Abstracts / Annals of Epidemiology 24 (2014) 682e702
Methods: Cross sectional hospital-based study was conducted in 4 major hospitals in Tanzania from 2012 -2014. Parents/guardians of 175 children were assessed about their socio-demographic profile by questionnaires. Children were clinically examined for cleft types and oral health status. Analysis elicited frequency distributions, cross tabulations and Chi-square statistics were used to assess bivariate relationships (P < 0.05). Results: Majority were: boys (54.3%), below 5-years (86.3%; 37.7% below 1year), rural dwellers (80.6%) and with low-educated parents (89.1%). Most children had complete-unilateral cleft lip (52%), more on left side (37.1%). Complete-bilateral clefts were common (18%) than incomplete (4%). Hardsoft palate clefts were found in 37.1% and soft palate-only in 2.9%. Malocclusion, caries and poor oral hygiene were found in 48.6%, 28.4% and 42.3% of children, respectively. Significant association (p<0.05) was such that; bilateral clefts were more in children below than above 1-year, cleft of the soft palate-only was more in children above than below 5-years and a cleft of both palates was more in urban than rural dwellers. Conclusion: Complete-unilateral cleft lip was the commonest type, malocclusion were the commonest oral- health impacts, rural dwellers and low-level of parents’ education were significant risk indicators for lip/palateclefts in Tanzanian children.
P72. Longer Breastfeeding Duration Reduces the Positive Relationships among Gestational Weight Gain, Birthweight, and Childhood Growth Yeyi Zhu PhD, Ladia M. Hernandez PhD, Yongquan Dong MS, John H. Himes PhD, Steven Hirschfeld PhD, MD, Michele R. Forman PhD. University of Texas at Austin Purpose: The relationship between gestational weight gain (GWG) and childhood growth remains controversial; thus an examination whether early life nutrition mediates this relationship may improve our understanding. We aimed to investigate whether the relationships among GWG, birthweight, and childhood anthropometrics were mediated through infant feeding practices (i.e., breastfeeding duration and age at introduction of solid foods). Methods: In a cross-sectional multi-ethnic study of 1387 mothers and their children aged 0-5.9 years in the US (2011-2012), perinatal and socio-demographic factors were collected from the mother via an interviewer-administered questionnaire. Child anthropometrics included age- and sex-specific z-scores for: weight-for-age (WAZ), height/length-for-age (HAZ), weight-forheight/length (WHZ) and body mass index-for-age (BMIZ), and ulnar length, a linear marker for skeletal growth. Mediation analysis was used to calculate standardized coefficients of the paths among GWG, birthweight, infant feeding, and anthropometrics. Results: Maternal GWG had a consistent positive effect on all anthropometrics mediated via birthweight, whereas longer breastfeeding duration reduced the positive associations of GWG and birthweight with WAZ, WHZ, and BMIZ (Î2 ¼ -0.076, -0.062, and -0.124, respectively) in non-Hispanics. Longer breastfeeding duration and a later age at introduction of solid foods were both positively associated with ulnar length (Î2 ¼ 0.021 and 0.030, respectively) not HAZ, suggesting for the first time a distinct effect of infant feeding practices on limb growth. Conclusions: Findings suggest promoting breastfeeding among women with excessive GWG may mitigate offspring obesity. Future prospective research investigating ethnic-specific interrelationships among GWG, birthweight, infant feeding, and postnatal growth are warranted.
Public Health P73. Performance-Enhancing Dietary Supplements and Other Health Related Factors Associated with Adverse Medical Outcomes in Active Duty U.S. Army Soldiers Esther O. Dada MPH, Chiping Nieh PhD, Theresa Jackson PhD, MPH, Samar DeBakey MD, MPH, Bruce H. Jones MD, MPH. U.S. Army Institute of Public Health, U.S. Army Public Health Purpose: Amid safety concerns over dietary supplements containing dimethylamylamine (DMAA), this study examined associations between DMAA use and other potential risk factors with experiencing adverse medical outcomes (AMOs) among Active Duty U.S. Army Soldiers.
699
Methods: A case-control study was conducted among 712 Army Soldiers with clinician-identified AMO (cardiac dysrhythmia, seizure, heat injury, rhabdomyolysis, cerebral hemorrhage, acute/subacute necrosis of liver, acute kidney failure) and 1,077 controls, ascertained from Defense Medical Surveillance System. Participants completed questionnaire on DMAA use, physical activity/fitness, sleep, medication and substance use. Multivariate logistic regression was performed to examine associations between DMAA use and other potential risk factors and AMOs. Results: No statistically significant association was observed between selfreported DMAA use and selected AMOs after controlling for relevant factors (adjusted odds ratio [aOR] 95% confidence interval [CI] of 0.85 (0.59-1.23)). However, Soldiers with multiple AMOs were twice as likely to have reported using DMAA compared to those with one reported outcome (OR 2.28 (95% CI 1.01-4.95)). Soldiers with prior AMO were almost 18 times more likely to have a current AMO (aOR 17.86 (95% CI 9.60-33.22)). Soldiers who reported antidepressant use were about three times more likely to experience an AMO (aOR 2.75 (95% CI 1.96-3.87)). Conclusions: While this study did not find that simply using DMAA increased the likelihood of Service members experiencing an AMO, study did show that those with multiple AMOs were twice as likely to use DMAA. The strongest predictor of AMOs was prior AMO suggesting the importance of prevention strategies to reduce reoccurrence.
P74. Risk Profiles of Younger and Older Pregnant Women Who Consume Alcohol During Pregnancy Panagiota Kitsantas PhD, Kathleen Gaffney, Carol Wu. George Mason University Purpose: Recent studies indicate that older women are more likely to consume alcohol during pregnancy. However, little is known about alcohol consumption during pregnancy within the context of maternal age. This study identifies risk profiles of younger and older women who are at risk of consuming alcohol during pregnancy using classification and regression trees (CART) analysis. Methods: We used the 2002-2009 Pregnancy Risk Assessment Monitoring System (PRAMS) dataset (311,428 records U.S. women). Logistic regression and classification trees were constructed separately for age groups, 24, 2529, and 30 years. Results: Alcohol consumption by age group was: 3.7% for 24, 5.7% for 2529, and 10.1% for 30 years of age. Women 24 years were at greater risk of consuming alcohol if they also smoked (5.8%). Among nonsmokers, higher levels of education and being Hispanic were associated with a 35% increase in alcohol use. Distinct high-risk subgroups emerged for the 25-29 year old group. Specifically, 12.8% of non-obese women who reported having experienced abuse during pregnancy also reported drinking alcohol in the last trimester. About 16% of women 30 years with at least 16 years of education, white or Hispanic with normal or underweight BMI, drank alcohol during their last trimester. Conclusions: Given limited health care resources for prevention and treatment, the early identification of high-risk groups for prenatal alcohol use is critical. This study provides evidence that risk factors contributing to alcohol consumption during pregnancy may differ by maternal age.
P75. Status Epilepticus and Subsequent Epilepsy Angela M. Malek PhD, Dulaney A. Wilson PhD, Braxton B. Wannamaker MD, Gabriel U. Martz MD, Gigi Smith PhD, APRN, CPNP-PC, Anbesaw W. Selassie DrPH. Medical University of South Carolina Purpose: The incidence of status epilepticus (SE), characterized by continuous seizures lasting >5 minutes or repeated seizures without regaining consciousness, ranges from 6.8-41 per 100,000 persons (DeLorenzo 1995; Wu 2002). Few studies have investigated SE and subsequent epilepsy. We evaluated risk of subsequent epilepsy after SE and described associated comorbid conditions Methods: This retrospective cohort study included South Carolina hospitalizations and emergency department (ED) visits from 2000-2011. All individuals with SE were identified and classified based on a subsequent epilepsy diagnosis. Individuals with known epilepsy prior to the diagnosis of SE and persons with epilepsia partialis continua were excluded. Descriptive statistics were used to compare demographic and clinical characteristics. Cox
700
Abstracts / Annals of Epidemiology 24 (2014) 682e702
proportional hazard models were used to generate hazard ratios (HRs) and corresponding 95% CIs for the risk of subsequent epilepsy. Results: Subsequent epilepsy was diagnosed in 35.7% of those with SE. 81.6% developed epilepsy within one year of SE, and an additional 10.4% the second year. The hazard of subsequent epilepsy was significantly increased for those with grand mal SE (HR¼2.02, 95% CI: 1.10-3.72); of Hispanic race/ethnicity (HR¼2.76, 95% CI: 1.44-5.28) and aged 13-18 years (vs. 0-5) (HR¼1.66, 95% CI: 1.09-2.54) after adjustment for covariables. Traumatic brain injury, cerebral palsy, tobacco use disorder, hyperlipidemia, >6 comorbidities, and more annual visits increased risk of subsequent epilepsy. Conclusion: A significant proportion of the cohort developed epilepsy within two years of SE with greater risk in those with grand mal SE. Certain demographic characteristics appear to heighten the risk of epilepsy after SE.
P76. Short-Term Exposure to Ambient Ozone and Acute Stroke Risk: A Case-Crossover Analysis Jessica Anne Montresor-Lopez MPH, Jeff D. Yanosky PhD, Murray Mittleman PhD, Amir Sapkota PhD, Xin He PhD, James Hibbert, Michael Wirth PhD, Robin C. Puett PhD. University of Maryland Purpose: Short-term exposure to ambient ozone air pollution may be associated with an increased risk of acute stroke onset; however, findings from previous studies have been inconsistent. We evaluated the association between short-term ozone exposure and stroke hospital admissions among residents of South Carolina. Methods: Data on all incident stroke hospitalizations (2002-2006) for adult residents of South Carolina (SC) were obtained from the SC Office of Research and Statistics. Ozone exposure data were obtained on a 12-km grid from the USEPA’s Hierarchical Bayesian Model. A semi-symmetric bidirectional casecrossover design was used to examine the association between ozone exposure on lag days 0, 1, and 2 and stroke hospitalization. Conditional logistic regression was used to estimate odds ratios (OR) and 95% confidence intervals (CI). Stratified analyses by gender, race and stroke subtype were performed. Results: No significant associations were observed between short-term ozone exposure and hospitalization for all stroke (e.g. lag day 0 OR: 0.99; 95% CI: 0.98, 1.01) or ischemic stroke (lag day 0 OR: 1.00; 95% CI: 0.98, 1.02.) For hemorrhagic stroke, stratified analyses suggest potential effect modification by race and gender. Risk of hospitalization for hemorrhagic stroke was higher among African-Americans than European-Americans and higher among males than females; however, these associations did not reach nominal statistical significance. Conclusions: Among adults in SC from 2002-2006, there was no evidence of an association between ozone exposure and increased risk of hospitalization for all stroke or ischemic stroke; however, African-Americans and males may have an increased risk of hemorrhagic stroke.
P77. Estimating the Effects of Sugar-Sweetened Beverage Taxes on Weight and Obesity in New York City Using Dynamic Loss Models Ryan Richard Ruff PhD, Chen Zhen PhD. New York University Purpose: Sugar-sweetened beverage taxes are frequently cited strategies to improve nutrition and reduce obesity at the population level. Few studies extrapolate calorie reductions from proposed public health interventions into expected weight change and the resulting shifts in the obesity distribution. Methods: Resting metabolic rates and body fat mass were estimated using the Mifflin-St. Jeor and Jackson equations, respectively. Expected calorie reductions due to simulated tax scenarios were derived previously using a fully endogenized distance-metric almost ideal demand system (DM-AIDS). One and ten-year simulations of body composition change were conducted using static and dynamic loss models. Results: The DM-AIDS model predicted that a .04-cent per calorie tax on sugar-sweetened beverages would reduce beverage energy intake by 5900 calories per person per year. Mean unweighted predicted body fat mass for the New York City adult population was 25.1 kg (SD¼12.83 kg). Obesity prevalence as determined by predicted body fat was 47% (95% CI¼45.5, 48.4) compared to 24.2% based on BMI (95% CI 22.8, 25.5). Preliminary results from
the dynamic loss model using population average data yielded a reduction of .4 kg of body fat in year one and .65 kg of body fat in year ten. Conclusions: Using initial body conditions and established metabolic parameters, dynamic weight loss models simulate the effect of dietary perturbations and energy expenditure on body composition change. This can lead to more informed intervention strategies to reduce the burden of obesity and associated noncommunicable diseases.
Reproductived Health P78. Urinary Bisphenol a and Semen Quality: The Life Study Alexandra E. Goldstone BS, Zhen Chen PhD, Melissa J. Perry ScD, MHS, Kurunthachalam Kannan PhD, Germaine M. Buck Louis PhD, MS. Milken Institute School of Public Health, George Washington University Purpose: Previous studies suggest bisphenol A (BPA) is negatively associated with semen quality in men recruited from clinical centers. This analysis builds on the existing literature by assessing the association between BPA and 35 semen quality parameters among men recruited from the general population. Methods: 501 couples discontinuing contraception to become pregnant were recruited preceding conception between 2005-2009 and followed for 12 months or until pregnant. Male partners provided urine and serum samples at enrollment for BPA and cotinine quantification, respectively. Following two days of abstinence, males collected two semen samples approximately one month later. Semen samples were assessed for 5 general characteristics, 8 motility measures, 6 sperm head measures, 14 morphology measures, and 2 sperm chromatin structure assay measures. Linear regression models assessed the association between BPA and individual semen quality parameters. Select semen parameters were dichotomized according to World Health Organization reference values and modeled using logistic regression. Generalized Estimating Equations accounted for repeated measures of semen quality, and adjusted models included age, abstinence time, alcohol consumption, body mass index, creatinine, education, income, previously fathered pregnancy, race/ethnicity, cotinine, and study site (n¼418). Results: Geometric mean BPA concentration was 0.55 ng/mL (95% CI 0.400.63). A negative relation between BPA and DNA fragmentation was observed in unadjusted (b¼ -0.0649, p¼0.002) and adjusted (b¼ -0.0544, p¼0.035) linear regression models. No other findings achieved statistical significance. Conclusions: Corroborating previous findings, our analysis suggests an association between BPA and semen quality, but only for DNA fragmentation and not other aspects of semen quality.
P79. Risk Factors for Small for Gestational Age (SGA) in Central Appalachia Counties with Mountain-Top Mining (MTM) Elisabeth Dissen MPH, Catherine Kruck, Nana Afari MPH, Hamid Ferdosi BS, Ji Li MD, PhD, Rusan Chen PhD, Manning Feinleib MD, DrPH, Steven H. Lamm MD, DTPH. John Hopkins Bloomberg School of Public Health Purpose: To determine whether residence in a county with mountain-top mining (MTM) activity is a risk factor for Small for Gestational Age (SGA) in central Appalachia. Methods: NCHS birth certificate data (1990-2003) for central Appalachia (Kentucky, Tennessee, Virginia, and West Virginia) included about 3.4 million live births. SGA criteria were from Oken (2003). Mining status of maternal residence (MTM mining, non-MTM, and non-mining) were from Ahern (2011). Tobacco use (a known risk factor for SGA that is recorded on the birth certificate) was examined by mining activity and by gestational age. SGA prevalences by gestational week were examined with graphic and Poisson analysis by mining activity of county of residence and tobacco use. Results: SGA prevalences were significantly greater for residents of counties with MTM activity and with non-MTM activity than those with no mining activity. Maternal tobacco use prevalences differed by mining activity (MTM 31%; non-MTM 25%; non-mining 16%) and remained constant over gestational ages. SGA prevalences were similar for those with and without tobacco use for gestational ages 22-32 weeks and then separated with rates for tobacco users reaching twice that of non-tobacco users at full-term (GA 3744). The inclusion of tobacco use in the Poisson analysis increased the explanatory power of the model from 4% to 92% with mining group having a negative coefficient.