharmacy.

4.2. Results

Our study showed that pregnant women were less likely to be experiencing polypharmacy compared to nonpregnant women, which is consistent with findings of a previous study.[22] A previous cross-sectional study conducted using the 1999 to 2006 NHANES databases found that medication use was reported by 22% of pregnant women and 47% of nonpregnant women.[22] This reflected the fact that real-world pregnant women are less likely to receive medications, or physicians are less likely to prescribe medications to pregnant women.[23,24] Our study updated the findings of that previous study,[22] by providing a longer study duration and containing a larger sample size of the study population, and was thus more representative of the US women of childbearing age and pregnant women.

We found a significant increasing trend of polypharmacy in pregnant women from 1999 to 2016, which was the same as findings of the prior study.[2] The increasing trend of polypharmacy could have resulted from the healthcare system, prescription patterns, and characteristics of pregnant women that have changed in recent years.[