Impact of Kidney Disease and Blood Pressure on the Development of Cardiovascular Disease
An Overview From the Japan Arteriosclerosis Longitudinal Study
Background— Kidney disease is associated with an increased risk of cardiovascular disease (CVD); however, there have been few well-designed prospective studies of this issue in Asian populations. Recent epidemiological studies have suggested that a lower blood pressure level may be associated with an increased risk of CVD in individuals with kidney dysfunction.
Methods and Results— Using data from 10 community-based cohort studies in Japan, we conducted follow-up on a total of 30 657 individuals 40 to 89 years of age without preexisting CVD or kidney failure and examined the relationship between reduced glomerular filtration rate (GFR) and the risk of CVD. During an average 7.4-year follow-up, 727 individuals experienced CVD. The age- and sex-adjusted incidence of CVD increased significantly in subjects with GFR of 60 to 89 mL · min−1 · 1.73 m−2 (4.3 per 1000 person-years, P=0.002) and in those with a GFR <60 mL · min−1 · 1.73 m−2 (6.5, P<0.001) compared with those with a GFR ≥90 mL · min−1 · 1.73 m−2 (2.9). Even after adjustment for potential confounding factors, subjects with a GFR <60 mL · min−1 · 1.73 m−2 had a 57% (95% CI 14% to 115%) greater risk of CVD than those with a GFR ≥90 mL · min−1 · 1.73 m−2. The multivariate-adjusted hazard ratios of CVD increased in a log-linear manner with elevations in blood pressure levels, regardless of GFR levels (all P for trend <0.01).
Conclusions— Our findings suggest that a reduced GFR is a significant risk factor for CVD in the general Japanese population. Additionally, a log-linear association of blood pressure level with CVD risk was observed, without evidence of a J-curve association, regardless of GFR levels.
Kidney disease is increasingly being recognized as a leading public health issue. Chronic kidney disease, most commonly defined by a reduction in glomerular filtration rate (GFR) or the presence of proteinuria, affects 10% to 15% of the adult population in Western countries1,2 and is associated with an increased risk of cardiovascular disease (CVD)3–5; however, there have been few well-designed large prospective studies in general Asian communities to date.6–8
Clinical Perspective p 2701
Blood pressure is an important determinant of the risk of CVD in the general population,9,10 in which it has been well established that treatment for high blood pressure prevents CVD.11 Blood pressure is commonly elevated in individuals with a reduced GFR,4,5 which suggests that lowering blood pressure may offer significant benefits in this population. Recent prospective cohort studies, however, have reported that the risk of stroke or death for individuals with a reduced GFR is greater among those with systolic blood pressure levels below 120 mm Hg than among those with higher levels.12,13 These data have raised concerns that lowering blood pressure may provide less benefit than previously believed, or may even be hazardous, in individuals with kidney dysfunction.
In the present report, we discuss the results of a pooling analysis from the Japan Arteriosclerosis Longitudinal Study–Existing Cohorts Combine (JALS-ECC), which is an overview of individual participant data from 21 community-based longitudinal observational studies in Japan.14 Our aims were to assess the impact of a reduction in GFR on the development of CVD in the general population and to examine the association of blood pressure with the risk of CVD in individuals with a reduced GFR.
The rationale, study design, and methods of the JALS-ECC have been described elsewhere.14 In brief, cohort studies were eligible for inclusion in this project if they satisfied the following criteria: (1) Japanese population; (2) prospective cohort study; (3) at least 3000 person-years of follow-up; (4) date of birth, sex, height, weight, blood pressure, and serum total cholesterol recorded at baseline; and (5) date of death or the age at death recorded during a follow-up. Quality control of the collected cohort data were performed at the JALS Coordinating Center. The individual records of 66 691 participants in 21 cohort studies were included in the present project, with 82.7% of the participants from 17 community-based cohorts and 17.3% from 4 work-site–based cohorts. Permission to submit each collection of cohort data to the JALS Coordinating Center was obtained from the relevant institutional review boards for ethical issues.
Of the 21 cohort studies, 11 cohorts were excluded from the present analysis for the following reasons: 4 were work-site–based cohorts, 3 did not include creatinine data, 3 lacked many values for relevant variables, and 1 included no event data for either stroke or myocardial infarction. From the remaining 10 cohorts, we excluded participants less than 40 years of age and those 90 years of age or older, those with unavailable examination data at baseline or unavailable event data, those with a history of CVD, and those with an estimated GFR <15 mL · min−1 · 1.73 m−2. For the analysis regarding CVD and stroke, the Niigata cohort was excluded, because information on stroke events was unavailable. A final total of 23 033 participants were enrolled in the CVD analysis, 23 084 in the stroke analysis, 30 657 in the myocardial infarction analysis, and 31 374 in the all-cause death analysis. The average follow-up period was 7.4 years.
The JALS Coordinating Center requested individual participant data from the collaborating investigators. Serum creatinine was measured by Jaffe’s method in 8 cohorts, by the enzymatic method in 1 cohort, and by both methods in 1 cohort. Serum creatinine values measured by the enzymatic method were corrected for Japanese subjects by the addition of 18.3 μmol/L.15 GFR was estimated with the 4-variable Modification of Diet in Renal Disease study equation.3 In accordance with the National Kidney Foundation Kidney Disease Outcomes Quality Initiative guidelines,16 GFR levels were classified in the following ranges: ≥90, 60 to 89, and <60 mL · min−1 · 1.73 m−2. Blood pressure was measured by a standard sphygmomanometer in all cohorts. Mean values were used in several cohorts that measured 2 or more blood pressure values. Blood pressure levels at baseline were classified into 4 categories (normal, prehypertension, stage 1 hypertension, and stage 2 hypertension) according to the criteria of the seventh report of the Joint National Committee on the Prevention, Detection, Evaluation, and Treatment of High Blood Pressure.17 Diabetes was defined as a fasting blood glucose level of ≥7.0 mmol/L, a casual blood glucose level of ≥11.1 mmol/L, current use of insulin or oral medication for diabetes, and/or a history of diabetes. Serum total cholesterol was determined enzymatically. Information on smoking habit was obtained through a standard questionnaire and classified as current habitual use or lack thereof.
In each cohort, vital status and the development of CVD were ascertained during follow-up by use of death certificates, hospital medical records, and/or questionnaire surveys. All outcomes were classified according to the International Classification of Diseases, 9th Revision (ICD-9). All events were recoded by coordinating center staff members. CVD was defined as the development of either stroke or myocardial infarction. Stroke was defined as an acute disturbance of focal neurological function with symptoms that lasted >24 hours or death caused by a stroke event (ICD-9 codes 430, 431, 433, 434, or 436). Myocardial infarction included both fatal and nonfatal myocardial infarction, which was diagnosed by use of an appropriate clinical history supported by ECG changes and/or elevations of cardiac enzymes or other biochemical markers of myocardial injury (ICD-9 410). Only the first event of the relevant outcome type was included in each analysis.
The SAS software package for Windows, release 9.13 (SAS Institute, Inc, Cary, NC) was used to perform all statistical analyses. The incidence rate of each outcome for the GFR subgroups was calculated by the person-year method and adjusted for the age and sex distribution of the overall population enrolled in the CVD analysis by the direct method, in which the subgroups and study population were subdivided into the same set of age groups (defined by decade) and the age- and sex-specific incidence rates were calculated within each subgroup.18 The hazard ratios (HRs) and their 95% CIs for the development of events were estimated with the Cox proportional hazards regression model. The cohort effect was adjusted as a fixed effect by taking the study as a strata variable, assuming only proportional hazards within each study and not between studies.18 Heterogeneity across cohorts was examined with the Cochran Q test and the I2 statistic.18 The risks of events according to blood pressure levels were also estimated with the Cox regression model. Trends in relationship between blood pressure levels and the risk of events were assessed by fitting models with a linear term for blood pressure categories according to kidney function status, and the heterogeneity of these relationships between kidney function status subgroups was estimated by the addition of an interaction term of a linear term for blood pressure levels and kidney function status to the relevant model. P<0.05 was considered statistically significant in all analyses.
The authors had full access to and take full responsibility for the integrity of the data. All authors have read and agree to the manuscript as written.
The characteristics of the 10 cohorts examined in the present study are shown in Table 1. Among all subjects, the mean age was 57.6 years, and the proportion of men was 38.0%. The mean value of serum creatinine was 78.6 μmol/L, and the frequency of GFR <60 mL · min−1 · 1.73 m−2 was 8.2%. During the average follow-up period of 7.4 years, a total of 727 subjects experienced CVD, 592 had strokes, 180 had myocardial infarctions, and a total of 2104 died.
Table 2 shows the baseline characteristics of the 23 033 subjects enrolled in the CVD analysis by sex. Their mean ages were 56.9 years for men and 58.2 years for women, and the frequency of GFR <60 mL · min−1 · 1.73 m−2 was 5.2% for men and 10.1% for women. The frequencies of normal blood pressure, prehypertension, stage 1 hypertension, and stage 2 hypertension were 19.6%, 41.8%, 26.0%, and 12.6% for men and 24.3%, 41.0%, 24.3%, and 10.4% for women, respectively. Similar findings were observed in subjects enrolled in the analyses of stroke, myocardial infarction, and all-cause death.
The age- and sex-adjusted incidences of CVD and stroke increased with declining GFR levels in the overall population (Table 3); the differences were statistically significant between subjects with a GFR ≥90 mL · min−1 · 1.73 m−2 and those with a GFR <90 mL · min−1 · 1.73 m−2 (all P<0.01). Subjects with a GFR <60 mL · min−1 · 1.73 m−2 showed a significantly higher age- and sex-adjusted incidence of myocardial infarction and all-cause mortality than those with a GFR ≥90 mL · min−1 · 1.73 m−2 (P<0.001). The age-adjusted incidences of CVD, stroke, and all-cause mortality were significantly higher in subjects with a GFR <60 mL · min−1 · 1.73 m−2 than in those with a GFR ≥90 mL · min−1 · 1.73 m−2 in both sexes (all P<0.05).
The risks of CVD, stroke, myocardial infarction, and all-cause death increased progressively with declining GFR levels in the overall population after adjustment for age and sex (Table 4). Even after adjustment for potential confounding factors, specifically age, sex, cohort, systolic blood pressure, diabetes, serum total cholesterol, body mass index, and current smoking status, the risks of CVD, myocardial infarction, and all-cause death were significantly higher in subjects with a GFR <60 mL · min−1 · 1.73 m−2 than in the overall population. There was no evidence of heterogeneity in these associations among study cohorts (all P for heterogeneity >0.6; Q=2.46, I2=0% for CVD; Q=4.06, I2=0% for stroke; Q=3.75, I2=0% for myocardial infarction; and Q=1.14, I2=0% for all-cause death). Subjects with a GFR <60 mL · min−1 · 1.73 m−2 had a significantly greater risk of myocardial infarction and death in men and of CVD, stroke, and death in women.
The Figure shows the log-linear relationship between blood pressure levels at baseline and the hazard of CVD, stroke, and all-cause death regardless of kidney function status after adjustment for potential confounding factors (all P for trend <0.01). There was no evidence of heterogeneity of the patterns in the association of blood pressure levels with the risk of outcomes between subgroups of kidney function status (all P for heterogeneity >0.7). The age- and sex-adjusted HR of myocardial infarction increased in a log-linear fashion with increasing blood pressure levels in the normal, prehypertension, stage 1 hypertension, and stage 2 hypertension groups in subjects with a GFR ≥60 mL · min−1 · 1.73 m−2 (HR 0.56 [95% CI 0.33 to 0.95], 1.00 [reference], 1.60 [1.08 to 2.37], and 1.75 [1.06 to 2.87]; P for trend 0.03) and in those with a GFR <60 mL · min−1 · 1.73 m−2 (0.19 [0.02 to 1.47], 1.00 [reference], 1.72 [0.80 to 3.70], and 2.36 [1.02 to 5.44]; P for trend 0.04). The number of myocardial infarctions in subjects with normal blood pressure levels was too small to assess reliably for multivariate-adjusted analysis.
We also performed sensitivity analyses to assess the risk of CVD according to GFR levels estimated by the MDRD formula corrected according to the Japanese coefficient of 0.881.15 The correction shifted the GFR distribution to a lower level. Consequently, more participants (21%) were assigned to the group whose GFR was <60 mL · min−1 · 1.73 m−2, and the age- and sex-adjusted risk of CVD among these subjects relative to those with a GFR ≥90 mL · min−1 · 1.73 m−2 was attenuated by 85% (95% CI 32% to 160%), although it was still significant. Similarly, a log-linear relationship between blood pressure levels and the risk of CVD was still observed in the subgroup whose GFR was <60 mL · min−1 · 1.73 m−2, even after correction with the Japanese coefficient (Data Supplement Figure).
In the present study, we demonstrated a clear association between reduced GFR and high risk of CVD. To the best of our knowledge, this is the first overview of this issue in a Japanese community-based longitudinal study. Furthermore, the relationship between blood pressure levels at baseline and CVD risk was found to be strong and continuous, regardless of kidney function status.
There have been few studies showing the association of reduced GFR with an increased risk of CVD or mortality in the general Japanese population.6–8 The findings of the Hisayama study revealed that a GFR <60 mL · min−1 · 1.73 m−2 was a significant risk factor for the development of coronary heart disease in men and of CVD and stroke in women.6 In a large cohort study conducted by Irie et al,7 reduced GFR was strongly associated with mortality due to CVD or stroke. A report from NIPPON DATA 90 also showed an association between a GFR <30 mL · min−1 · 1.73 m−2 and a high risk of cardiovascular death.8 In the present study, we demonstrated a clear association between reduced GFR and the risks of CVD, stroke, myocardial infarction, and death in an overview of 10 Japanese cohort studies. These results, therefore, highlight the importance of taking kidney function status into consideration in trying to reduce the burden of CVD in the general Japanese population.
There are several possible explanations for the association of reduced GFR with CVD.3 First, reduced GFR is associated with a high prevalence of traditional CVD risk factors, such as aging, hypertension, diabetes, smoking habits, and dyslipidemia.19 In the present study, reduced GFR was found to be a significant risk factor for the development of stroke after adjustment for demographic factors, but not after adjustment for potential traditional CVD risk factors, which suggests that an accumulation of traditional CVD risk factors in individuals with reduced GFR increases the risk of stroke. In contrast, the risks of CVD, myocardial infarction, and all-cause death in individuals with reduced GFR were also attenuated, although still significant, after adjustment for traditional CVD risk factors. Reduced GFR has been shown to be associated with increased levels of novel CVD risk factors, such as inflammation, asymmetric dimethylarginine, oxidative stress, and thrombogenic factors.19,20 Second, reduced GFR may be a marker of vascular disease; it is well recognized that renal arteriosclerosis and glomerular sclerosis are closely related to systemic atherosclerosis.21
In the present study, reduced GFR was associated with a high risk of stroke in men after adjustment for demographic factors but not after adjustment for potential confounding factors; however, this relationship was still observed in women even after adjustment for confounding factors. This sex difference may be a consequence of the effects of residual confounding factors, specifically, hypercoagulable states22 or gonadal steroids,23 in women. Furthermore, the lack of a significant association between reduced GFR and a high risk of myocardial infarction is probably due to the relatively small number of events.
In the present study, we demonstrated a clear log-linear association between blood pressure levels and the risks of CVD, stroke, and all-cause death, regardless of kidney function status. These findings are consistent with the results of other studies conducted in the general population.9,10 Recent publications of prospective cohort data suggest, however, that individuals with a reduced GFR and a systolic blood pressure below 120 mm Hg may be at increased risk of stroke or death.12,13 Other post hoc analyses of trials conducted on individuals with coronary heart disease24 and with diabetic nephropathy25 suggest an increased risk of coronary events at the lower achieved blood pressures. In the present study, however, no evidence of an increased risk of myocardial infarction was observed at the lower blood pressure level. One possible explanation for the J-curve association observed in the previous studies may be the phenomenon of reverse causality,26 in which extensive vascular disease or subclinical cardiac dysfunction is associated with lower blood pressure levels and reduced GFR and is associated independently with a relatively high risk of CVD, rather than with any adverse effects of low blood pressure itself.
Several limitations of the present study should be noted. First, the generalizability of our findings to some populations at high risk for CVD may be limited. The participants excluded from the analysis due to missing baseline examination data or event data were likely to have a higher cardiovascular risk, because they were older (mean 63 years), had higher blood pressure levels (mean 138/80 mm Hg), and had a greater prevalence of diabetes (8.7%) than the study population. This bias has the potential to alter our findings, which may therefore be conservative. Second, the present GFR estimates, which were made with a simplified prediction equation, may not be sufficiently correct, which possibly could lead to a certain number of misclassifications of estimated kidney function status. Such misclassifications would weaken the association found in the present study, biasing the results toward the null hypothesis. Third, we were unable to obtain information regarding the use of antihypertensive drugs, medication compliance, or blood pressure control during the follow-up period. The lack of this information may reduce the accuracy of our findings to some extent. Fourth, the applicability of the present results to populations with severe kidney dysfunction is limited, because very few of our subjects (0.1%) had a GFR <30 mL · min−1 · 1.73 m−2. Moreover, the absence of data on proteinuria in the present study makes it impossible to assess the effects of the earliest stages of kidney disease on the risk of CVD. Finally, creatinine measurement was conducted locally rather than at a central laboratory, which introduces a certain amount of variability that may reduce the reliability of the results.
In conclusion, the present findings suggest that a reduced GFR is associated significantly with a high risk of CVD in the general Japanese population. Furthermore, we observed a continuous relationship between blood pressure levels at baseline and the risk of CVD, regardless of kidney function status. The optimization of blood pressure control in individuals with kidney dysfunction is therefore likely to substantially reduce the burden of CVD in the general population.
The authors gratefully acknowledge the efforts of the JALS investigators, research coordinators, and committee members. A list of these individuals has been published previously14 and is provided in the Appendix (Data Supplement).
Sources of Funding
This study was supported by a research grant from Japan Arteriosclerosis Prevention Fund.
Chadban SJ, Briganti EM, Kerr PG, Dunstan DW, Welborn TA, Zimmet PZ, Atkins RC. Prevalence of kidney damage in Australian adults: the AusDiab kidney study. J Am Soc Nephrol. 2003; 14 (suppl 2): S131–S138.
Sarnak MJ, Levey AS, Schoolwerth AC, Coresh J, Culleton B, Hamm LL, McCullough PA, Kasiske BL, Kelepouris E, Klag MJ, Parfrey P, Pfeffer M, Raij L, Spinosa DJ, Wilson PW; American Heart Association Councils on Kidney in Cardiovascular Disease, High Blood Pressure Research, Clinical Cardiology, and Epidemiology and Prevention. Kidney disease as a risk factor for development of cardiovascular disease: a statement from the American Heart Association Councils on Kidney in Cardiovascular Disease, High Blood Pressure Research, Clinical Cardiology, and Epidemiology and Prevention. Circulation. 2003; 108: 2154–2169.
Irie F, Iso H, Sairenchi T, Fukasawa N, Yamagishi K, Ikehara S, Kanashiki M, Saito Y, Ota H, Nose T. The relationships of proteinuria, serum creatinine, glomerular filtration rate with cardiovascular disease mortality in Japanese general population. Kidney Int. 2006; 69: 1264–1271.
Nakamura K, Okamura T, Hayakawa T, Kadowaki T, Kita Y, Ohnishi H, Saitoh S, Sakata K, Okayama A, Ueshima H; the NIPPON DATA90 Research Group. Chronic kidney disease is a risk factor for cardiovascular death in a community-based population in Japan: NIPPON DATA90. Circ J. 2006; 70: 954–959.
MacMahon S, Peto R, Cutler J, Collins R, Sorlie P, Neaton J, Abbott R, Godwin J, Dyer A, Stamler J. Blood pressure, stroke, and coronary heart disease: part 1, prolonged differences in blood pressure: prospective observational studies corrected for the regression dilution bias. Lancet. 1990; 335: 765–774.
Weiner DE, Tighiouart H, Levey AS, Elsayed E, Griffith JL, Salem DN, Sarnak MJ. Lowest systolic blood pressure is associated with stroke in stages 3 to 4 chronic kidney disease. J Am Soc Nephrol. 2007; 18: 960–966.
Kovesdy CP, Trivedi BK, Kalantar-Zadeh K, Anderson JE. Association of low blood pressure with increased mortality in patients with moderate to severe chronic kidney disease. Nephrol Dial Transplant. 2006; 21: 1257–1262.
Imai E, Horio M, Nitta K, Yamagata K, Iseki K, Hara S, Ura N, Kiyohara Y, Hirakata H, Watanabe T, Moriyama T, Ando Y, Inaguma D, Narita I, Iso H, Wakai K, Yasuda Y, Tsukamoto Y, Ito S, Makino H, Hishida A, Matsuo S. Estimation of glomerular filtration rate by the MDRD study equation modified for Japanese patients with chronic kidney disease. Clin Exp Nephrol. 2007; 11: 41–50.
Chobanian AV, Bakris GL, Black HR, Cushman WC, Green LA, Izzo JL Jr, Jones DW, Materson BJ, Oparil S, Wright JT Jr, Roccella EJ; Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure. National Heart, Lung, and Blood Institute; National High Blood Pressure Education Program Coordinating Committee: seventh report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure. Hypertension. 2003; 42: 1206–1252.
Woodward M. Epidemiology: Study Design and Data Analysis. 2nd ed. Boca Raton, Fla: Chapman & Hall/CRC Press; 2005.
Tracy RP, Arnold AM, Ettinger W, Fried L, Meilahn E, Savage P. The relationship of fibrinogen and factors VII and VIII to incident cardiovascular disease and death in the elderly: results from the cardiovascular health study. Arterioscler Thromb Vasc Biol. 1999; 19: 1776–1783.
Berl T, Hunsicker LG, Lewis JB, Pfeffer MA, Porush JG, Rouleau JL, Drury PL, Esmatjes E, Hricik D, Pohl M, Raz I, Vanhille P, Wiegmann TB, Wolfe BM, Locatelli F, Goldhaber SZ, Lewis EJ; the Collaborative Study Group. Impact of achieved blood pressure on cardiovascular outcomes in the Irbesartan Diabetic Nephropathy Trial. J Am Soc Nephrol. 2005; 16: 2170–2179.
There have been several studies reporting a strong association between reduced kidney function and cardiovascular risk. The findings, however, have been inconsistent in Asian populations, and there has been no attempt to date to review the evidence. Hence, we conducted an overview of individual participant data from Japanese community-based cohort studies to reliably assess the impact of reduced kidney function on cardiovascular risk in the general Japanese population. Our findings suggest a clear association between reduced kidney function and a 57% greater risk of cardiovascular disease in the Japanese population, as well as a log-linear relationship between blood pressure levels and cardiovascular risk in individuals with reduced kidney function. The optimization of blood pressure control in individuals with reduced kidney function is therefore likely to substantially reduce the burden of cardiovascular disease in the general population. Given that the prevalence of reduced kidney function is ≈10% in the general population, we believe that these novel findings are significant in the areas of clinical and public health.
Continuing medical education (CME) credit is available for this article. Go to http://cme.ahajournals.org to take the quiz.
The online-only Data Supplement is available with this article at http://circ.ahajournals.org/cgi/content/full/118/25/2694/DC1.