Effects of ibuprofen and pentoxifylline on the cardiovascular response of normal humans to endotoxin G. DANIEL MARTICH, MARGARET M. PARKER, ROBERT E. CUNNION, AND ANTHONY F. SUFFREDINI Critical Care Medicine Department, National Institutes of Health, Bethesda, Maryland MARTICH,G.DANIEL,MARGARET M. PARKER,ROBERT E. CUNNION, AND ANTHONY F. SUFFREDINI. Effects of ibuprofen and pentoxifylline on the cardiovascular response of normal humans to endotoxin. J. Appl. Physiol. 73(3): 925-931, 1992.Endotoxin is a major mediator of the life-threatening cardiovascular dysfunction that characterizes Gram-negative sepsis. In animal models of endotoxemia, pretreatment with ibuprofen or pentoxifylline attenuates some of these cardiovascular changes. To evaluate the effects of these agents on the human cardiovascular response to endotoxemia, hemodynamic variables were measured serially in 24 normal subjects who were given intravenous endotoxin. The subjects were randomized to receive oral ibuprofen (n = 9), pentoxifylline (n = lo), or no medication before endotoxin administration (n = 5). The subjects were volume loaded 3-5 h after endotoxin administration, and hemodynamic measurements were reassessed. Core temperature after endotoxin alone or endotoxin-pentoxifylline approached a maximum at 3 h (238.6”C), while the endotoxinibuprofen group remained afebrile. At 3 and 5 h, all three groups had significant increases in heart rate, cardiac index, oxygen delivery, and oxygen consumption, while systemic vascular resistance index decreased significantly from baseline. The oxygen extraction ratio remained unchanged. After volume loading, the left ventricular ejection fraction and left ventricular end-diastolic and end-systolic volume indexes did not differ among the groups. The hyperdynamic cardiovascular response to endotoxin in humans occurs in the absence of fever and is not significantly ameliorated by oral cyclooxygenase or phosphodiesterase inhibition. hemodynamics; cytokines; prostaglandin inhibition; phosphodiesterase inhibition; tumor necrosis factor; sepsis; oxygen consumption; oxygen delivery

THE MORTALITY RATE from septic shock remains

high (25-75%), even with aggressive antibiotic and cardiovascular support in an intensive care setting (4). Endotoxin, a component of the Gram-negative bacterial cell wall, plays a central role in the pathogenesis of multiple organ failure due to septic shock (6). The development of organ failure is due, in part, to host inflammatory responses initiated by infection. Controlled trials have not shown benefit from corticosteroids, despite their broad spectrum of anti-inflammatory activities (4). Investigators have postulated that the use of specific anti-inflammatory agents might ameliorate organ dysfunction during shock (17). Two selective agents, the cyclooxygenase inhibitor ibuprofen and the phosphodiesterase inhibitor pentoxifylline, have been associated with improved hemodynamics and improved survival in animal models of

20892

endotoxemia (1,7,17,18,24). Ibuprofen is a potent antipyretic agent that diminishes the stress hormone and catecholamine responses of normal humans to intravenous endotoxin (20). Pentoxifylline can alter the acute cytokine response to intravenous endotoxin in humans, but its effects on the systemic cardiovascular response after endotoxin have not been characterized (16, 29). We previously reported that intravenous endotoxin administration to normal humans results in a hemodynamic response qualitatively similar to that of human septic shock: increased cardiac index, increased heart rate, decreased systemic vascular resistance index, and depressed left ventricular function (23). The current study was designed to assess whether ibuprofen or pentoxifylline alters the cardiovascular responses to endotoxin. To assess left ventricular responses to altered preload and to simulate the clinical therapy of sepsis, each subject was volume loaded with intravenous saline. Because of the integral role of tumor necrosis factor (TNFcy) in the pathogenesis of septic shock, we additionally evaluated the association between cardiovascular responses and TNF-a levels (25, 26). METHODS

healthy nonsmoking subjects Subjects. Twenty-four 19-40 yr old were studied. Nineteen were men and five were women. All had normal physical examinations, electrocardiograms, chest radiographs, and blood and urine analyses. The study was approved by the Institutional Review Board on Human Experimentation and performed in accordance with accepted ethical standards. Written informed consent was obtained from each subject. Endotoxin preparation and randomization. Purified lipopolysaccharide prepared from Escherichia coli 0113 (US Standard Reference Endotoxin, Bureau of Biologics, Food and Drug Administration, Bethesda, MD) was administered intravenously (4 rig/kg body wt) over 1 min and flushed with 10 ml of normal saline. The study was designed using a 2:2:1 randomization to evaluate the effects of ibuprofen and pentoxifylline on the hemodynamic response. The sample size was determined using a Monte Carlo simulation (to give a power of 0.7) based on the variability observed in left ventricular ejection fraction alterations measured in a previous study (23). The subjects were randomized into three groups: 1) no anti-inflammatory treatment before endotoxin (n = 5), 2) oral ibuprofen (800 mg; Upjohn, Kala925

Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (131.170.021.110) on October 24, 2018. Copyright © 1992 American Physiological Society. All rights reserved.

926

CARDIOVASCULAR

RESPONSE

mazoo, MI) 1.5 h before, at the time of, and 3 h after (4OO-mg endotoxin (n = 9), and 3) oral pentoxifylline extended-release tablets; Hoechst-Roussel, Somerville, NJ) every 8 h for five doses before endotoxin (n = 10). The dose of ibuprofen was chosen to attain plasma levels associated with effective cyclooxygenase inhibition (20). The dosing of pentoxifylline was based on its known serum half-life with the intent of achieving steady-state levels at the time of endotoxin administration and for the following 4 h (27). Plasma levels of ibuprofen in blood samples taken at 0 and 3 h after endotoxin administration (Pharmakinetics Laboratory, Baltimore, MD) and pentoxifylline in blood samples taken 0,1.5, and 4 h after endotoxin administration (Dr. M. Amantea, Pharmacy of Health) Development Service, National Institutes were measured using high-performance liquid chromatography and have been reported previously (16). Protocol. After an overnight fast, the subjects were admitted to the intensive care unit and remained at bed rest for r8 h. All su b‘jet t s underwent continuous electrocardiographic monitoring. Radial artery (Arrow International) and pulmonary artery catheters (VIP, American Edwards Laboratories) were placed percutaneously, and hemodynamic measurements were obtained as previously described (23). Core temperature was measured hourly with the thermistor of the pulmonary artery catheter. All catheters were removed 8 h after the administration of endotoxin. Radionuclide cineangiograms were obtained at baseline and after fluid loading by labeling the subjects’ erythrocytes with technetium-99m in vivo, as previously described (23). Baseline (O-h) measurements were obtained 30-60 min before the injection of endotoxin. Hemodynamic responses were evaluated at two subsequent time points: 1) 3 h after endotoxin administration, when previous studies have shown temperature and cardiac index to be maximal, and 2) 5 h after endotoxin and fluid loading, to assess ventricular performance with an increased preload. After hemodynamic measurements were obtained at 3 h, fluid loading was initiated by intravenous infusion of normal saline in graded increments to achieve an increase in the pulmonary arterial wedge pressure of 25 mmHg from baseline. TNF-ar ctssctys.The acute cytokine responses of the participants in this study have been reported (16). TNF-cw immunoreactivity was measured using a double-ligand immunoassay (R & D Systems, Minneapolis, MN), and TNF-cu bioactivity was evaluated using a cytotoxicity assay (WEHI 164, clone 13 mouse fibrosarcoma cell line) (16). Statistical analysis. Summary statistics are expressed as means t SE. To decrease statistical variability and to eliminate the influence of small differences at baseline between the groups that could bias comparisons made at subsequent time points, hemodynamic responses are reported as percent changes from baseline values. Paired t tests were used to evaluate changes from baseline within each group. One-way analysis of variance and unpaired t tests were used to assess differences among groups at 3 and 5 h. Total TNF-a production was measured by calculating the area under the curve for TNF-cu (16). Correlations within each group were assessed using Pearson’s r

TO

INTRAVENOUS

ENDOTOXIN

tests. Fisher’s method of combining independent tests was used to combine the separate results for each group into a single overall P value (15). Two-tailed P values are reported for each analysis. P < 0.05 is regarded as significant. Hemodynamic calculations. Measured hemodynamic variables included heart rate (HR), mean arterial pressure (MAP), peak systolic pressure (PSP), central venous pressure (CVP), pulmonary arterial pressure (PAP), pulmonary arterial wedge pressure (PAWP), mean pulmonary arterial pressure (MPAP), and cardiac output (CO). The following variables were calculated using standard formulas: cardiac index (CI) = CO/body surface area (1 min-l ms2), systemic vascular resistance index (SVRI) = 80 X (MAP - CVP)/CI (dyn s crnm5 mb2), pulmonary vascular resistance index (PVRI) = 80 X (MPAP - PAWP)/CI (dyn s cms5 ms2), stroke volume index (SVI) = CI/HR (ml beat-l me2), left ventricular stroke work index (LVSWI) = SVI X MAP X 0.0136 (g m mv2), oxygen content of arterial blood (Ca& = (hemoglobin X percent oxygen saturation of arterial blood X 1.34) + (arterial PO, X 0.003) (vol%), oxygen content of venous blood (Cv,,) = (hemoglobin X percent oxygen saturation of venous blood X .1.34) + (venous PO, X 0.003) (vol%), oxygen delivery (DO,) = CI X CaoZ (ml min-l mB2), oxygen consumption (v02) = CI X CCaO2 - Cvo2) (ml min-l mm2), oxygen extraction ratio (0,ER) = \j0,/00,, and arteriovenous oxygen difference [(a - v)O,] = Cao, - CvoZ (~01%). The left ventricular end-diastolic volume index (EDVI) and end-systolic volume index (ESVI) were calculated from simultaneously obtained thermodilution catheter-derived data and radionuclide-gated blood pool ejection fraction (EF) according to the following formulas: EDVI = SVI/EF (ml/m2) and ESVI = EDVI-SVI (ml/m2) (23). l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

RESULTS

Systemic effects and hemodynamic responses. All 15 subjects who received endotoxin alone or endotoxin-pentoxifylline developed chills within 1 h of endotoxin administration that lasted -15-30 min and were followed by “flu-like” symptoms. Only one of the nine subjects who received endotoxin-ibuprofen developed chills and flu-like symptoms. The temperatures of the endotoxin alone and endotoxin-pentoxifylline groups rose significantly from baseline and peaked at 3 h (Fig. lA, Table 1). Defervescence and improvement in symptoms occurred in these groups after 3 h. The endotoxin-ibuprofen group remained afebrile during the study. Mean levels of ibuprofen were 12 t 4 pg/ml at baseline and 39 t 13 pglml at 3 h (16). Levels of pentoxifylline and its active first metabolite were 403 t 97 and 453 t 120 pg/ml at baseline and 380 t 77 and 365 t 87 pg/ml at 4 h (16). Baseline and 3-h measurements of central venous pressure, pulmonary arterial wedge pressure, and mean pulmonary arterial pressure were similar among the groups. After the fluid load (5 h), these measurements rose to a similar degree in all groups. The amounts of saline required to raise the pulmonary arterial wedge pressure did not differ among the groups (mean 1,695 t

Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (131.170.021.110) on October 24, 2018. Copyright © 1992 American Physiological Society. All rights reserved.

CARDIOVASCULAR

ho-

RESPONSE

0 Endotoxin

TIME

POST ENDOTOXIN hxJrs)

FIG. 1. Changes in core temperature (A) and percent changes in heart rate (B), cardiac index (C), and systemic vascular resistance index (D) after intravenous endotoxin administration to normal humans.

131 ml/m2). Pulmonary vascular resistance indexes were similar among the groups at 0, 3, and 5 h (Table 1). The heart rate of each group rose significantly between baseline and 3 h, remaining elevated at 5 h (Fig. lB, Table 1). Each group had a significant increase in cardiac index between baseline and 3 h (Fig. lC,‘Table 1). At 5 h after volume infusion, the cardiac indexes had increased by U-69% compared with baseline, but no significant increase had occurred in any group compared with 3-h values. Changes in cardiac index did not differ significantly among the groups at 3 or 5 h. Mean arterial pressures were similar among the groups at 0 h; at 3 h mean arterial pressure had fallen by 6-18% in the groups (P =

TO INTRAVENOUS

ENDOTOXIN

927

NS) and remained at these levels at 5 h. Peak systolic pressures were similar in all groups at baseline and 5 h. Systemic vascular resistance indexes fell significantly at 3 h and were lowest at 5 h but did not differ among the groups (Fig. lD, Table 1). Left ventricular performance. Left ventricular function was evaluated by three methods using data obtained from the thermodilution pulmonary artery catheter and simultaneous radionuclide cineangiogram (Table 2). Left ventricular stroke work indexes were similar in all groups at baseline and by 5 h had fallen by 4-l 1% (P = NS). Left ventricular ejection fraction was similar in all groups at baseline and did not change significantly by 5 h. Enddiastolic and end-systolic volume indexes were similar among the groups at baseline, and neither changed significantly at 5 h in any group. Ratios of ventricular volume, pressure, and stroke work were evaluated, and the changes in these ratios from baseline to 5 h were compared (Table 2). No differences were found among the groups in change from baseline to 5 h in PAWP/EDVI, LVSWUEDVI, or PSP/ESVI. Oxygen consumption and delivery. Oxygen consumption and delivery, oxygen extraction ratios, and arteriovenous oxygen differences were similar in the three groups at 0 h (Table 3). At 3 h, oxygen consumption and oxygen delivery rose significantly from baseline values, from 35 to 45% and from 51 to 61%, respectively. Fluid loading was associated with modest decreases from 3 to 5 h in oxygen consumption and oxygen delivery, but oxygen consumption remained 27-39% greater than baseline and oxygen delivery was 39-44% greater than at 0 h. At baseline oxygen extraction ratios ranged from 0.20 to 0.23 but at 3 h fell to 0.18-0.21 and at 5 h remained below baseline values (0.19-0.21). Arteriovenous oxygen difference values fell from baseline to 3 h and fell further after fluid loading at 5 h. No intergroup differences were found in any of these variables. Correlation of hemodynamics with TNF-cu levels. TNFcy, measured by immunoassay or bioassay, was not detected in any subject at baseline and was highest at 1.5 h in all three groups (immunoassay mean: endotoxin 812, endotoxin-pentoxifylline 780, endotoxin-ibuprofen 3,194 pg/ml; bioassay mean: endotoxin 454, endotoxinpentoxifylline 262, endotoxin-ibuprofen 1,273 pg/ml) (16). In subjects given endotoxin-ibuprofen, total TNF-cu immunoreactivity was 4.2-fold higher than in those given endotoxin alone and 4.5fold higher than in those given endotoxin-pentoxifylline. Total TNF-cu bioactivity was also higher in the endotoxin-ibuprofen subjects than in those given endotoxin alone or endotoxin-pentoxifylline (16). Correlations were made between peak and total immunoreactive and bioactive TNF-a(- levels and percent changes in cardiac index, mean arterial pressure, heart rate, and systemic vascular resistance index at 3 and 5 h. No significant correlation was found between any hemodynamic measurement and peak immunoreactive or bioactive TNF-cu levels. Total immunoreactive TNF-CY correlated significantly with the percent change in cardiac index at 3 h (r = 0.67 for endotoxin alone, r = 0.07for endotoxin-pentoxifylline, and r = 0.74 for endotoxinibuprofen; overall P < 0.05) and 5 h (r = 0.41 for endo-

Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (131.170.021.110) on October 24, 2018. Copyright © 1992 American Physiological Society. All rights reserved.

928

CARDIOVASCULAR

TABLE 1. Effects of ibuprofen

Temp, *C Oh 3h 5h CVP, mmHg Oh 3h 5h PAWP, mmHg Oh 3h 5h MPAP, mmHg Oh 3h 5h HR, beats/min Oh 3h 5h CI, 1. min-’ . mm2 Oh 3h 5h MAP, mmHg Oh 3h 5h SVRI, dyn s 9crnm5. me2 Oh 3h 5h PVRI, dyn 9se crnm5. mm2 Oh 3h 5h

RESPONSE

and pentoxifylline

-49 68

39.1kO.2

5.5 4.3

38.3t0.2 6.420.9 3.2kl.O 9.4t0.9

INTRAVENOUS

on temperature

36.7t0.2 6.5 4.6

TO

ENDOTOXIN

and hemodynamics

36.620.1 38.6t0.2 38.2k0.2

1.4 1.5

6k0.9 -28 108

4t0.6 llkO.9

-25 57

-9

llk1.2 -36 44

8t2.2 1621.0

-36 63

lltl.O 720.9 16t0.7

6 65

14t0.7 15t1.7 23tl.O

-4 69

13t0.9 24k1.2

3 65

52 50

73k2.9 1 lOt4.0 108~1.6

53 61

61k2.4 92k3.5 97t2.7

40 54

15t1.4

3.97kO.3 52 57

6.41k0.3 6.45t0.3

70

51 69

-18 -13

96k2.7 79t4.5 84k5.5

-6 -9

-46 -47

1,7652138 940t49 931k92

-31 -42

3.59kO.2 5.39t0.4 5.99kO.4

after intravenous

36.&O. 1 37.3t0.2 37.4kO.2 6kO.6 4t0.7 8t0.7 10~1.0 8kO.8 15kO.6

endotoxin

E, P, 1 E, p, 1

E,P>I E,P>I

E, P P

None None

E, P E, R 1

None None

None E, p, 1

None None

E, R 1 E, P, 1

None None

E, p, 1 E, p, 1

None None

15t0.5

15kl.l 24k2.4 6722.7

9321.7 103k5.5

3.49t0.2 52 54

5.21tO.l 5.29t0.3

9Ok2.0

87k2.2 82k2.0 80t2.7

-11

82t3.2 80t2.9

E, p, 1 p, 1

None None

1,980?227 1,356*169 1,102+112

-39 -39

1,967+151 1,183+96 1,180*125

E, P, 1 E, p, 1

None None

None None

None None

-9

l

74 115

65k19 95t24 93k4

39 30

94k16 106t17 103t13

4 27

119k26 107k18 15Ok39

CVP, central venous pressure; PAWP, pulmonary arterial wedge pressure; MPAP, mean pulmonary arterial pressure; HR, heart rate; CI, cardiac index; M AP, mean arterial pressure; SVRI, systemic vascular resistance index; PVRI, pulmonary vascular resistance index; E, endotoxin; P, pentoxifylline; I, ibuprofen.

toxin alone, r = 0.35 for endotoxin-pentoxifylline, and r = 0.77 for endotoxin-ibuprofen; overall P < 0.05) as well as the percent change in mean arterial pressure at 5 h( r = 0.75 for endotoxin alone, r = 0.26 for endotoxinpentoxifylli ne, and r= 0.70 for endotoxi n -ibuprofen; overall P < 0.05). N .&ant correlation S were noted between total bioactive-TNF-cu and percent changes in cardiac index, mean arterial pressure, heart rate, or systemic vascular resistance index. DISCUSSION

This study demonstrates that prior treatment with oral ibuprofen or oral pentoxifylline does not alter the hyperdynamic cardiovascular response of normal humans to intravenous endotoxin. The hemodynamic response remains qualitatively identical to that seen in human sepsis. Although ibuprofen prevented the fever and constitutional symptoms associated with endotoxin administration, there was no significant diminution of the cardiovascular response. In this human model of endotoxin-induced fever, cyclooxygenase products do not ap-

pear to contribute significantly to the increased cardiac index, increased heart rate, and decreased systemic vascular resistance. Animal models of endotoxemia have shown attenuation of the hyperdynamic cardiovascular state with administration of cyclooxygenase inhibitors (17). These beneficial responses occur if ibuprofen is given before or after the administration of endotoxin or live E. coli infusion or after the creation of a septic focus (17). The improvements in hemodynamic parameters depended in part on the animal model (i.e., cardiac output and mean arterial pressure rose in hypodynamic models and decreased in hyperdynamic models) and on the dosages of ibuprofen administered (5-25 mg/kg) (9, 13, 17). In the current study, mean levels of ibuprofen at baseline were 12 -+ 4 pg/ml (16). Although this dose is associated in clinical studies with cyclooxygenase inhibition, higher sustained levels may be required to provide an anti-inflammatory effect that may ameliorate alterations in systemic hemodynamics after endotoxin administration (17, 21). A clinical trial of ibuprofen (3 doses of 800 mg administered as retention enema) in patients with sepsis

Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (131.170.021.110) on October 24, 2018. Copyright © 1992 American Physiological Society. All rights reserved.

CARDIOVASCULAR

RESPONSE

2. Effects of ibuprofen and pentoxifylline administration to normal humans

LVSWI, go m m2 Oh 3h 5h LVEF, % Oh 5h EDVI, ml/m2 Oh 5h ESVI, ml/m2 Oh 5h PAWP/EDVI, mmHg ml-’ mm2 Oh 5h A O-5 h LVSWI/EDVI, g. me ml-’ Oh 5h A O-5 h PSP/ESVI, mmHg ml-’ . rnw2 Oh 5h A O-5 h

(n = 5)

Pentoxifylline

%A

Mean k SE

%A

-11 -5

72k5 63t5 66k4

-4 -4

function

endotoxin Intergroup Comparison at 3 and 5 h (P < 0.05)

64k4 6324 57t5

None None

None None

48kl 49+3

None

None

-1

109k5 107t5

None

None

-3

57k3 55*4

None

None

E, R 1

None None

None

None None

None

None None

Ibuprofen

k SE

after intravenous Intragroup Comparison vs. 0 h (P < 0.05)

(n = 10) Mean

929

ENDOTOXIN

on measures of ventricular

TABLE

Endotoxin

TO INTRAVENOUS

(n = 9)

%A

Mean

+ SE

l

l

7Ok3 67k6 67t3

49+2 4

50t2

7

112k6 12os3

-1

4

57t4 6027

-5

0 -11

47k2 49k2

6

2

128M 126t7 68_t6 51t4

l

0.10_t0.01 0.14t0.01 0.04t0.01

35

0.09~0.01 0.13~0.01 0.04t0.01

77

0.57kO.04 0.53t0.03 -0.04t0.04

-9

2.1kO.2 1.7t0.4 -0.39t0.42

-2

51

-12

0.64kO.03 0.56kO.04 -0.08_tO.O5

-10

2.5kO.l 2.3k0.2 -0.24tO. 16

-13

0.09~0.01 0.14t0.01 0.05kO.02

0.59~0.02 0.54kO.04 -0.05t0.03

l

3

2.5k0.1 2.4t0.2 -0.07t0.20

LVSWI, left ventricular stroke work index; LVEF, left ventricular ejection fraction; EDVI, end-diastolic volume index; PAWP, pulmonary arterial wedge pressure; PSP, peak systolic pressure.

syndrome/septic shock has shown improvement in temperature, heart rate, and blood pressure (2). Previous studies of ibuprofen administration before endotoxin administration to normal volunteers showed attenuation of the rises in temperature, heart rate, and levels of stress hormones (corticotropin, cortisol, growth hormone, and catecholamines) (20). These data suggest that these aspects of the hormonal stress response are not critical to the hemodynamic alterations observed in the current study. 3. Effects of ibuprofen and pentoxifylline administration to normal humans Endotoxin %A

VO,, ml min-’ mm2 Oh 3h 5h IjO,, ml. min-’ . me2 Oh 3h l

5h 02ER Oh 3h 5h (a-v)O,, Oh 3h 5h

(n = 5) Mean

k SE

Pentoxifylline improves hemodynamic alterations and ameliorates organ damage in animal models of infection (24). Pentoxifylline has a variety of hemorrheologic, anti-inflammatory, and hemodynamic effects that may be important with regard to its use in clinical septic shock (24, 27). When pentoxifylline was given in large doses intravenously before Salmonella abortus equi endotoxin administration to normal humans, the resultant TNF-cu levels were lower than in controls given endotoxin alone (29). No improvements in symptoms or heart rate were

on oxygen transport

TABLE

Pentoxifylline %A

volume index; ESVI, end-systolic

variables after intravenous

(n = 10) Mean + SE

Ibuprofen %A

(n = 9) Mean

k SE

endotoxin

Intragroup Comparison vs. 0 h (P < 0.05)

Intergroup Comparison at 3 and 5 h (P < 0.05)

l

45 39

14124 205t6 197klO

45 29

147t9 210t14 185klO

35 27

138t5 185+19 174k12

E, p, 1 E, p, 1

None None

61 44

722k40 1,160&46 1,034+39

51 42

670t42 1 ,OOOa67 94Ok76

53 39

616k35 925+43 849+69

E, p, 1 E, R 1

None None

-12 -7

0.23t0.01 0.20t0.01 0.21t0.01

E None

None None

-11 -16

4.0t0.24 3.6t0.34 3.3k0.18

None p, 1

None None

-10 -3

0.20~0.01 0.18t0.01 0.19t0.01

-1 -7

-11 -15

3.7t0.03 3.220.07 3.1t0.22

-2 -23

0.22+0.01 0.21~0.01 0.20~0.01

~01%

Vo,, 0, consumption;

00,) 0, delivery; O,ER, 0, extraction

4.2&O. 18 3.9to. 19 3.1t0.13

ratio; (a-v)O,, arteriovenous

O2 difference.

Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (131.170.021.110) on October 24, 2018. Copyright © 1992 American Physiological Society. All rights reserved.

930

CARDIOVASCULAR

RESPONSE

noted in this study, and other hemodynamic parameters were not evaluated (29). Levels of pentoxifylline were limited in the current study by the maximum achievable blood levels with oral dosing (16), although the regimen used has been shown to alter some of the inflammatory responses associated with clinical illness (3). Pentoxifylline has been noted to have a mild inotropic effect in human and animal studies, yet this effect was not apparent in the current study (12, 27). Thus, with use of oral pentoxifylline, no significant effect was found on systemic hemodynamics or ventricular function changes after endotoxin administration. Although several studies have described various aspects of the human cardiovascular response to fever, data are limited concerning its effects on ventricular performance and oxygen transport variables (10). Humans given a bacterial pyrogen intravenously developed increased cardiac output, tachycardia, and decreased total peripheral resistance (5). This response was unchanged if fever was prevented by prior administration of amidopyrine (5), a pyrazolon derivative with anti-inflammatory properties similar to salicylates (28). Clinical studies of cardiac performance during febrile states suggested that fever alone may have adverse effects on ventricular function (11). In the current study, however, several observations confirm the presence of abnormalities in left ventricular function with and without fever after intravenous endotoxin (23). First, cardiac index failed to rise after volume infusion. Second, under conditions of increased pulmona .rY wedge pressure an .d heart rate and decreased system .lC vascula r resistance, the left ventricular ejection fraction failed to increase significantly. In contrast, normal subjects under similar filling conditions show a rise in the left ventricular ejection fraction (23). Third, with use of the ratio of left ventricular stroke work index to left ventricular end-diastolic volume index and the ratio of peak systolic pressure to left ventricular endsystolic volume index, which is relatively independent of filling conditions, no differences were demonstrated between the subjects given endotoxin alone and those pretreated with ibuprofen or pentoxifylline. Thus, alterations in ventricular function did not result from increases in core temperatures, because they occurred even in the afebrile endotoxin-ibuprofen subjects. Oxygen delivery and consumption increased significantly after the administration of endotoxin, and no significant changes occurred in the oxygen extraction ratio before or after fluid loading. These data suggest that, during the early cardiov ,ascular response of n ormal humans to endotoxin, the matchi ng of oxygen consumption and delivery is not impaired (22). The absence of fever did not significantly attenuate the increase in oxygen consumption after endotoxin administration. The changes in systemic hemodynamics, ventricular performance, and oxygen transport that follow intravenous endotoxin administration to normal humans are not dependent on fever alone and result from endotoxin itself or endogenous inflammatory mediators. TNF-cu is a major mediator of many of the pathophysiological changes that occur during septic shock (25). The relation of TNF-cu release to subsequent changes in cardiovascular function in the current study is complex. An

TO

INTRAVENOUS

ENDOTOXIN

association was found between total immunoreactive TNF-cu levels and changes in cardiac index at 3 and 5 h and mean arterial pressure at 5 h. However, mean arterial pressure did not significantly decrease from baseline in any of the three groups. Although statistically significant, the intergroup variability makes the importance of this correlation indeterminate. No association was found when bioactive TNF-cu levels were compared with cardiovascular responses. These data suggest that TNF-A! levels and cardiovascular responses may be independent variables. The levels of TNF-cu detected in our subjects were comparable to or higher than those reported in a series of patients with life-threatening infections measured with similar assays (26). Furthermore, augmented levels of TNF-cu were observed in subjects given endotoxin-ibuprofen compared with those given endotoxin alone or endotoxin-pentoxifylline, yet the endotoxinibuprofen subjects did not have more pronounced hemodynamic responses (16). Cyclooxygenase products serve as part of the inhibitory feedback loop that limits monocyte-macrophage TNF-a, production (16). Animal studies have demonstrated that cyclooxygenase inhibitors can improve the hypotension associated with endotoxin, Gram-negative bacteria, interleukin-1, or TNF-cu administration (8,144 17,19). Cyclooxygenase inhibition, in addition to having an effect on cytokine-producing cells, may have separate protective effects on peripheral vascular tone during exposure to high circulating levels of TNF-cu. Intravenous endotoxin administration to humans elicits cardiovascular responses that are relevant to the pathogenesis of septic shock, including a hyperdynamic cardiovascular profile with reversible left ventricular dysfunction (23). Prevention of endotoxin-induced symptoms and fever by the cyclooxygenase inhibitor ibuprofen did not attenuate this cardiovascular response. Similarly, the phosphodiesterase inhibitor pentoxifylline did not alter these responses. Thus the changes in systemic hemodynamics and oxygen transport variables after intravenous endotoxin administration to normal humans occur in the absence of fever and are not significantly ameliorated by oral cyclooxygenase or phosphodiesterase inhibition. The authors thank Dr. Robert A. Wesley for statistical analysis, Debra Reda for caring for the normal volunteers, and the Critical Care nurses and therapists of the Clinical Center. This study was presented in part at the annual meetings of the American Federation for Clinical Research, Washington, DC, May 1990 and the American Thoracic Society/American Lung Association, Anaheim, CA, May 1991 and was published, in part, in abstract form (Clin. Res. 38: 453A, 1990 and Am. Rev. Respir. Dis. 143: A204, 1991). Address for reprint requests: G. D. Martich, National Institutes of Health, Bldg. 10, Rm. 7-D-43, Bethesda, MD 20892. Received

18 September

1991;

accepted

in final

form

27 March

1992.

REFERENCES

BALK, R. A.,R.F. JACOBS, A.F. TRYKA, R. C. WALLS, AND R.C. BONE. Low dose ibuprofen reverses the hemodynamic alterations of canine endotoxin shock. Crit. Care Med. 16: 1128-1131, 1988. BERNARD,G. R.,H.D. REINES,~. A. METZ, P.V. HALUSHKA, B.B. SWINDELL, S. B. HIGGINS, P. E. WRIGHT, AND F. L. WATTS. Effects of a short course of ibuprofen in patients with severe sepsis (Abstract). Am. Rev. Respir. Dis. 137: 138, 1988. BIANCO, J.,J. ALMGREN, D.L. KERN, B. BALLARD, K. ROARK, F.

Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (131.170.021.110) on October 24, 2018. Copyright © 1992 American Physiological Society. All rights reserved.

CARDIOVASCULAR

RESPONSE

ANDREWS, J. NEMUNAITIS, T. SHIELDS, AND J. W. SINGER. Evidence that oral pentoxifylline reverses acute renal dysfunction in bone marrow transplant recipients receiving amphotericin B and cyclosporine. Transplantation 51: 925-927, 1991. 4. BONE, R. C., C. J. FISHER, JR., T. P. CLEMMER, G. J. SLOTMAN, C. A. METZ, R. A. BALK, AND THE METHYLPREDNISOLONE SEVERE SEPSIS STUDY GROUP. A controlled clinical trial of high-dose methylprednisolone in the treatment of severe sepsis and septic shock. N. Engl. J. Med. 317: 653-658, 1987. 5. BRADLEY, S. E., H. CHASIS, W. GOLDRING,

6.

7.

8.

9.

10.

11. 12.

13.

AND H. W. SMITH. Hemodynamic alterations in normotensive and hypertensive subjects during the pyrogenic reaction. J. CZin. Inuest. 24: 749-758, 1945. DANNER, R. L., R. J. ELIN, J. M. HOSSEINI, R. A. WESLEY, J. M. REILLY, AND J. E. PARRILLO. Endotoxemia in human septic shock. Chest 99: 169-175, 1991. DEMLING, R. H., C. LALONDE, AND M. E. P. GOAD. Effect of ibuprofen on the pulmonary and systemic response to repeated doses of endotoxin. Surgery 105: 421-429, 1988. EVANS, D. A., D. 0. JACOBS, A. REVHAUG, AND D. W. WILMORE. The effects of tumor necrosis factor and their selective inhibition by ibuprofen. Ann. Surg. 209: 312-321, 1989. FINK, M. P., T. J. MACVITTIE, AND L. C. CASEY. Inhibition of prostaglandin synthesis restores normal hemodynamics in canine hyperdynamic sepsis. Ann. Surg. 200: 619-626, 1984. GREISMAN, S. E. Cardiovascular alterations during fever. In: Fever: Basic Mechanisms and Management, edited by P. Mackowiak. New York: Raven, 1991, p. 143-165. HAUPT, M. T., AND E. C. RACKOW. Adverse effects of febrile state on cardiac performance. Am. Heart J. 105: 763-768, 1983. HEIDRICH, H., M. PAEPRER, D. BARCKOW, AND M. SCHARTL. The effect of pentoxifylline on central and peripheral haemodynamics: an experimental clinical study. 2. KardioL. 65: 385-391, 1975. JACOBS, E. R., M. E. SOULSBY, R. C. BONE, F. J. WILSON, JR., AND F. C. HILLER. Ibuprofen in canine endotoxin shock. J. Clin. Inuest.

70: 536-541, 14. KETTELHUT,

1982.

I. C., W. FIERS, AND A. L. GOLDBERG. The toxic effects of tumor necrosis factor in vivo and their prevention by cyclooxygenase inhibitors. Proc. Natl. Acad. Sci. USA 84: 4273-4277,

1987. 15. LITTELL,

R., AND J. L. FOLKS. Asymptotic optimality of Fisher’s method of combining independent tests. J. Am. Stat. Assoc. 66:

16.

TO INTRAVENOUS

C. A., AND J. N. SHEAGREN. Ibuprofen in animal models of septic shock. J. Crit. Care 5: 206-212, 1990. 18. NOEL, P., S. NELSON, R. BOKULIC, G. BAGBY, H. LIPPTON, G. LIPSCOMB, AND W. SUMMER. Pentoxifylline inhibits lipopolysaccharide-induced serum tumor necrosis factor and mortality. Life Sci. 17.

METZ,

47: 1023-1029, 1990. 19. OKUSAWA, S., J. A. GELFAND, T. IKEJIMA, R. J. CONNOLLY, C. A. DINARELLO. Interleukin-1 induces a shock-like state bits. J. Clin. Invest. 81: 1162-1172, 1988. 20. REVHAUG, A., H. R. MICHIE, J. M. MANSON, J. M. WATTERS, DINARELLO, S. M. WOLFF, AND D. W. WILMORE. Inhibition

AND

in rab-

C. A. of cyresponse to endotoxin in

clo-oxygenase attenuates the metabolic humans. Arch. Surg. 123: 162-170, 1988. 21. RINALDO, J., AND J. H. DAUBER. Effect of methylprednisolone and of ibuprofen, a nonsteroidal anti-inflammatory agent, on bronchoalveolar inflammation following endotoxemia. Circ. Shock 16: 195-203,1985. 22. SCHUMAKER,

sumption 23.

24.

P. T., AND R. W. SAMSEL. Oxygen supply and conin the adult respiratory distress syndrome. Clin. Chest

Med. 11: 715-722, 1990. SUFFREDINI, A. F., R. E. FROMM, J. A. KOVACS, R. A. WESLEY, AND

M. M. PARKER, M. BRENNER, J. E. PARRILLO. The cardiovascular response of normal humans to the administration of endotoxin. N. Engl. J. Med. 321: 280-287, TIGHE, D., R. Moss, J. HYND, HEATH, AND E. D. BENNETT.

1989.

S. BOGHOSSIAN, N. AL-SAADY, M. F. Pretreatment with pentoxifylline improves the hemodynamic and histologic changes and decreases neutrophil adhesiveness in a pig fecal peritonitis model. Crit. Care

Med. 18: 184-189, 1990. 25. TRACEY, K. J., Y. FONG, G. C. KUO, S. F. LOWRY,

D. G. HESSE, K. R. MANOGUE, A. T. LEE, AND A. CERAMI. Anti-cachectin/TNF antibodies prevent septic shock during lethal bacterae-

26.

27.

28.

802-806,1971. MARTICH,

G. D., R. L, DANNER, M. CESKA, AND A. F. SUFFREDINI. Detection of interleukin-8 and tumor necrosis factor in normal humans after intravenous endotoxin administration: the effect of anti-inflammatorv agents. J. Em. Med. 173: 1021-1024. 1991.

931

ENDOTOXIN

29.

monoclonal mia. Nature Lond. 330: 662-664, 1987. WAAGE, A., P. BRANDTZAEG, A. HALSTENSEN, P. KIERULF, AND T. ESPEVIK. The complex pattern of cytokines in serum from patients with meningococcal septic shock. J. Exp. Med. 169: 333-338,1989. WARD, A., AND S. P. CLISSOLD. Pentoxifylline: a review of its pharmacodynamic and pharmacokinetic properties, and its therapeutic efficacy. Drugs 34: 50-97, 1987. WOODBURY, D. M., AND E. FINGL. Pyrazolon derivatives. In: The Pharmacological Basis of Therapeutics (5th ed.), edited by L. S. Goodman and A. Gilman. New York: Macmillan, 1975, p. 347-348. ZABEL, P., M. M. SCH~NHARTING, D. T. WOLTER, AND U. F. SCHADE. Oxpentifylline in endotoxaemia. Lancet 2: 1474-1477, 1989.

Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (131.170.021.110) on October 24, 2018. Copyright © 1992 American Physiological Society. All rights reserved.