Relation of Left Ventricular Hypertrophy, Afterload, and Contractility to Left Ventricular Performance in Goldblatt Hypertension Massimo
To analyze the determinants of left ventricular (LV) performance (myocardial afterload, chamber size, mass, and contractility) in Goldblatt hypertension, 19 anesthetized one-kidney, one-clip (1K1C) and 28 two-kidney, one-clip (2K1C) male Wistar rats were studied 58 to 62 days after clipping, together with 19 sham-operated and 13 normal rats (controls), by M-mode echocardiography using necropsy-vali dated methods of measurement. The LV fractional shortening was inversely related to end-systolic stress in all groups (r = - 0 . 8 9 to - 0 . 9 5 , all Ρ < .00001): 7 2K1C (25%) and 9 1K1C (47%) had frac tional shortening above the upper confidence limit in control animals. Both 1K1C and 2K1C with high LV performance had severe hypertension, inade quate LV hypertrophy, with resultant high wall stress (both Ρ < .005), increased LV chamber dimen sion (P < .005 and Ρ < .05, respectively) and high afterload-corrected fractional shortening (both Ρ < .001); 2K1C also had high plasma renin activity and
M
aintenance of high blood pressure requires not only alterations in the arterial tree, but also increased cardiac force generation. En hancement of the forcefulness of cardiac
Received June 7, 1990. Accepted January 7, 1992. From The Department of Medicine and the Hypertension Center, The New York Hospital — Cornell Medical Center, New York, New York. This study was supported in part by grant HL18323 from the Na tional Heart, Lung, and Blood Institute, Bethesda, Maryland. Address correspondence and reprint requests to Dr. Richard B. De vereux, Division of Cardiology, Box 222, The New York Hospital — Cornell Medical Center, 525 East 68th Street, New York, NY 10021.
Volpe, Maria J.F. Camargo,
Donald
C.
atrial natriuretic factor levels (both Ρ < .01). Rats with normal LV performance exhibited mild hy pertension, adequate LV hypertrophy (normalizing wall stress), and normal LV chamber size and afterload-corrected fractional shortening. Thus, 8J weeks after clipping, adequate LV hypertrophy allows maintenance of normal LV function by nor malizing myocardial afterload in a majority of rats with Goldblatt hypertension, whereas increased LV contractility (and possibly use of preload reserve in 1K1C) maintains normal LV function in the pres ence of inadequate LV hypertrophy and elevated wall stress, in a substantial minority of rats that de veloped more severe Goldblatt hypertension. Am J Hypertens 1992;5:292-301
KEY WORDS: Left ventricular hypertrophy, afterload, contractility, plasma renin activity, atrial na triuretic factor.
pumping may, on theoretical grounds, be produced by 1) the use of preload reserve (Starling forces), 2) the provision of increased force-generating units (myocar dial hypertrophy), or 3) an increase in myocardial con tractility. In human hypertension a spectrum of ana tomic adaptations may occur, including the "classic" pattern of concentric left ventricular (LV) hypertrophy, eccentric hypertrophy (which may be associated with a subtle increase in cardiac output), or the combination of increased LV wall thicknesses but reduced LV chamber size and normal overall mass, a pattern we have termed concentric LV remodeling. Other human studies have suggested that enhanced myocardial contractility 1-3
Downloaded from http://ajh.oxfordjournals.org/ at New York University on March 31, 2016
Giovanni de Simone, Richard B. Devereux, Wallerson, and John H. Laragh
AJH-MAY
1992-VOL
5, NO. 5, PART 1
4-6
6
7
8
11
12-15
METHODS Animal Models Seventy-nine male Wistar rats (120 g initial body weight), were studied, under a protocol ap proved by the Institutional Animal Care and Use Com mittee of Cornell University Medical College. After they had been housed in a cage for 7 days, surgical proce dures were performed on the animals under anesthesia with ketamine (60 mg/kg) and acepromazine (0.3 m g / kg). In 47 animals, hypertension was induced by placing a silver clip (0.22 mm internal diameter) around the left renal artery, which had been exposed through a lumbar incision. The right kidney was removed in 19 rats (1K1C), whereas it was not touched in the other 28 (2K1C). The same surgical procedures except for the renal artery clipping were used to prepare 10 two-kid ney and 9 one-kidney shams. The other 13 rats were untouched and used as a normal control group. After surgery, the rats were maintained on normal sodium Purina (St. Louis, MO) rat chow and were al lowed to drink tap water ad libitum. Systolic blood pres sure (SBP) was measured weekly in awake animals by tail cuff sphygmomanometry (PE-300, Narco BioSystem, Houston, TX). Experimental Procedures All the rats underwent echocardiographic study to assess LV anatomy and func tion 58 to 62 days after surgery. Echocardiograms were performed the day after the last blood pressure determi nation, under light anesthesia with intramuscular keta mine (50 mg/kg body weight). This type of anesthesia made animals sleep for the 10 to 20 min required to perform echocardiograms. Rats were decapitated and
293
blood samples were collected for plasma renin activity (PRA), atrial natriuretic factor (ANF), and microhematocrit determinations 48 h after the echocardiogram. Plasma renin activity and plasma ANF immunoreactivity were determined by radioimmunoassay as previously reported in detail. ' Hematocrit was deter mined using standard analytical technique. 16 17
Echocardiographic Method A Hewlett-Packard (An do ver, Massachusetts) 770720 echocardiographic sys tem was used, equipped with a 5 mHz, shallow focus, 21211A phased-array transducer, placed on the left hemithorax with the rat in the partial left decubitus po sition. Phantom experiments and necropsy validation of in vivo LV measurements using this echocardiograph and transducer revealed a high level of accuracy in our previous report. Two-dimensionally targeted M-mode echocardio grams were obtained from short axis views of the left ventricle at or just below the tip of the mitral valve leaflets, and were recorded on strip-chart paper at 100 mm/sec. Only M-mode echocardiograms with well de fined continuous interfaces for septum and posterior wall were accepted for measurements (Figure l ) , a cri terion that resulted in the exclusion of 6 rats (8%). Trac ings were numerically coded and interpreted at the end of the study by two observers blind to knowledge of type of rat and weight. 18
1 9
Ί Septum LV chamber J Posterior wall
ί
Septum LV chamber
Posterior wall
FIGURE 1. Examples of M-mode echocardiographic tracings used for the study. The arrows identify the interfaces of interven tricular septum and posterior wall used to measure left ventricular internal dimension at end-diastole. End-systolic measurements were made at the maximum anterior motion of posterior wall.
Downloaded from http://ajh.oxfordjournals.org/ at New York University on March 31, 2016
may also help sustain hypertension in some patients with nearly normal LV m a s s . The duration and sever ity of blood pressure elevation have been proposed as an influences on the prevalence of these different pat terns. The type of stimulus might also influence the pattern and extent of LV hypertrophy: recently, we have shown that in rats with Goldblatt hypertension, 8 weeks after renal artery clipping and at comparable levels of blood pressure, LV hypertrophy is more prevalent in one-kidney, one-clip (1K1C) than in two-kidney, oneclip (2K1C) rats, due to the superimposition of a volume overload in the former model to the pressure overload of both models, consistent with the different mechanisms of blood pressure elevation. " Pump function and myocardial performance of the left ventricle have been extensively studied in Goldblatt hypertension in r a t s , but the relative contribution of hypertrophy, myocardial performance and LV chamber size to maintenance of in vivo LV performance in Gold blatt hypertensive rats has not been established. Accord ingly, we studied hemodynamic components of in vivo LV performance at rest in 2K1C and 1K1C Goldblatt hypertension by using M-mode echocardiography.
LV P E R F O R M A N C E IN GOLDBLATT HYPERTENSION
18
2
2
18
18
Normalization of Ventricular Dimension for Body Size In order to minimize the effect of body size, vari ables were normalized for body weight. In a separate previously studied group of 41 normal Wistar rats with a wide range of body weight (118 to 760 g) CSA and EDD were closely related to body weight (r = 0.90 and 0.64, respectively). Conventional normalization of the two measurements for the first power of body weight yielded strong residual relations to body weight (r = - 0 . 7 7 for CSA/weight and - 0 . 8 6 for EDD/weight), due to the different geometric dimensions of variables. A nonlinear regression analysis was therefore used to identify the correct power of the relations between LV dimensions and body weight, by means of allometric equations (y = a X x ), which describe the relations reg ulating body proportions in mammals. The power (b coefficient) of body weight that was most closely related to CSA was 0.53 while that for EDD was 0.21. The formulae (CSA/body w e i g h t ) and (EDD/body w e i g h t ) minimized the variability of CSA and EDD among normal animals [coefficient of variability was 18% and 13% for EDD and (EDD/body w e i g h t ) , re spectively; and 2 9 % and 14% for CSA and (CSA/body weight) ] and eliminated the residual relations of in dexed LV dimensions to body weight [r = 0.02 and 0.001 for (CSA/body w e i g h t ) and (EDD/body weight) - , respectively]. Accordingly, in the present study CSA and EDD were normalized for body weight to the powers of 0.53 and 0.21, respectively. 18
20
b
21
053
021
021
053
053
0
21
Left Ventricular Pump Performance and Afterload Fractional shortening was used as an index of LV func tion and was related to end-systolic stress, to assess the load dependence of LV systolic function. Meridional end-systolic wall stress,— an index of myocardial
afterload—was calculated at the endocardium by an invasively validated formula, using tail-cuff systolic blood pressure, measured in awake animals the day be fore the echocardiogram. To compute end-systolic stress, systolic LV internal dimension and posterior wall thickness were measured at the time of maximum poste rior wall thickness during systole. To quantify LV performance independently of the effects of myocardial afterload, fractional shortening was corrected to what it would have been at the mean end-systolic stress of the pooled groups of hypertensive, sham, and normal rats, by the equation cFS = FS — b X (logESS — logESSx), where cFS is afterloadcorrected fractional shortening, FS and logESS are ob served fractional shortening and the logarithm of end-systolic stress, respectively, logESSx is the mean logarithm of ESS and b the slope of the regression line relating FS to logESS. By making LV performance inde pendent of afterload, the effect of myocardial contractil ity and/or preload could be studied. 22
Statistical Analysis Data are expressed as mean ± one standard deviation in groups of rats. Since distribu tions of most variables in subgroups of rats were not normal, data concerning these subgroups are presented as the median (M) and the 9 5 % confidence limits of the median, defined by using the two-tailed value of " C " for a given number of observations, according to a bino mial distribution. One-way Kruskal-Wallis' rank-anal ysis of variance was used to detect differences between Goldblatt hypertensive rats falling within or above the upper limit of the confidence interval of the normal relation between FS and ESS and the control group, comprised of normal and sham-operated rats (as no dif ference was found among these groups in terms of body weight, blood pressure, hormonal pattern, LV anatomy and function, or relation of FS to ESS). Dunn's post-hoc test was used to detect differences among the sub groups. 23
The null hypothesis was rejected at Ρ < .05. RESULTS Goldblatt rats exhibited high blood pressure (163 ± 43 mm Hg in 2K1C, 173 ± 46 mm Hg in 1K1C, Ρ < .003 and .0003, respectively, ν 133 ± 13 mm Hg in controls). Plasma renin activity was high in 2K1C (7.6 ± 4.9 ng/ m L / h ) as compared to controls (4.8 ± 1.9, Ρ < .01), as was atrial natriuretic factor (56 ± 40 ν 34 ± 10 fmol/ mL, Ρ < .04), whereas the differences were not signifi cant in 1K1C (4.9 ± 2.2 n g / m L / h and 48 ± 36 fmol/ mL, respectively). No significant difference was noted in hematocrit (40% in 2K1C and controls and 3 9 % in 1K1C).
Downloaded from http://ajh.oxfordjournals.org/ at New York University on March 31, 2016
End-diastolic dimension (EDD) and interventricular septal and posterior wall thicknesses (IVS and PWT) were measured by the leading edge method at the time of maximum diastolic dimension, independently of the echocardiogram, because the high heart rate caused the R wave onset to occur during LV filing. End-systolic dimensions (ESD) were measured at the time of the maximum anterior motion of the posterior wall. Left ventricular meridional cross-sectional area (CSA) was calculated as an index of LV weight: CSA = 3.14 X {[(EDD + PWT + I V S ) / 2 ] - (EDD/2) }. The CSA determined by M-mode echocardiography corre lated closely with necropsy LV weight in 58 rats (r = 0.90; Ρ < .0001) and showed excellent sensitivity (89%) and specificity (100%) in detecting necropsy LV hy pertrophy. Diastolic relative wall thickness was cal culated as [(PWT + IVS)/2]/EDD and was used as an index of LV geometry. In addition, LV mass was also estimated by a previously reported, necropsy-validated equation.
AJH-MAY
1992-VOL
5, NO. 5, PART 1
LV P E R F O R M A N C E IN GOLDBLATT HYPERTENSION
Relation of LV Systolic Function to Myocardial Af terload Fractional shortening was closely related to ESS, most closely on a semilogarithmic scale, in all groups of rats. This relation was virtually identical in normal and in sham animals (FS = 95.1 — 12.4 X logESS ν FS = 94.6 - 12.4 X logESS, both r = - 0 . 9 4 ) , and the two groups of animals were pooled to generate normal confidence limits (Figure 2). As evident in Fig ures 3 and 4, 7 of 28 2K1C (25%) and 9 of 19 1K1C (47%) had FS above the upper limit of the normal 9 5 % confidence interval. These two subsets of rats were con sidered to have high LV performance.
0,21
053
Goldblatt rats with high LV performance (both Ρ < .05). Moreover, in contrast with expectations based on the values of wall stress, both subgroups of rats exhibited FS that was slightly but not significantly higher than con trols (Table 2). By design, in both the Goldblatt models the correction of FS for ESS resulted in values signifi cantly higher than in controls (both Ρ < .001). Increased performance was therefore associated with increased LV chamber size and afterload-corrected FS. Adjustment for ESS reduced the coefficient of varia bility of FS from 14, 15, and 16% in control animals, 2K1C, and 1K1C, respectively, to about 6% for all groups. Characteristics of Goldblatt Rats With Normal LV Performance Table 3 shows that blood pressure was moderately high in rats with normal LV performance in either model of Goldblatt hypertension (both Ρ < .05). No significant differences were found in plasma renin activity, atrial natriuretic factor, hematocrit, body weight, or heart rate. The LV end-diastolic dimension normalized for body w e i g h t was not statistically increased in either group of Goldblatt rats with normal performance, whereas relative wall thickness was high in 2K1C (P < .05) and normal in 1K1C. The CSA and CSA index as well as LV mass were increased in both subgroups (.05 < Ρ < .005). Table 4 shows that end-systolic stress and FS were normal in rats with normal LV performance in both Goldblatt models. In both subgroups, therefore, LV hy pertrophy appeared to be adequate to normalize wall stress in the presence of moderately high blood pres sure. The correction of FS for ESS resulted in no differ ence from the measured values. 021
CONTROLS
DISCUSSION Studies in human hypertension have demonstrated dif ferent hemodynamic and anatomic LV p a t t e r n s " ; in particular supernormal LV performance has been re lated to reduced myocardial afterload in the setting of severe concentric LV hypertrophy, or been shown to occur despite normal or increased afterload in patients with little or no hypertrophy. " Whether enhanced LV performance is sustained by use of preload reserve (Starling forces) or by increased intrinsic myocardial contractility is still controversial, because of the limitations of the current indices used for noninvasive assessment of myocardial contractility. However, estimates of contractility obtained from LV stress-short ening relations have been shown to be related to inva sive measurements obtained during hemodynamic load manipulation ' and to identify myocardial dysfunc tion associated with an adverse prognosis. By using this type of approach, previous studies have suggested 4-6,24
24
4
101
0.5
ι
ι
1.0
10.0
100.0 3
2
End-Systolic Stress (*10 dynes/mm ) FIGURE 2. Relation of left ventricular fractional shortening (vertical axis) to end-systolic stress (horizontal axis, logarithmic scale) in normotensive rats (control group). Solid lines represent the 95% confidence interval. Closed circles are sham-operated rats; open circles are normal rats.
6
26,27
2,27 28
29
25
Downloaded from http://ajh.oxfordjournals.org/ at New York University on March 31, 2016
Characteristics of Hypertensive Rats with High LV Performance Table 1 shows that blood pressure was significantly elevated in both subgroups with high LV performance (P < .001); moreover, in 2K1C high per formance was also associated with high PRA and ANF (both Ρ < .01). No significant differences were found in body weight, hematocrit, or heart rate. Left ventricular end-diastolic dimension was signifi cantly increased only in 1K1C (P < .05 ν controls); in dexation for body weight increased the difference in 1K1C (P < .005) and made it significant also in 2K1C (P < .05 ν controls). The (CSA/body w e i g h t ) was statistically higher in Goldblatt rats (P < .05), but no specific significant intergroup difference was identified by post hoc testing. The LV mass was not significantly different among the groups. Relative wall thickness was not significantly increased in either group of hyperten sive rats. Table 2 shows that ESS was moderately elevated in
295
TABLE 1. BLOOD PRESSURE, HORMONAL PROFILE, AND LEFT VENTRICULAR ANATOMY IN GOLDBLATT RATS WITH HIGH LEFT VENTRICULAR PERFORMANCE
Systolic BP (mm Hg) PRA (ng/mL/h) ANF (fmol/mL) Hematocrit (%) Body weight (g)
LV dimension (mm) LV dimension/kg (mm/kg ) 021
CSA (mm ) 2
CSA/kg (mm /kg - ) 2
0
53
LV mass (g) Relative wall thickness
2-Kidney, 1-Clip (n = 7)
1-Kidney, 1-Clip (n = 9)
129t (125-138) 4.6 (3.9-5.4) 30.7 (25.8-39.0) 41.5 (39.0-43.5) 436 (424-446) 484 (470-504) 6.20 (5.88-6.69) 7.37 (6.92-7.94) 38.16 (37.3-41.49) 89.48 (82.0-96.38) 0.78 (0.74-0.84) 0.50 (0.47-0.54)
213t (150-300) 7.8* (1.9-22.0) 50.6* (26.4-209.5) 40.0 (34.0-43.0) 390 (330-432) 488 (353-507) 6.85 (6.03-8.31) 8.20* (7.35-10.49) 43.08 (30.7-64.07) 67.81 (50.64-105.35) 0.84 (0.59-1.35) 0.42 (0.39-0.76)
175f (147-250) 4.1 (1.7-9.1) 39.9 (20.0-132.0) 40.0 (35.0-42.0) 413 (340-435) 467 (349-500) 7.16* (6.20-7.70) 8.87t (7.38-10.52) 51.19 (32.32-65.00) 134.54 (78.26-169.69) 1.07 (0.60-1.32) 0.47 (0.41-0.55)
Statistical significance: * Ρ < .05; f Ρ < .005 (Dunn's test). BP = blood pressure; LV = left ventricular; PRA — plasma renin activity; ANF = atrial natriuretic factor; CSA = cross-sectional area. Data are presented as medians of variables, with 95% confidence limits of the medians in parentheses.
that increased myocardial contractility may be an alter native mechanism to development of LV hypertrophy for sustaining LV performance in the face of hemody namic overload or that contractility becomes depressed when cardiac hypertrophy is p r e s e n t . " In this study we used a mathematical approach to derive an index of LV performance independent of af terload. This variable, therefore, may reflect the effect of myocardial contractility and/or preload reserve on LV performance. 4-6,30
the presence of compensatory LV hypertrophy; this compensation might be more difficult to achieve in 1K1C because of the additional volume overload.
33
Normal Left Ventricular Performance Both 2K1C and 1K1C with normal LV performance exhibited a de gree and pattern of LV hypertrophy sufficient, on aver age, to normalize wall stress. This anatomic compensa tion was systematically present in 2K1C, which exhibited concentric LV hypertrophy, but less consistent in 1K1C (with 4 / 1 0 rats having values of relative wall thickness below and 3 / 1 0 above the 9 5 % confidence limits of the control group), because, in the latter group the interaction of pressure and volume o v e r l o a d yields a greater variability of LV geometry. Thus, in the rats exhibiting normal load-adjusted shortening, arterial hypertension did not affect LV performance, because of 8-11
Enhanced Left Ventricular Performance Both sub groups of hypertensive rats with increased LV perform ance exhibited very high blood pressure associated with LV chamber enlargement (especially in 1K1C), but without the expected increase in LV wall thicknesses. Relative wall thickness was not significantly increased in either model, resulting in high values of wall stress (4 of 7 2K1C and 4 of 9 1K1C had values of end-systolic stress above the 9 5 % confidence limits of control ani mals). Although median values of CSA in these sub groups of rats with moderate hypertension were similar to those in the mildly-hypertensive subgroups with normal LV performance at the same interval after renal artery clipping (58 to 62 days), it is possible that they would develop, over a larger period of time, adequate LV hypertrophy to offset the higher pressure, a hypoth esis already suggested in human studies. These rats, indeed, maintained normal or slightly increased ejec tion-phase function, because of alternative compensa6
Downloaded from http://ajh.oxfordjournals.org/ at New York University on March 31, 2016
Heart rate (beats/min)
Controls (n = 32)
ONE - KIDNEY, ONE CLIP
0.5
1.0
10.0
100.0
36
2
End-Systolic Stress (*10 dynes/mm ) FIGURE 3. Relation of left ventricular fractional shortening (vertical axis) to end-systolic stress (horizontal axis, logarithmic scale) in one-kidney, one-clip rats. Solid lines represent the 9 5 % confidence interval of control animals. Closed circles represent rats with enhanced LV performance.
tory mechanisms. These mechanisms may include the use of myocardial contractility and/or preload reserve. The possibility of the use of the preload reserve might be suggested by the increase in LV diastolic dimension in the presence of normal to increased fractional shortening. Since compensation through the Starling forces is predominantly an adaptation to acute overload, it is more likely that increased values of afterload-corrected 34
35
TWO - KIDNEY, ONE CLIP
Comparison With Previous Studies Characteristics of LV function in rats with renovascular hypertension have been assessed by a variety of a p p r o a c h e s , such as examining the curves relating invasively measured LV end-diastolic pressure and cardiac output under a range of loading conditions. The ability to estimate end-systolic wall stress (the myocardial afterload that stops LV contraction) in vivo has been made possible in this study by M-mode echocardiography. Our study, therefore, evaluates for the first time LV ejection-phase systolic function in Goldblatt hypertension in rats in relation to myocardial afterload. Averill et a l reported reduction of cardiac functional response to volume infusions in ether-anesthetized open-chest 2K1C 6 to 22 weeks after clipping. In addition to the difference in experimental procedures and anesthesia, our study examines LV function only under relatively basal conditions in rats 58 to 62 days after clipping. Thus, while we have documented that both Goldblatt models preserve LV pumping ability at rest by undergoing hypertrophy and using contractile and/or preload reserves, it is possible that our animals would have resembled those of Averill et a l in being unable to respond normally to further hemodynamic overload. Left ventricular dysfunction has also been associated with severe LV hypertrophy. Kuwajima et al, using the LV end-diastolic pressure/cardiac output curve, demonstrated reduced LV function in ether-anesthetized 2K1C related to the presence of LV hypertrophy: when LV hypertrophy was reduced with methyldopa, without a comparable effect on blood pressure, LV function improved. Our results suggest that in Goldblatt rats high levels of myocardial performance are associated with small increases in myocardial mass that are inadequate to normalize wall stress. In our 2K1C animals with high LV performance the elevated plasma renin levels suggest a possible contribution from positive inotropic effects of angiotensin, ' whereas no specific mechanism of contractility enhancement is suggested by our findings in 1K1C rats. 12/13,37-39
12
12
12,38
FIGURE 4. Relation of left ventricular fractional shortening (vertical axis) to end-systolic stress (horizontal axis, logarithmic scale) in two-kidney, one-clip rats. Solid lines represent the 95% confidence interval of control animals. Closed circles represent rats with enhanced LV performance.
40 41
38
Downloaded from http://ajh.oxfordjournals.org/ at New York University on March 31, 2016
3
fractional shortening in these rats with a reasonably stable level of elevated blood pressure principally reflect enhanced myocardial contractility. On the other hand, persistent participation of the Frank-Starling mechanism to sustain the high level of LV performance is hard to exclude in these animals with mild LV chamber enlargement (especially in 1K1C, a model characterized by initially volume-dependent hypertension) in the absence of direct measurements of end-diastolic stress (the best measure of LV preload) or sarcomere length (indicative of increased preload on the contractile apparatus). In this regard, it is relevant that elevated end-diastolic stress had been observed in aortic regurgitation, a model of chronic volume overload with LV dilatation.
TABLE 2. LEFT VENTRICULAR FUNCTION IN GOLDBLATT RATS WITH HIGH LEFT VENTRICULAR PERFORMANCE
ESS Χ 10 (dynes/cm ) 3
2
Fractional shortening (%) Afterload-corrected shortening (%)
Controls (n = 32)
2-Kidney, 1-Clip (n = 7)
1-Kidney, 1-Clip (n = 9)
10.0 (7.6-14.8) 63 (61-69) 63 (61-66)
17.6* (8.8-36.3) 66 (62-74) 70t (69-74)
14.8* (13.1-31.1) 68 (59-74) 71f (69-74)
Statistical significance: * Ρ < 0.05; f Ρ < 0.001. ESS = end-systolic stress. Data are presented as medians of variables, with 95% confidence limits of the medians in parentheses.
neously with the echocardiogram, because measure ments of blood pressure during the echocardiographic session would have required prolongation of anesthe sia, potentially altering hemodynamics and increasing the risk of death during the procedure (with our proce dure no animal died during anesthesia). Although a blood pressure measurement simultaneous to the echo-
TABLE 3. BLOOD PRESSURE, HORMONAL PROFILE, AND LEFT VENTRICULAR ANATOMY IN GOLDBLATT RATS WITH NORMAL LEFT VENTRICULAR PERFORMANCE
Systolic BP (mm Hg) PRA (ng/mL/h) ANF (fmol/mL) Hematocrit (%) Body weight (g) Heart rate (beats/min) LV Dimension (mm) LV Dimension/kg (mm/kg ) 021
LV Mass (g) CSA (mm ) 2
CSA/kg (mm/kg ) 053
Relative wall thickness
Controls (n = 32)
2-Kidney, 1-Clip (n = 21)
1-Kidney, 1-Clip (n = 10)
129 (125-138) 4.6 (3.9-5.4) 30.7 (25.8-39.0) 41.5 (39.0-43.5) 436 (424-446) 484 (470-504) 6.20 (5.88-6.69) 7.37 (6.92-7.94) 0.78 (0.74-0.84) 38.16 (37.30-41.49) 89.48 (82.00-96.38) 0.50 (0.47-0.54)
150* (136-160) 5.5 (4.0-9.1) 44.1 (30.5-60.2) 41.0 (39.0-43.3) 445 (410-470) 465 (430-500) 6.28 (5.92-6.58) 7.46 (7.02-7.77) 0.91* (0.78-1.03) 45.26* (37.92-48.46) 101.72* (89.28-125.62) 0.55* (0.52-0.60)
159* (115-213) 5.8 (2.0-5.7) 22.4 (17.2-91.2) 40.5 (32.3-42.0) 375 (340-450) 464 (337-506) 6.74 (5.70-7.27) 8.23 (6.66-11.61) 1.05| (0.81-1.20) 49.37f (41.24-65.97) 116.00t (93.79-195.95) 0.51 (0.42-0.79)
Statistical significance: * Ρ < .05; j Ρ < .005 ν controls. For abbreviations see Table 1. Data are presented as medians of variables, with 95% confidence limits of the medians in parentheses.
Downloaded from http://ajh.oxfordjournals.org/ at New York University on March 31, 2016
Technical Limitations Potential limitations of this study need to be considered. The use of M-mode echo cardiography has the advantage of allowing in vivo measurements of LV wall stress, a procedure virtually impossible with other techniques. Unfortunately, endsystolic stress can be only approximated by using our method. Blood pressure was not measured simulta
TABLE 4. LEFT VENTRICULAR FUNCTION IN GOLDBLATT RATS WITH NORMAL LEFT VENTRICULAR PERFORMANCE
ESS Χ 10 (dynes/cm ) 3
2
Fractional shortening (%) Afterload-corrected shortening (%)
Controls (n = 32)
2-Kidney, 1-Clip (n = 21)
1-Kidney, 1-Clip (n = 10)
10.0 (7.6-14.8) 63 (61-69) 63 (61-66)
13.4 (10.8-22.3) 65 (57-67) 64 (62-65)
10.7 (5.93-24.4) 68 (55-72) 65 (62-68)
ESS = end-systolic stress. Data are presented as medians of variables, with 95% confidence limits of the medians in parentheses.
80
= Rats = Humans
4,5 42 43
44
45 46
4,5
c 'c
To analyze the determinants of left ventricular (LV) performance (myocardial afterload, chamber size, mass, and contractility) in Goldblatt hypertension, 19 anesthetized one-kidney, one-clip (1K1C) and 28 two-kidney, one-clip (2K1C) male Wistar rats were studied 58 to 62 days after clipping, together with 19 sham-operated and 13 normal rats (controls), by M-mode echocardiography using necropsy-vali dated methods of measurement. The LV fractional shortening was inversely related to end-systolic stress in all groups (r = - 0 . 8 9 to - 0 . 9 5 , all Ρ < .00001): 7 2K1C (25%) and 9 1K1C (47%) had frac tional shortening above the upper confidence limit in control animals. Both 1K1C and 2K1C with high LV performance had severe hypertension, inade quate LV hypertrophy, with resultant high wall stress (both Ρ < .005), increased LV chamber dimen sion (P < .005 and Ρ < .05, respectively) and high afterload-corrected fractional shortening (both Ρ < .001); 2K1C also had high plasma renin activity and
M
aintenance of high blood pressure requires not only alterations in the arterial tree, but also increased cardiac force generation. En hancement of the forcefulness of cardiac
Received June 7, 1990. Accepted January 7, 1992. From The Department of Medicine and the Hypertension Center, The New York Hospital — Cornell Medical Center, New York, New York. This study was supported in part by grant HL18323 from the Na tional Heart, Lung, and Blood Institute, Bethesda, Maryland. Address correspondence and reprint requests to Dr. Richard B. De vereux, Division of Cardiology, Box 222, The New York Hospital — Cornell Medical Center, 525 East 68th Street, New York, NY 10021.
Volpe, Maria J.F. Camargo,
Donald
C.
atrial natriuretic factor levels (both Ρ < .01). Rats with normal LV performance exhibited mild hy pertension, adequate LV hypertrophy (normalizing wall stress), and normal LV chamber size and afterload-corrected fractional shortening. Thus, 8J weeks after clipping, adequate LV hypertrophy allows maintenance of normal LV function by nor malizing myocardial afterload in a majority of rats with Goldblatt hypertension, whereas increased LV contractility (and possibly use of preload reserve in 1K1C) maintains normal LV function in the pres ence of inadequate LV hypertrophy and elevated wall stress, in a substantial minority of rats that de veloped more severe Goldblatt hypertension. Am J Hypertens 1992;5:292-301
KEY WORDS: Left ventricular hypertrophy, afterload, contractility, plasma renin activity, atrial na triuretic factor.
pumping may, on theoretical grounds, be produced by 1) the use of preload reserve (Starling forces), 2) the provision of increased force-generating units (myocar dial hypertrophy), or 3) an increase in myocardial con tractility. In human hypertension a spectrum of ana tomic adaptations may occur, including the "classic" pattern of concentric left ventricular (LV) hypertrophy, eccentric hypertrophy (which may be associated with a subtle increase in cardiac output), or the combination of increased LV wall thicknesses but reduced LV chamber size and normal overall mass, a pattern we have termed concentric LV remodeling. Other human studies have suggested that enhanced myocardial contractility 1-3
Downloaded from http://ajh.oxfordjournals.org/ at New York University on March 31, 2016
Giovanni de Simone, Richard B. Devereux, Wallerson, and John H. Laragh
AJH-MAY
1992-VOL
5, NO. 5, PART 1
4-6
6
7
8
11
12-15
METHODS Animal Models Seventy-nine male Wistar rats (120 g initial body weight), were studied, under a protocol ap proved by the Institutional Animal Care and Use Com mittee of Cornell University Medical College. After they had been housed in a cage for 7 days, surgical proce dures were performed on the animals under anesthesia with ketamine (60 mg/kg) and acepromazine (0.3 m g / kg). In 47 animals, hypertension was induced by placing a silver clip (0.22 mm internal diameter) around the left renal artery, which had been exposed through a lumbar incision. The right kidney was removed in 19 rats (1K1C), whereas it was not touched in the other 28 (2K1C). The same surgical procedures except for the renal artery clipping were used to prepare 10 two-kid ney and 9 one-kidney shams. The other 13 rats were untouched and used as a normal control group. After surgery, the rats were maintained on normal sodium Purina (St. Louis, MO) rat chow and were al lowed to drink tap water ad libitum. Systolic blood pres sure (SBP) was measured weekly in awake animals by tail cuff sphygmomanometry (PE-300, Narco BioSystem, Houston, TX). Experimental Procedures All the rats underwent echocardiographic study to assess LV anatomy and func tion 58 to 62 days after surgery. Echocardiograms were performed the day after the last blood pressure determi nation, under light anesthesia with intramuscular keta mine (50 mg/kg body weight). This type of anesthesia made animals sleep for the 10 to 20 min required to perform echocardiograms. Rats were decapitated and
293
blood samples were collected for plasma renin activity (PRA), atrial natriuretic factor (ANF), and microhematocrit determinations 48 h after the echocardiogram. Plasma renin activity and plasma ANF immunoreactivity were determined by radioimmunoassay as previously reported in detail. ' Hematocrit was deter mined using standard analytical technique. 16 17
Echocardiographic Method A Hewlett-Packard (An do ver, Massachusetts) 770720 echocardiographic sys tem was used, equipped with a 5 mHz, shallow focus, 21211A phased-array transducer, placed on the left hemithorax with the rat in the partial left decubitus po sition. Phantom experiments and necropsy validation of in vivo LV measurements using this echocardiograph and transducer revealed a high level of accuracy in our previous report. Two-dimensionally targeted M-mode echocardio grams were obtained from short axis views of the left ventricle at or just below the tip of the mitral valve leaflets, and were recorded on strip-chart paper at 100 mm/sec. Only M-mode echocardiograms with well de fined continuous interfaces for septum and posterior wall were accepted for measurements (Figure l ) , a cri terion that resulted in the exclusion of 6 rats (8%). Trac ings were numerically coded and interpreted at the end of the study by two observers blind to knowledge of type of rat and weight. 18
1 9
Ί Septum LV chamber J Posterior wall
ί
Septum LV chamber
Posterior wall
FIGURE 1. Examples of M-mode echocardiographic tracings used for the study. The arrows identify the interfaces of interven tricular septum and posterior wall used to measure left ventricular internal dimension at end-diastole. End-systolic measurements were made at the maximum anterior motion of posterior wall.
Downloaded from http://ajh.oxfordjournals.org/ at New York University on March 31, 2016
may also help sustain hypertension in some patients with nearly normal LV m a s s . The duration and sever ity of blood pressure elevation have been proposed as an influences on the prevalence of these different pat terns. The type of stimulus might also influence the pattern and extent of LV hypertrophy: recently, we have shown that in rats with Goldblatt hypertension, 8 weeks after renal artery clipping and at comparable levels of blood pressure, LV hypertrophy is more prevalent in one-kidney, one-clip (1K1C) than in two-kidney, oneclip (2K1C) rats, due to the superimposition of a volume overload in the former model to the pressure overload of both models, consistent with the different mechanisms of blood pressure elevation. " Pump function and myocardial performance of the left ventricle have been extensively studied in Goldblatt hypertension in r a t s , but the relative contribution of hypertrophy, myocardial performance and LV chamber size to maintenance of in vivo LV performance in Gold blatt hypertensive rats has not been established. Accord ingly, we studied hemodynamic components of in vivo LV performance at rest in 2K1C and 1K1C Goldblatt hypertension by using M-mode echocardiography.
LV P E R F O R M A N C E IN GOLDBLATT HYPERTENSION
18
2
2
18
18
Normalization of Ventricular Dimension for Body Size In order to minimize the effect of body size, vari ables were normalized for body weight. In a separate previously studied group of 41 normal Wistar rats with a wide range of body weight (118 to 760 g) CSA and EDD were closely related to body weight (r = 0.90 and 0.64, respectively). Conventional normalization of the two measurements for the first power of body weight yielded strong residual relations to body weight (r = - 0 . 7 7 for CSA/weight and - 0 . 8 6 for EDD/weight), due to the different geometric dimensions of variables. A nonlinear regression analysis was therefore used to identify the correct power of the relations between LV dimensions and body weight, by means of allometric equations (y = a X x ), which describe the relations reg ulating body proportions in mammals. The power (b coefficient) of body weight that was most closely related to CSA was 0.53 while that for EDD was 0.21. The formulae (CSA/body w e i g h t ) and (EDD/body w e i g h t ) minimized the variability of CSA and EDD among normal animals [coefficient of variability was 18% and 13% for EDD and (EDD/body w e i g h t ) , re spectively; and 2 9 % and 14% for CSA and (CSA/body weight) ] and eliminated the residual relations of in dexed LV dimensions to body weight [r = 0.02 and 0.001 for (CSA/body w e i g h t ) and (EDD/body weight) - , respectively]. Accordingly, in the present study CSA and EDD were normalized for body weight to the powers of 0.53 and 0.21, respectively. 18
20
b
21
053
021
021
053
053
0
21
Left Ventricular Pump Performance and Afterload Fractional shortening was used as an index of LV func tion and was related to end-systolic stress, to assess the load dependence of LV systolic function. Meridional end-systolic wall stress,— an index of myocardial
afterload—was calculated at the endocardium by an invasively validated formula, using tail-cuff systolic blood pressure, measured in awake animals the day be fore the echocardiogram. To compute end-systolic stress, systolic LV internal dimension and posterior wall thickness were measured at the time of maximum poste rior wall thickness during systole. To quantify LV performance independently of the effects of myocardial afterload, fractional shortening was corrected to what it would have been at the mean end-systolic stress of the pooled groups of hypertensive, sham, and normal rats, by the equation cFS = FS — b X (logESS — logESSx), where cFS is afterloadcorrected fractional shortening, FS and logESS are ob served fractional shortening and the logarithm of end-systolic stress, respectively, logESSx is the mean logarithm of ESS and b the slope of the regression line relating FS to logESS. By making LV performance inde pendent of afterload, the effect of myocardial contractil ity and/or preload could be studied. 22
Statistical Analysis Data are expressed as mean ± one standard deviation in groups of rats. Since distribu tions of most variables in subgroups of rats were not normal, data concerning these subgroups are presented as the median (M) and the 9 5 % confidence limits of the median, defined by using the two-tailed value of " C " for a given number of observations, according to a bino mial distribution. One-way Kruskal-Wallis' rank-anal ysis of variance was used to detect differences between Goldblatt hypertensive rats falling within or above the upper limit of the confidence interval of the normal relation between FS and ESS and the control group, comprised of normal and sham-operated rats (as no dif ference was found among these groups in terms of body weight, blood pressure, hormonal pattern, LV anatomy and function, or relation of FS to ESS). Dunn's post-hoc test was used to detect differences among the sub groups. 23
The null hypothesis was rejected at Ρ < .05. RESULTS Goldblatt rats exhibited high blood pressure (163 ± 43 mm Hg in 2K1C, 173 ± 46 mm Hg in 1K1C, Ρ < .003 and .0003, respectively, ν 133 ± 13 mm Hg in controls). Plasma renin activity was high in 2K1C (7.6 ± 4.9 ng/ m L / h ) as compared to controls (4.8 ± 1.9, Ρ < .01), as was atrial natriuretic factor (56 ± 40 ν 34 ± 10 fmol/ mL, Ρ < .04), whereas the differences were not signifi cant in 1K1C (4.9 ± 2.2 n g / m L / h and 48 ± 36 fmol/ mL, respectively). No significant difference was noted in hematocrit (40% in 2K1C and controls and 3 9 % in 1K1C).
Downloaded from http://ajh.oxfordjournals.org/ at New York University on March 31, 2016
End-diastolic dimension (EDD) and interventricular septal and posterior wall thicknesses (IVS and PWT) were measured by the leading edge method at the time of maximum diastolic dimension, independently of the echocardiogram, because the high heart rate caused the R wave onset to occur during LV filing. End-systolic dimensions (ESD) were measured at the time of the maximum anterior motion of the posterior wall. Left ventricular meridional cross-sectional area (CSA) was calculated as an index of LV weight: CSA = 3.14 X {[(EDD + PWT + I V S ) / 2 ] - (EDD/2) }. The CSA determined by M-mode echocardiography corre lated closely with necropsy LV weight in 58 rats (r = 0.90; Ρ < .0001) and showed excellent sensitivity (89%) and specificity (100%) in detecting necropsy LV hy pertrophy. Diastolic relative wall thickness was cal culated as [(PWT + IVS)/2]/EDD and was used as an index of LV geometry. In addition, LV mass was also estimated by a previously reported, necropsy-validated equation.
AJH-MAY
1992-VOL
5, NO. 5, PART 1
LV P E R F O R M A N C E IN GOLDBLATT HYPERTENSION
Relation of LV Systolic Function to Myocardial Af terload Fractional shortening was closely related to ESS, most closely on a semilogarithmic scale, in all groups of rats. This relation was virtually identical in normal and in sham animals (FS = 95.1 — 12.4 X logESS ν FS = 94.6 - 12.4 X logESS, both r = - 0 . 9 4 ) , and the two groups of animals were pooled to generate normal confidence limits (Figure 2). As evident in Fig ures 3 and 4, 7 of 28 2K1C (25%) and 9 of 19 1K1C (47%) had FS above the upper limit of the normal 9 5 % confidence interval. These two subsets of rats were con sidered to have high LV performance.
0,21
053
Goldblatt rats with high LV performance (both Ρ < .05). Moreover, in contrast with expectations based on the values of wall stress, both subgroups of rats exhibited FS that was slightly but not significantly higher than con trols (Table 2). By design, in both the Goldblatt models the correction of FS for ESS resulted in values signifi cantly higher than in controls (both Ρ < .001). Increased performance was therefore associated with increased LV chamber size and afterload-corrected FS. Adjustment for ESS reduced the coefficient of varia bility of FS from 14, 15, and 16% in control animals, 2K1C, and 1K1C, respectively, to about 6% for all groups. Characteristics of Goldblatt Rats With Normal LV Performance Table 3 shows that blood pressure was moderately high in rats with normal LV performance in either model of Goldblatt hypertension (both Ρ < .05). No significant differences were found in plasma renin activity, atrial natriuretic factor, hematocrit, body weight, or heart rate. The LV end-diastolic dimension normalized for body w e i g h t was not statistically increased in either group of Goldblatt rats with normal performance, whereas relative wall thickness was high in 2K1C (P < .05) and normal in 1K1C. The CSA and CSA index as well as LV mass were increased in both subgroups (.05 < Ρ < .005). Table 4 shows that end-systolic stress and FS were normal in rats with normal LV performance in both Goldblatt models. In both subgroups, therefore, LV hy pertrophy appeared to be adequate to normalize wall stress in the presence of moderately high blood pres sure. The correction of FS for ESS resulted in no differ ence from the measured values. 021
CONTROLS
DISCUSSION Studies in human hypertension have demonstrated dif ferent hemodynamic and anatomic LV p a t t e r n s " ; in particular supernormal LV performance has been re lated to reduced myocardial afterload in the setting of severe concentric LV hypertrophy, or been shown to occur despite normal or increased afterload in patients with little or no hypertrophy. " Whether enhanced LV performance is sustained by use of preload reserve (Starling forces) or by increased intrinsic myocardial contractility is still controversial, because of the limitations of the current indices used for noninvasive assessment of myocardial contractility. However, estimates of contractility obtained from LV stress-short ening relations have been shown to be related to inva sive measurements obtained during hemodynamic load manipulation ' and to identify myocardial dysfunc tion associated with an adverse prognosis. By using this type of approach, previous studies have suggested 4-6,24
24
4
101
0.5
ι
ι
1.0
10.0
100.0 3
2
End-Systolic Stress (*10 dynes/mm ) FIGURE 2. Relation of left ventricular fractional shortening (vertical axis) to end-systolic stress (horizontal axis, logarithmic scale) in normotensive rats (control group). Solid lines represent the 95% confidence interval. Closed circles are sham-operated rats; open circles are normal rats.
6
26,27
2,27 28
29
25
Downloaded from http://ajh.oxfordjournals.org/ at New York University on March 31, 2016
Characteristics of Hypertensive Rats with High LV Performance Table 1 shows that blood pressure was significantly elevated in both subgroups with high LV performance (P < .001); moreover, in 2K1C high per formance was also associated with high PRA and ANF (both Ρ < .01). No significant differences were found in body weight, hematocrit, or heart rate. Left ventricular end-diastolic dimension was signifi cantly increased only in 1K1C (P < .05 ν controls); in dexation for body weight increased the difference in 1K1C (P < .005) and made it significant also in 2K1C (P < .05 ν controls). The (CSA/body w e i g h t ) was statistically higher in Goldblatt rats (P < .05), but no specific significant intergroup difference was identified by post hoc testing. The LV mass was not significantly different among the groups. Relative wall thickness was not significantly increased in either group of hyperten sive rats. Table 2 shows that ESS was moderately elevated in
295
TABLE 1. BLOOD PRESSURE, HORMONAL PROFILE, AND LEFT VENTRICULAR ANATOMY IN GOLDBLATT RATS WITH HIGH LEFT VENTRICULAR PERFORMANCE
Systolic BP (mm Hg) PRA (ng/mL/h) ANF (fmol/mL) Hematocrit (%) Body weight (g)
LV dimension (mm) LV dimension/kg (mm/kg ) 021
CSA (mm ) 2
CSA/kg (mm /kg - ) 2
0
53
LV mass (g) Relative wall thickness
2-Kidney, 1-Clip (n = 7)
1-Kidney, 1-Clip (n = 9)
129t (125-138) 4.6 (3.9-5.4) 30.7 (25.8-39.0) 41.5 (39.0-43.5) 436 (424-446) 484 (470-504) 6.20 (5.88-6.69) 7.37 (6.92-7.94) 38.16 (37.3-41.49) 89.48 (82.0-96.38) 0.78 (0.74-0.84) 0.50 (0.47-0.54)
213t (150-300) 7.8* (1.9-22.0) 50.6* (26.4-209.5) 40.0 (34.0-43.0) 390 (330-432) 488 (353-507) 6.85 (6.03-8.31) 8.20* (7.35-10.49) 43.08 (30.7-64.07) 67.81 (50.64-105.35) 0.84 (0.59-1.35) 0.42 (0.39-0.76)
175f (147-250) 4.1 (1.7-9.1) 39.9 (20.0-132.0) 40.0 (35.0-42.0) 413 (340-435) 467 (349-500) 7.16* (6.20-7.70) 8.87t (7.38-10.52) 51.19 (32.32-65.00) 134.54 (78.26-169.69) 1.07 (0.60-1.32) 0.47 (0.41-0.55)
Statistical significance: * Ρ < .05; f Ρ < .005 (Dunn's test). BP = blood pressure; LV = left ventricular; PRA — plasma renin activity; ANF = atrial natriuretic factor; CSA = cross-sectional area. Data are presented as medians of variables, with 95% confidence limits of the medians in parentheses.
that increased myocardial contractility may be an alter native mechanism to development of LV hypertrophy for sustaining LV performance in the face of hemody namic overload or that contractility becomes depressed when cardiac hypertrophy is p r e s e n t . " In this study we used a mathematical approach to derive an index of LV performance independent of af terload. This variable, therefore, may reflect the effect of myocardial contractility and/or preload reserve on LV performance. 4-6,30
the presence of compensatory LV hypertrophy; this compensation might be more difficult to achieve in 1K1C because of the additional volume overload.
33
Normal Left Ventricular Performance Both 2K1C and 1K1C with normal LV performance exhibited a de gree and pattern of LV hypertrophy sufficient, on aver age, to normalize wall stress. This anatomic compensa tion was systematically present in 2K1C, which exhibited concentric LV hypertrophy, but less consistent in 1K1C (with 4 / 1 0 rats having values of relative wall thickness below and 3 / 1 0 above the 9 5 % confidence limits of the control group), because, in the latter group the interaction of pressure and volume o v e r l o a d yields a greater variability of LV geometry. Thus, in the rats exhibiting normal load-adjusted shortening, arterial hypertension did not affect LV performance, because of 8-11
Enhanced Left Ventricular Performance Both sub groups of hypertensive rats with increased LV perform ance exhibited very high blood pressure associated with LV chamber enlargement (especially in 1K1C), but without the expected increase in LV wall thicknesses. Relative wall thickness was not significantly increased in either model, resulting in high values of wall stress (4 of 7 2K1C and 4 of 9 1K1C had values of end-systolic stress above the 9 5 % confidence limits of control ani mals). Although median values of CSA in these sub groups of rats with moderate hypertension were similar to those in the mildly-hypertensive subgroups with normal LV performance at the same interval after renal artery clipping (58 to 62 days), it is possible that they would develop, over a larger period of time, adequate LV hypertrophy to offset the higher pressure, a hypoth esis already suggested in human studies. These rats, indeed, maintained normal or slightly increased ejec tion-phase function, because of alternative compensa6
Downloaded from http://ajh.oxfordjournals.org/ at New York University on March 31, 2016
Heart rate (beats/min)
Controls (n = 32)
ONE - KIDNEY, ONE CLIP
0.5
1.0
10.0
100.0
36
2
End-Systolic Stress (*10 dynes/mm ) FIGURE 3. Relation of left ventricular fractional shortening (vertical axis) to end-systolic stress (horizontal axis, logarithmic scale) in one-kidney, one-clip rats. Solid lines represent the 9 5 % confidence interval of control animals. Closed circles represent rats with enhanced LV performance.
tory mechanisms. These mechanisms may include the use of myocardial contractility and/or preload reserve. The possibility of the use of the preload reserve might be suggested by the increase in LV diastolic dimension in the presence of normal to increased fractional shortening. Since compensation through the Starling forces is predominantly an adaptation to acute overload, it is more likely that increased values of afterload-corrected 34
35
TWO - KIDNEY, ONE CLIP
Comparison With Previous Studies Characteristics of LV function in rats with renovascular hypertension have been assessed by a variety of a p p r o a c h e s , such as examining the curves relating invasively measured LV end-diastolic pressure and cardiac output under a range of loading conditions. The ability to estimate end-systolic wall stress (the myocardial afterload that stops LV contraction) in vivo has been made possible in this study by M-mode echocardiography. Our study, therefore, evaluates for the first time LV ejection-phase systolic function in Goldblatt hypertension in rats in relation to myocardial afterload. Averill et a l reported reduction of cardiac functional response to volume infusions in ether-anesthetized open-chest 2K1C 6 to 22 weeks after clipping. In addition to the difference in experimental procedures and anesthesia, our study examines LV function only under relatively basal conditions in rats 58 to 62 days after clipping. Thus, while we have documented that both Goldblatt models preserve LV pumping ability at rest by undergoing hypertrophy and using contractile and/or preload reserves, it is possible that our animals would have resembled those of Averill et a l in being unable to respond normally to further hemodynamic overload. Left ventricular dysfunction has also been associated with severe LV hypertrophy. Kuwajima et al, using the LV end-diastolic pressure/cardiac output curve, demonstrated reduced LV function in ether-anesthetized 2K1C related to the presence of LV hypertrophy: when LV hypertrophy was reduced with methyldopa, without a comparable effect on blood pressure, LV function improved. Our results suggest that in Goldblatt rats high levels of myocardial performance are associated with small increases in myocardial mass that are inadequate to normalize wall stress. In our 2K1C animals with high LV performance the elevated plasma renin levels suggest a possible contribution from positive inotropic effects of angiotensin, ' whereas no specific mechanism of contractility enhancement is suggested by our findings in 1K1C rats. 12/13,37-39
12
12
12,38
FIGURE 4. Relation of left ventricular fractional shortening (vertical axis) to end-systolic stress (horizontal axis, logarithmic scale) in two-kidney, one-clip rats. Solid lines represent the 95% confidence interval of control animals. Closed circles represent rats with enhanced LV performance.
40 41
38
Downloaded from http://ajh.oxfordjournals.org/ at New York University on March 31, 2016
3
fractional shortening in these rats with a reasonably stable level of elevated blood pressure principally reflect enhanced myocardial contractility. On the other hand, persistent participation of the Frank-Starling mechanism to sustain the high level of LV performance is hard to exclude in these animals with mild LV chamber enlargement (especially in 1K1C, a model characterized by initially volume-dependent hypertension) in the absence of direct measurements of end-diastolic stress (the best measure of LV preload) or sarcomere length (indicative of increased preload on the contractile apparatus). In this regard, it is relevant that elevated end-diastolic stress had been observed in aortic regurgitation, a model of chronic volume overload with LV dilatation.
TABLE 2. LEFT VENTRICULAR FUNCTION IN GOLDBLATT RATS WITH HIGH LEFT VENTRICULAR PERFORMANCE
ESS Χ 10 (dynes/cm ) 3
2
Fractional shortening (%) Afterload-corrected shortening (%)
Controls (n = 32)
2-Kidney, 1-Clip (n = 7)
1-Kidney, 1-Clip (n = 9)
10.0 (7.6-14.8) 63 (61-69) 63 (61-66)
17.6* (8.8-36.3) 66 (62-74) 70t (69-74)
14.8* (13.1-31.1) 68 (59-74) 71f (69-74)
Statistical significance: * Ρ < 0.05; f Ρ < 0.001. ESS = end-systolic stress. Data are presented as medians of variables, with 95% confidence limits of the medians in parentheses.
neously with the echocardiogram, because measure ments of blood pressure during the echocardiographic session would have required prolongation of anesthe sia, potentially altering hemodynamics and increasing the risk of death during the procedure (with our proce dure no animal died during anesthesia). Although a blood pressure measurement simultaneous to the echo-
TABLE 3. BLOOD PRESSURE, HORMONAL PROFILE, AND LEFT VENTRICULAR ANATOMY IN GOLDBLATT RATS WITH NORMAL LEFT VENTRICULAR PERFORMANCE
Systolic BP (mm Hg) PRA (ng/mL/h) ANF (fmol/mL) Hematocrit (%) Body weight (g) Heart rate (beats/min) LV Dimension (mm) LV Dimension/kg (mm/kg ) 021
LV Mass (g) CSA (mm ) 2
CSA/kg (mm/kg ) 053
Relative wall thickness
Controls (n = 32)
2-Kidney, 1-Clip (n = 21)
1-Kidney, 1-Clip (n = 10)
129 (125-138) 4.6 (3.9-5.4) 30.7 (25.8-39.0) 41.5 (39.0-43.5) 436 (424-446) 484 (470-504) 6.20 (5.88-6.69) 7.37 (6.92-7.94) 0.78 (0.74-0.84) 38.16 (37.30-41.49) 89.48 (82.00-96.38) 0.50 (0.47-0.54)
150* (136-160) 5.5 (4.0-9.1) 44.1 (30.5-60.2) 41.0 (39.0-43.3) 445 (410-470) 465 (430-500) 6.28 (5.92-6.58) 7.46 (7.02-7.77) 0.91* (0.78-1.03) 45.26* (37.92-48.46) 101.72* (89.28-125.62) 0.55* (0.52-0.60)
159* (115-213) 5.8 (2.0-5.7) 22.4 (17.2-91.2) 40.5 (32.3-42.0) 375 (340-450) 464 (337-506) 6.74 (5.70-7.27) 8.23 (6.66-11.61) 1.05| (0.81-1.20) 49.37f (41.24-65.97) 116.00t (93.79-195.95) 0.51 (0.42-0.79)
Statistical significance: * Ρ < .05; j Ρ < .005 ν controls. For abbreviations see Table 1. Data are presented as medians of variables, with 95% confidence limits of the medians in parentheses.
Downloaded from http://ajh.oxfordjournals.org/ at New York University on March 31, 2016
Technical Limitations Potential limitations of this study need to be considered. The use of M-mode echo cardiography has the advantage of allowing in vivo measurements of LV wall stress, a procedure virtually impossible with other techniques. Unfortunately, endsystolic stress can be only approximated by using our method. Blood pressure was not measured simulta
TABLE 4. LEFT VENTRICULAR FUNCTION IN GOLDBLATT RATS WITH NORMAL LEFT VENTRICULAR PERFORMANCE
ESS Χ 10 (dynes/cm ) 3
2
Fractional shortening (%) Afterload-corrected shortening (%)
Controls (n = 32)
2-Kidney, 1-Clip (n = 21)
1-Kidney, 1-Clip (n = 10)
10.0 (7.6-14.8) 63 (61-69) 63 (61-66)
13.4 (10.8-22.3) 65 (57-67) 64 (62-65)
10.7 (5.93-24.4) 68 (55-72) 65 (62-68)
ESS = end-systolic stress. Data are presented as medians of variables, with 95% confidence limits of the medians in parentheses.
80
= Rats = Humans
4,5 42 43
44
45 46
4,5
c 'c
