- Email: [email protected]
Relation of alcohol and cigarette consumption to blood pressure and serum creatinine levels
J Chron Dis Vol. 37, No. 8, pp. 617-623, 1984 Printed in Great Britain. All rights reserved
RELATION CONSUMPTION
ELLIOTT
SAVDIE,‘*
Copyright
0
0021-9681/84 $3.00 + 0.00 1984 Pergamon Press Ltd
OF ALCOHOL AND CIGARETTE TO BLOOD PRESSURE AND SERUM CREATININE LEVELS GEORGE
M.
GROSSLIGHT~~
and MICHAEL
A.
ADENA~$
‘Renal Unit, Department of Medical Research, Kanematsu Memorial Institute, Sydney Hospital, Sydney, NSW 2000, Australia, *Medicheck Referral Centre, Sydney and ‘Department of Population Biology, Australian National University, Canberra, ACT 2600, Australia (Received in revised form 9 January 1984)
Abstract-From the records of an automated multi-phasic health testing centre, daily drinkers in four alcohol consumption groups were each separately matched for sex, age and obesity to a single . . non-drinker control. Aiisubjects satisfied strict ehgiblhty criteria seiected to exclude the effects of other factors known to influence blood pressure or renal function or both. The 5500 pairs of subjects were compared for systolic and diastolic blood pressure and serum creatinine. After allowing for smoking, drinkers had significantly elevated blood pressure compared with their controls, and the elevation was greater the heavier the alcohol intake, except for the heaviest drinking females. This result was more pronounced in males than females, and for systolic than diastolic blood pressure. By contrast, smoking cigarettes was shown to be associated with lower blood pressure, independent of sex and drinking history. Smoking was also associated with a decreased serum creatinine concentration as was drinking three or more drinks per day. However, drinkers of two or fewer drinks dally had higher serum creatinine concentrations than their non-drinker controis.
INTRODUCTION
A POSITIVE association between ethanol consumption and hypertension has been noted in several studies [l-8]. Since numerous other factors may also influence blood pressure, these studies have used stratification, regression and standardisation to take into account the effects of sex and age; as well as controlling for other possible confounders, such as adiposity or weight [3-81, race [3], cigarette smoking [3,4,6-g], coffee use [3], educational attainment [3, 71, salt intake [6], family history of hypertension [6], exercise [7] and use of oral contraceptives [7]. The possibility of a dose-response between the quantity of alcohol imbibed and the absolute level of blood pressure had led several researchers to stratify their subjects by the amount of alcohol consumed [3, 5,6,8]. Alcohol consumption may also influence levels of hormones produced from both renal and extra-renal sources [9, IO], and, in an experimental rat model, has demonstrated toxic effects on both renal morphology and function [l 11.The aim of this study was to examine the relationship between alcohol consumption and both blood pressure and renal function (measured by serum creatinine concentration), while eliminating or controlling possible confounding factors. The design of this study has been to stratify for sex and alcohol consumption, to match for age and obesity and to estimate and control for the effects of cigarette smoking by regression. This has been coupled with stringent exclusions for factors known to affect blood pressure or renal function.
Present addresses: *Renal Unit, Royal North Shore Hospital, St Leonards, NSW 2065, Australia, 725 Daley Avenue, East Botany, NSW 2019, Australia and tIntstat Australia Pty. Ltd. 25 Market Street, Sydney, 2000 NSW, Australia. 6!?
618
ELLMT SUBJECTS
SAVDIE
AND
et al. METHODS
Data on drinking and non-drinking (control) subjects were drawn by computer from the records of the first visits of 27,435 women and 34,088 men who attended Medicheck Referral Centre in Sydney, Australia between July 1975 and October 1978. Patients attending this centre are referred by their family physician for an automated multiphasic health evaluation. Subjects were excluded for a number of criteria known to influence blood pressure or renal function or both; these were: (1) Self-reported consumption of analgesics, tranquillisers, anti-depressants, hallucinogens, central nervous system stimulants, narcotics, cannabis or diuretic drugs; (2) A history of treatment for cardiac or respiratory disease; (3) Self-reported diabetes, or a fasting serum glucose concentration greater than 6.6 mmol/l; (4) Abnormal thyroid function determined by a high (> 1.08) or low (< 0.89) effective thyroxine ratio measured by a modification of the “Res-O-Mat” ETR test (Mallinckrot Nuclear) usingv acid ---~- extraction; (5) Presence of syphilis as indicated by a positive VDRL test, and; (6) A past history of renal disease. In a computer-administered questionnaire, subjects were asked the following questions: (1) Do you drink alcohol? (Yes/Yes, but only a small amount/No), (2) In the past year, how often did you drink alcohol? (Every day or most days/A couple of times a week/Every week or two/Very rarely), and(3) How many drinks (wine, whisky, beer, cocktails etc.) did you usually have on each drinkinau dav in the nast ,vex? 8 &v i nr ~~, ___ .__. I--. ___. . Rota1 \- ---- of -- _9 &j&s -- mnreih -m----l” to 8 drinks a day/3 to 5 drinks a day/2 or less drinks a day). Drinking subjects gave either the first or second response to the first question and the first response to the second question, and were divided into four groups (Table 1) by their responses to the third question. A standard alcoholic drink, regardless of the type of beverage, contains about 10 g of ethanol. Control subjects answered the first question with the third response, and were not asked the two subsequent questions. Each drinking subject was randomly matched by computer to a non-drinking control of the same sex, age in years and obesity (measured by the Quetelet Index, weight divided by height squared, in kg/m2, and matched to the nearest even integer). We aimed to randomly select by computer as many pairs as possible of each sex for each of the four drinking categories. However, the paucity of female drinkers in groups 3 and 4 resulted in only 55 and 11 subject pairs for these groups respectively. The pool of non-drinkers was sufficient to ensure that the closeness of matching did not result in many drinkers being discarded for lack of a non-drinker control. The drinkers in each group were compared with their corresponding controls for the following variables: systolic and diastolic blood pressure (first and last Korotkoff sounds of the supine, resting subject measured by a trained nurse with a mercury sphygmomanometer; the procedure was repeated after a few minutes if either the systolic blood pressure was 160 mmHg or above, or the diastolic blood pressure was 90 mmHg or above), serum creatinine concentration (modified Jaffe reaction, Technicon method number SF4-OOllFH4), the presence of absence of protein in mid-stream urine specimen, the
TABLE 1. AGE AND CIBES~TY (MEAN + SE)
Daily alcohol consumption of drinkers (drinks) $2 3-5 6-8 a9
Number of pairs Females Males 797 486 55 11
iO3-l 1957 828 329
Age (~0 Females 44.8 45.0 47.5 46.0
+(14 * 0.5 * 1.6 k 3.0
OF SUBJECTS*
Males
43.3 42.9 42.7 41.7
& 04 IO.3 10.4 2 0.6
Obesity (kg/m’) Males Females 22.3 22.6 24.0 20.6
lo.! f 0.1 + 0.4 2 0.5
24.3 24.7 25.2 25.4
IO.! 2 0.1 + 0. I + 0.2
*As each subject was matched for age and obesity to a control, these mean measurements apply to both control and experimental subjects.
Alcohol,
Cigarettes,
Blood
Pressure
and Serum
619
Creatinine
presence or absence of micro-organisms in the urine specimen (more than, or less than, IO5 organisms per ml in pure growth, respectively), and the cellular and granular cast content of the spun urine sediment. ” .. suqects were classified as either heavy smokers (smoke ten or more cigarettes daiiyj, light smokers (smoke less than ten cigarettes daily), ex-smokers, or non-smokers. The effects of smoking on blood pressure and serum creatinine were each estimated, whilst controlling for alcohol consumption, sex, age and obesity. These results were analysed using parametric and non-parametric methods of matched pair analysis, one-way analysis of variance and two-way contingency tables. Analyses taking into account the effect of cigarette smoking of the drinking and control subjects relied on extensions of the parametric and non-parametric methods of matched pair analysis and involved fitting models analogous to the Bradley-Terry model for paired comparisons [ 121. These models allowed for the possibility that smokers and non-smokers might have shown different relationships between alcohol consumption and blood pressure or renal function; the models also investigated whether the sexes differed in these relationships. Such sex differences were apparent in the blood pressure analyses, but were absent in those for serum creatinine (see Results). The effects of smoking and drinking on blood pressure and serum creatinine concentrations were independent for all these analyses. RESULTS
The mean age and obesity of the subject groups are given in Table 1. Within each sex the mean ages of the groups were not significantly different, However, the groups were heterogeneous with respect to obesity. In males obesity was greater in the two groups of highest alcohol consumption (variance ratio, F,,,,,, = 20.6, p < 0.001). Females in the two groups of lowest alcohol consumption were less obese than those in the 6-8 drinks daily group, and females in the highest consumption group were the least obese (F,,,,,, = 9.51, p < 0.001). Mean systolic and diastolic blood pressure and mean serum creatinine concentrations of the drinkers and non-drinkers in each group are given in Table 2. Average matched pair differences, controlling for the possible confounding effect of smoking, are shown in Table 3. The systolic blood pressure of males who drank more than 9 drinks daily was significantly elevated by 10.6 mmHg (t = 8.1, p < 0.001) compared with their non-drinking age- and obesity-matched controls. For the groups of males who drank 6-8 drinks, 3-5 drinks and 2 or fewer drinks daily, the elevation in systolic blood pressure was 9.9 mmHg (t = 12, p < O.OOl), 5.2 mmHg (t = 9.5,~ < 0.001) and 2.3 mmHg (t = 3.0, 0.001
Daily alcohol consumption of drinkers (drinks)
Drinkers
Females Non-drinkers
Systolic BP (mmb)
<2 3-s 68 .9
122.5 125.6 132.0 124.0
t rt + +
0.7 1.0 3.0 5.0
122.8 123.3 129.0 123.0
+ 0.7 Li:0.9 Lt:3.0 i: 6.0
Diastolic BP (mmH&
$2 3-5 f&8 >9
15.7 77.0 80.6 77.0
* * * *
0.4 0.5 1.8 3.0
75.9 76.1 77.0 79.0
+ i tf 2
Serum creatinine (pm&b
<2 3-5 6-S 39
74.5 + 0.4 72.9 & 0.5 73.1 * 1.9 71.0+4.0
73.3 73.7 74.2 67.0
rt 0.4 rl_0.5 + 1.5 + 4.0
0.4 0.5 1.4 4.0
Drinkers 134.1 137.2 141.1 142.4
* 0.6 i 0.4 * 0.7 +_ 1.1
83.8 85.6 88.4 89.5
f 0.3 * 0.2 IO.4 f 0.7
92.4 t 0.4 91.5kO.3 90.2 k 0.5 86.7 ?r 0.7
Males Non-drinkers 132.1 132.4 132.0 132.9
rtO.6 k 0.4 + 0.6 + 0.9
83. I + 0.4 83.3 + 0.3 83.2 + 0.4 83.8 i 0.6 92.2 91.9 91.8 91.7
+ 0.4 * 0.3 +O.S io.7
620
ELLIOTT SAVDIE el al
Daily alcohol consumption of drinkers (drinks)
I
Females
Males
I
<2 3-5 6-8 29
0.1 3.1 5.0 2.0
+ * * *
0.8 1.1 3.0 7.0
A Systolic BP (mmHg) 0.0 (NS) 2.3 k 0.7 2.9** 5.2 * 0.5 1.7 (NS) 9.9 f 0.8 0.3 (NS) 10.6 +_ 1.3
3.08’ 9.5*** 12.0*** s.1***
$2
0.0 1.6 4.6 - 2.0
* f 5 +
0.5 0.6 1.8 4.0
A Diastolic BP (mmHg) 0.0 (NS) 0.7 * 0.4 2.7** 2.7 + 0.3 2.6* 5.7 * 0.5 0.5 (NS) 6.5 k 0.7
1.8 (NS) 9.0*** 11.4*** 9.3***
3-5 4-a 29
A Serum creatinine (pd/l) t 0.9 + 0.4 2.3* -0.1 kO.4 0.3 (NS) -1.1 +0.5 2.2’ -4.0 f 1.0 4.1.”
<2 3-5 &8 39
***p i
0.001 **p < 0.0 I *p < 0.05 (NS) not significant
compared
with non-drinkers.
In females, however, the differences in blood pressure between drinkers and nondrinkers were much less marked. No differences were evident in females drinking 2 or fewer drinks daily, although both blood pressures were significantly elevated in women drinking 3-5 drinks daily (systolic: 3.1 mmHg, t = 2.9, 0.001 < p < 0.01; diastolic: 1.6 mmHg, t = 2.7, 0.001
Smoking
history
Ex-smoker Light smoker Heavy smoker
***p < 0.001 ‘p < 0.05 (NS) not significant
A Systolic BP (mmHg)
f
-1.4* 1.0 -1.6+0.7 -2.5 f 0.6
1.4 (NS) 2.3’ 4.2”’
compared
with non-smokers
A Diastolic BP (mmHg) -0.1 -0.9 -2.0
+ 0.5 * 0.4 f 0.3
I 0.2 (NS) 2.2’ 6.7*‘*
A Serum creatinine (mow) -0.2 f 0.7 -1.1 kO.5 -1.9io.4
I 0.3 (NS) 2.2* 4.v**
Alcohol, Cigarettes, Blood Pressure and Serum Creatinine
621
cigarette exposure was greater. Because serum creatinine concentrations were measured only to the nearest 10 pmol/l, some caution is justified when interpreting those analyses. However, parallel non-parametric analysis yielded similar results. DISCUSSION
This study confirms the previously reported positive correlation between alcohol consumption and blood pressure. Its size, design and analysis give it good power. The sample of 5500 matched pairs of subjects compared favourably with unmatched crosssectional studies of 2100 [4], 3400 [5], 85 [6], 4800 [7], 20,920 [8] and 84,000 [3] subjects. In our study, the drinkers admitted to drinking alcohol daily or most days. Therefore. we have not addressed the question of the blood pressure and renal function of less frequent drinkers. One factor leading to possible overestimation of the effects of alcohol on blood pressure in this and other studies is that drinkers tend to understate their consumption of alcohol, either consciously or subconsciously. The magnitude of this effect has not been ascertained here. Since the subjects had fasted for at least 12 hours before examination it is possible that the blood pressure of the heavy drinkers was affected by abstinence from alcohol. Sympathetic overactivity, resulting from alcohol withdrawal of even a few hours’ duration, has been advanced as one explanation for the alcohol-blood pressure association in studies such as these [13], but has not been tested. However, 12 hours would have been insufficient time for blood pressure elevated by drinking to have returned toward the “normal” non-drinking level [14, 151. The patients attending the Medicheck Referral Centre are referred by a general practitioner, and are apparently well. They are, however, not a random sample of the general population. For example, attenders tend to be more affluent and better educated than non-attenders [16]. However, bias with respect to sex, age, obesity and admitted smoking habits have been explicitly eliminated by design and analysis in the work reported here;
and
bias
due to other
possible
factors
has been
abated
by the selection
of controls
the same subpopulation. In women, drug and alcohol abuse often occur together [17] so the paucity of females who admitted to drinking 6-8, or 9 or more drinks daily and who were not excluded from this study because of drug use in not surprising and warns of the unusual nature of those heavy drinking females who remained in the study. For example, the heaviest drinking females were much less obese than females in the other groups, whereas drinking and obesity were positively correlated in the males. In this context, the observed discrepancy between males and females of the blood pressure differences between drinkers and non-drinkers is of little value. When the heaviest drinking female group is excluded, there is a ciear dose-response reiationship ‘between aicohoi consumprion and biood pressure. Nevertheless, at least two other groups [3, 51, have shown that the elevation of blood pressure in women by alcohol is inconsistent and less than in males, raising the possibility that there is, in fact, a sexual factor modifying the effect of alcohol on blood pressure. Althouth the blood pressures of both drinkers and non-drinkers in both sexes were correlated with age and obesity, the matched pair differences were not. Therefore, there was no effect modification by age or obesity. The variances of these differences were homogeneous with obesity and age, except for that of the difference in systolic blood pressure which was positively correlated with age. This presumably is a consequence of the greater variance in systolic blood pressure shown by older subjects, whether drinkers or non-drinkers. However, this will not affect the conclusions previously drawn because the mean ages are similar for subjects from each group for both males and females. The work presented here supports that of Cooke et al. [8] that, at least in males, small but significant elevations in blood pressure are seen in those with even very modest alcohol consumption (2 or less drinks per day), arguing against the proposition that there is a “threshold” level of alcohol intake above which elevation will occur [3]. At the other end of the scale, our data showed that those males drinking 9 or more alcoholic units per day from
622
ELLIOTT SAVDIE et al.
had blood pressure elevations only slightly greater than those imbibing 68 units per day thus supporting the notion that there is a “maximal” effect of alcohol on blood pressure when about 9 units per day are ingested. Cooke et al., described a plateau at a similar level nf “I
onnrnmnt;r\n ~“LI~UI‘I~LI”LI
nf “I
Qhr\,.t LL”“UL
LCI ,,n;tr ,.P,. “” U‘AILO ybL
..mz.lr r*tAd*
IQ1 L”,.
This study also suggests that alcohol causes a greater elevation in systolic than in diastolic blood pressures supporting the observations of others [3, 5, 81. It has been shown in alcoholics that blood pressure is correlated with their mean daily alcohol intake over the preceding three months, and falls to normal after detoxification [ 141. Furthermore, the blood pressure will remain normal in those who continue to abstain, but rises again in those who recommence drinking. These data suggest that the effect of alcohol on blood pressure, though sustained, is reversible and not dependent on underlying pathological damage to the liver. They are supported by our data and those of others [3,8] showing a significant rise in blood pressure with an intake of alcohol in quantities well short of that which would be expected to cause liver disease. Klatsky et al. [3] hypothesized that “stress” or other psychosocial factors might be responsible for both high alcohol consumption on the one hand and high blood pressure on the other, but the fall in blood pressure in the detoxified alcoholics studied by Saunders [14] argues against such a postulate. It seems more likely that the hypertensive effect of ethanol is due to a haemodynamic effect on the circulation. Although some investigators have reported small increases in both systolic and diastolic blood pressures following the acute administration of ethanol [ 181, others have not [ 191. In the recovery phase from acute alcohol intoxication, renin and aldosterone production are stimulated [IO]. It is possible, therefore, that acute salt and water retention or the vasoconstrictor effects of angiotensin II could be responsible for the hypertensive effects of alcohol. The inverse relationship between alcohol consumption and serum creatinine concentration shown here, while statistically significant, is slight and not of clinical significance. Serum creatinine levels depend on creatinine production from the muscle mass, and renal clearance in this study we are reluctant to postulate that alcohol enhances creatinine clearance which is entirely due to glomerular filtration. Since we have not measured the creatinine clearance in this study we are reluctant to postulate that alcohol enhances creatinine clearance by the kidney, especially since studies in the rat have shown alcohol to have an adverse effect on both renal structure and function, leading to an increase in serum creatinine concentration [I 11. It is possible that the effect of alcohol on serum creatinine concentration observed here is a dilution phenomenon due to the retention of salt and water as a result of the perturbations in hormone levels mentioned above; but reported increases in blood volume in alcoholics are relatively small (approximately 5%) [20]. The explanation which we favour more is that subjects with a heavy alcohol intake become malnourished and develop muscle wasting and sometimes frank myopathy [21]. ,a:“+.,,L,.l,, ,.-..,A :.. t.,..r\ ..,,.,l+ ;.. A;m;..;nhcJ ,...-n+;*;s.3 . . . ..A fiA -..,,.1, 111~~~1~UISCUI ULIIILGGUU~U, 111 LULL,,IGJUIL 111a UIIIIILIIJIIGU ~LG~CIIIIIIC; p~uuiiCtiGfi ZiIb 2 decrease in the serum creatinine level independent of any change in plasma volume or glomerular filtration rate. Further studies involving measurement of both creatinine clearance and lead body mass would be needed to clearly delineate any effect of alcohol on renal function in man. The association of cigarette smoking with small decrements in blood pressure has been reported before [3,8,22], and appears to be independent of other variables. Because this effect is opposite to that of alcohol, and since alcohol and cigarette consumption are positively associated, cigarette consumption needs to be controlled for. Since nicotine is known to cause an elevation in blood pressure in acute experimental situations [23,24], the association of chronic smoking with decreased blood pressure is interesting and suggests the existence in cigarette smoke of substances other than nicotine which have an opposite, hypotensive action. The mechanism of the tiny decrease in serum creatinine concentrations caused by smoking is unclear. Acknowledgements-We would like to thank Dr H. G. Gallagher, Referral Centre, for assistance with this study.
formerly
Medical
Director
of the Medicheck
Alcohol,
Cigarettes,
Blood
Pressure
and Serum
Creatinine
623
REFERENCES 1. 2. 3. 4.
5. 6. 7.
8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24.
Clark VA. Chaoman JM, Coulson AH: Effects of various factors on systolic and diastolic blood pressure in the Los Angkles Heart Study. J Chron Dis 20: 571-581, 1967 Kannel WB, Sorlie P: Hypertension in Framingham. In Epidemiology and Control of Hypertension. Paul 0 (Ed) New York: Stratton Intercontinental Medical Book Corporation, 1974. pp 553-592 Klatsky AL, Friedman GD, Siegelaub AB, Gerard MJ: Alcohol consumption and blood pressure: Kaiser-Permanente Multiohasic Health Examination data. N Engl J Med 296: 1194-1200. 1977 Dyer AR, Stamler J, Paul 0, Berkson DM, Leeper MH, McKean-H, Shekelle RB, Lindberg HA, Garside D: Alcohol consumption, cardiovascular risk factors and mortality in two Chicago epidemiologic studies. Circulation 56: 1067-1074, 1977 Harburg E, Ozgoren F, Hawthorne VM, Schork MA: Community norms of alcohol usage and blood pressure: Tecumseh, Michigan. Am J Pub1 Hlth 70: 813-820, 1980 Mitchell PI, Morgan MJ, Boadle DJ, Batt JE, Marstrand JL, McNeil HP, Middleton C, Rayner K, Lickiss JN: Role of alcohol in the aetiology of hypertension. Med J Aust 2: 198-200, 1980 Criqui MH, Wallace RB, Mishkel MA, Barrett-Connor E, Heiss G: Alcohol Consumption and Systolic Blood Pressure--the Lipid Research Clinics Prevalence Study. Society for Epidemiological Research, 1980 (abstract) Cooke KM, Frost GW, Thorneil IR, Stokes GS: Alcohol consumption and blood pressure. Survey of the relationship at a health-screening clinic. Med J Aust 1: 65-69, 1982 Hauger-Kievene JH, Pinkas MBrDegrossi OJ: Paradoxical effect of ethanol on plasma renin activity levels in hyper-reninemic essentiai hypertension. Medicina Buenos Aires 39: 6i4-6i 8, i979 Linkola J, Fyhrquist F, Ylikahri R: Renin, aldosterone and cortisol during ethanol intoxication and hangover. Acta Physiol Stand 106: 75-82, 1979 van Thiel DH, Gavaler JS, Little JM, Lester R: Alcohol: its effect on the kidney. Metabolism 26: 857-866, 1977 Bradley RA, Terry ME: Rank analysis of incomplete block designs. I. The method of paired comparisons. Biometrika 39: 324-345, 1952 Dolinsky H. Letter to the Editor. N Engl J Med 297: 450, 1977 Saunders JB, Beevers DG, Paton A: Alcohol-induced hypertension. Lancet ii: 653-656, 1981 Ashley MJ, Rankin JG: Alcohol consumption and hypertension--the evidence for hazardous drinking and aicohoi popuiations. Aust NZ J Meri 9: 201-206, i979 Rawson G: Patient attitudes to screening by an automated multiphasic health testing facility. Aust Fam Phys 4: 219-222, 1975 Reynolds I, Harnas J, Gallagher H, Bryden D: Drinking and drug taking patterns of 8,516 adults in Sydney. Med J Aust 2: 782-785, 1976 Orlando J, Aronow NS, Cassidy J, et al: Effect of ethanol on angina pectoris. Ann Int Med 84: 652-655, 1976 Conway N: Haemodynamic effects of ethyl alcohol in patients with coronary heart disease. Br Heart J 30: 638-644, 1968. Beard JD. Knott DH: Fluid and electrolyte balance during acute withdrawal in chronic alcoholic patients. JAMA 204: 135-139, 1968 Knochel JP, Bilbrey GL, Fuller TJ, Carter NW: The muscle cell in chronic alcoholism: the possible role of phosphate depletion in alcoholic myopathy. Ann NY Acad Sci 252: 274-286, 1975 Arkwright PD, Beilin LJ, Rouse I, Armstrong B, Vandongen R: Alcohol and blood pressure in a working population. Clin Exp Pharmac Physiol 8: 451-454, 1981 Gebber GL: Neuroaenic basis for the rise in blood pressure evoked by nicotine in the cat. J Pharmac Exp Ther 166: 255-269,-1969 Roth GM, Shick RM: Effect of smoking on the cardiovascular system of man. Circulation 17: 443-459, 1958
RELATION CONSUMPTION
ELLIOTT
SAVDIE,‘*
Copyright
0
0021-9681/84 $3.00 + 0.00 1984 Pergamon Press Ltd
OF ALCOHOL AND CIGARETTE TO BLOOD PRESSURE AND SERUM CREATININE LEVELS GEORGE
M.
GROSSLIGHT~~
and MICHAEL
A.
ADENA~$
‘Renal Unit, Department of Medical Research, Kanematsu Memorial Institute, Sydney Hospital, Sydney, NSW 2000, Australia, *Medicheck Referral Centre, Sydney and ‘Department of Population Biology, Australian National University, Canberra, ACT 2600, Australia (Received in revised form 9 January 1984)
Abstract-From the records of an automated multi-phasic health testing centre, daily drinkers in four alcohol consumption groups were each separately matched for sex, age and obesity to a single . . non-drinker control. Aiisubjects satisfied strict ehgiblhty criteria seiected to exclude the effects of other factors known to influence blood pressure or renal function or both. The 5500 pairs of subjects were compared for systolic and diastolic blood pressure and serum creatinine. After allowing for smoking, drinkers had significantly elevated blood pressure compared with their controls, and the elevation was greater the heavier the alcohol intake, except for the heaviest drinking females. This result was more pronounced in males than females, and for systolic than diastolic blood pressure. By contrast, smoking cigarettes was shown to be associated with lower blood pressure, independent of sex and drinking history. Smoking was also associated with a decreased serum creatinine concentration as was drinking three or more drinks per day. However, drinkers of two or fewer drinks dally had higher serum creatinine concentrations than their non-drinker controis.
INTRODUCTION
A POSITIVE association between ethanol consumption and hypertension has been noted in several studies [l-8]. Since numerous other factors may also influence blood pressure, these studies have used stratification, regression and standardisation to take into account the effects of sex and age; as well as controlling for other possible confounders, such as adiposity or weight [3-81, race [3], cigarette smoking [3,4,6-g], coffee use [3], educational attainment [3, 71, salt intake [6], family history of hypertension [6], exercise [7] and use of oral contraceptives [7]. The possibility of a dose-response between the quantity of alcohol imbibed and the absolute level of blood pressure had led several researchers to stratify their subjects by the amount of alcohol consumed [3, 5,6,8]. Alcohol consumption may also influence levels of hormones produced from both renal and extra-renal sources [9, IO], and, in an experimental rat model, has demonstrated toxic effects on both renal morphology and function [l 11.The aim of this study was to examine the relationship between alcohol consumption and both blood pressure and renal function (measured by serum creatinine concentration), while eliminating or controlling possible confounding factors. The design of this study has been to stratify for sex and alcohol consumption, to match for age and obesity and to estimate and control for the effects of cigarette smoking by regression. This has been coupled with stringent exclusions for factors known to affect blood pressure or renal function.
Present addresses: *Renal Unit, Royal North Shore Hospital, St Leonards, NSW 2065, Australia, 725 Daley Avenue, East Botany, NSW 2019, Australia and tIntstat Australia Pty. Ltd. 25 Market Street, Sydney, 2000 NSW, Australia. 6!?
618
ELLMT SUBJECTS
SAVDIE
AND
et al. METHODS
Data on drinking and non-drinking (control) subjects were drawn by computer from the records of the first visits of 27,435 women and 34,088 men who attended Medicheck Referral Centre in Sydney, Australia between July 1975 and October 1978. Patients attending this centre are referred by their family physician for an automated multiphasic health evaluation. Subjects were excluded for a number of criteria known to influence blood pressure or renal function or both; these were: (1) Self-reported consumption of analgesics, tranquillisers, anti-depressants, hallucinogens, central nervous system stimulants, narcotics, cannabis or diuretic drugs; (2) A history of treatment for cardiac or respiratory disease; (3) Self-reported diabetes, or a fasting serum glucose concentration greater than 6.6 mmol/l; (4) Abnormal thyroid function determined by a high (> 1.08) or low (< 0.89) effective thyroxine ratio measured by a modification of the “Res-O-Mat” ETR test (Mallinckrot Nuclear) usingv acid ---~- extraction; (5) Presence of syphilis as indicated by a positive VDRL test, and; (6) A past history of renal disease. In a computer-administered questionnaire, subjects were asked the following questions: (1) Do you drink alcohol? (Yes/Yes, but only a small amount/No), (2) In the past year, how often did you drink alcohol? (Every day or most days/A couple of times a week/Every week or two/Very rarely), and(3) How many drinks (wine, whisky, beer, cocktails etc.) did you usually have on each drinkinau dav in the nast ,vex? 8 &v i nr ~~, ___ .__. I--. ___. . Rota1 \- ---- of -- _9 &j&s -- mnreih -m----l” to 8 drinks a day/3 to 5 drinks a day/2 or less drinks a day). Drinking subjects gave either the first or second response to the first question and the first response to the second question, and were divided into four groups (Table 1) by their responses to the third question. A standard alcoholic drink, regardless of the type of beverage, contains about 10 g of ethanol. Control subjects answered the first question with the third response, and were not asked the two subsequent questions. Each drinking subject was randomly matched by computer to a non-drinking control of the same sex, age in years and obesity (measured by the Quetelet Index, weight divided by height squared, in kg/m2, and matched to the nearest even integer). We aimed to randomly select by computer as many pairs as possible of each sex for each of the four drinking categories. However, the paucity of female drinkers in groups 3 and 4 resulted in only 55 and 11 subject pairs for these groups respectively. The pool of non-drinkers was sufficient to ensure that the closeness of matching did not result in many drinkers being discarded for lack of a non-drinker control. The drinkers in each group were compared with their corresponding controls for the following variables: systolic and diastolic blood pressure (first and last Korotkoff sounds of the supine, resting subject measured by a trained nurse with a mercury sphygmomanometer; the procedure was repeated after a few minutes if either the systolic blood pressure was 160 mmHg or above, or the diastolic blood pressure was 90 mmHg or above), serum creatinine concentration (modified Jaffe reaction, Technicon method number SF4-OOllFH4), the presence of absence of protein in mid-stream urine specimen, the
TABLE 1. AGE AND CIBES~TY (MEAN + SE)
Daily alcohol consumption of drinkers (drinks) $2 3-5 6-8 a9
Number of pairs Females Males 797 486 55 11
iO3-l 1957 828 329
Age (~0 Females 44.8 45.0 47.5 46.0
+(14 * 0.5 * 1.6 k 3.0
OF SUBJECTS*
Males
43.3 42.9 42.7 41.7
& 04 IO.3 10.4 2 0.6
Obesity (kg/m’) Males Females 22.3 22.6 24.0 20.6
lo.! f 0.1 + 0.4 2 0.5
24.3 24.7 25.2 25.4
IO.! 2 0.1 + 0. I + 0.2
*As each subject was matched for age and obesity to a control, these mean measurements apply to both control and experimental subjects.
Alcohol,
Cigarettes,
Blood
Pressure
and Serum
619
Creatinine
presence or absence of micro-organisms in the urine specimen (more than, or less than, IO5 organisms per ml in pure growth, respectively), and the cellular and granular cast content of the spun urine sediment. ” .. suqects were classified as either heavy smokers (smoke ten or more cigarettes daiiyj, light smokers (smoke less than ten cigarettes daily), ex-smokers, or non-smokers. The effects of smoking on blood pressure and serum creatinine were each estimated, whilst controlling for alcohol consumption, sex, age and obesity. These results were analysed using parametric and non-parametric methods of matched pair analysis, one-way analysis of variance and two-way contingency tables. Analyses taking into account the effect of cigarette smoking of the drinking and control subjects relied on extensions of the parametric and non-parametric methods of matched pair analysis and involved fitting models analogous to the Bradley-Terry model for paired comparisons [ 121. These models allowed for the possibility that smokers and non-smokers might have shown different relationships between alcohol consumption and blood pressure or renal function; the models also investigated whether the sexes differed in these relationships. Such sex differences were apparent in the blood pressure analyses, but were absent in those for serum creatinine (see Results). The effects of smoking and drinking on blood pressure and serum creatinine concentrations were independent for all these analyses. RESULTS
The mean age and obesity of the subject groups are given in Table 1. Within each sex the mean ages of the groups were not significantly different, However, the groups were heterogeneous with respect to obesity. In males obesity was greater in the two groups of highest alcohol consumption (variance ratio, F,,,,,, = 20.6, p < 0.001). Females in the two groups of lowest alcohol consumption were less obese than those in the 6-8 drinks daily group, and females in the highest consumption group were the least obese (F,,,,,, = 9.51, p < 0.001). Mean systolic and diastolic blood pressure and mean serum creatinine concentrations of the drinkers and non-drinkers in each group are given in Table 2. Average matched pair differences, controlling for the possible confounding effect of smoking, are shown in Table 3. The systolic blood pressure of males who drank more than 9 drinks daily was significantly elevated by 10.6 mmHg (t = 8.1, p < 0.001) compared with their non-drinking age- and obesity-matched controls. For the groups of males who drank 6-8 drinks, 3-5 drinks and 2 or fewer drinks daily, the elevation in systolic blood pressure was 9.9 mmHg (t = 12, p < O.OOl), 5.2 mmHg (t = 9.5,~ < 0.001) and 2.3 mmHg (t = 3.0, 0.001
Daily alcohol consumption of drinkers (drinks)
Drinkers
Females Non-drinkers
Systolic BP (mmb)
<2 3-s 68 .9
122.5 125.6 132.0 124.0
t rt + +
0.7 1.0 3.0 5.0
122.8 123.3 129.0 123.0
+ 0.7 Li:0.9 Lt:3.0 i: 6.0
Diastolic BP (mmH&
$2 3-5 f&8 >9
15.7 77.0 80.6 77.0
* * * *
0.4 0.5 1.8 3.0
75.9 76.1 77.0 79.0
+ i tf 2
Serum creatinine (pm&b
<2 3-5 6-S 39
74.5 + 0.4 72.9 & 0.5 73.1 * 1.9 71.0+4.0
73.3 73.7 74.2 67.0
rt 0.4 rl_0.5 + 1.5 + 4.0
0.4 0.5 1.4 4.0
Drinkers 134.1 137.2 141.1 142.4
* 0.6 i 0.4 * 0.7 +_ 1.1
83.8 85.6 88.4 89.5
f 0.3 * 0.2 IO.4 f 0.7
92.4 t 0.4 91.5kO.3 90.2 k 0.5 86.7 ?r 0.7
Males Non-drinkers 132.1 132.4 132.0 132.9
rtO.6 k 0.4 + 0.6 + 0.9
83. I + 0.4 83.3 + 0.3 83.2 + 0.4 83.8 i 0.6 92.2 91.9 91.8 91.7
+ 0.4 * 0.3 +O.S io.7
620
ELLIOTT SAVDIE el al
Daily alcohol consumption of drinkers (drinks)
I
Females
Males
I
<2 3-5 6-8 29
0.1 3.1 5.0 2.0
+ * * *
0.8 1.1 3.0 7.0
A Systolic BP (mmHg) 0.0 (NS) 2.3 k 0.7 2.9** 5.2 * 0.5 1.7 (NS) 9.9 f 0.8 0.3 (NS) 10.6 +_ 1.3
3.08’ 9.5*** 12.0*** s.1***
$2
0.0 1.6 4.6 - 2.0
* f 5 +
0.5 0.6 1.8 4.0
A Diastolic BP (mmHg) 0.0 (NS) 0.7 * 0.4 2.7** 2.7 + 0.3 2.6* 5.7 * 0.5 0.5 (NS) 6.5 k 0.7
1.8 (NS) 9.0*** 11.4*** 9.3***
3-5 4-a 29
A Serum creatinine (pd/l) t 0.9 + 0.4 2.3* -0.1 kO.4 0.3 (NS) -1.1 +0.5 2.2’ -4.0 f 1.0 4.1.”
<2 3-5 &8 39
***p i
0.001 **p < 0.0 I *p < 0.05 (NS) not significant
compared
with non-drinkers.
In females, however, the differences in blood pressure between drinkers and nondrinkers were much less marked. No differences were evident in females drinking 2 or fewer drinks daily, although both blood pressures were significantly elevated in women drinking 3-5 drinks daily (systolic: 3.1 mmHg, t = 2.9, 0.001 < p < 0.01; diastolic: 1.6 mmHg, t = 2.7, 0.001
Smoking
history
Ex-smoker Light smoker Heavy smoker
***p < 0.001 ‘p < 0.05 (NS) not significant
A Systolic BP (mmHg)
f
-1.4* 1.0 -1.6+0.7 -2.5 f 0.6
1.4 (NS) 2.3’ 4.2”’
compared
with non-smokers
A Diastolic BP (mmHg) -0.1 -0.9 -2.0
+ 0.5 * 0.4 f 0.3
I 0.2 (NS) 2.2’ 6.7*‘*
A Serum creatinine (mow) -0.2 f 0.7 -1.1 kO.5 -1.9io.4
I 0.3 (NS) 2.2* 4.v**
Alcohol, Cigarettes, Blood Pressure and Serum Creatinine
621
cigarette exposure was greater. Because serum creatinine concentrations were measured only to the nearest 10 pmol/l, some caution is justified when interpreting those analyses. However, parallel non-parametric analysis yielded similar results. DISCUSSION
This study confirms the previously reported positive correlation between alcohol consumption and blood pressure. Its size, design and analysis give it good power. The sample of 5500 matched pairs of subjects compared favourably with unmatched crosssectional studies of 2100 [4], 3400 [5], 85 [6], 4800 [7], 20,920 [8] and 84,000 [3] subjects. In our study, the drinkers admitted to drinking alcohol daily or most days. Therefore. we have not addressed the question of the blood pressure and renal function of less frequent drinkers. One factor leading to possible overestimation of the effects of alcohol on blood pressure in this and other studies is that drinkers tend to understate their consumption of alcohol, either consciously or subconsciously. The magnitude of this effect has not been ascertained here. Since the subjects had fasted for at least 12 hours before examination it is possible that the blood pressure of the heavy drinkers was affected by abstinence from alcohol. Sympathetic overactivity, resulting from alcohol withdrawal of even a few hours’ duration, has been advanced as one explanation for the alcohol-blood pressure association in studies such as these [13], but has not been tested. However, 12 hours would have been insufficient time for blood pressure elevated by drinking to have returned toward the “normal” non-drinking level [14, 151. The patients attending the Medicheck Referral Centre are referred by a general practitioner, and are apparently well. They are, however, not a random sample of the general population. For example, attenders tend to be more affluent and better educated than non-attenders [16]. However, bias with respect to sex, age, obesity and admitted smoking habits have been explicitly eliminated by design and analysis in the work reported here;
and
bias
due to other
possible
factors
has been
abated
by the selection
of controls
the same subpopulation. In women, drug and alcohol abuse often occur together [17] so the paucity of females who admitted to drinking 6-8, or 9 or more drinks daily and who were not excluded from this study because of drug use in not surprising and warns of the unusual nature of those heavy drinking females who remained in the study. For example, the heaviest drinking females were much less obese than females in the other groups, whereas drinking and obesity were positively correlated in the males. In this context, the observed discrepancy between males and females of the blood pressure differences between drinkers and non-drinkers is of little value. When the heaviest drinking female group is excluded, there is a ciear dose-response reiationship ‘between aicohoi consumprion and biood pressure. Nevertheless, at least two other groups [3, 51, have shown that the elevation of blood pressure in women by alcohol is inconsistent and less than in males, raising the possibility that there is, in fact, a sexual factor modifying the effect of alcohol on blood pressure. Althouth the blood pressures of both drinkers and non-drinkers in both sexes were correlated with age and obesity, the matched pair differences were not. Therefore, there was no effect modification by age or obesity. The variances of these differences were homogeneous with obesity and age, except for that of the difference in systolic blood pressure which was positively correlated with age. This presumably is a consequence of the greater variance in systolic blood pressure shown by older subjects, whether drinkers or non-drinkers. However, this will not affect the conclusions previously drawn because the mean ages are similar for subjects from each group for both males and females. The work presented here supports that of Cooke et al. [8] that, at least in males, small but significant elevations in blood pressure are seen in those with even very modest alcohol consumption (2 or less drinks per day), arguing against the proposition that there is a “threshold” level of alcohol intake above which elevation will occur [3]. At the other end of the scale, our data showed that those males drinking 9 or more alcoholic units per day from
622
ELLIOTT SAVDIE et al.
had blood pressure elevations only slightly greater than those imbibing 68 units per day thus supporting the notion that there is a “maximal” effect of alcohol on blood pressure when about 9 units per day are ingested. Cooke et al., described a plateau at a similar level nf “I
onnrnmnt;r\n ~“LI~UI‘I~LI”LI
nf “I
Qhr\,.t LL”“UL
LCI ,,n;tr ,.P,. “” U‘AILO ybL
..mz.lr r*tAd*
IQ1 L”,.
This study also suggests that alcohol causes a greater elevation in systolic than in diastolic blood pressures supporting the observations of others [3, 5, 81. It has been shown in alcoholics that blood pressure is correlated with their mean daily alcohol intake over the preceding three months, and falls to normal after detoxification [ 141. Furthermore, the blood pressure will remain normal in those who continue to abstain, but rises again in those who recommence drinking. These data suggest that the effect of alcohol on blood pressure, though sustained, is reversible and not dependent on underlying pathological damage to the liver. They are supported by our data and those of others [3,8] showing a significant rise in blood pressure with an intake of alcohol in quantities well short of that which would be expected to cause liver disease. Klatsky et al. [3] hypothesized that “stress” or other psychosocial factors might be responsible for both high alcohol consumption on the one hand and high blood pressure on the other, but the fall in blood pressure in the detoxified alcoholics studied by Saunders [14] argues against such a postulate. It seems more likely that the hypertensive effect of ethanol is due to a haemodynamic effect on the circulation. Although some investigators have reported small increases in both systolic and diastolic blood pressures following the acute administration of ethanol [ 181, others have not [ 191. In the recovery phase from acute alcohol intoxication, renin and aldosterone production are stimulated [IO]. It is possible, therefore, that acute salt and water retention or the vasoconstrictor effects of angiotensin II could be responsible for the hypertensive effects of alcohol. The inverse relationship between alcohol consumption and serum creatinine concentration shown here, while statistically significant, is slight and not of clinical significance. Serum creatinine levels depend on creatinine production from the muscle mass, and renal clearance in this study we are reluctant to postulate that alcohol enhances creatinine clearance which is entirely due to glomerular filtration. Since we have not measured the creatinine clearance in this study we are reluctant to postulate that alcohol enhances creatinine clearance by the kidney, especially since studies in the rat have shown alcohol to have an adverse effect on both renal structure and function, leading to an increase in serum creatinine concentration [I 11. It is possible that the effect of alcohol on serum creatinine concentration observed here is a dilution phenomenon due to the retention of salt and water as a result of the perturbations in hormone levels mentioned above; but reported increases in blood volume in alcoholics are relatively small (approximately 5%) [20]. The explanation which we favour more is that subjects with a heavy alcohol intake become malnourished and develop muscle wasting and sometimes frank myopathy [21]. ,a:“+.,,L,.l,, ,.-..,A :.. t.,..r\ ..,,.,l+ ;.. A;m;..;nhcJ ,...-n+;*;s.3 . . . ..A fiA -..,,.1, 111~~~1~UISCUI ULIIILGGUU~U, 111 LULL,,IGJUIL 111a UIIIIILIIJIIGU ~LG~CIIIIIIC; p~uuiiCtiGfi ZiIb 2 decrease in the serum creatinine level independent of any change in plasma volume or glomerular filtration rate. Further studies involving measurement of both creatinine clearance and lead body mass would be needed to clearly delineate any effect of alcohol on renal function in man. The association of cigarette smoking with small decrements in blood pressure has been reported before [3,8,22], and appears to be independent of other variables. Because this effect is opposite to that of alcohol, and since alcohol and cigarette consumption are positively associated, cigarette consumption needs to be controlled for. Since nicotine is known to cause an elevation in blood pressure in acute experimental situations [23,24], the association of chronic smoking with decreased blood pressure is interesting and suggests the existence in cigarette smoke of substances other than nicotine which have an opposite, hypotensive action. The mechanism of the tiny decrease in serum creatinine concentrations caused by smoking is unclear. Acknowledgements-We would like to thank Dr H. G. Gallagher, Referral Centre, for assistance with this study.
formerly
Medical
Director
of the Medicheck
Alcohol,
Cigarettes,
Blood
Pressure
and Serum
Creatinine
623
REFERENCES 1. 2. 3. 4.
5. 6. 7.
8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24.
Clark VA. Chaoman JM, Coulson AH: Effects of various factors on systolic and diastolic blood pressure in the Los Angkles Heart Study. J Chron Dis 20: 571-581, 1967 Kannel WB, Sorlie P: Hypertension in Framingham. In Epidemiology and Control of Hypertension. Paul 0 (Ed) New York: Stratton Intercontinental Medical Book Corporation, 1974. pp 553-592 Klatsky AL, Friedman GD, Siegelaub AB, Gerard MJ: Alcohol consumption and blood pressure: Kaiser-Permanente Multiohasic Health Examination data. N Engl J Med 296: 1194-1200. 1977 Dyer AR, Stamler J, Paul 0, Berkson DM, Leeper MH, McKean-H, Shekelle RB, Lindberg HA, Garside D: Alcohol consumption, cardiovascular risk factors and mortality in two Chicago epidemiologic studies. Circulation 56: 1067-1074, 1977 Harburg E, Ozgoren F, Hawthorne VM, Schork MA: Community norms of alcohol usage and blood pressure: Tecumseh, Michigan. Am J Pub1 Hlth 70: 813-820, 1980 Mitchell PI, Morgan MJ, Boadle DJ, Batt JE, Marstrand JL, McNeil HP, Middleton C, Rayner K, Lickiss JN: Role of alcohol in the aetiology of hypertension. Med J Aust 2: 198-200, 1980 Criqui MH, Wallace RB, Mishkel MA, Barrett-Connor E, Heiss G: Alcohol Consumption and Systolic Blood Pressure--the Lipid Research Clinics Prevalence Study. Society for Epidemiological Research, 1980 (abstract) Cooke KM, Frost GW, Thorneil IR, Stokes GS: Alcohol consumption and blood pressure. Survey of the relationship at a health-screening clinic. Med J Aust 1: 65-69, 1982 Hauger-Kievene JH, Pinkas MBrDegrossi OJ: Paradoxical effect of ethanol on plasma renin activity levels in hyper-reninemic essentiai hypertension. Medicina Buenos Aires 39: 6i4-6i 8, i979 Linkola J, Fyhrquist F, Ylikahri R: Renin, aldosterone and cortisol during ethanol intoxication and hangover. Acta Physiol Stand 106: 75-82, 1979 van Thiel DH, Gavaler JS, Little JM, Lester R: Alcohol: its effect on the kidney. Metabolism 26: 857-866, 1977 Bradley RA, Terry ME: Rank analysis of incomplete block designs. I. The method of paired comparisons. Biometrika 39: 324-345, 1952 Dolinsky H. Letter to the Editor. N Engl J Med 297: 450, 1977 Saunders JB, Beevers DG, Paton A: Alcohol-induced hypertension. Lancet ii: 653-656, 1981 Ashley MJ, Rankin JG: Alcohol consumption and hypertension--the evidence for hazardous drinking and aicohoi popuiations. Aust NZ J Meri 9: 201-206, i979 Rawson G: Patient attitudes to screening by an automated multiphasic health testing facility. Aust Fam Phys 4: 219-222, 1975 Reynolds I, Harnas J, Gallagher H, Bryden D: Drinking and drug taking patterns of 8,516 adults in Sydney. Med J Aust 2: 782-785, 1976 Orlando J, Aronow NS, Cassidy J, et al: Effect of ethanol on angina pectoris. Ann Int Med 84: 652-655, 1976 Conway N: Haemodynamic effects of ethyl alcohol in patients with coronary heart disease. Br Heart J 30: 638-644, 1968. Beard JD. Knott DH: Fluid and electrolyte balance during acute withdrawal in chronic alcoholic patients. JAMA 204: 135-139, 1968 Knochel JP, Bilbrey GL, Fuller TJ, Carter NW: The muscle cell in chronic alcoholism: the possible role of phosphate depletion in alcoholic myopathy. Ann NY Acad Sci 252: 274-286, 1975 Arkwright PD, Beilin LJ, Rouse I, Armstrong B, Vandongen R: Alcohol and blood pressure in a working population. Clin Exp Pharmac Physiol 8: 451-454, 1981 Gebber GL: Neuroaenic basis for the rise in blood pressure evoked by nicotine in the cat. J Pharmac Exp Ther 166: 255-269,-1969 Roth GM, Shick RM: Effect of smoking on the cardiovascular system of man. Circulation 17: 443-459, 1958