INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Multiorgan failure resulting from sepsis is a major cause of morbidity and mortality in intensive care units. Approximately 500,000 people develop sepsis and 175,000 die annually in the United States (1). New developments in antimicrobial therapy and other pharmaceutical agents during the past two decades have not decreased mortality. Increasing evidence suggests that an uncontrolled host response to infection, i.e., the systemic inflammatory response syndrome, initiates a cascade of events resulting in cell injury and death (2). Patients with sepsis often have a reversible decrease in myocardial function that is characterized by depression of left and right ventricular ejection fraction, ventricular dilatation, and altered Frank-Starling and diastolic pressure-volume relationships (3). It is postulated that ventricular dilatation is an appropriate compensatory mechanism that allows the heart to maintain stroke volume and, hence, cardiac output. Although sepsis-induced myocardial dysfunction can occasionally be profound, a low cardiac index is uncommon even in very late stages of septic shock (3). Given that cardiac index is usually maintained in sepsis, the degree to which myocardial failure contributes to morbidity and mortality in clinical sepsis is not clear, although some investigators believe it is a major cause of death.

Investigators have employed a variety of animal models (endotoxin, bacterial peritonitis, intravenous infusion of live bacteria, etc.) to examine putative mechanisms responsible for myocardial depression in sepsis (6). The decrease in contractility in sepsis does not appear to be due to a metabolic abnormality. Studies have shown that high energy phosphates (ATP and phosphocreatine) are well-maintained in hearts during sepsis (10, 11). Similarly, the concentrations of tricarboxylic acid intermediates are not different in septic versus sham rat hearts in a rat model of intra-abdominal sepsis (10, 11). Therefore, although the heart may shift its preferred fuel (12), i.e., substrate, in sepsis, the decrease in contractility is not due to altered metabolism or bioenergetic failure.

One possible mechanism for the decrease in cardiac contractility could be an abnormality in myocyte calcium handling. Ca²⁺ plays a central role in excitation-contraction by binding to troponin C and allowing interaction of actin and myosin. The decrease in cardiac relaxation observed in sepsis could also be related to Ca²⁺ because it is the active reuptake of Ca²⁺ into the sarcoplasmic reticulum that enables the heart to relax. Further substantiating a potential role for Ca²⁺ as a mediator of the myocardial dysfunction of sepsis is the characteristic effect of sepsis/endotoxemia to cause hypocalcemia, i.e., a decrease in the free ionized plasma Ca²⁺ concentration ([Ca²⁺]_e), a hallmark of the disorder (13). Similarly, although more controversial, endotoxemia and sepsis have been reported to induce an increase in the level of free intracellular calcium concentration ([Ca²⁺]_i) in certain types of cells, including smooth muscle cells, hepatocytes, and lymphocytes (14). There is debate about the proper treatment of the hypocalcemia that occurs in sepsis. Studies in a rat endotoxin model have shown that correction of the hypocalcemia increases mortality, possibly by aggravating intracellular Ca²⁺ overload (13). If Ca²⁺ administration causes Ca²⁺ overload in the heart, both systolic and diastolic function would deteriorate.

The purpose of this study was to determine if correction of the sepsis-induced hypocalcemia in the clinically relevant and widely utilized rat cecal ligation and puncture (CLP) model of sepsis would improve or worsen myocardial systolic and diastolic function. Secondly, we evaluated the effect of correction of afterload on myocardial function in sepsis. A decrease in afterload likely due to nitric-oxide-mediated vasodilatation is a frequent finding in sepsis (3). Failure to correct the afterload in sepsis when assessing myocardial performance can lead to spurious conclusions about the degree of myocardial depression (6). If afterload is decreased, the heart can eject more readily against the decreased pressure load, thereby reducing the maximal rate of pressure rise in the ventricle (+dP/dt) as well as decreasing the time-to-peak pressure and the peak pressure produced. Thus, cardiac function may appear falsely to be significantly depressed in sepsis if afterload is reduced (6). To obtain a better assessment of left ventricular (LV) contractility in sepsis, phenylephrine, a pure alpha agonist, was used to restore afterload (LV-developed pressure) in the septic rats to values comparable to sham-operated rats. Cardiac contractility was assessed by simultaneous hemodynamic and echocardiographic measurements of left ventricular pressure (LVP) using an intraventricular Millar catheter and LV geometry using two-dimensional (2-D) guided M-mode echocardiography, respectively.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Animal Model Cecal Ligation and Perforation

All animal procedures were approved by the Committee on the Humane Care of Laboratory Animals at Washington University. Male Sprague-Dawley rats weighing 300 to 400 g were purchased from Harlan (Omaha, NE), housed in a room with constant temperature (23° C), and exposed to a 12-h light-dark cycle for 7 d before surgery. The CLP model modified from Wichterman and coworkers (15) was used to induce sepsis. Briefly, rats were anesthetized with halothane, and a 2- to 3-cm abdominal incision was made to expose the cecum. A 4-0 silk ligature was placed around the base of the cecum after division of the vascular mesentery at this location. The cecum was devascularized, but the intestines were not obstructed. The cecum was punctured once with a 19-gauge needle and a small amount of feces extruded. The bowel was returned to the abdomen and the abdominal cavity was closed in two layers. Control rats underwent an abdominal laparotomy and cecal manipulation. Ten milliliters of 0.9% normal saline were administered subcutaneously on the dorsum of the back to both septic and sham-operated rats to adjust for intravascular fluid loss.

Assessment of Myocardial Contractility

Approximately 20 to 28 h after CLP or sham surgery, rats were anesthetized by an intraperitoneal injection of ketamine (55 mg/kg body weight) and medetomidine (0.375 mg/kg body weight). Septic rats required approximately two-thirds the amount of ketamine/medetomidine to achieve an adequate level of anesthesia. Oxygen was administered by a nose cone, and body temperature was maintained at 36 to 37° C by an infrared heating lamp. A micromanometer catheter (Model SPR-524, 2.5 F; Millar Instruments, Houston, TX) was inserted via the right carotid artery and passed retrogradely into the left ventricle. Confirmation of position was noted by the characteristic decrease in diastolic pressure that occurred with passage of the catheter across the aortic valve into the LV cavity. A left femoral arterial line was placed for infusion of calcium chloride solution (10 mg in 1 ml of sterile water, 90 µmoles). In a subgroup of septic rats, phenylephrine (1.23 to 2.46 nmoles/min) was infused to correct afterload via a left jugular venous line. Note that the afterload was considered equal to LV developed pressure. No significant pressure gradient was seen across the aortic valve and, therefore, LV developed pressure was considered equal to systemic arterial blood pressure.

After the confirmation of successful placement of the Millar catheter in the LV cavity, LVP was recorded on a high-fidelity FM tape recorder (TEAC, Model XR-510). As previously described (16, 17), the analog LVP data were then transferred off-line into a computer (Macintosh IIvx) and digitized at a 1-ms sampling rate. A wave form analysis program determined peak LVP, peak positive (+dP/dt), and negative (dP/dt) rate of change of LVP, and LV end-diastolic pressure (EDP). The LV isovolumic relaxation constant (Tau) was calculated with use of the least-squares method for the LVP interval starting at peak dP/dt and ending at 5 mm Hg above LVEDP, based on the model of exponential decay with a variable asymptote:

(1)

where P(t) = left ventricular pressure at time t, a = left ventricular pressure at t = 0 with respect to the variable asymptote, c = the value of the asymptote at t = infinity, b = the variable defining the rate of pressure decay, and e = the base of the natural logarithm. Tau is defined as the time required for the pressure at t = 0 to decline by 50%. This index has the advantage of being based on a model that provides statistically superior fits to the observed data, its calculation is not tied to the pressure asymptote, and it yields a value that, except in rare cases, occurs within the isovolumic relaxation period (18).

Echocardiography

Cardiac function was assessed simultaneously with measurements of LVP at baseline, after alteration of the afterload using phenylephrine, and after low and high dose CaCl₂ administrations. Echocardiograms were performed with a commercially available echocardiographic system equipped with a 12-MHz dynamic phased-array transducer (Hewlett Packard, Andover, MA). Images were obtained with the rats lying on their backs and extremities secured to the operating table. Care was taken to maintain adequate contact while avoiding excessive pressure on the chest wall. Two-dimensional parasternal short-axis images of the left ventricle were obtained at the level of the papillary muscles. The tip of the intraventricular Millar catheter was used to ensure that imaging was performed at comparable planes of the left ventricle. Two-dimensional targeted M-mode tracings were recorded through the anterior and posterior ventricular walls at a sweep speed of 100 mm/s. Echocardiograms were recorded on a S-VHS tape. Anterior and posterior end-diastolic and end-systolic wall thickness and LV internal dimensions were measured according to the recommendations of the American Society for Echocardiography leading-edge method from three consecutive cardiac cycles (19). Data were analyzed with a commercially available off-line image analysis system (Freeland Cardiac Workstation; Tomtec Imaging, Boulder CO) by experienced observers blinded to the experimental conditions. Percent LV fractional shortening (FS) was calculated as follows: FS = (LVIDd  LVIDs)/LVIDd × 100, where LVIDd and LVIDs are end-diastolic and end-systolic LV dimensions. Relative wall thickness was calculated as RWT = (PWd + IVSd)/LVIDd, where PWd and IVSd are end-diastolic posterior wall and interventricular septal thickness, respectively. The ejection time (ET) was measured from the pulsed-wave Doppler flow velocity envelope recorded across the aortic valve. The rate-corrected mean velocity of LV fiber shortening (Vcf_c), a measurement of ventricular function that incorporates the heart rate, was calculated as follows:

(2)

where RR is the interval between adjacent R waves on the ECG, and ET_c is the ejection time corrected for heart rate. In addition, left ventricular end-systolic meridional wall stress (_es) was calculated as follows:

(3)

where Ps is peak-systolic pressure, WTs is end-systolic wall thickness, and 0.337 is a conversion factor. The relationship between left ventricular end-systolic wall stress and rate-corrected velocity of fiber shortening is a clinically accepted index for evaluation of left ventricular contractile state (20): it is a preload- and heart rate-independent index of contractility and incorporates afterload into its analysis.

Measurement of Plasma Ionized Calcium

In a subgroup of septic and sham-operated rats, plasma ionized Ca²⁺ [Ca²⁺]_e was measured before and after calcium administration. Blood was withdrawn from a femoral arterial catheter into a heparinized syringe and ionized Ca²⁺ was determined using either a Nova Stat Profile (Nova, Waltham, MA) or, alternatively, the Radiometer ABL (Radiometer, Copenhagen, Denmark) in the Barnes Hospital Chemistry Laboratory.

To determine the effect of Ca²⁺ administration on myocardial performance, CaCl₂ (10 mg/ml) was administered by a Harvard Apparatus infusion pump (Harvard Apparatus, South Natick, MA). The effect of low dose CaCl₂ and high dose CaCl₂ were examined. A total of 90 µmoles of CaCl₂ were contained in 1,000 µl of sterile water. Measurement of cardiac contractility and relaxation was performed at baseline, after infusion of 250 µl of CaCl₂ solution (low dose), and after administration of an additional 750 µl of CaCl₂ solution (high dose). [Ca²⁺]_e was measured in a subgroup of septic and sham-operated rats at the end of the total 1,000 µl of CaCl₂ infused.

Statistical Analysis

Comparison of hemodynamic and echocardiographic values in septic versus sham-operated rats was performed by unpaired t tests using the Stat View statistical program. Results are expressed as mean ± SEM with a p < 0.05 chosen as statistically significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Confirmation of the Septic Model

Two of the 14 rats that underwent CLP died prior to examination. One CLP rat died during cardiac evaluation and was excluded from the study. No sham-operated rats died. At 20 to 28 h after surgical manipulations, the CLP-operated rats exhibited signs of sepsis, including piloerection, exudates around the eyes and nose, tachypnea, diarrhea, and decreased spontaneous movement. The sham-operated rats were active within their cages and appeared normal. Examination of the peritoneal cavity of septic rats demonstrated copious amounts of foul-smelling purulent peritoneal fluid, and the ligated portion of the cecum was grossly dilated and gray-black in color. Abdominal examination of control rats showed no noticeable odor, minimal peritoneal fluid, and the bowel was pink in color.

Hemodynamic Measurements

Parameters of systolic and diastolic function at baseline. Baseline assessment of cardiac indices revealed the characteristic changes noted in sepsis. Peak LVP was significantly lower in septic (98 ± 5 mm Hg, n = 9) than in sham-operated rats (136 ± 6 mm Hg, n = 9; p < 0.05) (Figure 1a). Septic rats had an increase in heart rate compared with sham-operated rats (273 ± 13 versus 238 ± 13 beats/min, respectively; p < 0.05) (Figure 1b). Peak +dP/dt and peak dP/dt were reduced in septic versus sham rats by 30.3% and 23.1%, respectively (p < 0.05) (Figures 1c and 1d). Left ventricular end diastolic pressure was slightly but significantly reduced in septic (n = 9) compared with sham (n = 9), 1.9 ± 0.6 versus 3.7 ± 0.6 mm Hg, respectively, p < 0.05. The isovolumic relaxation constant (Tau) was reduced in hearts of septic rats by 20%, (7.6 ± 0.4 versus 9.6 ± 0.5, p < 0.05).

View larger version (29K): [in this window] [in a new window]   Figure 1.   Hemodynamic measurements in septic and in sham rats. One group of septic rats (n = 9) received low dose CaCl2 infusion followed by high dose CaCl2 infusion in an identical manner to that of the sham rats (n = 9). A second group of septic rats (n = 5) received phenylephrine to increase peak LVP to a level comparable to that of the sham rats at baseline, i.e., pre-CaCl2 infusion. After attaining the desired peak LVP with phenylephrine, the septic rats then received low dose CaCl2 followed by high dose CaCl2. Values reported are mean ± SEM; *p < 0.05 for septic versus sham. (a) Effects of CaCl2 and phenylephrine administration on peak left ventricular pressure (LVP). Note the decreased peak LVP in septic (n = 9) versus sham (n = 9) rats at baseline (p < 0.05). (b) Heart rates (HR) in septic and those in sham rats. Neither CaCl2 nor phenylephrine had any significant effect on heart rate. At high dose CaCl2, the heart rates in septic rats were greater than those in sham rats (p < 0.05). (c ) Left ventricular contractility (+dP/dt) in septic and in sham rats. Baseline +dP/dt was decreased in septic versus sham rats (p < 0.05) but normalized after treatment with phenylephrine. (d  ) Effects of CaCl2 and phenylephrine administration on peak dP/dt. Baseline dP/dt was delayed in septic versus sham rats (p < 0.05) but normalized with phenylephrine.

Effect of CaCl₂ administration. The ionized plasma Ca²⁺ was decreased in septic compared to control rats, i.e., 4.9 ± 0.9 and 5.6 ± 0.01 mg/dl, respectively; p < 0.02 (n = 6). Administration of the entire 1,000 µl of CaCl₂ (90 µmoles) solution caused ionized Ca²⁺ to increase by ~ 75% in both septic and sham rats. (Note that ionized Ca²⁺ was determined at the end of the entire 1,000 µl of CaCl₂, i.e., high dose CaCl₂ and not after low dose CaCl₂ to avoid changes in hemodynamics caused by aspiration of blood. Cardiac contractility was assessed after administration of low dose CaCl₂ (250 µl of CaCl₂ solution) and high dose CaCl₂ (a total of 1,000 µl of CaCl₂ solution). CaCl₂ did not change heart rate significantly in either septic or sham rats (Figure 1b). Low dose CaCl₂ did not significantly increase LVP in either septic or sham rats. High dose CaCl₂ increased LVP in septic and sham rats by 15 ± 4 and 27 ± 8 mm Hg, but this increase was not statistically different from baseline (p > 0.1). Despite high dose CaCl₂, peak LVP remained decreased in septic versus sham animals (p < 0.05) (Figure 1a). High dose but not low dose CaCl₂ caused significant increases in +dP/dt and dP/dt in both septic and sham rats, but septic remained statistically reduced compared with sham (p < 0.05) (Figure 1, c and d).

Effect of phenylephrine and phenylephrine plus Ca²⁺. Phenylephrine was administered to septic rats and titrated to increase peak LVP to 134 ± 7 mm Hg, which was not statistically different from sham. Phenylephrine had no significant effect on either heart rate or left ventricular end-diastolic pressure in the septic rats. Importantly, correction of afterload with phenylephrine caused both +dP/dt and dP/dt to equalize in the septic and sham rats (Figure 1, c and d).

After augmentation of afterload in the septic rats with phenylephrine, Ca²⁺ administration resulted in equivalent changes in all parameters in septic and sham rats, i.e., +dP/dt, dP/dt, peak LVP, and Tau were indistinguishable in septic and sham animals (Figure 1).

Echocardiographic Measurements

To determine the effect of sepsis on the load-dependent parameters of LV geometry and systolic function, 2-D guided M-mode echocardiograms were performed simultaneously with the hemodynamic measurements at baseline, after phenylephrine and two different doses of CaCl₂ infusion.

Baseline measurements. Technically high quality images were obtained in all rats (Figures 2 and 3). The septic state caused the animals to develop a significantly smaller end-diastolic LV cavity size and a concomitantly increased relative wall thickness (Table 1 and Figure 3). These changes in LV geometry are likely the result of decreased preload and/or afterload, and they are consistent with the lower end-diastolic and peak systolic LV pressures obtained by the hemodynamic measurements (Figure 1). Load-dependent ejection phase indexes of LV systolic performance such as fractional shortening and percent wall thickening showed no significant differences between septic and sham-operated animals (Table 1). In contrast, the velocity of circumferential fiber shortening (another load-dependent index of systolic function that incorporates heart rate into the equation) was in fact elevated in the septic rats, which would imply enhanced systolic function but is more likely the result of decreased afterload of the septic state as evidenced by the significantly decreased LV systolic wall stress in the septic animals (Figure 4).

View larger version (118K): [in this window] [in a new window]   Figure 2.   Example of two-dimensional echocardiogram from a sham-operated rat. This parasternal short-axis image was obtained at the level of the papillary muscles (PM). The echogenic structure in the left ventricular cavity near the septum is the tip of the Millar micromanometer catheter (MC). Both of these structures were used as guides to ensure the consistency of the imaging plane.

View larger version (70K): [in this window] [in a new window]   Figure 3.   Examples of M-mode echocardiograms obtained with two-dimensional guidance from a short-axis midventricular view from a sham-operated (A) and a septic (B) rat. AW indicates anterior wall; PW indicates posterior wall. Note the decreased left ventricular cavity dimensions, increased wall thickness, and increased heart rate in the septic rat.

View larger version (10K): [in this window] [in a new window]   Figure 4.   Relationship between rate-corrected velocity of fiber shortening (Vcfc) and end-systolic meridional left ventricular wall stress (es) in sham-operated and septic rats. The relationship at two different levels of afterload (solid lines) represents the contractile state of the myocardium at baseline. There was no significant difference in the slope and y-intercept of these lines, demonstrating that the contractile state of the myocardium at baseline was the same in septic and control animals.

Effect of phenylephrine infusion. Phenylephrine was administered to sham-operated rats and titrated to increase peak left ventricular pressure by about 20 to 30 mm Hg. None of the echocardiographic parameters of LV geometry and systolic function changed significantly compared with those of the baseline values; however, at this slightly higher level of afterload there was a trend toward increased end-diastolic cavity size and decreased systolic function (Table 1). Administration of phenylephrine to septic rats resulted in changes of these echocardiographic parameters to levels that were not significantly different from those of the sham-operated animals.

Effect of CaCl₂ administration. To determine the effect of Ca²⁺ on LV geometry and systolic function, we measured the same echocardiographic parameters at two different doses of CaCl₂ in sham-operated rats without concurrent pharmacologic manipulation of afterload by phenylephrine administration and in septic rats with continuous infusion of phenylephrine at doses that elevated the baseline blood pressure to levels that were comparable to that of sham-operated control animals. There was a gradual increase in fractional shortening and percent wall thickening detected in both sham and septic animals with CaCl₂ administration that reached statistical significance only at high doses (Table 1). CaCl₂ infusion did not significantly affect LV geometry and the velocity of circumferential fiber shortening, but there was a trend toward decreased LVIDd and increased Vcf_c in both septic and control animals with low dose and even more so with high dose CaCl₂ compared with baseline measurements, but the degree of change in these and in all other parameters was not different between septic and sham-operated rats (Table 1). Once again, these echocardiographic parameters of LV structure and function confirm the results of the hemodynamic measurements.

Load-independent Measurement of LV Systolic Function

To correct for the significantly different loading conditions observed in the septic rats, we performed simultaneous LV pressure measurements and 2-D guided M-mode echocardiograms at two different levels of LV afterload (baseline and phenylephrine infusion) in both septic and sham-operated rats, and determined the relationship between the rate-corrected velocity of circumferential shortening and LV end-systolic wall stress. Left ventricular Vcf_c plotted against _es for the combined data of five septic and four sham animals is shown in Figure 4. The lines connecting the averaged data points represent baseline contractility. The slopes of these lines are identical for the septic and control rats but the data points of the two groups are located on a different portion of the same line, demonstrating that, although loading conditions are different (reduced afterload in sepsis), the baseline contractile states are indistinguishable under these experimental conditions.

To evaluate the effect of Ca²⁺ on the contractile state of the myocardium in sepsis we performed the same simultaneous LVP and echocardiographic measurements as described in the previous paragraph after the infusion of two different doses of CaCl₂. The averaged data of the Vcf_c-_es relationship at low and high dose CaCl₂ infusion is shown in Figure 5. The distance to the mean regression line represents the deviation in Vcf_c units at a given level of afterload. Because Vcf_c incorporates heart rate and is preload-independent, comparisons of the deviations of Vcf_c from the mean regression line can be used to assess relative LV contractility independent of changes in loading conditions induced by pharmacologic interventions. There was a trend toward an upward shift of the average data points in both groups, especially at high dose CaCl₂, but the differences did not reach statistical significance.

View larger version (34K): [in this window] [in a new window]   Figure 5.   Relative differences between Vcfc values obtained after CaCl2 infusion and baseline values extrapolated from the Vcfc-es relationship shown in Figure 4. A positive difference between these values represents increased contractility. The differences were not statistically significant; nevertheless, there was a trend toward increased contractility after CaCl2 infusion, especially in the septic group at high dose Ca2+ administration.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Perhaps the most important finding in the present study was the recognition that myocardial contractility in the rat CLP model was not depressed during sepsis. Prior to augmentation of afterload with phenylephrine, we presumed that sepsis induced major defects in both contractility and relaxation because of the reduction in peak left ventricular pressure, +dP/ dt, and dP/dt observed in septic animals. However, after correction of afterload with the pure alpha-agonist phenylephrine, there was no difference in any index of left ventricular systolic or diastolic function. Simultaneous echocardiographic assessment of ventricular function confirmed the results obtained via intraventricular pressure measurement. Once afterload was corrected in the septic animals, 2-D guided M-mode echocardiograms showed no difference in the percent fractional shortening of the left ventricle or the percent wall thickening of the myocardium. The LV loading conditions were found to be significantly different between septic and sham-operated rats as evidenced by diminished end-systolic wall stress (left ventricular afterload) and decreased end-diastolic internal left ventricular dimension and end diastolic pressure (left ventricular preload) in the septic rats compared with the sham-operated control rats; therefore, these load-dependent indexes of left ventricular performance do not reflect the true contractile state of the myocardium under septic conditions. Several relatively load-independent indexes of left ventricular performance such as the end-systolic wall stress-rate corrected velocity of fiber shortening relationship have been found to be useful in the assessment of the contractile state of the myocardium in humans (20, 21), and most recently in rodents (22, 23). Our results reported here show that this relationship is inverse and linear, similar to previously reported data (20). The experiments also demonstrate that although loading conditions are significantly different in sepsis, the baseline contractile state of the myocardium is not altered under these experimental conditions. Our results highlight the work of Abel (6) who noted that many of the studies examining myocardial function in sepsis and endotoxin shock were flawed because of failure to correct for changes in preload and afterload in the two conditions. As Abel stated, "If afterload is reduced, the heart can eject more readily against the reduced pressure load. For the left ventricle, this means that the aortic valve will open at a lower pressure, thus reducing the maximal rate of pressure rise in the ventricle, which is a function of the isovolumic pressure reached, as well as decreasing the time to peak pressure and the peak pressure produced." The decreased baseline peak LVP, +dP/dt, dP/dt observed in septic animals was presumably due to decreased afterload likely mediated by nitric-oxide-induced peripheral vasodilatation.

It should be emphasized that the septic model employed in the present study resulted in animals that demonstrated all the hallmarks of sepsis. Both hemodynamic changes (tachycardia, decreased peak LVP, +dP/dt and dP/dt depression, decreased left ventricular filling pressure by echocardiography) and physical examination (gross peritonitis) confirmed the septic state. Furthermore, our previous results using this rat CLP model demonstrate blood cultures positive for gram-negative and gram-positive organisms in septic but not in sham animals (24). In addition, plasma metabolites with this rat CLP model show typical metabolic changes of sepsis with increased lactate, decreased ketone bodies, and increased tyrosine. The new metabolic findings of the present study, i.e., the decreased plasma ionized Ca²⁺ in the septic rats, also support the establishment of the septic state using this model. Although the present study does not demonstrate myocardial depression in the rat CLP model at 20 to 28 h after onset of sepsis, it can be argued that evaluation of cardiac function at later time points may have done so. We selected the 20- to 28-h time point because our studies with this model of sepsis demonstrate that mortality from sepsis begins at approximately 24 h. Therefore, if no myocardial depression was present at 24 h, it would be unlikely that cardiac depression was a major cause of morbidity and mortality during the disorder. As noted previously, two of the 14 CLP rats died prior to the study and one died during the examination.

Investigators employing other models of sepsis/endotoxemia have demonstrated conclusively that myocardial depression occurs in their particular animal model (6, 25). Many of these investigators used endotoxin or intravenous infusions of live bacteria. These models cause enormous elevations in levels of circulating cytokines, e.g., TNF-, IL-1, IL-1, etc., which are orders of magnitude higher than occur in the CLP model of sepsis. Cytokines are potent depressors of myocardial function (28); one of the reasons for the different findings may therefore be the type of model employed. We selected the widely utilized rat CLP model instead of the endotoxin model because the CLP model is believed to more accurately reflect the clinical condition of sepsis that occurs in patients (15, 29). In this regard, the rat CLP model causes many of the same findings observed in patients, including increased plasma lactate, decreased ketone bodies, hypocalcemia, decreased peripheral resistance, increased muscle protein breakdown (29, 30), etc. However, our results indicate that the rat CLP model is not an appropriate model to examine the myocardial depression of clinical sepsis. Although the actual percentage of patients with sepsis who develop myocardial depression is not clear, clinical studies demonstrate conclusively that sepsis can cause moderate-to-severe myocardial depression in patients that is reversible with resolution of the disease (3).

Although the present study did not demonstrate any intrinsic defect in Ca²⁺ regulation in the heart during sepsis, it is important to note that the septic rats did have decreased [Ca²⁺]_e compared with the sham rats. The etiology of the hypocalcemia in sepsis is not known, but two mechanisms that have been postulated are: (1) an increase in circulating calcium-binding compounds during sepsis and (2) a net movement of extracellular Ca²⁺ into the intracellular compartment (13). There is debate about the proper therapy of hypocalcemia in sepsis (13), and studies have indicated that Ca²⁺ administration to correct the decreased [Ca²⁺]_e may be detrimental (13, 30). Our results do not indicate any adverse effect of Ca²⁺ administration on myocardial systolic or diastolic function even at supra-physiologic levels of [Ca²⁺]_e. Nevertheless, Ca²⁺ administration may have adverse effects on other organs during sepsis. One implication of the present study is that baseline [Ca²⁺]_i is unlikely to be increased in the rat heart in the CLP model of sepsis. If [Ca²⁺]_i was increased in cardiac myocytes during sepsis, the response to Ca²⁺ administration may not have been similar in sepsis and sham. However, the present study does not rule out an increase in [Ca²⁺]_i in other organs during sepsis. In fact, our previous work using 19F nuclear magnetic resonance NMR spectroscopy documented a greater than twofold increase in [Ca²⁺]_i in the perfused aorta from septic rats using a CLP model (31). Recent studies using in vivo 19F NMR also indicate that sepsis causes an approximately 50% increase in [Ca²⁺]_i in brain (unpublished data). Different types of cells, e.g., hepatocytes, smooth muscle, skeletal muscle, neurons, etc., have different mechanisms of controlling the cytosolic Ca²⁺ set point. For example, the Na⁺/Ca²⁺ exchanger plays an important role in cardiac and smooth muscle cell Ca²⁺ regulation, but it is not believed to function in hepatocytes (32).

Recently, Piper and associates (33) examined the effect of sepsis on structure-function relations in hearts from rats that underwent CLP. These investigators employed a retrogradely perfused nonworking rat heart model and found a decrease in left ventricular function as assessed by dP/dt, +dP/dt, and ventricular developed pressure. Hearts were examined also by electron microscopy, and wet-to-dry weight ratios were performed. These investigators found no evidence of cell injury or capillary edema at either 24 or 48 h after CLP. We also did not detect any morphologic changes in septic rat hearts examined by transmission electron microscopy and scanning electron microscopy (unpublished observations). Although we agree with Piper and associates on the lack of effect of sepsis on cell ultrastructure, we found no effect of sepsis on LV function (left ventricular developed pressure, +dP/dt, dP/dt) in a nonworking, retrogradely perfused rat heart (34). Therefore, the data in our previous study agree with our present report. There is no clear explanation for the different findings in the study of Piper and associates and the current work. Of note, peak left ventricular developed pressure was 140 ± 40 and 120 ± 50 mm Hg for sham and CLP hearts, respectively, in our previous study versus approximately 75 ± 10 and 45 ± 5 mm Hg for sham and CLP hearts, respectively, in the study of Piper and associates.

Potential Methodologic Limitations

Several potential limitations of these methods should be acknowledged. First, we used intravenous infusion of phenylephrine as a selective stimulant of the ₁-adrenergic receptors to manipulate afterload in order to approximate the hemodynamic status of the sham animals in the septic group and to establish end-systolic stress-shortening relationship and hence determine load-independent indices of LV contractility. The potential direct or indirect effects of this agent on the contractile state of the myocardium cannot entirely be excluded, but evidence indicates that, in the doses used in this study, the primary effect of the drug was a minimal degree of peripheral vasoconstriction. The average change in the blood pressure in response to phenylephrine was about 30 mm Hg, which is a relatively minor effect. No changes were noted in heart rate in either of the groups: there was neither tachycardia as a potential -effect nor bradycardia as evidence of reflex phenomenon. Furthermore, in recently published experiments validating the end-systolic stress-shortening relationship as a reliable parameter of contractility (22, 23, 35), the investigators used similar doses of the ₁-agent with similar hemodynamic effect as was shown in our study and found that the primary effect of phenylephrine was an increase in afterload without direct effect on contractility. Finally, in studies where the direct effect of ₁-stimulation of the myocardium was demonstrated (36), phenylephrine was used in a dose that was at least 280 times greater than in our investigation. Second, the anesthetic combination we used in our study has known cardiovascular effects that may influence the results of the experiments. Nevertheless, the animals were anesthetized to an equivalent level of anesthesia, and thus the differences noted in the septic versus the sham rats would be unlikely to be due to anesthesia alone. Furthermore, the cardiovascular effect of the anesthetic combination (mainly a slowing of the heart rate) was likely to be similar in the two groups as evidenced by the hemodynamic and echocardiographic data showing the expected changes of the septic state. Also, this anesthetic combination has much less effect on myocardial contractility than most other anesthetics.

Third, we used the end-systolic stress-shortening relationship to characterize the contractile state of the myocardium as opposed to the more widely excepted "gold standard" of end-systolic pressure-volume relationship. The LV end-systolic stress-shortening relation is a well-established and useful index of contractility that is relatively load-independent (21). This method has been used in a variety of experimental preparations and animal species, as well as in humans (37, 38). This end-systolic stress-shortening relation as a load-independent measure of contractility has recently been validated in rodents (22, 23). The relationship was shown to be linear over a wide range of afterloads and found to be sensitive to inotropic stimulation and depression. To the best of our knowledge, direct comparison of the end-systolic stress-shortening relation to the LV pressure-volume relation to determine the relative sensitivity of these indices in the characterization of the contractile state of the rodent myocardium has not been published. It is important to note that in a recent study both the end-systolic stress-shortening relation and the pressure-volume relation were used to determine the contractile function of the myocardium in a transgenic model of dilated cardiomyopathy (35). The investigators found that the analysis of both of these relationships showed very concordant results with similar level of significance (or even higher level of significance in favor of end-systolic stress-shortening relation) in the differences between wild type and transgenic animals. Furthermore, when we analyzed the end-systolic LV pressure- dimension relationship in our study in the sham and septic animals at two different levels of afterload (data not shown) the slopes of the lines were not significantly different, indicating that the contractile state of the myocardium is comparable in this model of sepsis. We would also note that the scatter of the data points in the pressure-dimension plot was greater then in the stress-shortening relationship; therefore, this method may actually be less sensitive in detecting alterations in contractility. A more automated technique for the simultaneous and continuous measurement of LV pressure and chamber dimensions, however, could yield more sensitive detection of altered contractility. Nevertheless, in our view, the relative importance of a minor potential impairment in LV contractility in contributing to the overall morbidity and mortality of sepsis would be minimal.

In conclusion, we found no evidence of adverse effects of Ca²⁺ administration on cardiac function in sepsis. Although the rat CLP model of sepsis reproduces many of the clinical abnormalities that occur in patients, it may not be an appropriate model to study cardiac dysfunction in the disorder.

Note added in proof: Zhou, Wang, and Chaudry recently reported that cardiac contractility and structure were not significantly compromised even during the late hypodynamic phase of sepsis using the rat cecal ligation and puncture model. (Zhou, M., P. Wang, and I. H. Chaudry. 1998. Cardiac contractility and structure are not significantly compromised even during the late hypodynamic stage of sepsis. Shock 9: 352-358.)