Selected Conference Proceedings
15th Annual Epidemiology, Biostatistics and Occupational Health Student Research Day


Message from the Epidemiology, Biostatistics and Occupational Health Student Society

Selected abstracts from the 15th Annual Epidemiology, Biostatistics and Occupational Health (EBOH) Research Day. The event was held on March 15, 2019 at the Montreal Neurological Institute, and was led by the Epidemiology, Biostatistics and Occupational Health Student Society. Research Day celebrates the research done by students in EBOH at McGill University. Special thanks to the abstract reviewers, presentation judges, administrative staff, faculty donors, and organizing committee, without whom this event would not have been possible.


Cardiotoxicity of Aromatase Inhibitors in Post-Menopausal Women with Breast Cancer: A Population-Based Cohort Study

Farzin Khosrow-Khavar1,2, Nathaniel Bouganim 3,4, Kristian B. Filion1,2,5, Samy Suissa12, Laurent Azoulay1,2,3
1Centre for Clinical Epidemiology, Lady Davis Institute, Jewish General Hospital, Montreal, Canada
2Department of Epidemiology, Biostatistics, and Occupational Health, McGill University, Montreal, Canada
3Gerald Bronfman Department of Oncology, McGill University, Montreal, Canada
4Department of Oncology, Cedar Cancer Center, McGill University Health Center, Montreal, Canada
5Division of Clinical Epidemiology, Department of Medicine, McGill University, Montreal, Canada
Oral Presentation

Background: Aromatase inhibitors (AIs) are widely used in treatment of estrogen-receptor positive breast cancer. However, randomized controlled trials have suggested that these drugs may be associated with adverse cardiovascular events. To examine this safety concern, we conducted a study in the real-world setting to determine whether AIs, in comparison with tamoxifen, are associated with increased risk of myocardial infarction, ischemic stroke, congestive heart failure, and cardiovascular death.

Methods: We conducted a population-based cohort study using the United Kingdom Clinical Practice Research Datalink linked to Hospital Episode Statistics hospitalization data and Office for National Statistics vital statistics. The study population consisted of post-menopausal women newly-diagnosed with breast cancer and newly-treated with either AIs or tamoxifen between April 1, 1998 and February 29, 2016. Cox proportional hazards model with inverse probability of treatment and censoring weights were used to estimate hazard ratios (HRs) with 95% confidence intervals (CIs) comparing AIs with tamoxifen for each of the study outcomes.

Results: The study cohort included 8,139 and 9,783 patients newly-treated with AIs and tamoxifen, respectively. The patients were followed for a median of 1.4 (SD: 1.7) years. Compared with tamoxifen, AIs were associated with a trend toward an increased risk of myocardial infarction (HR 1.38; 95% CI: 0.93-2.07), ischemic stroke (HR 1.23; 95% CI: 0.86-1.76), congestive heart failure (HR: 1.29; 95% CI: 0.79-2.12), and a significantly increased risk of cardiovascular death (HR: 1.44; 95% CI: 1.06-1.95).

Conclusion: In this large real-world setting study, AIs were associated with increased cardiovascular mortality compared with tamoxifen. There were also trends towards increased, though non-significant, risks of myocardial infarction, ischemic stroke, and congestive heart failure with AIs compared with tamoxifen. Further research is needed to ascertain whether the observed increased risks are due to deleterious effects of AIs or cardioprotective effects of tamoxifen.







Interactions of child maltreatment and genetic variations in adult psychopathology: A systematic review

Muzi Li1,2, Sibei Liu3,4, Tingting Gao4, Xiangfei Meng1,2
1Department of Psychiatry, Faculty of Medicine, McGill University, Montreal, Canada
2Douglas Mental Health University Institute, Montreal, Canada
3Mitacs Globalink Internship, Canada
4School of Public Health, Jilin University, Changchun, Jilin Province, China
Oral Presentation

Background: Although studies have found that childhood maltreatment significantly predicts the risk of adulthood psychopathology, susceptible genetic variations interact with childhood maltreatment increasing the risk with psychopathology. There is an increasing need to have an overview of genetic variations that interplay with maltreatment to develop adulthood psychopathology. This review aimed to systematically synthesize the interactions between childhood maltreatment and genetic variations in adult psychopathology among those with the exposure of childhood maltreatment.

Methods: Electronic databases and grey literature up to May 31st, 2018 were searched for English language articles with the eligibility criteria for psychopathology and childhood maltreatment. Data extraction and methodological quality assessment on key characteristics of the included studies were conducted independently. Both quantitative and qualitative approaches were used to systematically synthesize the findings.

Results: The initial search resulted in 8,225 articles. A total of 26 articles meeting the eligibility criteria were included in this review. Findings of this review included several most studied genes (5-HTTLR, CRHR1, BDNF, CREB1, FKBP5, MAOA, IL1B, NTRK2, and OXTR). Overall, childhood maltreatment increased the risk of adult psychopathology and effects of maltreatment were moderated by genetic variations.

Conclusion: This review firstly provides a comprehensive overview of interactions between childhood maltreatment and genetic variations in adult psychopathology. Future etiological explorations should target the sites identified above.







Evaluation of the Equivalency of the PHQ-8 and PHQ-9: A systematic review and individual participant data meta-analysis

Yin Wu1,2,3, Brooke Levis1,3, Andrea Benedetti3,4,5, Brett D. Thombs1,2,3,4,6,7 and the DEPRESsion Screening Data (DEPRESSD) Collaboration
1Lady Davis Institute for Medical Research, Jewish General Hospital, Montreal, Canada
2Department of Psychiatry, McGill University, Montreal, Canada
3Department of Epidemiology, Biostatistics, and Occupational Health, McGill University, Montreal, Canada
4Department of Medicine, McGill University, Montreal, Canada
5Respiratory Epidemiology and Clinical Research Unit, McGill University Health Centre, Montreal, Canada
6Department of Educational and Counselling Psychology, McGill University, Montreal, Canada
7Department of Psychology, McGill University, Montreal, Canada
Oral Presentation

Background: Item 9 of the Patient Health Questionnaire-9 (PHQ-9) queries about thoughts of death and self-harm, but not suicidality. Although it is sometimes used to assess suicide risk, most positive responses are not associated with suicidality. The PHQ-8, which omits Item 9, is increasingly used. We assessed equivalency of total score correlations and the diagnostic accuracy of the PHQ-8 and PHQ-9 to detect major depression.

Methods: We conducted an individual patient data meta-analysis. We fit bivariate random-effects models to assess diagnostic accuracy.

Results: 16,742 participants (2,097 major depression cases) from 54 studies were included. The correlation between PHQ-8 and PHQ-9 scores was 0·996 (95% confidence interval 0·996 to 0·996). The standard cutoff score of 10 for the PHQ-9 maximized sensitivity + specificity for the PHQ-8 among studies that used a semi-structured diagnostic interview reference standard (N = 27). At cutoff 10, the PHQ-8 was less sensitive by 0·02 (-0·06 to 0·00) and more specific by 0·01 (0·00 to 0·01) among those studies (N = 27), with similar results for studies that used fully structured interviews (N = 13) or the Mini International Neuropsychiatric Interview (N = 14). For all 54 primary studies combined, across all cutoffs, the PHQ-8 was less sensitive than the PHQ-9 by 0·00 to 0·05 (0·03 at cutoff 10), and specificity was within 0·01 for all cutoffs (0·00 to 0·01).

Conclusion: PHQ-8 and PHQ-9 total scores were similar. Sensitivity may be minimally reduced with the PHQ-8, but specificity is similar.







Efficacy of a carrageenan-based lubricant gel in increasing clearance of HPV infections in women: interim analysis of a double-blind, randomized, placebo-controlled trial

Sindy Magnan1, Joseph E. Tota2, Mariam El-Zein1, Ann N Burchell3, John T Schiller4, Alex Ferenczy5, Pierre-Paul Tellier6, Francois Coutlee7, Eduardo L. Franco1
1McGill University, Division of Cancer Epidemiology, Montreal, Canada
2National Cancer Institute, Division of Cancer Epidemiology and Genetics, Rockville, USA
3Li Ka Shing Knowledge Institute, Department of Family and Community Medicine, Toronto, Canada
4National Cancer Institute, Department of Health and Human Services, Bethesda, USA
5McGill University, Department of Pathology, Montreal, Canada
6McGill University, Department of Family Medicine, Montreal, Canada
7Centre Hospitalier de l’Université de Montréal, Département de Microbiologie et infectiologie, Montréal, Canada
Oral Presentation

Background: To evaluate the efficacy of a carrageenan-based lubricant gel in reducing incidence and prevalence of genital human papillomavirus (HPV) infections among sexually active women.

Methods: Between January 2013 and June 2017, 280 women aged 18+ years were randomly assigned to a carrageenan (n=141) or a placebo (n=139) gel to be self-applied every other day for the first month and prior to and following each intercourse during follow-up. Assessments were done at baseline, 0.5, 1, 3, 6, 9 and 12 months. Sociodemographic, behavioral and sexual history data were collected using computer-assisted self-administered questionnaires. We used Roche’s Linear Array assay to detect and genotype 36 HPV types in self-collected vaginal samples. The first primary outcome (reported previously) was incidence of HPV infections. The second primary outcome was clearance of HPV infections observed at baseline. We estimated hazard ratios (HR) and 95% confidence intervals (CI) using Cox models.

Results: 67 (48%) of the 141 participants in the carrageenan and 80 (58%) of the 139 participants in the placebo arm were HPV positive at baseline. Baseline and follow-up characteristics were well balanced between arms. The median follow-up time was 9.2 months (interquartile range: 2.6-13.2). 36 (54%) participants in the carrageenan and 34 (43%) participants in the placebo arm became HPV negative during follow-up (HR:1.45; 95% CI:0.92-2.32). Although the Kaplan-Meir curves eventually converged around 12 months, a significant difference was observed in the first few months with a clearance of 33.2% (95% CI:22.5-47.1) in the carrageenan vs. 9.6% (95% CI:4.7-19.0) in the placebo arm at 3 months.

Conclusion: In our trial’s interim analysis, the use of a carrageenan-based gel was associated with an increase in clearance of HPV infections in the first few months. This could be explained by the higher frequency of application (every other day) during the first month of the trial.







When adolescents face both Internet addiction and mood symptoms: A cross-sectional study of comorbidity and its predictors

Tingting Gao1,2,3, Muzi Li2,3, Yueyang Hu1, Zeying Qin1, Ruilin Cao1, Songli Mei1, Xiangfei Meng2,3
1Department of Social Medicine and Health Management, School of Public Health, Jilin University, Changchun, Jilin Province, China
2Department of Psychiatry, Faculty of Medicine, McGill University, Montreal, Canada
3Douglas Mental Health University Institute, Montreal, Canada
Poster Presentation

Background: Research is emerging in the field of comorbidity between mood disorders and symptoms and Internet addiction (IA). IA and mood disorders and symptoms tend to go together, and the comorbidity often predicts substantial burden and severe psychiatric symptomatology. This study was designed to explore the tendency of comorbidity between IA and mood symptoms, and to identify common and unique risk profiles for individual and comorbid mental health problems among adolescents.

Methods: FA total of 2272 high school students were included for this study. Participants were asked to complete paper-based questionnaires and scales on socio-demographic characteristics, lifestyle factors, Internet use, parent-child relationship, and IA and mood symptoms. Multivariate logistic regression was used to compare the risk profiles for IA, mood symptoms and their comorbidity. Bonferroni correction was used for multiple testing.

Results: The prevalence of IA alone, mood symptoms alone and their comorbidity in high school students was 7.0%, 27.0%, 13.6%, respectively. Those students who were current smokers, had more than 1.5 hours sleep deprivation on weekdays, more than 30 minutes of Internet use on weekdays, more than 240 minutes Internet use on weekends, perceived poor parent-child relationships, reported higher rates of comorbidity (P<0.05). Adolescents exposed to more risk factors would have high risk of IA, mood symptoms and their comorbidity.

Conclusion: IA frequently co-occurred with mood symptoms among adolescents. Attention should be paid to the above identified risk factors related to etiopathogenesis of mental disorders. Public mental illnesses prevention strategies should be able to effectively identify and intervene the co-occurrence of IA and mood symptoms.







Development and validation of a short-version of the CECA.Q3 among a population of mood disorder

Muzi Li1, Tamara Cassis1, Nancy Low1,2, Xiangfei Meng1,2
1Department of Psychiatry, Faculty of Medicine, McGill University, Montreal, Canada
2Douglas Mental Health University Institute, Montreal, Canada
Poster Presentation

Background: Childhood maltreatment has significant negative consequences on physical and mental health. The Childhood Experience of Care and Abuse Questionnaire 3 (CECA.Q3) interview as an extended version of the CECA.Q was developed to retrospectively measure childhood adversities. Although it is a reliable and valid measure to screen maltreatment exposures, the application of CECA.Q3 is limited due to its intensive and lengthy in-person interview. This paper aimed to develop and validate a short-version of CECA.Q3 to improve its application.

Data analyzed was from a clinical sample of 210 patients with mood disorders. Data was randomly split into training and testing datasets. The training dataset was used for scale reduction by using principle component factor analysis, while the testing one was used for cross-validation to examine if the short CECA.Q3 could have good yield of accuracy in the testing dataset. The optimal cut-off points of the short-version were also tested.

Methods: Purposive sampling was used whereby CHVs identified pregnant women and mothers with children less than one year old. Health providers in selected health facilities and members of the Sub-County Health Management Teams were also purposively selected. In total, 5 FGDs with CHVs, 13 FGDs with WoRA, and 17 key-informant interviews with health providers were conducted. All interviews were audio recorded, translated into English, and transcribed into word-processing files that were uploaded to NVivo 11 for coding and analyzed using a thematic framework approach.

Results: Overall, a total of 4 out of 8 subscales had items reduction without compromising their accuracy of measurements for childhood adversities. They are antipathy (4 items), neglect (5 items), psychological abuse (15 items), and role reversal (11 items). This short-version reduced about 35.0% items (35/100 items) from the full CECA.Q3. The accuracy of this short-version was validated by the testing dataset.

Conclusion: The short-version of CECA.Q3 offers a briefer but good accuracy of measurements for childhood maltreatment. This short-version could facilitate its application, as it significantly reduces the length of interview and therefore has better compliance from respondents.







Case fatality among Indian tuberculosis patients: a systematic review and meta-analysis

Sophie Huddart1,2, Anita Svadzian1,2, Vaidehi Nafade1,2, Srinath Satyanarayana3, Madhukar Pai1,2
1Department of Epidemiology, Biostatistics and Occupational Health, McGill University, Montreal, Canada
2McGill International TB Centre, Montreal, Canada
3Center for Operational Research, International Union Against Tuberculosis and Lung Disease, Delhi, India
Poster Presentation

Background: More than a quarter of the global TB deaths occur in India. Patient mortality is an important marker of care quality as prompt diagnosis and appropriate treatment should prevent deaths both during and after treatment. This systematic review seeks to estimate the case fatality ratio (CFR) for Indian TB patients.

Methods: We searched Medline, Embase and Global Health for eligible papers published between 2006 and 2017. The treatment and post-treatment CFRs were extracted and, when sufficiently homogeneous, pooled using Normal-Binomial Generalized Linear Mixed Models. Pooling was also performed in key patient subgroups.

Results: A total of 125 relevant studies were identified. The overall treatment CFR was 0.06 (95% CI: 0.04, 0.07). The CFR was higher for HIV+ [0.11 (0.08, 0.15)] and DR-TB patients [0.12 (0.08, 0.17)]. We found similar CFRs for adult [0.05 (0.03, 0.08)] and pediatric [0.04 (0.02, 0.09)] patients. The public sector CFR was 0.05 (0.04, 0.07) but only 4 of 125 (3.2%) papers described privately treated patients, precluding a pooled estimate for this stratum. Out of 125 studies, 78 (62.4%) had limited generalizability, 31 (24.8%) had selection bias, and 6 (4.8%) had short follow-up times.

Conclusion: Our study shows that overall, Indian TB patients experience a CFR equal to that called for in the WHO End TB strategy. However, the CFR is not well described or is unacceptably high for important vulnerable groups. This work highlights the need for more high-quality patient follow-up, especially in India’s large private healthcare sector.







Hepatitis C virus (HCV) seroprevalence and associated risk factors in Montreal men who have sex with men (MSM): results from three cross-sectional studies (2005, 2008-2009, 2018)

Charlotte Laniece Delaunay1, Joseph Cox1,2,3, Marina Klein2,4, Gilles Lambert3,5, Mathieu Maheu-Giroux1
1Department of Epidemiology, Biostatistics, and Occupational Health, McGill University, Montreal, Canada
2Division of Infectious Diseases/Chronic Viral Illness Service, Department of Medicine, McGill University Health Centre, Montreal, Canada
3Secteur Prévention et contrôle des maladies infectieuses, Direction régionale de santé publique, CIUSSS Centre-Sud-de-l'Île-de-Montréal, Montreal, Canada
4CIHR Canadian HIV Trials Network
5Unité des infections transmissibles sexuellement et par le sang, Institut national de santé publique du Québec, Montreal, Canada
Poster Presentation

Background: In the context of recent HCV epidemics among HIV-positive MSM, evidence suggests that sexual transmission of HCV occurs among HIV-positive MSM without a history of injection drug use (IDU), with high reinfection rates. This study aims to examine changes in HCV seroprevalence from 2005 to 2018 among Montreal MSM and to identify biological, social, and behavioural factors associated with HCV seropositivity.

Methods: We used data from three cross-sectional studies conducted in 2005 (Argus 1, n=1,808), 2008-2009 (Argus 2, n=1,260), and 2018 (Engage, n=1,086) among Montreal MSM. To allow for robust computation and comparison of prevalence estimates, the Argus surveys were standardized on age, income, sexual orientation, and first language to the Engage respondent-driven sample. We then pooled the three surveys and conducted logistic regression analyses to identify factors associated with HCV seropositivity.

Results: Our results show a relatively stable HCV seroprevalence among MSM from 6.3% (95% confidence interval (CI): 5.1–7.6%) in 2005, 7.2% in 2008-2009 (95% CI: 5.1–9.4%), and 7.6% (95% CI: 4.5–10.8%) in 2018. In multivariable logistic regressions, history of IDU (adjusted odds ratio (AOR): 18.0, 95% CI: 11.1–29.3), sexual orientation other than gay or homosexual (AOR: 4.7, 95% CI: 2.8–7.9), and HIV-positive status (AOR: 3.1, 95% CI: 1.8–5.4), were associated with increased odds of HCV seropositivity. Inversely, annual income >30,000 dollars (AOR: 0.4, 95% CI: 0.2–0.6) was associated with decreased odds of HCV seropositivity. No association was found between the odds of HCV seropositivity and age, first language other than French or English, self-reported risky sexual behaviour, number of sex partners in the last six months, current/past syphilis infection, or year of data collection.

Conclusion: These preliminary findings suggest a stable HCV seroprevalence among Montreal MSM from 2005 to 2018, and a strong association between HCV seropositivity and IDU.







Estimating Lifetime History of TBI in Canada: Results from the Canadian Longitudinal Study on Aging

Sacha Weill1,2, Bin Zhu1, Anoosh Poorian2, Christina Wolfson1,2
1Research Institute of the McGill University Health Center, Montreal, Canada
2Department of Epidemiology, Biostatistics and Occupational Health, McGill University, Montreal, Canada
Poster Presentation

Background: Traumatic Brain Injury (TBI) amongst older adults is a leading public health concern in developed nations. To be able to properly confront this issue, it is necessary to have an accurate understanding of the lifetime history of TBI in older adults. In Canada, as far as we are aware, there are no estimates of lifetime history of TBI in older adults.

Methods: We calculated unweighted and weighted estimates of older adults' lifetime history of TBI using data gathered on participants in the Canadian Longitudinal Study on Aging (CLSA). The CLSA is a Canadian population-based study of approximately 50 000 individuals, aged 45–85 years at baseline. We used sampling weights to adjust the data based on the demographic distribution of the participants included in the CLSA. Furthermore, we attempted to use previous knowledge about common TBI co-morbidities and ancillary data collected in the CLSA to provide additional evidence for the presence of TBI in participants screening positive for a lifetime history of TBI.

Results: The unweighted and weighted proportions of positive screens for lifetime history of TBI in older adults participating in the CLSA were found to be: 24.2% [95% CI: 23.73-24.74] and 25.8% [95% CI: 25.2-26.5]. The TBI status of participants was supported by higher rates of certain TBI correlates, such as post-traumatic stress disorder.

Conclusion: To the best of our knowledge, our results provide the first estimate of a lifetime history of TBI in older adults in Canada. The findings suggest that TBI is common in this population. Further research is needed to investigate the frequency, trends and causes of TBI in Canadian older adults and our findings provide evidence that the CLSA might be a suitable vehicle for the study of TBI in this population.







A comparison of privacy-preserving statistical methods for analyzing distributed data with binary outcome variables

Dominik Yang1, Paramita Saha Chaudhuri1
1Department of Epidemiology, Biostatistics, and Occupational Health, McGill University, Montreal, Canada
Poster Presentation

Background: Analysis of distributed databases is necessary for multi-centre etiologic studies. Data confidentiality regulations may restrict the ability to combine distributed data into a single dataset. This study will compare the strengths and limitations of three major privacy-preserving statistical methods (meta-analysis, micro-aggregation and distributed regression) employed to address this issue.

Methods: Data from the Canadian Longitudinal Study of Aging (CLSA) was used to estimate the association between restless legs syndrome (RLS) and three binary cardiovascular health outcomes (heart disease, peripheral vascular disease (PVD) and hypertension). Data from 11 collection sites across all provinces were compiled into one large dataset (gold standard). The three methods in question were performed as if the provinces were separate entities unable to share individual-level data between each other. Logistic regression was used for all analyses.

Results: Since 2009, CLSA has recruited 50,000 participants aged 45 and older. Participants classified as RLS positive or negative were included in our analysis and any missing values were excluded. Of the 16,672 subjects included, 3,520 had RLS (21.1%). The unadjusted odds ratio and 95% CI for PVD based on the entire dataset was 2.769 [2.337, 3.276]. Fixed effect meta-analysis provided 2.746 [2.316, 3.287], random effects meta-analysis returned 2.612 [2.014, 3.387] and micro-aggregation separated by provinces gave 2.794 [2.326, 3.372]. Similar results were observed with heart disease and hypertension as the outcomes of interest. From the unadjusted estimates, fixed effect meta-analysis provided the most similar results to the gold standard, marginally outperforming micro-aggregation.

Conclusion: Distributed regression is still a work in progress, as well as comparing the three methods after adjusting for potential confounders. Additional binary cardiovascular health outcomes will be explored as well. We will then compare these methods with a smaller sample size by examining the association between menopause and sleep disorders in women aged 45-60.







Lubricant Investigation in Men to Inhibit Transmission of HPV Infection (LIMIT-HPV): protocol for a phase IIB, placebo-controlled, randomized controlled trial

Cassandra Laurie1, Mariam El-Zein1, Pierre-Paul Tellier2, François Coutlée3, Alexandra de Pokomandy4, Eduardo L. Franco1
1Division of Cancer Epidemiology, McGill University, Montreal, Canada
2Department of Family Medicine, McGill University, Montreal, Canada
3Centre Hospitalier de l’Université de Montréal, Département de Microbiologie et infectiologie, Montréal, Canada
4The Research Institute of the McGill University Health Centre, Montreal, Canada
Poster Presentation

Background: Men who have sex with men (MSM), particularly HIV-positive men, have an increased risk of human papillomavirus (HPV) acquisition and HPV-associated diseases, such as anal cancer and anogenital warts. As HPV vaccines are not effective against all HPV genotypes, and condoms do not fully protect against HPV transmission, additional primary prevention measures are needed. A carrageenan-based lubricant has the potential to prevent HPV acquisition, thereby reducing the disease burden in this vulnerable population. The LIMIT-HPV study is a randomized controlled trial designed to evaluate the efficacy of a carrageenan-based lubricant in reducing type-specific anal HPV prevalence and incidence in sexually-active MSM.

Methods: LIMIT-HPV aims to recruit 380 MSM (110 HIV+ and 270 HIV-) living in Montreal. Volunteer MSM are randomly assigned 1:1 to the treatment (a self-applied anal microbicide gel with carrageenan) or placebo (a self-applied placebo gel) group. Anal specimens are collected at baseline and each consecutive follow-up visit (month 1, 2, 3, 6, 9, 12) and tested for 36 HPV types using the PGMY PCR in the Linear Array Assay. At each visit, computerized questionnaires are used to collect sociodemographic, clinical, lifestyle, and sexual behaviour characteristics, and monitor the safety and tolerability of the gels. HIV status is determined at baseline and at 12 months. Data analysis will be conducted according to intention-to-treat principles. Subgroup analyses by HIV status will be performed.

Results: Enrollment began in February 2016. Of the 296 MSM screened, 270 were eligible, and 215 have been enrolled. Of these, 50 MSM have completed all study visits and 66 were lost to follow up, with follow-up visits scheduled for 99 participants.

Conclusion: If the gel is proven effective, this research has the potential to improve the health of individuals in the MSM community by providing protection against all HPV genotypes, and ultimately reducing the risk of HPV-associated diseases.







The relationship between pregnancy weight gain by body mass index (BMI) category and adverse birth outcomes

Ugochinyere Vivian Ukah1, Sarka Lisonkova2
1Department of Epidemiology, Biostatistics and Occupational Health, McGill University, Montreal, Canada
2Department of Obstetrics and Gynaecology, University of British Columbia, Vancouver, Canada
Poster Presentation

Background: Optimal weight gain during pregnancy is recommended to reduce the risks of adverse maternal and perinatal outcomes. We sought to examine the association between sub-optimal pregnancy weight gain as recommended by the American College of Obstetricians and Gynecologists during pregnancy and adverse birth outcomes in relation to pre-pregnancy body mass index (BMI).

Methods: A population-based retrospective cohort study of women with singleton birth in Washington State, USA, 2003-2013 was used. Weight gain categories included low, optimal, and excess weight gain, based on the recommended weight gain by gestational age in each pre-pregnancy BMI category (underweight, normal BMI, overweight, and obese classes). The composite adverse outcome included maternal and perinatal death, and severe maternal and neonatal morbidity. Logistic regression yielded adjusted odds ratio (AOR) and 95% confidence intervals (CI).

Results: Among 722,839 women, 3.1% were underweight, 48.1% had normal BMI, 25.8% were overweight, and 22.9% obese. Approximately 45% of underweight women, 38% of normal weight, 25% of overweight, and 24% of obese women had optimal weight gain in pregnancy. Adverse birth outcomes occurred in 3.2% of women. Low weight gain compared with optimal weight gain in the same pre-pregnancy body mass index categories was significantly associated with higher adverse outcomes: adjusted OR 1.35 (95% CI, 1.13-1.62) for underweight, OR 1.26 (95% CI, 1.19-1.33) for normal weight, OR 1.39 (95% CI, 1.27-1.51) for overweight, and OR 1.32 (95% CI, 1.12-1.42) for obese. Similarly, excess weight gain was significantly associated with worse outcomes, except for the overweight group.

Conclusion: It is important to monitor both low and excess weight gain in pregnancy, especially in obese women, to improve nutritional and life-style counselling.







Testosterone replacement therapy and the risk of prostate cancer in men with late-onset hypogonadism.

Christina Santella1,2, Christel Renoux1,2,3, Hui Yin1, Oriana H. Y. Yu1,4, Laurent Azoulay1,2,5
1Centre for Clinical Epidemiology, Lady Davis Institute, Jewish General Hospital, Montreal, Canada
2Department of Epidemiology, Biostatistics, and Occupational Health, McGill University, Montreal, Canada
3Department of Neurology and Neurosurgery, McGill University, Montreal, Canada
4Division of Endocrinology, Jewish General Hospital, Montreal, Canada
5Gerald Bronfman Department of Oncology, McGill University, Montreal, Canada
Poster Presentation

Background: The association between the use of testosterone replacement therapy (TRT) and prostate cancer remains uncertain, with studies assessing this safety concern having important methodological limitations. Thus, the objective of this population-based study was to determine whether TRT is associated with an increased risk of prostate cancer in men with late-onset hypogonadism.

Methods: We used the United Kingdom Clinical Practice Research Datalink to assemble a cohort of 12,779 men, at least 45 years of age, newly-diagnosed with hypogonadism between 1 January 1995 and 31 August 2016, with follow-up until 31 August 2017. Time-dependent Cox proportional hazards models were used to estimate adjusted hazard ratios (HRs) with 95% confidence intervals (CIs) of prostate cancer associated with use of TRT, compared with non-use. Exposure to TRT was lagged by 1 year to account for cancer latency. Secondary analyses assessed the association by TRT formulation, number of prescriptions received, and time since initiation. We also conducted a propensity score-matched cohort analysis to further assess the impact of residual confounding.

Results: A total of 215 patients were newly-diagnosed with prostate cancer during 58,224 person-years of follow-up, generating an incidence rate of 3.7 per 1,000 person-years. Compared with non-use, use of TRT was not associated with an overall increased risk of prostate cancer (3.8 v 3.4 per 1,000 person-years, respectively; adjusted HR, 0.97; 95% CI, 0.71 to 1.32). Results remained consistent in secondary and sensitivity analyses, as well as in the propensity score-matched cohort analysis (HR, 0.87; 95% CI, 0.56 to 1.36).

Conclusion: The use of TRT was not associated with an increased risk of prostate cancer in men with late-onset hypogonadism.







Outdoor, indoor, and personal exposures to traffic-related air pollution in retired adults in urban Colombia

Kabisha Velauthapillai1, Mauricio Victor Herrera2, Skarlet Vasquez2, Kento Taro Magara Gomez3, Scott Weichenthal1, Jill Baumgartner1,4, Ellison Carter5
1Department of Epidemiology, Biostatistics, and Occupational Health, McGill University, Montreal, Canada
2Grupo Cardiología Preventiva, Universidad Autónoma de Bucaramanga, Bucaramanga, Colombia
3Department of Environmental Engineering, Pontificia Bolivariana University, Bucaramanga, Colombia
4Institute of Health & Policy, McGill University, Montreal, Canada
5Department of Civil and Environmental Engineering, Colorado State University, Fort Collins, USA
Poster Presentation

Background: Traffic-related air pollution (TRAP) is increasing due to urban expansion. Extensive research on TRAP has been conducted in high-income countries, but little is known about low and middle-income countries, where vehicle fleet, road paving, home-to-road distance, and ventilation practices may lead to very different exposures. We aimed to characterize the concentrations, chemical components, and sources of exposures to TRAP in urban Colombia.

Methods: Using a cross-sectional study design, we selected four neighbourhoods in Bucaramanga, Colombia based on vehicle braking frequency and traffic, diesel, and exhaust levels. In each neighbourhood, we randomly selected 19-20 homes with retired adults and measured 48-h indoor concentrations and personal exposures to fine particles (PM2.5), black carbon, carbon dioxide, and carbon monoxide. Community air pollution monitors measured PM2.5 and black carbon for 6 days in each neighbourhood.

Results: The high traffic, but low diesel neighbourhood, had the highest outdoor mean PM2.5 concentration: 14.8 (± 3.1) ug/m3, while the neighbourhood with high braking and diesel had the lowest with 10.1 (± 2.2) ug/m3. Indoor mean concentration and personal exposure to PM2.5 were highest in the high traffic, high diesel neighbourhood: 17.8 (± 1.6) ug/m3. Outdoor black carbon also comprised a higher proportion of total PM2.5 mass (e.g. 20-29%) compared with what has been observed in high-income settings.

Conclusion: We found that indoor levels exposures to particulate air pollutants are higher in high traffic neighbourhoods but are lower than the WHO’s 24-h guidelines (25 ug/m3). In each neighborhood, black carbon, a combustion pollution marker shown to be more strongly associated with adverse health outcomes than PM, made up a substantial proportion of measured particle mass. These results are important for further understanding TRAP in cities like Bucaramanga and will form the foundation for modeling urban cardiovascular health outcomes using existing health datasets collected by in-country partners.







Psychological Health of Canadian Veterans: Data from the Canadian Longitudinal Study on Aging

Jessica Maitland1,2, Istvan Molnar-Szakacs2, Geva Maimon2, Christina Wolfson1,2
1Department of Epidemiology, Biostatistics, and Occupational Health, McGill University, Montreal, Canada
2Research Institute of the McGill University Health Centre, Montreal, Canada
Poster Presentation

Background: It is generally accepted that military Veterans experience a higher frequency of mental health disorders, yet there is little research on the potential long-term impact of military service on psychological health. In this study, we describe the psychological health of older Veterans in Canada and how it compares to non-Veterans, using data from the Canadian Longitudinal Study on Aging (CLSA).

Methods: Canadian Veterans were identified among 51,338 CLSA participants, using a set of Veteran identifier questions asked at the CLSA baseline assessment. To describe psychological health at baseline, we used descriptive statistics for variables including life satisfaction, self-rated mental health, depression, self-reported anxiety and mood disorders, and Posttraumatic Stress Disorder (PTSD). As the Veterans were much more likely to be male and were older than non-Veterans, all results are age and sex-standardized.

Results: In the CLSA, 3,558 participants (7.1%), self-reported as Canadian Veterans with military join years as early as 1944 and release dates as recent as 2008. Our preliminary results suggest that Veterans and non-Veterans did not differ on measures of life satisfaction, self-rated mental health or depression. However, Veterans were slightly more likely to report mood disorders than non-Veterans (18% vs 16%); and anxiety disorders (10% vs 8%).

Conclusion: A preliminary analysis of the CLSA data suggests that on average Canadian Veteran and non-Veteran participants have comparable psychological health, with some indication of higher levels of mood and anxiety disorders among Veterans. Further analyses will be done to explore the extent of psychological health differences between Veterans and non-Veterans and whether these differences are affected by the duration of service and time since release from the military.