Preterm Labor Using Tocolysis as a Possible Risk Factor for Postpartum Depression: A 14-Year Population-Based Study in Taiwan

Postpartum depression (PPD) is associated with negative physical and mental health outcomes for the mother and infant. Women often experience elevated symptoms of PPD, and the incidence of PPD has increased in recent years. There were lack of studies to investigate the effects of medications during pregnancy. Herein, we focused on the most common obstetric medical therapies used in labor and determined whether the medical therapies cause mental stress in pregnant women. This 14-year retrospective population-based nationwide study was based on the National Health Insurance Research Database. Univariate and multivariate logistic regression analyses were used to evaluate unadjusted and adjusted odds ratios and 95% confidence intervals for each tocolytic and uterotonic treatments during pregnancy and common medical illnesses. In comparing the effects of tocolytic and uterotonic medications on maternal PPD, tocolysis with the injection form of ritodrine resulted in a significantly higher risk of PPD based on multivariate analysis. This study supports existing research demonstrating an association between tocolysis with ritodrine and PPD. Ritodrine treatment for preterm labor was a significant risk factor for PPD, especially the injection form. This information provides obstetricians and health policy providers to pay attention to maternal mental health outcomes among high-risk pregnant women.


Introduction
Postpartum depression (PPD) is an international public health priority. PPD is the most common cause of postnatal morbidity, with a prevalence as high as 19%, thus presenting a challenging target for prevention [1][2][3]. Postpartum mental health problems are recognized as a suitable target for preventive approaches, with the potential to avert the burden to women, their children and families, as well as the social and economic costs [4]. Women experience physical and mental changes during pregnancy and childbirth. These changes may cause a depressive mood. PPD is defined as a major depressive episode that begins during pregnancy or within 4 weeks after delivery [5]. PPD is widely reported worldwide with a prevalence from 0.5-60.8% [6,7]. Huang et al. [8] conducted a study, which showed a 19% and 18% prevalence of PPD in Taiwan and the United Kingdom, respectively. Many obstetric and pediatric factors are associated with PPD, including unplanned pregnancy, parity, newborns with medical illnesses, preterm labor, and complications during pregnancy [6,[9][10][11][12][13]. Some medical illnesses affecting the fetus, infant, and pregnant women need medications for treatment. Medications for these illnesses may cause mental stress on pregnant women [14].
The preterm birth rate in the United States had declined from 12.8% in 2006 to 11.7% in 2011, especially in the late-preterm category (34-36 weeks) [15]. In undeveloped and developing countries, the preterm birth rate was around 6% to 10% of all births, leading an important cause of perinatal mortality and morbidity [16]. In Taiwan, the preterm birth rate was around 8.56% of those who delivered between 32 and 37 weeks of gestation [17].
The perinatal morbidity was necrotizing enterocolitis, intracranial bleeding, and respiratory distress syndrome, and the neonatal mortalities were approximately two-thirds. It caused serious social and economic losses [18][19][20]. Therefore, it is important to improve perinatal outcomes by using maternal tocolytic treatments to prevent or delay preterm births. The U.S. Food and Drug Administration (FDA) had approved the beta-adrenergic receptor agonist ritodrine for preterm labor. Parenteral ritodrine confirmed a delivery delay up to 48 hours. The delay may aid maternal transport to tertiary care or permit fetal lung maturation with corticosteroids. The maternal effects and side effects of ritodrine are well known; pulmonary edema is a special concern. However, the possible adverse mental effects on pregnant women remain relatively unknown. In this present study, we have determined the clinical mental side effects of ritodrine, the most frequently used tocolytic agent, and evaluated the effects on PPD.
Oxytocin, as a uterotonic agent, is central to normal birth, lactation, and motherinfant attachment [20,21]. Oxytocin also plays a key role in regulating emotion, social interaction, and stress reactivity [22,23]. In addition, reductions in oxytocin measured in plasma [24,25] have been associated with PPD. Synthetic oxytocin is commonly used in labor management to induce and augment labor and to prevent postpartum hemorrhage. Higher synthetic oxytocin doses are associated with increased depressive, anxious, and somatization symptoms [26]. Inconsistencies in the reports of an association of PPD may depend on endogenous or synthetic oxytocin. The long-term consequences of oxytocin for maternal mental health and behavior are largely understudied. We determined the relationship between intrapartum synthetic oxytocin and postpartum maternal mental health.
Women often experience elevated symptoms of PPD, and the incidence of PPD has increased in recent years. There is also a lack of studies to investigate the effects of medications during pregnancy. Herein, we have focused on two of the most common obstetric medical therapies used in labor (tocolytics and uterotonic) and determined whether or not the medical therapies cause mental stress in pregnant women. Our secondary objective was to investigate the mental side effects of medications. Thus, we conducted a 14-year nationwide population-based study from 2000 to 2013 to determine the relationship between medications during pregnancy and PPD.

Data Source and Collection
This population-based study used the National Health Insurance (NHI) database. The NHI database covers 99.9% of the 23 million residents in Taiwan. The Taiwan NHI program began in 1995. The NHI database contains nationwide medical record and demographic information for inpatients and outpatients in Taiwan [27,28] We performed a retrospective study retrieved from the Longitudinal Health Insurance Database 2000 (LHID2000), a sub-dataset of the NHI research database (NHIRD), between January 2000 and December 2013. The LHID2000 randomly selected 1 million residents of the 23 million residents included in the NHIRD in 2000. The demographic distribution of residents in the LHID2000 and the original population in the NHIRD are similar [29]. The disease diagnosis was based on the International Classification of Diseases, 9th revision, Clinical Modification (ICD-9-CM) [30]. This retrospective study was approved by the Institutional Review Board of the Tri-Service General Hospital (approval no. B-104-20, Taipei, Taiwan), and the requirement to obtain informed consent was waived.

Study Population
The study subjects were selected from the LHID2000 database between January 2000 and December 2013. Figure 1 shows the selection process for the study population. We first selected females whose first delivery was between January 2000 and December 2013. The exclusion criteria included the following: (1) primiparas with a delivery date before 31 December 2000 or after 1 July 2013 (n = 15,112); (2) <20 years of age (n = 1962); (3) previous major depressive disorder or a history of psychosis (n = 3404); (4) follow-up period < 6 months (n = 251). Finally, we selected females who were newly diagnosed with major depressive disorder during the first 6 months after delivery (ICD-9-CM: 309, 311, 296.2, 296.3, 296.5, and 300.4). The diagnosis of major depressive disorder was established by psychiatrists who also provided treatment with antidepressants. The Charlson comorbidity index (CCI) score was used to evaluate the clinical status of the subjects. The CCI score contains 19 comorbid conditions based on the method of Quan et al. [30].
Board of the Tri-Service General Hospital (approval no. B-104-20, Taipei, Taiwan), and the requirement to obtain informed consent was waived.

Study Population
The study subjects were selected from the LHID2000 database between January 2000 and December 2013. Figure 1 shows the selection process for the study population. We first selected females whose first delivery was between January 2000 and December 2013. The exclusion criteria included the following: (1) primiparas with a delivery date before 31 December 2000 or after 1 July 2013 (n = 15,112); (2) < 20 years of age (n = 1962); (3) previous major depressive disorder or a history of psychosis (n = 3404); (4) follow-up period < 6 months (n = 251). Finally, we selected females who were newly diagnosed with major depressive disorder during the first 6 months after delivery (ICD-9-CM: 309, 311, 296.2, 296.3, 296.5, and 300.4). The diagnosis of major depressive disorder was established by psychiatrists who also provided treatment with antidepressants. The Charlson comorbidity index (CCI) score was used to evaluate the clinical status of the subjects. The CCI score contains 19 comorbid conditions based on the method of Quan et al. [30].

Covariates
Covariates were chosen in the current study based on the extant literature and clinical manifestations during pregnancy. We selected the following covariates: chronic pulmonary disease (ICD-9 The monthly income of subjects based on insurance data of NHI was also included. In addition, we chose two of the most common drug therapies during labor, which have a substantial influence on the mental status of women (tocolytics and uterotonics).

Statistical Analysis
Statistical analysis was performed by Microsoft SQL Server 2008 and IBM SPSS statistics software v20 (IBM SPSS, 2013). All medical records and baseline data were managed with Microsoft SQL Server 2008 and statistically analyzed using IBM SPSS statistics software v20 (IBM SPSS, 2013). Multivariate logistic regression was used to analyze the correlation between comorbidities, drug therapy during labor, and PPD. A two-sided p value < 0.05 was regarded as statistically significant.

Demographics
A total of 52,548 primiparas were enrolled in this study between January 2001 and June 2013. There were 326 women with newly diagnosed PPD within 6 months after delivery during the study period. We divided the 52,548 primiparas into two groups, as follows: PPD (n = 326) and non-PPD (n = 52,222). The demographic characteristics for the study groups are shown in Table 1. There were no differences in age and hospital stay in the two groups. The average age of women with PPD was 29.5 ± 4.6 years, and the average length of hospital stay was 3.8 ± 2.6 days. Women with PPD had significantly elevated CCI scores compared with women without PPD (p = 0.03, 0.2 ± 0.5 vs. 0.1 ± 0.4). Some comorbidities were significantly increased in the PPD group, including hypertension (p = 0.01), peptic ulcer disease (p = 0.0008), chronic kidney disease (p = 0.004), heart disease (p < 0.001), unstable lie (p = 0.006), and early delivery onset (p < 0.001). Monthly incomes were significantly different between the two groups. Tocolytic medications were significantly associated with PPD (p = 0.002). Both the oral and injection forms of tocolytic medications were associated with PPD. In contrast, uterotonic medications were not associated with PPD.

Discussion
Pregnant women are easily affected by physical, mental, or surrounding changes that cause a depressed mood. According to the fifth edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-5) [31], PPD is defined as a major depressive episode occurring during pregnancy or in the 4 weeks following delivery. This is the first study involving medications during pregnancy and PPD using a national population-based method. In the present study, it was determined that ritodrine, when used as a tocolytic agent in preterm labor, was associated with PPD. In particular, the injection form of ritodrine (p = 0.022, OR = 2.238, 95% CI = 1.122-4.463) was a significant risk factor for PPD after logistic regression analysis. In addition, chronic kidney disease, heart disease, an unstable lie, and lower monthly income were significant risk factors for PPD. Our results provide useful information for physicians and pregnant women with respect to the early prevention of PPD.
Based on univariate analysis, tocolytic medications were significantly associated with PPD. Pregnant women who require tocolytic medications are at risk for preterm labor. Preterm labor is a strong risk factor for PPD [32,33]. Adewuya et al. [34] reported that Nigerian women with preterm labor had 4.21 times the risk for PPD (OR = 4.21, CI = 2.78-6.39). Drewett et al. [35] concluded that women with preterm labor had 1.6 times the risk of PPD 8 weeks thereafter (adjusted relative risk = 1.6, 95% CI = 1.2-2.1). Logistic regression analysis showed that the injection form of tocolytic medications had 2.238 times the risk for PPD in the current study (OR = 2.238, 95% CI = 1.122-4.463). We suggest that pregnant women who require the injection form of tocolytic medications are under stress caused by the physical, mental, and surrounding changes leading to the development of PPD. In the current study, the oral form of tocolytic medications was not a risk factor for PPD based on logistic regression analysis. The injection versus oral forms of tocolytic medications are used depending on the severity of preterm labor. Indeed, a low probability of preterm labor only requires an oral form of tocolytic medication, hydration, and bed rest at home. In such cases, the stress associated with physical, mental, or surrounding changes is not sufficiently severe to affect the pregnant woman's mood. Different degrees of severity of preterm labor induces different stress on women with pregnancy. Vigod et al. [32] published a review article that showed women with preterm labor are at higher risk for PPD than women with term labor. Vigod et al. [32] also showed that women with very preterm labor had a continued risk of PPD throughout the first postpartum year. In addition, women with preterm labor have greater opportunity to receive cesarean sections, which are associated with PPD [36][37][38][39].
Uterotonic drugs, such as synthetic oxytocin, had no significant correlation with postpartum anxiety or PPD in the current study. Andersson et al. [40] reported that post-term labor had no correlation with postpartum anxiety or PPD, and oxytocin use had no significant association with postpartum anxiety or PDD. Nevertheless, higher synthetic oxytocin dose was associated with greater depressive, anxious, and somatization symptoms [26]. The inconsistencies in the reports of association of PPD may depend on ethnicity, background, the nature of behaviors involved, social salience, social environment, and individual social factors. In addition, post-term infants are usually healthier than preterm infants and imposed little mental stress on women. The induction of post-term infants leads to fewer neonatal deaths or complications [41].
This study showed that heart disease, an unstable lie, lower monthly income, and chronic kidney disease are significant risk factors for PPD. Yang et al. [36] also showed that heart disease is associated with PPD, but unstable lie is not. An unstable lie may need additional time for labor with an increased use of epidural analgesia. Chung et al. [42] reported that epidural analgesia is related to PPD. Several studies support that low income is associated with PPD [43,44]. Chronic kidney disease, such as kidney and urinary infections, might be associated with PPD symptoms and a PPD diagnosis [45].
Finally, it is obligatory to emphasize the situation of tocolytic agents, which is important for preterm labor, while discussing the results of our study on the effects of ritodrine on maternal PPD. There were several limitations to this study. This was a retrospective population-based study. Some data, such as the marital relationship, education status, and employment status, were not available in the NHIRD database, and these data are also important risk factors for PPD. The selected time interval of our study is limited due to the LHID2000 database being between January 2000 and December 2013, which could not be updated and did not contain the latest information. In the current study, we had 14-year study periods, and the study results were able to provide useful information. The diagnosis of PPD was made by psychiatrists in this study. It is evitable that every psychiatrist may have individual judgements, even under DSM V guidelines. Women who experienced mental stress might not consult psychiatrists; thus, PPD is underestimated. The number of patients was relatively small due to our strict inclusion criteria (primiparas only). We only chose two kinds of obstetric medications for the investigation. There are still other medications worthy of study. Further prospective studies are needed to evaluate the relationship between drug therapy during labor and PPD.

Conclusions
Although we know a lot about the pharmacokinetics and pharmacodynamics of these drugs, there are no data about their effects on maternal mental health. Tocolytic drugs are a significant risk factor for PPD, especially the injection form. In contrast, uterotonic drugs had no correlation with PPD. This result indicates that preterm labor causes much more mental stress on women than post-term labor. In addition, chronic kidney disease, heart disease, unstable lie, and low monthly income were also risk factors for PPD. Further prospective research to investigate the relationship between drug therapy during labor and PPD is recommended. We have provided useful information for gynecologists, obstetricians, and health policy providers that should pay attention to the influence of drug therapy during labor on mental stress.