Gillian Hanley

No abstract available.

No abstract available.

Background:Antibiotics are among the most used medications during pregnancy, and while their short-term benefits are clear, their potential long-term effects are underexplored. Antibiotic exposure induces changes in the maternal microbiome, which could in turn influence the development of the infant’s microbiome with implications for neurodevelopment. This thesis examined the associations of 1) intrapartum antibiotic exposure and 2) prenatal antibiotic exposure with autism spectrum disorder (ASD) in offspring.Methods:Using multiple linked population-level datasets, we examined everyone who delivered a live singleton term infant in British Columbia, Canada between April 1st 2000 and December 31st 2014. To examine the associations among pregnant individuals treated for the same indication, we studied sub-cohorts of those diagnosed 1) as group B streptococcus-positive, and 2) with urinary tract infections. Cox proportional hazards models were used to estimate unadjusted and adjusted hazard ratios. Sensitivity analyses were conducted to examine the impact of different classes of antibiotics, and to assess for dose-response and temporal relationships. Additionally, potential effect modification was examined based on sex and mode of delivery. To account for unmeasured confounding, we ran a conditional logistic regression of discordant sibling pairs.Results:We found no association between intrapartum antibiotic administration and ASD in offspring. In contrast, we observed a small statistically significant increase in risk of ASD associated with prenatal antibiotic exposure. In particular, highest risk was observed if exposure took place during the second trimester, to penicillins or other beta lactams, and for 15 days or longer. Among the cohort restricted to pregnant individuals diagnosed with UTIs, findings showed a positive association between prenatal antibiotics and ASD with an increased effect size; yet, this association was no longer statistically significant, possibly due to a reduced sample size. Therelationship between prenatal antibiotic exposure and ASD was no longer significant in the conditional logistic regression.Conclusions:Overall, our findings suggest that the risk of ASD should not factor into clinical decisions on whether to administer antibiotics both prenatally and during labour and delivery. Results may suggest a hypothesized role of impaired fetal metabolic programming by the maternal microbiome and metabolome in ASD development.

View record

BACKGROUND: More research is needed comparing post-surgical outcomes between patients who undergo hysterectomy for endometriosis with or without ovarian conservation. OBJECTIVE: To compare the rate of reoperation and use of other pain-related health services after hysterectomy for endometriosis, with or without ovarian conservationMETHODS: A population-based retrospective cohort study of 4489 patients aged 19-50 in British Columbia, Canada, undergoing hysterectomy for endometriosis between 2001 and 2016. Index surgeries were classified as: hysterectomy alone (conservation of both ovaries), hysterectomy with unilateral salpingoophorectomy (USO), or hysterectomy with bilateral salpingoophorectomy (BSO). Reoperation rate was the primary outcome. Secondary outcomes (measured at 3-12 months and 1-5 years after hysterectomy) included: physician visits for endometriosis and pelvic pain, prescriptions filled for opioids and hormonal suppression medications and hormone replacement therapy (HRT). RESULTS: 89.5% of patients remained reoperation free by the end of follow-up (median of 10 years, IQR = 6.1 to 14.3 years). Patients undergoing hysterectomy alone were more likely to undergo at least one reoperation compared to those having hysterectomy with BSO (13% vs 5%, p0.0001), most commonly oophorectomy and adhesiolysis. When oophorectomy as reoperation was removed in a sensitivity analysis, this difference was attenuated. Secondary outcomes including physician visits for endometriosis or pelvic pain and rates of opioid prescriptions filled were similar between groups. The rate of prescriptions filled for hormonal suppression medications was low for all groups. The rate of prescriptions filled for HRT after hysterectomy with BSO was suboptimal - 60.6% filling at least one prescription at 3-12 months after index surgery.CONCLUSION: Patients who underwent hysterectomy with BSO had a lower reoperation rate than those who had hysterectomy with conservation of one or both ovaries. However, there was little difference between the groups for the secondary outcomes measured, suggesting that persistent pelvic pain after hysterectomy for endometriosis may not differ significantly based on ovarian conservation status. Moreover, HRT use after hysterectomy with BSO was suboptimal, which may have significant health consequences for these individuals undergoing premature surgical menopause. Therefore, strong consideration should be given to ovarian conservation at the time of hysterectomy for endometriosis.

View record

Antenatal depression and anxiety (DEP-ANX) are highly prevalent conditions that have been associated with increased risk for adverse outcomes. Most studies to this point have employed simplistic definitions to describe preconception and antenatal DEP-ANX. This thesis had three aims: (1) use population-based data to operationalize variables describing mental health histories; (2) evaluate variables’ ability to predict adverse perinatal outcomes, postpartum DEP-ANX (PPD) and preterm birth (PTB); and (3) employ these variables to investigate the relationship between DEP-ANX and gestational diabetes mellitus (GDM). We examined mental health services use of all birth parents who delivered a live infant in British Columbia, Canada between April 1, 2000, and December 31, 2013, and were registered with provincial Medical Services Plan (MSP) for >100 days/year for the respective study period. We operationalized variables to proxy severity, persistence, and frequency of DEP-ANX from preconception through pregnancy, then constructed predictive regression models for PPD and PTB. Based on our findings, we ran unadjusted and adjusted logistic regression models to estimate odds ratios (ORs) for GDM given antenatal DEP-ANX. We then ran conditional logistic regression models that matched birth parents to themselves (in subsequent pregnancy) based on discordance of exposure and outcome. We found that PPD and PTB predictive models performed better with inclusion of more detailed mental health histories. Incorporating dichotomous DEP-ANX indicators across preconception markedly improved predictive power and model fit. Detailed measures of mental health service use predicted PPD better than PTB and the utility of individual measures depended on the outcome. We observed that those with antenatal DEP-ANX had significantly higher odds of GDM. Apparent risk for GDM given antenatal DEP-ANX was highest among those with no DEP-ANX history. Associations estimated by matched sibling analysis were non-significant, but effect size among the no history group was consistent with unmatched analysis. Overall, our findings suggest that more nuanced approaches to classify antenatal DEP-ANX within population-based data are feasible. By accounting for mental health treatment, mental health history, and current mental health, we can better control for differences in underlying conditions and better understand more complex associations between antenatal mental health and perinatal outcomes.

View record

Objectives: To prevent ovarian cancer among women at high genetic risk due to BRCA1/2 mutations, risk-reducing bilateral salpingo-oophorectomy (RRBSO) is recommended between 35-45 years. However, there is currently little evidence on long-term, non-cancer health outcomes among women who undergo this treatment. This thesis examined: 1) the risk of cardiovascular disease (CVD), of predisposing conditions to CVD, and the receipt of cardioprotective medications, and 2) the risk of bone fractures and osteoporosis, and the receipt of health services to prevent bone disease, following RRBSO among women with BRCA1/2 mutations.Methods: Using population-based data from British Columbia, Canada, between 1996 to 2017, I compared the hazard of CVD, and of fractures and osteoporosis among women with known deleterious BRCA1/2 mutations who underwent RRBSO before the age of 50 with two groups of age-matched women without known BRCA1/2 mutations: 1) women who underwent bilateral oophorectomy (BO) for benign gynecologic conditions; and, 2) women who underwent hysterectomy and/or salpingectomy but had intact ovaries.Results: Women with BRCA1/2 mutations were at increased risk for CVD compared to women without mutations and intact ovaries, and no significant increased risk for CVD when compared to women without mutations who underwent BO. Compared to both control groups, they were less likely to be diagnosed with predisposing conditions and to fill cardioprotective medications, although this did not reach statistical significance. Regarding bone health, women with BRCA1/2 mutations had a higher likelihood of having an osteoporosis diagnosis than women in both control groups, but they were not at an increased risk of fractures during the follow-up period. Although the likelihood of having a dual-energy x-ray absorptiometry scan was higher for women with BRCA1/2 mutations, under 50% of them were screened for bone loss by the end of follow-up. The rate of bisphosphonates use among women with osteoporosis was also low. Conclusions: These results suggest that the cardiovascular and bone health of women with BRCA1/2 mutations is worse than age-matched women with intact ovaries. We recommend that women with BRCA1/2 mutations receive focused care after RRBSO in order to prevent CVD, fractures and osteoporosis.

View record

Objectives. Epithelial Ovarian cancer (EOC) is composed of five distinct histologic subtypes. However, histotype- specific survival and mortality estimates among women with EOC are limited. Also, we lack information on long-term health conditions faced by ovarian cancer survivors. This thesis examined: 1) histotype- specific incidence and survival rates among women with EOC in British Columbia (BC); 2) causes of death among women with EOC by histotype; and, 3) compared these causes of death to age-standardized causes of death in the general population. Methods. Using population-based administrative datasets, I built two population-based cohorts of all women with EOC diagnosed in BC: 1. women diagnosed between 1980 and 2015(cohort 1) and women diagnosed between 1990 and 2014 (cohort 2). Cohort 1 was used to answer question 1, whereas cohort 2 was used to answer question 2 and 3. For question 2, I compared the causes of death within histotypes, by age at diagnosis, BRCA status, and time since diagnosis. For question 3, I calculated all-cause and cause-specific standardized mortality ratios. Results. Decreasing incidence rates and increasing survival rates were observed among women with EOC in BC. As expected, ovarian cancer was the most common cause of death among these women, which was first surpassed by other causes of deaths at 11 years after diagnosis. When stratified by serous and non- serous EOCs, ovarian cancer was the leading cause of death for 12 years and for 8 years respectively after diagnosis. Of particular interest, the number of deaths from other cancers (breast, colorectal and lung cancer) and external causes (falls) among long-term ovarian cancer survivors (5-9 and 10+ years post diagnosis) were higher than the expected. Conclusions. Although there was an improvement in the survival over time, ovarian cancer remains the leading cause of death for 11 years following diagnosis of EOC. My findings suggest that long-term survivors (those living 5-9 and 10+ years following diagnosis) are particularly vulnerable to deaths from other cancers and from falls in elderly survivors. Hence, these women may benefit from closer surveillance of other cancers and bone health.

View record