Atherosclerosis 220 (2012) 22–33
Contents lists available at ScienceDirect
Atherosclerosis journal homepage: www.elsevier.com/locate/atherosclerosis
Review
Postprandial lipemia and cardiovascular disease risk: Interrelationships between dietary, physiological and genetic determinants Kim G. Jackson a , Sally D. Poppitt b , Anne M. Minihane c,∗ a
Hugh Sinclair Unit of Human Nutrition, Department of Food and Nutritional Sciences, University of Reading, UK Human Nutrition Unit, School of Biological Sciences & Department of Medicine, University of Auckland, New Zealand c Department of Nutrition, Norwich Medical School, University of East Anglia, UK b
a r t i c l e
i n f o
Article history: Received 10 February 2011 Received in revised form 11 July 2011 Accepted 8 August 2011 Available online 9 September 2011 Keywords: Postprandial lipemia Non-fasting triglycerides Cardiovascular Genotype Fat composition Inflammation Vascular function Hemostasis
a b s t r a c t Although the independence of the association and causality has not been fully established, non-fasting (postprandial) triglyceride (TG) concentrations have emerged as a clinically significant cardiovascular disease (CVD) risk factor. In the current review, findings from three insightful prospective studies in the area, namely the Women’s Health Study, the Copenhagen City Heart Study and the Norwegian Counties Study, are discussed. An overview is provided as to the likely etiological basis for the association between postprandial TG and CVD, with a focus on both lipid and non-lipid (inflammation, hemostasis and vascular function) risk factors. The impact of various lifestyle and physiological determinants are considered, in particular genetic variation and meal fat composition. Furthermore, although data is limited some information is provided as to the relative and interactive impact of a number of modulators of lipemia. It is evident that relative to age, gender and body mass index (known modulators of postprandial lipemia), the contribution of identified gene variants to the heterogeneity observed in the postprandial response is likely to be relatively small. Finally, we highlight the need for the development of a standardised ‘fat tolerance test’ for use in clinical trials, to allow the integration and comparison of data from individual studies. © 2011 Elsevier Ireland Ltd. All rights reserved.
Contents 1. 2. 3. 4.
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Fasting triglycerides, postprandial lipemia and cardiovascular disease risk . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The independence of triglycerides as a cardiovascular disease risk factor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Physiological mechanisms underlying the impact of postprandial lipemia on cardiovascular disease risk . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1. Triglyceride-rich lipoprotein remnant infiltration into the arterial wall . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2. HDL and LDL concentration and subclass profile . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3. Inflammatory and oxidative status . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.4. Hemostasis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.5. Vascular function and reactivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5. Acute impact of dietary fat amount and composition on the postprandial lipemic response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6. Regulation of postprandial lipemia by genetic factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7. Relative impact of common gene variants on postprandial lipemia compared to established regulators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8. Regulation of postprandial inflammation, hemostasis, and vascular function by physiological factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.1. Inflammation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.2. Hemostasis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9. Vascular function and reactivity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10. Closing remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Author’s contribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conflict of interest . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
∗ Corresponding author. Tel.: +44 1603 593746; fax: +44 1603 593752. E-mail address:
[email protected] (A.M. Minihane). 0021-9150/$ – see front matter © 2011 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.atherosclerosis.2011.08.012
23 23 23 23 25 26 27 27 27 27 28 29 29 29 30 30 31 31 31 31
K.G. Jackson et al. / Atherosclerosis 220 (2012) 22–33
1. Introduction Due to the frequency of meal ingestion, individuals spend the majority of the day, approximately 18 h, in the fed (postprandial) state. The term given to the metabolic events that occur following the digestion and absorption of a meal that contains fat is postprandial lipemia. The magnitude and duration of the postprandial triglyceride (TG) response is influenced by a number of metabolic processes, including the rate of secretion of TG from the intestine and the liver, the activity of enzymes involved in the processing of TG-rich lipoproteins (TRLs; lipoprotein lipase (LPL) and hepatic lipase (HL)) and the rate of clearance of TRL remnants by receptormediated processes (Fig. 1). It is now well established that the postprandial lipemic response is influenced by both the amount and type of dietary fat given in a test meal [1]. However, the postprandial TG response to a standardised fat-rich meal has been shown to be highly variable between individuals [2] and there is considerable interest in understanding both physiological and genetic determinants of the lipemic response and the mechanisms that underpin its impact on the progression of cardiovascular disease (CVD). 2. Fasting triglycerides, postprandial lipemia and cardiovascular disease risk In 2007, Sarwar et al. [3] reported on the association between TG, predominately fasting, and risk of coronary heart disease (CHD), where original data from the EPIC-Norfolk (non-fasting TG) and Reykjavik studies were presented along with an updated metaanalysis which included 27 additional prospective studies. In the Reykjavik cohort, an age and gender adjusted odd ratios (OR) (95% confidence intervals) of 2.04 (1.78–2.32) was evident in those in the top vs. bottom tertile of fasting TG. An equivalent adjusted OR of 1.72 (1.56–1.90) was reported for the combined meta-analysis of all 29 studies. In a more recent output from The Emerging Risk Factor Collaboration (ERFC), which involved 300,000 individuals sourced from 68 separate long term prospective studies, a comparable age and gender adjusted hazard ratio (HR) for CHD was evident across the quartiles of TG concentration. However the association was lost (0.99, 0.94–1.05) following full adjustment of the model [4]. Differences in the strength of association observed between TG and CVD risk in the two-meta-analyses is likely to be attributable to the degree of correction for confounding factors, with full correction for high density lipoprotein-cholesterol (HDLC) (which is highly metabolically associated with TG, see below) in all studies included in the ERFC analysis likely to over-correct the model and underestimate the actual risk associated with circulating TG levels. Since Donald Zilversmit highlighted that atherosclerosis was a postprandial phenomenon over 30 years ago, numerous prospective case-control studies have qualitatively established postprandial TG as a risk factor for CVD (as reviewed in [1]). Three large prospective studies, namely the Women’s Health Study (WHS) [5], the Norwegian Counties Study (NCS) [6] and the Copenhagen City Heart Study (CCHS) [7] have confirmed the association and provided quantitative information about the relationship. In the WHS, unadjusted HR for incidence of CVD of 0.90 (0.47–1.72), 1.78 (1.02–3.10), 1.72 (0.99–2.98) and 2.81 (1.68–4.73) were reported across the quintiles (Q) of non-fasting TG levels (Table 1) [5]. In the NCS, comparable HRs for CVD deaths of 1.61 (1.28–2.03), 1.69 (1.35–2.12), 1.95 (1.36–2.43) and 3.27 (2.66–4.03) were observed in women with corresponding values of 1.04 (0.98–1.33), 1.17 (1.05–1.31), 1.39 (1.25–1.55) and 1.78 (1.61–1.98) in men (Table 1) [6]. This observation of an approximate 2-fold greater impact of non-fasting TG on cardiovascular events in women, also evident in the CCHS [7], is consistent with the HR
23
associated with fasting TG in women vs. men [8] (although the above mentioned meta-analysis of 29 prospective studies considering fasting TG failed to show this [3]). At present, the etiological basis of the differential effect of gender on the cardiovascular impact of non-fasting TG concentrations is unknown, as these differences remain following correction for all established CVD risk factors. Although non-fasting TG levels are in general considered more discriminatory with respect to CHD risk relative to fasting levels (for more extensive review see [9]), the issue remains somewhat controversial. In the WHS, although the minimally adjusted HRs were comparable, only non-fasting TG remained significantly associated with incidence CVD with HRs of 1.09 (0.85–1.41) and 1.98 (1.21–3.25) in tertile 3 of fasting and non-fasting TG, respectively [5]. In contrast, comparing the EPIC Reykjavik and Norfolk cohorts suggests no important differences in the predictive value of nonfasting (Norfolk) vs. fasting (Reykjavik) TG in the group as a whole, although in agreement with the WHS, non-fasting TG were more discriminatory in women [3]. 3. The independence of triglycerides as a cardiovascular disease risk factor Circulating TG is metabolically intimately linked with HDL; with elevated TG levels resulting in increased HL-mediated HDL hydrolysis and decreased HDL-C concentration. There is much ongoing debate regarding the independence of TG as a CVD risk indicator and its causal relevance, with many considering the impact of TG on CVD risk to be largely mediated through HDL-C. Consideration of the output from the three large prospective studies WHS, NCS and CCHS, provides insight into this discussion. In the WHS, a significant impact of non-fasting TG was evident even after full adjustment of the model, with a HR of cardiovascular events of 1.99 (1.05–3.78) in Q5 (Table 1) [5]. Although HDL-C was not included as a covariate in the primary analysis of the NCS and CCHS studies, a secondary analysis was conducted in which no significant impact of non-fasting TG remained in the NCS in either gender following adjustment for HDL-C [6] (Table 2). The output from the CCHS is somewhat more difficult to interpret as no fully adjusted model (including HDL-C and all other risk indicators) was presented. However, adjustment for age plus HDL-C only, had little impact on the HR values. Therefore, although the independence of non-fasting TG remains somewhat equivocal, some independent association is evident, in particular in women, with an attenuation rather than complete loss of risk following adjustment for HDL-C and other established CVD biomarkers. In principle, randomised controlled trials (RCTs) with TG-lowering agents such as fibrates, nicotinic acid or statins should be able to resolve the issue of causality of TG in CVD. However, in practice, this is not possible as such interventions affect other major lipids such as low density lipoprotein-cholesterol (LDL-C) and HDL-C. Although from a risk prediction perspective, the independence of non-fasting TG as a CVD risk marker is important, from a pathophysiological argument it is not. Non-fasting TG is a clinically significant CVD risk factor, which is influenced by, and influences (as will be reviewed in subsequent sections) many elements of the CVD risk phenotype, and therefore represents an important therapeutic target. 4. Physiological mechanisms underlying the impact of postprandial lipemia on cardiovascular disease risk Although, as discussed, a direct relationship between postprandial TG and CVD risk has now been established, the mechanisms by which TRLs exert their effect on the vascular wall are poorly understood. The sections below detail potential mechanisms of both lipid and non-lipid origin. Consensus is yet to be reached as to the effects
24
K.G. Jackson et al. / Atherosclerosis 220 (2012) 22–33
Table 1 Hazard ratios (95% confidence intervals) for cardiovascular events associated with non-fasting triglyceride levels. Quintiles of non-fasting TG Q1
Q2
Women’s Health Study [5] (WHS) (median follow up 11.4 years) TG, mg/dl (mmol/l) ≤85 (0.96) 86–113 (0.97–1.27) No. of participants 1273 1233 g 18 20 No. of CVD events a 1 0.90 (0.47–1.72) Model 1 1 0.83 (0.43–1.61) Model 2b 1 0.89 (0.45–1.75) Model 3c Norwegian Counties Study [6] (NCS) (mean follow up 27 years) Women TG, mean (range), mmol/l 0.66 (0.21–0.79) 0.90 (0.80–1.01) No. of participants 8585 8303 CVD deaths No. of events 110 204 1 1.61 (1.28–2.03) Model 4d Model 5e 1 1.37 (1.09–1.73) IHD deaths No. of events 47 92 Model 4d 1 1.70 (1.20–2.42) Model 5e 1 1.40 (0.99–2.00) Mene TG, mean (range), mmol/l No. of participants CVD deaths No. of events Model 4d Model 5e IHD deaths No. of events Model 4d Model 5e
P for trend Q3
Q4
Q5
114–154 (1.28–1.73) 1320 43 1.78 (1.02–3.10) 1.57 (0.89–2.78) 1.63 (0.90–2.96)
155–214 (1.74–2.40) 1273 47 1.72 (0.99–2.98) 1.41 (0.79–2.55) 1.60 (0.87–2.95)
≥215 (≥2.41) 1292 87 2.81 (1.68–4.73) 2.09 (1.13–3.86) 1.99 (1.05–3.78)
1.13 (1.02–1.27) 8576
1.46 (1.28–1.70) 8654
2.16 (1.71–18.72) 8482
233 1.69 (1.35–2.12) 1.28 (1.02–1.61)
280 1.95 (1.36–2.43) 1.34 (1.07–1.68)
469 3.27 (2.66–4.03) 2.02 (1.45–2.82)
