Collider Bias Is Only a Partial Explanation for the Obesity Paradox.

BACKGROUND: "Obesity paradox" refers to an association between obesity and reduced mortality (contrary to an expected increased mortality). A common explanation is collider stratification bias: unmeasured confounding induced by selection bias. Here, we test this supposition through a realistic generative model. METHODS: We quantify the collider stratification bias in a selected population using counterfactual causal analysis. We illustrate the bias for a range of scenarios, describing associations between exposure (obesity), outcome (mortality), mediator (in this example, diabetes) and an unmeasured confounder. RESULTS: Collider stratification leads to biased estimation of the causal effect of exposure on outcome. However, the bias is small relative to the causal relationships between the variables. CONCLUSIONS: Collider bias can be a partial explanation of the obesity paradox, but unlikely to be the main explanation for a reverse direction of an association to a true causal relationship. Alternative explanations of the obesity paradox should be explored. See Video Abstract at http://links.lww.com/EDE/B51.

The Obesity Paradox in Cancer: a Review.

There is a common perception that excess adiposity, commonly approximated by body mass index (BMI), is associated with reduced cancer survival. A number of studies have emerged challenging this by demonstrating that overweight and early obese states are associated with improved survival. This finding is termed the "obesity paradox" and is well recognized in the cardio-metabolic literature but less so in oncology. Here, we summarize the epidemiological findings related to the obesity paradox in cancer. Our review highlights that many observations of the obesity paradox in cancer reflect methodological mechanisms including the crudeness of BMI as an obesity measure, confounding, detection bias, reverse causality, and a specific form of the selection bias, known as collider bias. It is imperative for the oncologist to interpret the observation of the obesity paradox against the above methodological framework and avoid the misinterpretation that being obese might be "good" or "protective" for cancer patients.

Pioglitazone and bladder cancer risk: a multipopulation pooled, cumulative exposure analysis.

AIMS/HYPOTHESIS: The evidence on the association between pioglitazone use and bladder cancer is contradictory, with many studies subject to allocation bias. The aim of our study was to examine the effect of exposure to pioglitazone on bladder cancer risk internationally across several cohorts. The potential for allocation bias was minimised by focusing on the cumulative effect of pioglitazone as the primary endpoint using a time-dependent approach. METHODS: Prescription, cancer and mortality data from people with type 2 diabetes were obtained from six populations across the world (British Columbia, Finland, Manchester, Rotterdam, Scotland and the UK Clinical Practice Research Datalink). A discrete time failure analysis using Poisson regression was applied separately to data from each centre to model the effect of cumulative drug exposure on bladder cancer incidence, with time-dependent adjustment for ever use of pioglitazone. These were then pooled using fixed and random effects meta-regression. RESULTS: Data were collated on 1.01 million persons over 5.9 million person-years. There were 3,248 cases of incident bladder cancer, with 117 exposed cases and a median follow-up duration of 4.0 to 7.4 years. Overall, there was no evidence for any association between cumulative exposure to pioglitazone and bladder cancer in men (rate ratio [RR] per 100 days of cumulative exposure, 1.01; 95% CI 0.97, 1.06) or women (RR 1.04; 95% CI 0.97, 1.11) after adjustment for age, calendar year, diabetes duration, smoking and any ever use of pioglitazone. No association was observed between rosiglitazone and bladder cancer in men (RR 1.01; 95% CI 0.98, 1.03) or women (RR 1.00; 95% CI 0.94, 1.07). CONCLUSIONS/INTERPRETATION: The cumulative use of pioglitazone or rosiglitazone was not associated with the incidence of bladder cancer in this large, pooled multipopulation analysis.

Children's behavioural and emotional wellbeing during the COVID-19 pandemic: Findings from the Born in Bradford COVID-19 mixed methods longitudinal study.

Background: The COVID-19 pandemic led to a multitude of immediate social restrictions for many across the world. In the UK, the lives of children and young people were quickly impacted when COVID-19 restrictions led to school closures for most children and restrictions on social interactions. The Born in Bradford COVID-19 longitudinal research study explored the impact of the COVID-19 pandemic on the lives of children and their families living in Bradford. Methods: Surveys were administered during the first wave of the pandemic (March to June 2020) and compared to findings from before the pandemic. The current study examined the social and emotional wellbeing of children from before to during the pandemic, measured using the parent completed Strengths and Difficulties questionnaire (SDQ). Regression analyses looked at associations between a range of social determinants of health and changes in SDQ scores. Results: The results showed that those children most likely to experience difficulties during the pandemic were boys, younger children, those from White British ethnicity (compared to Pakistani heritage children) and those living in the most deprived areas. There were associations between experiencing difficulties and: food insecurity; financial worry; getting below recommended levels of physical activity; and having less than the recommended amount of sleep. Conclusions: The effect of COVID-19 restrictions are likely to have had negative consequences on children that could, in time, have long-lasting impacts on the health, wellbeing and development of children in the UK.

The COVID-19 pandemic caused immediate and long-lasting social restrictions to be implemented here in the UK and across the world. In the UK, children and young people were quickly affected by these restrictions that led to school closures and other restrictions that prevented these individuals from socialising in person with one another. This study explored the impact that the pandemic had on the wellbeing of children by comparing data from before the pandemic with data collected during the pandemic. The data that has been collected looks at the behavioural strengths and difficulties that children are displaying. Our exploration found that children that were most likely to experience difficulties during the pandemic were boys, younger children, those who were White British and those who lived in the most deprived areas. The effect of the COVID-19 restrictions are likely to have had a negative impact on children and young people which in time may impact the health and development of children living here in the UK.

Weight Changes in Type 2 Diabetes and Cancer Risk: A Latent Class Trajectory Model Study.

INTRODUCTION: Body mass index (BMI) is often elevated at type 2 diabetes (T2D) diagnosis. Using latent class trajectory modelling (LCTM) of BMI, we examined whether weight loss after diagnosis influenced cancer incidence and all-cause mortality. METHODS: From 1995 to 2010, we identified 7,708 patients with T2D from the Salford Integrated Record database (UK) and linked to the cancer registry for information on obesity-related cancer (ORC), non-ORC; and all-cause mortality. Repeated BMIs were used to construct sex-specific latent class trajectories. Hazard ratios (HRs) and 95% confidence intervals (CIs) were estimated using Cox regression models. RESULTS: Four sex-specific BMI classes were identified; stable-overweight, stable-obese, obese-slightly-decreasing, and obese-steeply-decreasing; comprising 41%, 45%, 13%, and 1% of women, and 45%, 37%, 17%, and 1% of men, respectively. In women, the stable-obese class had similar ORC risks as the obese-slightly-decreasing class, whereas the stable-overweight class had lower risks. In men, the obese-slightly-decreasing class had higher risks of ORC (HR = 1.86, 95% CI: 1.05-3.32) than the stable-obese class, while the stable-overweight class had similar risks No associations were observed for non-ORC. Compared to the stable-obese class, women (HR = 1.60, 95% CI: 0.99-2.58) and men (HR = 2.37, 95% CI: 1.66-3.39) in the obese-slightly-decreasing class had elevated mortality. No associations were observed for the stable-overweight classes. CONCLUSION: Patients who lost weight after T2D diagnosis had higher risks for ORC (in men) and higher all-cause mortality (both genders) than patients with stable obesity.

The Impact of the Pandemic on Mental Health in Ethnically Diverse Mothers: Findings from the Born in Bradford, Tower Hamlets and Newham COVID-19 Research Programmes.

Restrictions implemented by the UK Government during the COVID-19 pandemic have served to worsen mental health outcomes, particularly amongst younger adults, women, those living with chronic health conditions, and parents of young children. Studies looking at the impact for ethnic minorities have reported inconsistent findings. This paper describes the mental health experiences of mothers from a large and highly ethnically diverse population during the pandemic, using secondary analysis of existing data from three COVID-19 research studies completed in Bradford and London (Tower Hamlets and Newham). A total of 2807 mothers participated in this study with 44% White British, 23% Asian/Asian British Pakistani, 8% Other White and 7% Asian/Asian British Bangladeshi backgrounds. We found that 28% of mothers experienced clinically important depressive symptoms and 21% anxiety symptoms during the pandemic. In unadjusted analyses, mothers from White Other, and Asian/Asian British Bangladeshi backgrounds had higher odds of experiencing symptoms, whilst mothers from Asian/Asian British Indian backgrounds were the least likely to experience symptoms. Once loneliness, social support and financial insecurity were controlled for, there were no statistically significant differences in depression and anxiety by ethnicity. Mental health problems experienced during the pandemic may have longer term consequences for public health. Policy and decision makers must have an understanding of the high risk of financial insecurity, loneliness and a lack of social support on mother's mental health, and also recognise that some ethnic groups are far more likely to experience these issues and are, therefore, more vulnerable to poor mental health as a consequence.

Risk perception and disease knowledge in attendees of a community-based lung cancer screening programme.

OBJECTIVES: In England, a risk-based approach is used to determine eligibility for lung cancer screening. Ensuring effective communication and counselling of risk is therefore increasingly important. In this study, we explore the perception of lung cancer risk in attendees of a community-based screening service, located in socio-economically deprived areas of Manchester. We analyse responses based on demographic variables, calculated risk score and screening eligibility. MATERIALS AND METHODS: The Manchester Lung Health Check (LHC) programme invited ever smokers, age 55-80, to a lung cancer risk assessment in which their 6-year risk was calculated (using the PLCOM2012 model). Those at high risk (PLCOM2012 score ≥ 1.51%) were eligible for low dose CT (LDCT) screening. Prior to their assessment, attendees were invited to complete the study questionnaire, which assessed absolute and comparative risk perception, disease knowledge (incidence, survival, and risk factors), lung cancer specific worry, and mental health. RESULTS: 371 participants completed the questionnaire; 66% (n = 243) had linked clinical data. Perceived absolute risk was markedly higher than calculated risk (median: 20% vs. 1%; p < 0.001) and higher in women than men (25% vs. 15%; p = 0.001). There was no correlation between perceived absolute and calculated risk. Overall, 30% classified themselves at higher, and 21% at lower, lung cancer risk compared to others their age. Median PLCOM2012 score increased with perceived comparative risk (p = 0.004). Those eligible for screening were more likely to: classify themselves at higher comparative risk (41% vs. 21%; p < 0.0001), report lung cancer-specific worry (27% vs. 10%; p = 0.001) and have indications of depression (20% vs. 10%; p = 0.05). Family history of lung cancer was significantly associated with higher comparative risk (adjOR 4.03, 95%CI 1.74-9.3; p = 0.001). CONCLUSION: Employing comparative rather than absolute risk may assist risk counselling. Further research is required to determine the optimal approach to risk communication in this setting.

The relationship of ethnicity to the prevalence and management of hypertension and associated chronic kidney disease.

BACKGROUND: The effect of ethnicity on the prevalence and management of hypertension and associated chronic kidney (CKD) disease in the UK is unknown. METHODS: We performed a cross sectional study of 49,203 adults with hypertension to establish the prevalence and management of hypertension and associated CKD by ethnicity. Routinely collected data from general practice hypertension registers in 148 practices in London between 1/1/07 and 31/3/08 were analysed. RESULTS: The crude prevalence of hypertension was 9.5%, and by ethnicity was 8.2% for White, 11.3% for South Asian and 11.1% for Black groups. The prevalence of CKD stages 3-5 among those with hypertension was 22%. Stage 3 CKD was less prevalent in South Asian groups (OR 0.77, 95% CI 0.67 - 0.88) compared to Whites (reference population) with Black groups having similar rates to Whites. The prevalence of severe CKD (stages 4-5) was higher in the South Asian group (OR 1.53, 95% CI 1.17 - 2.0) compared to Whites, but did not differ between Black and White groups. In the whole hypertension cohort, achievement of target blood pressure (< 140/90 mmHg) was better in South Asian (OR 1.43, 95% CI 1.28 - 1.60) and worse in Black groups (OR 0.79, 95% CI 0.74 - 0.84) compared to White patients. Hypertensive medication was prescribed unequally among ethnic groups for any degree of blood pressure control. CONCLUSIONS: Significant variations exist in the prevalence of hypertension and associated CKD and its management between the major ethnic groups. Among those with CKD less than 50% were treated to a target BP of ≤ 130/80 mmHg. Rates of ACE-I/ARB prescribing for those with CKD were less than optimal, with the lowest rates (58.5%) among Black groups.

Prescribing in general practice for people with coronary heart disease; equity by age, sex, ethnic group and deprivation.

OBJECTIVE: Differences in drug prescribing for coronary heart disease have previously been identified by age, sex and ethnic group. Set in the UK, our study utilises routinely collected data from 98 general practices serving a socially diverse population in inner East London, to examine differences in prescribing rates among patients aged 35 years and over with coronary heart disease. DESIGN: 10,933 patients aged 35 years or more, with recorded coronary heart disease, from 98 practices in two Primary Care Trusts (PCT) in East London during 2009/2010 were included for this cross-sectional study. Multivariable logistic regression was used to assess the odds of prescribing for recommended coronary heart disease drugs by age, sex, ethnicity, social deprivation, co-morbidity and recorded reasons for not prescribing. RESULTS: Women are prescribed fewer recommended coronary heart disease drugs than men; Black African/Caribbean patients are prescribed fewer lipid modifying drugs and other cardiovascular drugs than White patients. Patients over age 84 are prescribed fewer lipid modifying drugs and beta blockers than patients aged 45-54. South Asian patients had the highest levels of prescribing and higher prevalence of coronary heart disease and diabetes co-morbidity. No difference in prescribing rates by social deprivation was found. DISCUSSION: Overall levels of prescribing are high but small differences between sex and ethnic groups remain and prescribing may be inequitable for women, for Black/African Caribbeans and at older ages. These differences were not explained by recorded intolerance, contraindications or declining treatment.

Measures of social position and cortisol secretion in an aging population: findings from the Whitehall II study.

OBJECTIVE: To examine whether dysregulation of the hypothalamic-pituitary-adrenal (HPA) axis associated with disadvantaged social position in working populations also occurs in older age groups. METHODS: This study examines the association of several indicators of social position with two measures of cortisol secretion, a product of the HPA axis. We examined the cortisol awakening response (CAR), and slope of the decline in cortisol secretion across the day. We examine whether the association is mediated by behavioral, psychosocial, and biological factors in 3992 participants of phase 7 (2002-2004) of the Whitehall II study, who provided six salivary cortisol samples across the day. RESULTS: In this older cohort (mean age = 61 years; range = 50-74 years), lowest social position (assessed by current or previous occupational grade and wealth) was associated with a flatter slope in the decline in cortisol secretion. For example, over the course of the day, men in the lowest employment grades had a reduction in their cortisol by 0.125 (nmol/L/h), which was a shallower slope than those in the high grades (-0.129 nmol/L/h). The difference in slopes by employment grade among men, but not women, was statistically significant (p = .003). The difference in slopes was explained primarily by poor health and sleep behaviors, although financial insecurity also played a role. No effects were apparent with the CAR or other measures of social position. CONCLUSIONS: In men, poorer health and sleep behaviors (primarily smoking and short sleep duration), and financial insecurity mediate the impact of occupational status and wealth on cortisol secretion.

The relationship between alcohol consumption and cortisol secretion in an aging cohort.

CONTEXT: Evidence for an association between alcohol consumption and activity of the hypothalamic-pituitary-adrenal (HPA) axis is inconclusive. OBJECTIVE: Our objective was to assess the relationship between indices of alcohol consumption and salivary cortisol concentration. DESIGN: This was a cross-sectional study of alcohol consumption and cortisol secretion from phase 7 (2002-2004) of the Whitehall II study. SETTING: An occupational cohort originally recruited in 1985-1987 was included in the study. PARTICIPANTS: A total of 2693 men and 977 women had information on cortisol levels and alcohol consumption. OUTCOME MEASURES: Saliva samples were taken on waking, waking+0.5, 2.5, 8, and 12 h, and bedtime for the assessment of cortisol. RESULTS: In men there was a positive association between cortisol and units of alcohol intake per week (3% increase in cortisol per unit of alcohol consumed; P=0.010). The slope of cortisol decline over the day in heavy drinkers was reduced (heavy drinkers beta=-0.155, moderate drinkers beta=-0.151), indicating reduced control of the HPA axis in heavy drinkers. In women the cortisol awakening response was greater in heavy drinkers 14.15 nmol/liter (9.12-19.17) compared with moderate drinkers 8.69 nmol/liter (7.72-9.67) (P=0.037). CONCLUSIONS: This study suggests that alcohol consumption is associated with activation of the HPA axis. These results are not due to alcohol consumption on the day, suggesting chronic changes of the HPA axis in heavy drinking groups.

Neuroendocrine and inflammatory factors associated with positive affect in healthy men and women: the Whitehall II study.

Positive affective states are associated with favorable health outcomes, but the underlying mechanisms are poorly understood. The authors assessed associations between positive affect, cortisol sampled over the day, and inflammatory markers (C-reactive protein and interleukin-6) among 2,873 healthy members of the Whitehall II study. Data for this study were collected in 2002-2004 in London, United Kingdom. Saliva free cortisol was assessed on waking, 30 minutes later, and four times over the day and evening. Positive affect was indexed by aggregating ecological momentary assessments of positive mood over the day. Salivary cortisol averaged over the day was inversely associated with positive affect after controlling for age, gender, income, ethnicity, body mass index, waist/hip ratio, smoking, paid employment, time of waking in the morning, and depression (p = 0.003). There was no association with cortisol responses to waking. The adjusted odds of C-reactive protein >/=3.00 mg/liter was 1.89 (95% confidence interval: 1.08, 3.31) in low- compared with high-positive-affect women, and plasma interleukin-6 was also inversely related to positive affect in women (p = 0.016). Neither inflammatory marker was related to positive affect in men. These results confirm findings from smaller studies relating cortisol with positive affect while suggesting that in women, positive affect is associated with reduced levels of inflammatory markers.

Psychological coping styles and cortisol over the day in healthy older adults.

Patterns of psychological coping are associated with a variety of health outcomes but the underlying pathways are not yet established. The purpose of this study was to assess the relationship between salivary cortisol output over the course of a day and coping style. Data were available from 350 men and 192 women with an average age of 60.9 years. Participants were drawn from the Whitehall II cohort, and had no history of cardiovascular disease. Individuals who were taking medication that might affect cortisol levels were also excluded. Saliva samples were provided on waking, then 0.5, 2.5, 8 and 12h after waking, and just before the participant went to sleep. Coping style was measured with a standard instrument, the COPE, and data were factor analysed to generate three factors: seeking social support, problem engagement and problem avoidance. The relationships between these factors and the cortisol awakening response (CAR), the slope of cortisol change over the day and total cortisol output over the day (excluding the waking period) were assessed using multiple linear regression. Cortisol output over the day was inversely associated with coping with stress by seeking social support (p=0.034) and by problem engagement (p=0.003), independently of age, gender, body mass index, smoking, depression, self-rated health, time of waking and income. Individuals who coped by problem engagement and seeking support had lower cortisol levels. Additionally, gender, BMI, smoking, self-rated health and time of waking were independently related to cortisol output over the day. There were no significant associations between coping and the CAR or cortisol slope over the day. The results indicate that adaptive coping styles are related to low levels of cortisol over the day, suggesting that neuroendocrine pathways may partly mediate relationships between psychological coping and health.

The relationship between smoking status and cortisol secretion.

CONTEXT: Evidence for an association of smoking status with cortisol secretion is mixed. OBJECTIVE: The objective of the study was to assess the relationship between smoking status and salivary cortisol. DESIGN: This was a cross-sectional study of smoking status and cortisol secretion from phase 7 (2002-2004) of the Whitehall II study. SETTING: An occupational cohort was originally recruited in 1985-1987. PARTICIPANTS: The study population consisted of 3103 men (1514 never-smokers, 1278 ex-smokers, and 311 smokers) and 1128 women (674 never-smokers, 347 ex-smokers, and 107 smokers). Information was collected on smoking status, average number of cigarettes smoked, and additional covariates. OUTCOME MEASURES: Saliva samples were taken on waking; waking + 0.5, 2.5, 8, and 12 h; and bedtime for the assessment of cortisol. RESULTS: Smoking status was significantly associated with increased salivary cortisol release throughout the day (P < 0.001) adjusted for covariates; this was apparent for the cortisol awakening response (P < 0.001) when examined separately. Compared with never-smokers, smokers had higher release of total cortisol (P = 0.002), whereas no difference was observed between never-smokers and ex-smokers (P = 0.594): mean release per hour (nanomoles per liter), never-smokers, 4.13 [confidence interval (CI) 4.02-4.24]; ex-smokers, 4.21 (CI 4.08-4.35); smokers, 4.63 (CI 4.35-4.93). There was no significant relationship between number of cigarettes smoked and total cortisol release. However, a difference was observed for the cortisol awakening response: mean release by tertiles of cigarettes smoked (nanomoles per liter): high, 13.49 (CI 10.74-16.23); medium, 9.58 (CI 7.40-11.76); low, 8.49 (CI 5.99-10.99), P = 0.029. CONCLUSION: Salivary cortisol is increased in current smokers, compared with nonsmokers; no differences were observed between ex-smokers and never-smokers, suggesting that smoking has a short-term effect on the neuroendocrine system.

Informative Observation in Health Data: Association of Past Level and Trend with Time to Next Measurement.

In routine health data, risk factors and biomarkers are typically measured irregularly in time, with the frequency of their measurement depending on a range of factors - for example, sicker patients are measured more often. This is termed informative observation. Failure to account for this in subsequent modelling can lead to bias. Here, we illustrate this issue using body mass index measurements taken on patients with type 2 diabetes in Salford, UK. We modelled the observation process (time to next measurement) as a recurrent event Cox model, and studied whether previous measurements in BMI, and trends in the BMI, were associated with changes in the frequency of measurement. Interestingly, we found that increasing BMI led to a lower propensity for future measurements. More broadly, this illustrates the need and opportunity to develop and apply models that account for, and exploit, informative observation.