Risk Factors for the Development of Left Ventricular Hypertrophy in a rospectively Followed Cohort of Dialysis I John D. Harnett2, Gloria M. Kent, aul E. Bane, Rhoda Taylor, and atrick S. arfrey J.D. Harnett, G.M. Kent,.S. arfrey, The Division of Nephrology, The Health Sciences Centre, Memorial University of Newfoundland, St. John s, Newfoundland, Canada.E. Barre, I?. Taylor, Royal Victoria Hospital, McGill University, Montreal, Quebec, Canada (J. Am. Soc. Nephrol. 1994; 4:1486-1490) ABSTRACT The objective of this study was to determine the role of hypertension, age, anemia, and hyperparathyroidism in the pathogenesis of left ventricular hypertrophy (LVH) developing after the initiation of dialysis for ESRD. A cohort of dialysis patients who were being treated for ESRD and whose initial echocardiograms after the start of dialysis therapy do not show LVH were studied. Three hundred and thirty-nine patients have been monitored at three centers since 1985. Serial echocardiograms have been performed with M-mode and two-dimensional echocardiography. Data on blood pressure, height, weight, hemoglobin, number and type of antihypertensive medications, and the presence of functioning vascular access have been collected prospectively. rospective data on serum calcium, serum phosphorus, alkaline phosphatase, and parathyroid hormone levels and skeletal x-rays have also been collected. By the use of set criteria and blinding to echocardiographic outcome, the presence and severity of hyperparathyroidism were graded by consensus. Fifty-one patients met eligibility criteria for inclusion; of these, 14 developed LVH (cases) and 37 did not (controls). had significantly higher systolic blood piessure ( = 0.009) and were older ( = 0.01) than controls. Systolic blood pressure correlated signifi- cantly with final posterior left ventricular wall thickness (r = 0.39; < 0.0 1). By the use of multivariate analysis, age and systolic blood pressure were significantly and independently associated with increased left ventricular mass index. The frequency of hyperparathyroidism was low and equal in both groups. There was a trend toward more severe anemia in cases that did not reach statistical significance. IVH developing after the initiation of dialysis for ESRD is associated with increased systolic blood pressure and older age. Key Words: Dialysis, left ventricular hypertrophy, risk factors L eft ventricular hypertrophy (LVH) is very common in patients with ESRD treated by dialysis, occurring in 68% of our nondiabetic dialysis patients (1). In patients without ESRD. electrocardiographicalby detected LVH is associated with a 5-yr mortality of 30% (2). Echocardiographically detected LVH is also associated with increased morbidity and mortality (3). Dialysis patients with LVH have an increased incidence of coronary events, stroke, and death when monitored prospectively (4,5). Among patients without ESRD, risk factors for LVH include increased body mass index, hypertension, gender, age, and diabetes mellitus (6). atients with ESRD have many additional potential risk factors for LVH including anemia (7). functioning vascubar access (8), hyperparathyroldism (9). aluminum accumulation ( 1 0). and perhaps an excess of NA-K- ATase inhibitors (1 1). In uremic subjects, the correlation between blood pressure and cardiac mass has not been impressive (1,12). In this study. we have prospectively monitored a cohort of dialysis patients initially free of LVH to answer the following question. Do nondiabetic dialysis patients who develop LVH differ from dialysis patients without LVH with respect to severity of hypertenslon, type of antihypertensive medication used, age, degree of anemia, and the presence and severity of hyperparathyroidism? I Received December 18. 1991. Accepted April 1, 1992. 2 Correspondence to Dr. J.D. Harnett, DIvialon of Nephrology. The Health sd- CncC$ Centre, rince hllllp Drive. St. John s, Newfoundland, AIB 3V6, Canada. 1046-6673/0407-1486$03.00/0 Journal of the American Society of Nephroloqy Copyright C 1994 by the American Society of Nephrology METHODS All dialysis patients at the Health Sciences Centre and the Grace General Hospital. St. John s, and the 1486 Volume 4. Number 7 #{149}1994
Harnett et al Royal Victoria Hospital, Montreal, have had annual echocardiography since 1 985 and prospective measurements of several relevant risk factors. atients were eligible for inclusion in this study if they: (1) had been on dialysis for at least 2 yr or had at least two routine echocardiograms; (2) were free of diabetes mellitus; or (3) had no echocardiographic evidence of LVH (posterior left ventricular wall thickness (LVWT] 1.2 cm) at the start of dialysis. atients were monitored prospectively, and those patients whose LVWT increased to 1.2 cm or more and whose LVWT increased by at least 0.2 cm were designated as cases. The remaining eligible patients formed the control group. Exposure data were collected prospectively on each patient blind to the final echocardiographic diagnosis. Blood pressure was measured routinely and documented in both hemodialysis and peritoneal dialysis continuous ambulatory peritoneal dialysis patients. For hemodialysis patients, mean predialysis blood pressure values were used, and monthly blood pressure values were documented in continuous ambulatory peritoneal dialysis patients. arterial pressures were calculated for all patients. Readings were recorded from the time of the initiation of dialysis therapy until the most recent echocardiogram. The number and type of antihypertensive medications used were recorded, and antihypertensive drugs were categorized as /3-blockers, vasodilators, angiotensin-converting enzyme inhibitors, centrally acting agents, calcium channel blockers, and others. Serial data on serum calcium, serum phosphorus, alkaline phosphatase, and parathyroid hormone (TH) levels and annual skeletal x-rays were recorded. These data were then reviewed by two nephrobogists (J.D. Harnett and.s. arfrey) blind to the echocardiographic diagnosis. Hyperparathyroidism was graded. by the use of set criteria, as absent, mild, or severe. Severe hyperparathyroidism was said to be present if the TH level was elevated and if skeletal x-rays showed definite changes of hyperparathyroidism. Mild disease was diagnosed if TH 1evels were elevated in association with raised alkaline phosphatase bevels and minimal or absent x-ray changes. Hyperparathyroidism was said to be absent if skeletal x-rays were normal and serum calcium and alkaline phosphatase levels were within the normal range; TH bevels could be normal or elevated. In reaching a diagnosis, therapy with 1.25 vitamin D and calcium supplements was taken into account in the interpretation of the laboratory values. All patients had age. sex, height, weight, body surface area, duration of dialysis therapy. and the type of underlying renal disease documented. In addition, monthly hemoglobin measurements were recorded. M-mode echocardiography was performed by cxperienced echocardiographers. in routine clinical conditions, using the criteria set forth by the American Society of Echocardiographers ( 1 3). Echocardiograms were usually performed within 24 h after a dialysis session in hemodialysis patients. Echocardiograms in both hemodialysis and peritoneal dialysis patients were performed at a time when they were felt to be close to dry weight. All data were entered on computer mainframe with Datatrieve software and were analyzed on the SSSX and BMD statistical software packages. Categorical variables were analyzed by the use of x2 tests or Fisher s exact test. Continuous variables were analyzed by use of the t test for unpaired variables. Multivariate analysis was performed by the use of multiple linear regression. A two-tailed value of 0.05 or less was considered significant. RESULTS Three hundred thirty-nine patients were reviewed for entry into the study. Forty-four were excluded because of diabetes, 1 65 were excluded because of abnormal echocardiography. 29 were excluded because transplantation occurred before a second echocardiogram, 1 1 died before a second echocardiogram, and 39 had incomplete echocardiographic measurements because of technical problems. Fiftyone nondiabetic patients with initially normal left ventricular (LV) wall thickness were entered into the study. The mean duration of follow-up was 29 mo (, 20). Fourteen subjects developed LVH (cases), and 37 patients did not (controls). Table 1 lists the demographic characteristics and echocardiographic measurements of both groups at baseline. Those patients who developed LVH were significantly older than TABLE I. Baseline demographic and echocardiographic data in case and control patients Cas es - Cont Age(yr) 58 17 45 13 0.01 LVWTinDiastole 10.4 0.5 10 1.1 0.08 (mm) IV End-Diastolic Diam- 49.8 5.5 53.1 8.9 0.12 eter (mm) IV Mass Index (g/m2) 140.5 40.6 149.4 37 0.49 IeftAtrium(mm) 35.2 6.6 37.3 7.6 0.34 DurationofFollow-Up 41.6 25.9 24.5 15.1 0.03 MaIe,N(%) 7 50 19 51 0.9 Hemodialysis, N(%) I I 77 23 62 0.4 rols Journal of the American Society of Nephrology 1487
Risk Factors for the Development of LVH those who did not, and the duration of follow-up was longer. Initial echocardiographic parameters were very similar in both groups. LVWT increased from a baseline value of 10.4 mm to a final value of 13.8 mm in cases, and their LV mass index increased from 140.5 to 215.9 g/m2. LVWT and LV mass did not change in control subjects (Tables 1 and 2). Because of differences in the duration of followup, we booked at echocardiographic parameters in both groups after approximately 1 8 mo of follow-up. These data are presented in Table 3 and demonstrate that LVWT was already significantly higher in cases than in controls at that time. had significantly higher systolic blood pressure than did controls (1 49.6 [, 1 3.4] versus 137.1 (. 15.7] mm Hg; = 0.0091. There were no statistically significant differences between cases and controls In diastolic blood pressure, mean arterial blood pressure, mean mnterdialytic weight gain, number of antihypertensive medications taken, and mean hemogbobin bevel (Table 4). Systolic blood pressure and final LVWT were significantly correlated (r = 0.39; < 0.01) (Figure 1). When hemodialysis and peritoneal dialysis patients were examined separately. similar findings were seen in both groups (Table 5). In peritoneal dialysis patients. the same trends, apart from increased age in cases. were present, but statistical significance was lacking because of the small sample size. There were no differences in the type of antihypertensive medications taken, although there was a trend toward a higher prevalence of centrally acting agents in the cases (Table 6). TABLE 4. Blood pressure and hemoglobin levels in cases and controls 0 B. blood pressure. b Hemodlalysis patients only. 18 I- r=0.39.!- E #{176} - 14 C a).. >81.. #{149}. 1-.C.91-10. 8 a. 6 80 (N = 14) (N 37) SystolicB(mmHg) 149.6 13.4 137.1 12.6 0.009 Arterial ressure I 04.3 10 99.8 1 1 0.17 (mm Hg) Diastolic B (mm Hg) 79.9 1 1.2 80.7 14.6 0.22 Hemoglobin(g/I) 84.9 14.1 93.4 15.5 0.07 Interdialytic Weight I.8 1.0 2. 1 0.85 0.51 Gain (kg)b 100 120 140 160 Systolic Blood ressure (mmhg) Figure 1. Systolic blood pressure and final LVWT were significantly correlated (< 0.01). 180 p TABLE 2. Indices of LVH at last follow-up IVWT(mm) 13.8 1.6 10.3 0.9 <0.001 IVMasslndex(g/ 215.9 62.4 153.9 65 0.003 m2) Increase in LVWT 3.4 1.6 0.2 0.9 <0.001 (mm) TABLE 3. LVWT in cases and controls at uniform follow-up time Cas es Cont IVWT(mm) 12.9 2.4 10.2 0.9 0.001 DurationofFollow-Up 19.4 6.8 16.3 7.2 0.17 rols Severe or mild hyperparathyroidism occurred in a minority of patients in both groups. and the proportion did not differ significantly between cases and controls (Table 6). Only one patient in each group was categorized as having severe hyperparathyroidism. The proportion of patients with functioning vascular access was similar in both groups. However, flow rate measurements were not available. Three of 1 4 cases were receiving erythropoietmn during the course of the study compared with 5 of 37 controls ( = 0.38; Fisher s exact test). Multiple linear regression used final LV mass as the dependent variable and systolic blood pressure, duration of follow-up, hyperparathyroidism, type of dialysis treatment, age. and hemoglobin levels as independent variables. Age ( = 0.03) and mean systolic blood pressure ( = 0.05) were the only variables independently and significantly associated with increased LV mass. The regression equation is presented in Table 7. 1488 Volume 4 Number 7. 1994
Harnett et al TABLE 5. Risk factors for IVH in different dialysis modalities Hemodialysis eritonea I Dialysis (N= 11) (N= 23) (N= 3) (N= 14) Age(yr) SystolicBloodressure (mm Hg) Hemoglobin(g/I) DurationofFollow-Up 61 15.8 148 13.5 84 15.4 48.5 24.9 43 11.9 138 19.1 90 17.2 26.5 18.2 0.0056 0.071 0.37 0.020 51 22.5 155 14.3 88 9.41 16.33 5.51 48 136 100.81 21.1 13.8 8.14 9.66 7.34 0.86 0.17 0.18 0.29 TABLE 6. arathyroid status and type of antihypertensive medications in case and control subjects N % N % Hyperparathyroidism Severeormild 3 23 11 31 0.45 Absent 10 77 25 69 Antihypertensive Treatment Ca channel blocker 5 36 13 35 1.0 9-blocker 5 36 16 43 0.86 Angiotensin-converting 3 2 1 8 2 1 0.65 enzyme inhibitor Vasodilator 3 21 6 16 0.50 Centrally active 3 21 3 8 0.19 a Fisher s exact test when N< 5; x2 when N> 5. TABLE 7. Multiple linear regression equation IVMass=2+7.6Tx+2.2Har+ 1.l2age - 0.05 FU + 1.04 sbp - 0.5 mhgb. where Tx = type of dialysis Har = hyperparathyroidism FU = duration of follow-up sbp = Systolic blood pressure mhgb = mean hemoglobin level DISCUSSION This study clearly establishes a relationship between systolic arterial hypertension and the devebopment of LVH in dialysis patients. In our previous study ( 1 ). there were a number of reasons why such an association may not have been found. The majority of patients already had LVH at the time of starting dialysis and predialysis blood pressures were not assessed. Cross-sectional blood pressure data might also be less accurate in predicting LVH than prospectively accumulated data. Although the relationship between systolic hypertension and LVH is clear in nonuremic individuals (6), this study is, to our knowledge, the first to clearly establish such a relationship among patients with ESRD. who developed LVH were significantly older than controls at baseline but nevertheless had virtually identical initial echocardiographic measurements. Our analysis shows that both age and systolic blood pressure were independent predictors of the development of LVH. Other authors have demonstrated an increased frequency of LVH with increasing age in nonuremic subjects (14-16). Our previous study in dialysis patients showed an association between age and severe LVH (1). It may be that the aging ventricle is more sensitive to the hypertrophic stimulus of an elevated systolic blood pressure. There was a tendency toward more severe anemia in the group who developed LVH that did not reach statistical significance. Given our sample size, however, we are unwilling to exclude an association between anemia and the development of LVH in dialysis patients on the basis of these data. were monitored for a significantly longer period than control subjects. However, LVH had already developed by 1 8 mo in the majority of cases. The duration of follow-up was entered as an independent variable in the multivariate analysis and did not emerge as being independently predictive of the development of LVH. Therefore, although the duration of follow-up is different in both groups. we do not believe that this affects our conclusions. Hyperparathyroidism was not common and does not appear in this study to be an important predictor of the development of LVH after the initiation of dialysis. Its robe in patients who already have LVH at the start of dialysis was not assessed by this study and remains controversial. It should be noted that our data only apply to those dialysis patients who developed LVH after dialysis Journal of the American Society of Nephrology 1489
Risk Factors for the Development of LVH has been initiated. This represents only a minority of prevalent cases of LVH in dialysis patients. Thus, these data should be generalized only to LVH developing in patients already on dialysis. In conclusion, our study shows that age and elevated systolic blood pressure are associated with the development of LVH after the initiation of dialysis. Hyperparathyroidism does not have major etiobogic Importance in the development of LVH in this group of patients. Although we cannot exclude an effect of anemia, hemoglobin differences were not major. We suggest that careful control of blood pressure levels in dialysis patients may have an important role in preventing the development of LVH. ACKNOWLEDGMENTS Supported by a grant from the Kidney Foundation of Canada. Dr. J.D. Harnett held a scholarship from the Kidney Foundation of Canada. REFERENCES 1. Harnett JD, arfrey S, Griffiths SM, Gault MB, Barre E, Guttman D: Left ventricular hypertrophy in endstage renal disease. Nephron 1988; 48:107-115. 2. Kannel WB, Gordon T, Offutt D: Left ventricular hypertrophy by electrocardiogram: revalence, incidence and mortality in the Framingham Study. Ann Intern Med 1969;71:89-105. 3. Levy D, Garrison RJ, Savage DD, Castelli WD: rognostic implications of echocardiographically determined left ventricular mass in the Framinham Heart Study. N Engl J Med 1990;322: 1561-1566. 4. arfrey S, Harnett JD, Griffiths SM, et at.: The clinical course of left ventricular hypertrophy in dialysis patients. Nephron 1990;55: 114-120. 5. Silberberg JS, Barre, richard S, Sniderman AD: Left ventricular hypertrophy: An independent determinant of survival in endstage renal failure. Kidney Int 1989;36:286-290. 6. Hammond IN, Devereux RB, Alderman MB, et at. : The prevalence and correlates of echocardiographic left ventricular hypertrophy among employed patients with uncomplicated hypertension. J Am Coll Cardiol 1986;7:639-650. 7. Neff MS, Kim KE, ersoff M, et at.: Hemodynamics of uremic anemia. Circulation 1 97 1 ;43: 876-883. 8. Anderson CB, Lodd JR, Graff RA, Grace MA, Harter MR. Newton WT: Cardiac failure and upper extremity arterlovenous dialysis fistula. Arch Intern Med 1 976;3:292-297. 9. London GM, de Vernejoul MC, Fabiani F, et at.: Secondary hyperparathyroidism and cardiac hypertrophy in hemodialysis patients. Kidney Int 1 987;32:900-927. 10. London GM, de Vernejoul MC, Fabiani F, et at.: Association between aluminum accumulation and cardiac hypertrophy in hemodialyzed patients. Am J Kidney Dis 1989;8:75-83. 1 1. Schreiber V, Korbel F: Does an endogenous digitalis-like immunoreactive factor participate in the development of cardiomegaly. Cor Vasa 1 982;24:228-232. 12. London GM, Fabiarn F, Mardrais SJ, et at.: Uremic cardiomyopathy: An inadequate left yentricular hypertrophy. Kidney mt i 987;3 1: 973-980. 1 3. Sahn DJ, De Maria A, Kisslo J: Recommendations regarding quantitation in M-mode echocardiography: Results of a survey of echocardiographic measurements. Circulation 1 978; 58: 1072-1083. 1 4. Gardin JM, Henry WL, Savage DD, Epstein SE: Echocardiographic evaluation of an older population without clinically apparent heart disease [Abstractj. Am J Cardiol 1977;39:277. 1 5. Gerstenblith G, Frederiksen J, Yin FC, Fortin NJ, Lakatta EG, Weisfeldt ML: Echocardiographic assessment of a normal aging population. Circulation 1 977;56:273-278. 1 6. Marcomichelakis J, Withers R, Newman GB, O Brien K, Emmanuel R: The relation of age to the thickness of the interventricular septum, the posterior left ventricular wall and their ratio. Int J Cardiol 1983;4:405-415. 1490 Volume 4 - Number 7 1994