The effect of misclassification in the presence of covariates

Approaches to harmonize mortality data sets in three diverse radiation worker cohorts

While there is a well-established link between ionizing radiation and cancer, there are uncertainties with effects following low doses delivered at low dose rates. To address these gaps, the ongoing Million Person Study of Radiation Workers and Veterans (MPS) is investigating the likelihood of a variety of cancer and non-cancer effects following chronic exposure to low dose-rate ionizing radiation. One challenge is and will be combining and harmonizing diverse cohorts with widely different measures of socio-economic status, birth cohorts, dose distributions and sex ratios. Herein, we have evaluated non-cancer mortality in three cohorts for which dose reconstructions have been completed: Rocketdyne (Atomics International, California, 1948-2008), Mound (Dayton, Ohio, 1944-2009) and nuclear weapons test participants (Atomic Veterans, 1945-2012). These three cohorts represent a small fraction of the overall MPS but provide valuable insight into methods of combining and harmonizing data from multiple diverse cohorts that can later be considered for all MPS cohorts. Heart disease mortality, including both underlying and contributing causes of death, was chosen for illustrating the statistical approaches. In all three cohorts, radiation dose estimates were distributed very differently by different measures of socio-economic status. Further, the effect of birth cohort was significantly different for heart disease mortality in all three cohorts, with all studies showing that later birth cohorts have lower rates of heart disease mortality than the earlier. The goal of this paper is not to quantify radiation effects based on these combined cohorts and it would be inappropriate to do so. Rather these cohorts are used to illustrate approaches for combining multiple data sets that incorporate the full set of individual confounder and cofactor information available from each cohort, though widely different. We identified five different methods to combine the results of these three datasets: the simple pooled analysis (PA), PA including study interactions, traditional stratified analysis, and both fixed and random effects meta-analysis. We describe the similarities and differences between the combined results using these approaches.

Association between planetary health diet and cardiovascular disease: a prospective study from the UK Biobank

AIMS

The Planetary Health Diet Index (PHDI) prioritizes the well-being of both individuals and the planet but has yielded mixed results on cardiovascular disease (CVD). The aim of this study was to assess the association between the PHDI and risk of CVD.

METHODS AND RESULTS

A cohort of 118 469 individuals aged 40-69 years from the UK Biobank, who were free of CVD at 2009-12 and followed up to 2021. The PHDI was calculated using at least two 24-h dietary assessments and included 14 food groups, with a possible range from 0 to 130 points. Cardiovascular disease incidence was defined as primary myocardial infarction or stroke and obtained from clinical records and death registries. During a 9.4-year follow-up, 5257 incident cases of CVD were ascertained. When comparing the highest (89.9-128.5 points) vs. the lowest (21.1-71.1 points) quartile of PHDI adherence, the multivariable-adjusted hazard ratio [95% confidence interval (CI)] was 0.86 (0.79, 0.94) for CVD, 0.88 (0.80, 0.97) for myocardial infarction, and 0.82 (0.70, 0.97) for stroke. The association was linear until a plateau effect was reached at 80 points of adherence to PHDI. Results remained robust when excluding participants with type 2 diabetes, including only those with three or more diet assessments, or excluding CVD cases in the first 3 years of follow-up. The food group components of the PHDI more strongly associated with a reduced CVD risk were higher consumption of whole grains, whole fruits, and fish and lower consumption of added sugars and fruit juices.

CONCLUSION

In this large cohort of middle-aged and older British adults, adherence to the PHDI was associated with a lower risk of CVD. These results provide empirical evidence that this dietary pattern, thought to be environmentally sustainable, benefits cardiovascular health.

Exploring bidirectional causality between religion and mental health: A longitudinal study using data from the parent generation of a UK birth cohort

Relations between religion and mental health have been studied extensively, yet whether associations are causal remains uncertain. Here, we use longitudinal data from the parental generation of the Avon Longitudinal Study of Parents and Children (ALSPAC), based in the UK, to assess: i) whether religiosity may cause subsequent depression and anxiety; ii) whether depression and anxiety may cause subsequent religiosity; and iii) whether there are gender differences in the above associations. All analyses were pre-registered, and adjusted for baseline confounders, exposures and outcomes in an attempt to rule out reverse causality and confounding bias. We found little conclusive evidence that religiosity was associated with subsequent mental health, or that mental health was associated with subsequent religiosity. Some weak associations were reported, but effect sizes were small and largely consistent with null effects. Small differences by gender were found, with religiosity marginally associated with better mental health in women and worse mental health in men, but the inconsistency of the results and the wide margins of error mean that firm conclusions cannot be made. In sum, in this UK population we find little evidence for bidirectional causation between religion and mental health, or for large differences in these associations by gender.

Familial confounding in the associations between maternal health and autism

Evidence suggests that maternal health in pregnancy is associated with autism in the offspring. However, most diagnoses in pregnant women have not been examined, and the role of familial confounding remains unknown. Our cohort included all children born in Denmark between 1998 and 2015 (n = 1,131,899) and their parents. We fitted Cox proportional hazard regression models to estimate the likelihood of autism associated with each maternal prenatal ICD-10 diagnosis, accounting for disease chronicity and comorbidity, familial correlations and sociodemographic factors. We examined the evidence for familial confounding using discordant sibling and paternal negative control designs. Among the 1,131,899 individuals in our sample, 18,374 (1.6%) were diagnosed with autism by the end of follow-up. Across 236 maternal diagnoses we tested (prevalence ≥0.1%), 30 were significantly associated with autism after accounting for sociodemographic factors, disorder chronicity and comorbidity, and correction for multiple testing. This included obstetric, cardiometabolic and psychiatric disorders (for example, diabetes in pregnancy (hazard ratio (HR) 1.19, 95% confidence interval (CI) 1.08-1.31) and depression (HR 1.49, 95% CI 1.27-1.75)), previously shown to be associated with autism. Family-based analyses provided strong evidence for familial confounding in most of the observed associations. Our findings indicate pervasive associations between maternal health in pregnancy and offspring autism and underscore that these associations are largely attributable to familial confounding.

Causal Inference in Oncology: Why, What, How and When

Oncologists are faced with choosing the best treatment for each patient, based on the available evidence from randomized controlled trials (RCTs) and observational studies. RCTs provide estimates of the average effects of treatments on groups of patients, but they may not apply in many real-world scenarios where for example patients have different characteristics than the RCT participants, or where different treatment variants are considered. Causal inference defines what a treatment effect is and how it may be estimated with RCTs or outside of RCTs with observational - or 'real-world' - data. In this review, we introduce the field of causal inference, explain what a treatment effect is and what important challenges are with treatment effect estimation with observational data. We then provide a framework for conducting causal inference studies and describe when in oncology causal inference from observational data may be particularly valuable. Recognizing the strengths and limitations of both RCTs and observational causal inference provides a way for more informed and individualized treatment decision-making in oncology.

Impact of observability period on the classification of COPD diagnosis timing among Medicare beneficiaries with lung cancer

BACKGROUND

Investigators often use claims data to estimate the diagnosis timing of chronic conditions. However, misclassification of chronic conditions is common due to variability in healthcare utilization and in claims history across patients.

OBJECTIVE

We aimed to quantify the effect of various Medicare fee-for-service continuous enrollment period and lookback period (LBP) on misclassification of COPD and sample size.

METHODS

A stepwise tutorial to classify COPD, based on its diagnosis timing relative to lung cancer diagnosis using the Surveillance Epidemiology and End Results cancer registry linked to Medicare insurance claims. We used 3 approaches varying the LBP and required continuous enrollment (i.e., observability) period between 1 to 5 years. Patients with lung cancer were classified based on their COPD related healthcare utilization into 3 groups: pre-existing COPD (diagnosis at least 3 months before lung cancer diagnosis), concurrent COPD (diagnosis during the -/+ 3months of lung cancer diagnosis), and non-COPD. Among those with 5 years of continuous enrollment, we estimated the sensitivity of the LBP to ascertain COPD diagnosis as the number of patients with pre-existing COPD using a shorter LBP divided by the number of patients with pre-existing COPD using a longer LBP.

RESULTS

Extending the LBP from 1 to 5 years increased prevalence of pre-existing COPD from ~ 36% to 51%, decreased both concurrent COPD from ~ 34% to 23% and non-COPD from ~ 29% to 25%. There was minimal effect of extending the required continuous enrollment period beyond one year across various LBPs. In those with 5 years of continuous enrollment, sensitivity of COPD classification (95% CI) increased with longer LBP from 70.1% (69.7% to 70.4%) for one-year LBP to 100% for 5-years LBP.

CONCLUSION

The length of optimum LBP and continuous enrollment period depends on the context of the research question and the data generating mechanisms. Among Medicare beneficiaries, the best approach to identify diagnosis timing of COPD relative to lung cancer diagnosis is to use all available LBP with at least one year of required continuous enrollment.

A Comparison of Green, Delta, and Monte Carlo Methods to Select an Optimal Approach for Calculating the 95% Confidence Interval of the Population-attributable Fraction: Guidance for Epidemiological Research

OBJECTIVES

This study aimed to compare the Delta, Greenland, and Monte Carlo methods for estimating 95% confidence intervals (CIs) of the population-attributable fraction (PAF). The objectives were to identify the optimal method and to determine the influence of primary parameters on PAF calculations.

METHODS

A dataset was simulated using hypothetical values for primary parameters (population, relative risk [RR], prevalence, and variance of the beta estimator ) involved in PAF calculations. Three methods (Delta, Greenland, and Monte Carlo) were used to estimate the 95% CIs of the PAFs. Perturbation analysis was performed to assess the sensitivity of the PAF to changes in these parameters. An R Shiny application, the "GDM-PAF CI Explorer," was developed to facilitate the analysis and visualization of these computations.

RESULTS

No significant differences were observed among the 3 methods when both the RR and p-value were low. The Delta method performed well under conditions of low prevalence or minimal RR, while Greenland's method was effective in scenarios with high prevalence. Meanwhile, the Monte Carlo method calculated 95% CIs of PAFs that were stable overall, though it required intensive computational resources. In a novel approach that utilized perturbation for sensitivity analysis, was identified as the most influential parameter in the estimation of CIs.

CONCLUSIONS

This study emphasizes the necessity of a careful approach for comparing 95% CI estimation methods for PAFs and selecting the method that best suits the context. It provides practical guidelines to researchers to increase the reliability and accuracy of epidemiological studies.

Race adjustments in clinical algorithms can help correct for racial disparities in data quality

Despite ethical and historical arguments for removing race from clinical algorithms, the consequences of removal remain unclear. Here, we highlight a largely undiscussed consideration in this debate: varying data quality of input features across race groups. For example, family history of cancer is an essential predictor in cancer risk prediction algorithms but is less reliably documented for Black participants and may therefore be less predictive of cancer outcomes. Using data from the Southern Community Cohort Study, we assessed whether race adjustments could allow risk prediction models to capture varying data quality by race, focusing on colorectal cancer risk prediction. We analyzed 77,836 adults with no history of colorectal cancer at baseline. The predictive value of self-reported family history was greater for White participants than for Black participants. We compared two cancer risk prediction algorithms-a race-blind algorithm which included standard colorectal cancer risk factors but not race, and a race-adjusted algorithm which additionally included race. Relative to the race-blind algorithm, the race-adjusted algorithm improved predictive performance, as measured by goodness of fit in a likelihood ratio test (P-value: <0.001) and area under the receiving operating characteristic curve among Black participants (P-value: 0.006). Because the race-blind algorithm underpredicted risk for Black participants, the race-adjusted algorithm increased the fraction of Black participants among the predicted high-risk group, potentially increasing access to screening. More broadly, this study shows that race adjustments may be beneficial when the data quality of key predictors in clinical algorithms differs by race group.

Multigenerational association between smoking and autism spectrum disorder: findings from a nationwide prospective cohort study

Animal studies have shown that exposure to cigarette smoke during pregnancy can induce neurobehavioral anomalies in multiple subsequent generations. However, little work has examined such effects in humans. We examined the risk of grandchild autism spectrum disorder (ASD) in association with grandmother's smoking during pregnancy, using data from 53 562 mothers and grandmothers and 120 267 grandchildren in Nurses' Health Study II. In 1999, Nurses' Health Study II participants with children reported on their mothers' smoking. Grandchildren's ASD diagnoses were reported by the mothers in 2005 and 2009. Among grandmothers, 13 383 (25.0%) smoked during pregnancy, and 509 (0.4%) grandchildren were diagnosed with ASD. The adjusted odds ratio for ASD for grandmother smoking during pregnancy was 1.52 (95% CI, 1.06-2.20). Results were similar with direct grandmother reporting in 2001 of her smoking during pregnancy from the Nurses' Mothers Cohort Study subgroup (n = 22 167 grandmothers, n = 49 917 grandchildren) and were stronger among grandmothers who smoked ≥15 cigarettes per day during pregnancy (adjusted odds ratio = 1.93 [95% CI, 1.10-3.40]; n = 1895 grandmothers, n = 4212 grandchildren). Results were similar when we adjusted for mother's smoking during pregnancy. There was no association with grandfather's smoking as reported by the grandmother. Our results suggest a potential persistent impact of gestational exposure to environmental insults across 3 generations.

Evaluation of Insurance Type as a Proxy for Socioeconomic Status in the Pediatric Emergency Department: A Pilot Study

STUDY OBJECTIVE

To determine whether insurance status can function as a sufficient proxy for socioeconomic status in emergency medicine research by examining the concordance between insurance status and direct socioeconomic status measures in a sample of pediatric patients.

METHODS

We conducted a cross-sectional pilot study of patients aged 5 to 17 years in the emergency department of a quaternary care children's hospital. Socioeconomic status was measured using the highest level of the caregiver's education (low: less than bachelor's degree; high: bachelor's or greater) and previous year household income (low: <$75,000; high: ≥$75,000). We calculated the misclassification rate of insurance status (low: public; high: private) using education and income as reference standards. Results were expressed as percentages with 95% confidence intervals.

RESULTS

In total, 300 patients were enrolled (median age 11 years, 44% female). Insurance status misclassified 23% (95% CI 18% to 28%) and 14% (95% CI 10% to 19%) of patients when using caregiver education and income, respectively, as reference standards.

CONCLUSIONS

Insurance status misclassified socioeconomic status in up to 23% of pediatric patients, as measured by caregivers' education and income. Emergency medicine studies of pediatric patients using insurance as a covariate to adjust for socioeconomic status may need to consider this misclassification and the resulting potential for bias. These findings require confirmation in larger, more diverse samples, including adult patients.

ABCG2 polymorphisms and susceptibility to ARV-associated hepatotoxicity

BACKGROUND

The ABCG2 421C/A polymorphism contributes significantly to the distribution and absorption of antiretroviral (ARV) regimens and is associated with the undesirable side effects of efavirenz.

METHODS

To investigate this, we examined ABCG2 34G/A (rs2231137) and 421C/A (rs2231142) genetic variations in 149 HIV-infected patients (116 without hepatotoxicity, 33 with ARV-induced hepatotoxicity) and 151 healthy controls through the PCR-restriction fragment length polymorphism (PCR-RFLP) technique.

RESULTS AND DISCUSSION

The ABCG2 34GA genotype and 34A allele indicated a risk for antiretroviral therapy-associated hepatotoxicity development (p = 0.09, OR = 1.58, 95% CI: 0.93-2.69; p = 0.06, OR = 1.50, 95% CI: 0.98-2.30). The haplotype GA was associated with hepatotoxicity (p = 0.042, OR = 2.37, 95% CI: 1.04-5.43; p = 0.042, OR = 2.49, 95% CI: 1.04-5.96). Moreover, when comparing HIV patients with hepatotoxicity to healthy controls, the haplotype GA had an association with an elevated risk for the development of hepatotoxicity (p = 0.041, OR = 1.73, 95% CI: 1.02-2.93). Additionally, the association of the ABCG2 34GA genotype with the progression of HIV (p = 0.02, OR = 1.97, 95% CI: 1.07-3.63) indicated a risk for advanced HIV infection. Furthermore, the ABCG2 421AA genotype was linked to tobacco users and featured as a risk factor for the progression of HIV disease (p = 0.03, OR = 11.07, 95% CI: 1.09-270.89).

CONCLUSION

The haplotype GA may enhance the risk of hepatotoxicity development and its severity. Individuals with the ABCG2 34A allele may also be at risk for the development of hepatotoxicity. Additionally, individuals with an advanced stage of HIV and the ABCG2 34GA genotype may be at risk for disease progression.

The Southern European Atlantic diet and all-cause and cause-specific mortality: a European multicohort study

AIMS

The Southern European Atlantic diet (SEAD) is the traditional dietary pattern of northwestern Spain and northern Portugal, but it may resemble that of central, eastern, and western European countries. The SEAD has been found associated with lower risk of myocardial infarction and mortality in older adults, but it is uncertain whether this association also exists in other European populations and if it is similar as that found in its countries of origin.

METHODS AND RESULTS

We conducted a prospective analysis of four cohorts with 35 917 subjects aged 18-96 years: ENRICA (Spain), HAPIEE (Czechia and Poland), and Whitehall II (United Kingdom). The SEAD comprised fresh fish, cod, red meat and pork products, dairy, legumes and vegetables, vegetable soup, potatoes, whole-grain bread, and moderate wine consumption. Associations were adjusted for sociodemographic variables, energy intake, lifestyle, and morbidity. After a median follow-up of 13.6 years (range = 0-15), we recorded 4 973 all-cause, 1 581 cardiovascular, and 1 814 cancer deaths. Higher adherence to the SEAD was associated with lower mortality in the pooled sample. Fully adjusted hazard ratios and 95% confidence interval per 1-standard deviation increment in the SEAD were 0.92 (0.89, 0.95), 0.91 (0.86, 0.96), and 0.94 (0.89, 0.99) for all-cause, cardiovascular, and cancer mortality, respectively. The association of the SEAD with all-cause mortality was not significantly different between countries [Spain = 0.93 (0.88, 0.99), Czechia = 0.94 (0.89,0.99), Poland = 0.89 (0.85, 0.93), United Kingdom = 0.98 (0.89, 1.07); P for interaction = 0.16].

CONCLUSION

The SEAD was associated with lower all-cause, cardiovascular, and cancer mortality in southern, central, eastern, and western European populations. Associations were of similar magnitude as those found for existing healthy dietary patterns.

The practice of consanguineous marriage and the risk of diabetes among offspring in the province of Tetouan (Morocco)

Purpose

The practice of consanguinity has been culturally preferred in most Arab countries, including Morocco. This behavior leads to an increase in genetic abnormalities, such as hypertension and diabetes. This paper examines the prevalence and determinants of first-cousin marriages and their impact on diabetes among offspring.

Design/methodology/approach

Data on 882 couples were collected through face-to-face interview via a pre-established questionnaire based on the variables selected within the objectives of this study. The authors used the multiple logistic regression modeling procedure in this study.

Findings

The results of the study indicate that the prevalence of first-cousin marriages were 15% among students’ parents. From the multiple logistic regression modeling, the authors found a significant effect of paternal and maternal grandparents’ first-cousins marriage on that of parents (aOR = 3.27 and aOR = 3.36, respectively). However, an 11-fold higher risk of first relative marriages among parents once the paternal and maternal grandparents were first-cousins and the father was illiterate (aOR = 11.01). Moreover, the authors reported a diabetes risk of more than 14 times when the effects of first-cousin maternal grandparents and parents and the hypertension among mother or her sibling were combined (aOR = 14.48) or when the effects of first-cousins maternal grandparents, first-cousin parents and mother’s age at marriage between 21 and 29 years were combined (aOR = 14.56).

Originality/value

First-cousin marriage depends on the father’s illiteracy and the consanguinity of grandparents’ factors. The cumulative effect of first-cousin marriage among grandparents, parents and a family history of hypertension among mother or her sibling increase the risk of diabetes among these mothers.

Diagnosis of behavioral symptoms as a predictor of institutionalization among Medicaid patients with dementia

OBJECTIVES

Behavioral symptoms are commonly observed in the course of dementia. This study aimed to assess the association of the diagnosis of a cluster of behavioral symptoms (e.g., agitation, aggression, psychotic symptoms, and delirium/wandering) with the likelihood of subsequent institutionalization.

METHODS

A retrospective cohort study of adults aged 65 and above diagnosed with dementia identified in the IBM® MarketScan® Multistate Medicaid database between October 01, 2015, and September 30, 2019, was conducted. The index date was defined as the first diagnosis date of dementia. The presence or absence of behavioral symptoms was identified in the 6 months prior to the index date (baseline). Institutionalization was evaluated 12 months (follow-up) post the index date. The association between diagnosed behavioral symptoms during the baseline period and institutionalization in the follow-up period was assessed using a multivariable logistic regression, adjusting for baseline sociodemographic and clinical characteristics.

RESULTS

The study cohort included 40,714 patients with dementia. A diagnosis of behavioral symptoms was found among 2,067 (5.1%) patients during the baseline period. An increased likelihood of institutionalization was found during the follow-up among patients with agitation and aggression in baseline (OR = 1.51 (95% CI: 1.18-1.92)) compared to patients without these symptoms at baseline. Patients with psychotic symptoms in baseline had significantly higher odds of getting institutionalized during the follow-up compared to patients without psychotic symptoms in baseline (OR = 1.36 (95% CI: 1.20-1.54)). Similarly, patients with symptoms of delirium and wandering in baseline had a higher likelihood of institutionalization than patients without these symptoms at baseline (OR = 1.61 (95% CI: 1.30-1.99)).

CONCLUSION

Several diagnosed behavioral symptoms were associated with a higher risk of institutionalization among older adults with dementia and should be considered when planning treatment strategies for the effective management of the condition.

Doxycycline post exposure prophylaxis could select for cross-resistance to other antimicrobials in various pathogens: An in silico analysis

BACKGROUND

A number of randomized controlled trials have found that doxycycline post exposure prophylaxis (PEP) can reduce the incidence of gonorrhoea, chlamydia and syphilis in men who have sex with men (MSM). If tetracycline resistance is associated with resistance to other antimicrobials in a range of bacterial species, then doxycycline PEP could have the unintended effect of selecting for resistance to other antimicrobials in these bacterial species.

METHODS

Antimicrobial susceptibility data were retrieved from two sources: pubMLST (https://pubmlst.org/) and Pathogenwatch (https://pathogen.watch/) for the following bacterial pathogens: Klebsiella pneumoniae, Salmonella enterica subsp. Enterica serovar Typhi, Campylobacter jejuni, Staphylococcus aureus, Streptococcus pneumoniae and Streptococcus pyogenes. We assessed if tetracycline resistance was associated with resistance to six relevant antimicrobials.

RESULTS

We found evidence of cross resistance to various antimicrobials in all six bacterial species assessed. Cross resistance was found in 4 of 5 antimicrobials for K. pneumoniae, 1 of 2 for C. jejuni, 3 of 5 for S. enterica subsp. Enterica serovar Typhi, 5 of 5 for S. aureus, 5 of 6 for S. pneumoniae and 2 of 3 for S. pyogenes. These associations include a higher prevalence of methicillin resistance in tetracycline resistant S. aureus, penicillin resistance in S. pneumoniae, macrolide and clindamycin resistance in S. pyogenes, fluoroquinolone resistance in S. enterica subsp. Enterica serovar Typhi and third-generation cephalosporin resistance in K. pneumoniae.

CONCLUSION

These results suggest that studies evaluating the effects of doxycycline PEP should include the effects of doxycycline on resistance not only to doxycycline but also to other antimicrobials and in a broader array of bacterial species than has been included in doxycycline PEP studies thus far.

Long-term exposure to low-level particulate air pollution and Parkinson's disease diagnosis - A Finnish register-based study

BACKGROUND

There is mixed evidence for an association between particulate matter air pollution and Parkinson's disease despite biological plausibility.

OBJECTIVES

We studied the association between particulate air pollution, its components and Parkinson's disease (PD) risk.

METHODS

We conducted a nested case-control study within the population of Finland using national registers. A total of 22,189 incident PD cases diagnosed between 1996 and 2015 were matched by age, sex and region with up to seven controls (n = 148,009) per case. Time weighted average air pollution exposure to particulate matter and its components was modelled at the residential addresses, accounting for move history, for the 16 years preceding diagnosis. Conditional logistic regression analysis was used to evaluate the association between air pollution and PD. Different exposure periods (6-16 years, 11-16 years, 5-10 years, 0-5 years) before the index date (date of PD diagnosis) were applied.

RESULTS

Time-weighted average exposures were relatively low at 12.1 ± 6.5 μg/m3 (mean ± SD) for PM10 and 7.7 ± 3.2 μg/m3 for PM2.5. No associations were found between PM2.5 or PM10 exposure 6-16 years before index date and PD (OR: 0.99; 95% CI: 0.96, 1.02; per IQR of 3.9 μg/m3 and OR: 0.99; 95% CI: 0.96, 1.01; per IQR of 7.8 μg/m3, respectively). However, inverse associations were observed for the same exposure period with black carbon (OR: 0.96; 95% CI: 0.93, 0.99; per IQR of 0.6 μg/m3), sulphate (OR: 0.79; 95% CI: 0.68, 0.92; per IQR of 1.2 μg/m3), secondary organic aerosols (OR: 0.86; 95% CI: 0.80, 0.93; per IQR of 0.1 μg/m3) and sea salt (OR: 0.92; 95% CI: 0.87, 0.98; per IQR of 0.1 μg/m3).

DISCUSSION

Low-level particulate matter air pollution was not associated with increased risk of incident PD in this Finnish nationwide population. The observed weak inverse associations with specific particle components should be investigated further.

Exploring causality from observational data: An example assessing whether religiosity promotes cooperation

Causal inference from observational data is notoriously difficult, and relies upon many unverifiable assumptions, including no confounding or selection bias. Here, we demonstrate how to apply a range of sensitivity analyses to examine whether a causal interpretation from observational data may be justified. These methods include: testing different confounding structures (as the assumed confounding model may be incorrect), exploring potential residual confounding and assessing the impact of selection bias due to missing data. We aim to answer the causal question 'Does religiosity promote cooperative behaviour?' as a motivating example of how these methods can be applied. We use data from the parental generation of a large-scale (n = approximately 14,000) prospective UK birth cohort (the Avon Longitudinal Study of Parents and Children), which has detailed information on religiosity and potential confounding variables, while cooperation was measured via self-reported history of blood donation. In this study, there was no association between religious belief or affiliation and blood donation. Religious attendance was positively associated with blood donation, but could plausibly be explained by unmeasured confounding. In this population, evidence that religiosity causes blood donation is suggestive, but rather weak. These analyses illustrate how sensitivity analyses can aid causal inference from observational research.

Role of APOC3 3238C/G polymorphism in HIV-associated neurocognitive disorder

Apolipoprotein not only have a role in cholesterol metabolism but also play a role in normal brain function. Apolipoprotein gene polymorphisms are known risk factors for a number of mental and neurological disorders. The expression of brain apolipoproteins is significantly altered in several brain disorders. Therefore, we assed ApoC33238 C/G polymorphism in a total of 248 patient infected with HIV (45 with HAND, 89 without HAND, 114 without ART) and 134 healthy controls using PCR-RFLP. ApoC3 3238CG, 3238 GG genotypes and 3238G allele showed a non-significant increased risk for severity of HAND (P = 0.16, OR = 1.83; P = 0.32, OR = 2.78; P = 0.10, OR = 1.65) while comparing individuals with and without HAND. ApoC3 3238 GG genotype and 3238G allele revealed an increased risk for disease progression when compared between HIV patients with and without ART (P = 0.55, OR = 1.76; P = 0.65, OR = 1.12) though risk could not reach statistical significance. ApoC3 3238 GG genotype and 3238G allele were associated with the reduced risk of acquiring HIV infection when comparing HIV patients who are not on ART with healthy controls (P = 0.05, OR = 0.29; P = 0.04, OR = 0.66). In HIV patients on ART,ApoC3 3238 GG genotype showed an increased susceptibility to development of HAND (P = 0.48, OR = 2.24) when comparing alcohol drinkers and non-drinkers however risk could not reach statistical significance. In conclusion, the genotype ApoC33238GG displayed an inclination of risk for the severity of HAND and HIV disease progression. The polymorphism of APOC3 3238C/G may have a role to reduce the risk for acquisition of HIV infection. ApoC33238GG genotype in presence of alcohol may increase susceptibility to development of HAND.

The rate of hepatitis C reinfection in Canadians coinfected with HIV and its implications for national elimination

BACKGROUND

The World Health Organisation (WHO) has set targets for the rate of new infections as a way to measure progress towards the elimination of hepatitis C virus (HCV) as a public health threat. As more people are successfully treated for HCV, a higher proportion of new infections will be reinfections. We consider whether the reinfection rate has changed since the interferon era and what we can infer about national elimination efforts from the current reinfection rate.

METHODS

The Canadian Coinfection Cohort is representative of HIV HCV coinfected people in clinical care. We selected cohort participants successfully treated for a primary HCV infection either in the interferon era or in the era of direct acting antivirals (DAAs). Selected participants were followed from 12 weeks after completing a successful treatment until the end of 2019 or until their last measured HCV RNA. We estimated the reinfection rate in each treatment era, overall and in participant subgroups, using proportional hazard models appropriate for interval censored data.

RESULTS

Among 814 successfully treated participants with additional HCV RNA measurements, there were 62 reinfections. The overall reinfection rate was 2.6 (95% confidence interval, CI, 1.2-4.1) /100 person years (PY) in the interferon era and 3.4 (95% CI 2.5-4.4) /100 PY in the DAA era. The rate in those reporting injection drug use (IDU) was much higher: 4.7 (95% CI 1.4-7.9) /100 PY and 7.6 (95% CI 5.3-10) /100 PY in the interferon and DAA eras respectively.

CONCLUSION

The overall reinfection rate in our cohort is now above the WHO target set for new infections in people who inject drugs. The reinfection rate in those reporting IDU has increased since the interferon era. This suggests Canada is not on track to achieve HCV elimination by 2030.

Longitudinal effects of lipid indices on incident cardiovascular diseases adjusting for time-varying confounding using marginal structural models: 25 years follow-up of two US cohort studies

Background

This study assesses the effect of blood lipid indices and lipid ratios on cardiovascular diseases (CVDs) using inverse probability-of-exposure weighted estimation of marginal structural models (MSMs).

Methods

A pooled dataset of two US representative cohort studies, including 16736 participants aged 42-84 years with complete information at baseline, was used. The effect of each lipid index, including low-density lipoprotein cholesterol (LDL-C), high-density lipoprotein cholesterol (HDL-C), triglyceride (TG), ratios of TC/HDL-C, LDL-C/HDL-C, and TG/HDL-C on coronary heart disease (CHD) and stroke were estimated using weighted Cox regression.

Results

There were 1638 cases of CHD and 1017 cases of stroke during a median follow-up of 17.1 years (interquartile range: 8.5 to 25.7). Compared to optimal levels, the risk of CVD outcomes increased substantially in high levels of TC, LDL-C, TC/HDL-C, and LDL-C/HDL-C. If everyone had always had high levels of TC (≥240 mg/dL), risk of CHD would have been 2.15 times higher, and risk of stroke 1.35 times higher than if they had always had optimal levels (<200 mg/dL). Moreover, if all participants had been kept at very high (≥190 mg/dL) levels of LDL-C, risk of CHD would have been 2.62 times higher and risk of stroke would have been 1.92 times higher than if all participants had been kept at optimal levels, respectively. Our results suggest that high levels of HDL-C may be protective for CHD, but not for stroke. There was also no evidence of an adverse effect of high triglyceride levels on stroke.

Conclusions

Using MSM, this study highlights the effect of TC and LDL-C on CVD, with a stronger effect on CHD than on stroke. There was no evidence for a protective effect of high levels of HDL-C on stroke. Besides, triglyceride was not found to affect stroke.

Schizophrenia, but not depression or bipolar affective disorder, adds additional risk of aspiration pneumonia among stroke survivors: A national cohort study in Taiwan

OBJECTIVE

Few studies have assessed the sex-specific and age-specific risk of aspiration pneumonia (AP) in patients with stroke and evaluated whether mental disorders may increase this risk. In this population-based cohort study, we investigated the sex-specific and age-specific risk of AP in association with stroke and the joint effects of stroke and mental disorders on the risk of AP.

METHODS

We included 23,288 patients with incident stroke admitted between 2005 and 2017 and 68,675 matched nonstroke controls. Information on mental disorders was obtained from medical claims data within the 3 years before the stroke incidence. Cox proportional hazards models considering death as a competing risk event were constructed to estimate the hazard ratio of AP incidence by the end of 2018 associated with stroke and selected mental disorders.

RESULTS

After ≤14 years of follow-up, AP incidence was higher in the patients with stroke than in the controls (11.30/1000 vs. 1.51/1000 person-years), representing a covariate-adjusted subdistribution hazard ratio (sHR) of 3.64, with no significant sex difference. The sHR significantly decreased with increasing age in both sexes. Stratified analyses indicated schizophrenia but not depression or bipolar affective disorder increased the risk of AP in the patients with stroke.

CONCLUSION

Compared with their corresponding counterparts, the patients with schizophrenia only, stroke only, and both stroke and schizophrenia had a significantly higher sHR of 4.01, 5.16, and 8.01, respectively. The risk of AP was higher in younger stroke patients than those older than 60 years. Moreover, schizophrenia was found to increase the risk of AP in patients with stroke.

Role of MMP-13-77A/G polymorphism in HIV-associated neurocognitive disorders patients

Many diseases including HIV-Associated Neurocognitive Disorder (HAND) are impacted by matrix metalloproteinases (MMPs). MMP-13 play a role to cleave the collagen. MMP-13 contributes to peripheral neuropathy and induces unmyelinated axon degeneration. MMP-13-77A/G polymorphism has been associated to a lower level of MMP-13. MMP-13 have been linked to increased expression in a number of diseases including neurological disease. Hence we analyzed the effect of MMP-13-77A/G polymorphism in pateints with and without HAND. The PCR-Restriction fragment length polymorphism approach was used to genotype MMP-13-77A/G polymorphism. The MMP-13-77AG genotype was shown to be more prevalent in HAND patients than in controls and showed a risk for severe HAND (44.4% vs. 34.8%, P = 0.16, OR = 1.79). When compared to healthy controls, the MMP-13-77AG genotype was found to be prevalent in HAND patients (44.4 %vs. 38.2%, P = 0.66, OR = 1.26). MMP-13-77AG genotype was overrepresented (51.5% vs. 38.2%, OR = 1.70, P = 0.29) in HAND patients who had advanced HIV disease. In without HAND patients, the MMP-13-77AG genotype was found be lessor in advanced stage of HIV disease when compared with healthy controls and it was associated with a reduced risk for advancement in disease (38.2% vs. 11.82%, P = 0.03, OR = 0.18). Smokers were more likely to have the MMP-13-77AG genotype than non-smokers, indicating an elevated risk of HAND severity (60.0% vs. 40.0%, P = 0.50, OR = 2.29, 95%). In patients with and without HAND, alcohol intake enhanced the risk for developing HAND and its severity when the MMP-13-77GG genotype was present (P = 0.78, OR = 2.10, P = 0.78, OR = 2.10). In conclusion, Individuals with alcohol usage and the MMP-13-77GG genotype may have additive effect on HAND development and its severity. Individuals of without HAND and MMP-13-77AG genotype showed reduced risk for advancement of HIV disease.

Misconceptions About the Direction of Bias From Nondifferential Misclassification

Measurement error is pervasive in epidemiologic research. Epidemiologists often assume that mismeasurement of study variables is nondifferential with respect to other analytical variables and then rely on the heuristic that "nondifferential misclassification will bias estimates towards the null." However, there are many exceptions to the heuristic for which bias towards the null cannot be assumed. In this paper, we compile and characterize 7 exceptions to this rule and encourage analysts to take a more critical and nuanced approach to evaluating and discussing bias from nondifferential mismeasurement.

Smoking During Pregnancy and Risk of Attention-deficit/Hyperactivity Disorder in the Third Generation

BACKGROUND

Animal experiments indicate that environmental factors, such as cigarette smoke, can have multigenerational effects through the germline. However, there are little data on multigenerational effects of smoking in humans. We examined the associations between grandmothers' smoking while pregnant and risk of attention-deficit/hyperactivity disorder (ADHD) in her grandchildren.

METHODS

Our study population included 53,653 Nurses' Health Study II (NHS-II) participants (generation 1 [G1]), their mothers (generation 0 [G0]), and their 120,467 live-born children (generation 2 [G2]). In secondary analyses, we used data from 23,844 mothers of the nurses who were participants in the Nurses' Mothers' Cohort Study (NMCS), a substudy of NHS-II.

RESULTS

The prevalence of G0 smoking during the pregnancy with the G1 nurse was 25%. ADHD was diagnosed in 9,049 (7.5%) of the grandchildren (G2). Grand-maternal smoking during pregnancy was associated with increased odds of ADHD among the grandchildren (adjusted odds ratio [aOR] = 1.2; 95% confidence interval [CI] = 1.1, 1.2), independent of G1 smoking during pregnancy. In the Nurses' Mothers' Cohort Study, odds of ADHD increased with increasing cigarettes smoked per day by the grandmother (1-14 cigarettes: aOR = 1.1; 95% CI = 1.0, 1.2; 15+: aOR = 1.2; 95% CI = 1.0, 1.3), compared with nonsmoking grandmothers.

CONCLUSIONS

Grandmother smoking during pregnancy is associated with an increased risk of ADHD among the grandchildren.

Resting heart rate and risk of left and right heart failure in 0.5 million Chinese adults

OBJECTIVES

To compare the shape and strength of the associations of resting heart rate (RHR) with incident heart failure (HF) and pulmonary heart disease (PHD) in Chinese adults.

METHODS

The prospective China Kadoorie Biobank recruited >0.5 million adults from 10 geographically diverse regions (5 urban, 5 rural) of China during 2004-2008. After an 11-year follow-up, 6082 incident cases of HF and 5572 cases of PHD, were recorded among 491 785 participants with no prior history of heart disease or use of beta-blockers at baseline. Cox regression yielded HRs for each disease associated with usual RHR after adjustment for confounding factors.

RESULTS

The mean (SD) baseline RHR was 79 (12) (men 78 (12); women 80 (11)) bpm, and these decreased with increasing age (by about 1 bpm per 10 years). Usual RHR showed J-shaped associations with HF and log-linear associations PHD. For HF, each 10 bpm higher usual RHR was associated with an adjusted HR of 1.25 (95% CI 1.17 to 1.34) for RHR>75 bpm. For PHD, each 10 bpm higher RHR was associated with HR of 1.74 (1.67-1.81) across the full range of usual RHR. For HF at RHR>75 bpm but not PHD, the HRs per 10 bpm higher RHR were approximately halved by further adjustment for diabetes and hypertension.

CONCLUSIONS

RHR was strongly positively associated with PHD throughout the range studied, but was only associated with HF at RHR>75 bpm, and the strength of the associations with HF were only one-third of those with PHD.

Testing the specificity of environmental risk factors for developmental outcomes

Developmental theories often assume that specific environmental risks affect specific outcomes. Canonical Correlation Analysis was used to test whether 28 developmental outcomes (measured at 11-15 years) share the same early environmental risk factors (measured at 0-3 years), or whether specific outcomes are associated with specific risks. We used data from the UK Millennium Cohort Study (N = 10,376, 51% Female, 84% White) collected between 2001 and 2016. A single environment component was mostly sufficient for explaining cognition and parent-rated behavior outcomes. In contrast, adolescents' alcohol and tobacco use were specifically associated with their parents', and child-rated mental health was weakly associated with all risks. These findings suggest that with some exceptions, many different developmental outcomes share the same early environmental risk factors.

Individual treatment effect estimation in the presence of unobserved confounding using proxies: a cohort study in stage III non-small cell lung cancer

Randomized Controlled Trials (RCT) are the gold standard for estimating treatment effects but some important situations in cancer care require treatment effect estimates from observational data. We developed “Proxy based individual treatment effect modeling in cancer” (PROTECT) to estimate treatment effects from observational data when there are unobserved confounders, but proxy measurements of these confounders exist. We identified an unobserved confounder in observational cancer research: overall fitness. Proxy measurements of overall fitness exist like performance score, but the fitness as observed by the treating physician is unavailable for research. PROTECT reconstructs the distribution of the unobserved confounder based on these proxy measurements to estimate the treatment effect. PROTECT was applied to an observational cohort of 504 stage III non-small cell lung cancer (NSCLC) patients, treated with concurrent chemoradiation or sequential chemoradiation. Whereas conventional confounding adjustment methods seemed to overestimate the treatment effect, PROTECT provided credible treatment effect estimates.

The confounder matrix: A tool to assess confounding bias in systematic reviews of observational studies of etiology

Systematic reviews and meta-analyses are essential for drawing conclusions regarding etiologic associations between exposures or interventions and health outcomes. Observational studies comprise a substantive source of the evidence base. One major threat to their validity is residual confounding, which may occur when component studies adjust for different sets of confounders, fail to control for important confounders, or have classification errors resulting in only partial control of measured confounders. We present the confounder matrix-an approach for defining and summarizing adequate confounding control in systematic reviews of observational studies and incorporating this assessment into meta-analyses. First, an expert group reaches consensus regarding the core confounders that should be controlled and the best available method for their measurement. Second, a matrix graphically depicts how each component study accounted for each confounder. Third, the assessment of control adequacy informs quantitative synthesis. We illustrate the approach with studies of the association between short interpregnancy intervals and preterm birth. Our findings suggest that uncontrolled confounding, notably by reproductive history and sociodemographics, resulted in exaggerated estimates. Moreover, no studies adequately controlled for all core confounders, so we suspect residual confounding is present, even among studies with better control. The confounder matrix serves as an extension of previously published methodological guidance for observational research synthesis, enabling transparent reporting of confounding control and directly informing meta-analysis so that conclusions are drawn from the best available evidence. Widespread application could raise awareness about gaps across a body of work and allow for more valid inference with respect to confounder control.

Grandmothers' endocrine disruption during pregnancy, low birth weight, and preterm birth in third generation

BACKGROUND

Diethylstilbestrol (DES) is an endocrine-disrupting pharmaceutical prescribed to pregnant women to prevent pregnancy complications between the 1940s and 1970s. Although DES has been shown in animal studies to have multigenerational effects, only two studies have investigated potential multigenerational effects in humans on preterm birth (PTB), and none on low birthweight (LBW)-major determinants of later life health.

METHODS

Nurses' Health Study (NHS) II participants (G1; born 1946-64) reported their mothers' (G0) use of DES while pregnant with them. We used cluster-weighted generalized estimating equations to estimate odds ratios (OR) and 95% confidence intervals (CI) for risk of LBW and PTB among the grandchildren by grandmother use of DES. G1 birthweight and gestational age were considered to explore confounding by indication.

RESULTS

Among 54 334 G0-G1/grandmother-mother pairs, 973 (1.8%) G0 used DES during pregnancy with G1. Of the 128 275 G2 children, 4369 (3.4%) were LBW and 7976 (6.2%) premature. Grandmother (G0) use of DES during pregnancy was associated with an increased risk of G2 LBW [adjusted OR (aOR) = 3.09; 95% CI: 2.57, 3.72], that was reduced when restricted to term births (aOR = 1.59; 95% CI: 1.08, 2.36). The aOR for PTB was 2.88 (95% CI: 2.46, 3.37). Results were essentially unchanged when G1 birthweight and gestational age were included in the model, as well as after adjusting for other potential intermediate variables, such as G2 pregnancy-related factors.

CONCLUSIONS

Grandmother use of DES during pregnancy is associated with an increased risk of LBW, predominantly through an increased risk of PTB. Results when considering G1 birth outcomes suggest this does not result from confounding by indication.

KDR polymorphism (1192G/A, 1719A/T) and modulation of ARV drug-induced hepatotoxicity

Kinase insert Domain containing Receptor (KDR)/Vascular Endothelial Growth Factor Receptor (VEGFR-2) participate in endothelial dysfunction, which can lead to chronic liver disease. KDR reflects naturally against the toxicants from the damaged liver cells. Association of KDR polymorphism has been reported with many diseases including liver disease, but its role has not been described in ARV induced hepatotoxicity. Hence, we examined the exonic regions KDR (1192G/A, 1719A/T) polymorphism from 165 HIV-infected individuals (34/165 had ARV induced hepatotoxicity, 131/165 were with no hepatotoxicity) and 160 normal uninfected individuals by PCR-RFLP. In univariate analysis, KDR 1719 TT genotype presented at greater frequency from all HIV positive individuals in contrast with normal uninfected individuals (7.87% vs. 4.4%, OR = 1.72, P = 0.38). Individuals with KDR 1719 TT genotype had a risk for increasing hepatotoxicity and its severity (OR = 1.91, P = 0.38). Individuals with haplotype AT had risk for increasing hepatotoxicity and its severity (OR = 1.60, P = 0.50; OR = 2.35, P = 0.27). Whereas haplotype AA was associated with reduced risk of developing hepatotoxicity (OR = 0.40, P = 0.04). Individuals with KDR 1719 TT genotype were at greater risk of advancement of HIV disease (OR = 2.31, P = 0.23). Individuals with KDR 1719 TT genotype had more vulnerability for developing hepatotoxicity among alcohol users (OR = 2.57, P = 0.23). Individuals with KDR 1719 TT genotype were at higher risk of developing hepatotoxicity and its severity among nevirapine and alcohol consumers (OR = 2.47, P = 0.24; OR = 5.42, P = 0.42). In multivariate analysis, hepatotoxicity patients taking ART inclusive of nevirapine was associated with the severity of hepatotoxicity (OR = 4.82, P = 0.002). In conclusion, KDR 1719 TT genotype and haplotype AT may have a risk for development of hepatotoxicity and its severity. Haplotype AA may have influence to reduce the risk of developing hepatotoxicity.

Invited Commentary: Dealing With the Inevitable Deficiencies of Bias Analysis-and All Analyses

Lash et al. (Am J Epidemiol. 2021;190(8):1604-1612) have presented detailed critiques of 3 bias analyses that they identify as "suboptimal." This identification raises the question of what "optimal" means for bias analysis, because it is practically impossible to do statistically optimal analyses of typical population studies-with or without bias analysis. At best the analysis can only attempt to satisfy practice guidelines and account for available information both within and outside the study. One should not expect a full accounting for all sources of uncertainty; hence, interval estimates and distributions for causal effects should never be treated as valid uncertainty assessments-they are instead only example analyses that follow from collections of often questionable assumptions. These observations reinforce those of Lash et al. and point to the need for more development of methods for judging bias-parameter distributions and utilization of available information.

A systematic review of quantitative bias analysis applied to epidemiological research

BACKGROUND

Quantitative bias analysis (QBA) measures study errors in terms of direction, magnitude and uncertainty. This systematic review aimed to describe how QBA has been applied in epidemiological research in 2006-19.

METHODS

We searched PubMed for English peer-reviewed studies applying QBA to real-data applications. We also included studies citing selected sources or which were identified in a previous QBA review in pharmacoepidemiology. For each study, we extracted the rationale, methodology, bias-adjusted results and interpretation and assessed factors associated with reproducibility.

RESULTS

Of the 238 studies, the majority were embedded within papers whose main inferences were drawn from conventional approaches as secondary (sensitivity) analyses to quantity-specific biases (52%) or to assess the extent of bias required to shift the point estimate to the null (25%); 10% were standalone papers. The most common approach was probabilistic (57%). Misclassification was modelled in 57%, uncontrolled confounder(s) in 40% and selection bias in 17%. Most did not consider multiple biases or correlations between errors. When specified, bias parameters came from the literature (48%) more often than internal validation studies (29%). The majority (60%) of analyses resulted in >10% change from the conventional point estimate; however, most investigators (63%) did not alter their original interpretation. Degree of reproducibility related to inclusion of code, formulas, sensitivity analyses and supplementary materials, as well as the QBA rationale.

CONCLUSIONS

QBA applications were rare though increased over time. Future investigators should reference good practices and include details to promote transparency and to serve as a reference for other researchers.

An underappreciated misclassification mechanism: implications of nondifferential dependent misclassification of covariate and exposure

Misclassification is a pervasive problem in assessing relations between exposures and outcomes. While some attention has been paid to the impact of dependence in measurement error between exposures and outcomes, there is little awareness of the potential impact of dependent error between exposures and covariates, despite the fact that this latter dependency may occur much more frequently, for example, when both are assessed by questionnaire. We explored the impact of nondifferential dependent exposure-confounder misclassification bias by simulating a dichotomous exposure (E), disease (D) and covariate (C) with varying degrees of non-differential dependent misclassification between C and E. We demonstrate that under plausible scenarios, an adjusted association can be a poorer estimate of the true association than the crude. Correlated errors in the measurement of covariate and exposure distort the covariate-exposure, covariate-outcome and exposure-outcome associations creating observed associations that can be greater than, less than, or in the opposite direction of the true associations. Under these circumstances adjusted associations may not be bounded by the crude association and true effect, as would be expected with nondifferential independent confounder misclassification. The degree and direction of distortion depends on the amount of dependent error, prevalence of covariate and exposure, and magnitude of true effect.

The Southern European Atlantic Diet and all-cause mortality in older adults

BACKGROUND

The Southern European Atlantic Diet (SEAD) is the traditional diet of Northern Portugal and North-Western Spain. Higher adherence to the SEAD has been associated with lower levels of some cardiovascular risk factors and reduced risk for myocardial infarction, but whether this translates into lower all-cause mortality is uncertain. We hence examined the association between adherence to the SEAD and all-cause mortality in older adults.

METHODS

Data were taken from the Seniors-ENRICA-1 cohort, which included 3165 individuals representative of the non-institutionalized population aged ≥ 60 years in Spain. Food consumption was assessed with a validated diet history, and adherence to the SEAD was measured with an index comprising 9 food components: fresh fish, cod, red meat and pork products, dairy products, legumes and vegetables, vegetable soup, potatoes, whole-grain bread, and wine. Vital status was ascertained with the National Death Index of Spain. Statistical analyses were performed with Cox regression models and adjusted for the main confounders.

RESULTS

During a median follow-up of 10.9 years, 646 deaths occurred. Higher adherence to the SEAD was associated with lower all-cause mortality (fully adjusted hazard ratio [95% confidence interval] per 1-SD increment in the SEAD score 0.86 [0.79, 0.94]; p-trend < 0.001). Most food components of the SEAD showed some tendency to lower all-cause mortality, especially moderate wine consumption (hazard ratio [95% confidence interval] 0.71 [0.59, 0.86]). The results were robust in several sensitivity analyses. The protective association between SEAD and all-cause death was of similar magnitude to that found for the Mediterranean Diet Adherence Screener (hazard ratio [95% confidence interval] per 1-SD increment 0.89 [0.80, 0.98]) and the Alternate Healthy Eating Index (0.83 [0.76, 0.92]).

CONCLUSIONS

Adherence to the SEAD is associated with a lower risk of all-cause death among older adults in Spain.

Prenatal Methamphetamine Exposure and Adverse Neonatal Outcomes: A Nationwide Cohort Study

BACKGROUND

There is limited knowledge on the adverse outcomes in newborns after maternal methamphetamine (MA) use during pregnancy.

OBJECTIVES

To compare neonatal outcomes in newborns exposed to MA with the newborns of opioid-exposed mothers and of mothers from the general population (GP).

METHOD

A cohort study using nationwide registries in Czechia (2000-2014). Women hospitalized with a main diagnosis of MA use disorder during pregnancy (n = 258) and their newborns were defined as MA-exposed. The comparison groups consisted of women (n = 199) diagnosed with opioid use disorder during pregnancy, defined as opioid-exposed, and women (n = 1,511,310) with no substance use disorder diagnosis (GP). The neonatal outcomes studied were growth parameters, gestational age, preterm birth, and Apgar score. To explore the associations between MA exposure and neonatal outcomes, regression coefficients (b) and odds ratios from multivariable linear and binary logistic regression were estimated.

RESULTS

MA-exposed women had similar socio-economic characteristics to opioid-exposed, both of which were worse than in the GP. After adjustment, MA exposure was associated with a more favourable birthweight when compared to the opioid-exposed (adjusted mean differences [aMD] b = 122.3 g, 95% CI: 26.0-218.5) and length (aMD b = 0.6 cm, 0.0-1.1). Unadjusted results from the comparison with the GP showed that the MA group had poorer neonatal outcomes, especially in the growth parameters. Adjustment for background characteristics had a profound effect on the comparison with the GP. After adjustment, MA exposure was associated only with a slightly reduced birthweight (aMD b = -63.0 g, -123.0 to -3.1) and birth length (aMD b = -0.3 cm, -0.6 to 0.0).

CONCLUSIONS

Although the observed negative outcomes were large in the MA-exposed newborns, the adjustment had a profound effect on the comparison with the GP, indicating the large influence of lifestyle and socio-economic factors in these high-risk pregnancies. MA-exposed newborns had better neonatal outcomes compared to opioids-exposed.

Classification of drug use patterns

Characterizing long‐term prescription data is challenging due to the time‐varying nature of drug use. Conventional approaches summarize time‐varying data into categorical variables based on simple measures, such as cumulative dose, while ignoring patterns of use. The loss of information can lead to misclassification and biased estimates of the exposure‐outcome association. We introduce a classification method to characterize longitudinal prescription data with an unsupervised machine learning algorithm. We used administrative databases covering virtually all 1.3 million residents of Manitoba and explicitly designed features to describe the average dose, proportion of days covered (PDC), dose change, and dose variability, and clustered the resulting feature space using K‐means clustering. We applied this method to metformin use in diabetes patients. We identified 27,786 metformin users and showed that the feature distributions of their metformin use are stable for varying the lengths of follow‐up and that these distributions have clear interpretations. We found six distinct metformin user groups: patients with intermittent use, decreasing dose, increasing dose, high dose, and two medium dose groups (one with stable dose and one with highly variable use). Patients in the varying and decreasing dose groups had a higher chance of progression of diabetes than other patients. The method presented in this paper allows for characterization of drug use into distinct and clinically relevant groups in a way that cannot be obtained from merely classifying use by quantiles of overall use.

Accounting for Expected Adjusted Effect

The point that adjustment for confounders do not always guarantee protection against spurious findings and type 1-errors has been made before. The present simulation study indicates that for traditional regression methods, this risk is accentuated by a large sample size, low reliability in the measurement of the confounder, and high reliability in the measurement of the predictor and the outcome. However, this risk might be attenuated by calculating the expected adjusted effect, or the required reliability in the measurement of the possible confounder, with equations presented in the present paper.

An Introduction to Causal Diagrams for Anesthesiology Research

Making good decisions in the era of Big Data requires a sophisticated approach to causality. We are acutely aware that association ≠ causation, yet untangling the two remains one of our greatest challenges. This realization has stimulated a Causal Revolution in epidemiology, and the lessons learned are highly relevant to anesthesia research. This article introduces readers to directed acyclic graphs; a cornerstone of modern causal inference techniques. These diagrams provide a robust framework to address sources of bias and discover causal effects. We use the topical question of whether anesthetic technique (total intravenous anesthesia vs. volatile) affects outcome after cancer surgery as a basis for a series of example directed acyclic graphs, which demonstrate how variables can be chosen to statistically control confounding and other sources of bias. We also illustrate how controlling for the wrong variables can introduce, rather than eliminate, bias; and how directed acyclic graphs can help us diagnose this problem.

This is a rapidly evolving field, and we cover only the most basic elements. The true promise of these techniques is that it may become possible to make robust statements about causation from observational studies—without the expense and artificiality of randomized controlled trials.

Prevalence of ABCC3-1767G/A polymorphism among patients with antiretroviral-associated hepatotoxicity

Background

Plasma concentrations of antiretrovirals (ARVs) regimens have considerably varied in individuals of human immunodeficiency virus (HIV) because of variations in the expression of drug‐metabolizing and transporter genes. Transporter genes play an important role in the disposition of drugs. Polymorphism in transporter gene (ABCC3) affects the MRP3 expression and varies the treatment outcome.

Method

We examined the polymorphism of ABCC3‐1767G/A gene in a total of 165 HIV patients (out of 165 HIV patients, 34 were with and 131 were without hepatotoxicity) and 156 healthy individuals using the polymerase chain reaction–restriction fragment length polymorphism method.

Results

In univariate analysis, we found a decreased prevalence of ABCC3 1767GA, 1767GA+AA genotypes, and 1767A allele in patients with hepatotoxicity as compared to patients without hepatotoxicity (23.5% vs. 28.2% and 23.5% vs. 30.53%; 11.76% vs. 16.41%), while a higher prevalence of 1767AA genotype was observed in HIV patients in comparison with healthy controls (2.3% vs. 1.3%, odds ratio [OR] = 1.71, 95% confidence interval [CI]: 0.23–15.03, p = .89). The frequency of ABCC3‐1767AA genotype was dispersed higher in individuals with early and advanced HIV disease stage in comparison with healthy controls (5.3% vs. 1.3%, OR = 4.73, p = .70; 8.9% vs. 1.3%, OR = 1.89, p = .91). A higher occurrence of ABCC3‐1767AA genotype was found in tobacco using HIV patients without hepatotoxicity compared with nonusers (4.7% vs. 1.1%, OR = 4.28, p = .52). The distribution of ABCC3‐1767GA genotype was higher in nevirapine receiving HIV patients irrespective of their hepatotoxicity status as compared to nonusers (30.4% vs. 9.1%, OR = 3.34, p = .22; 29.4% vs. 16.7%, OR = 1.69, p = .77). In multivariate analysis, HIV patients receiving nevirapine and with hepatotoxicity was found to have a significant risk for severity of hepatotoxicity (OR = 4.56, 95% CI: 1.60–12.99, p = .004).

Conclusion

ABCC3 1767G/A polymorphism was not significantly associated with susceptibility to ARV‐associated hepatotoxicity, although ABCC3 1767AA genotype designated a risk for acquisition of hepatotoxicity and advancement of the disease. Nevirapine usage emerged as an independent risk factor for hepatotoxicity severity.

Reflection on modern methods: five myths about measurement error in epidemiological research

Epidemiologists are often confronted with datasets to analyse which contain measurement error due to, for instance, mistaken data entries, inaccurate recordings and measurement instrument or procedural errors. If the effect of measurement error is misjudged, the data analyses are hampered and the validity of the study's inferences may be affected. In this paper, we describe five myths that contribute to misjudgments about measurement error, regarding expected structure, impact and solutions to mitigate the problems resulting from mismeasurements. The aim is to clarify these measurement error misconceptions. We show that the influence of measurement error in an epidemiological data analysis can play out in ways that go beyond simple heuristics, such as heuristics about whether or not to expect attenuation of the effect estimates. Whereas we encourage epidemiologists to deliberate about the structure and potential impact of measurement error in their analyses, we also recommend exercising restraint when making claims about the magnitude or even direction of effect of measurement error if not accompanied by statistical measurement error corrections or quantitative bias analysis. Suggestions for alleviating the problems or investigating the structure and magnitude of measurement error are given.

Improving measurement of binary covariates in claims data: A simulation study

PURPOSE

When investigators have two claims-based definitions for a binary confounder, it is unclear whether to prefer the more sensitive or more specific definition. Our objective was to compare adjusting for the sensitive or specific definition alone vs two novel approaches combining both definitions: a "two-algorithm indicator" and a "two-algorithm restriction" approach.

METHODS

Each simulated patient had a binary exposure, outcome, and confounder. We created two nested, misclassified versions of the confounder using validated heart failure definitions. The sensitive definition had a sensitivity/specificity of 0.98/0.83, while the specific definition had a sensitivity/specificity of 0.77/0.99. Patients were classified into 3 groups: group 0 did not meet either definition, group 1 met the sensitive but not specific definition, and group 2 met both. The two-algorithm indicator approach adjusted using indicators for groups 1 and 2, while the two-algorithm restriction approach excluded patients in group 1 and adjusted using an indicator for group 2. Adjusted exposure odds ratios (ORs) were estimated for each approach using logistic regression.

RESULTS

The crude OR was 1.33 (95% CI, 1.07-1.63). Adjusting for the specific or sensitive definitions resulted in ORs of 1.09 (95% CI, 0.87-1.35) and 1.14 (95% CI, 0.91-1.40). The two-algorithm indicator method returned an OR of 1.07 (95% CI, 0.86-1.33). The two-algorithm restriction approach returned an OR of 1.02 (95% CI, 0.79-1.29) but excluded 20% of the cohort.

CONCLUSIONS

The two-algorithm indicator approach may improve adjustment for claims-based confounders by returning a point estimate at least as unbiased as the better of the two component definitions.

β2-adrenoceptor agonists and antagonists and risk of Parkinson's disease

BACKGROUND

β2-adrenoreceptors have recently been identified as regulators of the α-synuclein gene, which is implicated in the pathogenesis of Parkinson's disease.

OBJECTIVE

The objectives of this study were to assess the association between use of β2-agonists and β-antagonists and the risk of developing PD.

METHODS

We conducted a nested case-control study in a cohort of 1,762,164 adults without a diagnosis of PD. They were identified on January, 1, 2004, from the electronic medical records of the largest health care provider in Israel. Participants were followed up until June 30, 2017, for the occurrence of PD. Ten randomly selected controls were matched to each case of PD on age, sex, ethnic group, and duration of follow-up.

RESULTS

During follow-up 11,314 patients were newly diagnosed with PD and were matched with 113,140 controls. An increased risk of PD was seen with the use of nonselective β-antagonists (RR, 2.04 [1.90-2.20]) but not with the use of selective β1-antagonists (RR, 1.00 [0.95-1.05]). Use of β2-agonists was associated with reduced risk of PD (RR, 0.89 [0.82-0.96] for short-acting; RR, 0.84 [0.76-0.93] for long-acting; and RR, 0.49 [0.25-0.92] for ultra-long-acting β2-agonists). In an analysis of individual drugs, propranolol and salbutamol were significantly associated with PD risk, even when these drugs were ascertained 5 years prior to the index date, compared with nonusers (RR, 1.31 [1.08-1.58] and 1.89 {1.53-2.33]) in patients who filled <6 and ≥6 propranolol prescriptions, respectively; the corresponding RRs for salbutamol were 0.95 (0.83-1.08) and 0.65 (0.45-0.94), respectively.

CONCLUSIONS

Use of propranolol appears to be associated with an increased risk of PD, whereas use of β2-agonists is associated with a decreased risk of PD. © 2018 International Parkinson and Movement Disorder Society.

Tissue Inhibitor of Metalloproteinase-2 Polymorphisms and Risk for HIV-Associated Neurocognitive Disorder

The imbalance between MMPs and TIMPs is associated with the HIV dissemination tissue damage pathology neurodegenerative disorders, including HAND. Genetic variations in the TIMP gene may modulate the neurocognitive disorder in HIV patients. Hence, we evaluated the genetic variants of TIMP-2 (-418G/C, 303G/A) gene with the risk of HAND. Genotyping of TIMP-2 polymorphism was performed in 50 patients with HAND, 100 no HAND, and 154 healthy controls by PCR-RFLP. TIMP-2 -418GC and 303AA genotypes represented a predominant risk for HAND severity (OR = 1.55, P = 0.30; OR = 4.58, P = 0.24). The variant -418CC genotype, -418A allele, had exhibited a significant risk for the acquisition of HAND (OR = 12.55, P = 0.026; OR = 2.66, P = 0.004). TIMP-2 303GA, 303AA genotype, and 303A allele evinced a higher risk for HAND severity (OR = 1.82, P = 0.14; OR = 1.70, P = 0.63; and OR = 1.68, P = 0.12). In HIV patients, TIMP-2 -418CC genotype and -418C allele significantly occurred in comparison to healthy controls (OR = 10.10, P = 0.006; OR = 2.02, P = 0.009). In the intermediate and early HIV disease stage, TIMP-2 -418CC genotype was significantly increased compared with healthy controls (11.1% vs. 1.3%, OR = 14.63, P = 0.01; 16.9% vs. 1.3%, OR = 14.51, P = 0.002). In patients with HAND among tobacco and alcohol users, TIMP-2 -418CC genotype displayed a risk for HAND severity (OR = 3.96, P = 0.26; OR = 4.83, P = 0.19). On multivariate logistic regression, TIMP-2 303AA genotype, advanced stage, and gender had a risk for HAND severity (OR = 28.98, P = 0.02; OR = 2.35, P = 0.070; and OR = 2.36, P = 0.04). In conclusion, TIMP-2 -418G/C polymorphism independently, along with alcohol and tobacco, may have an impact on the acquisition of HAND and its severity. TIMP-2 303G/A polymorphism bare a risk for HAND severity.

Interleukin-10 polymorphisms and susceptibility to ARV associated hepatotoxicity

Interleukin-10 (IL-10) is an anti-inflammatory cytokine associated with the inhibition of HIV replication. IL-10 polymorphisms were found to be linked to drug-induced hepatotoxicity. Hence we examined the prevalence of IL-10 (-819C/T,-1082A/G) polymorphisms in a total of 165 HIV patients which included 34 patients with hepatotoxicity, 131 without hepatotoxicity and 155 healthy controls by the PCR-RFLP method. In HIV patients with hepatotoxicity, the IL-10-819TT genotype increased the risk of ARV associated hepatotoxicity severity (OR = 1.61, P = 0.35). IL-10-819TT genotype was overrepresented in patients with hepatotoxicity as compared to healthy controls (26.5% vs. 13.5%, OR = 1.61, P = 0.46). IL-10 -819CT genotype was associated with advance HIV disease stage (OR = 0.49, P = 0.045). In HIV patients without hepatotoxicity, the IL-10-819TT genotype was more prevalent in patients consuming tobacco as compared to non-users (OR = 1.60, P = 0.41). In HIV patients without hepatotoxicity using both alcohol + efavirenz along with IL-10 -819CT genotype resulted in increased risk for the acquisition of ARV associated hepatotoxicity (OR = 4.00, P = 0.36). In multivariate logistic regression, taking nevirapine was associated with the risk hepatotoxicity severity (OR = 0.23, P = 0.005). In conclusion, an insignificant association between IL-10 polymorphisms and susceptibility to ARV associated hepatotoxicity.

Occurrence of CYP2B6 516G>T polymorphism in patients with ARV-associated hepatotoxicity

Background

Hepatic enzyme cytochrome P450 2B6 (CYP2B6) plays a role in the metabolism of efavirenz drugs. CYP2B6 516G>T variation showed an implication for HIV treatment.

Methods

CYP2B6 516G>T polymorphism was genotyped in a total 165 HIV patients that include 34 with and 131 without hepatotoxicity and 155 healthy individuals by the PCR‐RFLP.

Results

In patients with hepatotoxicity, the prevalence of CYP2B6 516TT genotype was higher as compared to healthy individuals (35.3% vs. 30.5%, OR = 1.74). Patients with hepatotoxicity using tobacco had a higher prevalence of genotypes CYP2B6 516GT, 516TT, 516GT+TT as compared to healthy individuals (28.57% vs. 25.93%; 57.14% vs. 29.63%; 85.71% vs. 55.56%). Likewise, hepatotoxicity in patients consuming alcohol showed higher distributions of CYP2B6 516GT, 516TT, 516GT+TT genotypes (57% vs. 25.93%; 42.86% vs. 33.33%; 71.43% vs. 59.26%). Nevirapine users with hepatotoxicity overrepresented genotypes CYP2B6 TT and 516GT+TT as compared to efavirenz users (47.83% vs. 45.45%, OR = 6.88, 65.22% vs. 54.55%, OR = 1.56). Similarly, in nevirapine +alcohol users with hepatotoxicity, the frequency of CYP2B6 516GT, 516GT+TT genotypes was higher than with nevirapine +alcohol nonusers (40.0% vs. 11.11%, OR = 8.00, 80.0% vs. 27.78%, OR = 4.00). In HIV patients, nevirapine users had higher frequency of CYP2B6 516GT, 516GT+TT genotypes as compared to efavirenz users (42.02% vs. 25.00%, OR = 2.53; 72.27% vs. 58.33%, OR = 1.86). Likewise, in HIV patients, genotypes CYP2B6 516GT, 516GT+TT were predominant with nevirapine +alcohol users as compared to nevirapine +alcohol nonusers (57.89% vs. 34.57%, OR = 2.46; 78.95% vs. 69.14%, OR = 1.67). In multivariate logistic regression, taking nevirapine had a protection for severity of ARV‐associated hepatotoxicity (OR = 0.23, p = 0.005).

Conclusions

No significant association was detected between CYP2B6 516G>T polymorphism and susceptibility to ARV‐associated hepatotoxicity.

Investigating gateway effects using the PATH study

BACKGROUND

A recent meta-analysis of nine cohort studies in youths reported that baseline ever e-cigarette use strongly predicted cigarette smoking initiation in the next 6-18 months, with an adjusted odds ratio of 3.62 (95% confidence interval 2.42-5.41).  A recent review of e-cigarettes agreed there was substantial evidence for this "gateway effect".  However, the number of confounders considered in the studies was limited, so we investigated whether the effect might have resulted from inadequate adjustment, using Waves 1 and 2 of the Population Assessment of Tobacco and Health study.

METHODS

Our main analyses considered Wave 1 never cigarette smokers who, at Wave 2, had information available on smoking initiation.  We constructed a propensity score for ever e-cigarette use from Wave 1 variables, using this to predict ever cigarette smoking.  Sensitivity analyses accounted for use of other tobacco products, linked current e-cigarette use to subsequent current smoking, or used propensity scores for ever smoking or ever tobacco product use as predictors.  We also considered predictors using data from both waves to attempt to control for residual confounding from misclassified responses.

RESULTS

Adjustment for propensity dramatically reduced the unadjusted odds ratio (OR) of 5.70 (4.33-7.50) to 2.48 (1.85-3.31), 2.47 (1.79-3.42) or 1.85 (1.35-2.53), whether adjustment was made as quintiles, as a continuous variable or for the individual variables.  Additional adjustment for other tobacco products reduced this last OR to 1.59 (1.14-2.20).  Sensitivity analyses confirmed adjustment removed most of the gateway effect.  Control for residual confounding also reduced the association.

CONCLUSIONS

We found that confounding is a major factor, explaining most of the observed gateway effect.  However, our analyses are limited by small numbers of new smokers considered and the possibility of over-adjustment if taking up e-cigarettes affects some predictor variables.  Further analyses are intended using Wave 3 data which should avoid these problems.

Investigating gateway effects using the PATH study

BACKGROUND

A recent meta-analysis of nine cohort studies in youths reported that baseline ever e-cigarette use strongly predicted cigarette smoking initiation in the next 6-18 months, with an adjusted odds ratio of 3.62 (95% confidence interval 2.42-5.41).  A recent review of e-cigarettes agreed there was substantial evidence for this "gateway effect".  However, the number of confounders considered in the studies was limited, so we investigated whether the effect might have resulted from inadequate adjustment, using Waves 1 and 2 of the Population Assessment of Tobacco and Health study.

METHODS

Our main analyses considered Wave 1 never cigarette smokers who, at Wave 2, had information available on smoking initiation.  We constructed a propensity score for ever e-cigarette use from Wave 1 variables, using this to predict ever cigarette smoking.  Sensitivity analyses accounted for use of other tobacco products, linked current e-cigarette use to subsequent current smoking, or used propensity scores for ever smoking or ever tobacco product use as predictors.  We also considered predictors using data from both waves to attempt to control for residual confounding from misclassified responses.

RESULTS

Adjustment for propensity dramatically reduced the unadjusted odds ratio (OR) of 5.70 (4.33-7.50) to 2.48 (1.85-3.31), 2.47 (1.79-3.42) or 1.85 (1.35-2.53), whether adjustment was made as quintiles, as a continuous variable or for the individual variables.  Additional adjustment for other tobacco products reduced this last OR to 1.59 (1.14-2.20).  Sensitivity analyses confirmed adjustment removed most of the gateway effect.  Control for residual confounding also reduced the association.

CONCLUSIONS

We found that confounding is a major factor, explaining most of the observed gateway effect.  However, our analyses are limited by small numbers of new smokers considered and the possibility of over-adjustment if taking up e-cigarettes affects some predictor variables.  Further analyses are intended using Wave 3 data which should avoid these problems.

Principles of confounder selection

Selecting an appropriate set of confounders for which to control is critical for reliable causal inference. Recent theoretical and methodological developments have helped clarify a number of principles of confounder selection. When complete knowledge of a causal diagram relating all covariates to each other is available, graphical rules can be used to make decisions about covariate control. Unfortunately, such complete knowledge is often unavailable. This paper puts forward a practical approach to confounder selection decisions when the somewhat less stringent assumption is made that knowledge is available for each covariate whether it is a cause of the exposure, and whether it is a cause of the outcome. Based on recent theoretically justified developments in the causal inference literature, the following proposal is made for covariate control decisions: control for each covariate that is a cause of the exposure, or of the outcome, or of both; exclude from this set any variable known to be an instrumental variable; and include as a covariate any proxy for an unmeasured variable that is a common cause of both the exposure and the outcome. Various principles of confounder selection are then further related to statistical covariate selection methods.

Negative control exposure studies in the presence of measurement error: implications for attempted effect estimate calibration

Background

Negative control exposure studies are increasingly being used in epidemiological studies to strengthen causal inference regarding an exposure-outcome association when unobserved confounding is thought to be present. Negative control exposure studies contrast the magnitude of association of the negative control, which has no causal effect on the outcome but is associated with the unmeasured confounders in the same way as the exposure, with the magnitude of the association of the exposure with the outcome. A markedly larger effect of the exposure on the outcome than the negative control on the outcome strengthens inference that the exposure has a causal effect on the outcome.

Methods

We investigate the effect of measurement error in the exposure and negative control variables on the results obtained from a negative control exposure study. We do this in models with continuous and binary exposure and negative control variables using analysis of the bias of the estimated coefficients and Monte Carlo simulations.

Results

Our results show that measurement error in either the exposure or negative control variables can bias the estimated results from the negative control exposure study.

Conclusions

Measurement error is common in the variables used in epidemiological studies; these results show that negative control exposure studies cannot be used to precisely determine the size of the effect of the exposure variable, or adequately adjust for unobserved confounding; however, they can be used as part of a body of evidence to aid inference as to whether a causal effect of the exposure on the outcome is present.