It is not well understood how air pollution leads to adverse pregnancy outcomes. One pathway may be through C-reactive protein, a biomarker of systemic inflammation that has been reported to increase the risk of preterm delivery. We examined whether air pollution influences serum concentrations of C-reactive protein in early pregnancy.
We studied 1696 pregnant women in Allegheny County, PA, from 1997 through 2001. C-reactive protein concentrations were assayed in blood collected before the 22nd week of gestation. We estimated levels of particles of less than 10 μm (PM10) and less than 2.5 μm diameter (PM2.5), carbon monoxide, nitrogen dioxide, sulfur dioxide, and ozone at the maternal zip code using Kriging interpolation for measurements obtained from ambient stations. Associations between air pollution and high C-reactive protein concentrations (≥8 ng/mL) were evaluated using logistic regression.
Among nonsmokers, an observed 9.2 μg/m3 increase in PM10 (averaged over 28 days prior to the blood sample) was associated with an odds ratios of 1.41 for high C-reactive protein concentrations (95% confidence interval = 0.99–2.00). Similarly, a 4.6 μg/m3 increase in PM2.5 was associated with an odds ratio of 1.47 (1.05–2.06). The odds ratio was 1.49 (0.75–2.96) per 7.9 ppb increase in ozone during summer. There were no associations in smokers or for other air pollutants, and there was no evidence for effect-measure modification by obesity.
PM10, PM2.5, and ozone exposures were associated with increased C-reactive protein concentrations in early pregnancy, suggesting that these air pollutants contribute to inflammation and thereby possibly to adverse pregnancy outcomes.
Recent studies have estimated the reduction in HIV-1 infectiousness with antiretroviral therapy (ART), but high-quality studies such as randomized control trials, accompanied by rigorous adherence counselling, are likely to overestimate the effectiveness of treatment-as-prevention in real-life settings.
We attempted to summarize the effect of ART on HIV transmission by undertaking a systematic review and meta-analysis of HIV-1 infectiousness per heterosexual partnership (incidence rate and cumulative incidence over study follow-up) estimated from prospective studies of discordant couples. We used random-effects Poisson regression models to obtain summary estimates. When possible, the analyses were further stratified by direction of transmission (man-to-woman or woman-to-man) and economic setting (high- or low-income countries). Potential causes of heterogeneity of estimates were explored through subgroup analyses.
Fifty publications were included. Nine allowed comparison between ART and non-ART users within studies (ART-stratified studies), where summary incidence rates were 3.6/100 person-years (95% confidence interval= 2.0-6.5) and 0.2/100 person-years (0.07-0.7) for non-ART- and ART-using couples, respectively (p<0.001), constituting a 91% (79%-96%) reduction in per-partner HIV-1 incidence rate with ART use. The 41 studies that did not stratify by ART use provided estimates with high levels of heterogeneity (I2 statistic) and few reported levels of ART use, making interpretation difficult. Nevertheless, estimates tended to be lower with than without ART use. Infectiousness tended to be higher for low-income than high-income settings, but there was no clear pattern by direction of transmission (man-to-woman and woman-to-man).
ART substantially reduces HIV-1 infectiousness within discordant couples, based on observational studies, and could play a major part in HIV-1 prevention efforts. However the non-zero risk from partners receiving ART demonstrates that appropriate counselling and other risk reduction strategies for discordant couples are still required. Additional estimates of ART effectiveness by adherence level from real-life settings will be important, especially for persons starting treatment early without symptoms.
The parametric g-formula can be used to estimate the effect of a policy, intervention, or treatment. Unlike standard regression approaches, the parametric g-formula can be used to adjust for time-varying confounders that are affected by prior exposures. To date, there are few published examples in which the method has been applied.
We provide a simple introduction to the parametric g-formula and illustrate its application in analysis of a small cohort study of bone marrow transplant patients in which the effect of treatment on mortality is subject to time-varying confounding.
Standard regression adjustment yields a biased estimate of the effect of treatment on mortality relative to the estimate obtained by the g-formula.
The g-formula allows estimation of a relevant parameter for public health officials: the change in the hazard of mortality under a hypothetical intervention, such as reduction of exposure to a harmful agent or introduction of a beneficial new treatment. We present a simple approach to implement the parametric g-formula that is sufficiently general to allow easy adaptation to many settings of public health relevance.
Extremes of temperature have been associated with short-term increases in daily mortality. We identified subpopulations with increased susceptibility to dying during temperature extremes, based on personal demographics, small-area characteristics and preexisting medical conditions.
We examined Medicare participants in 135 U.S. cities and identified preexisting conditions based on hospitalization records prior to their deaths, from 1985–2006. Personal characteristics were obtained from the Medicare records, and area characteristics were assigned based on zip-code of residence. We conducted a case-only analysis of over 11 million deaths, and evaluated modification of the risk of dying associated with extremely hot days and extremely cold days, continuous temperatures, and water-vapor pressure. Modifiers included preexisting conditions, personal characteristics, zip-code-level population characteristics, and land-cover characteristics. For each effect modifier, a city-specific logistic regression model was fitted and then an overall national estimate was calculated using meta-analysis.
People with certain preexisting conditions were more susceptible to extreme heat, with an additional 6% (95% confidence interval= 4% – 8%) increase in the risk of dying on an extremely hot day in subjects with previous admission for atrial fibrillation, an additional 8% (4%–12%) in subjects with Alzheimer disease, and an additional 6% (3%–9%) in subjects with dementia. Zip-code level and personal characteristics were also associated with increased susceptibility to temperature.
We identified several subgroups of the population who are particularly susceptible to temperature extremes, including persons with atrial fibrillation.
Telomere length is a marker of cellular aging that varies by the individual, is inherited, and is highly correlated across somatic cell types within persons. Inter-individual telomere length variability may partly explain differences in reproductive aging rates. We examined whether leukocyte telomere length was associated with menopausal age.
We evaluated the relationship between leukocyte telomere length and age at natural menopause in 486 white women aged 65 years or older. We fit linear regression models adjusted for age, income, education, body mass index, physical activity, smoking, and alcohol intake. We repeated the analysis in women with surgical menopause. We also performed sensitivity analyses excluding women (1) with unilateral oophorectomy, (2) who were nulliparous, or (3) reporting menopausal age <40 years, among other exclusions.
For every one kilobase (kb) increase in leukocyte telomere length, average age at natural menopause increased by 10.2 months (95% confidence interval= 1.3 to 19.0). There was no association in 179 women reporting surgical menopause. In all but one sensitivity analysis, the association between leukocyte telomere length and age at menopause became stronger. However, when excluding women with menopausal age <40 years, the association decreased to 7.5 months (−0.4 to 15.5).
Women with the longest leukocyte telomere length underwent menopause three years later than those with the shortest leukocyte telomere length. If artifactual, an association would likely also have been observed in women with surgical menopause. If these results are replicated, leukocyte telomere length may prove to be a useful predictor of age at menopause.
Although ambient concentrations of particulate matter ≤10μm (PM10) are often used as proxies for total personal exposure, correlation (r) between ambient and personal PM10 concentrations varies. Factors underlying this variation and its effect on health outcome-PM exposure relationships remain poorly understood.
We conducted a random-effects meta-analysis to estimate effects of study, participant and environmental factors on r; used the estimates to impute personal exposure from ambient PM10 concentrations among 4,012 non-smoking, diabetic participants in the Women’s Health Initiative clinical trial; and then estimated the associations of ambient and imputed personal PM10 concentrations with electrocardiographic measures such as heart rate variability.
We identified fifteen studies (in years 1990-2009) of 342 participants in five countries. The median r was 0.46 (range = 0.13 to 0.72). There was little evidence of funnel-plot asymmetry but substantial heterogeneity of r, which increased 0.05 (95% confidence interval [CI]= 0.01 to 0.09) per 10 μg/m3 increase in mean ambient PM10 concentration. Substituting imputed personal exposure for ambient PM10 concentrations shifted mean percent changes in electrocardiographic measures per 10μg/m3 increase in exposure away from the null and decreased their precision, e.g. −2.0% (95% CI= −4.6% to 0.7%) versus −7.9% (−15.9% to 0.9%) for the standard deviation of normal-to-normal RR interval duration.
Analogous distributions and heterogeneity of r in extant meta-analyses of ambient and personal PM2.5 concentrations suggest that observed shifts in mean percent change and decreases in precision may be generalizable across particle size.
Heat is recognized as one of the deadliest weather-related phenomena. Although the impact of high temperatures on mortality has been a subject of extensive research, few previous studies have assessed the impact of population adaptation to heat.
We examined adaptation patterns by analyzing daily temperature and mortality data spanning more than a century in New York City. Using a distributed-lag non linear model, we analyzed the heat-mortality relationship in adults age 15 years or older in New York City during two periods: 1900 to 1948 and 1973 to 2006, in order to quantify population adaptation to high temperatures over time.
During the first half of the century, the decade-specific relative risk of mortality at 29 °C vs. 22 °C ranged from 1.30 (95% confidence interval=1.25 to 1.36) in the 1910s to 1.43 (1.37 to 1.49) in the 1900s. Since the 1970s, however, there was a gradual and substantial decline in the relative risk, from 1.26 (1.22 to 1.29) in the 1970s to 1.09 (1.05 to 1.12) in the 2000s. Age-specific analyses indicated a greater risk for people age 65 years and older in the first part of the century but there was less evidence for enhanced risk among this older age group in more recent decades.
The excess mortality with high temperatures observed between 1900 and 1948 was substantially reduced between 1973 and 2006, indicating population adaption to heat in recent decades. These findings may have implications for projecting future impacts of climate change on mortality.
Obesity is associated with increased mortality in the general population but, paradoxically, with decreased mortality in individuals with diabetes.
Among 88,373 French women participating in the E3N-EPIC study who were free of diabetes in 1990, we estimated the mortality hazard ratios (HR) and 95% confidence intervals (CI) for body-mass index (BMI) levels by diabetes status.
During an average 16.7 years of follow-up, 3,750 deaths and 2,421 incident diabetes cases occurred. In overweight/obese versus normal weight women, the mortality HR (95%CI) was 1.42 (1.32,1.53) in women without diabetes, and 0.69 (0.40,1.18) in women with incident diabetes. Mortality increased with BMI among women without diabetes and decreased as BMI increased in women with diabetes.
We found a direct association between BMI and mortality among women without diabetes but not among those with incident diabetes in the same population. Selection bias may be a simple explanation for this “paradox”.
The average effect of an infectious disease intervention (e.g., a vaccine) varies across populations with different degrees of exposure to the pathogen. As a result, many investigators favor a per-exposure effect measure that is considered independent of the population level of exposure and that can be used in simulations to estimate the total disease burden averted by an intervention across different populations. However, while per-exposure effects are frequently estimated, the quantity of interest is often poorly defined, and assumptions in its calculation are typically left implicit. In this paper, we build upon work by Halloran and Struchiner (Epidemiology. 1995; 6:172-151) to develop a formal definition of the per-exposure effect and discuss conditions necessary for its unbiased estimation. With greater care paid to the parameterization of transmission models, their results can be better understood and can thereby be of greater value to decision-makers.
Net reclassification indices have recently become popular statistics for measuring the prediction increment of new biomarkers. We review the various types of net reclassification indices and their correct interpretations. We evaluate the advantages and disadvantages of quantifying the prediction increment with these indices. For pre-defined risk categories, we relate net reclassification indices to existing measures of the prediction increment. We also consider statistical methodology for constructing confidence intervals for net reclassification indices and evaluate the merits of hypothesis testing based on such indices. We recommend that investigators using net reclassification indices should report them separately for events (cases) and nonevents (controls). When there are two risk categories, the components of net reclassification indices are the same as the changes in the true-positive and false-positive rates. We advocate use of true- and false-positive rates and suggest it is more useful for investigators to retain the existing, descriptive terms. When there are three or more risk categories, we recommend against net reclassification indices because they do not adequately account for clinically important differences in shifts among risk categories. The category-free net reclassification index is a new descriptive device designed to avoid pre-defined risk categories. However, it suffers from many of the same problems as other measures such as the area under the receiver operating characteristic curve. In addition, the category-free index can mislead investigators by overstating the incremental value of a biomarker, even in independent validation data. When investigators want to test a null hypothesis of no prediction increment, the well-established tests for coefficients in the regression model are superior to the net reclassification index. If investigators want to use net reclassification indices, confidence intervals should be calculated using bootstrap methods rather than published variance formulas. The preferred single-number summary of the prediction increment is the improvement in net benefit.
Air pollution may be related to adverse birth outcomes. Exposure information from land-based monitoring stations often suffers from limited spatial coverage. Satellite data offer an alternative data source for exposure assessment.
We used birth certificate data for births in Connecticut and Massachusetts, U.S. (2000-2006). Gestational exposure to PM2.5 was estimated from US Environmental Protection Agency monitoring data and from satellite data. Satellite data were processed and modeled using 2 methods – denoted satellite (1) and satellite (2) – before exposure assessment. Regression models related PM2.5 exposure to birth outcomes while controlling for several confounders. Birth outcomes were mean birth weight at term birth, low birth weight at term (LBW <2500g), small for gestational age (SGA, <10th percentile for gestational age and sex), and preterm birth (<37 weeks).
Overall, the exposure assessment method modified the magnitude of the effect estimates of PM2.5 on birth outcomes. Change in birth weight per inter-quartile range (2.41 μg/m3)-increase in PM2.5 was -6g (95% confidence interval = -8 to -5), -16g (-21 to -11) and -19g (-23 to -15), using the monitor, satellite (1) and satellite (2) methods, respectively. Adjusted odds ratios, based on the same 3 exposure methods, for term LBW were 1.01 (0.98 to 1.04), 1.06 (0.97 to 1.16), and 1.08 (1.01 to 1.16); for SGA, 1.03 (1.01 to 1.04), 1.06 (1.03 to 1.10) and 1.08 (1.04 to 1.11); and for preterm birth, 1.00 (0.99 to 1.02), 0.98 (0.94 to 1.03) and 0.99 (0.95 to 1.03).
Under exposure assessment methods, we found associations between PM2.5 exposure and adverse birth outcomes particularly for birth weight among term births and for SGA. These results add to the growing concerns that air pollution adversely affects infant health and suggest that analysis of health consequences based on satellite-based exposure assessment can provide additional useful information.
Previous studies have found relationships between DNA methylation and various environmental contaminant exposures. Associations with weather have not been examined. Because temperature and humidity are related to mortality even on non-extreme days, we hypothesized that temperature and relative humidity may affect methylation.
We repeatedly measured methylation on long interspersed nuclear elements (LINE-1), Alu, and 9 candidate genes in blood samples from 777 elderly men participating in the normative aging Study (1999–2009). We assessed whether ambient temperature and relative humidity are related to methylation on LINE-1 and Alu, as well as on genes controlling coagulation, inflammation, cortisol, DNA repair, and metabolic pathway. We examined intermediate-term associations of temperature, relative humidity, and their interaction with methylation, using distributed lag models.
Temperature or relative humidity levels were associated with methylation on tissue factor (F3), intercellular adhesion molecule 1 (ICAM-1), toll-like receptor 2 (TRL-2), carnitine O-acetyltransferase (CRAT), interferon gamma (IFN-γ), inducible nitric oxide synthase (iNOS), and glucocorticoid receptor, LINE-1, and Alu. For instance, a 5°c increase in 3-week average temperature in ICAM-1 methylation was associated with a 9% increase (95% confidence interval: 3% to 15%), whereas a 10% increase in 3-week average relative humidity was associated with a 5% decrease (−8% to −1%). The relative humidity association with ICAM-1 methylation was stronger on hot days than mild days.
DNA methylation in blood cells may reflect biological effects of temperature and relative humidity. Temperature and relative humidity may also interact to produce stronger effects.
We present results that allow the researcher in certain cases to determine the direction of the bias that arises when control for confounding is inadequate. The results are given within the context of the directed acyclic graph causal framework and are stated in terms of signed edges. Rigorous definitions for signed edges are provided. We describe cases in which intuition concerning signed edges fails and we characterize the directed acyclic graphs that researchers can use to draw conclusions about the sign of the bias of unmeasured confounding. If there is only one unmeasured confounding variable on the graph, then non-increasing or non-decreasing average causal effects suffice to draw conclusions about the direction of the bias. When there are more than one unmeasured confounding variable, non-increasing and non-decreasing average causal effects can be used to draw conclusions only if the various unmeasured confounding variables are independent of one another conditional on the measured covariates. When this conditional independence property does not hold, stronger notions of monotonicity are needed to draw conclusions about the direction of the bias.
A key question in many studies is how to divide the total effect of an exposure into a component that acts directly on the outcome and a component that acts indirectly, i.e. through some intermediate. For example, one might be interested in the extent to which the effect of diet on blood pressure is mediated through sodium intake and the extent to which it operates through other pathways. In the context of such mediation analysis, even if the effect of the exposure on the outcome is unconfounded, estimates of direct and indirect effects will be biased if control is not made for confounders of the mediator-outcome relationship. Often data are not collected on such mediator-outcome confounding variables; the results in this paper allow researchers to assess the sensitivity of their estimates of direct and indirect effects to the biases from such confounding. Specifically, the paper provides formulas for the bias in estimates of direct and indirect effects due to confounding of the exposure-mediator relationship and of the mediator-outcome relationship. Under some simplifying assumptions, the formulas are particularly easy to use in sensitivity analysis. The bias formulas are illustrated by examples in the literature concerning direct and indirect effects in which mediator-outcome confounding may be present.
The reliability of retrospective time to pregnancy (TTP) has been established, but its validity has been assessed in only 1 study, which had a short follow-up.
Ninety-nine women enrolled a decade earlier in a prospective TTP study were queried by means of mailed questionnaires about the duration of time they had required to become pregnant. Their responses were compared with their earlier data from daily diaries (gold standard).
One-third of women could not recall their earlier TTP either in menstrual cycles or calendar months. Only 17%-19% of women recalled their TTP exactly. Agreement increased to 41%-51%, 65%-72%, and 72%-77% when defined as ±1, ±2, and ±3 months, respectively. Women with longer observed TTPs or previous pregnancies were more likely to under-report their TTP.
The findings raise questions about the commonly assumed validity of self-reported TTP. Recalled TTP may introduce error when estimating fecundability or classifying couples’ fecundity status.
During the 2009 influenza pandemic, uncertainty surrounding the seriousness of human infections with the H1N1pdm09 virus hindered appropriate public health response. One measure of seriousness is the case fatality risk, defined as the probability of mortality among people classified as cases.
We conducted a systematic review to summarize published estimates of the case fatality risk of the pandemic influenza H1N1pdm09 virus. Only studies that reported population-based estimates were included.
We included 77 estimates of the case fatality risk from 50 published studies, about one-third of which were published within the first 9 months of the pandemic. We identified very substantial heterogeneity in published estimates, ranging from less than 1 to more than 10,000 deaths per 100,000 cases or infections. The choice of case definition in the denominator accounted for substantial heterogeneity, with the higher estimates based on laboratory-confirmed cases (point estimates= 0–13,500 per 100,000 cases) compared with symptomatic cases (point estimates= 0–1,200 per 100,000 cases) or infections (point estimates=1–10 per 100,000 infections). Risk based on symptomatic cases increased substantially with age.
Our review highlights the difficulty in estimating the seriousness of infection with a novel influenza virus using the case fatality risk. In addition, substantial variability in age-specific estimates complicates the interpretation of the overall case fatality risk and comparisons among populations. A consensus is needed on how to define and measure the seriousness of infection before the next pandemic.
Studies of hypertension and cognition variously report adverse, null
and protective associations. We evaluated evidence supporting three
potential explanations for this variation: an effect of hypertension
duration, an effect of age at hypertension initiation, and selection bias
due to dependent censoring.
The Normative Aging Study is a prospective cohort study of men in the
greater Boston area. Participants completed study visits, including
hypertension assessment, every 3-5 years starting in 1961. 758 of 1284 men
eligible at baseline completed cognitive assessment between 1992 and 2005;
we used the mean age-adjusted cognitive test z-score from their first
assessment to quantify cognition. We estimated how becoming hypertensive and
increasing age at onset and duration since hypertension initiation affect
cognition. We used inverse probability of censoring weights to reduce and
quantify selection bias.
A history of hypertension diagnosis predicted lower cognition.
Increasing duration since hypertension initiation predicted lower mean
cognitive z-score (-0.02 standard units per year increase
[95% confidence interval= -0.04 to -0.001]),
independent of age at onset. Comparing participants with and without
hypertension, we observed noteworthy differences in mean cognitive score
only for those with a long duration since hypertension initiation,
regardless of age at onset. Age at onset was not associated with cognition
independent of duration. Analyses designed to quantify selection bias
suggested upward bias.
Previous findings of null or protective associations between
hypertension and cognition likely reflect the study of persons with short
duration since hypertension initiation. Selection bias may also contribute
to cross-study heterogeneity.
Despite the serious biases that characterize self-rated health, researchers rely heavily on these ratings to predict mortality. Using newly collected survey data, we examine whether simple ratings of participants' health provided by interviewers and physicians can markedly improve mortality prediction.
We use data from a prospective cohort study based on a nationally representative sample of older adults in Taiwan. We estimate proportional hazard models of all-cause mortality between the 2006 interview and 30 June 2011 (mean 4.7 years follow-up).
Interviewer ratings were more strongly associated with mortality than physician or self-ratings, even after controlling for a wide range of covariates. Neither respondent nor physician ratings substantially improve mortality prediction in models that include interviewer ratings. The predictive power of interviewer ratings likely arises in part from interviewers' incorporation of information about the respondents' physical and mental health into their assessments.
The findings of this study support the routine inclusion of a simple question at the end of face-to-face interviews, comparable to self-rated health, asking interviewers to provide an assessment of respondents' overall health. The costs of such an undertaking are minimal and the potential gains substantial for demographic and health researchers. Future work should explore the strength of the link between interviewer ratings and mortality in other countries and in surveys that collect less detailed information on respondent health, functioning, and well-being.
Air pollution, particularly from vehicle exhaust, has been shown to influence hormonal activity. However, at present, it is unknown whether air pollution exposure is associated with the occurrence of uterine leiomyomata, a hormonally sensitive tumor of the uterus.
Proximity to major roadways and outdoor levels of PM less than 10 microns (PM10) or 2.5 microns (PM2.5) or between 10 and 2.5 microns (PM10–2.5) in diameter were determined for all residential addresses from September 1989 to May 2007 for 85,251 women aged 25–42 at enrollment in the Nurses’ Health Study II who were alive and responding to questionnaires, premenopausal with intact uteri, without diagnoses of cancer, or prevalent uterine leiomyomata. Incidence of ultrasound- or hysterectomy-confirmed uterine leiomyomata and covariates were reported on biennial questionnaires sent through May 2007. Multivariable time-varying Cox proportional hazard models were used to estimate the relation between distance to road or PM exposures and uterine leiomyomata risk.
During 837,573 person-years of follow-up, there were 7,760 incident cases. Living close to a major road and exposures to PM10 or PM10–2.5 were not associated with an increased risk of uterine leiomyomata. However, each 10 µg/m3 increase in 2-year average, 4-year average, or cumulative average PM2.5 was associated with an adjusted hazard ratio (HR) of 1.08 (95% confidence interval (CI):1.00–1.17), 1.09 (95%CI:0.99–1.19) and 1.11 (95%CI:1.03–1.19), respectively.
Chronic exposure to PM2.5 may be associated with a modest increased risk of uterine leiomyomata.
Air pollution; Distance to road; Uterine Fibroids; Leiomyoma; Prospective cohort
In environmental epidemiology, we are often faced with two challenges. First, an exposure prediction model is needed to estimate the exposure to an agent of interest, ideally at the individual level. Second, when estimating the health-effect associated with the exposure, confounding adjustment is needed in the health-effects regression model. The current literature addresses these two challenges separately. That is, methods that account for measurement error in the predicted exposure often fail to acknowledge the possibility of confounding, while methods designed to control confounding often fail to acknowledge that the exposure has been predicted. In this paper, we consider exposure prediction and confounding adjustment in a health-effects regression model simultaneously. By using theoretical arguments and simulation studies, we show that the bias of a health-effect estimate is influenced by the exposure prediction model, the type of confounding adjustment used in the health-effects regression model, and the relationship between these two. Moreover, we argue that even with a health-effects regression model that properly adjusts for confounding, the use of a predicted exposure can bias the health-effect estimate unless all confounders included in the health-effects regression model are also included in the exposure prediction model. While these results of this paper were motivated by studies of environmental contaminants, they apply more broadly to any context where an exposure needs to be predicted.