Risk of lung cancer and physical activity by smoking status and body mass index, the Norwegian Women and Cancer Study

Gaps in the usage and reporting of multiple imputation for incomplete data: findings from a scoping review of observational studies addressing causal questions

BACKGROUND

Missing data are common in observational studies and often occur in several of the variables required when estimating a causal effect, i.e. the exposure, outcome and/or variables used to control for confounding. Analyses involving multiple incomplete variables are not as straightforward as analyses with a single incomplete variable. For example, in the context of multivariable missingness, the standard missing data assumptions ("missing completely at random", "missing at random" [MAR], "missing not at random") are difficult to interpret and assess. It is not clear how the complexities that arise due to multivariable missingness are being addressed in practice. The aim of this study was to review how missing data are managed and reported in observational studies that use multiple imputation (MI) for causal effect estimation, with a particular focus on missing data summaries, missing data assumptions, primary and sensitivity analyses, and MI implementation.

METHODS

We searched five top general epidemiology journals for observational studies that aimed to answer a causal research question and used MI, published between January 2019 and December 2021. Article screening and data extraction were performed systematically.

RESULTS

Of the 130 studies included in this review, 108 (83%) derived an analysis sample by excluding individuals with missing data in specific variables (e.g., outcome) and 114 (88%) had multivariable missingness within the analysis sample. Forty-four (34%) studies provided a statement about missing data assumptions, 35 of which stated the MAR assumption, but only 11/44 (25%) studies provided a justification for these assumptions. The number of imputations, MI method and MI software were generally well-reported (71%, 75% and 88% of studies, respectively), while aspects of the imputation model specification were not clear for more than half of the studies. A secondary analysis that used a different approach to handle the missing data was conducted in 69/130 (53%) studies. Of these 69 studies, 68 (99%) lacked a clear justification for the secondary analysis.

CONCLUSION

Effort is needed to clarify the rationale for and improve the reporting of MI for estimation of causal effects from observational data. We encourage greater transparency in making and reporting analytical decisions related to missing data.

Non-occupational physical activity and risk of cardiovascular disease, cancer and mortality outcomes: a dose-response meta-analysis of large prospective studies

OBJECTIVE

To estimate the dose-response associations between non-occupational physical activity and several chronic disease and mortality outcomes in the general adult population.

DESIGN

Systematic review and cohort-level dose-response meta-analysis.

DATA SOURCES

PubMed, Scopus, Web of Science and reference lists of published studies.

ELIGIBILITY CRITERIA

Prospective cohort studies with (1) general population samples >10 000 adults, (2) ≥3 physical activity categories, and (3) risk measures and CIs for all-cause mortality or incident total cardiovascular disease, coronary heart disease, stroke, heart failure, total cancer and site-specific cancers (head and neck, myeloid leukaemia, myeloma, gastric cardia, lung, liver, endometrium, colon, breast, bladder, rectum, oesophagus, prostate, kidney).

RESULTS

196 articles were included, covering 94 cohorts with >30 million participants. The evidence base was largest for all-cause mortality (50 separate results; 163 415 543 person-years, 811 616 events), and incidence of cardiovascular disease (37 results; 28 884 209 person-years, 74 757 events) and cancer (31 results; 35 500 867 person-years, 185 870 events). In general, higher activity levels were associated with lower risk of all outcomes. Differences in risk were greater between 0 and 8.75 marginal metabolic equivalent of task-hours per week (mMET-hours/week) (equivalent to the recommended 150 min/week of moderate-to-vigorous aerobic physical activity), with smaller marginal differences in risk above this level to 17.5 mMET-hours/week, beyond which additional differences were small and uncertain. Associations were stronger for all-cause (relative risk (RR) at 8.75 mMET-hours/week: 0.69, 95% CI 0.65 to 0.73) and cardiovascular disease (RR at 8.75 mMET-hours/week: 0.71, 95% CI 0.66 to 0.77) mortality than for cancer mortality (RR at 8.75 mMET-hours/week: 0.85, 95% CI 0.81 to 0.89). If all insufficiently active individuals had achieved 8.75 mMET-hours/week, 15.7% (95% CI 13.1 to 18.2) of all premature deaths would have been averted.

CONCLUSIONS

Inverse non-linear dose-response associations suggest substantial protection against a range of chronic disease outcomes from small increases in non-occupational physical activity in inactive adults. PROSPERO registration number CRD42018095481.

Physical activity and risk of lung cancer: A systematic review and dose-response meta-analysis of cohort studies

Objective

To synthesize the knowledge about the association of total physical activity (TPA), leisure-time physical activity (LTPA), occupational physical activity (OPA) and lung cancer risk and explore the dose-response relationship between LTPA level and lung cancer.

Methods

PubMed and Web of Science were searched up to 17 November 2021. The summary relative risks (RRs) and 95% confidence intervals (CIs) were calculated by random-effects or fixed-effects model. The dose-response analysis was conducted with restricted cubic splines.

Results

We identified 25 articles (42 cohort studies) that assessed the physical activity-lung cancer association, including 9,983,295 study participants and 85,988 incident cases of lung cancer. When comparing the highest to the lowest level of TPA and LTPA, lung cancer risk reduced 22% (RR, 0.78; 95% CI: 0.70, 0.86) and 12% (RR, 0.88; 95% CI: 0.83, 0.93), respectively. We found an approximately U-shaped association between LTPA and lung cancer (P non-linearity < 0.001), with the lowest risk at 15 metabolic equivalent of task hours per week (h/wk) of LTPA. Compared to participants with sitting occupations, lung cancer risk significantly increased among those being unemployed (RR, 1.33; 95% CI: 1.17, 1.51) or with standing occupations (RR, 1.37; 95% CI: 1.15, 1.63), but not among those with light or high OPA.

Conclusions

Our meta-analysis supported a protective effect of TPA and LTPA, but not OPA, on lung cancer risk. The novel finding of a U-shaped association between LTPA and lung cancer risk warrants further investigation.

Evaluating risk factors for lung cancer among never-smoking individuals using two Australian studies

PURPOSE

Lung cancer (LC) in never-smoking individuals would rank as Australia's eighth most deadly cancer, yet risk factors remain uncertain. We investigated demographic, lifestyle and health-related exposures for LC among never-smoking Australians.

METHODS

Using the prospective 45 and Up Study with 267,153 New South Wales (NSW) residents aged ≥ 45 years at recruitment (2006-2009), we quantified the relationship of 20 potential exposures with LC among cancer-free participants at baseline who self-reported never smoking. Adjusted hazard ratios (HR) and 95% confidence intervals (CI) for incident LC were estimated using Cox regression. The NSW Cancer, Lifestyle and Evaluation of Risk (CLEAR) Study, a case-control study including 10,781 NSW residents aged ≥ 18 years (2006-2014), was used to examine 16 potential LC exposures among cancer-free never-smoking participants. Adjusted odds ratios (OR) and 95% CI of LC were estimated using logistic regression.

RESULTS

There were 226 LC cases among 132,354 cancer-free 45 and Up Study participants who reported never smoking, with a median follow-up of 5.41 years. The CLEAR Study had 58 LC cases and 1316 cancer-free controls who had never smoked. Analyses of both datasets showed that Asian-born participants had a higher risk of LC than those born elsewhere: cohort, adjusted HR = 2.83 (95% CI 1.64-4.89) and case-control, adjusted OR = 3.78 (1.19-12.05). No significant association with LC was found for other exposures.

CONCLUSION

Our findings support the growing evidence that never-smoking, Asian-born individuals are at higher risk of developing LC than those born elsewhere. Ethnicity could be considered when assessing potential LC risk among never-smoking individuals.

Physical activity in Sami and non-Sami populations in rural Northern Norway, the SAMINOR 2 Clinical Survey

Background

The Sami people is an indigenous minority population living in the northern parts of Norway and mainly in rural areas. We lack data of contemporary levels of physical activity (PA) in rural regions of Northern Norway and in the Sami population in particular. We aimed to describe the PA levels and investigate whether PA levels differs between Sami and non-Sami and between coastal and inland areas.

Methods

We used data from the second survey of the Population-based Study on Health and Living Conditions in Regions with Sami and Norwegian Populations – the SAMINOR 2 Clinical Survey (2012–2014) that includes the adult population in 10 municipalities in the counties Troms, Finnmark and Nordland. Participants self-reported on PA, ethnicity and modifiable lifestyle factors. Twelve thousand four hundred fifty-five individuals were invited with a response rate of 48.2% (n = 6004 participants). We tested differences using chi-square tests, two sample t-tests and linear regression models.

Results

Among 5628 participants, 41.1 and 40.9% of men and women, respectively, were defined as Sami. We found no ethnic differences in PA in men overall. However, Sami men living in Tana, and Nesseby reported higher PA compared to non-Sami men in the same area. For Sami women there was overall lower PA levels compared to non-Sami women, especially pronounced in Kautokeino/ Karasjok.

Conclusion

This study showed small differences in PA levels between Sami and non-Sami men. Sami women had lower PA levels compared to their non-Sami counterparts. It is important to identify whether there are differences in various ethnic populations, together with other predictors for PA in future planning of public health interventions.

Objectives, design and main findings until 2020 from the Rotterdam Study

The Rotterdam Study is an ongoing prospective cohort study that started in 1990 in the city of Rotterdam, The Netherlands. The study aims to unravel etiology, preclinical course, natural history and potential targets for intervention for chronic diseases in mid-life and late-life. The study focuses on cardiovascular, endocrine, hepatic, neurological, ophthalmic, psychiatric, dermatological, otolaryngological, locomotor, and respiratory diseases. As of 2008, 14,926 subjects aged 45 years or over comprise the Rotterdam Study cohort. Since 2016, the cohort is being expanded by persons aged 40 years and over. The findings of the Rotterdam Study have been presented in over 1700 research articles and reports. This article provides an update on the rationale and design of the study. It also presents a summary of the major findings from the preceding 3 years and outlines developments for the coming period.

The relation between healthy lifestyle changes and decrease in systemic inflammation in patients with stable cardiovascular disease

BACKGROUND AND AIMS

Pharmacological lowering of inflammation has proven effective in reducing recurrent cardiovascular event rates. Aim of the current study is to evaluate lifestyle changes (smoking cessation, weight loss, physical activity level increase, alcohol moderation, and a summary lifestyle improvement score) in relation to change in plasma C-reactive protein (CRP) concentration in patients with established cardiovascular disease.

METHODS

In total, 1794 patients from the UCC-SMART cohort with stable cardiovascular disease and CRP levels ≤10 mg/L, who returned for a follow-up study visit after median 9.9 years (IQR 5.4-10.8), were included. The relation between changes in smoking status, weight, physical activity, alcohol consumption, a summary lifestyle improvement score and change in plasma CRP concentration was evaluated with linear regression analyses.

RESULTS

Smoking cessation was related to a 0.40 mg/L decline in CRP concentration (β-coefficient -0.40; 95%CI -0.73,-0.07). Weight loss (per 1SD = 6.4 kg) and increase in physical activity (per 1 SD = 48 MET hours per week) were related to a decrease in CRP concentration (β-coefficients -0.25; 95%CI -0.33,-0.16 and -0.09; 95%CI -0.17,-0.01 per SD). Change in alcohol consumption was not related to CRP difference. Every point higher in the summary lifestyle improvement score was related to a decrease in CRP concentration of 0.17 mg/L (β-coefficient -0.17; 95%CI -0.26,-0.07).

CONCLUSIONS

Smoking cessation, increase in physical activity, and weight loss are related to a decrease in CRP concentration in patients with stable cardiovascular disease. Patients with the highest summary lifestyle improvement score have the most decrease in CRP concentration. These results may indicate that healthy lifestyle changes contribute to lowering systemic inflammation, potentially leading to a lower cardiovascular risk in patients with established cardiovascular disease.

Association Between Physical Activity and Lower Risk of Lung Cancer: A Meta-Analysis of Cohort Studies

Background: Epidemiological evidences regarding the association between physical activity and the risk of lung cancer are still controversial.

Objectives: We aimed to investigate the relationship between physical activity and risk of lung cancer in men and women, as well as other high-risk populations such as cigarette smokers.

Methods: We conducted a meta-analysis of cohort studies to evaluate the association between physical activity and risk of lung cancer. Relevant studies were identified by searching PubMed and Web of Knowledge through August 2018. Study-specific relative risk (RR) with 95% confidence interval (CI) were pooled using random effect model when significant heterogeneity was detected.

Results:Twenty cohort studies with a total of 2,965,811 participants and 31,807 lung cancer cases were included. There was an inverse association between the physical activity and risk of lung cancer. Compared with the low level of physical activity, the pooled RR was 0.83 (95%CI: 0.77, 0.90), with significant heterogeneity (I² = 62.6%, P_{heterogeneity} < 0.001). The corresponding pooled RRs were 0.90 (95%CI: 0.82, 0.99) for women and 0.81 (95%CI: 0.73, 0.90) for men. Smokers with a high level of physical activity were associated with a 10% lower risk for lung cancer (RR = 0.90, 95% CI: 0.84, 0.97), while the association was not significant among non-smokers (RR = 0.95, 95% CI: 0.88, 1.03). Subgroups analysis stratified by whether the studies adjusted for smoking intensity and durations yielded the same magnitude of RR. However, the RR for subgroups without adjustment for dietary factors was 0.74 (95%CI: 0.71, 0.77), which was significantly lower than that with dietary factors adjusted (RR = 0.89, 95%CI: 0.84, 0.95).

Conclusions:Increased physical activity might be associated with lower risk of lung cancer. Such inverse association was identified among smokers rather than non-smokers. Large interventional studies are expected to further verify these findings.