JAGS 39~260-164,1997

The Effect of Caffeine on Postprandial HvPotension in the Elderlv J I

J

D. Heseltine, MRCP,* M. Dakkak, MRCP," K. Woodhouse, MD,t 1. A. Macdonald, PhD,S and J. F. Potter, DMS

In a double-blind, randomized trial the effects of caffeinated and decaffeinated drinks on postprandial hemodynamic and neurohumoral changes were studied in seven fit, elderly subjects after a standard 2.4MJ meal. There was a significant difference in supine postprandial systolic blood pressure between the placebo and caffeine phases (P < 0.01); at 60 minutes, supine systolic blood pressure had fallen 14 mmHg [95% confidence interval (CI) -7 to -21 mmHg, p < 0.01) after placebo, but was unchanged after caffeine (+9 mmHg, CI 0 to 18 mmHg, NS]. Similar differences between placebo and caffeine

ostprandial blood pressure falls in fit and frail elderly subjects,'s2 the greatest postprandial changes occurring after a high carbohydrate A content meal.3 The mechanisms responsible for these changes are not fully understood, although glucose-mediatedimpairment of baroreflex sensitivity4 and increased splanchnic blood flow mediated by insulin, vasoactive gastrointestinal hormone^,^ or adenosine,6 with a fall in systemic vascular resistance, may contribute to it. Acute caffeine ingestion is known to increase blood especially in the elderly? and to prevent the postprandial fall in blood pressure in patients with autonomic failure." However, in healthy elderly subjects, caffeine given 60 minutes before food failed to prevent the postprandial fall in blood pressure." The conflicting results regarding the effect of caffeine on postprandial blood pressure may relate to the timing of caffeine intake in relation to meal ingestion. Thus, the present study was designed to assess the effects of

P

From the Departments of *Geriatric Medicine, Kingston General Hospital, Hull and tRoyal Victoria Infirmary, Newcastle upon Tyne; the $Departmentof Physiology and Pharmacology, Queens Medical Centre, Nottingham; and §University Department of Medicine for the Elderly, Leicester General Hospital, Leicester, UK. Address correspondence to J. F. Potter, D.M. M. R. C. ,.'l University Department of Medicine for the Elderly, Leicester General Hospital, Leicester LE5 4PW, UK.

1991 by the American Geriatrics Society

were seen in erect systolic and diastolic blood pressure (P

c 0.01), although orthostatic tolerance was maintained throughout each study period. Postprandial plasma noradrenaline levels were higher (P < 0.02) and the increase greater (P < 0.02) after caffeine than after placebo. Caffeine administered at the end of a standard test meal prevents the postprandial fall in blood pressure in fit, elderly subjects. The clinical relevance of this finding has yet to be determined, but it may offer a simple remedy for patients with symptomatic postprandial hypotension. J Am Geriat SOC39:160-164,1991

caffeine, given immediately after completion of a test meal, on the postprandial blood pressure changes in healthy elderly subjects.

MATERIALS AND METHODS Subjects. Seven fit, independent subjects (five women and two men) drawn from local pensioners clubs participated in the study (mean age 67.4 years, range 64 to 72 years). All were within 10% of their ideal body weight [body mass index 24.4 kg/m2 2 3.1 (SD)]and were selected according to the following criteria: normotensive [systolic blood pressure (SBP) < 160 mmHg and diastolic blood pressure (DBP) < 9 0 mmHg, checked on at least two separate occasions prior to the study], no evidence of postural hypotension (fall in SBP > 20 mmHg), and no history of symptomatic postprandial hypotension, ischaemic heart disease, cardiac failure, cerebrovascular disease, diabetes mellitus, chronic respiratory disease, malignancy or mental deterioration. No subject was taking any medication and all were non-smokers who were accustomed to a mean daily intake of 447 -C 48 mg (range 335 to 700 mg) of caffeine, ingested as coffee and/or tea (no chocolate or carbonated cola drinks; caffeine content of instant coffee taken as 65 mg/cup, and that of tea, 45 mg/cup12). All subjects gave informed consent, and the study was approved by the local Health District Ethical Committee. 10002-8614/91/$3.50

JAGS-FEBRUARY 1991-VOL. 39, NO. 2

CAFFEINE PREVENTS POSTPRANDIAL HYPOTENSION

Methods Each subject was studied on two occasions, one week apart. After overnight abstinence from alcohol and caffeine, subjects were allowed a light breakfast before arriving for the study at 11:30 a.m. A full medical history was obtained and a physical examination, along with an EKG, was performed on all subjects prior to entry into the study. The subjects then rested supine, and a 19 gauge butterfly needle was placed into a dorsal hand vein to administer heparin sodium saline (Hepsal CP Pharmaceuticals Clwyd, Wales). Arterialed blood samples were obtained by warming the hand to 55 OC in a thermostatically controlled box, all blood samples were taken without hemostasis. Blood pressures were measured using a Hawksley Random Zero Sphygmomanometer (DBP phase V), and the mean of two recordings, both supine and after 1 minute of standing unassisted, was recorded. A standard electrocardiograph recorded heart rate in both supine and erect positions. Autonomic function was assessed before entry into the study using standard methods described by Ewing and Clarke.13 Heart rate variation (R-R interval) was measured with the paper recording EKG set at 5 cm/s during deep breathing timed at 6 cycles/min (taking the mean of two maximum expiratory/minimum inspiratory R-R ratios). At 1:OO p.m., subjects consumed a standard 2.4MJ meal (585 kcal) consisting of roast chicken, boiled potatoes and carrots, banana, custard powder, skimmed milk and sucrose (71% carbohydrate, 8% fat, 21% protein). Meals were eaten over 15 minutes, and on completion, subjects were given either caffeinated coffee (Nescafe Blend 37), weighed to contain 200 mg of caffeine, or decaffeinated coffee (Nescafe Gold Blend Decaffeinated), dissolved in 100 ml of water in a double-blind, random order, crossover fashion. Before and after the meal, subjects rested supine. Supine blood pressure and heart rate recordings were taken at 15-minute intervals for 30 minutes preprandially, and up to 60 minutes postprandially, the final reading being taken 90 minutes after completion of the meal and the caffeinated or decaffeinated drink. Erect blood pressure and heart rate recordings were taken immediately preprandially and at 30, 60 and 90 minutes postprandially. Autonomic function and blood samples were assessed before the meal (T = 0) and 60 minutes postprandially (T = 60), the time of the maximum post-meal fall in blood pressure as found in a previous study using a similar protoc01.~Blood samples were taken after 60 minutes of supine rest for determination of serum osmolality and packed cell volume (PCV), blood glucose, plasma caffeine, catecholamines and insulin levels. Assay Methods. Plasma osmolality was assessed by an automatic micro-osmometer (CAMLAB), blood hematocrit by Coulter counter, and blood glucose by

161

the glucose oxidase method (CAMLAB). Plasma caffeine was measured by high performance liquid chromatography (HPLC) [interassaycoefficient of variation (CV) 5%], plasma catecholamines by HPLC with electrochemical detection (interassay CV noradrenaline 8.4% and adrenaline 13.9%),14and insulin by a radioimmunoassay (interassay CV 7%).”

Statistics All results are expressed as mean +- SEM, and 95% confidence intervals (CIS) are also given where indicated. The differences between blood pressure and heart rate responses following caffeine and placebo were analyzed by calculatingthe area under the curve (AUC) for that parameter (by the trapezoid method) divided by the time, which gave an overall summary statistic for each subject for that study period; these area/time values for each subject were corrected for baseline values, and the differences following caffeine and placebo were compared using the Wilcoxon signed rank test.16 If an overall significant difference was found, differences between individual time points were compared by paired Student’s t-tests. Significance values were taken at P < 0.025 to allow for multiple comparisons. Pearson’s correlation coefficient was also calculated. RESULTS All subjects consumed the whole meal on both occasions along with the caffeinated or decaffeinated drink, and none experienced any adverse symptoms throughout the study.

Hemodynamic Effects The mean postprandial changes in supine SBP, DBP and heart rate with placebo and caffeine are shown in Figure 1 and the individual changes in supine SBP are shown in Figure 2. There was a significant difference in supine SBP response postprandially between placebo and caffeine phases over the 90-minute study period ( P C 0.01). After the non-caffeinated drink, supine SBP fell in six subjects and showed no change in the other subject (see Figure 2); the maximal fall in SBP occurred 60 minutes after the meal (-14 mmHg, 95% CI = -7 to -21 mmHg, p < 0.01). After caffeine, SBP increased in five subjects and showed no change in two subjects (Figure 2), with maximal changes again occurring 60 minutes after the meal (+9 mmHg, 0 to 18 mmHg, NS). There was no difference in the changes in supine DBP and heart rate between the two phases. The changes in erect SBP, DBP and heart rate with placebo and caffeine are shown in Figure 3 and individual changes in erect SBP in Figure 2. There was a significant difference in erect SBP between the placebo and caffeine treatment (P < 0.01). Erect SBP fell in five subjects and showed no change in two subjects, the

162

JAGS-FEBRUARY 1991-VOL. 39, NO. 2

HESELTINE ET AL

-

MEAL

I

Caffeine Placebo

A

SUPINE SBP(mmHg)

A ERECT SBP(mmHg)

130 pc 0.01 0

110 -

I1 0 0 90 -I NS

n m

60 J

0

NS

I

l

l

-30-15 0

a 1

1

1

1

% 15 30 45 60

I

90

TIME (min)

FIGURE 1. Changes in supine systolic blood pressure (SBP), diastolic blood pressure (DBP) and heart rate (HR) during the 30-minute run-in period and for 90 minutes after the high carbohydrate test meal with placebo (A)and caffeine (0)in seven fit, normotensive subjects. Values are means f SEM.

maximal fall occurring 90 minutes after the meal with placebo (-17 mmHg, -5 to -20 mmHg, P C 0.02), whereas after caffeine, erect SBP increased in six subjects and showed no changes in one subject, maximal changes occurring at 60 minutes postprandially (+12 mmHg, +4 to +20 mmHg, P C 0.02). There was also a significant difference in the postprandial changes in erect DBP between placebo and caffeine groups (P < 0.05, see Figure 3). After placebo, the greatest fall in erected BP occurred 60 minutes postprandially (-10 mmHg, CI -2 to -18 mmHg, P < 0.01), whereas after caffeine, DBP was unchanged (+2 mmHg CI -6 to +10 mmHg, NS). There was no significant change in erect heart rate postprandially .during either phase. Before the meal there was no orthostatic difference in SBP between placebo and caffeine treatment groups (-1 f 2 mmHg and -3 +. 2 mmHg, respectively), and postprandially, no postural differences in SBP occurred individually or in the group as a whole. There was no relationship between maximum SBP changes with caffeine or placebo and subjects' mean daily consumption of caffeine. Postprandially, no postural differences in DBP o r heart rate developed in either group.

FIGURE 2. Indlvidual postprandial changes from baseline in supine and erect systolic blood pressure (SBP) after placebo (A)and caffeine (0).Values in parentheses indicate the subjects' usual daily caffeine intake.

phase. Plasma osmolality and PCV at baseline and postprandially were similar in the placebo and caffeine phases (Table 1).

Blood Glucose and Plasma Insulin No differences in baseline blood glucose or plasma insulin levels were found between the two phases. As expected, a significant increase (P < 0.01) in blood glucose occurred after the meals, although values at 60 minutes were similar on both occasions (Table 1).The postprandial increase in plasma insulin levels was not influenced by caffeine consumption (Table 1).

Autonomic Function Parasympathetic function was assessed using the standard method of measuring Plasma Caffeine, Osmolality and PCV Changes the expiratory and inspiratory R-R ratio,l3and baseline Plasma caffeine levels postprandially increased from values were within the normal standardized age limits 0.1 f 0.1 (not recorded) in five subjects) to 2.6 mg/L in all subject^.'^ There was no significant difference in at 60 minutes after the caffeine phase, whereas caffeine the R-R ratios postprandially following the two phases levels were 0.1 f 0.1 mg/L (not recorded in five sub- of the study. Plasma noradrenaline levels were similar jects) and remained unchanged after the decaffeinated in the two phases at time 0, but a significant postpran-

JAGS-FEBRUARY 1991-VOL. 39, NO, 2

-1

P $

130 120 110

-

Caffeine Placebo

I..M-EA:-.*-

I'

100

90

4

a m n

6o-j

u

65 60

I

CAFFEINE PREVENTS POSTPRANDIAL HYPOTENSION

T

T

TIME (min)

FIGURE 3. Changes in postprandial erect SBP, DBP and heart rate after placebo (A)and caffeine (0).Values are means f SEM.

TABLE 1. BASELINE AND 60-MINUTE POSTPRANDIAL VALUES FOR THE HEMODYNAMIC AND NEUROHUMORAL CHANGES AFTER PLACEBO AND CAFFEINE Preprandial Poshmandial Expiration/ Inspiration R-R ratio Packed cell volume Osmolality (mosmol/L) Glucose (mmol/L) Insulin (m-units/L) Noradrenaline (nmol/L) Adrenaline (nmol/L)

Placebo Caffeine

mean k SEM 1.26 f 0.10 1.27 f 0.07 1.31 & 0.12 1.27 f 0.11

Placebo 0.388 f 0.012 0.385 f 0.014 Caffeine 0.402 f 0.021 0.401 ?z 0.021 293 f 1 298 f 1 Placebo 296 & 1 Caffeine 291 f 2 4.5 f 0.1 8.3 ?z 0.6** Placebo 4.5 f 0.1 7.9 f 0.7** Caffeine 5.6 k 1.3 50.3 f 5.6** Placebo 7.0 & 2.7 55.8 f 10.7** Caffeine 2.63 f 0.31 Placebo 2.09 f 0.16 3.89 k 0.52* Caffeine 2.21 f 0.19 0.17 f 0.04 Placebo 0.15 f 0.04 0.15 f 0.04 Caffeine 0.12 & 0.03

Significance values for comparisons between pre- and postprandial values. * P < 0.02. ** P < 0.01.

dial increase was only seen after caffeine (1.68 nmol/ L, 0.84 to 2.52 nmol/L, P < 0.02). Postprandial noradrenaline levels were higher after caffeine than after the non-caffeinated phase (1.25 nmol/L, 0.24 to 2.26 nmol/L, P < 0.02; see Table 1). Plasma adrenaline levels were unchanged throughout the study.

163

DISCUSSION The main findings of this double-blind, randomized study were that caffeine administered immediately after a high carbohydrate meal prevented the postprandial fall in supine and erect blood pressure in fit, elderly patients. This was associated with a postprandial increase in sympathetic nervous system (SNS) activity (as gauged by changes in plasma noradrenaline levels) but not in heart rate. No patient exhibited orthostatic hypotension during either phase of the study, and all remained asymptomatic despite the postprandial fall in blood pressure during the no-caffeine phase. Caffeine is known to increase blood pressure, the pressor effect being more marked with increasing age.' In patients with autonomic failure, caffeine administered 30 minutes before a test meal prevented the postprandial blood pressure fall.",' However, 250 mg of caffeine given 60 minutes preprandially to fit elderly subjects failed to attenuate the postprandial fall in blood pressure," a finding that contrasts with our results. This difference may be explained simply by the timing of caffeine administration, peak plasma caffeine levels and the resulting maximal pressor effect occuring 60 minutes post-meal in our study," the time of the maximal postprandial blood pressure fall.3Presumably, caffeine is only likely to antagonize the hypotensive effect of the meal when ingested with or immediately after the meal. Subjects in this present study were given only 200 mg caffeine (corresponding to 3 cups of coffee"). Although plasma caffeine levels were not as high as in previous studies8,",20and all subjects were regular caffeine users (and thus may have been partially tolerant to the pressor effect of caffeine), the postprandial fall in blood pressure was still prevented. It has been suggested that the pressor effect of caffeine is mediated by Stimulation of the SNS.8*'9-21 This is supported by the present observation of a greater postprandial increase in plasma NA levels after caffeine than after placebo, although it must be acknowledged that plasma noradrenaline levels are only an indirect index of SNS activity. However, in patients with autonomic failure and an impaired SNS, caffeine still prevented the postprandial blood pressure Furthermore, it has been suggested that the observed increases in plasma catecholamines are unrelated to the pressor effect of Alternatively, caffeine may exert its pressor effect by activation of the renin-angiotensin system," but no study has demonstrated a close association between the caffeine-induced changes in plasma renin activity and blood pressure. It is also unlikely that caffeine prevents postprandial hypotension by its actions on cardiac autonomic function, plasma volume, glucose or insulin levels, as no differences were observed in any

164

HESELTINE ET AL

JAGS-FEBRUARY 1991-VOL 39, NO. 2

10. Onrot J, Goldberg MR, Biaggioni I, et aI. Hemodynamic and humoral effects ‘ of caffeine in autonomic failure: therapeutic implications for postprandial hypotension. N Engl J Med 1988; 313:549. 11. Lenders JWM, Morre HLC, Smits P, et al. The effects of caffeine on the postprandial fall of blood pressure in the elderly. Age Ageing 1988;17936. 12. Bunker ML, McWilliams M: Caffeine content of common beverages. J Am Diet Assoc 1979;7428. 13. Ewing DJ, Clarke BF: Diagnosis and management of diabetic autonomic neuropathy. Br Med J 1982;295:916. 14. Macdonald IA, Lake DM. An improved technique for extracling catecholamines from body fluids. J Neurosd Methods 1985; 13:239. 15. Soeldner JS,Slone D. Critical variables in the radioimmunoasay of serum insulin using double antibody technic. Diabetes 1965; 14:771. 16. Matthews J, Altman DE, Campben M, et al. AnaIysis of serial measurement in medical research. Br Med J 1990;300:230. 17. Clark CV, Mapstone R. Age adjusted normal tolerancelint& for cardiovascular autonomic function assessment in the elderly. Age Ageine 1986;15221. 18. Smits P, Thien Th, Van‘t Laar A Circulatory effects of cdfee in relation to the pharmacokinetics of caffeine. Am J Cardid 1985; 56:958. 19. Smits P, Hoffman H,Thien Th, et aI. Hemodynamicand humoral effects of coffee after Beta-1-selective and non-selective Beta blockade. Clin Pharmacol Ther 1983;34153. 20. Robertson D, Wade D, Workman R, et al. Tolerance to the humoral and hemodynamic effects of caffeine in man. J Clin Invest 1981;67:1111. 21. Smits P, Pieters G, Thien Th. The role of epinephrine m the REFERENCES circulatory effects of coffee. Clin Pharmacol Ther 1986;a 4 3 1 . 22. Beavo JA, Rogers NL, Crofford 0 8 , et al. Effects of xanthiie 1. Lipsitz LA, Nyquist RP, Wei JY, et al. Postprandial reduction in derivatives on lipolysis and on adenosine 3’,5’-monophosphate blood pressure in the elderly. N Engl J Med 1983;309:81. phosphodiesterase activity. Mol Pharmacol 1970;6:597. 2. Westenend M, Lenders JWM, Thien Th. The course of blood pressure after a meal; a difference between young and elderly 23. Fredholm BS: Are methylxanthine effects clue to the a n t a p i s m of endogenous adenosine? Trends Pharmacol Sci 1980;1:129. subjects. J Hypertension 1985; 3 (suppl 3):s417. 3. Potter JF, Heseltine D, Matthews J, et al. Effects of meal com- 24. Bruns RF, Daly JW, Snyder SH. Adenosine receptors in brain membranes binding of N6-cyclohexy [3HJ adenosine and 1,3position on the postprandal blood pressure, catecholamine and diethyl-8-[3H]phenylxanthine.Proc Natl Acad S a USA 1980; insulin changes in elderly subjects. Clin Sci 1989; 77:226. 775547. 4. Appenzeller 0,Goss JE. Glucose and baroreceptor function: effects of oral administration of glucose on baroreceptor function 25. Robertson D, Hollister AS, Kincaid D, et al. Caffeine and hypertension. Am J Med 1984; 7754. in cerebrovascular disease and in other disorders with barore26. Berne RM, Knutt RM, Ely SW, et al. Adenosine in the local ceptor reflex block. Arch Neurol 1970; 23:137. regulation of blood flow: a brief overview. Fed Prac 1983; 5. Minaker KL, Rowe JW, Young JB, et al. Effect of age on insulin 423136. stimulation of sympathetic nervous system activity in man. Me27. Granger HI, Noms SCP. Role of adenosine in local confx~lof tabolism 1982;31:1181. intestinal circulation in the dog. Circ Res 1980;46:764. 6. Granger DN, Valleau JD, Parker RE, et al. Effects of adenosine on intestinal hemodynamics, oxygen delivery, and capillary fluid 28. Scholthert J, Lochner W, Renn H, et al. Die wirking von Noradrenalin, Adrenalin, Isoprotenarol and Adenosin auf die Durchexchange. Am J Physiol 1978;235:H707. blutung der heber und des Splanchniausgebietes des Hundes. 7. Smits P, Thien Th, Van’t Laar A. The cardiovascular effects of Pflugers Arch 1967;293:129. regular and decaffeinated coffee. Br J Clin Phannacol 1985; 29. Lagerkranser M, Irestedt L, Sollevi A, et al. Central and splanch19:852, nic hemodynamics in the dog during controlled hypotension 8. Robertson D,Frohlich JC, Cam RK: Effects of caffeine on plasma with adenosine. Anesthesiology 1984; 60:547. renin activity, catecholamines and blood pressure. N Engl J Med 30. Sollevi A, Lagerkranser M, Irestedt L, et al. Controlled hypoten1978;298:181. sion with adenosine in cerebral aneurysm surgery. Anes?hesiol9. Izzo JL, Ghosal A, Kwong T, et al: Age and prior caffeine use ogy 1984;61:400. alter the cardiovascular and adreno-medullary response to oral caffeine. Am J Cardiol 1983;52:769.

of these parameters between the placebo and caffeine phases. Two of the main pharmacological effects of caffeine are inhibition of cyclic AMP phosphodiesterase2’ and antagonism of adenosine receptors.23p24 The plasma concentrations of caffeine that raise blood pressure seem to be below the threshold for inhibition of phosph~diesterase,’~ but within the range for antagonism of adenosine receptors.26Adenosine is a potent vasodilator,26and it can increase splanchnic and hepatic blood flow in dogs27-29and lower blood pressure in man.3o Thus, it is possible that caffeine, by blocking the vasodilatory adenosine receptors, prevents splanchnic vasodilation and attenuates the postprandial fall in blood pressure. In conclusion, in healthy elderly subjects without symptoms of postprandial hypotension, caffeine administered at the end of a high carbohydrate test meal prevented the postprandial fall in blood pressure. Future studies should be directed to determine whether the attenuation of a postprandial fall in blood pressure by caffeine can be reproduced in frail subjects with symptomatic post-meal hypotension.