High plasma aldosterone and low renin predict blood pressure increase and hypertension in middle-aged Caucasian populations

Transcription

1 (2008) 22, & 2008 Macmillan Publishers Limited All rights reserved /08 $ ORIGINAL ARTICLE High plasma aldosterone and low renin predict blood pressure increase and hypertension in middle-aged Caucasian populations P Meneton 1, P Galan 2, S Bertrais 2, D Heudes 3, S Hercberg 2 and J Ménard 3 1 INSERM U872, Centre de Recherche des Cordeliers, Paris, France; 2 INSERM U557/INRA U1125/CNAM/ Paris 13 SMBH, Centre de Recherche en Nutrition Humaine d Ile-de-France, Bobigny, France and 3 Département de Santé Publique et d Informatique Médicale, Université Paris Descartes, Paris, France and renin levels have been associated with blood pressure increase and 3 4 year incidence of hypertension in a middle-aged North American community in Framingham. To confirm these findings in a different population, a nested case control study was performed in a national sample of 1984 French non-hypertensive volunteers aged year and followed for 5 years. Cases and controls (individuals becoming hypertensive or remaining non-hypertensive on follow-up) were individually matched on sex, diastolic and systolic pressures at baseline. Multivariable regression models show that plasma aldosterone and renin are respectively positively and negatively associated with the increase in systolic pressure (P ¼ 0.01 and 0.001) and the risk of hypertension (22% increase and 16% decrease per s.d. increment in the log, P ¼ 0.04 and 0.07). These associations are mostly observed in the lowest tertiles of dietary sodium and potassium intakes where plasma aldosterone is positively associated with the increase in systolic pressure (P ¼ 0.01 and 0.08) and the risk of hypertension (59 and 69% increase per s.d. increment in the log, P ¼ 0.02 and 0.01), whereas plasma renin is negatively associated with the increase in systolic pressure (P ¼ and 0.004) and the risk of hypertension (31 and 28% decrease per s.d. increment in the log, P ¼ 0.03 and 0.05). These results reinforce the hypothesis that high plasma aldosterone and low plasma renin levels precede blood pressure increase and the occurrence of hypertension in middle-aged Caucasian populations. (2008) 22, ; doi: /jhh ; published online 1 May 2008 Keywords: blood pressure; plasma aldosterone and renin; population Introduction Despite the extensive use of plasma aldosterone and renin and of their ratio in hypertension clinics to establish the diagnosis of secondary hypertension and to select appropriate antihypertensive treatments, 1 4 the number of community-based longitudinal studies reporting these parameters was rather modest until recently. It was made of one English cohort selected in a general population and one American cohort chosen at the worksite. 5 7 Both studies tested the possibility to predict coronary events by measuring plasma renin and have concluded in opposite directions. 5 7 In the past few Correspondence: Professor J Ménard, Département de Santé Publique et d Informatique Médicale, Université Paris Descartes, 15 rue de l Ecole de Médecine, Paris 75006, France. Joelle.Schlama@spim.jussieu.fr Received 22 February 2008; revised 10 March 2008; accepted 16 March 2008; published online 1 May 2008 years, the investigators of the Framingham cohort have analysed the association of multiple biomarkers with blood pressure increase and incident hypertension They showed that the highest quartile of serum aldosterone, relative to the lowest, was associated with a 60% rise in the risk of blood pressure increase and hypertension at 4 years. 8 They also observed that plasma renin was associated with a drop in the risk of blood pressure increase and hypertension on a 3-year follow-up. 10 However, these authors subsequently found in the same cohort that only three biomarkers among nine eligible were retained in a multivariable model: plasma c-reactive protein and plasminogen activator inhibitor-1 and urinary albumin/creatinine ratio whereas neither serum aldosterone nor plasma renin entered the model. 9 These observations prompted us to investigate plasma aldosterone and renin in a national sample of middle-aged French healthy volunteers. 11 Our first aim was to verify that these hormonal

2 levels were associated with blood pressure changes and the risk of hypertension in a cohort showing substantial departures from the Framingham population. Our second aim was to test whether individuals with identical pressure levels had different odds of pressure increase and hypertension on a 5-year follow-up depending on their plasma aldosterone and renin levels. This latter point was important because different pressure levels trigger chronic cardiovascular and renal adaptations that can strongly influence the risk of pressure increase and hypertension and it is unclear whether adjustment for initial pressure levels in multivariable models can correct for these adaptations. 12 For this purpose, we chose to perform a nested case control study that also had the interest of maximizing the statistical power while minimizing the number of hormonal measurements. Our third aim was to analyse the potential modifying effect of habitual dietary sodium and potassium intakes on the ability of plasma aldosterone and renin to predict the risk of pressure increase and hypertension as these dietary factors influence both blood pressure and the plasma levels of these hormones. 13,14 Materials and methods Study population The SUpplémentation en VItamines et Minéraux AntioXydants study originally aimed to test whether adequate intake of antioxidants reduces the incidence of cancers and ischaemic cardiovascular diseases in middle-aged population. 11 A large majority (495%) of the candidates who volunteered for the study after the media campaign were of Caucasian origin. Among them, returned a completed questionnaire and informed consent and were eligible on the basis of the absence of overt disease that could have hindered participation to the study or threatened survival. Finally, participants (age range year) were enrolled in 1994/1995 throughout the metropolitan French territory and randomly allocated to receive either a combination of antioxidants at nutritional doses or a placebo for 7 years. In the present analysis, the supplemented and placebo groups were merged as antioxidant supplementation has been shown to have no effect on the risk of developing hypertension and on the incidence of cardiovascular disease during the course of the study. 11,15 Participants were invited to attend a clinical visit with blood sampling and physical examination in 1995/1996 (baseline) and 2001/2002 (mean followup ¼ 5.4±0.3 years). The number of participants at the first visit was 9243, including 6470 nonhypertensive. Among these non-hypertensive participants, 4011 attended the second visit (Supplementary Figure 1). Part of this attrition rate was explained by the fact that 174 participants died, 1009 withdrew consent and 736 were lost. The number of deaths of cardiovascular origin was 28 for a total of 583 cardiovascular events (74 and 29 of these events respectively occurred in the hypertensive and non-hypertensive groups that were defined for the present analysis). Blood pressure measurement During clinical visits that were performed by trained investigators in mobile units, blood pressure was taken once on each arm with a standard mercury sphygmomanometer and appropriate-sized cuff after 10 min of rest in a supine position. If mean systolic and diastolic pressures were below 160 and 100 mm Hg respectively, they were used for the analyses. Otherwise, blood pressure was measured again on each arm after additional 10 min of rest and the lowest mean was retained for the analyses. Participants also completed questionnaires reporting medication intake, health event, consultation and hospitalization. 11 Hypertension was defined by a systolic pressurex140 mm Hg and/or a diastolic pressurex90 mm Hg and/or the use of antihypertensive medication. Measurement of plasma renin and aldosterone levels Blood sampling was performed at the first clinical visit in recumbent participants who had been fasting for 12-h overnight. Measurement of plasma renin activity by radioimmunoassay (estimated by the amount of 125 I-angiotensin I generated during 1 h at 37 1C and ph 5.7) has been purposely selected instead of immunoreactive active renin quantification to avoid the uncertainties existing in some of these assays in the low range of plasma renin values. 16,17 was measured by radioimmunoassay with 125 I-aldosterone as a tracer (Coat-A-Count Aldosterone, Diagnostic Products Corporation, France). Detection thresholds were ng AI per ml per h for plasma renin activity and ng ml 1 for plasma aldosterone. Interassay (n ¼ 52) and intra-assay (n ¼ 6) coefficients of variation were 31 and 16% in the low concentration range and 25 and 9% in the high concentration range for plasma renin activity. The coefficients of variation were respectively 15 and 5% and 10 and 2% for plasma aldosterone. Estimation of habitual sodium and potassium intakes During follow-up, participants were requested to provide 24-h dietary record every 2 months by using an instruction manual for the codification of foods and the estimation of portion size. 18 The day of the record was randomly allocated to 2 weekend days and 4 weekdays in such a way that by the end of the study records were available for each day of the week in all seasons. The records collected between 551

3 552 the inclusion in the study in 1994/1995 and the first clinical visit in 1995/1996 (average of 8.0±3.5 records per individual) were used to estimate habitual dietary sodium and potassium intakes of participants at baseline (Supplementary Figure 2). During this period, intra-individual variability was very similar to inter-individual variability (s.d. and mean s.d. ¼ 62.4 and 67.4 mmol per 24 h and 21.9 and 21.3 mmol per 24 h for sodium and potassium intakes, respectively). Sodium and potassium intakes were calculated from food consumption using a food composition table issued by the French Centre Informatique sur la Qualité des Aliments. 19 Although sodium intake did not include discretionary salt, for which no reliable information was available, sodium and potassium intakes were significantly correlated with 24-h urinary excretions in a subset of 149 participants who provided three dietary records and 24-h urinary collections over a week for the purpose of validating food frequency questionnaire (r ¼ 0.27, P ¼ 0.05 for sodium and r ¼ 0.35, P ¼ 0.02 for potassium). The same records were used to estimate habitual alcohol consumption at baseline that was entered in multivariable models as another important confounder. These records included data on 37 alcoholic beverages that allowed to calculate alcohol consumption as g per 24-h. Statistical analyses At baseline, the associations of plasma aldosterone and renin and their ratio with the other characteristics of participants were tested by computing Pearson s unadjusted correlation coefficients and partial correlation coefficients that were obtained after adjusting for the other covariates. To test whether baseline plasma aldosterone and renin levels predict hypertension, cases and controls (participants becoming hypertensive or remaining non-hypertensive on follow-up) were individually matched on sex, and diastolic and systolic blood pressures (using 5 mm Hg classification) at baseline. Among 4011 non-hypertensive participants at baseline who attended the clinical visit at the end of follow-up, we were able to match 992 of 1058 cases to 992 controls randomly selected from 2953 participants who remained non-hypertensive (Supplementary Figure 1). Logistic regression models were used to examine the association of plasma aldosterone, renin and aldosterone/renin ratio with the risk of developing hypertension. Multivariable models were adjusted for sex, age, body mass index, habitual sodium and potassium intakes and alcohol consumption at baseline, percentage of body mass index change on follow-up, and included plasma aldosterone and renin separately or jointly as linear predictors. Multiple regression models were also used in the whole cohort (n ¼ 1984) to explore whether plasma aldosterone and renin predict systolic and diastolic blood pressure changes on follow-up. They were adjusted for the same factors than the logistic models and included plasma aldosterone and renin separately or jointly as predictors. Regression coefficients were standardized with the same scale of measurement to facilitate their comparison. To further test for effect modification by sodium and potassium intakes on blood pressure changes and on the risk of hypertension, these analyses were stratified by tertiles of sodium and potassium intakes. and renin levels and their ratio were considered as continuous variables and natural log transformed for the analyses because of positively skewed distributions (Supplementary Figure 2). All statistical analyses were performed with the statistical discovery software JMP 7 (SAS, Cary, NC, USA). Results Study participants The characteristics of participants at baseline and at the end of follow-up are shown in Table 1. Of importance for the aim of the present study, mean diastolic and systolic pressures at baseline were not different between control and case groups (P ¼ 0.19 Table 1 Characteristics of participants Case Control N Men (%) Baseline Age (year) 52.1± ±6.1 BMI (kg m 2 ) 24.6± ±3.1 SBP (mm Hg) 122.6± ±7.7 DBP (mm Hg) 78.7± ± ± ±8.1 (ng per 100 ml) Plasma renin 2.0± ±1.4 (ng AI per ml per h) Dietary Na intake 149.2± ±59.2 (mmol per 24 h) Dietary K intake 77.6± ±21.3 (mmol per 24 h) Alcohol consumption (g per 24 h) 18.5± ±19.0 End of follow-up BMI (kg m 2 ) 25.5± ±3.4 DBMI (kg m 2 ) 1.0± ±1.4 SBP (mm Hg) 139.5± ±8.0 DSBP (mm Hg) 16.9± ±9.2 DBP (mm Hg) 87.6± ±6.1 DDBP (mm Hg) 9.0± ±7.0 Antihypertensive treatment (%) Abbreviations: BMI, body mass index; DBP, diastolic blood pressure; SBP, systolic blood pressure. Values are means±s.d. Participants were randomly allocated to a case or a control group (individuals becoming hypertensive or remaining non-hypertensive on follow-up) using sex, SBP and DBP as matching factors.

4 and P ¼ 0.89, respectively) as a consequence of the individual matching of participants in these groups on sex and blood pressure levels. Correlations among the characteristics of participants at baseline The correlations computed in the whole cohort (n ¼ 1984) are reported in Table 2. After adjustment for the other variables, plasma aldosterone is positively correlated with plasma renin (Po0.0001), diastolic pressure (P ¼ 0.003) and potassium intake (P ¼ 0.04) and negatively correlated with sodium intake (Po0.0001). Plasma renin is negatively correlated with age (Po0.0001). Plasma aldosterone/renin ratio is positively associated with age (Po0.0001) and diastolic pressure (P ¼ 0.01) and negatively associated with sodium intake (P ¼ 0.001). In addition (not shown in Table 2), diastolic pressure is positively correlated with body mass index (r ¼ 0.14, Po0.0001) and sodium intake (r ¼ 0.06, P ¼ 0.03) and systolic pressure is positively correlated with age (r ¼ 0.16, Po0.0001) and diastolic pressure (r ¼ 0.56, Po0.0001). Prediction of blood pressure increase and the risk of hypertension by plasma aldosterone and renin levels at baseline is positively associated with systolic pressure increase on follow-up in multivariable models incorporating plasma aldosterone and renin together (P ¼ 0.01) but not in models including them separately (Table 3). Plasma renin is negatively associated with systolic pressure increase (P ¼ and in models incorporating plasma aldosterone and renin separately or together) whereas plasma aldosterone/renin ratio is positively associated with systolic pressure increase (P ¼ 0.009). Significant associations with the risk of hypertension are also observed in models incorporating plasma aldosterone and renin together but not in models including them separately. An s.d. increment in log aldosterone is associated with a 22% increase (P ¼ 0.04) in the odds of developing hypertension on follow-up. Although not reaching conventional statistical significance, an s.d. increment in log renin would be rather associated with a 16% decrease (P ¼ 0.07) in the risk of hypertension. Multivariable models incorporating plasma aldosterone/renin ratio show that an s.d. increment in the log ratio is associated with a 21% increase (P ¼ 0.02) in the risk of hypertension. Modifying effect of dietary sodium and potassium intakes on the ability of plasma aldosterone and renin to predict blood pressure increase and hypertension Multivariable models jointly incorporating plasma aldosterone and renin or their ratio as predictors of blood pressure changes and of the risk of hypertension on follow-up were tested by tertiles of baseline Table 2 Correlations of plasma aldosterone and renin levels with the other characteristics of participants at baseline Plasma renin Age BMI SBP DBP Dietary Na intake Dietary K intake Alcohol consumption Pairwise 0.37 (o0.0001) 0.05 (0.02) 0.05 (0.03) 0.03 (0.12) 0.06 (0.008) 0.10 (o0.0001) 0.01 (0.77) 0.03 (0.27) Partial 0.36 (o0.0001) 0.03 (0.23) 0.04 (0.11) 0.03 (0.27) 0.08 (0.003) 0.13 (o0.0001) 0.08 (0.004) 0.03 (0.32) Plasma rennin Pairwise 0.12 (o0.0001) 0.05 (0.03) 0.01 (0.64) 0.01 (0.84) 0.02 (0.46) 0.05 (0.04) 0.06 (0.01) Partial 0.13 (o0.0001) 0.02 (0.48) 0.03 (0.20) 0.03 (0.21) 0.03 (0.32) 0.03 (0.22) 0.04 (0.10) /rennin Pairwise 0.09 (o0.0001) 0.01 (0.91) 0.01 (0.54) 0.05 (0.02) 0.10 (o0.0001) 0.05 (0.03) 0.04 (0.11) Partial 0.11 (o0.0001) 0.01 (0.71) 0.04 (0.15) 0.07 (0.01) 0.09 (0.001) 0.02 (0.48) 0.02 (0.42) Abbreviations: BMI, body mass index; DBP, diastolic blood pressure; SBP, systolic blood pressure. Values are Pearson s unadjusted correlation coefficients or partial correlation coefficients computed in the whole cohort (n ¼ 1984) after adjustment for sex and all the other variables shown in the table. The statistical significance (P) is given in parenthesis. 553

5 554 Table 3 BP changes and risk of hypertension at 5 year according to plasma aldosterone and renin levels at baseline DSBP DDBP Hypertension Model b (s.e.) P b (s.e.) P OR (95% CI) P Age, sex 0.02 (0.69) (0.34) ( ) 0.62 Multivariable 0.02 (0.62) (0.42) ( ) 0.21 Multivariable, renin 0.06 (0.41) (0.47) ( ) 0.04 Plasma renin Age, sex 0.10 (0.47) o (0.32) ( ) Multivariable 0.09 (0.57) (0.38) ( ) 0.27 Multivariable, aldosterone 0.10 (0.61) (0.41) ( ) 0.07 /renin Age, sex 0.07 (0.44) (0.30) ( ) Multivariable 0.07 (0.53) (0.36) ( ) 0.02 Abbreviations: DBP, diastolic blood pressure; SBP, systolic blood pressure. b (s.e.) are standardized regression coefficients with the standard error and are computed in the whole cohort (n ¼ 1984). Odds ratios with 95% confidence interval (OR, 95% CI) are per s.d. increment in the natural log of plasma aldosterone, renin or aldosterone/renin ratio. Multivariable models include aldosterone and renin as predictors either separately or jointly and are adjusted for sex, age, BMI, dietary Na and K intakes and alcohol consumption at baseline and percentage of BMI change on follow-up. sodium and potassium intakes in the search for a potential modifying effect of these dietary factors (Table 4). and aldosterone/renin ratio are positively associated and plasma renin is negatively associated with systolic pressure change only in the lowest sodium intake tertile (P ¼ 0.01, and , respectively). Significant associations with diastolic pressure change (P ¼ 0.01 for plasma renin and P ¼ 0.04 for plasma aldosterone/ renin ratio) and with the risk of hypertension (59% increase (P ¼ 0.02), 31% decrease (P ¼ 0.03) and 31% increase (P ¼ 0.05) per s.d. increment in log aldosterone, log renin and log aldosterone/renin, respectively) are also observed in this tertile. Similarly, significant associations with systolic pressure change (P ¼ for plasma renin and P ¼ for plasma aldosterone/renin ratio) and with the risk of hypertension (69% increase (P ¼ 0.01), 28% decrease (P ¼ 0.05) and 49% increase (P ¼ 0.01) per s.d. increment in log aldosterone, log renin and log aldosterone/renin, respectively) are only observed in the lowest potassium intake tertile. Discussion The associations found in the Framingham cohort between plasma aldosterone and renin levels, blood pressure increase and the risk of hypertension are confirmed by the present study that was performed in a nationwide sample of French healthy volunteers differing from the North American cohort in several aspects. In particular, the French population sample is slightly younger (51.2 vs 55.6 years) and has a much lower body mass index (24.1 vs 27.0 kg m 2 ). Despite the difficulties expected in longitudinal studies exploring a single initial measurement of rapidly changing hormone levels (random errors increasing the noise, biasing the results towards the null hypothesis and opening the possibility of chance finding) and the use of different experimental and analytical designs, the two studies are in remarkable agreement to demonstrate that high plasma aldosterone and low renin levels (and consequently a high aldosterone/renin ratio) are present, even in individuals with similar pressure levels, before the rise in blood pressure occurs a few years later. Adjusted odds ratios per s.d. increment in log aldosterone and log renin are almost identical in both cohorts and indicate that a twofold increment in plasma aldosterone and renin is approximately associated with a 20% increase and 15% decrease in the odds of developing hypertension on a 3- to 5-year follow-up. 8,10 Significant associations of plasma aldosterone and renin and of their ratio with blood pressure increase are also reported in the two cohorts. In the present study, these associations are observed with both systolic and diastolic pressure increases and are partly confirmed by the cross-sectional analysis at baseline showing that plasma aldosterone is independently correlated with diastolic pressure. Although repeated measurements of daily sodium and potassium intakes and alcohol consumption should provide a more precise estimation of these dietary factors than does a single blood sampling to assess plasma aldosterone and renin levels, it is the instantaneous investigation of the hormonal status that best predicts the risk of hypertension and not sodium and potassium intakes or alcohol consumption that are not by themselves associated with the risk of hypertension (data not shown). This would

6 Table 4 Modifying effect of dietary Na and K intakes at baseline on BP changes and the risk of hypertension at 5 year 555 Dietary Na intake (mmol per 24 h) T1 (89.9±21.4) P T2 (139.2±11.9) P T3 (211.8±58.7) P DSBP (b, s.e.) 0.14 (1.27) (1.16) (1.17) 0.83 DDBP (b, s.e.) 0.03 (0.84) (0.78) (0.81) 0.65 Hypertension (OR, 95% CI) 1.59 ( ) ( ) ( ) 0.25 Plasma rennin DSBP (b, s.e.) 0.20 (1.19) (1.03) (0.98) 0.36 DDBP (b, s.e.) 0.15 (0.78) (0.69) (0.68) 0.91 Hypertension (OR, 95% CI) 0.69 ( ) ( ) ( ) 0.72 /rennin DSBP (b, s.e.) 0.20 (1.01) (0.90) (0.87) 0.58 DDBP (b, s.e.) 0.11 (0.67) (0.60) (0.60) 0.76 Hypertension (OR, 95% CI) 1.31 ( ) ( ) ( ) 0.27 Dietary K intake (mmol per 24 h) T1 (54.5±11.7) P T2 (76.8±5.1) P T3 (101.1±13.6) P DSBP (b, s.e.) 0.09 (1.15) (1.29) (1.18) 0.26 DDBP (b, s.e.) 0.01 (0.76) (0.86) (0.83) 0.31 Hypertension (OR, 95% CI) 1.69 ( ) ( ) ( ) 0.89 Plasma rennin DSBP (b, s.e.) 0.16 (1.16) (1.07) (0.96) 0.29 DDBP (b, s.e.) 0.05 (0.76) (0.72) (0.68) 0.73 Hypertension (OR, 95% CI) 0.72 ( ) ( ) ( ) 0.44 /rennin DSBP (b, s.e.) 0.15 (0.95) (0.97) (0.86) 0.81 DDBP (b, s.e.) 0.02 (0.62) (0.65) (0.61) 0.82 Hypertension (OR, 95% CI) 1.49 ( ) ( ) ( ) 0.62 b (s.e.) are standardized regression coefficients with the standard error and are computed in the whole cohort (n ¼ 1984). Odds ratios with 95% confidence interval (OR, 95% CI) are per s.d. increment in the natural log of plasma aldosterone, renin or aldosterone/renin ratio. Analyses are computed by tertiles of Na and K intakes (mean±s.d. of each tertile in parenthesis). Multivariable models include both aldosterone and renin as predictors and are adjusted for sex, age, BMI, alcohol consumption at baseline, percentage of BMI change on follow-up and either Na intake (when tested by tertiles of K intake) or K intake (when tested by tertiles of Na intake) at baseline. indicate that in the dietary environment shared by the present cohort, plasma renin and aldosterone levels are more important determinants of blood pressure changes with time, at least in middle-aged individuals. However, a high sodium intake is a critical factor in many studies that have reported inflammatory, fibrotic and thrombotic effects of aldosterone on the heart, brain and kidneys associated with or independently of the effect of dietary sodium on blood pressure. 20 In the Framingham study, the sodium/creatinine ratio obtained in a spot urine sample was explored as a simple assessment of exposure to salt. 8 No association was found with the incidence of hypertension, but the association of serum aldosterone with blood pressure changes was strengthened for persons whose urine sodium index was at or above the median, whereas it was no more statistically significant for those whose index was below the median. 8 In the present study, on the contrary, the associations of plasma aldosterone and renin levels with blood pressure increase and the risk of hypertension are only observed in the lowest tertile of sodium intake (o119 mmol per 24 h), as well as in the lowest tertile of potassium intake (o68 mmol per 24 h). The origin of this discrepancy between the two studies is unclear but could be related to the spot urine protocol used in the Framingham study that is unlikely to give a precise estimation of habitual sodium intake. The fact that the associations are only present when sodium and potassium intakes are low is difficult to interpret due to the strong and positive correlation (r ¼ 0.65) between these two dietary factors and to their opposite effect on aldosterone and renin release in the adrenals and kidneys. 14 The internal validity of hormonal measurements is emphasized by the magnitude of the expected positive correlation observed at baseline between two independently measured parameters, plasma aldosterone and its main regulatory factor, plasma

7 556 renin. This strong correlation probably explains why the associations of these hormonal parameters with blood pressure increase and the risk of hypertension are maximized in models including them jointly but not separately, a phenomenon that is also observed in the Framingham study. 10 Another expected negative correlation observed at baseline is between plasma renin and age and between plasma aldosterone and sodium intake. The absence of correlation between plasma renin and sodium intake is not surprising when sodium intake is mostly between 100 and 200 mmol per 24 h as in the present cohort (Supplementary Figure 2). The association described in several studies between plasma renin and sodium intake is based on the rapid increase in plasma renin observed when normal subjects are shifted from a very high to a very low sodium intake 21 and on the exponential relationship reported between plasma renin and sodium elimination in a single 24-h urinary collection on the day before blood sampling. 22,23 In our study, no urinary measurement was performed at the time of the initial exam, but the exposure to dietary sodium and potassium was periodically measured for 2 years. Although it misses discretionary salt added to food, this methodology that aims at estimating habitual intakes should greatly minimize intra-individual variability. Thus, whereas a third of the hypertensive patients referred to a hypertension clinic have values of urinary sodium less than 50 mmol per 24 h, such low sodium intakes were only observed in 2% of the individuals in the present cohort (Supplementary Figure 2). Many studies have reported plasma renin in relatively small groups of patients referred to care centres. 5,24 27 Most frequently, these cross-sectional studies have found a decrease of plasma renin in hypertensive subjects compared to normotensive controls, at least in adults. 28 To imply a participation of low renin levels to the occurrence of high blood pressure, a concept of inappropriate renin level in relation to blood pressure 14,29 and plasma angiotensin 30 has been proposed and vigorously challenged as a theoretical approach used to contradict evidence. 31 In this context, it is worth noting that the investigation of longitudinal blood pressure changes in community samples seems to confirm these clinical findings. This shows that despite some limitations inherent to their blood sampling conditions that are usually much less controlled than in hypertension clinics, especially concerning posture and diet, population studies can bring complementary information to that obtained in clinical investigations. The findings in the present cohort and in the Framingham population should not necessarily be interpreted to encourage the use of plasma aldosterone and renin measurements for estimating the risk of hypertension as pointed out before Rather, their main contribution is to add to the understanding of the role of the renin/aldosterone system among other factors in the progressive rise of blood pressure that occurs with age. Even though it is important to recall that observational studies describe statistical associations that do not imply causation, these data support the hypothesis that abnormally high or low but inappropriate plasma aldosterone or renin levels are primary aetiological factors of hypertension. This view is supported by the fall in blood pressure consistently induced by all the drugs that inhibit the renin cascade or block the mineralocorticoid receptor or the aldosterone-induced proteins, even in the presence of a high sodium intake. 35 Mechanistically, the opposite effect of plasma aldosterone and renin levels on the subsequent rise in blood pressure remains to be explained. The endothelial dysfunction with the accompanying decrease in nitric oxide production that occurs with age in a context of inflammation might be an interesting path to explore. 36 Indeed, nitric oxide production has been shown both to increase renin release and to decrease aldosterone secretion and a deficiency in its production may therefore lead to a pattern of high plasma aldosterone and low renin levels. 37,38 Both studies also point out to a stronger participation of aldosterone in the development of human hypertension than expected from the investigations in animal models that have more emphasized the role of renin and angiotensin. 39 They suggest that a pharmacological therapy aimed at decreasing aldosterone synthesis would be useful if the correction of a high aldosterone status present before blood pressure rise is able to prevent it, as does a renin angiotensin system pharmacological blockade. 40 What is known about this topic K and renin levels have been associated with blood pressure increase and 3 4 year incidence of hypertension only once in a middle-aged North American community. K These findings remained uncertain due to the large intraindividual variability of plasma aldosterone and renin and blood pressure and to the fact that single measurements were used to estimate these phenotypes. K In the same community, only three biomarkers (plasma c- reactive protein and plasminogen activator inhibitor-1 and urinary albumin/creatinine ratio) among nine eligible (including plasma aldosterone and renin) were retained in a multivariable model used to predict the risk of hypertension. What this study adds K The present study confirms in a French community sample that plasma aldosterone and renin are predictors of blood pressure increase and of the risk of hypertension at 5 year. K It shows that these correlations are observed in individuals with low dietary sodium and potassium intakes, that is, in conditions where the renin aldosterone system is upregulated.

8 Acknowledgements We are grateful to Dr Tanh Tam Guyenne (INSERM U872) for performing aldosterone assays and to Mrs Marie-Françoise Gonzalez (INSERM U872) for performing plasma renin activity assays. We thank the Fondation Robert Debré and the Institut National pour la Santé et la Recherche Médicale for their funding. References 1 Laragh JH. Vasoconstriction-volume analysis for understanding and treating hypertension: the use of renin and aldosterone profiles. Am J Med 1973; 55(3): Laragh JH. Renin profiling for diagnosis, risk assessment, and treatment of hypertension. Kidney Int 1993; 44(5): Blaufox MD, Lee HB, Davis B, Oberman A, Wassertheil-Smoller S, Langford H. Renin predicts diastolic blood pressure response to nonpharmacologic and pharmacologic therapy. JAMA 1992; 267(9): Preston RA, Materson BJ, Reda DJ, Williams DW, Hamburger RJ, Cushman WC et al. Age-race subgroup compared with renin profile as predictors of blood pressure response to antihypertensive therapy. JAMA 1998; 280(13): Meade TW, Imeson JD, Gordon D, Peart WS. The epidemiology of plasma renin. Clin Sci (London) 1983; 64(3): Meade TW, Cooper JA, Peart WS. Plasma renin activity and ischemic heart disease. N Engl J Med 1993; 329(9): Alderman MH, Madhavan S, Ooi WL, Cohen H, Sealey JE, Laragh JH. Association of the renin-sodium profile with the risk of myocardial infarction in patients with hypertension. N Engl J Med 1991; 324(16): Vasan RS, Evans JC, Larson MG, Wilson PW, Meigs JB, Rifai N et al. Serum aldosterone and the incidence of hypertension in nonhypertensive persons. N Engl J Med 2004; 351(1): Wang TJ, Gona P, Larson MG, Levy D, Benjamin EJ, Tofler GH et al. Multiple biomarkers and the risk of incident hypertension. Hypertension 2007; 49(3): Newton-Cheh C, Guo CY, Gona P, Larson MG, Benjamin EJ, Wang TJ et al. Clinical and genetic correlates of aldosterone-to-renin ratio and relations to blood pressure in a community sample. Hypertension 2007; 49(4): Hercberg S, Galan P, Preziosi P, Bertrais S, Mennen L, Malvy D et al. The SU.VI. MAX study: a randomized, placebo-controlled trial of the health effects of antioxidant vitamins and minerals. Arch Intern Med 2004; 164(21): de Wardener HE, MacGregor GA. Harmful effects of dietary salt in addition to hypertension. J Hum Hypertens 2002; 16(4): Meneton P, Jeunemaitre X, de Wardener HE, MacGregor GA. Links between dietary salt intake, renal salt handling, blood pressure, and cardiovascular diseases. Physiol Rev 2005; 85(2): Sealey JE, Laragh JH. The renin angiotensin aldosterone system for normal regulation of blood pressure and sodium and potassium homeostasis.in: Laragh JH, Brenner BM (eds). Hypertension: Pathophysiology, Diagnosis and Management. Raven Press: New York, 1995,pp Czernichow S, Bertrais S, Blacher J, Galan P, Briancon S, Favier A et al. Effect of supplementation with antioxidants upon long-term risk of hypertension in the SU.VI. MAX study: association with plasma antioxidant levels. JHypertens2005; 23(11): Menard J, Guyenne TT, Corvol P, Pau B, Simon D, Roncucci R. Direct immunometric assay of active renin in human plasma. J Hypertens Suppl 1985; 3(Suppl 3): S275 S Sealey JE. Plasma renin activity and plasma prorenin assays. Clin Chem 1991; 37(10 Part 2): Mennen LI, Bertrais S, Galan P, Arnault N, Potier de Couray G, Hercberg S. The use of computerised 24 h dietary recalls in the French SU.VI. MAX study: number of recalls required. Eur J Clin Nutr 2002; 56(7): Feinberg M, Favier CJ, Ireland-Ripert J. Répertoire général des aliments. Technique & Documentation- Lavoisier: Paris, 1991, 282pp. 20 Dluhy RG, Williams GH. Aldosterone villain or bystander? N Engl J Med 2004; 351(1): Tuck ML, Dluhy RG, Williams GH. Sequential responses of the renin angiotensin aldosterone axis to acute postural change: effect of dietary sodium. J Lab Clin Med 1975; 86(5): Brunner HR, Laragh JH, Baer L, Newton MA, Goodwin FT, Krakoff LR et al. Essential hypertension: renin and aldosterone, heart attack and stroke. N Engl J Med 1972; 286(9): Rosei EA, Brown JJ, Cumming AM, Fraser R, Semple PF, Lever AF et al. Is the sodium index a useful way of expressing clinical plasma renin, angiotensin and aldosterone values? Clin Endocrinol (Oxf) 1978; 8(2): Birkenhager WH, Schalekamp MA, Krauss XH, Kolsters G, Schalekamp-Kuyken MP, Kroon BJ et al. Systemic and renal haemodynamics, body fluids and renin in benign essential hypertension with special reference to natural history. Eur J Clin Invest 1972; 2(2): Grim CE, Weinberger MH, Henry DP, Luft FC, Fineberg NS. Biochemical correlates of the increase in blood pressure with age. Clin Sci Mol Med Suppl 1978; 4: 377s 379s. 26 Thomas GW, Ledingham JG, Beilin LJ, Stott AN, Yeates KM. Reduced renin activity in essential hypertension: a reappraisal. Kidney Int 1978; 13(6): Thurston H, Bing RF, Pohl JE, Swales JD. Renin subgroups in essential hypertension: an analysis and critique. Q J Med 1978; 47(187): Vetter H, Glanzer K, Vetter W. Essential hypertension: relationship between renin and blood pressure. Clin Exp Hypertens 1980; 2(3 4): Alderman MH, Cohen HW, Sealey JE, Laragh JH. Plasma renin activity levels in hypertensive persons: their wide range and lack of suppression in diabetic and in most elderly patients. Am J Hypertens 2004; 17(1): Williams GH, Hollenberg NK, Moore TJ, Dluhy RG, Bavli SZ, Solomon HS et al. Failure of renin suppression by angiotensin II in hypertension. Circ Res 1978; 42(1):

9 Swales JD. On the inappropriate in hypertension research. Lancet 1977; 2(8040): Wald NJ, Hackshaw AK, Frost CD. When can a risk factor be used as a worthwhile screening test? BMJ 1999; 319(7224): Pepe MS, Janes H, Longton G, Leisenring W, Newcomb P. Limitations of the odds ratio in gauging the performance of a diagnostic, prognostic, or screening marker. Am J Epidemiol 2004; 159(9): D Agostino RB. Risk prediction and finding new independent prognostic factors. J Hypertens 2006; 24(4): Azizi M, Bissery A, Bura-Riviere A, Menard J. Dual renin angiotensin system blockade restores blood pressure-renin dependency in individuals with low renin concentrations. J Hypertens 2003; 21(10): Bautista LE. Inflammation, endothelial dysfunction, and the risk of high blood pressure: epidemiologic and biological evidence. J Hum Hypertens 2003; 17(4): Kurtz A, Wagner C. Role of nitric oxide in the control of renin secretion. Am J Physiol 1998; 275(6 Part 2): F849 F Muldowney III JA, Davis SN, Vaughan DE, Brown NJ. NO synthase inhibition increases aldosterone in humans. Hypertension 2004; 44(5): Freel EM, Connell JM. Mechanisms of hypertension: the expanding role of aldosterone. J Am Soc Nephrol 2004; 15(8): Julius S, Nesbitt SD, Egan BM, Weber MA, Michelson EL, Kaciroti N et al. Feasibility of treating prehypertension with an angiotensin-receptor blocker. N Engl J Med 2006; 354(16): Supplementary Information accompanies the paper on the website (