- Email: [email protected]
potassium interventions
PREVENTIVE
MEDICINE
13, 233-244
(1984)
Dietary Approaches to the Reduction of Blood Pressure: The Independence of Weight and Sodium/Potassium Interventions’ RENA R. WING,*-* ARLENE W. CAGGIULA,+ MARY PATRICIA NOWALK,* RANDI KOESKE,* SEUNG LEE,* AND HERBERT LANGFORDS *University of Pittsburgh School of Medicine, and tGraduate School of Public of Pittsburgh, Pittsburgh, Pennsylvania 15213: and #University of Mississippi Jackson, Mississippi 39216
Health. Medical
University Center,
This study was designed to determine the feasibility of teaching mildly hypertensive individuals to select a diet, using normally available food products, that either would produce a 5% reduction in percentage of overweight (without altering sodium (Na), potassium (K), or Na:K ratio) or would decrease Na to <70 mEq and increase K to >I00 mEq (without affecting weight) and to compare the resulting changes in blood pressure. Fifty-two participants with mild hypertension were randomly assigned to either a weight-loss or a &a:K intervention. Blood pressure, weight, 3-day diaries, and 24-h urinary excretion of Na and K were measured before and after an I-week intervention. Participants in the weight-loss intervention had significantly greater changes in weight and calorie intake than those in the Na:K intervention, while changes in Na:K ratio were greatest in the intervention targeted for that change. The percentage of participants who were able to meet the dietary goals is presented and the implications of these data for the selection of dietary goals are discussed.
INTRODUCTION While estimates of the prevalence of hypertension vary, it is clear that hypertension is a risk for premature cardiovascular disease for millions of Americans (4). The basic dilemma facing the medical and public health community today regards how best to treat these individuals, particularly those with mild hypertension. Recent clinical trials have found that treating mild hypertension reduces mortality (2), but there is concern about placing large portions of the population on drug therapy, since drug therapy is costly and may have risks of its own (9). Consequently, attention has focused on nonpharmacological approaches to blood pressure reduction, such as relaxation and diet modification. There are several different types of nutrition intervention which may be effective as the sole therapy for mild hypertension or as an adjunct to drug therapy: moderate sodium (Na) restriction (1,5,7,8, 1I), increases in potassium (K) intake (6), changes in the Na:K ratio (IO, 14, 15), and weight reduction (12, 16). While there is evidence that nutrition interventions reduce blood pressure, there are very few studies comparing the effectiveness of different types of nutrition intervention. Such comparisons may be difficult, because the various aspects of the diet are interrelated. For example, changes may occur simultaneously ’ The research reported in this paper was supported in part by Grant 5 ROI HL24369-03 from the National Institutes of Health to Dr. Langford. ’ To whom requests for reprints should be addressed at: Western Psychiatric Institute and Clinic, 381I O’Hara Street, Pittsburgh, Pa 15213. 233
0091-7435/84 Copyright All nghls
$3.00
G IY84 by Acadcmlc Press. Inc. of rcproduclion in any form reserved.
234
WING
ET AL.
in both caloric intake and in the sodium:potassium content of the diet, making it difficult to assess the effect of either change independently. A few studies have attempted to keep the changes separate by providing food to the participants (14, placing participants on a low-calorie liquid formula regimen with known sodium concentration (16), encouraging participants in a weight-reduction program to eat salty foods (121, or adding a potassium supplement to a normal diet (6). However, it is not known whether these dietary changes can be made independently when individuals self-select their diet. Further, in some of these studies body weight and electrolyte excretion were not measured both prior to and following the dietary intervention to confirm the independence of the changes. The purpose of the present study was to determine whether it is possible to modify weight and sodium:potassium ratio independently by dietary intervention in free-living adults. The goal was to teach hypertensive individuals to self-select a diet, using normally available food products, that either would produce a 5% reduction in body weight without altering sodium, potassium, or sodium:potassium ratio or would decrease sodium to 70 mEq and increase potassium to 100 mEq without affecting weight. The study was conducted as a feasibility test of the intervention program designed for the Dietary Intervention Study of Hypertension (DISH). METHOD Subjects
Participants were recruited from the Pittsburgh area by radio and newspaper advertisements. Eligibility was based on blood pressure measurements taken at two screening visits scheduled one week apart. Blood pressure measurements were made by a nurse, certified according to standardized procedures from the Multiple Risk Factor Intervention Trial (MRFIT) protocol. A random zero sphygmomanometer was used. Three blood pressure readings were taken with the participant seated after 5-10 min of rest. The first reading was disregarded and the second and third averaged to form the blood pressure measurement. Appropriately sized cuffs were used to account for differences in arm circumference. The baseline blood pressure was the average of the blood pressure readings across the two screening sessions. Individuals whose average diastolic blood pressure (DBP) over the two sessions was a85 or who were being treated with anti-hypertensive medication and had DBP 380 were accepted into the study. Fifty-two individuals met these eligibility criteria and were randomly assigned to an intervention, 27 to the weight-control intervention and 25 to the Na:K intervention. Procedures
Eligible participants were randomly assigned to one of two intervention conditions: Weight Loss or Na:K Change. The goal of the Na:K intervention was to decrease sodium intake to less than 70 mEq/day and increase potassium intake to greater than 100 mEq/day, while keeping body weight constant. The weight intervention was designed to produce a 5% reduction in body weight with no change in Na, K, or Na:K ratio.
DIETARY
INTERVENTIONS
FOR THE REDUCTION
OF BP
235
Both interventions were conducted in a group format, with 10 to 14 participants per group. Meetings were held once a week for 8 weeks. At the start of each meeting, patients were weighed individually and a team of four nutritionists reviewed food diaries. This was followed by the presentation and discussion of a new strategy for changing sodium and potassium intake or reducing weight. The discussion was led by a nutritionist and a behavior therapist. At the first meeting, participants were given a booklet from which they could determine the sodium and potassium content of common foods or the calorie content of these foods. They were instructed to monitor their intake aiming to achieve the Na:K goals (~70 mEq Na and >lOO mEq K) or an individualized calorie goal (based on entry weight x 12 - 1,000 calories/day, with a minimum calorie goal of 1,000 calories per day). The Na:K intervention focused first on decreasing sodium by eliminating table salt, using alternative seasonings, and modifying food choices. The goal of increasing potassium by eating foods high in this mineral was added by Week 4. The weight-control intervention emphasized reducing consumption of high-fat and other high-calorie foods and increasing caloric expenditure through walking. Measures
Procedures for blood pressure measurement are described above. The same procedures were used to obtain blood pressure measurements at Weeks 4 and 8 of the intervention. Weight was assessed at both screening meetings and at each week of intervention using a balance-beam scale. Participants were weighed wearing indoor clothing and without shoes. Baseline weight was the average weight across the two screening sessions. Height was determined at the second screening. Percentage of overweight was computed using the Metropolitan Life Insurance norms (1959). Participants were trained by nutritionists to keep accurate food records, and recorded their intake for 3 consecutive days at the beginning and end of the intervention. The baseline diaries were completed before participants knew their treatment assignment, and the final diaries were brought to the meetings at Week 8 of the intervention. All food diaries were documented by trained nutritionists and coded for nutrient analysis, using a computerized tape of USDA Handbook 456 (17). Additional information of Na:K content of low-sodium foods was added to this tape for processing of the diaries. Participants were also asked to collect 24-h urine samples that coincided with the third day of recorded food intake at the beginning and end of the program. Oral and written instructions for the collection of a 24-h specimen were provided. Urine samples were analyzed for creatinine, using an adaptation of Folin and Wu, Hoffmann, Peters, and Phillips from Fister Manual Method C43A.l and for sodium and potassium by standard methods of flame photometry using Instrumentation Laboratories, Inc., Model 143. The urinary data were analyzed by laboratory personnel who were blind to the hypothesis of the study and to the participant’s treatment assignment.
236
WING ET AL.
Data Analysis
Repeated-measure analyses of variance were computed on each dependent variable, using intervention condition (weight, Na:K) as a between-group factor and time as a repeated measure (Pre, Post or Pre, Mid, Post). Additional analyses were conducted using either sex or use of hypertension medication as a betweengroup factor. Chi-square analyses were done to compare the percentage of subjects in each condition who achieved the dietary goals. RESULTS
Forty-one participants completed the 8-week intervention program, 19 (70%) in the weight-control intervention and 22 (88%) in the Na:K intervention. A x2 analysis indicated that the rate of attrition was not significantly different for the two interventions. Participants who completed the program ranged in age from 35 to 70 years. Sixty-five percent were on medication and the majority (65%) were female. The demographic characteristics of the subjects in the two intervention conditions are presented in Table 1. The changes in weight, percentage of overweight, and self-reported caloric intake are presented in Table 2.. Changes in body mass index (weight/height*) were also analyzed and yielded similar results. The time x intervention interaction was significant for all weight and weight-related variables. Participants in the weight intervention showed greater changes than those in the Na:K intervention in all of these variables. Analyses of variance computed on each condition separately indicated that the weight changes were significant over time for both the weight (z = -9.61 lbs, -6.97%) and the Na:K intervention (x = -2.77 lbs, - 1.91%). However, only those in the weight intervention had significant decreases in self-reported caloric intake. Men had greater changes in weight, F(2,72) = 8.41, P < 0.001; body mass index, F(2,72) = 4.74, P < 0.01; and selfreported intake, F(1,30) = 12.61, P < 0.001; than females. No significant differences in weight loss or calorie reduction were observed for those on hypertension medication compared with those not on hypertension medication. While the weight intervention experienced the greatest changes in weight-related variables, participants in the Na:K intervention experienced the greatest changes in electrolyte intake and excretion (Fig. 1). The electrolyte data are presented in Tables 3 and 4. Participants in both the weight and the Na:K interTABLE 1 PATIENT CHARACTERISTICS(MEAN k SD)
Weight intervention N % Female % on Hypertension Medication Age SBP DBP % Overweight Weight (lbs)
19 73.7 57.9 56.53 f 155.74 f 93.47 f 30.49 5 178.68 5
6.08 14.40 6.28 25.22 31.74
Na:K intervention 22 63.6 68.2 52.77 ” 147.00 ” 93.59 * 29.89 f 185.91 k
9.84 20.24 8.86 18.69 37.65
DIETARY
INTERVENTIONS
FOR THE REDUCTION
237
OF BP
TABLE 2 CHANGES IN WEIGHT AND SELF-REPORTED CALORIC INTAKE (MEAN 2 SD)
Weight intervention Body wt (Ibs) N Pre Mid Post Change % Overweight N Pre Mid Post Change Food record calories N Pre Post Change
Na:K intervention
18 178.44 2 32.65 174.28 + 33.43 168.83 2 31.91 -9.61**
22 185.91 f 37.65 184.14 f 38.59 183.14 k 37.77 -2.77**
18 31.05 2 25.82 21.97 t 26.06 24.08 2 25.52 -6.97**
22 29.88 2 18.69 28.59 2 19.47 27.97 t- 19.32 - 1.91**
16 1918.06 2 694.99 1127.09 ? 383.93 -790.97**
18 1986.16 r 885.17 1695.00 -r- 405.51 -290.92
Significance of time X intervention interaction
P < 0.001
P < 0.001
P < 0.05
** P < 0.01.
ventions significantly decreased their reported intake of sodium (Table 3). In contrast, the changes over time in reported intake of potassium were significantly different for the two interventions, F( 1,32) = 11.17; P < 0.01. Participants in the weight intervention reported significant decreases in K intake, while those in the Na:K intervention reported slight, but not significant increases in K. Since participants in the weight intervention reported significant decreases in both sodium and potassium, there was no change in the overall Na:K ratio. However, there was a very dramatic change in the ratio for subjects in the Na:K intervention, from 1.77 at pretreatment to 0.92 at post-treatment. The interaction of time x intervention for Na:K ratio from self-reported intake was highly significant. The urinary excretion data (Table 4) provided much higher estimates of sodium intake than did the self-report data. The diary data accounted for 87% of the excretion at baseline and 89% at the end of treatment. Comparable underestimation occurred in the weight and Na:K interventions. The correlation between urinary excretion of Na and self-reported intake of Na was 0.49 (P < 0.01) pretreatment and 0.41 (P < 0.05) post-treatment. For potassium, the correlation between intake and excretion was 0.44 (P < 0.05) pretreatment and 0.62 (P < 0.01) post-treatment. The diary data overestimated potassium excretion by 7% at pretreatment and by 9% at post-treatment. The urinary excretion data showed a pattern similar to the diary data, but none of the time x intervention interactions were significant. Na excretion decreased significantly over time for the Na:K intervention, but the sodium excretion change
238
WING ET AL.
WEIGHT
NA EXCRETION
+3
-+20 EXCkTlON -+I0
+2 +I 2 -1 m co -3 ;
-4
5
r;
d
-7 -2 1
-a
-9 -10 1
WT
m
NIX
,NTER”ENTlON
INTERVENTION
FIG. 1. Weight loss and change in electrolyte excretion for weight-control and Na:K interventions.
was not significant for the weight intervention. Potassium excretion slightly increased in the Na:K intervention and decreased in the weight intervention, but neither change nor the interaction was significant. Again, the weight intervention experienced no change in the Na:K ratio, while there was a significant decrease from a 2.5 ratio at pretreatment to a 1.52 ratio at post-treatment for the Na:K intervention. The time x condition interaction for Na:K ratio approached significance, P = 0.07. Excretion of creatinine showed no change over time. However, there was a highly significant effect of intervention condition with higher levels of creatinine in the Na:K intervention than in the weight intervention. A further analysis of the urinary data was conducted, excluding subjects who had 24-h creatinine levels below 600 mg for either urine sample (N = 6) or experienced more than a 33% variation in creatinine excretion between the 0- and &week samples (N = 8, including 5 of the subjects noted above). These criteria were imposed to exclude participants who provided incomplete urine collections and led to the exclusion of a total of 9 participants (7 from the weight intervention and 2 from the Na:K intervention). When these participants were excluded, there were highly significant interactions of the time x condition for K excretion, F(1,27) = 5.39, P < 0.05, and for Na:K ratio, F(1,26) = 7.36, P = 0.01. Goal Achievement
Another way to assess the adequacy of the treatment interventions for producing dietary change and keeping the modifications in weight and Na:K independent is to look at the number of participants who achieved the goals prescribed for each intervention condition (Table 5). A 5% reduction in body weight was
DIETARY
INTERVENTIONS
FOR THE REDUCTION
239
OF BP
TABLE 3 CHANGE IN SELF-REPORTED INTAKE OF SODIUM AND POTASSIUM (MEAN 2 SD)
Weight intervention
Na:K intervention
Food record Na (mEq) N Pre Post Change
16 121.75 + 62.22 88.38 f 52.49 - 33.37*
18 126.48 2 61.70 66.67 + 31.57 -59.81**
Food record K (mEq) N Pre Post Change
16 63.09 f 22.12 48.40 t 14.76 - 14.69**
18 72.97 t 25.92 79.83 f 19.08 +6.86
16 1.91 f 0.70 1.91 2 1.09 0
18 1.77 f 0.67 0.92 -t 0.53 -0.85**
Food record Na:K N Pre Post Change
Significance of time X intervention interaction
NS
P < 0.01
P < 0.01
* P < 0.05. ** P < 0.01.
achieved by 58% of participants in the weight-loss intervention, but not by any of the participants in the Na:K intervention. Achievement of the sodium goal occurred in a comparable proportion of the two intervention conditions, but that proportion is far lower if one considers the urinary data rather than self-reported intake. The goal of raising K above 100 was achieved by only 3 participants in the Na:K condition using self-report data and by only 1 participant using urinary excretion data. However, if adherence to a goal of 75 mEq of K is assessed, 70% of participants in the Na:K condition achieved the goal based on self-reported intake and 55% based on urinary excretion. These proportions are significantly higher than those found in the weight intervention. It is interesting to note that 5 participants in the weight intervention were excreting ~75 mEq of K at baseline, and that 4 of these 5 were no longer achieving the goal at the end of the weightreduction period. In contrast, 9 patients in the Na:K intervention were above this goal initially and 11 at the end of the program. The percentage of participants who achieved a Na:K ratio of <0.7, based on self-report, was greater in the intervention targeted for that change than in the weight intervention, but there were no differences in percentages of participants who achieved the goal based on urinary output. Blood Pressure
Four participants stopped taking their hypertension medication during the 8week program. In two of these, the changes were made without the physician’s consent. The four participants have been excluded from the blood pressure anal-
WING ET AL. TABLE 4 CHANGE IN URINARY EXCRETION OF SODIUM AND POTASSIUM (MEAN t SD)
Weight intervention Urinary Na (mEq) N Pre Post Change Urinary K (mEq) N” Pre Post Change Urinary Na:K NO Pre Post Change Urinary creatinine (mg) N Pre Post Change
Na:K intervention
19 141.75 2 69.27 139.87 2 107.62 -7.88
19 168.87 5 96.01 107.72 k 48.09 -61.15**
18 59.09 k 17.61 52.49 -t 22.33 - 6.60
20 76.11 k 32.69 82.34 k 35.22 +6.23
18 2.61 2.69 +0.08
19 2.50 1.52 -0.98**
19 971.11 913.79 -57.32
21 1242.38 1227.90 - 14.48
Significance of time x intervention interaction
NS
NS
P = 0.07
NS
a Two subjects on K supplements were omitted from this analysis. ** P < 0.01.
yses. Blood pressure data for the remaining participants are presented in Table 6. Changes in systolic (SBP) and diastolic (DBP) blood pressure from pretreatment to Week 8 were significant for both interventions. Time x intervention interactions did not reach conventional levels of significance for DBP or SBI? The blood pressure data from this study must be interpreted with caution because of the small sample size, short treatment duration, and the lack of a control group asked to make no changes in diet. DISCUSSION
The results of the present study suggest that participants who were taught nutritional and behavioral strategies to lose weight and given no information or advice on Na:K intake lost slightly more than 1 lb/week over an g-week period. These individuals experienced small, nonsignificant decreases in urinary excretion of both Na and K, resulting in no change in the Na:K ratio. In contrast, participants whose program focused on sodium and potassium had a significant decrease in Na excretion, from 169 mEq before treatment to 108 mEq at the end of the program. There was a slight, nonsignificant increase in K excretion. The Na:K ratio decreased markedly from 2.5 at pretreatment to 1.5 post-treatment. Participants in the Na:K intervention lost significantly less weight than those in
DIETARY
INTERVENTIONS
FOR THE REDUCTION
241
OF BP
TABLE 5 PERCENTAGE OF PATIENTS ACHIEVING WEIGHT AND Na:K GOALS
Percentage achieving goal Na:K intervention
Significance of x2
Intervention goal
Weight intervention
5% Reduction body wt
58
0
P < 0.001
Na ~70 mEq diary urine
50
32
70 16
NS NS
K >lOO mEq diary urine
0 0
15 5
NS NS
K >75 mEq diary urine
0 17
70 55
P < 0.001 P < 0.05
Na:K CO.7 diary urine
6 6
55 16
P < 0.01
NS
the weight-control intervention, but their weight loss of almost 3 lbs was significant. It would thus appear that it is possible to keep the changes in weight and Na:K ratio relatively independent of each other. There may, however, be a small, and possibly significant, weight loss in clients targeted for Na:K change. Some of this weight loss may be due to diuresis, not calorie restriction. Over half of the participants in the weight-control intervention were successful in reaching their prescribed goal of a 5% reduction in body weight. In contrast, TABLE 6 CHANGES IN DIASTOLIC AND SYSTOLIC BLOOD PRESSURE(MEAN 2 SD)
Weight intervention (N = 17)
Na:K intervention (N = 20)
Diastolic blood pressure Pre Mid Post Change
92.47 f 5.81 86.29 2 6.03 84.76 k 12.39 -7.71**
94.45 5 8.74 90.90 k 12.35 90.35 -t 10.29 -4.1**
Systolic blood pressure Pre Mid Post Change
155.23 2 14.10 142.65 2 15.62 141.29 k 14.01 - 13.94***
146.70 f 20.24 140.50 k 24.31 140.50 f 23.68 -6.20**
** P < 0.01. *** P < 0.001.
242
WING
ET AL.
only 3 of the 18 participants in the Na:K intervention achieved the goal of ~70 mEq of sodium based on urinary excretion, and only 1 achieved the goal of > 100 mEq of potassium. If a more modest goal had been selected, for example 275 mEq potassium, then over half the patients would have been successful. However, for many, this would have represented maintenance of current potassium intake rather than an increase in potassium. Although increasing potassium intake would appear to be an attractive alternative to stringent sodium restriction for the mildly hypertensive population, the present study suggests that it may be difficult to modify this aspect of dietary behavior. The difficulty of increasing potassium intake may have resulted from the fact that this aspect of the dietary modification was introduced late in the program (Week 4) and represented a second change in diet, in addition to sodium restriction. Participants in the present study had a high level of potassium intake at baseline; this may also have made it difficult to produce further elevation in potassium intake. It may also be more difficult to increase intake of specified foods than to decrease intake. These data raise more general issues regarding the selection of dietary goals in intervention studies. The electrolyte goals for the present study were chosen because they were used in previous studies where they were associated with a reduction in blood pressure (1, 7, 8, 10). However, some of these earlier studies did not assess urinary excretion before and after treatment to determine adherence to the goal (1, 7, g), and, in others, adherence seems to have been quite poor. Adherence might be improved by setting individualized goals for electrolyte intake and excretion, rather than using general target levels. Since there are marked differences between clients in baseline intake and excretion, and since total caloric intake may affect the difficulty of achieving prescribed goals, it may be appropriate to encourage clients to reduce sodium intake (or increase potassium intake) by a given percentage, rather than asking all clients to attain a preset target level. However, it is not known whether it is the absolute level of Na or K or the percentage of change that is important in lowering blood pressure. In addition, it is difficult to know whether to determine adherence by selfreported intake or by urinary excretion, since there are problems with both measures. In the present study, estimates of sodium intake based on 3-day diaries were about 20% lower than those obtained from urinary excretion. There was much less discrepancy between reported intake and output for potassium, with self-reports overestimating excretion by only 7%. The discrepancies in this clinical sample are greater than those found for research subjects. Schachter ef al. (13) found that 3-day food records kept by highly motivated adults, knowledgeable about nutrition, underestimated urinary sodium excretion by only 11% and potassium excretion by less than 1%. Interestingly, in the present study, the discrepancy between intake and output was not significantly greater in participants who had been specifically targeted for Na:K change, although these participants might have been inclined to distort their intake to appear more compliant with the program goals. The discrepancy between the intake and excretion measures of sodium and potassium may result from various factors, including incomplete diaries, failure
DIETARY
INTERVENTIONS
FOR THE REDUCTION
OF BP
243
to accurately estimate added salt, and incomplete collection of urines. Six participants in the present study had creatinine excretion of less than 600 mg, suggesting that the urine samples they provided were incomplete. A decision to exclude such subjects has been made in other reports (19). It would be helpful to be able to use the variability in 24-h creatinine excretion to identify incomplete samples. However, Vestergaard and Leverett (18) and Joossens ef al. (3) found that the variation in 24-h creatinine excretion is quite marked in some individuals. In conclusion, the finding that various dietary changes can be made independently of each other in free-living adults should enable us in subsequent studies to compare the effectiveness of these dietary approaches for the reduction of blood pressure. REFERENCES 1. Carney, S., Morgan, T., Wilson, M., Matthews, G., and Roberts, R. Sodium restriction and thiazide diuretics in the treatment of hypertension. Med. J. Austr. 1, 803-807 (1975). 2. Hypertension Detection and Follow-up Program Cooperative Group: Five-Year Findings of the Hypertension Detection and Follow-up Program. JAMA 242, 2562-2567 (1979). 3. Joosens, .I. V., Claessens, J., Geboers, .I., and Claes, J. H. Electrolytes and creatinine in multiple 24-hour urine collections (1970-1974) in “Epidemiology of Arterial Blood Pressure” (H. Kesteloot and J. V. Joosens, Eds.), p. 45. Boston, 1980. 4. Kaplan, N. M. Hypertension: Prevalence, risks and effect of therapy. Ann. Intern. Med. 98,705709 (1983). 5. MacGregor, G. A., Best, E E., Cam, J. M., Markandu, N. D., Elder, D. M., Sagnella, G. A., and Squires, M. Double-blind randomized crossover trial of moderate sodium restriction in essential hypertension. Lancer 1, 351-355 (1982). 6. MacGregor, G. A., Markandu, N. D., Smith, S. J., Banks, R. A., and Sagnella, G. A. Moderate potassium supplementation in essential hypertension. Lancer 2, 567-570 (1982). 7. Magnani, B., Ambrosioni, E., Agosta, R., and Racco, F. Comparison of the effects of pharmacological therapy and a low sodium diet on mild hypertension. C/in. Sci. Mol. Med 51, 625s626s (1976). 8. Morgan, T., Gillies, A., Morgan, G., Adam, W., Wilson, M., and Carney, S. Hypertension treated by salt restriction. Lancet 1, 227-230 (1978). 9. Multiple Risk Factor Intervention Trial Research Group. Multiple Risk Factor Intervention Trial: Risk factor and mortality results. JAMA 248, 1465-1477 (1982). 10. Parfrey, P. S., Wright, P., Goodwin, F. J., Vandenburg, M. J., Holly, J. M. P., Evans, S. J. W., and Ledingham, J. M. Blood pressure and hormonal changes following alteration in dietary sodium and potassium in mild essential hypertension. Lancet 1, 59-63 (1981). 11. Parijs, J., Joosens, J. V., Linden, L., Verstreken, G., and Amery, A. Moderate sodium restriction and diuretics in the treatment of hypertension. Amer. Heart J. 85, 22-34 (1973). 12. Reisin, E., Abel, R., Modan, M., Silverberg, D., Eliahou, H. E., and Modan, B. Effect of weight loss without salt restriction on the reduction of blood pressure in overweight hypertensive patients. New Engl. J. Med. 298, l-6 (1978). 13. Schachter, J., Harper, P. H., Radin, M. E., Caggiula, A. W., McDonald, R. H., and Diven, W. F. Comparison of sodium and potassium intake with excretion. Hypertension 2, 695-699 (1980). 14. Skrabal, F., Aubock, J., and Hortnagl, H. Low sodium/high potassium diet for prevention of hypertension: Probable mechanisms of action. Lancer 2, 895-902 (1981). 15. Skrabal, F., Aubock, J., Hortnagl, H., and Braunsteiner, H, Effect of moderate salt restriction and high potassium intake on pressor hormones response to noradrenaline and baroreceptor function in man. Clin. Sci. 59, 157s- 160s (1980). 16. Tuck, M. L., Sowers, J., Dornfeld, L., Kledzik, G., and Maxwell, M. The effect of weight reduction on blood pressure, plasma renin activity and plasma aldosterone levels in obese patients. New Engl. J. Med. 304, 930-934 (1981).
244
WING ET AL.
17. United States Department of Agriculture. “Nutritive Value of American Foods in Common Units,” Agricultural Handbook 456, U.S. Govt. Printing Office, Washington, D.C., 1975. 18. Vestergaard, P., and Leverett, R. Constancy of urinary creatinine excretion. J. Lab. Clin. Med. 51, 211-218 (1958). 19. Wassertheil-Smoller, S., Langford, H. G., Blaufox, M. D., Oberman, A., Raiford, Y., Lucido, D., Caggiula, A., and Wylie, J. Sodium and potassium in hypertensive patients. Paper presented at 21st Conference on Cardiovascular Epidemiology, American Heart Association, March 1981.
MEDICINE
13, 233-244
(1984)
Dietary Approaches to the Reduction of Blood Pressure: The Independence of Weight and Sodium/Potassium Interventions’ RENA R. WING,*-* ARLENE W. CAGGIULA,+ MARY PATRICIA NOWALK,* RANDI KOESKE,* SEUNG LEE,* AND HERBERT LANGFORDS *University of Pittsburgh School of Medicine, and tGraduate School of Public of Pittsburgh, Pittsburgh, Pennsylvania 15213: and #University of Mississippi Jackson, Mississippi 39216
Health. Medical
University Center,
This study was designed to determine the feasibility of teaching mildly hypertensive individuals to select a diet, using normally available food products, that either would produce a 5% reduction in percentage of overweight (without altering sodium (Na), potassium (K), or Na:K ratio) or would decrease Na to <70 mEq and increase K to >I00 mEq (without affecting weight) and to compare the resulting changes in blood pressure. Fifty-two participants with mild hypertension were randomly assigned to either a weight-loss or a &a:K intervention. Blood pressure, weight, 3-day diaries, and 24-h urinary excretion of Na and K were measured before and after an I-week intervention. Participants in the weight-loss intervention had significantly greater changes in weight and calorie intake than those in the Na:K intervention, while changes in Na:K ratio were greatest in the intervention targeted for that change. The percentage of participants who were able to meet the dietary goals is presented and the implications of these data for the selection of dietary goals are discussed.
INTRODUCTION While estimates of the prevalence of hypertension vary, it is clear that hypertension is a risk for premature cardiovascular disease for millions of Americans (4). The basic dilemma facing the medical and public health community today regards how best to treat these individuals, particularly those with mild hypertension. Recent clinical trials have found that treating mild hypertension reduces mortality (2), but there is concern about placing large portions of the population on drug therapy, since drug therapy is costly and may have risks of its own (9). Consequently, attention has focused on nonpharmacological approaches to blood pressure reduction, such as relaxation and diet modification. There are several different types of nutrition intervention which may be effective as the sole therapy for mild hypertension or as an adjunct to drug therapy: moderate sodium (Na) restriction (1,5,7,8, 1I), increases in potassium (K) intake (6), changes in the Na:K ratio (IO, 14, 15), and weight reduction (12, 16). While there is evidence that nutrition interventions reduce blood pressure, there are very few studies comparing the effectiveness of different types of nutrition intervention. Such comparisons may be difficult, because the various aspects of the diet are interrelated. For example, changes may occur simultaneously ’ The research reported in this paper was supported in part by Grant 5 ROI HL24369-03 from the National Institutes of Health to Dr. Langford. ’ To whom requests for reprints should be addressed at: Western Psychiatric Institute and Clinic, 381I O’Hara Street, Pittsburgh, Pa 15213. 233
0091-7435/84 Copyright All nghls
$3.00
G IY84 by Acadcmlc Press. Inc. of rcproduclion in any form reserved.
234
WING
ET AL.
in both caloric intake and in the sodium:potassium content of the diet, making it difficult to assess the effect of either change independently. A few studies have attempted to keep the changes separate by providing food to the participants (14, placing participants on a low-calorie liquid formula regimen with known sodium concentration (16), encouraging participants in a weight-reduction program to eat salty foods (121, or adding a potassium supplement to a normal diet (6). However, it is not known whether these dietary changes can be made independently when individuals self-select their diet. Further, in some of these studies body weight and electrolyte excretion were not measured both prior to and following the dietary intervention to confirm the independence of the changes. The purpose of the present study was to determine whether it is possible to modify weight and sodium:potassium ratio independently by dietary intervention in free-living adults. The goal was to teach hypertensive individuals to self-select a diet, using normally available food products, that either would produce a 5% reduction in body weight without altering sodium, potassium, or sodium:potassium ratio or would decrease sodium to 70 mEq and increase potassium to 100 mEq without affecting weight. The study was conducted as a feasibility test of the intervention program designed for the Dietary Intervention Study of Hypertension (DISH). METHOD Subjects
Participants were recruited from the Pittsburgh area by radio and newspaper advertisements. Eligibility was based on blood pressure measurements taken at two screening visits scheduled one week apart. Blood pressure measurements were made by a nurse, certified according to standardized procedures from the Multiple Risk Factor Intervention Trial (MRFIT) protocol. A random zero sphygmomanometer was used. Three blood pressure readings were taken with the participant seated after 5-10 min of rest. The first reading was disregarded and the second and third averaged to form the blood pressure measurement. Appropriately sized cuffs were used to account for differences in arm circumference. The baseline blood pressure was the average of the blood pressure readings across the two screening sessions. Individuals whose average diastolic blood pressure (DBP) over the two sessions was a85 or who were being treated with anti-hypertensive medication and had DBP 380 were accepted into the study. Fifty-two individuals met these eligibility criteria and were randomly assigned to an intervention, 27 to the weight-control intervention and 25 to the Na:K intervention. Procedures
Eligible participants were randomly assigned to one of two intervention conditions: Weight Loss or Na:K Change. The goal of the Na:K intervention was to decrease sodium intake to less than 70 mEq/day and increase potassium intake to greater than 100 mEq/day, while keeping body weight constant. The weight intervention was designed to produce a 5% reduction in body weight with no change in Na, K, or Na:K ratio.
DIETARY
INTERVENTIONS
FOR THE REDUCTION
OF BP
235
Both interventions were conducted in a group format, with 10 to 14 participants per group. Meetings were held once a week for 8 weeks. At the start of each meeting, patients were weighed individually and a team of four nutritionists reviewed food diaries. This was followed by the presentation and discussion of a new strategy for changing sodium and potassium intake or reducing weight. The discussion was led by a nutritionist and a behavior therapist. At the first meeting, participants were given a booklet from which they could determine the sodium and potassium content of common foods or the calorie content of these foods. They were instructed to monitor their intake aiming to achieve the Na:K goals (~70 mEq Na and >lOO mEq K) or an individualized calorie goal (based on entry weight x 12 - 1,000 calories/day, with a minimum calorie goal of 1,000 calories per day). The Na:K intervention focused first on decreasing sodium by eliminating table salt, using alternative seasonings, and modifying food choices. The goal of increasing potassium by eating foods high in this mineral was added by Week 4. The weight-control intervention emphasized reducing consumption of high-fat and other high-calorie foods and increasing caloric expenditure through walking. Measures
Procedures for blood pressure measurement are described above. The same procedures were used to obtain blood pressure measurements at Weeks 4 and 8 of the intervention. Weight was assessed at both screening meetings and at each week of intervention using a balance-beam scale. Participants were weighed wearing indoor clothing and without shoes. Baseline weight was the average weight across the two screening sessions. Height was determined at the second screening. Percentage of overweight was computed using the Metropolitan Life Insurance norms (1959). Participants were trained by nutritionists to keep accurate food records, and recorded their intake for 3 consecutive days at the beginning and end of the intervention. The baseline diaries were completed before participants knew their treatment assignment, and the final diaries were brought to the meetings at Week 8 of the intervention. All food diaries were documented by trained nutritionists and coded for nutrient analysis, using a computerized tape of USDA Handbook 456 (17). Additional information of Na:K content of low-sodium foods was added to this tape for processing of the diaries. Participants were also asked to collect 24-h urine samples that coincided with the third day of recorded food intake at the beginning and end of the program. Oral and written instructions for the collection of a 24-h specimen were provided. Urine samples were analyzed for creatinine, using an adaptation of Folin and Wu, Hoffmann, Peters, and Phillips from Fister Manual Method C43A.l and for sodium and potassium by standard methods of flame photometry using Instrumentation Laboratories, Inc., Model 143. The urinary data were analyzed by laboratory personnel who were blind to the hypothesis of the study and to the participant’s treatment assignment.
236
WING ET AL.
Data Analysis
Repeated-measure analyses of variance were computed on each dependent variable, using intervention condition (weight, Na:K) as a between-group factor and time as a repeated measure (Pre, Post or Pre, Mid, Post). Additional analyses were conducted using either sex or use of hypertension medication as a betweengroup factor. Chi-square analyses were done to compare the percentage of subjects in each condition who achieved the dietary goals. RESULTS
Forty-one participants completed the 8-week intervention program, 19 (70%) in the weight-control intervention and 22 (88%) in the Na:K intervention. A x2 analysis indicated that the rate of attrition was not significantly different for the two interventions. Participants who completed the program ranged in age from 35 to 70 years. Sixty-five percent were on medication and the majority (65%) were female. The demographic characteristics of the subjects in the two intervention conditions are presented in Table 1. The changes in weight, percentage of overweight, and self-reported caloric intake are presented in Table 2.. Changes in body mass index (weight/height*) were also analyzed and yielded similar results. The time x intervention interaction was significant for all weight and weight-related variables. Participants in the weight intervention showed greater changes than those in the Na:K intervention in all of these variables. Analyses of variance computed on each condition separately indicated that the weight changes were significant over time for both the weight (z = -9.61 lbs, -6.97%) and the Na:K intervention (x = -2.77 lbs, - 1.91%). However, only those in the weight intervention had significant decreases in self-reported caloric intake. Men had greater changes in weight, F(2,72) = 8.41, P < 0.001; body mass index, F(2,72) = 4.74, P < 0.01; and selfreported intake, F(1,30) = 12.61, P < 0.001; than females. No significant differences in weight loss or calorie reduction were observed for those on hypertension medication compared with those not on hypertension medication. While the weight intervention experienced the greatest changes in weight-related variables, participants in the Na:K intervention experienced the greatest changes in electrolyte intake and excretion (Fig. 1). The electrolyte data are presented in Tables 3 and 4. Participants in both the weight and the Na:K interTABLE 1 PATIENT CHARACTERISTICS(MEAN k SD)
Weight intervention N % Female % on Hypertension Medication Age SBP DBP % Overweight Weight (lbs)
19 73.7 57.9 56.53 f 155.74 f 93.47 f 30.49 5 178.68 5
6.08 14.40 6.28 25.22 31.74
Na:K intervention 22 63.6 68.2 52.77 ” 147.00 ” 93.59 * 29.89 f 185.91 k
9.84 20.24 8.86 18.69 37.65
DIETARY
INTERVENTIONS
FOR THE REDUCTION
237
OF BP
TABLE 2 CHANGES IN WEIGHT AND SELF-REPORTED CALORIC INTAKE (MEAN 2 SD)
Weight intervention Body wt (Ibs) N Pre Mid Post Change % Overweight N Pre Mid Post Change Food record calories N Pre Post Change
Na:K intervention
18 178.44 2 32.65 174.28 + 33.43 168.83 2 31.91 -9.61**
22 185.91 f 37.65 184.14 f 38.59 183.14 k 37.77 -2.77**
18 31.05 2 25.82 21.97 t 26.06 24.08 2 25.52 -6.97**
22 29.88 2 18.69 28.59 2 19.47 27.97 t- 19.32 - 1.91**
16 1918.06 2 694.99 1127.09 ? 383.93 -790.97**
18 1986.16 r 885.17 1695.00 -r- 405.51 -290.92
Significance of time X intervention interaction
P < 0.001
P < 0.001
P < 0.05
** P < 0.01.
ventions significantly decreased their reported intake of sodium (Table 3). In contrast, the changes over time in reported intake of potassium were significantly different for the two interventions, F( 1,32) = 11.17; P < 0.01. Participants in the weight intervention reported significant decreases in K intake, while those in the Na:K intervention reported slight, but not significant increases in K. Since participants in the weight intervention reported significant decreases in both sodium and potassium, there was no change in the overall Na:K ratio. However, there was a very dramatic change in the ratio for subjects in the Na:K intervention, from 1.77 at pretreatment to 0.92 at post-treatment. The interaction of time x intervention for Na:K ratio from self-reported intake was highly significant. The urinary excretion data (Table 4) provided much higher estimates of sodium intake than did the self-report data. The diary data accounted for 87% of the excretion at baseline and 89% at the end of treatment. Comparable underestimation occurred in the weight and Na:K interventions. The correlation between urinary excretion of Na and self-reported intake of Na was 0.49 (P < 0.01) pretreatment and 0.41 (P < 0.05) post-treatment. For potassium, the correlation between intake and excretion was 0.44 (P < 0.05) pretreatment and 0.62 (P < 0.01) post-treatment. The diary data overestimated potassium excretion by 7% at pretreatment and by 9% at post-treatment. The urinary excretion data showed a pattern similar to the diary data, but none of the time x intervention interactions were significant. Na excretion decreased significantly over time for the Na:K intervention, but the sodium excretion change
238
WING ET AL.
WEIGHT
NA EXCRETION
+3
-+20 EXCkTlON -+I0
+2 +I 2 -1 m co -3 ;
-4
5
r;
d
-7 -2 1
-a
-9 -10 1
WT
m
NIX
,NTER”ENTlON
INTERVENTION
FIG. 1. Weight loss and change in electrolyte excretion for weight-control and Na:K interventions.
was not significant for the weight intervention. Potassium excretion slightly increased in the Na:K intervention and decreased in the weight intervention, but neither change nor the interaction was significant. Again, the weight intervention experienced no change in the Na:K ratio, while there was a significant decrease from a 2.5 ratio at pretreatment to a 1.52 ratio at post-treatment for the Na:K intervention. The time x condition interaction for Na:K ratio approached significance, P = 0.07. Excretion of creatinine showed no change over time. However, there was a highly significant effect of intervention condition with higher levels of creatinine in the Na:K intervention than in the weight intervention. A further analysis of the urinary data was conducted, excluding subjects who had 24-h creatinine levels below 600 mg for either urine sample (N = 6) or experienced more than a 33% variation in creatinine excretion between the 0- and &week samples (N = 8, including 5 of the subjects noted above). These criteria were imposed to exclude participants who provided incomplete urine collections and led to the exclusion of a total of 9 participants (7 from the weight intervention and 2 from the Na:K intervention). When these participants were excluded, there were highly significant interactions of the time x condition for K excretion, F(1,27) = 5.39, P < 0.05, and for Na:K ratio, F(1,26) = 7.36, P = 0.01. Goal Achievement
Another way to assess the adequacy of the treatment interventions for producing dietary change and keeping the modifications in weight and Na:K independent is to look at the number of participants who achieved the goals prescribed for each intervention condition (Table 5). A 5% reduction in body weight was
DIETARY
INTERVENTIONS
FOR THE REDUCTION
239
OF BP
TABLE 3 CHANGE IN SELF-REPORTED INTAKE OF SODIUM AND POTASSIUM (MEAN 2 SD)
Weight intervention
Na:K intervention
Food record Na (mEq) N Pre Post Change
16 121.75 + 62.22 88.38 f 52.49 - 33.37*
18 126.48 2 61.70 66.67 + 31.57 -59.81**
Food record K (mEq) N Pre Post Change
16 63.09 f 22.12 48.40 t 14.76 - 14.69**
18 72.97 t 25.92 79.83 f 19.08 +6.86
16 1.91 f 0.70 1.91 2 1.09 0
18 1.77 f 0.67 0.92 -t 0.53 -0.85**
Food record Na:K N Pre Post Change
Significance of time X intervention interaction
NS
P < 0.01
P < 0.01
* P < 0.05. ** P < 0.01.
achieved by 58% of participants in the weight-loss intervention, but not by any of the participants in the Na:K intervention. Achievement of the sodium goal occurred in a comparable proportion of the two intervention conditions, but that proportion is far lower if one considers the urinary data rather than self-reported intake. The goal of raising K above 100 was achieved by only 3 participants in the Na:K condition using self-report data and by only 1 participant using urinary excretion data. However, if adherence to a goal of 75 mEq of K is assessed, 70% of participants in the Na:K condition achieved the goal based on self-reported intake and 55% based on urinary excretion. These proportions are significantly higher than those found in the weight intervention. It is interesting to note that 5 participants in the weight intervention were excreting ~75 mEq of K at baseline, and that 4 of these 5 were no longer achieving the goal at the end of the weightreduction period. In contrast, 9 patients in the Na:K intervention were above this goal initially and 11 at the end of the program. The percentage of participants who achieved a Na:K ratio of <0.7, based on self-report, was greater in the intervention targeted for that change than in the weight intervention, but there were no differences in percentages of participants who achieved the goal based on urinary output. Blood Pressure
Four participants stopped taking their hypertension medication during the 8week program. In two of these, the changes were made without the physician’s consent. The four participants have been excluded from the blood pressure anal-
WING ET AL. TABLE 4 CHANGE IN URINARY EXCRETION OF SODIUM AND POTASSIUM (MEAN t SD)
Weight intervention Urinary Na (mEq) N Pre Post Change Urinary K (mEq) N” Pre Post Change Urinary Na:K NO Pre Post Change Urinary creatinine (mg) N Pre Post Change
Na:K intervention
19 141.75 2 69.27 139.87 2 107.62 -7.88
19 168.87 5 96.01 107.72 k 48.09 -61.15**
18 59.09 k 17.61 52.49 -t 22.33 - 6.60
20 76.11 k 32.69 82.34 k 35.22 +6.23
18 2.61 2.69 +0.08
19 2.50 1.52 -0.98**
19 971.11 913.79 -57.32
21 1242.38 1227.90 - 14.48
Significance of time x intervention interaction
NS
NS
P = 0.07
NS
a Two subjects on K supplements were omitted from this analysis. ** P < 0.01.
yses. Blood pressure data for the remaining participants are presented in Table 6. Changes in systolic (SBP) and diastolic (DBP) blood pressure from pretreatment to Week 8 were significant for both interventions. Time x intervention interactions did not reach conventional levels of significance for DBP or SBI? The blood pressure data from this study must be interpreted with caution because of the small sample size, short treatment duration, and the lack of a control group asked to make no changes in diet. DISCUSSION
The results of the present study suggest that participants who were taught nutritional and behavioral strategies to lose weight and given no information or advice on Na:K intake lost slightly more than 1 lb/week over an g-week period. These individuals experienced small, nonsignificant decreases in urinary excretion of both Na and K, resulting in no change in the Na:K ratio. In contrast, participants whose program focused on sodium and potassium had a significant decrease in Na excretion, from 169 mEq before treatment to 108 mEq at the end of the program. There was a slight, nonsignificant increase in K excretion. The Na:K ratio decreased markedly from 2.5 at pretreatment to 1.5 post-treatment. Participants in the Na:K intervention lost significantly less weight than those in
DIETARY
INTERVENTIONS
FOR THE REDUCTION
241
OF BP
TABLE 5 PERCENTAGE OF PATIENTS ACHIEVING WEIGHT AND Na:K GOALS
Percentage achieving goal Na:K intervention
Significance of x2
Intervention goal
Weight intervention
5% Reduction body wt
58
0
P < 0.001
Na ~70 mEq diary urine
50
32
70 16
NS NS
K >lOO mEq diary urine
0 0
15 5
NS NS
K >75 mEq diary urine
0 17
70 55
P < 0.001 P < 0.05
Na:K CO.7 diary urine
6 6
55 16
P < 0.01
NS
the weight-control intervention, but their weight loss of almost 3 lbs was significant. It would thus appear that it is possible to keep the changes in weight and Na:K ratio relatively independent of each other. There may, however, be a small, and possibly significant, weight loss in clients targeted for Na:K change. Some of this weight loss may be due to diuresis, not calorie restriction. Over half of the participants in the weight-control intervention were successful in reaching their prescribed goal of a 5% reduction in body weight. In contrast, TABLE 6 CHANGES IN DIASTOLIC AND SYSTOLIC BLOOD PRESSURE(MEAN 2 SD)
Weight intervention (N = 17)
Na:K intervention (N = 20)
Diastolic blood pressure Pre Mid Post Change
92.47 f 5.81 86.29 2 6.03 84.76 k 12.39 -7.71**
94.45 5 8.74 90.90 k 12.35 90.35 -t 10.29 -4.1**
Systolic blood pressure Pre Mid Post Change
155.23 2 14.10 142.65 2 15.62 141.29 k 14.01 - 13.94***
146.70 f 20.24 140.50 k 24.31 140.50 f 23.68 -6.20**
** P < 0.01. *** P < 0.001.
242
WING
ET AL.
only 3 of the 18 participants in the Na:K intervention achieved the goal of ~70 mEq of sodium based on urinary excretion, and only 1 achieved the goal of > 100 mEq of potassium. If a more modest goal had been selected, for example 275 mEq potassium, then over half the patients would have been successful. However, for many, this would have represented maintenance of current potassium intake rather than an increase in potassium. Although increasing potassium intake would appear to be an attractive alternative to stringent sodium restriction for the mildly hypertensive population, the present study suggests that it may be difficult to modify this aspect of dietary behavior. The difficulty of increasing potassium intake may have resulted from the fact that this aspect of the dietary modification was introduced late in the program (Week 4) and represented a second change in diet, in addition to sodium restriction. Participants in the present study had a high level of potassium intake at baseline; this may also have made it difficult to produce further elevation in potassium intake. It may also be more difficult to increase intake of specified foods than to decrease intake. These data raise more general issues regarding the selection of dietary goals in intervention studies. The electrolyte goals for the present study were chosen because they were used in previous studies where they were associated with a reduction in blood pressure (1, 7, 8, 10). However, some of these earlier studies did not assess urinary excretion before and after treatment to determine adherence to the goal (1, 7, g), and, in others, adherence seems to have been quite poor. Adherence might be improved by setting individualized goals for electrolyte intake and excretion, rather than using general target levels. Since there are marked differences between clients in baseline intake and excretion, and since total caloric intake may affect the difficulty of achieving prescribed goals, it may be appropriate to encourage clients to reduce sodium intake (or increase potassium intake) by a given percentage, rather than asking all clients to attain a preset target level. However, it is not known whether it is the absolute level of Na or K or the percentage of change that is important in lowering blood pressure. In addition, it is difficult to know whether to determine adherence by selfreported intake or by urinary excretion, since there are problems with both measures. In the present study, estimates of sodium intake based on 3-day diaries were about 20% lower than those obtained from urinary excretion. There was much less discrepancy between reported intake and output for potassium, with self-reports overestimating excretion by only 7%. The discrepancies in this clinical sample are greater than those found for research subjects. Schachter ef al. (13) found that 3-day food records kept by highly motivated adults, knowledgeable about nutrition, underestimated urinary sodium excretion by only 11% and potassium excretion by less than 1%. Interestingly, in the present study, the discrepancy between intake and output was not significantly greater in participants who had been specifically targeted for Na:K change, although these participants might have been inclined to distort their intake to appear more compliant with the program goals. The discrepancy between the intake and excretion measures of sodium and potassium may result from various factors, including incomplete diaries, failure
DIETARY
INTERVENTIONS
FOR THE REDUCTION
OF BP
243
to accurately estimate added salt, and incomplete collection of urines. Six participants in the present study had creatinine excretion of less than 600 mg, suggesting that the urine samples they provided were incomplete. A decision to exclude such subjects has been made in other reports (19). It would be helpful to be able to use the variability in 24-h creatinine excretion to identify incomplete samples. However, Vestergaard and Leverett (18) and Joossens ef al. (3) found that the variation in 24-h creatinine excretion is quite marked in some individuals. In conclusion, the finding that various dietary changes can be made independently of each other in free-living adults should enable us in subsequent studies to compare the effectiveness of these dietary approaches for the reduction of blood pressure. REFERENCES 1. Carney, S., Morgan, T., Wilson, M., Matthews, G., and Roberts, R. Sodium restriction and thiazide diuretics in the treatment of hypertension. Med. J. Austr. 1, 803-807 (1975). 2. Hypertension Detection and Follow-up Program Cooperative Group: Five-Year Findings of the Hypertension Detection and Follow-up Program. JAMA 242, 2562-2567 (1979). 3. Joosens, .I. V., Claessens, J., Geboers, .I., and Claes, J. H. Electrolytes and creatinine in multiple 24-hour urine collections (1970-1974) in “Epidemiology of Arterial Blood Pressure” (H. Kesteloot and J. V. Joosens, Eds.), p. 45. Boston, 1980. 4. Kaplan, N. M. Hypertension: Prevalence, risks and effect of therapy. Ann. Intern. Med. 98,705709 (1983). 5. MacGregor, G. A., Best, E E., Cam, J. M., Markandu, N. D., Elder, D. M., Sagnella, G. A., and Squires, M. Double-blind randomized crossover trial of moderate sodium restriction in essential hypertension. Lancer 1, 351-355 (1982). 6. MacGregor, G. A., Markandu, N. D., Smith, S. J., Banks, R. A., and Sagnella, G. A. Moderate potassium supplementation in essential hypertension. Lancer 2, 567-570 (1982). 7. Magnani, B., Ambrosioni, E., Agosta, R., and Racco, F. Comparison of the effects of pharmacological therapy and a low sodium diet on mild hypertension. C/in. Sci. Mol. Med 51, 625s626s (1976). 8. Morgan, T., Gillies, A., Morgan, G., Adam, W., Wilson, M., and Carney, S. Hypertension treated by salt restriction. Lancet 1, 227-230 (1978). 9. Multiple Risk Factor Intervention Trial Research Group. Multiple Risk Factor Intervention Trial: Risk factor and mortality results. JAMA 248, 1465-1477 (1982). 10. Parfrey, P. S., Wright, P., Goodwin, F. J., Vandenburg, M. J., Holly, J. M. P., Evans, S. J. W., and Ledingham, J. M. Blood pressure and hormonal changes following alteration in dietary sodium and potassium in mild essential hypertension. Lancet 1, 59-63 (1981). 11. Parijs, J., Joosens, J. V., Linden, L., Verstreken, G., and Amery, A. Moderate sodium restriction and diuretics in the treatment of hypertension. Amer. Heart J. 85, 22-34 (1973). 12. Reisin, E., Abel, R., Modan, M., Silverberg, D., Eliahou, H. E., and Modan, B. Effect of weight loss without salt restriction on the reduction of blood pressure in overweight hypertensive patients. New Engl. J. Med. 298, l-6 (1978). 13. Schachter, J., Harper, P. H., Radin, M. E., Caggiula, A. W., McDonald, R. H., and Diven, W. F. Comparison of sodium and potassium intake with excretion. Hypertension 2, 695-699 (1980). 14. Skrabal, F., Aubock, J., and Hortnagl, H. Low sodium/high potassium diet for prevention of hypertension: Probable mechanisms of action. Lancer 2, 895-902 (1981). 15. Skrabal, F., Aubock, J., Hortnagl, H., and Braunsteiner, H, Effect of moderate salt restriction and high potassium intake on pressor hormones response to noradrenaline and baroreceptor function in man. Clin. Sci. 59, 157s- 160s (1980). 16. Tuck, M. L., Sowers, J., Dornfeld, L., Kledzik, G., and Maxwell, M. The effect of weight reduction on blood pressure, plasma renin activity and plasma aldosterone levels in obese patients. New Engl. J. Med. 304, 930-934 (1981).
244
WING ET AL.
17. United States Department of Agriculture. “Nutritive Value of American Foods in Common Units,” Agricultural Handbook 456, U.S. Govt. Printing Office, Washington, D.C., 1975. 18. Vestergaard, P., and Leverett, R. Constancy of urinary creatinine excretion. J. Lab. Clin. Med. 51, 211-218 (1958). 19. Wassertheil-Smoller, S., Langford, H. G., Blaufox, M. D., Oberman, A., Raiford, Y., Lucido, D., Caggiula, A., and Wylie, J. Sodium and potassium in hypertensive patients. Paper presented at 21st Conference on Cardiovascular Epidemiology, American Heart Association, March 1981.