سطح سرمی لیپیدها و واکنش پذیری قلبی عروقی به استرس

Abstract Several studies have reported an association between serum lipid levels and cardiovascular reactivity to laboratory stressors. Their findings, however, are equivocal. The inconsistencies may be due to shortcomings such as the small number of subjects, the inclusion of patient groups, no control for medication, and no control for age effects. Two studies are presented investigating the relationship in large groups of adolescent and middle-aged males and females. Cholesterol, triglycerides and HDL were measured. Subjects were exposed to mental stressors, and in one study also to a cold pressor test. In addition to heart rate and blood pressure, in one study impedance cardiography was used to measure pre-ejection period, stroke volume and total peripheral resistance. Canonical correlation analysis suggested an association between triglycerides and decreased cardiac reactivity to mental stressors in middle-aged females. Trends in the same direction were found in both middle-aged males and females with respect to reactivity to the cold pressor. These associations, however, were not confirmed when the extreme deciles of the triglyceride distributions were compared with respect to stress reactivity. The fact that associations were completely absent in youngsters but sometimes showed up in older persons suggested an age dependency of the association. In post hoc analyses, indeed, some evidence was found for stronger cardiac responsivity being associated with cholesterol specifically in relatively older males. In females, in contrast to this, the association between triglycerides and cardiac responsivity was stronger in the younger group. More detailed measurement techniques, of specifically vascular processes, may be needed to explore further the effects of sex and age on the association between lipids and stress reactivity.

Introduction The association between serum lipid levels and the risk of coronary heart disease (CHD) is well established. Lipids, including serum cholesterol and triglycerides, are transported in the circulation bound to (apo)proteins, the cluster together forming the lipoproteins. The different lipoproteins can be separated on the basis of their densities (lipid/protein ratios). The major classes of lipoproteins are very low-density lipoprotein (VLDL), mainly consisting of triglycerides, low-density lipoprotein (LDL), the main transporter of cholesterol, and high-density lipoprotein (HDL). HDL exhibits a reverse relationship with CHD risk. Serum lipid and lipoprotein levels have a substantial genetic/constitutional component (Heller et al., 1993 and Snieder et al., 1997) and are also influenced by behavioral factors like diet (Hopkins, 1992), exercise (Tran et al., 1983) and smoking (Freeman et al., 1993). Periods of episodic or chronic stress have shown to be associated with elevated cholesterol levels. This has, for example, been demonstrated for undergoing an earthquake (Trevisan et al., 1992), job loss (Kasl et al., 1968), threat of unemployment (Mattiasson et al., 1990), perceived job insecurity (Siegrist et al., 1988), perceived workload (McCann et al., 1990) or examinations (van Doornen and van Blokland, 1987). A rise in cholesterol level during stress is, however, not a universal finding (Niaura et al., 1991). The mechanism underlying a stress-induced cholesterol elevation is supposed to be the sympatho-adrenergic system, because adrenaline is a potent lipolytic agent. Exaggerated cardiovascular reactivity to stress has been proposed as a risk factor for future CHD (Matthews et al., 1986 and van Doornen, 1991). The propensity of individuals to be sympathetically reactive to stress is generally assessed by exposing subjects to short-term laboratory stressors. Large individual differences in cardiovascular reactivity are generally observed, which seem to reflect differences in adrenergic mobilisation (Eisenhofer et al., 1985). Because the sympatho-adrenergic system is both involved in lipolysis and in stress reactions several studies have explored the association between cardiovascular reactivity to stress and serum lipids. According to this idea lipid levels and stress reactivity are associated because of a common mechanism: sympatho-adrenal activity. There is yet another mechanism that might be responsible for a possible association. Hypercholesteremia has been shown to be associated with an augmented response to vasoconstrictors (Heistad et al., 1984) and impaired endothelial-dependent vascular relaxation (Casino et al., 1993). This points to a more peripheral vascular explanation for the association between stress reactivity and lipids. The idea of a common sympathetic mechanism would show up as a relationship between cardiac responsivity and lipids, the more peripheral explanation would predict lipids to be associated with a stronger rise in peripheral resistance in response to stress. To date, seven studies have investigated the relationship between lipid and lipoprotein levels and cardiovascular reactivity to laboratory stressors. The studies have been empirical in their approach in the sense that relationships between reactivity and lipids were studied without paying much attention to the physiological basis to expect such a relationship. Some refer briefly to the possible role of sympathetic activity. This might apply to an association of reactivity with cholesterol or triglyceride levels, but for HDL there is no clear physiological basis to expect an association with reactivity. We will first summarize the empirical evidence for the existence of a relationship, and in case associations are observed, whether they support a common sympathetic basis or a peripheral explanation. A schematic overview of the studies is presented in Table 1. Table 1. Overview of studies to stress reactivity and lipid levels Study (1) Subjects Tasks (1) Cardiovascular variables (1) Analysis Results (2) Age (2) Lipids (2) Age correction (3) Medication McKinney et al. (1987) (1) 49 m 1 f; hypertension+gastrointest. patients Video game; mental arithmetic; cold pressor (1) SBP, DBP, MAP, HR, TPR, SV (1) Correlative Video game: No associations (2) 46.5 (±10.4) (2) Cholesterol, HDL, triglycerides (2) No Mental arithmetic: Trigl×DBP r=+0.32 Trigl×MAP r=+0.35 HDL×DBP r=−0.46 HDL×TPR r=−0.35 (3) 48% Cold pressor: Chol×DBP r=+0.33 HDL×TPR r=−0.44 Jorgensen et al. (1988) (1) 59 m; mild hypertension Video game, Stroop (1) HR, SBP, DBP (1) Median split HR; response aggregated across tasks High HR reactors (2) 28–69, median 48 (2) Cholesterol, triglycerides (2) Median split age Higher cholesterol and triglycerides (3) No No age interaction Fredrikson and Blumenthal (1988) (1) 42 m MI patients Mental arithmetic (1) HR, SBP, DBP (1) Median split cholesterol/HDL No effects (2) 28–66, median 52 (2) Cholesterol, HDL, triglycerides Task levels (3) ±70%, β-block; ±50% calcium antag. (2) Yes Fredrikson et al. (1991) (1) 30 m 29 f Mirror drawing; mental arithmetic; stroop; cold pressor; handgrip; type A-interv. (1) HR, SBP, DBP (1) Median split lipids High chol>SBP resp. (2) 30–55, m=41.2 (2) Cholesterol, LDL, HDL Responses aggregated across tasks High LDL>HR resp. (3) No Suarez et al. (1991) (1) 24 m Mental arithmetic; word identification (1) HR, SBP, DBP, FBF (1) Correlative Mental arithmetic: Chol×DBP, r=−0.40 (2) 35–50 (2) Cholesterol, HDL, LDL, VLDL, triglycerides (2) No (3) No Owens et al. (1993) (1) 15 m 34 f Handgrip; mirror drawing; speech task (1) HR, SBP, DBP (1) Correlative Speech task: Chol×SBP, r=0.28 (2) 40–55 (2) Cholesterol, HDL, LDL, triglycerides (2) Yes Chol×DBP, r=0.28 (3) No Burker et al. (1994) (1) 62 m 37 f Mental arithmetic; public speaking; cold pressor; video game (1) SBP, DBP, HR (1) Median split SBP DBP, HR Total group: High SBP reactors, higher Chol, LDL, apo-B High DBP reactors, higher LDL, apo-B, lower HDL (2) 45–60 (m=44.5 (m); m=47 (f)) hypertensives (2) Cholesterol, LDL, HDL, VLDL, apo-AI, Apo-AII, Apo-B, Lp(a) Responses aggregated across tasks Women: High HR reactors, higher Apo-AI/A-II ratio (3) Discontinued (2) No Table options McKinney et al. (1987)measured cholesterol, HDL and triglycerides in 50 physicians and exposed them to two mental stressors and the cold pressor test. Cholesterol and triglycerides correlated positively with the diastolic blood pressure (DBP) and total peripheral resistance (TPR) response to the cold pressor and to one of the mental tasks. For HDL, the reverse pattern was observed. Lipid parameters were unrelated to the heart rate (HR) and systolic blood pressure (SBP) reactions to the stressors. The correlations with DBP and TPR suggest an association between vascular responsiveness and lipids. Jorgensen et al. (1988)showed in a group of 59 mild hypertensives that those above the median with respect to the average HR response to two mental stressors had significantly higher total cholesterol and triglyceride levels than those below the median of HR reactivity. Results for a median split on the basis of SBP or DBP reactivity were not mentioned and so probably were not significant. Fredrikson and Blumenthal (1988)found no association between total cholesterol/HDL ratio and HR and BP responses to a mental arithmetic task in a group of 42 infarction patients. Fredrikson et al. (1991)exposed 60 healthy subjects in their 40s to six stressors. Compared to low-cholesterol subjects, high-cholesterol subjects showed a stronger SBP response, but the same DBP response across the tasks. HDL was unrelated to reactivity. Suarez et al. (1991)observed a negative correlation (−0.40) between the DBP response to mental arithmetic and cholesterol level in 24 healthy males. The SBP and HR responses were unrelated to cholesterol. HDL, LDL, VLDL and triglycerides levels were unrelated to reactivity. No relations were observed with the reactivity to another task. Owens et al. (1993)observed positive correlations between cholesterol level and both systolic and diastolic reactivity to the delivery of a short speech in a group of 48 subjects. No correlations were observed for two other tasks. In the most recent study by Burker et al. (1994), those above the median of blood pressure reactivity to a series of four stressors had higher levels of cholesterol and LDL and lower levels of HDL. These results suggest some connection between lipid parameters and cardiovascular reactivity but the pattern of results is confusing, not consistent, and often contradictory between studies. A peculiar aspect is that within studies only for certain tasks an association was observed, whereas for other tasks the results were negative. The results do not allow a conclusion concerning the question whether lipids are associated with cardiac or with vascular responding. Only McKinney et al. (1987)measured the response of the total peripheral resistance and indeed observed a stronger response to be associated with elevated lipids. Other studies, however, showed HR and SBP reactions to be related to lipid levels, supporting an association of lipids with cardiac responding. Limitations of these studies are the small number of subjects and the use of clinical groups in some of them, sometimes without control for medication. In McKinney's study half of the subjects were on antihypertensive medication, which probably comprised of β-blockers. In the study of Fredrikson and Blumenthal (1988)study more than two thirds of the subjects were using β-blockers. β-Blockers are known to influence both lipid profile and HR and DBP reactivity to stress. Other points of concern are the influence of age, sex, and female contraceptive use. In several studies the age range of the subjects was large. This may influence the size and the direction of the correlations, because serum cholesterol is known to rise with age, whereas HR responsivity declines with age ( Snieder et al., 1995). Some studies corrected for age effects, whereas others did not. Because there are sex differences in lipoprotein levels and in cardiovascular reactivity it is indicated to analyse males and females separately. Contraceptive use is known to influence lipid profile thus contraceptive users should be excluded. In the present study the relationship between cardiovascular reactivity and plasma lipid and lipoprotein levels was explored in four groups of middle-aged subjects, and in two subject groups of pubertal age. Males and females, and the older and younger groups will be analysed separately. To find out to what extent age is of influence on the association between lipids and reactivity, in case age is significantly associated with lipid level within a group, relations will be presented with and without age correction. Subjects on medication and female contraceptive users were excluded. In addition to heart rate and blood pressure the cardiac pre-ejection period (PEP) was measured as a specific sympathetic index and total peripheral resistance was measured in part of the groups. These parameters will give some clue to decide between the two explanatory options for an association between lipids and reactivity: either a common sympathetic mechanism or a more peripheral explanation

Results 3.1. Study 1 Lipid levels and the average cardiovascular reactions to the stressors of the four groups of subjects are presented in Table 2. Table 2. Lipid levels and the average cardiovascular responses to two mental stressors (study 1) Middle aged males (n=146) Middle aged females (n=141) Boys (n=160) Girls (n=135) Age 47.9 (6.4) 46.0 (5.9) 16.8 (1.8) 16.2 (1.8) (35–65) (35–59) (14–21) (13–21) Cholesterol (mmol/l) 5.8 (1.0) 5.6 (1.1) 4.1 (0.6) 4.3 (0.7) HDL (mmol/l) 1.1 (0.3) 1.4 (0.3) 1.2 (0.2) 1.4 (0.3) Triglycerides (mmol/l) 1.4 (0.7) 0.9 (0.4) 0.7 (0.3) 0.7 (0.2) IBI (ms) −62.9 (51.6) −72.7 (51.5) −82.2 (52.8) −82.8 (49.6) SBP (mmHg) 9.1 (6.1) 7.8 (7.1) 8.3 (5.4) 7.9 (6.1) DBP (mmHg) 5.0 (3.2) 3.5 (3.7) 6.3 (3.6) 5.6 (3.5) PEP (ms) −6.4 (11.2) −7.0 (10.4) −4.8 (9.7) −5.9 (9.4) Table options Means and standard deviations of cholesterol and HDL of these middle-aged males and females closely resemble the values from a large-scale Dutch epidemiological study including data from 5622 males and 6062 females in the age range from 40 to 49 years (Verschuren et al., 1994). The tendency for males of this age to have somewhat higher cholesterol and triglyceride levels, but lower HDL levels than females, is in line with this Dutch study and with findings abroad (Brunner et al., 1993 and Schaefer et al., 1994). The values for the boys and girls, and the tendency of boys to have somewhat lower cholesterol and HDL values at the age of 16, correspond to reports from other studies on this age group in The Netherlands and abroad (Lauer et al., 1988, Srinivasan et al., 1993 and Twisk et al., 1995). In middle-aged males age was unrelated to lipid values and reactivity to the tasks. HDL was associated with a somewhat stronger HR reaction to the tasks (r=0.18, p<0.05). In middle-aged females, increasing age was associated with higher cholesterol (r=0.37, p<0.01), triglycerides (r=0.21, p<0.05) and HDL (r=0.17, p<0.05). A higher cholesterol level was associated with a smaller cardiac sympathetic reaction as indicated by the PEP, but this only after age correction (r=−0.17, p<0.05). Higher triglyceride levels were associated with a stronger SBP response (r=0.18, p<0.05), though this correlation became non-significant after age correction. In the two young groups, no significant associations emerged between lipid values and reactivity. Thus the evidence for an association between lipids and reactivity in these rather large groups is meager. The scarce significant correlations are low and may well be chance findings. Yet in a multivariate approach we found some evidence for an association. In the group of middle-aged females, canonical analysis (on the age-corrected values) revealed a significant multivariate effect (F(12,331)=1.83, p=0.04). The loadings of the variables on the one significant canonical variate are shown in Table 3. Table 3. Canonical variate–variable loadings of lipids and cardiovascular responses to the mental stressors in middle-age females (study 1) Canonical variate Cholesterol 0.87 Triglycerides 0.83 HDL −0.14 IBI −0.45 PEP −0.51 SBP 0.36 DBP −0.10 Table options This canonical variate is mainly defined by cholesterol and triglycerides, their loadings being 0.87 and 0.83, respectively. The canonical correlation was 0.33. This indicates that the set of lipid variables and the set of reactivity variables share 10% common variance. The same analysis on the non-age-corrected correlations showed a canonical correlation of 0.36. The negative loadings of IBI and PEP on this canonical variate suggest smaller cardiac responsivity to be associated with higher cholesterol and triglycerides in these middle-aged females. Our second study, comprising middle-aged males and females, allows a check of the replicability of this finding. 3.2. Study 2 Lipid levels and the cardiovascular reactions to the stressors are presented in Table 4. Table 4. Lipid levels and the average cardiovascular responses to three mental stressors and to the cold pressor (study 2) Males (n=178) Females (n=169) Age 43.1 (6.2) 45.1 (6.7) (35–63) (65–61) Cholesterol (mmol/l) 5.4 (1.0) 5.5 (1.1) HDL (mmol/l) 1.1 (0.3) 1.4 (0.3) Triglycerides (mmol/l) 1.3 (0.8) 1.0 (0.4) Mental stressors IBI (ms) −118.5 (59.7) −103.3 (64.0) SBP (mmHg) 12.0 (7.9) 10.4 (8.8) DBP (mmHg) 6.7 (5.7) 6.1 (5.4) PEP (ms) −9.0 (11.6) −10.3 (9.0) SV (%) 5.5 (25.9) 11.7 (25.8) CO (%) 20.8 (31.6) 27.2 (33.9) TPR (%) −1.2 (26.7) −6.7 (26.0) Cold pressor IBI (ms) −49.1 (76.0) −66.2 (62.3) SBP (mmHg) 13.2 (7.8) 14.9 (9.7) DBP (mmHg) 12.1 (7.6) 11.5 (7.2) PEP (ms) −1.5 (10.1) −1.2 (7.0) SV (%) −8.8 (16.8) −7.1 (15.4) CO (%) −3.8 (17.3) 0.7 (18.9) TPR (%) 23.8 (41.3) 17.2 (23.7) Table options The response pattern to the mental stressors contrasted in the expected way with the one to the cold pressor. Mental stress led to a predominantly cardiac response and a slight decrease in peripheral resistance, whereas the cold pressor mainly triggered a vascular response. In Table 5 the correlations between lipid values and the cardiovascular responses are presented. Table 5. Correlations between lipid values and the average cardiovascular responses to three mental stressors, and to the cold pressor (study 2) (age-corrected values between brackets) Age Chol. HDL Trigl. Males (n=178) Mental stressors Age 0.19** IBI −0.16* SBP DBP −0.13 (−0.18*) PEP 0.15* (0.15*) SV% 0.19** CO% TPR% −0.17* Cold pressor IBI SBP DBP −0.17* (−0.20*) PEP 0.19** 0.16* (0.13) SV% 0.16* CO% 0.17* TPR% 0.17* (0.17*) Females (n=165) Mental stressors Age 0.48** 0.19* 0.19* IBI SBP DBP PEP SV% 0.20** 0.17* (0.08) 0.20** (0.17*) CO% 0.20** 0.25** (0.22**) TPR% −0.19** −0.15* (−0.14) Cold pressor IBI −0.19* (−0.19*) SBP 0.20** DBP PEP −0.08 (−0.15*) SV% CO% −0.09 (−0.15*) TPR% *p<0.05; **p<0.01 (two-sided). Table options In both males and females, age was associated with both the lipid values and the reactivity measures. Because of this, we will focus on the age-corrected correlations further. In males, cholesterol showed a positive correlation with the PEP response to the mental tasks. HDL was associated with smaller DBP reactivity to both the mental tasks and the cold pressor. Triglyceride level was related to a slightly stronger rise in TPR in response to the cold pressor. In females, HDL was associated with stronger SV and CO reactions to the mental tasks. Cholesterol level was associated with smaller responses of PEP and CO to the cold pressor, and triglycerides with a smaller heart rate response. For comparability with study 1, canonical analyses were run for the mental tasks and the cold pressor separately. As in study 1, canonical analysis only evidenced an association between the sets of lipids and reactivity to the mental tasks for the females. The association, however, was weaker and statistically only a trend (F(21,442)=1.42, p=0.10). The canonical correlation was 0.30. Age correction does not seem to have been of any influence, the canonical correlation on the non-age-corrected data being 0.31. Table 6 (left column) shows the loadings of the variables on the canonical dimension. Table 6. Canonical variate–variable loadings of lipids and cardiovascular responses to the mental stressors in females and to the cold pressor in males and females (study 2) Females, mental stress Males, cold pressor Females, cold pressor Cholesterol 0.33 0.01 0.46 Triglycerides −0.60 0.74 1.00 HDL 0.73 0.04 −0.41 IBI 0.36 −0.49 −0.68 SBP 0.33 −0.54 −0.16 DBP 0.70 −0.13 −0.40 PEP 0.42 −0.05 −0.16 SV% 0.53 −0.37 0.28 CO% 0.65 −0.58 −0.03 TPR% −0.33 0.59 −0.40 Table options HDL and the DBP, CO and SV reactivity showed the highest loadings in the same direction, suggesting a stronger cardiac reaction being associated with high HDL. Triglycerides load negatively on this dimension, suggesting an association between high triglycerides and lower cardiac reactivity. In study 1 the canonical analysis also suggested such a relationship. One should, however, be careful in comparing the results of the analyses because the pattern of loadings of the lipid parameters on the canonical dimensions differ between the two studies. A more direct test of the association between vascular responsiveness and lipids can be furnished by canonical analyses on the cold pressor results. For both males and females the canonical analyses showed a trend. (F(21,482)=1.44, p=0.09, and F(21,442)=1.48, p=0.08, respectively)). Both canonical correlations were 0.28. Again the canonical correlations on the non-age-corrected data were about the same: 0.29 and 0.30 for males and females, respectively. In Table 6 (middle and right column) the loadings of the variables on the canonical variates are presented. In males high triglycerides seem to be associated with a stronger rise in TPR, a smaller rise (or stronger decrement) in CO, and smaller heart rate and SBP reactivity. The pattern of loadings in females is different. Here, also, triglycerides have the highest loading (even 1.00) on the canonical dimension. As in males, high triglycerides seem to be associated with a smaller heart rate reaction. The sign of TPR is, however, negative, which suggests an opposite relation between triglycerides and vascular reactivity in males and females. We should interpret the results of these canonical analyses with caution because the multivariate associations were only trends.