- Email: [email protected]
Vasoactive Hormones in the Renal Response to Systemic Sepsis
Vasoactive Hormones in the Renal Response to Systemic Sepsis Allan D. Cumming, MS, MRCP(UK), Albert A. Driedger, MD, FRCP(C), John W. D. McDonald, MD, FRCP(C), Robert M. Lindsay, MD, FRCP(E), FRCP(C), Kim Solez, MD, and Adam L. Linton, MS, FRCP(E), FRCP(C) • The pathophysiology of renal dysfunction in generalized sepsis remains unknown. In this study, 24 hours after surgical induction of peritonitis in 20 volume-loaded sheep, three patterns of renal function were seen. In group 1 (n = 8), glomerular filtration rate (GFR) decreased by 70%, urine volume by 85%, absolute sodium excretion by 95%, and fractional sodium excretion by 83%. Group 2 (n = 4) exhibited similar sodium retention but GFR did not fall. Group 3 (n = 8) showed no change in GFR or urine volume and only minimally reduced sodium excretion. Mean arterial pressure fell 17% in group 1 only; central venous pressure, pulmonary capillary wedge pressure, and plasma volume were maintained at or above presepsis values in all groups. Cardiac index was either increased or unchanged, and renal plasma flow was maintained in all groups; there was thus no hemodynamic evidence to suggest volume contraction. Histologic examination showed only minor changes with no consistent pattern. Renal functional changes correlated with other manifestations of severe sepsis-GFR and sodium retention correlated significantly with increased cardiac index, decreased systemic vascular resistance, pulmonary arterial hypertension, leukopenia, hypoproteinemia, and hypoglycemia. All of these changes were most marked in group 1. In groups 1 and 2, plasma renin activity (PRA) increased and urinary kallikrein excretion decreased. PRA correlated inversely with GFR, urine volume, and sodium excretion; urinary kallikrein excretion correlated positively with urine volume and sodium excretion. Urinary excretion of 6-keto-PGF1a was increased in groups 1 and 2 and correlated inversely with mean arterial pressure in group 1 animals. During sepsis, urinary thromboxane 8 2 excretion continued at presepsis values in all groups. The results suggest that unusual reciprocal changes in activity of the renin-angiotensin and renal kallikrein-kinin systems may play a role in the renal response to sepsis. PGI 2 synthesis is increased and may affect systemic hemodynamics and renal function; the role of thromboxane A2 in this context is unknown. © 1988 by the National Kidney Foundation, Inc. INDEX WORDS: Sepsis; acute renal failure; vasoactive hormones; prostaglandins; kinins; venin.
T
HE LINK between sepsis and renal functional changes including classical acute renal failure (ARF) , is clinically important and poorly understood. 1.2 While investigating the renal response to sepsis,3 we have recently established a model of severe systemic sepsis in sheep. Circulating volume is maintained by aggressive fluid administration, and hypotension is avoided. At an early stage, we recognized various patterns of renal response, probably reflecting minor differences in surgical technique and constitutional variation between animals. 4 More recently, we have largely eliminated this variability; we did, however, undertake a detailed analysis of the different response patterns in our initial group of 20 sheep, hoping to gain information regarding the mechanism of renal functional changes and, particularly, the role of vasoactive hormones acting on the kidney. We found a correlation between severity of sepsis, judged by hemodynamic, hematologic, and biochemical disturbance, and the type of renal response. In this study we have defined this relationship and attempted to correlate different patterns of renal functional change with alterations in me-
diator systems that might influence kidney function in sepsis. It has been suggested that vasoactive substances may be important in the pathogenesis of some types of ARP and that vasodilator prostaglandins (PGs) protect renal function against a variety of insults. 6 In systemic sepsis, increased plasma renin activity (PRA) and increased plasma concentrations of 6-keto-PGFla and of thromboxane B2 (TXB 2) have been documented. 3,7 The possible influence of these compounds on renal function in sepsis has not been defined. We have, therefore, measured PRA, urinary 6-keto-PGFla and TXB 2 From the Department of Medicine, The University of Western Ontario, London, Ontario, Canada; and the Department of Pathology, Johns Hopkins University School of Medicine, Baltimore, Supported in part by NATO Grant No, RG.85/0043, The present address of Dr Solez is University Department of Medicine, Royal Infirmary, Edinburgh, Scotland. Address reprint requests to Adam L. Linton, MB, Department of Medicine, Victoria Hospital, P.O. Box 5375, London, Ontario, Canada N6A 4G5. © 1988 by the National Kidney Foundation, Inc. 0272-6386/88/1101 -0005$3.00/0
American Journal of Kidney Diseases, Vol XI, No 1 (January), 1988: pp 23-32
23
24
CUMMING ET AL
in septic sheep. 8.9 We also measured urinary kallikrein excretion, since the renal kallikrein-kinin system (KKS) interacts with the renin-angiotensin system (RAS) and renal PGs. JO-12 We have tested correlations between these variables and other aspects of the septic state, particularly changes in systemic hemodynamics and renal function. MATERIALS AND METHODS The technique for surgical induction of peritonitis is modified from that of Wichterman et al 13 and has been previously described in detail. 3 In brief, 20 sheep, aged 9 to 18 months and weighing 35 to 50 kg, underwent general anesthesia and cannulation of the aorta via the common carotid artery, and of the pulmonary artery via the external jugular vein with a triplelumen Swan-Ganz catheter. The bladder was catheterized per urethra. After recovery from anesthesia, the animals were housed in a metabolic cage and volume-loaded with 8 L of Ringer's lactate intravenously (IV) over 24 hours. After control hemodynamic measurements and blood and urine sampling, animals were anesthetized and peritonitis was induced by devascularization and proximal ligation of 6 to 8 cm of distal cecum, combined with partial omentectomy and a 2-cm puncture in the cecal tip. The abdomen was closed and the animal allowed to recover from anesthesia. Postoperatively, all sheep received 50 mg Demerol (Merperidine, Abbott Labs, Toronto) and 25 mg Acepromazine IV (Atra-Vet, Ayerst Laboratories, Montreal) and were continued thereafter on IV Demerol, 50 mg/6 to 8 h. IV infusion of Ringer's lactate, 8 Ll24 h, continued for the duration of the experiment. Follow-up hemodynamic measurements and blood and urine collections were made 24 and, where possible, 48 hours after induction of sepsis. Animals were continuously under the care of qualified animal health technicians during the study; during sepsis, the sheep showed lethargy, anorexia, pyrexia, and increased respiratory rate. Any animal observed to be restless, agitated, or showing other evidence of discomfort or distress despite Demerol was killed immediately by IV injection of pentobarbital (Euthanol, Euthanol Forte, MTC Pharmaceuticals, Mississauga, Canada). Mortality was, therefore, not used as an endpoint in the study. Animals surviving through 48 hours were killed as previously described; an open renal biopsy was taken from all sheep at the time of death. The study was approved by the institutional ethical review process of The University of Western Ontario, and the care of the animals conformed with the conditions set down in the "Guide to the Care and Use of Experimental Animals of the Canadian Council on Animal Care." Hemodynamic parameters measured were mean arterial pressure (MAP), mean pulmonary artery pressure (PAP), and pulmonary capillary wedge pressure (PCWP) (mm Hg). Central venous pressure (CVP) was measured in centimeters of H 20 and thermodilution cardiac output in liters per minute, as previously described. 3 Cardiac index (CI) and systemic vascular resistance index (SVRI) were derived by standard formulae. Blood samples were taken from the aortic cannula; WBC was measured with a Coulter cell counter (Model S., Coulter Electronics, Burglinton, Ontario), and serum creatinine (SCr), total protein, albumin, glucose, sodium, and potassium measured by standard autoanalyzer methods. PRA was measured by ra-
dioimmunoassay14 and results expressed in nanograms per milliliter per hour of generated angiotensin 1. Urine was collected for four hours on -the morning of each study day and the volume recorded; creatinine, sodium, potassium, and osmolarity were measured by standard methods. Urinary kallikrein was assayed by the method of Amundsen et al,15 measuring the ability of urine to hydrolyze the synthetic tripeptide chromogenic substrate, S2266 (Kabi Diagnostica, Stockholm). Urinary 6-ketoPGFla and TXB2 were measured by radioimmunoassay.16.17 Plasma volume was measured as the immediate distribution space of 1-125 albumin l8 ; effective renal plasma flow was measured as the clearance of orthoiodinated hippuran labeled with 1-131 19 ; clearance of Tc-99m-DTPA was measured as an additional indicator of glomerular filtration rate (GFR).20 In each case, the isotope-labeled compound was injected via the venous catheter, and blood samples were taken from the aortic cannula. Kidney tissue was examined by light microscopy using standard methods. 21 Only results at baseline and after 24 hours of sepsis were analyzed since only one of eight sheep in the most severely septic group survived through 48 hours. Statistical analysis was performed using the Data Analysis Interactive System (Rainbow Computing, Inc, Northridge, CAl. Comparison between groups at baseline and 24 hours was by the Mann-Whitney U test; comparisons, within groups, between baseline and 24 hour data were by the Wilcoxon Rank Sum test. Studies of correlation were made for 24-hour postsepsis data from the whole animal population by Spearman Rank coefficients. Values of P less than .05 were held to be significant; results are expressed as mean ± SEM.
RESULTS
For analysis, sheep were divided into three groups on the basis of their renal function after induction of generalized sepsis. Sheep showing a significant decline in GFR after 24 hours of sepsis, as evidenced by a rise in Ser and decreased clearance of endogenous creatinine and DTPA, were allocated to group 1 (n = 8) (Fig 1). These animals also exhibited major decrements in urine volume, sodium excretion rate, and fractional sodium excretion (Fig 2). Sheep showing oliguria and sodium retention without reduction in GFR were allocated to group 2 (n = 4) (Figs 1 and 2). Sheep that exhibited no change in glomerular or tubular function apart from a modest decrease in sodium excretion were allocated to group 3 (n = 8) (Figs 1 and 2). Renal plasma flow decreased during sepsis in group 1 only, but this did not achieve significance (Fig 1). Urinary osmolarity remained around 400 mosm/kg after induction of sepsis, except in group 2 where all four animals concentrated urine to > 800 mosm/kg. Indicators of Circulating VOlume Status
Comparison of hemodynamic measurements at baseline with those 24 hours after induction of sep-
25
RENAL RESPONSE TO SEPSIS 150
~
2000
~
1600
3
Z
!
.p=
>
~
,!!
1i
" U
800
~
E
"
Ii
(/)
0
0
'2 200 E
!
.";;; ~ . :&
150
U
0
u
~Il
~~
c
~i ..
100
rZ_ ~
c
..
.;
1i
~
.~ i
Il
"
"uc
OJ
.S... iu
...
~
(/)o & ~
13 100 -< II.
15
~-
..
.2
50
U
ol:
0
1200
!l
t1.
J ]
800
400 200
o Baaellne
24 hr.
Group 1
Ba.ellne
24 hr.
Group 2
Baaellne
24 hr.
Group 3
Fig 1. Serum creatinine, creatinine clearance, DTPA clearance, and renal plasma flow at baseline and after 24 hours of sepsis in 3 groups of septic sheep (group 1, n = 8; group 2, n = 4; group 3, n = 8; mean ± SEM).
sis revealed that CVp, PCWp, and plasma volume were maintained in all groups (Fig 3). There was a small (17 %) decline in MAP in group 1 animals that was statistically significant, but no change in MAP in groups 2 and 3. Indicators of Severity of Sepsis
In all sheep, a polymicrobial peritonitis and bacteremia was confirmed at 24 hours by blood culture. Organisms grown included Serratia marcescens, Enterobacter cloaceae, Pseudomonas, Bacteroides sp, and Escherichia coli. Autopsyexamination revealed distended loops of bowel and an inflammatory mass in the right lower quadrant.
B..ellne
24 hr.
Group 1
Ba.ellne
24 hr.
Group 2
Base....
24 hr.
Group 3
Fig 2. Urine volume, sodium excretion, fractional sodium excretion, and urinary osmolarity in septic sheep (as in Fig 1).
All animals developed pyrexia and tachypnea. WBC counts fell in all groups, but 24-hour values were significantly lower in group 1 than in group 3 (Fig 4). Similarly, while serum total protein, albumin, and glucose levels fell in all three groups, these changes were greatest in group 1 (Fig 5). Significant hemodynamic changes occurred only in group 1 with increased CI, reduced SVRI, and increased PAP (Fig 4). Serum potassium fell in all groups (Fig 5). Hormonal Changes
Both PRA and urinary excretion of 6-ketoPGFla increased significantly during sepsis in groups 1 and 2, but not in group 3. Urinary kallikrein excretion decreased significantly in groups 1
26
CUMMING ET AL
.. .. 4:0
10
!;
• p
8
....
N
6
.5 ~ u c:
i E >,3
4
OJ OJ
WI
6 :J:
~
.. .. .. ct
~
'ON
.!! '0 ~
'E
, .J
~~
2
C
()
0 15
!; OJ OJ
. a
'0
10
~
a
:J:
.. E E
~
..
ci.~
5
()
.il:l
0
0..
!..
..". ~
.t"
2000
~ E
E .a 1500
..
.~.5
0
>
.. a:"
E
'"
,.,~
~ .. :J:
15
c
.so
1000
5
~
500
12 10
. £
150
8
!:>
6
~ 100
4
a
Ii :J: 'i E
t: E
..
00(
c:
:!
~
2
50
o Baseline Baseline 24 hr. 24 hr. Group 1 Group :!
o Baseline
24 hr. Group 1
Baseline
24 hr. Group 2
Baseline
24 hr. Group 3
Fig 3. Central venous pressure, pulmonary capillary wedge pressure, plasma volume and mean arterial pressure in septic sheep (as in Fig 1).
and 2, but not in group 3, while urinary TXB 2 excretion was unchanged during sepsis in all three groups (Fig 6). Renal Histology Various histologic features were observed in kidney biopsies taken at time of animal death. Kidney tissue from three animals was unsuitable for examination for technical reasons. In the other 17 sheep, the changes observed included hyperplasia of the juxtaglomerular apparatus (JGA) (six animals), mild tubular dilatation (eight animals), very
Baseline 24 hr. Group 3
Fig 4. Cardiac index, systemic vascular resistance index, pulmonary artery pressure and white cell count in septic sheep (as in Fig 1).
occasional foci of tubular regeneration (four animals), patchy interstitial infiltrate of mononuclear cells (13 animals), and small foci of medullary calcification (12 animals). With the exception of JGA hyperplasia,4 the pathologic changes were minor and evenly distributed between groups of animals, irrespective of the pattems of renal functional response. Studies of Correlation The three measurements reflecting GFR, ie, serum creatinine (SCr)' creatinine clearance (CCr), and DTPA clearance (C DTPA), correlated well during sepsis (SCr v CCr: r = _.. 84, P < .001;
27
RENAL RESPONSE TO SEPSIS • p
10
.~ 50
e
Q,
40
1i 30
o
.... 20
e
t
VI
10
0
.~
.5 ;;:
20
~OJ
VI
-S ~
6 5
4
Baaellne Baaellne 24 hr. 24 hr. Group 1 Group 2
Baseline 24 hr. Group 3
Fig 5. Serum total protein, albumin, glucose, and potassium in septic sheep (as in Fig 1).
CCT v C DTPA : r = - .77, P < .001; SCT v CDTPA : r = .75, P < .001). Other correlations are shown in Table 1. Measured plasma volume and PCWP significantly correlated, indicating maintained intravascular volume. Degree of renal impairment, as measured by SCT levels, correlated with various indicators of severity of sepsis, including CI, SVRI, WBC, and serum albumin; PRA correlated inversely with urine volume and sodium excretion. Urinary kallikrein excretion correlated positively with urine volume and sodium excretion. In group 1 sheep only (n = 8), there were significant inverse correlations between MAP and PCWP (r = - .72, P < .05), plasma volume (r
Baaellne 24 hr. Group 1
Baseline 24 hr. Group 2
Baseline 24 hr. Group 3
Fig 6. Plasma renin activity and urinary excretion rates of kallikrein, 6-keto-PGF1c<, and thromboxane 8 2 in septic sheep (as in Fig 1). = - .99, P < .05), and urinary excretion of 6keto-PGFla (r = - .94, P < .05). Urinary kallikrein excretion correlated with serum albumin (r = .79, P < .05), urine volume (r = .76, P < .05) and CCT (r = .77, P < .05). Other correlations did not achieve significance.
DISCUSSION
Multiple organ failure in generalized sepsis carries a high morbidity and mortality, particularly when the kidneys are involved;'the link between sepsis and changes in renal function is, therefore, of considerable clinical importance.1.2 Classical
28 Table 1.
CUMMING ET AL
Correlations Between Renal and Hemodynamic Parameters Found in the Whole Group of 20 Sheep PCWP (mm Hg)
Ser (JLmoIlL) Vp (mL) CDTPA (mUmin) Vu (mL)
CI (Umin/m2)
SVRI (d/sec/cm- 5/m 2)
(x 10"/L)
WBC
Total Protein (giL)
.49*
-.55*
-.52*
- .45*
PRA (ng/ml/h)
UK'I
UNa
(nkat/4h)
(mmol/4 h)
.50* - .41-
.44* -.52*
-.67t
.74t
-.74t
.57*
-.52-
.75t
UNa
(mmoll4 h) SAlb
(gIL)
.73t
Numbers shown are r values (Spearman Rank). *P < .05. tP < .01. Abbreviations: UKal , urinary kallikrein excretion; UNa' urinary sodium; V p' plasma volume; V u , urine volume; SAlb' serum albumin.
studies describe the inappropriate polyuria of sepsis,22 but oligoanuric renal failure with sodium retention resistant to plasma volume expansion is also commonly observed. 23 Renal biopsy in the established phase mayor may not show acute tubular necrosis and thus gives little help in understanding the pathophysiologic events that lead to renal failure, especially in the initiation phase. The importance of endotoxin has been emphasized2 but studies involving injection of endotoxin or bacteria to animals are complicated by the short survival after such insults and the usual development of a hypotensive low cardiac-output state 2.24 that does not resemble the hyperdynamic state more commonly observed in human sepsis. 25 Surgical induction of fecal peritonitis in sheep, combined with aggressive volume loading, produces a normotensive high cardiac-output vasodilated state, and we have used this model to investigate the changes in renal function induced by systemic sepsis over 24 hours. Twenty-four hours after induction of sepsis, three patterns of renal functional response were observed. Eight animals (group 1) showed increased Scp 60% reduction in GFR, marked sodium and water retention despite volume loading, with positive fluid balance of approximately 6 L of isotonic fluid per animal over 24 hours. This group showed manifestations of severe sepsis, ie, increased CI, reduced SVRI, and increased PAP. Furthermore, the decreases in WBC counts, serum total protein, albumin, and glucose were greatest
in this group. Four animals (group 2) showed a partial renal response in which GFR was preserved, but sodium and water retention comparable to group 1 was observed, indicating that significant changes in renal tubular function can occur in this situation in the absence of altered GFR. Eight animals (group 3) showed no alteration in renal function apart from minor sodium retention. Groups 2 and 3 did not demonstrate significant hemodynamic disturbance. The link between severity of sepsis and renal functional change was strengthened by studies of correlation, with significant correlations between SCr and CI, SVRI, WBC count, and serum total protein. Thus, it seems likely that changes in renal function are due to the septic state rather than to any other factor such as anesthesia, performance of a laparotomy, or analgesic administration, which were identical in all animals. The mechanism of renal response to this type of sepsis is undoubtedly complex. Those animals that developed a decline in GFR showed a small (17%), but significant, decrease in MAP; however, this change in isolation should not adversely affect renal function. 26 Similarly, the 20% decrease in renal plasma flow, which did not achieve statistical significance, seems unlikely to cause a drop in GFR and avid sodium reabsorption. 27 Other workers have documented increased renal blood flow in sepsis despite hypotension 2,28 and in ARF after shock. 29 All animals, particularly groups 1 and 2, were in positive fluid balance (4 to
RENAL RESPONSE TO SEPSIS
6 Ll24 h) and CVp, PCWp, and plasma volume were maintained at presepsis levels. The inverse correlation between plasma volume and sodium excretion makes it unlikely that the sodium retention reflects circulating volume contraction. Taking these factors into account, we believe that renal blood flow is not markedly reduced in this experimental situation and that hypovolemia does not explain the reduced GFR or the avid sodium retention. The renal functional changes could be explained by structural renal parenchymal damage. While various renal histologic features were observed, they were mild and uniformly distributed among animals showing totally different patterns of renal function. There was little or no evidence of structural damage to renal tubular cells, and immunofluorescent staining showed no evidence of immune complex deposition. There was also no evidence of intrarenal fibrin deposition with capillary microthrombi, characteristically seen in animals after endotoxin infusion. 30 We, therefore, suggest that, in this experimental model, the renal response to systemic sepsis is primarily mediated by vasoactive hormone systems known to affect kidney function, particularly the RAS, the renal KKS, and PGs. The marked increase in PRA in animals developing sodium retention (groups 1 and 2) indicates a major stimulus to renin release. Several mechanisms may be involved. Baroreceptor-mediated renin release could be influenced by the small reduction in MAP in group 1.31 The sympathetic nervous system is stimulated in sepsis32; increased circulating catecholamines and renal sympathetic nervous activity could stimulate renin release. 33 Altered sodium or chloride concentration at the macula densa could activate tubular glomerular feedback 34 ; also we observed increased urinary excretion of 6-ketoPGFla, a major metabolite of PGI 2, which stimulates renin release. 35 Angiotensin II (AlI) could play a major role in the renal response to sepsis. Increasing concentrations of AlI initially increase GFR by efferent arteriolar constriction, followed by a decrease in GFR due to a decline in the glomerular ultrafiltration coefficient and afferent arteriolar constriction. 36,37 The view that the RAS has a role in reducing GFR in this model is supported by the positive correlation of PRA with SCr and the inverse correlation with COTPA- The action of AlI to stimulate tubular
29
sodium reabsorption 38 may also be important. Since activation of the RAS in sepsis can be seen as appropriate to preserve the systemic circulation, it may prove difficult to confirm this by pharmacologic intervention; our preliminary experience using a converting-enzyme inhibitor indicates that severe hypotension and death may ensue (unpublished observations). Although the RAS has long been suspected of being involved in the genesis of a number of types of acute renal failure, 39 extensive investigations have failed to produce a conclusive answer.4O We believe that the RAS plays a significant part in the genesis of renal failure in this model. The renal effects of RAS activation are modulated by counterregulatory mechanisms, including the renal KKS10,1I and vasodilator PGs. 8,41 The renal KKS is primarily a local tissue hormone system; renal kinins are vasodilators and stimulate PG synthesis, and antagonize the renovascular and tubular effects of AlI and vasopressin, thereby facilitating water excretion. 10 ,1I,42 Since urinary kallikrein is identical to renal tissue kallikrein, urinary excretion of kallikrein is regarded as indicating activity of the renal KKS under most conditions. 10,43 The RAS and renal KKS are closely linked; most maneuvers that increase PRA also increase kallikrein excretion (eg, hypotensive hemorrhage 26 and sodium deprivation). 44 Saline infusion, however, suppresses PRA while increasing kallikrein excretion and causes natriuresis and diuresis. 45 In septic animals with sodium and water retention, we observed a discordant response: despite volumeloading, PRA increased and kallikrein excretion fell. This was not observed in animals showing no change in renal function. PRA correlated inversely with urinary sodium excretion and volume; urinary kallikrein excretion correlated positively with these parameters. These correlations would be predicted if reciprocal changes in activity of the RAS and renal KKS mediated changes in sodium and water excretion during sepsis. A similar pattern is produced by catecholamine infusion or stimulation of renal sympathetic nerves 46.48 ; sympathetic activity is increased in sepsis32 and could trigger reciprocal changes in renin and kallikrein. Such an imbalance between a vasoconstrictor sodium-retaining mechanism and a system with vasodilator natriuretic properties could contribute to the decreased GFR, and sodium and water retention seen in sepsis. Since both the RAS and the
30
CUMMING ET AL Endotoxin -
t
t
SEPSIS-
Circulating catecholamlnes Renal nerve activity
I
Systemic kinin.
/~
t PGI2
,.-----: -t /
PRA
~
I
- , BP, 'SVRI
-
Renal _ , kallikrein
tAnglOIl Renal 'kinin.
Thromboxane A2 --...:!to--_ _ _ _ _ _.../
renal KKS are located in the renal cortex,10 such imbalance might cause a greater decrease in cortical than medullary blood flow; such redistribution has been described in sepsis48 and could explain the marked reduction in GFR with relative preservation of total renal plasma flow. The urinary excretion rate of 6-keto-PGFla was increased, despite oliguria, in sheep showing a renal response to sepsis. It is known that elevated plasma levels of 6-keto-PGFla occur in sepsis7 and this may influence urinary excretion. 49 Possible stimuli to P0I 2 synthesis in sepsis include AII,50 norepinephrine,51 bradykinin,35.52 and hypokalemia. 53 PGI 2 is released from distended blood vessels 54; this may enable the vasculature to dilate to accommodate a volume load. 45 Saline infusion in humans increases excretion of the PGI 2 metabolite, dinor-6-keto-PGFla. 45 In group 1 animals, MAP correlated inversely with PCWp, plasma volume, and excretion of 6-keto-PGFla, suggesting that vascular PGI 2 synthesis in response to rapid fluid administration combined with oliguria and sodium retention may reduce BP by its potent vasodilator effect. 55 This effect may be augmented by endotoxin-induced generation of kinins, which are also potent vasodilators. 56 During sepsis, urinary excretion of TXB 2, the stable metabolite of the vasoconstrictor and proaggregatory prostanoid (TXA 2), was maintained at baseline levels. 7,8 Urinary TXB 2 in disease states may be derived more from systemic than renal TXA 2 synthesis 57; unchanged urinary TXB 2 excretion rate, despite 75 % reduction in GFR and an 85 % reduction in urine volume, therefore suggests increased systemic TXA2 generation. Increased plasma TXB 2 has been described in endotoxemia 58 and bacteremia. 7 In addition to vasoconstriction, TXA 2 causes contraction of mesangial cells, a property shared with All; this could contribute to the reduction in GFR in septic animals. 59
'GFR
-'Vu , UNaV
Fig 7. Hypothetical interaction of factors that may contribute to renal dysfunction in sepsis. Vu , urine volume; UNaV, urinary sodium excretion rate.
Many other mechanisms undoubtedly contribute to the renal response to sepsis. Plasma antidiuretic hormone (ADH) levels are elevated in sepsis,60 although increased urinary osmolarity was observed in a minority of animals (group 2). Increased sympathetic nervous system activity may have a direct effect to reduce GFR and sodium excretion in sepsis; these actions may be potentiated by AII.61 The hypoglycemia and hypokalemia that we observed during sepsis suggests insulin-stimulated transport of potassium and glucose into cells; insulin could also affect renal function. 62 Endotoxin-induced complement activation may also be important. 63 Differing patterns of renal response to sepsis probably reflect the net effect of several regulatory and counterregulatory mechanisms, many of which may function primarily to maintain an effective systemic circulation. Some hypothetical interactions are shown in Fig 7, and several of these should be testable experimentally. Our results suggest that reciprocal changes in the activities of the RAS and the renal KKS may be important in the renal response to sepsis; PGI 2 generation is increased, with possible systemic and renal effects; and TXA 2 could also contribute to the renal response to sepsis. ACKNOWLEDGMENT The authors thank Andrea Neal for technical assistance and Lyn MacLean for secretarial help. Allan D. Cumming was a Clinical Research Fellow of the Kidney Foundation of Canada.
REFERENCES 1. Werb R, Linton AL: Aetiology, diagnosis, treatment and prognosis of acute renal failure in an intensive care unit. Resuscitation 7:95-100, 1979 2. Wardle N: Acute renal failure in the 1980s; the importance of septic shock and of endotoxernia. Nephron 30:193200, 1982 3. Richmond 1M, Walker IF, Avila A, et al: Renal and cardiovascular response to nonhypotensive sepsis in a large animal model with peritonitis. Surgery 97:205-213, 1985
RENAL RESPONSE TO SEPSIS 4. Walker JF, Cumming AD, Lindsay RM, et al: The renal response produced by non-hypotensive sepsis in a large animal model. Am J Kidney Dis 8:88-97, 1986 5. Thurau K, Vogt C, Dahlheim H: Renin activity in the juxtaglomerular apparatus of the rat kidney during post-ischemic acute renal failure. Kidney Int 1O:S177-S182, 1976 6. Oken DE: Hypothesis on the role of prostaglandins in the pathogenesis of acute renal failure. Lancet 1:1319-1322, 1975 7. Butler RR Jr, Wise WC, Halushka PV, et al: Thromboxane and prostacyclin production during septic shock. Adv Shock Res 7:133-145, 1982 8. Gerber JG, Anderson RJ, Schrier RW, et al: Prostaglandins and the regulation of renal circulation and function. Adv Prostaglandin Thromboxane Leukotriene Res 10:227-254, 1982 9. Patrono C, Ciabattoni G, Patrignani P, et al: Evidence for a renal origin of urinary thromboxane B2 in health and disease. Adv Prostaglandin Thromboxane Leukotriene Res 11 :493-498, 1983 10. Levinsky NG: The renal kallikrein-kinin system. Circ Res 44:441-451, 1979 11. Mills IH: The renal kallikrein-kinin system and sodium excretion. Q J Exp Physiol 67:393-399, 1982 12. Abe K, Sato M, Imai Y, et al: Renal kallikrein-kinin: Its relation to renal prostaglandins and renin-angiotensinaldosterone in man. Kidney Int 19:869-880, 1981 13. Wichterman KA, Baur AE, Chaudry IH: Sepsis and septic shock-A review of laboratory models and a proposal. J Surg Res 29:189-201, 1980 14. Haber E, Koerner T, Paige LB, et al: Application of a radioimmunoassay for angiotensin I to the physiologic measurement of plasma renin activity in normal human subjects. J Clin Endocrinol Metab 29: 1349-1355, 1969 15. Amundsen E, Putter J, Friberger P, et al: Methods for the determination of glandular kallikrein by means of a chromogenic tripeptide substrate. Adv Exp Med Bioi 120A:83-95, 1979 16. Ali M, McDonald JWD: Synthesis of thromboxane B2 and 6-keto-prostaglandin Fla by bovine gastric mucosal and muscle microsomes. Prostaglandins 20:245-254, 1980 17. Ali M, Zamecnik J, Cerskus AL, et a1: Synthesis of thromboxane B2 and prostaglandins by bovine gastric mucosal microsomes. Prostaglandins 14:819, 1977 18. International Committee for Standardization in Hematology: Standard technique for the estimation of red cell and plasma volume. Br J HaematoI25:801-814, 1973 19. Dubofsky EV, Russell CD: Quantitation of renal function with glomerular and tubular agents. Semin Nucl Med 12:308-329, 1982 20. Bianchi C, Bonadio M, Donadio C, et a1: Measurement of glomerular filtration rate in man using DTPA_ 99mTc. Nephron 24:174-178, 1979 21. Solez K, Ideura T, Silvia CB, et a1: Clonidine after renal ischemia to lessen acute renal failure and microvascular damage. Kidney Int 18:309-322, 1980 22. Hermreck AS, Berg RA, Ruhlen JR, et a1: Renal response in sepsis. Arch Surg 107:169-175, 1973 23. Vaz AJ: Low fractional excretion of urine sodium in acute renal failure due to sepsis. Arch Intern Med 143:738739, 1983 24. Hinshaw LB, Benjamin B, Holmes DD, et al: Responses
31
of the baboon to live E. coli organisms and endotoxins. Surg Gynecol Obstet 145:1-11,1977 25. Siegel JH, Cirra FB, Coleman B, et al: Physological and metabolic correlations in human sepsis. Surgery 86:163-193, 1979 26. Maier M, StarIinger M, Wagner M, et al: The effect of hemorrhagic hypotension on urinary kallikrein excretion, renin activity, and renal cortical blood flow in the pig. Circ Res 48:386-392, 1981 27. Schnermann J, Briggs JP: Participation of renal cortical prostaglandins in the regulation of glomerular filtration rate. Kidney Int 19:802-815, 1981 28. Cronenwett JL, Lindenauer SM: Hemodynamic effects of cecal ligation sepsis in dogs. J Surg Res 33:324-331, 1982 29. Reubi F, Vorburger C: Renal hemodynamics in acute renal failure after shock in man. Kidney Int 1O:S137-SI43, 1976 30. Westwick J, Fletcher MS, Kakkar VV: Inhibition of thromboxane formation prevents endotoxin-induced renal fibrin deposition in jaundiced rats. Adv Prostaglandin Thromboxane Leukotriene Res 12:83-91, 1983 31. Blaine EH, Zimmerman MB: Renin secretion during dynamic changes in renal perfusion pressure. Am J Physiol 236:F546-F551, 1979 32. Sugerman HJ, Austin G, Newsome HH, et al: Hemodynamics, oxygen consumption and serum catecholamine changes in progressive, lethal peritonitis in the dog. Surg Gynecol Obstet 154:8-12, 1982 33. Henrich WL, Schrier RW, BerI T: Mechanisms of renin secretion during hemorrhage in the dog. J Clin Invest 64:1-7, 1969 34. Baylis C, Ichikawa I, Willis WT, et al: Dynamics of glomerular ultrafiltration. IX. Effects of plasma protein concentration. Am J PhysioI232:F58-F71, 1977 35. Mullane KM, Moncada S: Prostacyclin release and the modulation of some vasoactive hormones. Prostaglandins 20:25-49, 1980 36. Hall JE, Guyton AC, Jackson TE, et al: Control of glomerular filtration rate by renin-angiotensin system. Am J Physiol 233:F366-F372, 1977 37. Ishikawa I, Hollenberg NK: Pharmacologic interruption of the renin-angiotensin system in myohemoglobinuric acute renal failure. Kidney Int 1O:S183-S190, 1976 38. Levens NR, Peach MJ, Carey RM: Role of the intrarenal renin-angiotensin system in the control of renal function. Circ Res 48:157-167, 1981 39. Brown JJ, Gleadle RI, Lawson DH, et al: Renin and acute renal failure: Studies in man. Br Med J 1:253-258, 1970 40. Oken DE, Cotes SC, Flamenbaum W, et al: Active and passive immunization to angiotensin in experimental acute renal failure. Kidney Int 7: 12-18, 1985 41. Henrich WL, BerI T, McDonald KM, et al: Angiotensin II, renal nerves and prostaglandins in renal hemodynamics during hemorrhage. Am J Physiol 235:F46-F51, 1978 42. Levy SB, Lilley JJ, Frigon RP, et al: Urinary kallikrein and plasma renin activity as determinants of renal blood flow. J Clin Invest 60: 129-138, 1977 43. Pisano JJ: Observations on the kallikrein-kinin system in the kidney. Adv Exp Med Bioi 156B:929-938, 1983 44. Margolius HS, Horwitz D, Pisano JJ, et a1: Urinary kallikrein excretion in hypertensive man. Relationships to sodium
32 intake and sodium retaining steroids. Circ Res 35:820-825, 1974 45. Watson ML, Cumming AD, Lambie AT, et al: Urinary kallilkrein and systemic protacyc1in synthesis during sodium chloride infusion in normal man. Clin Sci 68:537-543, 1985 46. Olsen UB: Changes of urinary kallikrein and kinin excretions induced by adrenaline infusion in conscious dogs. Scand J Clin Lab Invest 40: 173-178, 1980 47. Albertini R, Ladron DE, Guevara R, et al: Effect of renal nerve stimulation on urine and tissue kininogenase activity in cats. Hypertension 3:50-54, 1981 48. Stone AM, Stein T, Lafortune J, et al: Changes in intrarenal blood flow during sepsis. Surg Gynecol Obstet 148:731-734, 1979 49. Rosenkranz B, Kitajima W, Frolich JC: Relevance of urinary 6-keto-prostaglandin Fla determination. Kidney Int 19:755-759, 1981 50. Shebuski RJ, Aiken JW: Angiotensin II-induced renal prostacyclin release suppresses platelet aggregation in the anesthetized dog. Adv Prostaglandin Thromboxane Res 7:11491152, 1980 51. Vandonger R, O'Dwyer J, Barden A: Effect of alphaadrenergic stimulation on prostacyclin and renin release in the rat kidney. Prostaglandins 24:815-825, 1982 52. O'Donnell TF, Clowes GHA, Talamo RC, et al: Kinin activation in the blood of patients with sepsis. Surg Gynecol Obstet 143:538-545, 1976 53. Gullner HF,' Graf AK, Gill JR, et al: Hypokalemia stimulates prostaglandin synthesis in the rat. Clin Sci 65:43-46, 1983
CUMMING ET AL 54. Ritter JM, Blair lA, Barrow SE, et al: Release of prostacyclin in vivo and its role in man. Lancet 1:317-319, 1983 55. Harris RH, Zmudka M, Maddox Y, et al: Relationships of TXB z and 6-keto PGFla to the hemodynamic changes during baboon endotoxic shock. Adv Prostaglandin Thromboxane Res 7:843-849, 1980 56. Regoli D, Barabe J: Pharmacology of bradykinin and related kinins. Pharmacol Rev 32: 1-46, 1980 57. Zipser RD, Kronborg I, Rector W, et al: Therapeutic trial of thromboxane synthesis inhibition in the hepatorenal syndrome. Gastroenterology 87: 1228-1232, 1984 58. Winn R, Harlan J, Nadir B, et al: Thromboxane A z mediates lung vasoconstriction but not permeability after endotoxin. J Clin Invest 72:911-918, 1983 59. Mene P, Simonson M, Rettberg C, et al: Thromboxane A z and endoperoxide analogs contract cultured rat glomerular mesangial cells. Am Soc Nephrol 191A, 1985 (abstr) 60. Wilson MF, Brackett DJ, Tompkins P, et al: Elevated plasma vasopressin concentrations during endotoxin and E. coli shock. Adv Shock Res 6: 16-26, 1981 61. Lanier SM, Malik KU: Attenuation by prostaglandins of the facilitatory effect of angiotensin II at the adrenergic prejunctional sites in the isolated Krebs-perfused rat heart. Circ Res 51:594-601, 1982 62. Gregg CM, Cohen JJ, Black AH, et al: Effects of glucose and insulin on metabolism and function of perfused rat kidney. Am J Physiol 235:F52-F61, 1978 63. Barnes J, Osgood RW, Pinckard RN, et al: The effect of complement depletion on the post-ischemic model of acute renal failure. Kidney Int 14:721, 1978
T
HE LINK between sepsis and renal functional changes including classical acute renal failure (ARF) , is clinically important and poorly understood. 1.2 While investigating the renal response to sepsis,3 we have recently established a model of severe systemic sepsis in sheep. Circulating volume is maintained by aggressive fluid administration, and hypotension is avoided. At an early stage, we recognized various patterns of renal response, probably reflecting minor differences in surgical technique and constitutional variation between animals. 4 More recently, we have largely eliminated this variability; we did, however, undertake a detailed analysis of the different response patterns in our initial group of 20 sheep, hoping to gain information regarding the mechanism of renal functional changes and, particularly, the role of vasoactive hormones acting on the kidney. We found a correlation between severity of sepsis, judged by hemodynamic, hematologic, and biochemical disturbance, and the type of renal response. In this study we have defined this relationship and attempted to correlate different patterns of renal functional change with alterations in me-
diator systems that might influence kidney function in sepsis. It has been suggested that vasoactive substances may be important in the pathogenesis of some types of ARP and that vasodilator prostaglandins (PGs) protect renal function against a variety of insults. 6 In systemic sepsis, increased plasma renin activity (PRA) and increased plasma concentrations of 6-keto-PGFla and of thromboxane B2 (TXB 2) have been documented. 3,7 The possible influence of these compounds on renal function in sepsis has not been defined. We have, therefore, measured PRA, urinary 6-keto-PGFla and TXB 2 From the Department of Medicine, The University of Western Ontario, London, Ontario, Canada; and the Department of Pathology, Johns Hopkins University School of Medicine, Baltimore, Supported in part by NATO Grant No, RG.85/0043, The present address of Dr Solez is University Department of Medicine, Royal Infirmary, Edinburgh, Scotland. Address reprint requests to Adam L. Linton, MB, Department of Medicine, Victoria Hospital, P.O. Box 5375, London, Ontario, Canada N6A 4G5. © 1988 by the National Kidney Foundation, Inc. 0272-6386/88/1101 -0005$3.00/0
American Journal of Kidney Diseases, Vol XI, No 1 (January), 1988: pp 23-32
23
24
CUMMING ET AL
in septic sheep. 8.9 We also measured urinary kallikrein excretion, since the renal kallikrein-kinin system (KKS) interacts with the renin-angiotensin system (RAS) and renal PGs. JO-12 We have tested correlations between these variables and other aspects of the septic state, particularly changes in systemic hemodynamics and renal function. MATERIALS AND METHODS The technique for surgical induction of peritonitis is modified from that of Wichterman et al 13 and has been previously described in detail. 3 In brief, 20 sheep, aged 9 to 18 months and weighing 35 to 50 kg, underwent general anesthesia and cannulation of the aorta via the common carotid artery, and of the pulmonary artery via the external jugular vein with a triplelumen Swan-Ganz catheter. The bladder was catheterized per urethra. After recovery from anesthesia, the animals were housed in a metabolic cage and volume-loaded with 8 L of Ringer's lactate intravenously (IV) over 24 hours. After control hemodynamic measurements and blood and urine sampling, animals were anesthetized and peritonitis was induced by devascularization and proximal ligation of 6 to 8 cm of distal cecum, combined with partial omentectomy and a 2-cm puncture in the cecal tip. The abdomen was closed and the animal allowed to recover from anesthesia. Postoperatively, all sheep received 50 mg Demerol (Merperidine, Abbott Labs, Toronto) and 25 mg Acepromazine IV (Atra-Vet, Ayerst Laboratories, Montreal) and were continued thereafter on IV Demerol, 50 mg/6 to 8 h. IV infusion of Ringer's lactate, 8 Ll24 h, continued for the duration of the experiment. Follow-up hemodynamic measurements and blood and urine collections were made 24 and, where possible, 48 hours after induction of sepsis. Animals were continuously under the care of qualified animal health technicians during the study; during sepsis, the sheep showed lethargy, anorexia, pyrexia, and increased respiratory rate. Any animal observed to be restless, agitated, or showing other evidence of discomfort or distress despite Demerol was killed immediately by IV injection of pentobarbital (Euthanol, Euthanol Forte, MTC Pharmaceuticals, Mississauga, Canada). Mortality was, therefore, not used as an endpoint in the study. Animals surviving through 48 hours were killed as previously described; an open renal biopsy was taken from all sheep at the time of death. The study was approved by the institutional ethical review process of The University of Western Ontario, and the care of the animals conformed with the conditions set down in the "Guide to the Care and Use of Experimental Animals of the Canadian Council on Animal Care." Hemodynamic parameters measured were mean arterial pressure (MAP), mean pulmonary artery pressure (PAP), and pulmonary capillary wedge pressure (PCWP) (mm Hg). Central venous pressure (CVP) was measured in centimeters of H 20 and thermodilution cardiac output in liters per minute, as previously described. 3 Cardiac index (CI) and systemic vascular resistance index (SVRI) were derived by standard formulae. Blood samples were taken from the aortic cannula; WBC was measured with a Coulter cell counter (Model S., Coulter Electronics, Burglinton, Ontario), and serum creatinine (SCr), total protein, albumin, glucose, sodium, and potassium measured by standard autoanalyzer methods. PRA was measured by ra-
dioimmunoassay14 and results expressed in nanograms per milliliter per hour of generated angiotensin 1. Urine was collected for four hours on -the morning of each study day and the volume recorded; creatinine, sodium, potassium, and osmolarity were measured by standard methods. Urinary kallikrein was assayed by the method of Amundsen et al,15 measuring the ability of urine to hydrolyze the synthetic tripeptide chromogenic substrate, S2266 (Kabi Diagnostica, Stockholm). Urinary 6-ketoPGFla and TXB2 were measured by radioimmunoassay.16.17 Plasma volume was measured as the immediate distribution space of 1-125 albumin l8 ; effective renal plasma flow was measured as the clearance of orthoiodinated hippuran labeled with 1-131 19 ; clearance of Tc-99m-DTPA was measured as an additional indicator of glomerular filtration rate (GFR).20 In each case, the isotope-labeled compound was injected via the venous catheter, and blood samples were taken from the aortic cannula. Kidney tissue was examined by light microscopy using standard methods. 21 Only results at baseline and after 24 hours of sepsis were analyzed since only one of eight sheep in the most severely septic group survived through 48 hours. Statistical analysis was performed using the Data Analysis Interactive System (Rainbow Computing, Inc, Northridge, CAl. Comparison between groups at baseline and 24 hours was by the Mann-Whitney U test; comparisons, within groups, between baseline and 24 hour data were by the Wilcoxon Rank Sum test. Studies of correlation were made for 24-hour postsepsis data from the whole animal population by Spearman Rank coefficients. Values of P less than .05 were held to be significant; results are expressed as mean ± SEM.
RESULTS
For analysis, sheep were divided into three groups on the basis of their renal function after induction of generalized sepsis. Sheep showing a significant decline in GFR after 24 hours of sepsis, as evidenced by a rise in Ser and decreased clearance of endogenous creatinine and DTPA, were allocated to group 1 (n = 8) (Fig 1). These animals also exhibited major decrements in urine volume, sodium excretion rate, and fractional sodium excretion (Fig 2). Sheep showing oliguria and sodium retention without reduction in GFR were allocated to group 2 (n = 4) (Figs 1 and 2). Sheep that exhibited no change in glomerular or tubular function apart from a modest decrease in sodium excretion were allocated to group 3 (n = 8) (Figs 1 and 2). Renal plasma flow decreased during sepsis in group 1 only, but this did not achieve significance (Fig 1). Urinary osmolarity remained around 400 mosm/kg after induction of sepsis, except in group 2 where all four animals concentrated urine to > 800 mosm/kg. Indicators of Circulating VOlume Status
Comparison of hemodynamic measurements at baseline with those 24 hours after induction of sep-
25
RENAL RESPONSE TO SEPSIS 150
~
2000
~
1600
3
Z
!
.p=
>
~
,!!
1i
" U
800
~
E
"
Ii
(/)
0
0
'2 200 E
!
.";;; ~ . :&
150
U
0
u
~Il
~~
c
~i ..
100
rZ_ ~
c
..
.;
1i
~
.~ i
Il
"
"uc
OJ
.S... iu
...
~
(/)o & ~
13 100 -< II.
15
~-
..
.2
50
U
ol:
0
1200
!l
t1.
J ]
800
400 200
o Baaellne
24 hr.
Group 1
Ba.ellne
24 hr.
Group 2
Baaellne
24 hr.
Group 3
Fig 1. Serum creatinine, creatinine clearance, DTPA clearance, and renal plasma flow at baseline and after 24 hours of sepsis in 3 groups of septic sheep (group 1, n = 8; group 2, n = 4; group 3, n = 8; mean ± SEM).
sis revealed that CVp, PCWp, and plasma volume were maintained in all groups (Fig 3). There was a small (17 %) decline in MAP in group 1 animals that was statistically significant, but no change in MAP in groups 2 and 3. Indicators of Severity of Sepsis
In all sheep, a polymicrobial peritonitis and bacteremia was confirmed at 24 hours by blood culture. Organisms grown included Serratia marcescens, Enterobacter cloaceae, Pseudomonas, Bacteroides sp, and Escherichia coli. Autopsyexamination revealed distended loops of bowel and an inflammatory mass in the right lower quadrant.
B..ellne
24 hr.
Group 1
Ba.ellne
24 hr.
Group 2
Base....
24 hr.
Group 3
Fig 2. Urine volume, sodium excretion, fractional sodium excretion, and urinary osmolarity in septic sheep (as in Fig 1).
All animals developed pyrexia and tachypnea. WBC counts fell in all groups, but 24-hour values were significantly lower in group 1 than in group 3 (Fig 4). Similarly, while serum total protein, albumin, and glucose levels fell in all three groups, these changes were greatest in group 1 (Fig 5). Significant hemodynamic changes occurred only in group 1 with increased CI, reduced SVRI, and increased PAP (Fig 4). Serum potassium fell in all groups (Fig 5). Hormonal Changes
Both PRA and urinary excretion of 6-ketoPGFla increased significantly during sepsis in groups 1 and 2, but not in group 3. Urinary kallikrein excretion decreased significantly in groups 1
26
CUMMING ET AL
.. .. 4:0
10
!;
• p
8
....
N
6
.5 ~ u c:
i E >,3
4
OJ OJ
WI
6 :J:
~
.. .. .. ct
~
'ON
.!! '0 ~
'E
, .J
~~
2
C
()
0 15
!; OJ OJ
. a
'0
10
~
a
:J:
.. E E
~
..
ci.~
5
()
.il:l
0
0..
!..
..". ~
.t"
2000
~ E
E .a 1500
..
.~.5
0
>
.. a:"
E
'"
,.,~
~ .. :J:
15
c
.so
1000
5
~
500
12 10
. £
150
8
!:>
6
~ 100
4
a
Ii :J: 'i E
t: E
..
00(
c:
:!
~
2
50
o Baseline Baseline 24 hr. 24 hr. Group 1 Group :!
o Baseline
24 hr. Group 1
Baseline
24 hr. Group 2
Baseline
24 hr. Group 3
Fig 3. Central venous pressure, pulmonary capillary wedge pressure, plasma volume and mean arterial pressure in septic sheep (as in Fig 1).
and 2, but not in group 3, while urinary TXB 2 excretion was unchanged during sepsis in all three groups (Fig 6). Renal Histology Various histologic features were observed in kidney biopsies taken at time of animal death. Kidney tissue from three animals was unsuitable for examination for technical reasons. In the other 17 sheep, the changes observed included hyperplasia of the juxtaglomerular apparatus (JGA) (six animals), mild tubular dilatation (eight animals), very
Baseline 24 hr. Group 3
Fig 4. Cardiac index, systemic vascular resistance index, pulmonary artery pressure and white cell count in septic sheep (as in Fig 1).
occasional foci of tubular regeneration (four animals), patchy interstitial infiltrate of mononuclear cells (13 animals), and small foci of medullary calcification (12 animals). With the exception of JGA hyperplasia,4 the pathologic changes were minor and evenly distributed between groups of animals, irrespective of the pattems of renal functional response. Studies of Correlation The three measurements reflecting GFR, ie, serum creatinine (SCr)' creatinine clearance (CCr), and DTPA clearance (C DTPA), correlated well during sepsis (SCr v CCr: r = _.. 84, P < .001;
27
RENAL RESPONSE TO SEPSIS • p
10
.~ 50
e
Q,
40
1i 30
o
.... 20
e
t
VI
10
0
.~
.5 ;;:
20
~OJ
VI
-S ~
6 5
4
Baaellne Baaellne 24 hr. 24 hr. Group 1 Group 2
Baseline 24 hr. Group 3
Fig 5. Serum total protein, albumin, glucose, and potassium in septic sheep (as in Fig 1).
CCT v C DTPA : r = - .77, P < .001; SCT v CDTPA : r = .75, P < .001). Other correlations are shown in Table 1. Measured plasma volume and PCWP significantly correlated, indicating maintained intravascular volume. Degree of renal impairment, as measured by SCT levels, correlated with various indicators of severity of sepsis, including CI, SVRI, WBC, and serum albumin; PRA correlated inversely with urine volume and sodium excretion. Urinary kallikrein excretion correlated positively with urine volume and sodium excretion. In group 1 sheep only (n = 8), there were significant inverse correlations between MAP and PCWP (r = - .72, P < .05), plasma volume (r
Baaellne 24 hr. Group 1
Baseline 24 hr. Group 2
Baseline 24 hr. Group 3
Fig 6. Plasma renin activity and urinary excretion rates of kallikrein, 6-keto-PGF1c<, and thromboxane 8 2 in septic sheep (as in Fig 1). = - .99, P < .05), and urinary excretion of 6keto-PGFla (r = - .94, P < .05). Urinary kallikrein excretion correlated with serum albumin (r = .79, P < .05), urine volume (r = .76, P < .05) and CCT (r = .77, P < .05). Other correlations did not achieve significance.
DISCUSSION
Multiple organ failure in generalized sepsis carries a high morbidity and mortality, particularly when the kidneys are involved;'the link between sepsis and changes in renal function is, therefore, of considerable clinical importance.1.2 Classical
28 Table 1.
CUMMING ET AL
Correlations Between Renal and Hemodynamic Parameters Found in the Whole Group of 20 Sheep PCWP (mm Hg)
Ser (JLmoIlL) Vp (mL) CDTPA (mUmin) Vu (mL)
CI (Umin/m2)
SVRI (d/sec/cm- 5/m 2)
(x 10"/L)
WBC
Total Protein (giL)
.49*
-.55*
-.52*
- .45*
PRA (ng/ml/h)
UK'I
UNa
(nkat/4h)
(mmol/4 h)
.50* - .41-
.44* -.52*
-.67t
.74t
-.74t
.57*
-.52-
.75t
UNa
(mmoll4 h) SAlb
(gIL)
.73t
Numbers shown are r values (Spearman Rank). *P < .05. tP < .01. Abbreviations: UKal , urinary kallikrein excretion; UNa' urinary sodium; V p' plasma volume; V u , urine volume; SAlb' serum albumin.
studies describe the inappropriate polyuria of sepsis,22 but oligoanuric renal failure with sodium retention resistant to plasma volume expansion is also commonly observed. 23 Renal biopsy in the established phase mayor may not show acute tubular necrosis and thus gives little help in understanding the pathophysiologic events that lead to renal failure, especially in the initiation phase. The importance of endotoxin has been emphasized2 but studies involving injection of endotoxin or bacteria to animals are complicated by the short survival after such insults and the usual development of a hypotensive low cardiac-output state 2.24 that does not resemble the hyperdynamic state more commonly observed in human sepsis. 25 Surgical induction of fecal peritonitis in sheep, combined with aggressive volume loading, produces a normotensive high cardiac-output vasodilated state, and we have used this model to investigate the changes in renal function induced by systemic sepsis over 24 hours. Twenty-four hours after induction of sepsis, three patterns of renal functional response were observed. Eight animals (group 1) showed increased Scp 60% reduction in GFR, marked sodium and water retention despite volume loading, with positive fluid balance of approximately 6 L of isotonic fluid per animal over 24 hours. This group showed manifestations of severe sepsis, ie, increased CI, reduced SVRI, and increased PAP. Furthermore, the decreases in WBC counts, serum total protein, albumin, and glucose were greatest
in this group. Four animals (group 2) showed a partial renal response in which GFR was preserved, but sodium and water retention comparable to group 1 was observed, indicating that significant changes in renal tubular function can occur in this situation in the absence of altered GFR. Eight animals (group 3) showed no alteration in renal function apart from minor sodium retention. Groups 2 and 3 did not demonstrate significant hemodynamic disturbance. The link between severity of sepsis and renal functional change was strengthened by studies of correlation, with significant correlations between SCr and CI, SVRI, WBC count, and serum total protein. Thus, it seems likely that changes in renal function are due to the septic state rather than to any other factor such as anesthesia, performance of a laparotomy, or analgesic administration, which were identical in all animals. The mechanism of renal response to this type of sepsis is undoubtedly complex. Those animals that developed a decline in GFR showed a small (17%), but significant, decrease in MAP; however, this change in isolation should not adversely affect renal function. 26 Similarly, the 20% decrease in renal plasma flow, which did not achieve statistical significance, seems unlikely to cause a drop in GFR and avid sodium reabsorption. 27 Other workers have documented increased renal blood flow in sepsis despite hypotension 2,28 and in ARF after shock. 29 All animals, particularly groups 1 and 2, were in positive fluid balance (4 to
RENAL RESPONSE TO SEPSIS
6 Ll24 h) and CVp, PCWp, and plasma volume were maintained at presepsis levels. The inverse correlation between plasma volume and sodium excretion makes it unlikely that the sodium retention reflects circulating volume contraction. Taking these factors into account, we believe that renal blood flow is not markedly reduced in this experimental situation and that hypovolemia does not explain the reduced GFR or the avid sodium retention. The renal functional changes could be explained by structural renal parenchymal damage. While various renal histologic features were observed, they were mild and uniformly distributed among animals showing totally different patterns of renal function. There was little or no evidence of structural damage to renal tubular cells, and immunofluorescent staining showed no evidence of immune complex deposition. There was also no evidence of intrarenal fibrin deposition with capillary microthrombi, characteristically seen in animals after endotoxin infusion. 30 We, therefore, suggest that, in this experimental model, the renal response to systemic sepsis is primarily mediated by vasoactive hormone systems known to affect kidney function, particularly the RAS, the renal KKS, and PGs. The marked increase in PRA in animals developing sodium retention (groups 1 and 2) indicates a major stimulus to renin release. Several mechanisms may be involved. Baroreceptor-mediated renin release could be influenced by the small reduction in MAP in group 1.31 The sympathetic nervous system is stimulated in sepsis32; increased circulating catecholamines and renal sympathetic nervous activity could stimulate renin release. 33 Altered sodium or chloride concentration at the macula densa could activate tubular glomerular feedback 34 ; also we observed increased urinary excretion of 6-ketoPGFla, a major metabolite of PGI 2, which stimulates renin release. 35 Angiotensin II (AlI) could play a major role in the renal response to sepsis. Increasing concentrations of AlI initially increase GFR by efferent arteriolar constriction, followed by a decrease in GFR due to a decline in the glomerular ultrafiltration coefficient and afferent arteriolar constriction. 36,37 The view that the RAS has a role in reducing GFR in this model is supported by the positive correlation of PRA with SCr and the inverse correlation with COTPA- The action of AlI to stimulate tubular
29
sodium reabsorption 38 may also be important. Since activation of the RAS in sepsis can be seen as appropriate to preserve the systemic circulation, it may prove difficult to confirm this by pharmacologic intervention; our preliminary experience using a converting-enzyme inhibitor indicates that severe hypotension and death may ensue (unpublished observations). Although the RAS has long been suspected of being involved in the genesis of a number of types of acute renal failure, 39 extensive investigations have failed to produce a conclusive answer.4O We believe that the RAS plays a significant part in the genesis of renal failure in this model. The renal effects of RAS activation are modulated by counterregulatory mechanisms, including the renal KKS10,1I and vasodilator PGs. 8,41 The renal KKS is primarily a local tissue hormone system; renal kinins are vasodilators and stimulate PG synthesis, and antagonize the renovascular and tubular effects of AlI and vasopressin, thereby facilitating water excretion. 10 ,1I,42 Since urinary kallikrein is identical to renal tissue kallikrein, urinary excretion of kallikrein is regarded as indicating activity of the renal KKS under most conditions. 10,43 The RAS and renal KKS are closely linked; most maneuvers that increase PRA also increase kallikrein excretion (eg, hypotensive hemorrhage 26 and sodium deprivation). 44 Saline infusion, however, suppresses PRA while increasing kallikrein excretion and causes natriuresis and diuresis. 45 In septic animals with sodium and water retention, we observed a discordant response: despite volumeloading, PRA increased and kallikrein excretion fell. This was not observed in animals showing no change in renal function. PRA correlated inversely with urinary sodium excretion and volume; urinary kallikrein excretion correlated positively with these parameters. These correlations would be predicted if reciprocal changes in activity of the RAS and renal KKS mediated changes in sodium and water excretion during sepsis. A similar pattern is produced by catecholamine infusion or stimulation of renal sympathetic nerves 46.48 ; sympathetic activity is increased in sepsis32 and could trigger reciprocal changes in renin and kallikrein. Such an imbalance between a vasoconstrictor sodium-retaining mechanism and a system with vasodilator natriuretic properties could contribute to the decreased GFR, and sodium and water retention seen in sepsis. Since both the RAS and the
30
CUMMING ET AL Endotoxin -
t
t
SEPSIS-
Circulating catecholamlnes Renal nerve activity
I
Systemic kinin.
/~
t PGI2
,.-----: -t /
PRA
~
I
- , BP, 'SVRI
-
Renal _ , kallikrein
tAnglOIl Renal 'kinin.
Thromboxane A2 --...:!to--_ _ _ _ _ _.../
renal KKS are located in the renal cortex,10 such imbalance might cause a greater decrease in cortical than medullary blood flow; such redistribution has been described in sepsis48 and could explain the marked reduction in GFR with relative preservation of total renal plasma flow. The urinary excretion rate of 6-keto-PGFla was increased, despite oliguria, in sheep showing a renal response to sepsis. It is known that elevated plasma levels of 6-keto-PGFla occur in sepsis7 and this may influence urinary excretion. 49 Possible stimuli to P0I 2 synthesis in sepsis include AII,50 norepinephrine,51 bradykinin,35.52 and hypokalemia. 53 PGI 2 is released from distended blood vessels 54; this may enable the vasculature to dilate to accommodate a volume load. 45 Saline infusion in humans increases excretion of the PGI 2 metabolite, dinor-6-keto-PGFla. 45 In group 1 animals, MAP correlated inversely with PCWp, plasma volume, and excretion of 6-keto-PGFla, suggesting that vascular PGI 2 synthesis in response to rapid fluid administration combined with oliguria and sodium retention may reduce BP by its potent vasodilator effect. 55 This effect may be augmented by endotoxin-induced generation of kinins, which are also potent vasodilators. 56 During sepsis, urinary excretion of TXB 2, the stable metabolite of the vasoconstrictor and proaggregatory prostanoid (TXA 2), was maintained at baseline levels. 7,8 Urinary TXB 2 in disease states may be derived more from systemic than renal TXA 2 synthesis 57; unchanged urinary TXB 2 excretion rate, despite 75 % reduction in GFR and an 85 % reduction in urine volume, therefore suggests increased systemic TXA2 generation. Increased plasma TXB 2 has been described in endotoxemia 58 and bacteremia. 7 In addition to vasoconstriction, TXA 2 causes contraction of mesangial cells, a property shared with All; this could contribute to the reduction in GFR in septic animals. 59
'GFR
-'Vu , UNaV
Fig 7. Hypothetical interaction of factors that may contribute to renal dysfunction in sepsis. Vu , urine volume; UNaV, urinary sodium excretion rate.
Many other mechanisms undoubtedly contribute to the renal response to sepsis. Plasma antidiuretic hormone (ADH) levels are elevated in sepsis,60 although increased urinary osmolarity was observed in a minority of animals (group 2). Increased sympathetic nervous system activity may have a direct effect to reduce GFR and sodium excretion in sepsis; these actions may be potentiated by AII.61 The hypoglycemia and hypokalemia that we observed during sepsis suggests insulin-stimulated transport of potassium and glucose into cells; insulin could also affect renal function. 62 Endotoxin-induced complement activation may also be important. 63 Differing patterns of renal response to sepsis probably reflect the net effect of several regulatory and counterregulatory mechanisms, many of which may function primarily to maintain an effective systemic circulation. Some hypothetical interactions are shown in Fig 7, and several of these should be testable experimentally. Our results suggest that reciprocal changes in the activities of the RAS and the renal KKS may be important in the renal response to sepsis; PGI 2 generation is increased, with possible systemic and renal effects; and TXA 2 could also contribute to the renal response to sepsis. ACKNOWLEDGMENT The authors thank Andrea Neal for technical assistance and Lyn MacLean for secretarial help. Allan D. Cumming was a Clinical Research Fellow of the Kidney Foundation of Canada.
REFERENCES 1. Werb R, Linton AL: Aetiology, diagnosis, treatment and prognosis of acute renal failure in an intensive care unit. Resuscitation 7:95-100, 1979 2. Wardle N: Acute renal failure in the 1980s; the importance of septic shock and of endotoxernia. Nephron 30:193200, 1982 3. Richmond 1M, Walker IF, Avila A, et al: Renal and cardiovascular response to nonhypotensive sepsis in a large animal model with peritonitis. Surgery 97:205-213, 1985
RENAL RESPONSE TO SEPSIS 4. Walker JF, Cumming AD, Lindsay RM, et al: The renal response produced by non-hypotensive sepsis in a large animal model. Am J Kidney Dis 8:88-97, 1986 5. Thurau K, Vogt C, Dahlheim H: Renin activity in the juxtaglomerular apparatus of the rat kidney during post-ischemic acute renal failure. Kidney Int 1O:S177-S182, 1976 6. Oken DE: Hypothesis on the role of prostaglandins in the pathogenesis of acute renal failure. Lancet 1:1319-1322, 1975 7. Butler RR Jr, Wise WC, Halushka PV, et al: Thromboxane and prostacyclin production during septic shock. Adv Shock Res 7:133-145, 1982 8. Gerber JG, Anderson RJ, Schrier RW, et al: Prostaglandins and the regulation of renal circulation and function. Adv Prostaglandin Thromboxane Leukotriene Res 10:227-254, 1982 9. Patrono C, Ciabattoni G, Patrignani P, et al: Evidence for a renal origin of urinary thromboxane B2 in health and disease. Adv Prostaglandin Thromboxane Leukotriene Res 11 :493-498, 1983 10. Levinsky NG: The renal kallikrein-kinin system. Circ Res 44:441-451, 1979 11. Mills IH: The renal kallikrein-kinin system and sodium excretion. Q J Exp Physiol 67:393-399, 1982 12. Abe K, Sato M, Imai Y, et al: Renal kallikrein-kinin: Its relation to renal prostaglandins and renin-angiotensinaldosterone in man. Kidney Int 19:869-880, 1981 13. Wichterman KA, Baur AE, Chaudry IH: Sepsis and septic shock-A review of laboratory models and a proposal. J Surg Res 29:189-201, 1980 14. Haber E, Koerner T, Paige LB, et al: Application of a radioimmunoassay for angiotensin I to the physiologic measurement of plasma renin activity in normal human subjects. J Clin Endocrinol Metab 29: 1349-1355, 1969 15. Amundsen E, Putter J, Friberger P, et al: Methods for the determination of glandular kallikrein by means of a chromogenic tripeptide substrate. Adv Exp Med Bioi 120A:83-95, 1979 16. Ali M, McDonald JWD: Synthesis of thromboxane B2 and 6-keto-prostaglandin Fla by bovine gastric mucosal and muscle microsomes. Prostaglandins 20:245-254, 1980 17. Ali M, Zamecnik J, Cerskus AL, et a1: Synthesis of thromboxane B2 and prostaglandins by bovine gastric mucosal microsomes. Prostaglandins 14:819, 1977 18. International Committee for Standardization in Hematology: Standard technique for the estimation of red cell and plasma volume. Br J HaematoI25:801-814, 1973 19. Dubofsky EV, Russell CD: Quantitation of renal function with glomerular and tubular agents. Semin Nucl Med 12:308-329, 1982 20. Bianchi C, Bonadio M, Donadio C, et a1: Measurement of glomerular filtration rate in man using DTPA_ 99mTc. Nephron 24:174-178, 1979 21. Solez K, Ideura T, Silvia CB, et a1: Clonidine after renal ischemia to lessen acute renal failure and microvascular damage. Kidney Int 18:309-322, 1980 22. Hermreck AS, Berg RA, Ruhlen JR, et a1: Renal response in sepsis. Arch Surg 107:169-175, 1973 23. Vaz AJ: Low fractional excretion of urine sodium in acute renal failure due to sepsis. Arch Intern Med 143:738739, 1983 24. Hinshaw LB, Benjamin B, Holmes DD, et al: Responses
31
of the baboon to live E. coli organisms and endotoxins. Surg Gynecol Obstet 145:1-11,1977 25. Siegel JH, Cirra FB, Coleman B, et al: Physological and metabolic correlations in human sepsis. Surgery 86:163-193, 1979 26. Maier M, StarIinger M, Wagner M, et al: The effect of hemorrhagic hypotension on urinary kallikrein excretion, renin activity, and renal cortical blood flow in the pig. Circ Res 48:386-392, 1981 27. Schnermann J, Briggs JP: Participation of renal cortical prostaglandins in the regulation of glomerular filtration rate. Kidney Int 19:802-815, 1981 28. Cronenwett JL, Lindenauer SM: Hemodynamic effects of cecal ligation sepsis in dogs. J Surg Res 33:324-331, 1982 29. Reubi F, Vorburger C: Renal hemodynamics in acute renal failure after shock in man. Kidney Int 1O:S137-SI43, 1976 30. Westwick J, Fletcher MS, Kakkar VV: Inhibition of thromboxane formation prevents endotoxin-induced renal fibrin deposition in jaundiced rats. Adv Prostaglandin Thromboxane Leukotriene Res 12:83-91, 1983 31. Blaine EH, Zimmerman MB: Renin secretion during dynamic changes in renal perfusion pressure. Am J Physiol 236:F546-F551, 1979 32. Sugerman HJ, Austin G, Newsome HH, et al: Hemodynamics, oxygen consumption and serum catecholamine changes in progressive, lethal peritonitis in the dog. Surg Gynecol Obstet 154:8-12, 1982 33. Henrich WL, Schrier RW, BerI T: Mechanisms of renin secretion during hemorrhage in the dog. J Clin Invest 64:1-7, 1969 34. Baylis C, Ichikawa I, Willis WT, et al: Dynamics of glomerular ultrafiltration. IX. Effects of plasma protein concentration. Am J PhysioI232:F58-F71, 1977 35. Mullane KM, Moncada S: Prostacyclin release and the modulation of some vasoactive hormones. Prostaglandins 20:25-49, 1980 36. Hall JE, Guyton AC, Jackson TE, et al: Control of glomerular filtration rate by renin-angiotensin system. Am J Physiol 233:F366-F372, 1977 37. Ishikawa I, Hollenberg NK: Pharmacologic interruption of the renin-angiotensin system in myohemoglobinuric acute renal failure. Kidney Int 1O:S183-S190, 1976 38. Levens NR, Peach MJ, Carey RM: Role of the intrarenal renin-angiotensin system in the control of renal function. Circ Res 48:157-167, 1981 39. Brown JJ, Gleadle RI, Lawson DH, et al: Renin and acute renal failure: Studies in man. Br Med J 1:253-258, 1970 40. Oken DE, Cotes SC, Flamenbaum W, et al: Active and passive immunization to angiotensin in experimental acute renal failure. Kidney Int 7: 12-18, 1985 41. Henrich WL, BerI T, McDonald KM, et al: Angiotensin II, renal nerves and prostaglandins in renal hemodynamics during hemorrhage. Am J Physiol 235:F46-F51, 1978 42. Levy SB, Lilley JJ, Frigon RP, et al: Urinary kallikrein and plasma renin activity as determinants of renal blood flow. J Clin Invest 60: 129-138, 1977 43. Pisano JJ: Observations on the kallikrein-kinin system in the kidney. Adv Exp Med Bioi 156B:929-938, 1983 44. Margolius HS, Horwitz D, Pisano JJ, et a1: Urinary kallikrein excretion in hypertensive man. Relationships to sodium
32 intake and sodium retaining steroids. Circ Res 35:820-825, 1974 45. Watson ML, Cumming AD, Lambie AT, et al: Urinary kallilkrein and systemic protacyc1in synthesis during sodium chloride infusion in normal man. Clin Sci 68:537-543, 1985 46. Olsen UB: Changes of urinary kallikrein and kinin excretions induced by adrenaline infusion in conscious dogs. Scand J Clin Lab Invest 40: 173-178, 1980 47. Albertini R, Ladron DE, Guevara R, et al: Effect of renal nerve stimulation on urine and tissue kininogenase activity in cats. Hypertension 3:50-54, 1981 48. Stone AM, Stein T, Lafortune J, et al: Changes in intrarenal blood flow during sepsis. Surg Gynecol Obstet 148:731-734, 1979 49. Rosenkranz B, Kitajima W, Frolich JC: Relevance of urinary 6-keto-prostaglandin Fla determination. Kidney Int 19:755-759, 1981 50. Shebuski RJ, Aiken JW: Angiotensin II-induced renal prostacyclin release suppresses platelet aggregation in the anesthetized dog. Adv Prostaglandin Thromboxane Res 7:11491152, 1980 51. Vandonger R, O'Dwyer J, Barden A: Effect of alphaadrenergic stimulation on prostacyclin and renin release in the rat kidney. Prostaglandins 24:815-825, 1982 52. O'Donnell TF, Clowes GHA, Talamo RC, et al: Kinin activation in the blood of patients with sepsis. Surg Gynecol Obstet 143:538-545, 1976 53. Gullner HF,' Graf AK, Gill JR, et al: Hypokalemia stimulates prostaglandin synthesis in the rat. Clin Sci 65:43-46, 1983
CUMMING ET AL 54. Ritter JM, Blair lA, Barrow SE, et al: Release of prostacyclin in vivo and its role in man. Lancet 1:317-319, 1983 55. Harris RH, Zmudka M, Maddox Y, et al: Relationships of TXB z and 6-keto PGFla to the hemodynamic changes during baboon endotoxic shock. Adv Prostaglandin Thromboxane Res 7:843-849, 1980 56. Regoli D, Barabe J: Pharmacology of bradykinin and related kinins. Pharmacol Rev 32: 1-46, 1980 57. Zipser RD, Kronborg I, Rector W, et al: Therapeutic trial of thromboxane synthesis inhibition in the hepatorenal syndrome. Gastroenterology 87: 1228-1232, 1984 58. Winn R, Harlan J, Nadir B, et al: Thromboxane A z mediates lung vasoconstriction but not permeability after endotoxin. J Clin Invest 72:911-918, 1983 59. Mene P, Simonson M, Rettberg C, et al: Thromboxane A z and endoperoxide analogs contract cultured rat glomerular mesangial cells. Am Soc Nephrol 191A, 1985 (abstr) 60. Wilson MF, Brackett DJ, Tompkins P, et al: Elevated plasma vasopressin concentrations during endotoxin and E. coli shock. Adv Shock Res 6: 16-26, 1981 61. Lanier SM, Malik KU: Attenuation by prostaglandins of the facilitatory effect of angiotensin II at the adrenergic prejunctional sites in the isolated Krebs-perfused rat heart. Circ Res 51:594-601, 1982 62. Gregg CM, Cohen JJ, Black AH, et al: Effects of glucose and insulin on metabolism and function of perfused rat kidney. Am J Physiol 235:F52-F61, 1978 63. Barnes J, Osgood RW, Pinckard RN, et al: The effect of complement depletion on the post-ischemic model of acute renal failure. Kidney Int 14:721, 1978