Clinical Science(l986)71,743-747
743
Prostacyclin in the circulation of patients with vascular disorders
undergoing surgery
J. M. RITTER, G. HAMILTON*, S. E. BARROW, D. J. HEAVEY, N. E. HICKLING,
K. M. TAYLOR, K. E. F. HOBBS* AND C. T. DOLLERY
Departments of Clinical Pharmacology and Surgery, Royal Postgraduate Medical School, London and *Academic
Department of Surgery, Royal Free Hospital, London
(Received 18 June 1986; accepted 16 July 1986)
Summary
1. The object of this study was to investigate
clinical conditions in which increased Droduction of
prostacyclin
has been reported* 6-ox0prostaglandin F,, (6-oxo-PGF1,)is the stable hydrolysis product of PGI, and was measured in plasma
frompatients undergoing hepatic or cardiac surgery
and in unoperated patients with
and
hepatic disease, using gas chromatography/mass
soectrometrv.
' 2. Blood- obtained simultaneously from portal
and peripheral veins, during emergency surgery for
bleeding oesophageal varices in six patients with
cirrhosis of the liver, contained very high concentrations of 6-0x0-PGFIa(range 99-11485 pg/ml of
plasma). 6-Oxo-PGFlawas higher in portal than in
peripheral blood in five out of six patients.
3. Six unoperated patients with cirrhosis and
oesophageal varices which were not bleeding all had
normal peripheral plasma concentrations of 6-0x0PGF,, < 2 p g / d (normal up to 5 pg/ml).
4. Seventeen patients with severe vascular
disease had normal basal plasma 6-oxo-PGF1, concentrations ( < 2 p g / d ) .
5. Eighteen subjects with atheromatous coronary artery disease underwent aorta-coronary
artery grafting, and plasma concentrations of 6-0x0PGF,, were markedly elevated during surgery
(range 55-1207 pg/ml).
6. We conclude that surgery stimulates PGI, production substantially, and argue that the function of
PGI, may be to limit intravascular extension of
thrombus from sites of haemostasis. Inappropriate
PGI, synthesis may contribute to the massive
haemorrhage characteristic of oesophageal variceal
bleeding.
Key words: atheroma, portal hypertension, prostacyclin, surgery.
Abbreviations: GC/NICIMS, gas chromatography/
negative ion chemical ionization mass spectrometry;
PG, prostaglandin; PGI,, prostacych.
Introduction
Prostacyclin (PGI,) relaxes vascular smooth muscle
and is the most potent known endogenous inhibitor
of platelet aggregation [I, 21. It activates platelet
adenylate cyclase [3,4]thereby antagonizing several
different pathways of platelet activation [5]. It is
synthesized at very high rates by vascular tissue in
vitro [6]but its concentration in circulating blood of
healthy subjects is much lower than that causing
systemic effects during PGI, infusion [7-91. We
found that intravascular trauma caused by catheters, distension and hypertonic solutions stimulated
local PGI, synthesis in vivo [9, lo], and have argued
that PGI, is a local hormone that may limit thrombus extension at sites of vascular injury.
An elevated concentration of PGI,, measured as
its hydrolysis product 6-0x0-PGF],, has been
reported in the blood of patients undergoing cardiopulmonary by-pass [11-131. These measurements
were made by radioimmunoassay, a method that has
many pitfalls and can overestimate grossly the true
concentration of 6-oxo-PGF1, in plasma [141. The
744
J. M. Ritter et al.
a urinary metabolite of PGI, has also been reported
in patients with extensive vascular disease [16].The
object of this study was to investigate clinical conditions in which increased synthesis of PGI, might
occur, using a highly specific and sensitive method
[7] based on gas chromatography/negative ion
chemical ionization mass spectrometry (GC/
NICIMS). Plasma concentrations of 6-oxo-PGF1a
were therefore measured in patients with portal
hypertension, patients undergoing cardiopulmonary
by-pass surgery and patients with extensive vascular
disease.
Methods
Subjects
None of the patients (ages 26-77 years) had
taken aspirin or other non-steroidal anti-inflammatory drugs in the 2 weeks before blood sampling.
Portal hypertension. All 12 patients with portal
hypertension and oesophageal varices had cirrhosis
of the liver. Six were not bleeding from their varices
at the time of the study, and were in hospital for
investigation of their liver disease. Six were actively
bleeding from their varices when studied.
Vascular disease. Seventeen patients were studied
under resting conditions. Nine of these had triple
vessel coronary artery disease, and were also
studied during cardiac surgery. Of the remaining
patients, four had accelerated hypertension, defined
by the presence of hypertensive haemorrhages and/
or cotton wool spots on fundoscopy with or without
proteinuria, two had extensive atheromatous
disease, one had proliferative diabetic retinopathy,
and one had Behcet's disease. Aortic aneurysms
were confirmed by ultrasound. Symptomatic
atheroma was diagnosed from the history and
physical examination confirmed by digital subtraction angiography or coronary angiography. Behcet's
disease was diagnosed on clinical features including
a febrile illness with mouth ulcers, erythema nodosum, multiple venous thromboses with pulmonary
embolism, arterial thrombosis causing cerebral
infarction and an abdominal aortic aneurysm. Proliferative retinopathy was confirmed by fluorescein
angiography.
An additional nine patients with triple vessel
coronary artery disease were studied during cardiac
surgery.
who were not actively bleeding. Peripheral venous
blood was also sampled from three of the patients
with variceal bleeding before surgery. Simultaneous
peripheral and portal samples were drawn immediately after opening the abdomen from all six
actively bleeding patients before any major handling of bowel.
Patients with vascular disease. Blood from two of
the patients with accelerated hypertension was
obtained before treatment and from the other two
shortly after sublingual nifedipine. A pair of blood
samples was obtained from each patient undergoing
cardiac surgery, one immediately before and one
during surgery. The two samples were taken
through the same peripheral venous cannula (nine
patients) or central venous cannula (nine patients).
The samples obtained during surgery were drawn
immediately on termination of pulsatile flow bypass.
Heparinized blood (20 ml) was centrifuged
immediately at 4"C, plasma was separated and 2 ng
of tetradeuterated 6-oxo-PGFIa added to an exact
volume (usually 5 ml), to act as an internal standard.
Subsequent extraction, derivatization and assay by
capillary column GC/NICIMS were performed as
described previously [7]. The limit of sensitivity of
the assay was 2 p g / d (1 pg/ml when a 10 ml
plasma sample was assayed).
Analysis
Paired data were compared by the Wilcoxon
signed-rank test, unpaired data by the rank-sum
test. Differences were considered significant when
a < 0.05.
Results
Portal hypertension
Blood sampling
All six unoperated patients with oesophageal
varices and portal hypertension due to hepatic
cirrhosis had normal peripheral plasma concentrations of 6-oxo-PGF1. ( < 2 pg/ml) (Table 1). In
contrast, very high concentrations were present in
both portal and peripheral blood sampled immediately after opening the abdomen (range
99-11485 pg/ml, a<0.05) in six patients with
variceal bleeding. Peripheral blood samples
obtained from three of these patients immediately
before operation, but after insertion of cannulae for
blood transfusion, contained moderately increased
6-oxo-PGF,. (25,26 and 31 p g / d ) .
Peripheral blood was sampled by venepuncture
of an antecubital vein.
Vascular disease
Patients with portal hypertension. Peripheral
venous blood was obtained from all stable patients
Plasma concentrations of 6-oxo-PGF, o1 were normal ( < 2 pg/ml) in 17 patients with various severe
Prostacyclin in surgery
745
TABLE1. Plasma 6-oxo-PGF1, in patients with oesophageal varices and portal hyper
tension due to cirrhosis with and without surgery
Patient Condition Peripheral venous 6-oxo-PGF,, (pg/ml) Peripheral venous Portal venous
~-oxo-PGF,,
~-oxo-PGF,,
No surgery*
Pre-surgery?
during surgery$ during surgery$
(pg/ml)
(pg/ml)
1
2
3
4
5
6
7
8
9
10
11
12
Stable
Stable
Stable
Stable
Stable
Stable
Bleeding
Bleeding
Bleeding
Bleeding
Bleeding
Bleeding
<2
<2
<2
<2
<2
<2
-
25
26
31
740
748
99
1842
440
137
796
1820
41 1
11485
177
217
*Blood sampled from stable uninstrumented patients with oesophageal varices.
?Blood sampled from patients with variceal bleeding before surgery, but after insertion of intravascular catheters and blood transfusion.
$Blood obtained simultaneouslyfrom portal and peripheral veins immediately after opening the
peritoneal cavity, and after minimal handling of bowel and peritoneum.
TABLE2. Plasma 6-oxo-PGF1,in vascular disease
Patient
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Diagnosis
Accelerated hypertension*
Accelerated hypertension*
Accelerated hypertension
Accelerated hypertension
Aortic, renal and femoral
atheroma
Aortic aneurysm, femoral
atheroma
Aortic aneurysm, Behcet’s
disease?
Proliferative diabetic retinopathy
Coronary artery atheroma
Coronary artery atheroma
Coronary artery atheroma
Coronary artery atheroma
Coronary artery atheroma
Coronary artery atheroma
Coronary artery atheroma
Coronary artery atheroma
Coronary artery atheroma
Plasma
6-0xo-PGFl
(pg/ml)
(I
<2
<2
<2
<2
(2
<2
<2,2
<1
(2
<2
<2
(2
<2
(2
<2
<2
<2
*Blood sampled after emergency treatment with sublingual nifedipine.
?Two blood samples from the same patient: one (2 pg/
ml) at presentation with venous and arterial vasculitis, on
no treatment; the other ( < 2 pg/ml) when an abdominal
aortic aneurysm had developed 1 month later while
receiving prednisolone 40 mg daily.
vascular disorders under resting conditions (Table
2). Eighteen patients with triple vessel coronary
artery disease were studied before and during cardiac surgery. Peripheral venous blood was obtained
from nine of them and blood from a central venous
cannula in the remaining nine. In all cases, the basal
peripheral venous blood contained < 2 pg of
6-oxo-PGF1,/ml (Table 2). The basal central
venous blood contained higher concentrations of
6-oxo-PGF1, than these basal peripheral blood
samples in all cases: 6-24 pg/ml of plasma in eight
of nine subjects and an outlying higher value of 124
pg/ml in the remaining subject.In all 18 patients the
plasma concentration of 6-oxo-PGF1, during surgery was higher than the basal value (a<
(Fig. 1).
Discussion
The most striking feature of these observations is
the very large increase in plasma 6-oxo-PGF1, in
response to surgery. Concentrations of 6-0x0PGF,, of 55 pg/ml-1.8 ng/ml were present in
venous plasma during surgery. This compares with
human venous plasma concentrations of 291-5 10
pg/ml during infusion of prostacyclin at 8 ng min-’
kg-’ [9], a dose that causes marked haemodynamic
and platelet effects in man [9].In five of six patients
undergoing surgery for bleeding varices, the plasma
concentrations of 6-oxo-PGF1, were higher in the
portal than in the peripheral venous sample, despite
minimal handling of the bowel before sampling
(Table 1).This confirms a previous report [17] and
suggests that some of the PGI, could be derived
either from the portal vascular tree, or from visceral
organs which can synthesize PGI,, such as the
stomach [18]. The question of whether there is a
small local increase of PGI, in portal venous blood
J. M.Ritter et al.
746
I2O0
1100
1
/
1000
900
z.
M
a
800
700
v
I
5
%
$
600
500
W
400
300
200
100
0
Baseline
Inter-op.
FIG. 1. Plasma concentrations of 6-0x0-PGF,
before and during cardiac surgery (Inter-op.) in 18
patients undergoing coronary artery by-pass grafting.
of unoperated patients with portal hypertension
remains open. It is attractive to speculate that such
an increase, if present, could contribute to the
increased splanchnic blood flow of portal hypertension, to the development of portal-systemic
collaterals and perhaps to the severity of variceal
haemorrhage, which is often massive [ 151. However,
if increased PGI, is present in the portal venous
blood of unoperated patients with portal hypertension, it is not reflected in an increase in 6-0x0PGF,, in peripheral venous blood (Table 1).Indeed,
concentrations of 6-oxo-PGFla in plasma of unoperated patients with cirrhosis and oesophageal
varices are normal ( < 2 pg/ml). This finding contrasts with a previous report of 6-oxo-PGF,, concentrations of around 1 ng/ml in preoperative
peripheral venous plasma of patients with portal
hypertension measured by radioimmunoassay [ 171.
The > 500-fold difference reflects the specificity of
the GClNICIMS assay and re-emphasizes the unreliability of radioimmunoassay of 6-oxo-PGF1, in
plasma [14].
We did not detect increased 6-0x0-PGF,, in the
blood of patients with severe vascular disease
(Table 2). FitzGerald et al. [16]inferred that there is
increased PGI, synthesis in such patients, from the
observation that, in subjects with severe atheroma,
excretion of the major urinary metabolite of PGI, is
increased approximately fourfold. Since the plasma
concentration of 6-oxo-PGFla in healthy subjects is
frequently below the detection limit of our assay [7],
it is perhaps not surprising that we failed to detect
an increase in the patients with vascular disease.
Patients with accelerated hypertension were
included because of reports [19, 201 that PGI,
synthesis is increased in the blood vessels of hypertensive rats. However, we did not find an increase in
concentration of 6-oxo-PGF1, in the peripheral
blood of these subjects.
We did, however, find a very marked increase in
plasma 6-0x0-PGF,, concentrations in patients
coming off by-pass after coronary artery surgery
(Fig. l ) , in agreement with earlier reports [ll-131. It
is of interest that these patients with extensive
atheroma were able to produce large quantities of
PGI, in response to surgery. The very high concentrations synthesized in response to both thoracic
and abdominal surgery raise the question of what
function PGI, could have in sharp trauma where a
vasodilator/anti-aggegatory substance might be
expected to cause increased blood loss. Further, it is
possible that local production at sites of venous
rupture in patients with bleeding varices contributes
to the severity of this type of bleeding. However,
PGI, does not inhibit platelet adhesion to damaged
blood vessels at concentrations that prevent platelet
aggregation [211. In therapeutic use, including
cardiac surgery [22, 231, PGI, has been strikingly
free of bleeding complications [24]. We therefore
argue that the function of PGI, released locally at
sites of incision may be to prevent thrombus extension from points of initiation of haemostasis, and
that PGI, may also act systemically in this circumstance to oppose excessive vasoconstriction and
counter acutely the hypercoagulable state associated with surgery and trauma.
Acknowledgments
We thank Dr Eva Kohner for letting us study her
patient with diabetic retinopathy. B. Edinborough
provided excellent secretarial assistance. D.J.H. was
an MRC Training Fellow. The work was supported
by the Medical Research Council and British Heart
Foundation.
References
1. Moncada, S., Gryglewski, R., Bunting, S. & Vane, J.R.
(1976)An enzyme isolated from arteries transforms
Prostacyclin in surgery
prostaglandin endoperoxides to an unstable substance that inhibits platelet aggregation. Nature
(London),263,663-665.
2. Gryglewski, R.J., Bunting, S., Moncada, S.,Flower,
R.J. & Vane, J.R. (1976)Arterial walls are protected
against deposition of platelet thrombi by a substance
(prostaglandinX) which they make from prostaglandin endoperoxides. Prostaglandins, 12,685-708.
3. Tateson, J.E., Moncada, S. & Vane, J.R. (1977)
Effects of prostacyclin (PGX) on cyclic AMP concentrations in human platelets. Prostaglandins, 13,
389-397.
4. Gorman, R.R., Bunting, S. & Miller, O.V. (1977)
Modulation of human platelet adenylate cyclase by
prostacyclin (PGX).Prostaglandins, 13,377-388.
5. Moncada, S. & Vane, J.R. (1981) Discovery, biological significance,and therapeutic potential of prostacyclin. In: Clinical Pharmacology of Prostacyclin, pp.
1-8. Ed. Lewis, P.J. & OGrady, J. Raven Press, New
York.
6. Bunting, S., Grylewski, R., Moncada, S. & Vane, J.R.
(1976) Arterial walls generate from prostaglandin
endoperoxides a substance (prostaglandin X) which
relaxes strips of mesenteric and coeliac arteries and
inhibits platelet aggregation. Prostaglandins, 12,
897-913.
7. Blair, LA., Barrow, S.E., Waddell, K.A., Lewis, P.J. &
Dollery, C.T. (1982) Prostacyclin is not a circulating
hormone in man. Prostaglandins, 23,579-589.
8. Patrono, C., Pugliese, F., Ciabattoni, G., Patrignani,
P., Maseri, A., Chierchia, S., Paskar, B.A., Cinotti,
G.A., Simonetti, B.M. & Pierucci, A. (1982) Evidence for a direct stimulatory effect of prostacyclin
on renin release in man. Journal of Clinical Investigation, 69,231-239.
9. Dollery, C.T., Barrow, S.E., Blair, LA., Lewis, P.J.,
MacDemot, J., Orchard, M.A., Ritter, J.M., Robinson, C., Shepherd, G.L., Waddell, K.A. & Allison,
D.J. (1983) Role of prostacyclin. h Atherosclerosis:
Mechanisms and Approaches to Therapy, pp.
105-123. Ed. Miller, N.E. Raven Press, New York.
10. Ritter, J.M., Barrow, S.E., Blair, LA. & Dollery, C.T.
(1983)Release of prostacyclin in vivo and its role in
man. Lancet, i, 317-319.
11. Watkins, W.D., Peterson, M.B., Kong, D.L., Kono, K.,
Buckley, M.J., Levine, F.H. & Philbm, D.M. (1982)
Thromboxane and prostacyclin changes during
cardiopulmonary bypass with and without pulsatile
flow. Journal of Thoracic and Cardiovascular Surgeiy,
84,250-256.
12. Ylikorkala, O., Saarela, E. & Viinikka, L. (1981)
Increased prostacyclin and thromboxane production
747
in man during cardiopulmonary bypass. Journal of
Thoracic and Cardiovascular Sueery, 82,245-251.
13. Uotila, P., Suves, M., Heikkila, H. & Jalonen, J.
(1984)Prostanoids and hemodvnamicsin man before
Ad during cardiopulmonary dypass. Prostaglandins,
28,497-508.
14. Morris, H.G., Sherman, N.A. & Shepperdson, F.T.
(1981) Variables associated with radioimmunoassay
of prostaglandins in plasma. Prostaglandins, 21,
771-788.
15. Hamilton, G., Phing, R.C.F., Hutton, R.A., Dandona,
P. & Hobbs, K.E.F. (1982)The relationship between
prostacyclin activity and pressure in the portal vein.
Hepatology, 2,236-242.
16. FitzGerald, G.A., Smith, B., Pedersen, A.K. & Brash,
A.R. (1984) Increased prostacyclin biosynthesis in
patients with severe atherosclerosis and platelet
activation. New England Journal of Medicine, 310,
1065-1068.
17. Hashizume, M., Inokuchi, K. & Tanaka, K. (1985)
The role of prostacyclin in patients with portal
hypertension. Liver, 5,89-93.
18. Whittle, B. (1981) Temporal relationship between
cyclo-oxygenase inhibition, as measured by prostacyclin biosynthesis, and the gastrointestinal damage
induced by indomethacinin the rat. Gastroenterology,
80,94-98.
19. Pace-Asciak, C.R., Carrara, M.C., Rangaraj, G. &
Nicolaou, K.C. (1978) Enhanced formation of PGI,,
a potent hypotensive substance, by aortic rings and
homogenates of the spontaneously hypertensive rat.
Prostaglandins, 15,1005-1012.
20. Botha, J.H., Leary, W.P. & Asmal, A.C. (1980)
Enhanced release of a ‘prostacyclin-like’substance
from aortic strips of spontaneously hypertensive rats.
Prostaglandins, 19,285-290,
21. Higgs, E.A., Moncada, S., Vane, J.R., Caen, J.P.,
Michel, H. & Tobelem, G. (1978) Effect of prostacyclin (PGI,) on platelet adhesion to rabbit arterial
subendothelium. Prostaglandins, 16,17-22.
22. Longmore, D.B., Bennett, J.G., Hoyle, P.M., Smith,
M.A., Gregory, A., Osivand, T. & Jones, W.A. (1981)
Prostacyclin administration during cardiopulmonary
bypass in man. Lancet, i, 800-804.
23. Faichney, A., Davidson, K.G., Wheatley, D.J., Davidson, J.F. & Walker, I.D. (1982)Prostacyclinin cardiopulmonary bypass operations. Journal of Thoracic
and Cardiovascular Surgery, 84,601-608.
24. Pickles, H. & OGrady, J. (1982) Side effects occurring during administration of epoprostenol (prostacyclin, PGI,), in man. British Journal of Clinical
Pharmacology, 14,177- 185.