Contribution of remnant cholesterol to cardiovascular risk.

BACKGROUND: Remnant cholesterol in triglyceride-rich lipoproteins is associated observationally and genetic, causally with increased risk of atherosclerotic cardiovascular disease in healthy individuals. OBJECTIVES: We tested the hypothesis that an unmet medical need exists in individuals with high nonfasting remnant cholesterol and prior atherosclerotic cardiovascular disease. METHODS: From amongst 109 574 individuals in a prospective cohort study of the Danish general population, we included 2973 individuals aged 20-80 with baseline diagnoses of myocardial infarction/ischaemic stroke ascertained from national Danish health registries. RESULTS: The recurrent major cardiovascular event (MACE) incidence rates per 1000 person-years were 39 (95% confidence interval: 30-50) for individuals with remnant cholesterol levels ≥ 1.5 mmol L-1 (≥58 mg dL-1 ), 31 (26-37) for 1-1.49 mmol L-1 (39-57 mg dL-1 ), 27 (24-31) for 0.5-0.99 mmol L-1 (19-38 mg dL-1 ) and 23 (19-27) for individuals with remnant cholesterol < 0.5 mmol L-1 (<19 mg dL-1 ). Compared to individuals with remnant cholesterol < 0.5 mmol L-1 (<19 mg dL-1 ), the subhazard ratio for recurrent MACE was 1.23 (95% CI: 0.98-1.55) for individuals with remnant cholesterol levels of 0.5-0.99 mmol L-1 (19-38 mg dL-1 ), 1.48 (1.14-1.92) for 1-1.49 mmol L-1 (39-57 mg dL-1 ) and 1.79 (1.28-2.49) for ≥ 1.5 mmol L-1 (≥58 mg dL-1 ). The recurrent MACE incidence rates per 1000 person-years for individuals with remnant cholesterol levels < 0.5 mmol L-1 (<19 mg dL-1 ) and ≥ 1.5 mmol L-1 (≥58 mg dL-1 ) were 10 (6.6-15) and 31 (21-47) for those below age 65 and correspondingly 25 (21-30) and 43 (32-59) for those with LDL cholesterol levels < 3 mmol L-1 (<116 mg dL-1 ), respectively. For a 20% recurrent MACE risk reduction in secondary prevention, an estimated remnant cholesterol lowering of 0.83 mmol L-1 (32 mg dL-1 ) would be needed. CONCLUSIONS: In individuals with a diagnosis of myocardial infarction/ischaemic stroke, a lower remnant cholesterol of 0.8 mmol L-1 (32 mg dL-1 ) was estimated to reduce recurrent MACE by 20% in secondary prevention. Our data indicate an unmet medical need for secondary prevention in individuals with high nonfasting remnant cholesterol levels.

Coffee intake protects against symptomatic gallstone disease in the general population: a Mendelian randomization study.

BACKGROUND AND OBJECTIVES: Coffee intake is associated with low risk of symptomatic gallstone disease (GSD). We tested the hypothesis that high coffee intake causally protects against symptomatic GSD using a Mendelian randomization design. METHODS: First, we tested whether high coffee intake was associated with low risk of GSD in 104 493 individuals from the general population. Mean follow-up was 8 years (range: <1-13 years). Secondly, we tested whether two genetic variants near CYP1A1/A2 (rs2472297) and AHR (rs4410790), combined as an allele score, were associated with higher coffee intake measured as a continuous variable. Thirdly, we tested whether the allele score was associated with lower risk of GSD in 114 220 individuals including 7294 gallstone events. Mean follow-up was 38 years (range: <1-40 years). RESULTS: In observational analysis, those with coffee intake of >6 cups daily had 23% lower risk of GSD compared to individuals without coffee intake [hazard ratio (HR) = 0.77 (95% confidence interval: 0.61-0.94)]. In genetic analysis, there was a stepwise higher coffee intake of up to 41% (caffeine per day) in individuals with 4 (highest) versus 0 (lowest) coffee intake alleles (P for trend = 3 x 10-178 ) and a corresponding stepwise lower risk of GSD up to 19%[HR = 0.81 (0.69-0.96)]. The estimated observational odds ratio for GSD for a one cup per day higher coffee intake was 0.97 (0.96-0.98), equal to 3% lower risk. The corresponding genetic odds ratio was 0.89 (0.83-0.95), equal to 11% lower risk. CONCLUSION: High coffee intake is associated observationally with low risk of GSD, and with genetic evidence to support a causal relationship.

Cardiac chamber volumes and left ventricular mass in people living with HIV and matched uninfected controls.

OBJECTIVES: People living with HIV (PLWH) have increased risk of cardiovascular diseases compared with uninfected populations. We assessed structural cardiac abnormalities and their associated risk factors in well-treated PLWH and uninfected controls using multidetector computed tomography (MDCT). METHODS: People living with HIV and age- and sex-matched uninfected controls underwent MDCT to determine left atrial volume (LAV), left ventricular diastolic volume (LVDV), right ventricular diastolic volume (RVDV) and left ventricular mass (LVM). All outcomes were indexed to body surface area (BSA) (LAVi, LVDVi, RVDVi and LVMi). RESULTS: A total of 592 PLWH and 1184 uninfected controls were included in the study. PLWH had smaller mean (SD) LAVi [40 (8) vs. 41 (9) mL/m2 ; P = 0.002] and LVDVi [61 (13) vs. 65 (14) mL/m2 ; P < 0.001] but larger RVDVi [89 (18) vs. 86 (17) mL/m2 ; P < 0.001] than uninfected controls. HIV was independently associated with 7 mL (95% CI: -10 to -3) smaller LVDV, and with 12 mL (95% CI: 8-16) larger RVDV, and 4 g (95% CI: 1-6) larger LVM after adjustment for cardiovascular risk factors and BSA. Large RVDV in PLWH was not associated with obstructive lung function. CONCLUSIONS: HIV was independently associated with smaller LVDV and larger RVDV and LVM. Alterations in cardiac chamber volumes in PLWH were mainly minor. The clinical impact of these findings is uncertain, but it seems unlikely that alterations in cardiac chamber volumes explain the increased burden of cardiovascular disease previously observed in PLWH.

Prevalence of impaired renal function in virologically suppressed people living with HIV compared with controls: the Copenhagen Comorbidity in HIV Infection (COCOMO) study.

OBJECTIVES: While renal impairment is reported more frequently in people living with HIV (PLWH) than in the general population, the PLWH samples in previous studies have generally been dominated by those at high renal risk. METHODS: Caucasian PLWH who were virologically suppressed on antiretroviral treatment and did not have injecting drug use or hepatitis C were recruited from the Copenhagen Comorbidity in HIV Infection (COCOMO) study. Sex- and age-matched controls were recruited 1:4 from the Copenhagen General Population Study up to November 2016. We defined renal impairment as one measurement of estimated glomerular filtration rate ≤ 60 mL/min/1.73 m2 , and assessed associated factors using adjusted logistic regression models. The impact of HIV-related factors was explored in a subanalysis. RESULTS: Among 598 PLWH and 2598 controls, the prevalence of renal impairment was 3.7% [95% confidence interval (CI) 2.3-5.5%] and 1.7% (95% CI 1.2-2.2%; P = 0.0014), respectively. After adjustment, HIV status was independently associated with renal impairment [odds ratio (OR) 3.4; 95% CI 1.8-6.3]. In addition, older age [OR 5.4 (95% CI 3.9-7.5) per 10 years], female sex [OR 5.0 (95% CI 2.6-9.8)] and diabetes [OR 2.9 (95% CI 1.3-6.7)] were strongly associated with renal impairment. The association between HIV status and renal impairment became stronger with older age (P = 0.02 for interaction). Current and nadir CD4 counts, duration of HIV infection and previous AIDS-defining diagnosis were not associated with renal impairment among virologically suppressed PLWH. CONCLUSIONS: The prevalence of renal impairment is low among low-risk virologically suppressed Caucasian PLWH, but remains significantly higher than in controls. Renal impairment therefore remains a concern in all PLWH and requires ongoing attention.

Chylomicronemia risk factors ranked by importance for the individual and community in 108 711 women and men.

BACKGROUND: Hypertriglyceridemia prevalence is increasing as more individuals become obese, and chylomicronemia risk factors for the individual and community have not been described previously. OBJECTIVE: To describe chylomicronemia risk factors in the general population for individuals and community. METHODS: A total of 108 711 individuals from the Copenhagen General Population Study were grouped as unlikely chylomicronemia (nonfasting triglycerides <2 mmol L-1 (177 mg dL-1 )), possible chylomicronemia (2-4.99 mmol L-1 (177-442 mg dL-1 )), probable chylomicronemia (5-9.99 mmol L-1 (443-885 mg dL-1 )) and definite chylomicronemia (≥10 mmol L-1 (≥ 886 mg dL-1 )). Relative risk (RR) from Poisson regression ranked dichotomized chylomicronemia risk factors for individuals, and population attributable fractions (PAF) for the community: type 2 diabetes, alcohol intake, obesity, fat intake, hypothyroidism, kidney function, education, sedentary lifestyle, menopause and hormone replacement (women). RESULTS: For women and men, chylomicronemia was unlikely in 81% and 64%, possible in 18% and 33%, probable in 1% and 3% and definite in 0.03% and 0.14%, respectively. For the individual, the three top-ranked risk factors for probable/definite versus unlikely chylomicronemia in women were type 2 diabetes (RR: 4.21; 95% confidence interval: 3.30-5.36), menopause (RR: 3.74; 2.62-5.36) and obesity (RR: 3.44; 2.81-4.21). Corresponding top-ranked risk factors in men were obesity (RR: 3.86; 3.46-4.30), type 2 diabetes (RR: 1.88; 1.61-2.19) and reduced kidney function (RR: 1.86; 1.48-2.34). For the community, top-ranked risk factors in women were menopause (PAF: 63%), obesity (PAF: 29%) and type 2 diabetes (PAF: 15%). Corresponding top-ranked risk factors in men were obesity (PAF: 29%), type 2 diabetes (PAF: 6.4%) and sedentary lifestyle (PAF: 6.0%). CONCLUSIONS: Obesity and type 2 diabetes were the most important modifiable chylomicronemia risk factors in women and men, both for the individual and community. This could influence chylomicronemia prevention and help design randomized trials aimed at reducing triglycerides.

Lactase persistence, milk intake, hip fracture and bone mineral density: a study of 97 811 Danish individuals and a meta-analysis.

BACKGROUND: Whether a causal relationship exists between milk intake and reduced risk of fractures is unclear. OBJECTIVES: We tested the hypothesis that genetically determined milk intake reduces the risk of fractures and increases bone mineral density (BMD). METHODS: We investigated the association between milk intake, LCT-13910 C/T (rs4988235), which is associated with lactase persistence (TT/TC) in Northern Europeans, and hip fractures in three Danish prospective studies (N = 97 811, age ≥20 years). We added meta-analyses of LCT-13910 and fractures and BMD from five published Northern European population studies. RESULTS: In the Danish studies, the adjusted hazard ratio (HR) for hip fracture per one glass per week higher milk intake was 1.00 (95% CI: 0.99-1.01). The per T-allele milk intake was 0.58 (0.49-0.68) glasses per week, but HR was 1.01 (0.94-1.09) for hip fracture. In meta-analyses of Danish studies with published Northern European population studies, the random effects odds ratio for any fracture was 0.86 (0.61-1.21; I2 = 73%) for TT vs. CC and 0.90 (0.68-1.21; I2 = 63%) for TC vs. CC. The standardized mean difference in femoral neck BMD was 0.10 (0.02-0.18; I2 = 0%) g cm-2 for TT vs. CC and 0.06 (-0.04 to 0.17; I2 = 17%) g cm-2 for TC vs. CC. There were no differences in lumbar spine or total hip BMD comparing TT or TC with CC. CONCLUSION: Genetically lifelong lactase persistence with high milk intake was not associated with hip fracture in Danish population-based cohorts. A meta-analysis combining Danish studies with published Northern European population studies also showed that lactase persistence was not associated with fracture risk. Genetic lactase persistence was associated with a higher femoral neck BMD, but not lumbar spine or total hip BMD.

Diagnostic performance of clinical characteristics to detect airflow limitation in people living with HIV and in uninfected controls.

OBJECTIVES: Chronic obstructive pulmonary disease (COPD) is underdiagnosed in the general population and possibly also in people living with HIV (PLWH). We evaluated the diagnostic performance of symptoms and risk factors for assessment of airflow limitation in PLWH and in uninfected controls. METHODS: Spirometry was performed in the Copenhagen Comorbidity in HIV Infection (COCOMO) study and Copenhagen General Population Study (CGPS), and airflow limitation was defined by forced expiratory volume in 1 s/forced vital capacity < lower limit of normal. We calculated the sensitivity, specificity, predictive values and area under the curve (AUC) of symptoms and risk factors for assessment of airflow limitation in PLWH and uninfected controls. RESULTS: A total of 1083 PLWH and 12 074 uninfected controls were included in the study. The sensitivity for sputum, chronic cough, breathlessness, wheezing, current and cumulative smoking and self-reported COPD was higher, but the specificity lower, in PLWH than in uninfected controls. The negative and positive predictive values were largely similar between the groups. The AUCs were similar or slightly higher in PLWH and highest for > 20 pack-years smoked [0.65; 95% confidence interval (CI) 0.58-0.72] and wheezing (0.64; 95% CI 0.57-0.71). A summed score for five variables was associated with slightly higher AUC in PLWH compared with uninfected controls [0.71 (95% CI 0.63-0.79) versus 0.65 (95% CI 0.63-0.68), respectively; P = 0.06]. CONCLUSIONS: Clinical variables were relatively poor discriminators of airflow limitation in PLWH and uninfected controls. Active COPD case finding by screening for symptoms and relevant exposures, as recommended in the general population, is likely to yield similar diagnostic power in PLWH.

HIV infection is independently associated with a higher concentration of alpha-1 antitrypsin.

OBJECTIVES: Alpha-1 antitrypsin (AAT) deficiency is associated with an increased risk of chronic obstructive pulmonary disease and has been related to CD4 T-cell count decline in people living with HIV (PLWH). We determined whether HIV status is associated with AAT concentrations and assessed associations between AAT concentration, pulmonary function and immunological status. METHODS: Alpha-1 antitrypsin was measured and spirometry performed in 1011 PLWH from the Copenhagen Comorbidity in HIV Infection (COCOMO) study and in 11 962 age- and sex-matched uninfected controls. We studied associations between AAT concentration, HIV status, pulmonary function, and current and nadir CD4 T-cell counts using multivariate linear regression analyses. RESULTS: The mean age of PLWH was 50.7 [standard deviation (SD) 11.3] years and 98.6% were receiving combination antiretroviral therapy (cART). The mean current CD4 T-cell count was 718 (SD 284) cells/µL. PLWH had a higher median AAT concentration than uninfected controls [1.4 (interquartile range (IQR) 1.3-1.6) versus 1.3 (IQR 1.2-1.4) g/L; P < 0.0001] and HIV infection was independently associated with higher AAT concentration [adjusted ß = 0.10 g/L; 95% confidence interval (CI) 0.08; 0.11 g/L; P < 0.001]. Low AAT concentration (< 1.0 g/L) was not more common in PLWH with airflow limitation (defined as forced expiratory volume in 1 second/forced vital capacity (FEV1 /FVC) < 0.7 with FEV1 -predicted < 80%) compared with uninfected controls with airflow limitation, and the effect of AAT on FEV1 %-predicted was comparable to that in uninfected controls (P-interaction = 0.66). AAT concentration was not associated with current or nadir CD4 T-cell count. CONCLUSIONS: HIV infection was independently associated with a higher concentration of AAT through unknown mechanisms. However, AAT does not seem to contribute to lower pulmonary function or to low CD4 T-cell counts in PLWH.

High plasma 25-hydroxyvitamin D and high risk of nonmelanoma skin cancer: a Mendelian randomization study of 97 849 individuals.

BACKGROUND: High plasma 25-hydroxyvitamin D [25(OH)D] concentration has been associated observationally with a high risk of nonmelanoma skin cancer (NMSC), whereas many studies suggest that vitamin D could have a protective effect against cancer. The true association between vitamin D and risk of skin cancer remains unclear. OBJECTIVES: To test the hypothesis that genetically high plasma 25(OH)D protects against NMSC. METHODS: We included 103 084 individuals from the Danish general population, of whom 35 298 had plasma 25(OH)D measured and 97 849 were genotyped for four genetic variants near DHCR7 and CYP2R1 associated with 25(OH)D concentrations. We tested the association between plasma 25(OH)D levels and NMSC observationally and between genetically determined 25(OH)D levels and NMSC, using an instrumental variable approach. RESULTS: Multivariate-adjusted hazard ratios of NMSC were 3·27 [95% confidence interval (CI) 2·22-4·84] for plasma 25(OH)D ≥ 50 nmol L-1 vs. < 25 nmol L-1 . Genetic variants around DHCR7 and CYP2R1 were associated with up to 8·2 nmol L-1 higher 25(OH)D concentrations (F = 314). The odds ratio (OR) for a genetically determined 20 nmol L-1 higher plasma 25(OH)D was 1·11 (95% CI 0·91-1·35) for NMSC, with a corresponding observational multivariable adjusted OR of 1·13 (95% CI 1·10-1·17). CONCLUSIONS: Genetically determined high 25(OH)D levels did not appear to protect against NMSC, whereas high plasma 25(OH)D concentrations were associated with an observational high risk of NMSC. Thus, the observational association likely reflects confounding by sun exposure rather than causality.

Common breast cancer risk alleles and risk assessment: a study on 35 441 individuals from the Danish general population.

Background: We hypothesized that common breast cancer risk alleles are associated with incidences of breast cancer and other cancers in the general population, and identify low risk women among those invited for screening mammography. Participants and Methods: About 35 441 individuals from the Danish general population were followed in Danish health registries for up to 21 years after blood sampling. After genotyping 72 breast cancer risk loci, each with 0­2 alleles, the sum for each individual was calculated. We used the simple allele sum instead of the conventional polygenic risk score, as it is likely more sensitive in detecting associations with risks of other endpoints than breast cancer. Results: Breast cancer incidence in the 19 010 women was increased across allele sum quintiles (log-rank trend test; P = 1×10 − 12), but not incidence of other cancers (P = 0.41). Age- and study-adjusted hazard ratio for the fifth versus the first allele sum quintile was 1.82 (95% confidence interval; 1.53­2.18). Corresponding hazard ratios per allele were 1.04 (1.03­1.05) and 1.05 (1.02­1.08) for breast cancer incidence and mortality, similar across risk factors. In 50-year-old women, the starting age for screening mammography in Denmark, the average 5-year breast cancer risk was 1.5%, overall and 1.1%, 1.4%, 1.6%, 1.7%, 2.1%, for the first through fifth quintile, respectively. Based on age, nulliparity, familial history, and allele sum, 25% of women aged 50­69 years, and 94% of women aged 40­49 years, had absolute 5-year breast cancer risks ≤ 1.5%. Using polygenic risk score led to similar results. Conclusion: Common breast cancer risk alleles are associated with incidence and mortality of breast cancer in the general population, but not with other cancers. After including breast cancer allele sum in risk assessment, 25% of women currently being offered screening mammography had an absolute 5-year risk below the cutoff of average risk for a 50-year-old woman.

Asthma, other atopic conditions and risk of infections in 105 519 general population never and ever smokers.

BACKGROUND: Individuals with atopic conditions may have increased susceptibility to infections outside the organs directly affected by their atopic condition. OBJECTIVE: We tested the hypothesis that atopic conditions overall, and stratified by smoking history, are associated with increased risk of hospitalization for infections. METHODS: We collected information on smoking history and self-reported atopic conditions from 105 519 individuals from the general population and followed them for up to 23 years for infectious disease hospitalizations and deaths. For asthma, we focused on never smokers with asthma diagnosed before age 50 (early asthma) to minimize confounding by chronic obstructive pulmonary disease. RESULTS: During follow-up, 11 160 individuals had infections. Never smokers with early asthma versus no atopic conditions had significantly increased risks of any infection (hazard ratio 1.65; 95% confidence interval 1.40-1.94), pneumonia (2.44; 1.92-3.11) and any non-respiratory tract infection (1.36; 1.11-1.67); results were similar in ever smokers. Never smokers with any asthma had significantly increased risks of any infection (1.44; 1.24-1.66) and pneumonia (1.99; 1.62-2.44). Neither atopic dermatitis (1.00; 0.91-1.10) nor hay fever (1.00; 0.93-1.07) was associated with risk of any infection. In never smokers, risk estimates for any infection were comparable between asthma and diabetes, as were the population attributable fractions of 2.2% for any asthma and 2.9% for diabetes. CONCLUSION: Early asthma was associated with significantly increased risks of any infection, pneumonia and any non-respiratory tract infection in never and ever smokers. In never smokers, risk estimates as well as population attributable fractions for any infection were comparable between asthma and diabetes, suggesting that asthma may be a substantial risk factor for infections in the general population.

Association between the antioxidant uric acid and depression and antidepressant medication use in 96 989 individuals.

OBJECTIVE: In the last decade, several studies have suggested that depression is accompanied by increased oxidative stress and decreased antioxidant defenses. We tested the hypothesis that high levels of the antioxidant uric acid are associated with lower risk of hospitalization with depression and use of prescription antidepressant medication. METHOD: We examined plasma levels of the antioxidant uric acid in 96 989 individuals from two independent cohort studies. Logistic regression and Cox proportional hazards regression models were multivariable adjusted for age, gender, alcohol, smoking, income, body mass index, C-reactive protein, hemoglobin, triglycerides, cardiovascular disease, diabetes, and intake of meat and vegetables. Results were performed separately in each study and combined in a meta-analysis. RESULTS: In both studies, high uric acid was associated with lower risk of hospitalization as in-patient or out-patient with depression and antidepressant medication use. A doubling in uric acid was associated with an effect estimate of 0.57 (95% CI 0.49-0.65) and 0.77 (0.73-0.81) for hospitalization with depression and antidepressant medication use. The association was consistent across strata of all covariates. Results were attenuated in Cox regression analyses with less statistical power. CONCLUSION: High plasma levels of uric acid were associated with low risk of depression hospitalization and antidepressant medication use.

Role of inflammatory marker YKL-40 in the diagnosis, prognosis and cause of cardiovascular and liver diseases.

This review summarizes present evidence for the role of YKL-40 in the diagnosis, prognosis and cause of cardiovascular and alcoholic liver disease. The question of whether YKL-40 is merely a marker or a causal factor in the development of cardiovascular and liver disease is addressed, with emphasis on the Mendelian randomization design. The Mendelian randomization approach uses genetic variants associated with lifelong high plasma YKL-40 levels that are largely unconfounded and not prone to reverse causation. Thus, the approach mimics a controlled double-blind randomized trial, but it uses genetic variants rather than a drug and placebo, and like a blinded trial, it allows inference about causality. Moreover, the review also covers background on the molecular biology and functions of YKL-40, YKL-40 levels in healthy individuals and reference range, and the role of YKL-40 as a biomarker of cardiovascular and alcoholic liver disease. YKL-40 is a plasma protein named after its three N-terminal amino acids, Y (tyrosine), K (lysine) and L (leucine), and its molecular weight of 40 kDa. It is produced by local inflammatory cells in inflamed tissues, such as lipid-laden macrophages inside the vessel wall and perhaps also hepatic stellate cells. Observational studies show that plasma YKL-40 levels are elevated in patients with cardiovascular and liver disease and are associated with disease severity and prognosis. Furthermore, elevated plasma YKL-40 levels in apparently healthy individuals are associated with a 2-fold increased risk of future ischemic stroke and venous thromboembolism, but not with myocardial infarction, suggesting that YKL-40 could play a role in the formation of embolisms rather than atherosclerosis per se. Further, elevated YKL-40 levels combined with excessive alcohol consumption are associated with 10-years risk of alcoholic liver cirrhosis of up to 7%, suggesting that YKL-40 can be used as a strong noninvasive marker of predicting alcoholic liver cirrhosis. Importantly, in Mendelian randomization studies, genetically elevated plasma YKL-40 levels were not associated with risk of cardiovascular and alcoholic liver disease, thus suggesting that plasma YKL-40 does not play a causal role in the development of these diseases. Despite this, plasma YKL-40 levels may play a role in disease progression after diagnosis, and inhibition of YKL-40 activity might be a novel therapy in some cardiovascular and liver diseases.

IgE and risk of cancer in 37 747 individuals from the general population.

BACKGROUND: Immunoglobulin E (IgE) is produced by plasma cells, often as part of an allergic immune response. It is currently unknown whether plasma IgE levels are associated with risk of cancer in individuals from the general population. We tested the hypothesis that high levels of plasma total IgE are associated with overall risk of cancer and with risk of specific cancers. MATERIALS AND METHODS: Plasma total IgE was measured in 37 747 individuals from the general population, and the participants were followed prospectively for up to 30 years. All statistical tests were two-sided. RESULTS: During a mean follow-up of 7 years, a first cancer was diagnosed in 3454 participants. The multivariable adjusted hazard ratio for a 10-fold higher level of IgE was 1.05 [95% confidence interval (CI) 1.00-1.11; P = 0.04] for any cancer, 0.44 (0.30-0.64; P = 0.00002) for chronic lymphocytic leukemia (CLL), 0.53 (0.33-0.84; P = 0.007) for multiple myeloma, 1.54 (1.04-2.29; P = 0.03) for other non-Hodgkin lymphoma, 1.38 (1.04-1.84; P = 0.03) for cancer of the oral cavity and pharynx, and 1.12 (1.00-1.25; P = 0.05) for lung cancer. The findings for CLL and multiple myeloma were generally robust; however, after correcting for 27 multiple comparisons only the finding for CLL remained significant. CONCLUSION: High levels of plasma total IgE were associated with low risk of CLL and possibly of multiple myeloma, without convincing evidence for high risk of any cancer type.

Obesity as a causal risk factor for deep venous thrombosis: a Mendelian randomization study.

OBJECTIVE: To test the hypothesis that obesity is causally associated with deep venous thrombosis (DVT). DESIGN: A Mendelian randomization design. SETTING: The Copenhagen General Population Study and the Copenhagen City Heart Study combined. SUBJECTS: Body mass index (BMI) measurements were available for 87, 574 individuals of Danish descent from the adult general population. All subjects completed questionnaires and were genotyped for the FTO rs9939609 variant. MAIN OUTCOME MEASURE: First events of DVT with or without pulmonary embolism (PE). ANALYSIS: The results were assessed using Cox regression, instrumental variable analysis and Poisson regression. RESULTS: Observationally, the risk of DVT increased with increasing BMI (P-trend < 0.0001). The multivariable-adjusted hazard ratio [95% confidence interval (CI)] for DVT was 1.3 (1.1-1.6) in overweight, 1.8 (1.4-2.2) in moderately obese and 3.4 (2.6-4.6) in severely obese compared with normal-weight individuals. For DVT complicated by PE, corresponding hazard ratios (95% CI) were 1.2 (0.8-1.8), 2.1 (1.3-3.5) and 5.1 (2.8-9.2). FTO AA versus TT genotype was associated with a 2.4% increase in BMI with hazard ratios (95% CI) of 1.09 (0.95-1.25) for DVT and 1.54 (1.12-2.10) for DVT complicated by PE. In instrumental variable analysis, the causal odds ratio (95% CI) for an increase in BMI of 1 kg m(-2) was 1.13 (0.92-1.39) for DVT alone and 1.86 (1.14-3.02) for DVT complicated by PE. The absolute 10-year risk of DVT in a high-risk group (i.e. those aged >60 years and homozygous for Factor V Leiden) was 35% in obese individuals and 18% in normal-weight individuals. CONCLUSION: A strong observational association between obesity and DVT with or without PE, supported by a direct genetic association between the obesity-specific locus FTO and DVT with PE, implies that obesity is likely to be causally associated with DVT.

Plasma testosterone in the general population, cancer prognosis and cancer risk: a prospective cohort study.

BACKGROUND: Testosterone is an important anabolic hormone in humans and in vitro testosterone stimulates growth of lung and colon cancer cells. We tested the hypothesis that plasma testosterone associate with increased risk of cancer and with increased risk of early death after cancer. MATERIALS AND METHODS: Plasma testosterone was measured in 8771 20- to 94-year-old men and women who participated in a prospective study of the general population. Participants were included in 1981-1983 and followed for a median of 22 years (range: 0-30 years). RESULTS: During follow-up, 1140 men and 809 women developed cancer. For risk of early death after cancer, for men, after adjustment for age at diagnosis, tumour stage at diagnosis, and time since blood-sampling, the hazard ratio was 1.30 [95% confidence interval (CI) 1.03-1.65] for the 2nd quintile, 1.31 (1.02-1.67) for the 3rd quintile, 1.52 (1.19-1.93) for the 4th quintile, and 1.52 (1.20-1.91) for the 5th quintile, versus the 1st quintile. For women, corresponding hazard ratios were 1.09 (0.81-1.46), 1.17 (0.86-1.59), 1.03 (0.76-1.39), and 1.80 (1.32-2.46). For risk of cancer, multifactorially adjusted hazard ratios for risk of any cancer were 1.07 (95% CI 0.98-1.18) and 1.06 (0.93-1.22) for men and women, respectively, when testosterone doubled. For both men and women, a doubling of testosterone was not associated with risk of any cancer type. CONCLUSIONS: In this prospective study of 8771 men and women from the general population followed for >30 years, increased levels of testosterone were associated with a 30%-80% increased risk of early death after cancer, but unchanged risk of incident cancer.

Genetically elevated bilirubin and risk of ischaemic heart disease: three Mendelian randomization studies and a meta-analysis.

BACKGROUND: Elevated plasma levels of bilirubin, an endogenous antioxidant, have been associated with reduced risk of ischaemic heart disease (IHD) and myocardial infarction (MI). Whether this is a causal relationship remains unclear. OBJECTIVE: We tested the hypothesis that elevated plasma bilirubin is causally related to decreased risk of IHD and MI. DESIGN: We used a Mendelian randomization approach and three independent studies from Copenhagen, Denmark. We measured bilirubin in 43 708 white individuals from the general population, and genotyped rs6742078 G>T in the uridine diphosphate glucuronosyltransferase 1A1 (UGT1A1) gene in 67 068 individuals, of whom 11 686 had IHD. RESULTS: Third versus first tertile of baseline bilirubin levels was associated with 134% increased bilirubin levels, with sex- and age-adjusted hazard ratios (HRs) of 0.86 [95% confidence interval (CI), 0.76-0.98; P = 0.02] for IHD and 0.81 (95% CI, 0.66-0.99; P = 0.04) for MI, but with corresponding multifactorially adjusted HRs of 0.93 (95% CI, 0.82-1.06; P = 0.29) and 0.90 (95% CI, 0.73-1.12; P = 0.35). UGT1A1 rs6742078 TT versus GG genotype was associated with 95% increased bilirubin levels (P < 0.001); TT versus GG genotype was associated with odds ratios (ORs) of 1.03 (95% CI, 0.96-1.11; P = 0.73) for IHD and 1.01 (95% CI, 0.92-1.12; P = 0.68) for MI. Finally, in a meta-analysis of the present three studies and eight previous studies including a total of 14 711 cases and 60 324 controls, the random effects OR for ischaemic cardiovascular disease for genotypes with approximately 100% increased bilirubin levels versus reference genotypes was 1.01 (95% CI, 0.88-1.16). CONCLUSION: These data suggest that plasma bilirubin is not causally associated with risk of IHD.

Cancer risk by combined levels of YKL-40 and C-reactive protein in the general population.

BACKGROUND: YKL-40 and C-reactive protein (CRP) are biomarkers that may reflect cancer-related subclinical inflammation. We assessed elevated YKL-40 and CRP levels as combined risk predictors for cancer. METHODS: We measured plasma YKL-40 and CRP at baseline in 8706 individuals from the Danish general population. RESULTS: Hazard ratio (HR) of gastrointestinal cancer for a doubling of YKL-40 levels was 1.37 (95% CI: 1.17-1.61) and indifferent to adjustment for CRP levels. Hazard ratio of lung cancer for a doubling of CRP levels was 1.35 (1.17-1.56) and indifferent to adjustment for YKL-40 levels. Compared to individuals with both low CRP (<1.7 mg l(-1)) and YKL-40 (<154 µg l(-1)), individuals with high YKL-40 but low CRP had an HR of gastrointestinal cancer of 3.36 (1.70-6.64), whereas individuals with high CRP but low YKL-40 had an HR of lung cancer of 2.19 (1.24-3.87). The area under the receiver operating characteristic (ROC) curve was 0.68 for the ability of YKL-40 to predict gastrointestinal cancer and 0.67 for the ability of CRP to predict lung cancer. CONCLUSION: Elevated YKL-40 levels are associated with increased risk of gastrointestinal cancer, independently of CRP levels, whereas elevated CRP levels are associated with increased risk of lung cancer, independently of YKL-40 levels.