The impact of gestational diabetes mellitus on postpartum urinary incontinence: a longitudinal cohort study on singleton pregnancies

Transcription

1 DOI: /j x Epidemiology The impact of gestational diabetes mellitus on postpartum urinary incontinence: a longitudinal cohort study on singleton pregnancies C-M Chuang, a,b, I-F Lin, a, H-C Horng, b Y-H Hsiao, b I-L Shyu, b P Chou a a Institute of Public Health, School of Medicine, National Yang-Ming University, Taipei b Department of Obstetrics and Gynecology, Taipei Veterans General Hospital, Taipei, Taiwan Correspondence: Dr P Chou, No.1, Sec., Linong Street, Taipei 11, Taiwan. cmjuang@gmail.com These authors contributed equally to this work. Accepted 1 May 1. Published Online August 1. Objective To determine whether gestational diabetes mellitus (GDM) is an independent risk factor for postpartum urinary incontinence in singleton pregnancies. Design A longitudinal cohort study. Setting A single tertiary-care hospital in Taiwan. Population Pregnant women with term deliveries between and 7 (n = ) were consecutively recruited. Methods Logistic regression models were fitted based on generalised estimating equation methods to derive odds ratios for occurrences of type-specific urinary incontinence in the third trimester and at four time-points over years during the postpartum period. Main outcome measures Evaluation of whether GDM is an independent risk factor for postpartum urinary incontinence. Results The full model analysis revealed that GDM was an independent risk factor for all type-specific urinary incontinence (odds ratio [9% confidence interval]: 1.97 [1..1],.11 [.1.] and.7 [1.7.] for stress, urge and mixed incontinence, respectively]. Compared with women without GDM, women with GDM tended to exhibit more severe symptoms of stress incontinence for up to years postpartum, whereas for urge or mixed incontinence, more severe symptoms were found only for months postpartum. Evaluation of quality of life using the Incontinence Impact Questionnaire 7 suggested that women with GDM requiring insulin treatment had a higher likelihood of functional impairment than women with GDM requiring conservative treatment only or women without GDM (P <., by the chi-square test for trend). Conclusions GDM was found to be an independent risk factor for postpartum urinary incontinence and had a significant impact on quality of life. Women with GDM should be provided with timely consultation and support once urinary incontinence occurs. Keywords Gestational diabetes mellitus, mixed urinary incontinence, postpartum, stress urinary incontinence, urge urinary incontinence. Please cite this paper as: Chuang C, Lin I, Horng H, Hsiao Y, Shyu I, Chou P. The impact of gestational diabetes mellitus on postpartum urinary incontinence: a longitudinal cohort study on singleton pregnancies. BJOG 1;119:1 1. Introduction Urinary incontinence is a highly prevalent disorder affecting nearly % of middle-aged and older women. 1 Epidemiological studies conducted in various populations have identified several potential risk factors for urinary incontinence, including age, parity, obesity, hysterectomy, diabetes mellitus, neurological disorders, lung disorders and genetic or ethnic factors., Diabetes mellitus is a worldwide disorder that is increasing in prevalence as a consequence of lifestyle changes, increasing obesity rates and population aging. Women with diabetes mellitus are % more likely to experience urinary incontinence compared with women with normal glucose levels. Childbearing is an established risk factor for urinary incontinence among young and middle-aged women. The likelihood of urinary incontinence increases with the number of children delivered; however, the greatest influence is the birth of the first child. 7 Pregnant women experience multifactorial mechanical and hormonal changes that are linked to urinary incontinence. Peripartum factors that 1 ª 1 The Authors BJOG An International Journal of Obstetrics and Gynaecology ª 1 RCOG

2 Gestational diabetes mellitus and urinary incontinence exacerbate postpartum urinary incontinence are an increased prepregnancy body mass index, higher parity, vaginal delivery and a prolonged second stage of labour.,9 These studies were mainly cross-sectional in design, with relatively short-term postpartum follow up. 7 1 Gestational diabetes mellitus (GDM) is a condition of carbohydrate intolerance resulting from inadequate insulin supply first recognised or developed during pregnancy, with varying perinatal morbidity and mortality. 1 Although the associations between diabetes mellitus and urinary incontinence, and between pregnancy and urinary incontinence are well established, the associations between GDM and urinary incontinence have rarely been investigated. Currently, there is only one relevant report, which revealed a positive relationship between GDM and stress urinary incontinence. 1 In the light of the complex inter-relationships among diabetes mellitus, pregnancy and urinary incontinence, in the current study we hypothesised that GDM may be associated with urinary incontinence. We adopted a prospective-cohort, repeated-measure design to investigate the effects of GDM on the occurrence, severity and impact on quality of life (QoL) of postpartum urinary incontinence over a -year follow-up period. Methods Study design Women who had term deliveries between January and December 7 at a tertiary-care hospital were consecutively recruited. The study was approved by the Institutional Review Board of Taipei Veterans General Hospital. The procedures used in the study were in accordance with the guidelines of the Helsinki Declaration on human experimentation. The diagnosis of GDM was defined according to the criteria recommended by the National Diabetes Data Group 1 derived from O Sullivan s criteria. Briefly, each pregnant woman underwent a glucose challenge test at between and weeks of gestation. For measurement of glucose levels, a venous blood sample was drawn 1 hour after intake of a -g oral glucose load without regard to fasting state. A positive glucose challenge test was defined as a venous plasma glucose level 7. mmol/l. Women with a positive glucose challenge test were then recruited for a 1-g, -hour oral glucose tolerance test. The diagnosis was confirmed when two or more venous plasma glucose values reached or exceeded the following thresholds: fasting,. mmol/l; 1 hour, 1. mmol/l; hours, 9. mmol/l; hours,. mmol/l. Urinary incontinence was defined according to the International Continence Society definition of urinary incontinence. 17 Types of urinary incontinence were classified as stress urinary incontinence (SUI; involuntary leakage on effort or exertion, or on sneezing or coughing), urge urinary incontinence (UUI; involuntary leakage accompanied by or immediately preceded by urgency), and mixed urinary incontinence (MUI; involuntary leakage associated with urgency and with exertion, effort, sneezing or coughing). Inclusion and exclusion criteria Inclusion criteria were that the participant was a pregnant woman aged 1 years with a gestational age of >7 weeks in the index pregnancy and singleton gestation; was willing to undergo all study-related assessments; had read the information provided on the study and given written consent to participate. Exclusion criteria were as follows: preterm delivery (<7 weeks of gestation); known type 1 or type diabetes mellitus; pre-gestational hypertension; multifetal gestation; known fetal anomaly; any clinical condition that in the opinion of the investigators may have jeopardised the health status of the woman. Data collection Individuals who satisfied the inclusion criteria were invited to participate in the study after informed consent had been provided. Baseline information was collected at two timepoints. First, at between 7 and 1 weeks of gestation, information was obtained on maternal characteristics, including demographics, anthropometrics, previous obstetric history, and incontinence status at prenatal visits. Second, a form completed immediately after the birth was used to record the labour process, the mode of delivery and the neonatal birth profile. A ten-item composite questionnaire was used to assess the type, severity and impact on QoL of urinary incontinence. This questionnaire was also used in our previous report. 1 At weeks, months, 1 and years postpartum, participants were followed by face-to-face or telephone interview by trained nurses, as appropriate, to assess incontinence status. At weeks and 1 year postpartum, participants were also interviewed to complete the Incontinence Impact Questionnaire 7 (IIQ-7; see Outcome assessment section for a description). Outcome assessment The primary outcome assessment was the evaluation of whether GDM is an independent risk factor for postpartum urinary incontinence. The secondary outcome assessments were an evaluation of the severity of urinary incontinence during the postpartum follow-up period and the impact on QoL. The severity of incontinence was assessed using a validated two-question Sandvik Incontinence Severity Index. Incontinence was classified as slight (score 1 ), moderate ª 1 The Authors BJOG An International Journal of Obstetrics and Gynaecology ª 1 RCOG 1

3 Chuang et al. (score ), or severe (score ) based on composite scores. 19 Quality of Life was evaluated using a validated IIQ-7, which is a seven-item questionnaire that assesses four domains: physical activity, social relationships, travel and emotional health. It has a four-point scale: = not at all, 1 = slightly, = moderately, and = greatly; so a composite score could be computed, with a higher score indicating poorer QoL. Statistical analysis Descriptive statistics were used to compare participants with and without GDM. Data for women who became pregnant during the -year follow-up period were excluded from the analysis. The proportions of GDM and non-gdm participants with moderate/severe symptoms of postpartum urinary incontinence according to the Sandvik Incontinence Severity Index were compared using the z test. Data obtained from the IIQ-7 QoL questionnaire were analysed using the chi-square test for trend to identify trends of functional impairment across GDM (insulin treatment), GDM (conservative treatment), and non-gdm groups in all domains of the IIQ-7. We used a logistic generalised estimating equation method to handle correlated data. 1 To reflect the natural course of pregnancy from the prepartum to intrapartum to postpartum stages, three models were serially constructed: model 1 (only introduces prepartum factors), model (comprises factors in model 1 plus intrapartum factors), and model (comprises factors in model plus postpartum factors; this is the full model). The model was adjusted for incontinence status during the index pregnancy as a confounder. Additionally, because the mode of delivery and postpartum time interval are well established as factors that affect postpartum urinary incontinence, we also tested the interaction effects of GDM with mode of delivery, GDM with time (postpartum duration in months), and mode of delivery with time (postpartum duration in months). An autoregressive structure was specified for intrasubject correlations because correlations were assumed to be an exponential function of the time lag. Data were analysed using STATA version 11. for Windows (STATA Corporation, College Station, TX, USA). Results Description of study cohort Of the 91 participants initially screened and interviewed, were eligible for inclusion in the final analysis, of whom 9 had complete and had incomplete followup data. Table 1 shows the demographic characteristics and clinical parameters of the participants. The characteristics of included and excluded participants were similar, except for the distribution of parity (see Table S1); however, several variables differed between the eligible GDM and non-gdm participants. These variables were adjusted in the data analysis. Longitudinal patterns for type-specific urinary incontinence Of the participants with complete follow up (n = 9), there were in the GDM group and 7 in the non- GDM group. In the third-trimester assessment, there were (79.%) women without urinary incontinence, 9 (1.9%) with SUI, 17 (.%) with UUI, and (1.1%) with MUI in the GDM group, whereas in the non-gdm group there were 7 (.%) women without urinary incontinence, 71 (1.%) with SUI, 7 (1.%) with UUI, and 9 (.%) with MUI. At weeks postpartum, there were 7 (11.9%) de novo cases of SUI, (.%) de novo cases of UUI, and 1 (.7%) de novo cases of MUI in the GDM group, whereas in the non-gdm group there were 7 (.%) de novo cases of SUI, 91 (1.%) de novo cases of UUI, and 1 (.%) de novo cases of MUI. Further followup data are shown in Figure 1. Unadjusted and mutually adjusted generalised estimating equation results As shown in Table, GDM remained a significant influential factor for all three types of urinary incontinence (odds ratio [9% confidence interval]: 1.97 [1..1] for SUI,.11 [.1.] for UUI, and.7 [1.7.] for MUI) in the full model analysis (model ), indicating that women with GDM had a higher likelihood of developing postpartum urinary incontinence than women without GDM, after adjustment for prepartum, intrapartum and postpartum factors. The significance of the interaction terms of GDM with mode of delivery, GDM with length of postpartum period, and mode of delivery with length of postpartum period was negligible. The full model analysis further identified that, in addition to GDM, parity and the mode of delivery were also independent risk factors for the occurrence of all types of urinary incontinence. Evaluation of severity of urinary incontinence For SUI, the proportions of moderate/severe urinary incontinence were higher in women with GDM than in those without GDM at all four time-points postpartum. For both UUI and MUI, significant differences were observed at weeks and months postpartum, but at 1 and years postpartum, similar values were observed. To summarise, women with GDM tended to exhibit more severe symptoms of SUI for longer durations (up to years postpartum), whereas for UUI and MUI, more severe symptoms were only found for months postpartum (Figure ). 1 ª 1 The Authors BJOG An International Journal of Obstetrics and Gynaecology ª 1 RCOG

4 Gestational diabetes mellitus and urinary incontinence Table 1. Demographic and clinical characteristics of study participants, 7 (total n = ) GDM (n = ) Non-GDM (n = 7) P-value n % n % Age (years) Mean (SD).1 (.). (.) <.1 < BMI (kg/m ) <1. 1. < Parity < Education level High school. 1. <.1 Undergraduate.7 7. Postgraduate Employment status Unemployed <.1 Daytime only Night shift.. Smoking Current smoker Ex-smoker. 9. Nonsmoker Pregnancy-induced hypertension Yes <.1 No Mode of delivery Vaginal delivery <.1 Caesarean section Birthweight (g) < < Breastfeeding Yes No Postpartum PFMT Yes No BMI, body mass index; PFMT, pelvic floor muscle training; SD, standard deviation. All P-values are two-sided. Recorded at last prenatal check-up. Weight (kg)/height (m). Vaginal deliveries included spontaneous and instrumental deliveries. Including exclusive and almost exclusive breastfeeding. Performing 1 intensive contractions of the pelvic floor muscle at home twice a day for at least months postpartum. ª 1 The Authors BJOG An International Journal of Obstetrics and Gynaecology ª 1 RCOG 17

5 Chuang et al. 91 initially screened analyzed GDM (+) = GDM ( ) = 7 Excluded - Multiple gestation (n = ) - Known type 1 or type DM (n = 1) - Miscellaneous (n = 7) - Incomplete follow-up (n = ) - Complete follow-up (n = 9) GDM (+) n = GDM ( ) n = 7 Normal SUI UUI MUI MUI UUI SUI Normal rd-t Delivery Normal SUI UUI MUI MUI UUI SUI Normal weeks months a 1 year b years Figure 1. Flow chart indicating the distributions of type-specific incontinence in study participants at each scheduled time-point. Numbers in black denote the number of participants without urinary incontinence (without disease), whereas numbers in red denote the number of participants with urinary incontinence (with disease). a One participant developed de novo stress urinary incontinence (SUI) 1 year postpartum. b Two participants developed de novo SUI 1 year postpartum. DM, diabetes mellitus; GDM, gestational diabetes mellitus; MUI, mixed urinary incontinence; UUI, urge urinary incontinence. Assessment of QoL The proportions of moderate/severe functional impairment for each domain of IIQ-7 at weeks and 1 year postpartum are shown in Figure for GDM and non-gdm participants with urinary incontinence. Trend tests revealed that functional impairment was most severe in women with GDM requiring insulin treatment, less severe in women with GDM requiring conservative treatment only, and least severe in women without GDM in all four domains of the IIQ-7 questionnaire. Discussion In this hospital-based repeated-measure cohort study of pregnant women, three novel findings were obtained. First, women with GDM had a higher likelihood of developing all three types of urinary incontinence. Second, at the end of the follow-up period, at years postpartum, there were still higher proportions of moderate/severe symptoms of SUI in women diagnosed with GDM (Figure ), indicating that the impact of GDM on postpartum genitourinary function was relatively prolonged in these women, which contradicts the common belief that the effects of GDM usually vanish soon after delivery. Third, QoL was generally poorer among women diagnosed with GDM in the postpartum period (Figure ). Although studies in the current literature demonstrate positive relationships between pregnancy and urinary incontinence, these findings may be affected by two 1 ª 1 The Authors BJOG An International Journal of Obstetrics and Gynaecology ª 1 RCOG

6 Gestational diabetes mellitus and urinary incontinence Table. Unadjusted and mutually adjusted risks for each type-specific UI (n = ) Characteristics SUI UUI MUI Unadjusted OR (9% CI) Mutually adjusted OR (9% CI) Unadjusted Mutually adjusted OR (9% CI) Unadjusted Mutually adjusted OR (9% CI) OR (9% CI) OR (9% CI) Model 1 Model Model Model 1 Model Model Model 1 Model Model GDM No Yes (1..1) (1.9.) (1..) (1..1) (..) (1.9.99) (..1) (.1.) (1..) (1..9) (1.7.) (1.7.) Prepartum factors Age (1. 1.7) (1. 1.) (1. 1.7) (1. 1.7) (1. 1.7) (.9 1.) (.9 1.) (.9 1.) (.9 1.) (.9 ) (.9 1.) (.9 1.) BMI (.9 1.) (1. 1.) (1. 1.) (1. 1.) ( 1.) (.99 ) (.99 ) (.99 ) (.99 1.) (.9 1.) (.99 1.) (.99 1.) Smoking No Yes (.9.) (. 1.9) (.1.) (..) (.7.9) (..9) (..) (..9) (. 9.9) (.7.) (..) (..7) PIH No Yes ( ) ( 1.97) (1.7.) (1.7.) (.7.7) (.7.1) (.1.) (.7.) (. 1.) (. 1.) (.7 1.7) (. 1.9) Parity ( ) (1.7 1.) (1. 1.) (1. 1.7) ( 1.7) ( ) ( ) ( ) (1..1) (1..1) (1..7) (1..1) (1..) (1.1 1.) (1.1 1.) (1.1 1.) (1..) (1..) (1.1.1) (1.1.) (1.7.) (1.7.7) (1..7) (1.9.7) Intrapartum factors Mode of delivery CS VD (1.9.) (.1.) (..) (.7 1.) (.9 1.7) ( ) (1..) (1.1.) (1..) Neonatal birthweight < g g (1.1.1) (. 1.7) (.7 1.) (.9.9) (..7) (.7.) (.7) (.7.7) (.7.7) GDM mode of delivery (.9 1.) (. 1.) (. 1.9) (.19 1.) (.9 1.) (. 1.1) Postpartum factors Breast feeding No Yes (1. 1.7) ( ) ( ) (1. 1.9) (.7 1.) ª 1 The Authors BJOG An International Journal of Obstetrics and Gynaecology ª 1 RCOG 19

7 Chuang et al. Table. (Continued) Characteristics SUI UUI MUI Mutually adjusted OR (9% CI) Unadjusted Mutually adjusted OR (9% CI) Unadjusted Mutually adjusted OR (9% CI) OR (9% CI) OR (9% CI) Model 1 Model Model Model 1 Model Model Model 1 Model Model Unadjusted OR (9% CI) Postpartum PFMT No Yes (1. 1.1) (.7 1.) (.7 1.) (.1 ) (. 1.1) BMI, body mass index; CS, caesarean section; PIH, pregnancy induced hypertension; VD, vaginal delivery (including spontaneous vaginal delivery and instrumental delivery). Model 1, adjusted for GDM plus prepartum factors only. Model, adjusted for GDM plus prepartum and intrapartum factors (including interaction term GDM Mode of delivery). Model, adjusted for GDM plus prepartum, intrapartum (including interaction term GDM Mode of delivery), and postpartum factors (full model). BMI was calculated based on body weight and height collected on the day of admission for delivery. Combine current smoker and ex smoker. Parity in the index pregnancy. Therefore denotes nulliparity in the index pregnancy. potential problems. First, the postpartum follow-up period is relatively short in these studies, limited usually to months or sometimes to 1 months, leading to potentially incomplete evaluation of postpartum factors. Second, only two studies focused on type-specific urinary incontinence, 11,1 whereas the others considered all types of urinary incontinence together, hindering application of the results to specific types of urinary incontinence. Also, although a previous report revealed a positive relationship between GDM and urinary incontinence, it only addressed stress incontinence, and was cross-sectional in design without a control group. 1 A cross-sectional study can only demonstrate a relationship between a prediction variable and an outcome variable. In contrast, the current study expanded the outcome measures to all type-specific urinary incontinence, and had a prospective, longitudinal design; hence, the present findings lend further support to the hypothesis that there is an association between GDM and postpartum urinary incontinence. Although the underlying pathophysiological aetiologies linking GDM and postpartum urinary incontinence cannot be elucidated from the results of this study, some evidence can be extracted from previous studies investigating type diabetes mellitus and urinary incontinence. Factors potentially contributing to urinary incontinence in diabetes mellitus include microvascular damage, insulin injection and disease duration., Furthermore, diabetic bladder dysfunction has been linked to aetiologies of urinary incontinence. The early phase of diabetic bladder dysfunction presents as storage problems, such as urgency and urge incontinence, whereas the late phase manifests as voiding problems, leading to high residual urine and overflow incontinence. Further stratified by type of urinary incontinence, associations between type diabetes mellitus and type-specific urinary incontinence can be scenario dependent. In one study, type diabetes mellitus was associated with UUI, but not SUI or MUI, whereas in another study, type diabetes mellitus was associated with MUI, but not SUI or UUI., In contrast, GDM was associated with all types of urinary incontinence in the current study, with the highest likelihood for MUI in the full model assessment. Plausible explanations for the contradictions between the results of our study and those of the previous studies are that the current study had more confounding variables in addition to the diabetes factor: prepartum, intrapartum and postpartum factors. In addition to these factors, varying age distributions and different ethnic backgrounds are also potential contributing factors. In the current work, in addition to GDM, parity and mode of delivery were also found to be important risk factors for all types of urinary incontinence, which is relatively consistent with published results.,7 Moreover, age, 1 ª 1 The Authors BJOG An International Journal of Obstetrics and Gynaecology ª 1 RCOG

8 Gestational diabetes mellitus and urinary incontinence % SUI GDM (+) GDM ( ) rd-t weeks months 1 year years UUI MUI Figure. The proportions of moderate/severe symptoms evaluated using the Sandvik Incontinence Severity Index for type-specific incontinence among women with gestational diabetes mellitus (GDM) and women without GDM. rd-t, third trimester; MUI, mixed urinary incontinence; SUI, stress urinary incontinence; UUI, urge urinary incontinence. % Physical activity GDM (+, insulin treatment) GDM (+, conservative treatment) GDM ( ) Social relationship Travel Emotional health weeks 1 year Figure. The proportions of women with moderate/severe functional impairment according to a quality of life (QoL) assessment at weeks and 1 year postpartum. Participants with gestational diabetes mellitus (GDM) were dichotomised into insulin treatment and conservative treatment subgroups. A trend test revealed a steady decline of functional impairment for all four domains (P <., by chi-square test for trend). body mass index and pregnancy-induced hypertension were associated with SUI. Although age and body mass index are well-established risk factors for SUI,,,9 we could not find any association of pregnancy-induced hypertension and SUI in the literature. However, a-methyldopa, a centrally acting sympatholytic commonly used for the treatment of essential hypertension, especially pregnancy-induced hypertension, can cause decreased urethral pressure. Nearly % (7/1) of participants who had pregnancy-induced hypertension had ever used a-methyldopa in the current study. Whether the observed association between pregnancy-induced hypertension and SUI was attributable to an effect of a-methyldopa-induced decreased urethral pressure needs to be further investigated. We also noted that women who breastfed had a higher likelihood of experiencing postpartum UUI. Although breastfeeding is known to induce uterine contraction postpartum, we could not find a correlation between breastfeeding and UUI in the published literature. Nonetheless, common parasympathetic innervations from sacral roots to the bladder detrusor muscle and uterus may plausibly predispose lactating women to concurrent uterine and detrusor contraction, resulting in a higher likelihood of UUI. 9 ª 1 The Authors BJOG An International Journal of Obstetrics and Gynaecology ª 1 RCOG 11

9 Chuang et al. The current study has several limitations. First, the monocentric design may potentially result in an over-represention of high-risk cases with resultant selection bias. Second, blood sugar levels were not monitored in the postpartum period, with the consequence that diagnoses of type diabetes mellitus postpartum could have been missed. A meta-analysis indicated that women with GDM had a 7.7-fold higher risk of developing type diabetes mellitus postpartum compared with those who experienced a normoglycaemic pregnancy. Whether the increased prevalence of SUI in women with GDM in this study was attributable to type diabetes mellitus postpartum should be explored further. Third, weight gain or loss, which is an important factor for urinary incontinence, was not recorded in the gestational or postpartum period. 1 Although our findings concur with those of Kim et al. 1 in that GDM was associated with SUI, both studies lacked adequate postpartum body weight recording. We cannot explain why the effect of GDM on SUI disappears with time; however, precise recording of postpartum body weight, treated as a time-varying covariate, may elucidate the effect of changes in postpartum body weight on SUI. Fourth, urinary incontinence is a medical, psychological, social, economic and hygienic problem. Many relevant factors, such as depression, socio-economic status, and type of caesarean section (elective or emergent), were not measured, which may have biased the results. Perinatal morbidities induced by GDM (e.g. giant baby and shoulder dystocia) are well established in the literature; the current work adds the novel finding of a higher likelihood of postpartum urinary incontinence induced by GDM, although the disease per se usually resolves soon after delivery. Care providers should offer targeted education to women with GDM. Moreover, women with GDM are strongly recommended to have follow-up blood sugar tests in the postpartum period to explore the relationships between type diabetes mellitus and urinary incontinence, in view of the higher risk of progression to type diabetes mellitus in these women. The epidemiological value of the current research lies in the establishment of a positive relationship between GDM and urinary incontinence, in addition to the established positive relationships between diabetes mellitus and urinary incontinence, and between gestation/delivery and urinary incontinence. This finding suggests that diabetes and gestation/delivery may have additive or synergistic effects on postpartum urinary incontinence. However, the underlying biological pathway was not investigated in the current research, and merits further in-depth study. Disclosure of interests The authors have no conflict of interest with regard to any drugs or materials relevant to this paper. Contribution to authorship CMC planned the study, collected and analysed the data, interpreted the results and wrote the paper. IFL interpreted the results and revised the manuscript. HCH, YHH ILS and HCP provided additional results for the study. PC abstracted data, assessed the quality of studies and revised the manuscript. Details of ethics approval This study was approved by the Institutional Review Board of Taipei Veterans General Hospital (VGHIRB.: 11IC). Funding This work was supported by the Taiwan National Science Council (99-1-B-7-) and National Health Research Institute (PS9-1). Acknowledgements We are indebted to the research panel: Drs Lilly Wen (Catholic Cardinal Tien Hospital, Taipei), Nan-Ni Chen (Municipal Hospital for Women and Children, Taipei), Kuo-Chang Wen, Wei-Min Hu, Hsiao-Wen Tsai, Pi-Lin Sung, Wei-Lun Hsu, Chih-Yu Chen, Jen-Yu Huang, Chia- Ming Chang, Peng-Hui Wang, Nae-Fong Twu, Hsiang-Tai Chao and Kuan-Chung Chao (Taipei Veterans General Hospital, Taipei). Supporting Information Additional Supporting Information may be found in the online version of this article: Table S1. Comparison of demographic and clinical characteristics of included and excluded study participants. Please note: Wiley-Blackwell are not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author. j References 1 Wetle T, Scherr P, Branch LG, Resnick NM, Harris T, Evans D, et al. Difficulty with holding urine among older persons in a geographically defined community: prevalence and correlates. J Am Geriatr Soc 199;:9. Hunskaar S, Arnold EP, Burgio K, Diokno AC, Herzog AR, Mallett VT. Epidemiology and natural history of urinary incontinence. Int Urogynecol J ;11:1 19. Monz B, Chartier-Kastler E, Hampel C, Samsioe G, Hunskaar S, Espuna-Pons M. Patient characteristics associated with quality of life in European women seeking treatment for urinary incontinence: results from PURE. Eur Urol 7;1:17. Tuomelihto J, Lindstrom J, Eriksson JG, Valle TT, Hamalainen H, Ilanne-Parikka P. Prevention of type diabetes mellitus by changes 1 ª 1 The Authors BJOG An International Journal of Obstetrics and Gynaecology ª 1 RCOG

10 Gestational diabetes mellitus and urinary incontinence in life style among subjects with impaired glucose tolerance. N Engl J Med 1;:1. Lifford KL, Curban GC, Hu FB, Barbieri RL, Grodstein F. Type diabetes mellitus and risk of developing urinary incontinence. J Am Geriatr Soc ;:11 7. Rortveit G, Hannestad YS, Daltveit AK, Hunskaar S. Age and typedependent effects of parity on urinary incontinence: the Norwegian EPICONT study. Obstet Gynecol 1;9: Gasquet I, Tcherny-Lessenot S, Gaudebout P, Bosio Le Goux B, Klein P, Haab F. Influence of the severity of stress urinary incontinence on quality of life, health care seeking, and treatment: a national crosssectional survey. Eur Urol ;:1. Diez-Itza I, Arrue M, Ibanez L, Murgiondo A, Paredes J, Sarasqueta C. Factors involved in stress urinary incontinence 1 year after first delivery. Int Urogynecol J 1;1:9. 9 Farrell SA, Allen VM, Baskett TF. Parturition and urinary incontinence in primiparas. Obstet Gynecol 1;97:. 1 Solans-Domenech M, Sanchez E, Espuaa-Pons M; Pelvic Floor Research Group (Grup de Pecerca del Sol Pelvia; GRESP). Urinary and anal incontinence during pregnancy and postpartum: incidence, severity, and risk factors. Obstet Gynecol 1;11:1. 11 Wesnes SL, Hunskaar S, Bo K, Rortveit G. Urinary incontinence and weight change during pregnancy and postpartum: a cohort study. Am J Epidemiol 1;17:1. 1 Arrue M, Ibanez L, Paredes J, Murgiondo A, Belar M, Sarasqueta C, et al. Stress urinary incontinence six months after first vaginal delivery. Eur J Obstet Gynecol Reprod Biol 1;1: Burgio KL, Zyczynski H, Locher JL, Richter HE, Redden DT, Wright KC. Urinary incontinence in the 1-month postpartum period. Obstet Gynecol ;1: Buchana TA, Xiang AH. Gestational diabetes mellitus. J Clin Invest ;11: Kim C, Mcewen LN, Sarma AV, Piette JD, Herman WH. Stress urinary incontinence in women with a history of gestational diabetes mellitus. J Womens Health ;17: National Diabetes Data Group. Classification and diagnosis of diabetes mellitus and other categories of glucose intolerance. Diabetes 1979;: Abrams P, Cardozo L, Fall M, Griffiths D, Rosier P, Ulmsten U. The standardization of terminology of lower urinary tract function: report from the standardization Sub-committee of the International Continence Society. Am J Obstet Gynecol ;17:11. 1 Juang CM, Yu KJ, Chou P, Yen MS, Twu NF, Horng HC, et al. Efficacy analysis of trans-obturator tension-free vaginal tape (TVT-O) plus modified Ingelman Sundberg procedure versus TVT-O alone in the treatment of mixed urinary incontinence: a randomized study. Eur Urol 7;1: Sandvik H, Seim A, Vanvik A, Hunskaar S. A severity index for epidemiological surveys of female urinary incontinence: comparison with -hour pad-weighing tests. Neurourol Urodyn ;19:17. Uebersax JS, Wyman JF, Shumaker SA, McClish DK, Fantl JA. Short forms to assess life quality and symptom distress for urinary incontinence in women: the Incontinence Impact Questionnaire and the Urogenital Distress Inventory. Neurourol Urodyn 199;1: Zeger SL, Liang KY, Albert PS. Models for longitudinal data: a generalized estimating equation approach. Biometrics 19;:19. Brown JS, Wessells H, Chancellor MB, Howards S, Stamm WE, Stapleton AE. Urologic complications of diabetes. Diabetes Care ;:177. Jackson SL, Scholes D, Boyko EJ, Abraham L, Fihn SD. Urinary incontinence and diabetes in postmenopausal women. Diabetes Care ;:17. Daneshgari F, Liu G, Birder G, Hanna-Mitchell AT, Chacko S. Diabetic bladder dysfunction: current translational knowledge. J Urol 9;1(Suppl ):S1. Danforth KN, Townsend MK, Curhan GC, Resnick NM, Grodstein F. Type diabetes and risk of stress, urge, and mixed urinary incontinence. J Urol 9;11:19 7. Waetjen LE, Liao S, Johnson WO, Sampselle CM, Sternfield B, Harlow SD, et al. Factors associated with prevalent and incident urinary incontinence in a cohort of midlife women: a longitudinal analysis of data: study of women s health across the nation. Am J Epidemiol 7;1: Parazzini F, Chiaffarino F, Lavezzari M, Giambanco V. Risk factors for stress, urge or mixed urinary incontinence in Italy. BJOG ;11:97. Fine PM, Antonini TG, Appell RA. Clinical evaluation of women with lower urinary tract dysfunction. Clin Obstet Gynecol ;1:. 9 De Groat WC. Integrative control of the lower urinary tract: preclinical perspective. Br J Pharmacol ;17(Suppl ):S. Bellamy L, Casas JP, Hingorani AD, Williams D. Type diabetes mellitus after gestational diabetes: a systemic review and meta-analysis. Lancet 9;7: Subak LL, Wing R, West DS, Franklin F, Vittinghoff E, Creasman JM, et al. Weight loss to treat urinary incontinence in overweight and obese women. N Engl J Med 9;:1 9. Santacreu M, Fernandez-Ballesteros R. Evaluation of a behavioral treatment for female urinary incontinence. Clin Interv Aging 11;:1 9. ª 1 The Authors BJOG An International Journal of Obstetrics and Gynaecology ª 1 RCOG 1