What are the probable predictors of urinary incontinence during pregnancy?

Inclusion and exclusion criteria

Pregnant women above 18 years of age without any acute or chronic disease were included in the study. Pregnant women in the high-risk category were excluded. Individuals were also excluded based on the following: the presence of any systemic or chronic diseases, such as diabetes mellitus or any condition of increased blood glucose levels or disturbed glucose states; hypertension (blood pressure over 125/85 mmHg); hepatitis or any state with elevated liver enzymes; any neurological disease; Cushing’s disease; asthma; cardiac failure; central nervous system disorders or urinary tract infection or stones, etc. Women with previous urogynecologic diseases and obvious neuropathies leading to UI were also excluded from the study. Other risk factors for UI were also asked, including smoker status and use of medications (such as alpha-blockers and cholinergic or anticholinergic drugs), sedatives, myorelaxants, diuretics and angiotensin-converting enzyme inhibitors. Patients who smoked and used such substances were also excluded from participation.

Statistical analysis

All data were analyzed using SPSS Version 19.0 for Windows (SPSS Inc., Chicago, Illinois, USA). Categorical variables were presented as frequencies and percentages, and continuous variables were expressed as mean ± SD. The normality of the distribution of continuous variables was tested using the Shapiro–Wilk test. Differences in continuous variables between groups were examined using the independent sample t-test or nonparametric Mann–Whitney U test. The comparison of results between three or more groups was made using the Kruskal–Wallis test. The Dunn’s test was used as a post hoc test if the Kruskal–Wallis test was statistically significant. Categorical values were compared using a chi-square test. Multivariate logistic regression analysis was performed to assess independent risk factors. A p-value of <0.05 was considered statistically significant.

Socio-demographic features

Age: The majority of participants were young pregnant women in the age group of 21–29 years. The mean age of all participants was 27.5 ± 5.1 years. Two age groups were formed: 18–35 and ≥35 years. There was no significant difference in terms of existence of UI with respect to age groups (p = 0.146, Table 1). Age was a possible predictor in developing UI, but there was no statistically significant relationship between the existence of UI and age (OR = 0.845, 95% CI [0.268–2.669], p = 0.782, Table 6).

Height: There was a significant difference between UI-present and UI-absent groups according to body height (p = 0.037, Table 1).

BMI: There was no significant difference between UI-present and UI-absent groups according to BMI values (p = 0.881, Table 1). BMI was in favor of the onset of UI, but not significant statistically in logistic regression analyses (p = 0.998, OR = 1.000, 95% CI [1.000–1.000], Table 6).

Occupational status: Over two-thirds of the participants were housewives (that is, they had no occupation other than carrying out housework) in the UI-present group (n = 39, 69.6%) and UI-absent group (n = 53, 69.7%). There was no significant difference in terms of the occurrence of UI between those working and those not working (p = 0.122, Table 1). Logistic regression analyses revealed that occupational status might be a predictor of the occurrence of UI (OR = 0.897, 95% CI [0.392–2.055]), but it was not statistically significant (p = 0.798, Table 6).

Educational level: Having graduated at a higher level than primary school was a significant feature between UI-present and absent groups (p = 0.016, Table 1) The number of primary school graduates in UI absent group (n = 32) was significantly more than UI-present (n = 11, p = 0.01). The educational levels in the 2013 census of Zonguldak have been revealed that 6% were illiterate, 16% were literate but did not graduate at any school, 26% graduated from primary school, 23% graduated from intermediary school, 18% graduated from high school and 11% for university degree graduates, whereas in the current study, similar results were found but not any for the illiteracy category (Table 1). In the present study, the number of primary school graduates in UI-present (19.6%) and intermediate school graduates in UI absent (6.6%) were lower compared to general distribution in the area. Also, university degree graduates had a different distribution. All participants in the study had some degree of graduation. The number of primary school graduates in the UI absent group was significantly more than UI-present.

Place of living: The location (rural vs. urban) of one’s residence was significant in the logistic regression analyses (p = 0.020, 95% CI 0.887 and 3.653, Table 6). Hence, living in a rural area favored the occurrence of UI. Zonguldak is a city with population size of 326,374 (58.4%) in rural areas and 232,200 (41.6%) in urban areas. Our study population had a rural predominance (76/56). The ratio of rural vs. urban was 36/20 in UI-present group, and 36/36 in UI absent. The table of random numbers was used to overcome selection bias. Pregnant women who came from rural areas might have chosen the hospital because of some economic, social or social security preferences instead of special hospitals, or other health care providers.

Gestational features: With respect to parity, gestational weeks, multiple pregnancies, interval between pregnancies and the occurrence of miscarriage or anemia, the results of statistical analyses are presented in Table 2. There was no significant difference according to gestational weeks or trimesters because all participants were in their third trimester (p = 0.908). We did not encounter any significant difference between UI-present and UI absent groups in the occurrence of UI according to parity values (p = 0.358), history of multiple pregnancies (p = 0.747) or interval between previous pregnancies (p = 0.283, Table 2) either. There were statistically significant relationships according to history of miscarriage (p = 0.002, OR = 1.219, 95% CI [0.588–2.825]), and also parity values (p = 0.006, OR = 0.519, 95% CI [0.325–0.829], Table 6) in logistic regression analyses.

Previous observations have suggested that parity, or pregnancy itself, might contribute to the onset of UI independent of the mode of delivery. Consistent with the literature, the present study denoted that parity was statistically a predictor of the onset of UI (Table 6) (Fritel et al., 2012; Cetinel et al., 2007).

Trimesters: All participants in the present study were in the same (third) trimester. Most of the patients had been referred to the health center, a tertiary level of care, by their primary or secondary healthcare providers in the region. This referral is often carried out at a time near the suspected birth date. As stated above, no statistical difference was found with respect to trimesters because all of the participants were in their third trimester (p = 0.09, Table 2). Patients may have preferred to visit the health center shortly before delivery.

Miscarriage: The relationship between history of miscarriage and presence of UI was statistically significant (p = 0.041 in chi-square test, Table 2 and p = 0.002 in logistic regression analysis, Table 6).

When the history of previous pregnancies was further analyzed, no statistically significant relationship was found with respect to history of preterm labor (p = 0.474), anomalous babies (p = 0.827), chronic disease—if present—during previous pregnancies (p = 0.828), or anemia (p = 0.862, Table 2). Regarding the present pregnancy, there was no significant difference in the occurrence of UI according to vitamin usage (p = 0.166), weight gain (p = 0.995), exercise (p = 0.099), sexual intercourse (p = 0.366).

Anemia: Participants who had mean blood hemoglobin values below 11.5 mg/dl were accepted as anemic during the study. There was no significant difference in the presence of anemia between UI-present and absent groups (p = 0.862, Table 2). Prior to the onset of the study, any chronic disease patients were excluded from participation. Anemia presented in the current study most likely developed during pregnancy due to insufficient iron intake, though women deficient in iron, folate and vitamin B12 were prescribed supplements beforehand in order to participate in the study. Logistic regression analysis revealed that anemia was indicated in favor of the onset of UI; as the anemia worsened, the possibility of developing UI increased (Table 6). However, it was not significant in logistic analyses (Table 6).

Impact of UI on QoL: With respect to frequency of UI reported in the UI-present group, there were occasions of urinary leakage once a week or less in 18 participants (32.1%), twice or thrice a week in eight participants (14.3%), once a day in five participants (8.9%), a few times a day in 14 participants (25.0%) and constantly throughout the day in eight participants (14.3%, Table 3).

Among those in the UI-present group, mainly a small amount of urine leaked in 33 participants (58.9%), a moderate amount leaked in four participants (7.1%) and large amount in four participants (7.1%, Table 3). The amount of urine leakage was assigned by the number of hygienic pads used.

With respect to Wagner’s QoL scores, the majority of UI-present women (n = 33, 58.9%) experienced mild urinary incontinence symptoms, some of them experienced moderate symptoms (n = 4, 7.1%) and 7.1% of women (n = 4) experienced severe symptoms, whereas urinary incontinence did not affected their QoL in 26.8% of women with UI (n = 15, Table 4). There were statistically significant relationships between QoL scores and frequency of UI as well as the amount of leakage (p = 0.002 and p = 0.002, respectively, Kruskal–Wallis test). Thus, in general, the majority reported having mild UI.

UI impacted the daily life activities of women by making them less likely to undertake activities outside their homes (23.2%), by affected their working performance and friendships (8.9%), their daily home activities (7.1%), their general health status (12.5%), their sexual relations (12.5%), by making them more nervous or anxious (10.7%) and by the need to wear pads or protectors (25%) (Table 5).

Whether the outcomes were clinically relevant in the sense that the statistically relevant differences that we found were also of meaning in the clinic, we tested with the minimal important difference method. Therefore, we used the standard deviations given in Table 4 to calculate whether the found relationships were clinically relevant as well. We transformed standard deviations given in to standard errors, and performed an ANOVA test for the statistics, then we found that there were significance between them (p = 0.001, Table 7). Then, we performed the Tukey and Tamhane tests: the statistically significant results gained from Kruskall Wallis analyses (Table 4) for impact of urinary incontinence on QoL of pregnant women were relevant with clinical outcomes; they were all significant (Table 8). Thus, we could state that there was a statistical relevance between results of QoL tests and clinical outcomes.

Statistical importance tests were also performed for history of preterm labor (p = 0.341), babies small for gestational age (p = 1.000), anomalous babies (p = 1.000), alcohol intake (no participant had alcohol intake), vitamin intake (p = 0.166), exercise (p = 0.099), age of first birth (p = 0.390) and sexual intercourse (p = 0.366). None of these variables were significantly related to the occurrence of UI.

According to logistic regression analyses, the following factors were designated to favor the existence of UI: age (OR = 0.845, 95% CI [0.268–2.669]), occupational status (OR = 1.800, 95% CI [0.850–3.810]), anemia (OR = 0.939, 95% CI [0.464–1.901]), parity (OR = 0.519, 95% CI [0.325–0.829]), miscarriage in previous pregnancies (OR = 1.219, 95% CI [0.588–2.825]) and place of living (rural vs. urban, OR = 1.8, 95% CI [0.887–3.653]).

Miscarriage, parity and place of living (living in a rural settlement) were statistically significant predictors of the occurrence of UI (p = 0.002, p = 0.006 and p = 0.020 respectively: Table 6).

Obstetrical features

In the current study, parity and miscarriage were significant predictors of onset of UI, but trimester was not. Hansen et al. (2012) demonstrated that, with adjustment for potential risk factors, UI in pregnant women was 3.3 times more prevalent than UI occurring in a control group of nulliparous women. In a study by Abdullah et al. (2016), frequency of UI was 34.3% and trimester was not a significant factor for occurrence of UI. In the study by Seshan and Muliira, the participants with UI had one or more miscarriages in the past (79% of the total participants, p < 0.01), supporting the current findings. Thus, miscarriage was a predictor of the occurrence of UI (Seshan & Muliira, 2013). Findik et al. (2012) stated in their study that among women who had experienced miscarriage, the rate of stress incontinence was significantly high. In addition, as the number of miscarriages increased, the rate of stress incontinence also increased, but the rate of urgent UI was not influenced by miscarriage (Townsend et al., 2007). However, in the current study, distinctions between types of UI were not made.

In the present study, anemia was a predictor for the presence of UI but not significant. We do not know if anemia led to or why and how anemia might lead to UI. Most probably it was not a primary cause but could be regarded as a secondary in development. In the study by Pahor, Manini & Cesari (2009), they stated that anemia was associated with reduced physical performance and muscle strength. They showed the decreases in skeletal muscular strength measures occurred in the presence of anemia (Pahor, Manini & Cesari, 2009). Sangsawang & Sangsawang (2016) stated that any derangement of pelvic floor muscles were important in development of UI. Walker & Gunasekera (2011) said that the overall prevalence of UI was 28.7% (ranging from 5–70%) and together with Scherf et al. (2002) and Bodner-Adler, Shrivastava & Bodner (2007) they stated that the common factors that have been reported to be associated with UI in women included standards of living, poor nutrition (causing poor tissue tensile strength), anemia and regular physical heavy work. Moderate and severe anemia also appears to be a risk factor associated with pelvic organ prolapse leading to UI. The other factors included age, obesity, and menopause. We could say anemia, in addition to other poor socioeconomic features, might lead to a decrease in pelvic floor muscle performance or pelvic organ prolapse at some level and then to the onset of UI (Walker & Gunasekera, 2011; Scherf et al., 2002; Bodner-Adler, Shrivastava & Bodner, 2007).

Contrary to Zhu et al. (2009), the authors could not state any decision about significance with respect to alcohol intake because no participant had used alcohol.

QoL scores: Among social complaints impacted by UI, those related to wearing pads or protectors were the most encountered requirements leading to increase in economic cost consistent with the study by Kocak et al. (2005). There were significant relationships between QoL scores and both the frequency and the amount of UI (p = 0.002 and p = 0.002, respectively), whereas in the study by Kocaoz, Talas & Atabekoglu (2010) there was no significant relationship between the amount of UI and mean QoL score; however, there was a statistically considerable relationship between frequency of UI and mean QoL score, signifying that increased frequency of UI significantly impacted women’s QoL. Minassian, Drutz & Al-Badr (2003) stated that UI had moderate to severe impact on QoL in 10%–22% of the individuals. Abdullah et al. (2016) stated that more than 50% of women at their third trimester felt that UI did not impact on their daily activities at all; however 10% of women felt greatly affected. Seshan & Muliira (2013) found that the majority of women with UI experienced symptoms at a moderate level (78%) and others rated the symptoms as mild (22%). Adamczuk et al. (2015) studied stress UI and its impact on QoL. They found that UI turned out to be a depressing factor, and it was associated with lower QoL.

Restrictions of the study: All of the participants were in their third trimester. This is probably because they came to the hospital at a time shortly before giving birth due to social, cultural or most probably economic reasons. Some women were also unwilling to talk about their symptoms because of being ashamed and had thus chosen not to participate. Therefore, the rate of UI cases might have been underestimated.

Treatment options: Treatment options, such as pelvic floor muscle exercises, have been available for UI and discussed in the relevant literature, though they were not investigated in the current study. Pelvic floor muscle exercise is a safe and effective treatment for UI during pregnancy, without significant adverse effects (Sangsawang & Sangsawang, 2013). Hence, dealing with UI in pregnancy is important with respect to daily healthcare services from a therapeutic point of view (Sangsawang & Sangsawang, 2016).