Remnant Cholesterol, Low-Density Lipoprotein Cholesterol, and Blood Pressure as Mediators From Obesity to Ischemic Heart DiseaseNovelty and Significance
Rationale: Obesity leads to increased ischemic heart disease (IHD) risk, but the risk is thought to be mediated through intermediate variables and may not be caused by increased weight per se.
Objective: To test the hypothesis that the increased IHD risk because of obesity is mediated through lipoproteins, blood pressure, glucose, and C-reactive protein.
Methods and Results: Approximately 90 000 participants from Copenhagen were included in a Mendelian randomization design with mediation analyses. Associations were examined using conventional measurements of body mass index and intermediate variables and using genetic variants associated with these. During ≤22 years of follow-up 13 945 participants developed IHD. The increased IHD risk caused by obesity was partly mediated through elevated levels of nonfasting remnant cholesterol and low-density lipoprotein cholesterol, through elevated blood pressure, and possibly also through elevated nonfasting glucose levels; however, reduced high-density lipoprotein cholesterol and elevated C-reactive protein levels were not mediators in genetic analyses. The 3 intermediate variables that explained the highest excess risk of IHD from genetically determined obesity were low-density lipoprotein cholesterol with 8%, systolic blood pressure with 7%, and remnant cholesterol with 7% excess risk of IHD. Corresponding observational excess risks using conventional body mass index were 21%, 11%, and 20%, respectively.
Conclusions: The increased IHD risk because of obesity was partly mediated through elevated levels of nonfasting remnant and low-density lipoprotein cholesterol and through elevated blood pressure. Our results suggest that there may be benefit to gain by reducing levels of these risk factors in obese individuals not able to achieve sustained weight loss.
- cardiovascular diseases
- Mendelian randomization analysis
- myocardial ischemia
- risk factors
It has been shown previously that obesity is associated with increased risk of ischemic heart disease (IHD),1–3 and that the association between obesity and IHD is likely to be causal4; however, because it is probably not increased weight per se that leads to the increased risk, it is likely that this risk is mediated through variables such as lipoprotein levels, blood pressure, glucose levels, and inflammation. Obesity is an increasing problem for public health in many countries, and although elevated body mass index (BMI) is a modifiable risk factor, it is difficult to encourage individuals to lose weight and even more difficult to maintain weight loss. Therefore, it is important to delineate which of the intermediate variables from obesity to IHD are likely to be mediators.
Editorial, see p 570
In this study, we use Mendelian randomization in the analysis of data from participants from the Copenhagen General Population Study (CGPS; n=69 535), the Copenhagen City Heart Study (n=10 099), and the Copenhagen Ischemic Heart Disease Study (n=5050) to test the hypothesis that the increased risk of IHD because of obesity is mediated through (1) elevated nonfasting remnant cholesterol, (2) reduced high-density lipoprotein (HDL) cholesterol, (3) elevated low-density lipoprotein (LDL) cholesterol, (4) elevated systolic blood pressure, (5) elevated diastolic blood pressure, (6) elevated glucose levels, and (7) elevated C-reactive protein (CRP). In Mendelian randomization, the random assortment of alleles at conception is used to circumvent confounding and reverse causation that can bias observational associations. Assumptions for Mendelian randomization are first, that the genetic variants are associated with the phenotype; second, that the association between the genetic variants and the outcome is not confounded by the same factors that confound the observational association; third, that the genetic variants are not in linkage disequilibrium with other genetic variants associated with the outcome; and finally, that the genetic variants are not associated with the outcome through other ways than through the phenotype, that is, there is no pleiotropy.5
To assess potential benefits if these intermediate variables are targeted in obese individuals unable to obtain sustained weight loss, we also performed mediation analyses to assess how much excess risk of IHD from obesity each intermediate variable mediates. Besides conventional mediation analyses using observational data, we used the novel approach of genetic mediation analyses to quantify how much of the risk of IHD, because of genetically determined obesity, is mediated through the different intermediate variables as this is presently not known. Each association was first examined in observational analyses, that is, analyses prone to confounding and reverse causation, and second in genetic analyses not prone to the same confounding and reverse causation as the observational studies.
Studies were approved by the Herlev Hospital and Danish ethical committees (H-KF-01-144/01; KF-100.2039/91; KF-01-144/01; KA-93125; and KA-99039) and were conducted according to the Declaration of Helsinki with informed consent from participants. All participants were white and of Danish descent, and none appeared in >1 study.
The CGPS is a prospective study of the general population initiated in 2003 with ongoing enrollment.4 Participants were randomly selected from the National Danish Civil Registration System to reflect the adult population aged 20 to ≥100 years. Data collection included a questionnaire, a physical examination, and blood sampling for biochemical analyses and DNA extraction.
The Copenhagen City Heart Study
The Copenhagen City Heart Study is a prospective study of the general population initiated in 1976 to 1978 with follow-up examinations in 1981 to 1983, 1991 to 1994, and 2001 to 2003.6 Participants were recruited and examined exactly as in the CGPS. Blood samples for biochemical measurements and DNA extraction were drawn at the 1991 to 1994 and 2001 to 2003 examinations.
The Copenhagen Ischemic Heart Disease Study
This study comprises 5050 patients referred for coronary angiography to Rigshospitalet, Copenhagen University Hospital during the period 1991 to 2011. Beside a diagnosis of IHD as described below, patients also had stenosis/atherosclerosis on coronary angiography and a positive exercise electrocardiography test.
Ischemic Heart Disease
Information on a diagnosis of IHD (International Classification of Diseases [ICD]-8: 410–414, ICD-10: I20–I25) was collected and verified from 1977 until April 2013 by reviewing all hospital admissions and diagnoses entered in the National Danish Patient Registry and all causes of death entered in the National Danish Causes of Death Registry, as described.6 These registers contain information on all participants and none were lost to follow-up. During ≤22 years of follow-up 13 945 participants developed IHD.
Standard hospital assays measured nonfasting total cholesterol, triglycerides, HDL cholesterol, high-sensitivity CRP, and glucose (Boehringer Mannheim, Dako, Dade Behring, and Konelab). LDL cholesterol was calculated using the Friedewald equation when plasma triglycerides were ≤4 mmol/L, and otherwise measured directly (Konelab). Nonfasting remnant cholesterol was nonfasting total cholesterol minus HDL cholesterol minus LDL cholesterol.6–8
Using information from our previous studies and from genome-wide association studies, we chose well-known genetic variants for each allele score so that the allele score was associated with BMI or the particular intermediate variable of interest, and ideally with nothing else: (1) for BMI, we chose the 5 genetic variants, FTO rs9939609, MC4R rs17782313, TMEM18 rs6548238, BDNF rs10767664, and GNPDA2 rs10938397, with the largest reported effects on BMI.9; (2) for nonfasting remnant cholesterol, we chose TRIB1 rs2954029, GCKR rs1260326, and APOA5 rs6518217,10; (3) for HDL cholesterol, we chose LIPC −480C/T, ABCA1 N1800H, and ABCA1 R2144X7,11,12; (4) for LDL cholesterol, we chose APOB rs5742904, LDLR W23X, LDLR W66G, LDLR W556S, and PCSK9 rs115911477,13,14; (5) for systolic and diastolic blood pressures, we chose ATP2B1 rs2681472 and CYP17A1 rs1119154815; (6) for glucose, we chose GCK rs4607517, G6PC2 rs560887, ADCY5 rs11708067, DGKB rs2191349, and ADRA2A rs1088512216; and (7) for CRP, we chose CRP rs1205, CRP rs1130864, CRP rs3091244, and CRP rs3093077.17
Genotyping was by TaqMan (Applied Biosystems) or by restriction enzyme assays (details available in the Online Data Supplement). Genotypes were verified by genotyping of randomly selected samples of each variant by 2 different methods (TaqMan plus sequencing or restriction enzyme assay). Call rates for genotypes were >99% for all assays. ATP2B1 rs2681472 and CYP17A1 rs11191548 were determined by LGC Genomics using a competitive allele-specific polymerase chain reaction system (KASPar).
BMI was measured as weight (kg) divided by measured height squared (m2). Systolic and diastolic blood pressures were measured at study inclusion. Smokers were current smokers. Lipid-lowering therapy was self-reported.
We conducted analyses according to the order shown in Figure 1. We first analyzed the observational association between BMI and the 7 intermediate variables (Figure 1, No. 1). Second, we investigated how a BMI allele score associated causally with levels of the intermediate variables (Figure 1, No. 2). Third, we combined information from Nos. 1 and 2 in instrumental variable analyses, to examine whether observational associations between BMI and intermediate variables are causal, and how large a difference in levels of intermediate variables an observational and a causal 10 kg/m2 higher BMI was associated with, corresponding to the difference between being normal weight and obese (Figure 1, No. 3). Fourth, using the genetic variants only associated with BMI as well as other well-known genetic variants associated only with individual intermediate variables, we investigated how much the observational and causal change in the intermediate variable (caused by a 10 kg/m2 higher BMI) would translate into in observational and causal risk of IHD (Figure 1, No. 4a and b). Finally and fifth, we combined all information from Nos. 1 to 4 and previous knowledge from randomized clinical intervention trials to conduct observational and causal mediation analyses to examine how much excess risk of IHD caused by obesity is mediated through the 7 intermediate variables (Figure 1, No. 5a and b).
For a detailed description of the statistical analyses performed, please see the Online Data Supplement.
The Table shows characteristics of 84 684 participants in the Copenhagen City Heart Study, CGPS, and CHIDS (Copenhagen Ischemic Heart Disease Study); however, the number of participants varies between the different analyses according to the availability of genotypes, intermediate variables, and covariates. Online Table II shows baseline characteristics of participants in the CGPS by genotypes and shows that the allele scores used as instruments for the different intermediate variables were not associated with other conventional risk factors for IHD, except for the expected associations of the remnant cholesterol allele score and the LDL cholesterol allele score with lipid-lowering therapy, that is, the allele scores were largely without pleiotropic effects on other cardiovascular risk factors. Genotype distributions for all studies were in Hardy–Weinberg equilibrium (P values >0.1).
Observational and Genetic Associations: BMI and IHD
As described previously using the same studies,4 a 4 kg/m2 higher BMI was associated with increased risk of IHD with an observational odds ratio of 1.26 (95% CI, 1.19–1.34) and a genetically derived odds ratio of 1.52 (1.12–2.05), corresponding to an observational odds ratio of 1.78 (1.54–2.08) and a genetically derived odds ratio of 2.85 (1.33–6.02) for a 10 kg/m2 higher BMI. Also corresponding to the previous findings, BMI in quintiles was associated with a stepwise increased risk of IHD (Online Figure I, left), and the BMI allele score was associated with stepwise increased risk of IHD for increasing number of alleles (Online Figure I, right; Figure 1, Pa and Pb).
Observational Associations: BMI and Intermediate Variables
Increasing BMI levels were observationally associated with higher levels of nonfasting remnant cholesterol, LDL cholesterol, glucose, and CRP, with higher systolic and diastolic blood pressures, and with lower levels of HDL cholesterol (Figure 1, No. 1; Figure 2).
Genetic Associations: BMI Allele Score and Intermediate Variables
The BMI allele score was associated with higher levels of nonfasting remnant cholesterol, and CRP, with higher systolic and diastolic blood pressures, and with lower HDL cholesterol levels for increasing number of BMI increasing alleles (Figure 3). The BMI allele score did not show strong evidence for an association with LDL cholesterol or nonfasting glucose levels; however, there was a tendency toward higher levels for increasing number of BMI alleles. Online Figure II shows similar results as Figure 3 for the 5 genetic variants used in the BMI allele score separately (Figure 1, No. 2).
Combined Observational and Genetic Associations: BMI and Intermediate Variables
Figure 4 shows that both an observational and a genetically determined 10 kg/m2 higher BMI was associated with higher levels of remnant cholesterol, LDL cholesterol, glucose, and CRP, with higher systolic and diastolic blood pressures, and with lower HDL cholesterol levels. Although P values for comparison showed significant differences between observational and genetic estimates for remnant cholesterol, HDL cholesterol, and diastolic blood pressure, all genetic estimates were in the same direction as observational estimates indicating causality (Figure 1, No. 3).
Online Figure III shows results for individuals with a BMI <30 kg/m2 and Online Figure IV results for those with a BMI >30 kg/m2.
Observational and Genetically Determined Changes in Intermediate Variables, Associated With a 10 kg/m2 Higher BMI, Translated Into Risk of IHD
Figure 5 shows observational and genetic risk estimates for IHD for levels of the intermediate variables corresponding to the magnitude of the observational and genetic associations of a 10 kg/m2 higher BMI with the intermediate variables shown in Figure 4, that is, Figure 5 shows what the difference in intermediate variable associated with an observational or a genetically determined 10 kg/m2 higher BMI translated into in risk of IHD. For nonfasting remnant cholesterol and LDL cholesterol, both observational and genetic estimates for IHD risk were in the same direction indicating that the associations may be causal, with observational hazard ratios of 1.10 (1.08–1.11) and 1.05 (1.04–1.06) for remnant cholesterol and LDL cholesterol, respectively, and with corresponding genetically derived risk ratios of 1.20 (1.07–1.34) and1.28 (1.20–1.36), respectively. Genetic estimates were higher than observational estimates, which can probably be explained by genetic variants causing lifelong higher levels of remnant and LDL cholesterol, whereas observational estimates are based on a single measurement of remnant and LDL cholesterol (Figure 1, No. 4).
For HDL cholesterol levels, systolic and diastolic blood pressures, and for glucose levels, the observational estimates indicate increased IHD risk; however, the genetically derived risk ratios were not statistically different from 1 (Figure 5), indicating that the association may not be causal, and that the observational estimate may be confounded. For elevated CRP levels there was an increased observational risk of IHD, but the genetically derived risk ratio was in the opposite direction.
The allele scores for the intermediate variables were combined from genetic variants known to be associated with the intermediate variable of interest so that they had as large as possible an effect and with the effect mainly on the intermediate variable of interest and not on the other intermediate variables (Online Figure V).
All intermediate variables were associated observationally with risk of IHD (Online Figure VI); however, only allele scores for remnant cholesterol, LDL cholesterol, and glucose were associated with risk of IHD (Online Figure VII).
Observational and Genetic Mediation Analyses
For all intermediate variables, the observational excess risk (Figure 6, left) was higher than the genetically determined excess risk (Figure 6, right). P values for comparison showed significant differences between observational and genetic estimates for remnant cholesterol, LDL cholesterol, and diastolic blood pressure; however, estimates were in the same direction, and the differences between observational and genetic estimates could be explained by the genetic variants explaining only part of the variation in the phenotypes and the observational estimates being prone to confounding. Genetically determined excess risk is only shown for the intermediate variables with results suggesting genetically determined risk for IHD in Figure 5, except for blood pressure and glucose where previous randomized clinical intervention trials or Mendelian randomization studies suggest causality,16,18,19 and where the lack of genetic association in Figure 5 may be explained by lack of power because of weak instruments. The 3 intermediate variables that explained the highest excess risk of IHD from genetically determined obesity were LDL cholesterol with 8%, systolic blood pressure with 7%, and remnant cholesterol with 7% excess risk of IHD, with corresponding observational excess risks of 21%, 11%, and 20%, respectively. Results were similar when using the percent excess risk mediated method,20 except that the excess risk for LDL cholesterol was only 1% (Online Figure VIII; Figure 1, No. 5).
Online Table III shows P values for interaction between the intermediate variables and BMI or BMI allele score on predicting risk of IHD. After correction for multiple comparisons, there was a significant interaction between systolic blood pressure and BMI in the observational association.
In this study, we show that the previously established increased risk of IHD because of obesity4 is partly mediated through elevated levels of nonfasting remnant cholesterol and LDL cholesterol and through elevated blood pressure and possibly also through elevated nonfasting glucose levels; however, HDL cholesterol and CRP levels did not seem to be causal mediators.
We and others have examined previously the association between BMI and IHD and the association between BMI and some of the intermediate variables4,21–23; however, we now use the novel approach of genetic mediation analyses to quantify how much of the risk of IHD, because of genetically determined obesity, is mediated through the different intermediate variables. Our data contribute important information because IHD is a growing problem worldwide, partly because of increased prevalence of obesity. BMI is a modifiable risk factor, but it can be difficult to obtain sustained weight loss, and thus an alternative approach could be risk reduction of IHD by modifying the intermediate risk factors. Results from this study, integrating results from randomized intervention trials of blood pressure lowering,18 indicate that there may be a potential for reducing IHD risk by reducing levels of remnant cholesterol, LDL cholesterol, and blood pressure in obese individuals unable to achieve sustained weight loss. Thereby, the findings from this study reinforce existing recommendations on lowering of LDL cholesterol and blood pressure and stress the need for large clinical intervention trials examining whether a lowering of remnant cholesterol in individuals with elevated levels will reduce IHD risk.
Genetically elevated levels of nonfasting remnant and LDL cholesterol and genetically elevated blood pressure explained only 27% of the excess IHD risk from obesity in our study, and most of the risk remained unexplained. This could be because both the lipoprotein metabolism and the blood pressure control are complicated processes influenced by many different pathways as well as by the environment, and although we carefully chose genetic variants associated with the phenotypes, they only explained a small fraction of the overall phenotypes. The observational estimates for levels of nonfasting remnant and LDL cholesterol and elevated blood pressure explained a larger proportion of the mediated risk with 63% excess IHD risk; however, these estimates can be confounded by other cardiovascular risk factors correlated with the phenotypes and influenced by reverse causation.
Mechanistically, the explanation for elevated nonfasting remnant cholesterol and LDL cholesterol causing IHD most likely is that the lipoproteins enter and get trapped in the arterial wall,24–26 followed by cholesterol accumulation and development of atherosclerosis. Remnant cholesterol is the cholesterol content of the triglyceride-rich lipoproteins composed of very-LDLs and some intermediate-density lipoproteins in the fasting state and of these 2 lipoproteins together with chylomicron remnants in the nonfasting state. We have previously found elevated levels of remnant cholesterol to be associated with elevated levels of CRP in both observational and genetic analyses.8 This is probably because remnant cholesterol causes low-grade inflammation in the arterial wall after accumulation in the arterial intima and elsewhere in the body.
Triglyceride and remnant cholesterol levels are highly correlated.7 In this study, we calculate remnant cholesterol as nonfasting total cholesterol minus LDL and HDL cholesterol. In participants with triglyceride levels <4 mmol/L, LDL cholesterol was calculated from the Friedewald equation and calculated remnant cholesterol is therefore a rescaling of triglycerides; however, in participants with triglyceride levels >4 mmol/L, LDL cholesterol was directly measured and the correlation between triglycerides and remnant cholesterol for these participants cannot be explained by this and is probably explained by the composition of remnant; that is, that triglycerides and remnant cholesterol are part of the same molecules, that is, remnants. However, because most cells can degrade triglycerides, and no cells can degrade cholesterol, it is plausible that it is the cholesterol content of remnants that causes atherosclerosis and IHD development. In the fasting state where there is little cholesterol in chylomicrons, the Friedewald equation estimates LDL cholesterol as total cholesterol minus HDL and estimated very-LDL cholesterol from triglycerides; however, participants in our study were nonfasting, which means that our estimation of remnant cholesterol as total cholesterol minus LDL and HDL cholesterol also includes cholesterol in chylomicron remnants and also the part of intermediate-density lipoprotein cholesterol that is not included in the LDL fraction.
We found no genetic association of elevated systolic and diastolic blood pressure with increased risk of IHD; however, because randomized clinical trials have consistently shown that lowering of blood pressure is associated with reduced risk of IHD,18 lack of genetic association in this study was probably explained by weak instruments. We chose 2 top-hits from genome-wide association studies,15 but they only explained a small fraction of the variation in blood pressure.
A potential limitation to our study is that results were reported for an increase in BMI of 10 kg/m2 as this corresponds to going from being normal weight to being obese. Another increase in BMI could have been chosen instead; however, this would not have changed the overall interpretation of the results, but only the scaling of results. Another potential limitation in our study is the adjustment for lipid-lowering therapy and antihypertensive therapy as categorical covariates. A more detailed adjustment for different types/doses of lipid-lowering therapy and antihypertensive therapy might have made our estimates more accurate. Also, repeated measures of the phenotypes could have made our observational estimates more accurate, and this is why we also use genetic estimates that are presumably more accurate, because genetic variants are associated with lifelong altered levels of the phenotypes.
An assumption for mediation analysis is lack of interaction between exposure and mediators on the end point. We found an interaction between systolic blood pressure and BMI in the observational association, and estimates from the observational mediation analysis for systolic blood pressure should therefore be interpreted with some caution.
Mendelian randomization studies are a way of circumventing confounding and reverse causation seen in observational epidemiology; however, some limitations apply such as canalization, population stratification, linkage disequilibrium, and pleiotropy.27 The most important for this study is pleiotropy, which is when the genetic variant used as an instrument for a given risk factor is associated with yet other risk factors. To minimize influence from pleiotropy, we carefully selected genetic instruments and used several genetic variants combined in allele scores for BMI and for the intermediate variables.28 However, because of the lipoprotein metabolism with many complex interactions, it is difficult to find genetic variants associated with a specific lipoprotein only. By combining several genetic variants for each type of lipoprotein cholesterol, we have minimized pleitropy on other types of lipoprotein cholesterol, but not eliminated it completely; however, these pleitropic effects on other types of lipoprotein cholesterol were only small and can probably not explain our results. To avoid population stratification, we only studied white individuals of Danish descent, which potentially could affect the generalizability of our results; however, we are not aware of data, suggesting that our results should not apply to most races and countries where obesity is prevalent.
In conclusion, we found that the increased risk of IHD because of genetically determined obesity was partly mediated through elevated levels of nonfasting remnant cholesterol and LDL cholesterol and through elevated blood pressure. This indicates that there may be benefit to gain by reducing levels of these risk factors in obese individuals not able to achieve sustained weight loss. Thereby, the findings from this study reinforce existing recommendations on lowering of LDL cholesterol and blood pressure and stress the need for large clinical intervention trials examining if lowering of remnant cholesterol in individuals with elevated levels will reduce IHD risk.
We are indebted to staff and participants of the Copenhagen General Population Study, Copenhagen City Heart Study, and Copenhagen Ischemic Heart Disease Study for their important contributions.
Sources of Funding
This study was supported by The Danish Council for Independent Research, Medical Sciences (FSS); The Danish Heart Foundation; Herlev Hospital, Copenhagen University Hospital; Copenhagen County Foundation; and Chief Physician Johan Boserup and Lise Boserup’s Fund, Denmark. G.D. Smith and N.J. Timpson work within a unit supported by the MRC and the University of Bristol (MC_UU_12013/1, MC_UU_12013/3). The funding sources of the study had no role in the study design, collection, analysis, or interpretation of data and had no role in writing of the article or the decision to submit the article for publication.
B.G. Nordestgaard has received lecture and consultancy honoraria from Omthera, Sanofi-Aventis/Regeneron, Aegerion, AstraZeneca, Merck, Pfizer, Fresenius, and ISIS Pharmaceuticals. G.D. Smith has received lecture/consultancy honoraria from Merck. The other authors report no conflicts.
In October, 2014, the average time from submission to first decision for all original research papers submitted to Circulation Research was 16 days.
The online-only Data Supplement is available with this article at http://circres.ahajournals.org/lookup/suppl/doi:10.1161/CIRCRESAHA.116.304846/-/DC1.
- Nonstandard Abbreviations and Acronyms
- body mass index
- Copenhagen General Population Study
- C-reactive protein
- high-density lipoprotein
- ischemic heart disease
- low-density lipoprotein
- Received July 18, 2014.
- Revision received November 13, 2014.
- Accepted November 19, 2014.
- © 2014 American Heart Association, Inc.
- Ni Mhurchu C,
- Rodgers A,
- Pan WH,
- Gu DF,
- Woodward M
- Nordestgaard BG,
- Palmer TM,
- Benn M,
- Zacho J,
- Tybjaerg-Hansen A,
- Davey Smith G,
- Timpson NJ
- Varbo A,
- Benn M,
- Tybjærg-Hansen A,
- Nordestgaard BG
- Jørgensen AB,
- Frikke-Schmidt R,
- West AS,
- Grande P,
- Nordestgaard BG,
- Tybjærg-Hansen A
- Frikke-Schmidt R,
- Nordestgaard BG,
- Stene MC,
- Sethi AA,
- Remaley AT,
- Schnohr P,
- Grande P,
- Tybjaerg-Hansen A
- Whelton PK,
- He J,
- Appel LJ,
- Cutler JA,
- Havas S,
- Kotchen TA,
- Roccella EJ,
- Stout R,
- Vallbona C,
- Winston MC,
- Karimbakas J
- Ray KK,
- Seshasai SR,
- Wijesuriya S,
- Sivakumaran R,
- Nethercott S,
- Preiss D,
- Erqou S,
- Sattar N
- Shaikh M,
- Wootton R,
- Nordestgaard BG,
- Baskerville P,
- Lumley JS,
- La Ville AE,
- Quiney J,
- Lewis B
- Nordestgaard BG,
- Tybjaerg-Hansen A,
- Lewis B
- Nordestgaard BG,
- Wootton R,
- Lewis B
- Davey Smith G,
- Ebrahim S
Novelty and Significance
What Is Known?
Obesity is an increasing problem in many countries, and it is difficult to encourage individuals to lose weight and even more difficult to maintain weight loss.
Obesity leads to increased risk of ischemic heart disease (IHD).
However, it may not be obesity per se that causes IHD, but rather changes in mediating factors such as levels of lipoproteins, blood pressure, glucose levels, and low-grade inflammation caused by obesity.
What New Information Does This Article Contribute?
In this study, we combine observational and genetic evidence from ≈90 000 individuals from the Danish general population to delineate which mediating factors contribute causally to the increased risk of IHD from obesity.
Obesity is a growing problem and although elevated body mass index is a modifiable risk factor, it is challenging to achieve persistent weight loss in obese individuals. It is therefore important to delineate which intermediate factors from obesity to IHD that are causal to be treated with benefit. We used the novel approach of combining observational analyses, Mendelian randomization, and observational and genetic mediation analyses to circumvent confounding and reverse causation, both of which are sources of bias in conventional observational epidemiology. The top 3 intermediate factors in this study that mediated the increased risk of IHD from obesity were elevated levels of low-density lipoprotein cholesterol, elevated levels of remnant cholesterol, and elevated blood pressure. Current guidelines on IHD prevention include recommendations for treatment of elevated levels of low-density lipoprotein cholesterol and elevated blood pressure; however, there is no consensus on whether elevated levels of remnant cholesterol should be treated, mainly because of a lack of evidence from randomized clinical intervention trials. Results from this study, combined with previous studies, highlights the need for randomized clinical intervention trials aimed at reducing levels of remnant cholesterol in individuals with elevated levels.