Journal Pre-proof Severity and impact of accidental bowel leakage two decades after no, one and two sphincter injuries

tract symptoms was observed ( P .73 and P .69).

this study. 14,20,21 AI was defined as isolated gas incontinence (IGI) or FI with or without co-score, the sum ranging from 0 (continent) to 20 (complete incontinence) and here further 179 divided into mild (1-3), moderate (4-8) and severe (≥9). 14 Symptom specific bother of AI, FI, 180 and IGI was dichotomized into Not bothersome (=No problem and A minor nuisance), and 181 Bothersome (=Some bother, Much bother, and A major problem). Urinary incontinence (UI), 182 subtypes of UI (stress, urgency, and mixed UI), as well as overactive bladder (OAB) were 183 defined according to the IUGA/ICS definition. 22 Symptoms of pelvic organ prolapse (sPOP) 184 were defined according to IUGA/ICS by the question "Do you have a sensation of tissue 185 protrusion (a vaginal bulge) from your vagina?", 21 between the cohorts assumed that women without OASI were least affected, followed by 202 Logistic regression analysis was used to compare the BMI and age-adjusted odds ratio (aOR) 204 for outcomes with 95% CI and P value. To estimate the degree of bother according to the 205 Jorge-Wexner score, four separate logistic regression models were used. All statistical testing 206 was two-sided, and the significance level was set to P<.05 and was performed using SAS 9.4 207 (SAS Institute Inc., Cary, NC, USA). A power calculation was performed for the primary 208 study and has been published in a previous report. 17 209 210

Cohort characteristics 212
Selection of the study population, cohort characteristics, response rates, and use of the web-213 form have been published previously. 17 Women with no OASI were somewhat younger, 214 whereas infant birth weight ≥ 4kg was higher in women with OASIs. Vacuum extraction rates 215 were doubled and tripled in women with one and two OASIs than without OASI (Trends P< 216 .0001). 17 The overall rate of missing data was 1.6%, lower in women with two OASIs (0.8%). 217 218 Impact and severity 219 The severity of incontinence increased after one and two OASIs for all components of 220 leakage, i.e., solid and liquid stool and concomitant gas. High frequency of solid stool leakage 221 was 1.4%, 2.3%, and 4.6% after no, one and two OASIs and 2.7%, 6.4%, and 10.5% for 222 liquid stool (Trends P<.0001, Supplemental Table S1; Figure 1). Bothersome incontinence 223 increased 3-fold and 5-fold for FI from 3.3% in women without OASI to 10.4% in those with 224 one OASI (aOR, 3.25; 95% CI, 2.23-4.73) and 16.5 % with two OASIs (aOR, 5.16; 95% CI  Figure S1). The use of protective products/pads and the effect on daily life increased with one 231 and two OASIs. Any use of protective products/pads occurred in 2.3% (no OASI), 7.1% (1 232 OASI), and 8.4% (2 OASIs), and more frequently in women with one or two OASIs 233 compared to those without OASI (Trend P <.0001, Supplemental Table S1; Figure 3, Figure  234 4). Any impact of incontinence on daily life occurred in 8.6% (no OASI), 19.7% (one OASI), 235 and 29.6% (two OASIs), and more frequently in those with one or two OASIs (Trend P 236 <.0001, Supplemental Table S1, Figure 3, Figure 4). 237 The mean Jorge-Wexner score was 1.27 (SD 1.97), 2.25 (SD 2.90), and 2.98 (SD 3.04) with 238 0, 1, and 2 OASIs, respectively (Trend P<.0001) (Supplemental Table S3 and Figure S2). Out 239 of all the women in this study with a score of six, more than 50% perceived "some bother", 240 and more than 50% of those with a score of 12 had "much bother" ( Figure 5). Few women 241 had received any treatment for incontinence, yet the rate increased from 1.3% to 3.4%, and 242 4.1% with 0, 1, and 2 OASIs (Trend P <.0001) (Supplemental Table S1, Figure 3, Figure 4). 243

244
The overall effect of one versus two OASIs on measures of anal incontinence was 2.19 (95% 245 CI 1.68-2.85, P<.0001) for one OASI and 3.91 (95% CI 3.06-5.00, P<.0001) for two OASI 246 OASIs were compared with 1384 women with two vaginal births and an OASI at the first 294 birth. All women were secundiparous with no further births and were followed for ~8 years 295 from the last birth. The condition-specific impact on quality of life ("a lot" and "some" 296 bother) was higher after two OASIs (34.9%, 37/106) versus one (24.2%, 335/1384). Still, the 297 difference was not significant in an adjusted regression analysis. 28 The findings of this study 298 are unique in that it evaluated the 20-year clinical outcomes of one and two OASIs in 299 comparison with that in two parous women without OASI of the severity and impact of ABL. supports this approach, which has led to a preference for the Jorge-Wexner score, 306 supplemented with e few additional items, depending on the target group at hand. 15 307 308

Clinical implications 309
Avoiding sphincter injury at vaginal delivery would most probably significantly reduce the 310 severity and impact of FI and AI. This information, that bowel incontinence after one or two 311 OASIs was more prevalent and more severe should be communicated to obstetricians, 312 midwives, and pregnant women as part of the antenatal consultation. This calls for a critical 313 attitude toward current clinical practices and pursuing the best preventive strategy to avoid 314 OASIs. 315

Research implications 317
In the future, efforts to avoid OASIs have to involve the development of an individualized 318 risk prediction instrument based on high-quality national medical birth registers that can be 319 used by pregnant women antenatally and guide obstetricians and midwives. 320 321

Strengths 322
The main strengths of this study were that all women were strictly 2-para, retrieved from a 323 high-quality national register, making the control group the theoretically most appropriate 324 cohort for analyses. 29 The first cohort (1992)(1993)(1994)(1995)(1996)(1997)(1998)      J o u r n a l P r e -p r o o f J o u r n a l P r e -p r o o f