thrombocytopenia post liver transplantation

16 Journal of Hepatology, 1992; 16:16-22 01992 Elsevier Scientific Publishers Ireland Ltd. All rights reserved. 0168-8278/92/$05.00

HEPAT 01201

Thrombocytopenia post liver transplantation

Correlations with pre-operative platelet count, blood transfusion requirements, allograft function and outcome

G.W. McCaughan", R. Herkes b, B. Powers a, K. Rickard c, N.D. Gallagher a, J.F. Thompson d and A.G.R. Sheil d

"A. W. Morrow Gastroenterology and Liver Centre, blntensive Care Unit, CHaematology Department and dDepartment of Transplant Surgery, Royal Prince Alfred Hospital and University of Sydney, NSW, Australia

(Received 12 December 1990)

This study reports that thrombocytopenia is a universal phenomenon post hepatic transplantation. In 53 consecutive adult patients undergoing liver transplantation the platelet count fell by a mean of 63% (157 x 109/1 to 50 x 109/1). The platelet count reached a nadir at Day 5 post-transplant but returned to pre-operative levels by Day 14. Non-parametric regression analysis found that pre-operative platelet count, blood transfusion requirements and maximum post- operative ALT values were independent predictors of the percentage fall in platelet count. No correlation was seen with length of graft cold ischaemic time or the use of University of Wisconsin (UW) solution. The nadir day correlated with maximum post-operative bilirubin and ALT, graft ischaemic time and use of UW solution. Maximum post- operative ALT was also an independent predictor of nadir platelet count. It was observed that patients who did not survive the hospital admission had lower post-operative platelet counts and these did not return to pre-operative levels by Day 14. The percentage fall in platelet count was an iridependent predictor of survival. Severe thrombocytopenia was associated with cerebral haemorrhage in 3 patients. This report provides evidence that allograft dysfunction (maximum post-operative bilirubin and/or AST/ALT) was the most consistent independent predictor of the nadir platelet count, nadir day and percentage fall in platelet count post liver transplantation although the exact mechanism(s) of the platelet changes remain uncertain.

Key words: Thrombocytopenia; Liver; Transplantation

Although thrombocytopenia is a recognised post- operative complication following hepatic transplantation (1,2) there has been only a limited analysis of this phenomenon (3-5). Thrombocytopenia post liver transplant may contribute to morbidity and mortality by (i) exacerbating continued post-surgical bleeding, (ii) in conjunction with post-operative hypertension increasing the risk of cerebral haemorrhage and (iii) preventing the use of percutaneous liver biopsy in the early post-operative period to diagnose acute allograft rejection.

In this report we have undertaken a study to look at

the extent and time-course of thrombocytopenia in adult patients undergoing liver transplantation. Regression analysis was used to try and identify major contributing factors to the thrombocytopenia and also to examine the association between thrombocytopenia and short- term outcome.

Materials and Methods

Over the 4-year period 1986-1989 54 adult patients underwent hepatic transplantation. The diseases for which these patients were transplanted were chronic

Correspondence to: Dr. G. W. McCaughan, A. W. Morrow Gastroenterology and Liver Centre, Royal Prince Alfred Hospital, Missenden Road, Camperdown, NSW 2050, Australia.

THROMBOCYTOPENIA POST LIVER TRANSPLANTATION 17

active hepatitis/cryptogenic cirrhosis in 16 patients, primary sclerosing cholangitis in 10 patients, primary biliary cirrhosis in 11 patients, metabolic disorders in 7 patients and other diseases in 10 patients. The details on one patient have been lost, leaving 53 patients for analysis. Immunosuppressive therapy was based on triple therapy (corticosteroids, imuran I mg/kg/day, cyclosporin initially 12 mg/kg/day). Allografts for Patients 1-34 were stored in routine preservation fluid (6) whilst allografts for Patients 35-54 were stored in University of Wisconsin (UW) solution (7). The time- course of the thrombocytopenia was examined with daily platelet counts over the first two weeks following transplantation. During this time many patients had platelet transfusions because of thrombocytopenia. Intra-operative and post-operative platelet transfusions were given usually when there was significant thrombocytopenia (<50 x 109/1) and/or persistent bleeding. The lowest platelet count on each day was used for analysis. The nadir day was defined as the post-operative day at the time of the lowest platelet count. Intra-operative platelet transfusion and red blood cell transfusions as well as post-operative platelet transfusion were recorded. Analysis was undertaken between nadir platelet count, percentage fall in platelet count, nadir day, pre-operative platelet count, blood and platelet transfusion requirements, graft ischaemic time, use of UW solution, maximum bilirubin in the first post-operative week, maximum alanine aminotransferase (ALT) and aspartate aminotransferase (AST) levels in the first post-operative week and short-term survival. Short- term survival was defined as those patients surviving the initial hospital admission. Initial analysis of the data used Spearman's rank correlation tests. An analysis of independent predictors of the nadir platelet count, nadir day and percentage fall in platelet count was undertaken using several methods. The nadir platelet count was normalised using a square root function and analysed with linear regression. The nadir day was analysed using Poisson regression analysis. Percentage fall in platelet count was normally distributed and was analysed with stepwise linear regression. Logistic regression was used to determine the independent predictors of survival. These analyses were performed with the NCSS Programme and SPIDA.

Results

Time-course and degree of thrombocytopenia As can be seen from Fig. 1 there was a dramatic fall

in the mean platelet count following liver transplanta-

=-.. O

¢-

_o 0 .

c o

250

200

150

100

50

0 = 0

* p<0.0OS.

- - All Patients (54).. Non-Survivors (11).

I I I l I I I

2 4 6 8 10 12 14 Days Post Liver Transplantation.

Fig. 1. Time-course of thrombocytopenia post hepatic transplantation. The data represent mean + S.E.

tion. This fell from 157 x 109/1 pre-operatively to a nadir of 50 x 109/1 on Day 5. In the patient group taken as a whole the mean platelet count had returned to approxi- mate pre-operative levels (174 x 109/1) by Day 14. Only 3 patients had a nadir platelet count of >100 x109/1 (101 x109/1=20% fall, 101 x109/1=55% fall and 234 x 109/1= 60% fall). The minimum percentage fall was 20% with a maximum fall of 98%. A comparison of the platelet fall in non-survivors vs. the patient group as a whole was also undertaken. This is seen in Fig. 1 and Table 3. The causes of death in the 11 non-survivors were cerebral haemorrhage (3 patients), persisting sepsis (3 patients), uncontrolled rejection (3 patients), recurrent hepatitis B (1 patient) and cerebral anoxia (1 patient). The data indicate that although the pre-operative platelet count was similar in this group of patients the nadir platelet count was lower (23 x 109/1) and tended to peak at Day 6 rather than Day 5 (not statistically significant). It is also noted that unlike the patient group as a whole, non-survivors had only a small rise in their platelet count following the nadir at Day 6 so that by Day 14 the platelet count was 88 x 109/1 in non-survivors versus 174x109/1 in the patient group as a whole (p<0.01). During this time 3 patients died from intracerebral haemorrhage. The fall in platelet count in these patients were 450x109/1 to 9x109/1 at Day 5, 135x109/1 to 25 x 109/1 at Day 7, 183 x 109/1 to 28 x 109/1 at Day 7. All 3 patients were hypertensive at the time of cerebral haemorrhage. Four patients required re-operation to stop post-operative intra-abdominal haemorrhage. The

18

mean nadir platelet count in these 4 patients was no different to the group as a whole.

Correlation analysis Non-parametric regression analysis was undertaken

between various factors which were thought likely to influence the time-course and degree of thrombocytopenia. These analyses are shown in Table 1 and Figs. 2-4. Independent predictors of the nadir platelet count were the pre-operative platelet count (p < 0.0001), total platelet transfusion (p<0.0001) and maximum post- operative ALT (p<0.001) (Table 2). This analysis was by stepwise linear regression analysis of the square root of the nadir platelet count. The nadir day was analysed using two models. Maximum post-operative bilirubin and AST emerged as independent predictors in both models (bilirubin p<0.0001 and AST <0.05, stepwise regression; bilirubin p<0.02, AST <0.05, exhaustive search analysis on Poisson regression analysis). With the use of exhaustive search analysis, graft ischaemic time (p <0.01) and use of UW solution were additional independent predictors (p<0.01). We have included these data as it is indirect evidence for the role of initial graft injury in some of the changes we are observing. Graft ischaemic time was inversely correlated to nadir day. Stepwise analysis also included total platelet transfusion requirements as an independent predictor of nadir day (p<0.05). An analysis of independent predictors for percentage fall in platelet count revealed pre-operative platelet count (p<0.02), nadir day (p<0.05), blood transfusion requirements (p<0.02) and maximum ALT (p <0.05) to be significant.

An analysis of these factors in the non-surviving vs. the surviving group revealed there was a statistically significant difference between the nadir platelet count, percentage fall, Day 14 platelet count and blood transfusion requirements in survivors versus non-survivors (Table 3). There was no correlation between the pre- operative platelet count or the nadir day with survival. Stepwise logistic regression analysis of short-term sur-

G.W. M c C A U G H A N et al.

TABLE 2

Independent predictors of the nadir platelet count ~

Name Estimate S.E. p-Value

Constant 6.91 0.42 < 0.0001 Preoperative 0.0092 0.0017 < 0.0001

platelet count Total platelet -0 .0411 0.0069 < 0.0001

transfusion ALT - 0.0004 0.0001 < 0.001

Analysis by stepwise linear regression analysis of the square root of the nadir platelet count. Degrees of freedom =49; Model R2 = 0.63.

TABLE 3

Data in survivors and non-survivors

Survivors Non-survivors p-Value a

Nadir platelet count (I 09/dl) 57 +- 5 b 23 _+ 4 0.0001

% Fall 57 +- 2 85 _ 2 0.0000 c Platelet count Day 14 189+_13 88+_17 0.0021 Preplatelet count 152 _+ 18 179 +- 29 N.S. Nadir day 4.04 +- 0.38 5.9 _+ 1.2 N.S. Blood transfusion

(units) 18.7 +- 1.7 35.8 +_ 8.5 0.0268

"Analysis was by Mann-Whi tney two-sample test. b The data represent mean + S.E. ° Stepwise logistic regression analysis indicated that the % fall was the

only independent predictor of short- term survival. N.S. = not significant.

vival showed that only the percentage fall of platelet count was an independent predictor.

Discussion

This paper indicates that post-operative thrombocytopenia is almost a universal phenomenon following human liver transplantation with an average fall in the platelet count of 63% that reached a peak on the fifth post-operative day. These findings are very similar to those already published although the time-course of recovery was substantially quicker in our group of patients (3,4).

TABLE I

Correlates with platelet changes

Platelet Pre-op. Blood transf. Total Nadir Graft Max imum Max imum Max imum platelet requirements a platelet day a ischaemic bilirubin a ALl" ~ AST" count transfusion time"

Post-operative platelet count 0.38 ~ - 0 . 3 5 c - 0 . 6 3 ~ - -0 .22 a - 0 . 0 3 d - 0 . 3 6 ~ - -0 .27 b 0.25 d

Percentage fall 0.44 c 0.17 a 0.13 a 0.39 d - 0 . 1 5 d 0.40 ~ 0.29 b 0.34 b Nadir day 0.I 7 d 0.0 ld - 0.17 a - - - 0.03 a 0.29 b r = 0.27 b 0.25 a

a Values shown indicate R-values derived from regression analysis (Spearman's rank correlation); b p < 0.05; C p < 0.01; d Not significant.


i=i '30 IOO

30

o o o o

° oo

i i i l

230

o

0 i | I I

100

0 100 200 300 400 500 600 700 800

Pro-Ol~lrotive Plotelet Count (lOS/I).

O0

~ 80 g

o

_~ 60

.c

o u. 40

o

20 n

0 0

0 0 °o @ o o

O 0 0 0 ~) 0 0

0 0 0 0 0 0 0

~o~ @d~o

o @ 0 0

o o o

0

0

0 I I I I I I

0 1 00 200 ,300 400 500 600 Bilirubin (/unol/I)

:50

o v

100 c

50 o

o..

z 00 ~0 0 0

o o o o 50 0 "~ 0 ~. o o q , , ,

I

o ,o 2'0 ~o :o 5'0 6'0 ,o ~o ,o ,oo o Blood Tronsfused (units).

O oo

0

8o 6) 0°° o o o o

~ o ~ 60 ~0 0 0

~ o 0 clDo

o 0

20 0 0 0 J;-~'." J

0 * * ' * ' * I I I I

0 I0 20 30 40 50 60 70 80 90 100

Blood tronsfumKI (units).

50

- I00 c

i30 o . .,=

z 00

o 50 I

G .

o o

0

° o o ~ 0 0 0 0

0 0 ~ 0

% o ~@ ~ o ¢ o , , _ ~ , , 100 200 ~ 400 ~

Bilirubin ( # m o l / l ) .

Fig. 2. Graph showing post-operative nadir platelet count versus pre- operative platelet count (top), intra-operative blood transfusion requirements (middle) and maximum post-operative serum bilirubin

level (bottom).

100 0

~ 80

o -~ 6o o.

.e

'? 4O

~ 20

@ O

QII~@ c o 0 0 O0 0

0 %~ oo a:~o co ~o

o ~

0

~ o

I I l I I I I I

0 100 200 300 400 500 600 700 800

Pre-0perotive Plotelet Count (10g/I).

Fig. 3. Graph showing percentage fall in the post-operative platelet count versus pre-operative platelet count (bottom), intra-operative blood transfusion requirements (middle) and maximum post-operative

serum bilirubin level (top).

20 G.W. McCAUGHAN et al.

c)

14

t2

1o

8

6

4

2

o o

0

0 0

o o o o

o ~) o

o ( 3oo (13o o

o o o o o

o o o o

o o o

( D O o o o CDO O

10 ' ' 80 2 40 60

0

0

o

Percentoge Foil in Plotelet Count.

o 0 0

0 o

I

1oo

14

12

1 0

8

6

4

2

0 0

(DO 0

O0 0 0

OlXmO 0 0 O0

000 0 0

(3O 0 CO

0 O0

CO ~ O0 0 0

I I I I

1 oo 2oo 3oo 400 Bilirubin ~OI/I) .

I

5O0 I

600

1

t2[

10 O D

8!

0 0 0

6 D 00

t o l e D O

4 q l ) O

D O

2 DO

m m O O

0 0 2O00

0

0

0

0

O O I | I I I

4000 6000 8000 10000 12000

ALT (unlts/I).

Fig. 4. Graph showing day of lowest post-operative platelet count (nadir day) vs. percentage fall in platelet count Itop), maximum post- operative serum bilirubin level (middle) and maximum post-operative

ALl" level (bottom).

Our study is unique in that it has analysed various variables that may have affected the degree and time course of the thrombocytopenia. The data indicate that there are several imprecise correlations with the degree of thrombocytopenia in the post-operative period. Regression analysis in search of independent predictors revealed that allograft dysfunction (maximum bilirubin, ALT or AST) was a significant factor in determining the nadir platelet count and percentage fall in platelet count. Furthermore, in one model the length of graft ischaemic time negatively correlated with the nadir day. We have included this analysis as it supports the concept that initial graft injury may be important in some of the post-operative platelet changes. However, it should be noted that significance values between 0.05 and 0.01 may be suspect as multiple statistical analyses were undertaken. If p-values <0.01 are only considered then the percentage fall in platelet count and nadir platelet count only correlated with pre-op platelet count, maximum post-op bilirubin, maximum post-op ALT and total platelet transfusions. A recent report by Castaldo et al. is similar in design to our study but is only available in abstract form (8). In that report thrombocytopenia post transplant correlated with use of UW solution, pre-operative platelet count, haematocrit, serum albumin, duration of surgery and intra-operative transfusion requirements. In addition, post-operative thrombocytopenia correlated with ultimate survival.

The above analysis suggests that the function of the hepatic allograft is an important determinant of the degree and timing of post-transplant thrombocytopenia. Other evidence also suggests that the platelet changes are caused by consumption in the recently implanted liver allograft (3). However, splenic consumption of platelets may also be important and it is possible that the damaged platelets are consumed in the spleen as well as the allograft (9). It is of interest that in the single patient of our group that underwent splenectomy during the transplant operation, the platelet count still fell from a pre-operative level of 96 × 109/1 to 61 x 109/1 at Day 3 post-transplant. The correlation in our study with allograft ischaemic times and use of UW solution for preservation concerned only the nadir day and was only present using exhaustive search Poisson regression analysis, and therefore may not be statistically significant. This is in contrast with preliminary data from Castaldo et al. (8) and Williams and O'Grady (10) who both found a correlation between lower post-operative platelet counts and use of UW solution. It should be noted however that the 'non-UW' preservation solution


may differ between transplant units, therefore making

such comparisons difficult. Other workers have suggested that endotoxaemia may

be an important factor in causing thrombocytopenia post liver transplantation (5). In that study endotoxin

levels post-transplant correlated with the requirement for post-operative platelet transfusion (R=0.462). Thrombocytopenia may also have been due to impaired

production secondary to the use of imuran as an immunosuppressive agent in 40 of our 54 patients. However, bone marrow aspirates were performed on 10 patients in our study and these all indicated that there was marked megakaryocyte hyperplasia consistent with peripheral destruction and/or consumption. A further

intriguing clue to the possible aetiology of the thrombocytopenia comes from a study by Siemensma and col-

leagues (11). They showed a 50% fall in the platelet count in rats following four-fifths partial hepatectomy. Although there was evidence of platelet accumulation in

the regenerating liver the main mechanism was suppres- sion of entry of platelets into the blood. Perhaps such a mechanism could exist during recovery from other

forms of liver injury. Early studies had shown significant thrombocytopenia following massive liver resection in dogs and humans (12,13) and also following other forms

of major surgery (14). Thrombocytopenia in the post-operative period is not

simply an academic observation. The combination of thrombocytopenia with hypertension is a potential cata-

strophic event. Hypertension in the immediate post- operative period was common in our patients (over 60%) and in combination with thrombocytopenia led to the

fatal complication of cerebral haemorrhage in 3 patients. This occurred early in our programme and since then we have used nitroprusside in the Intensive Care Unit

to attain fine tuning of blood pressure control. Persistent thrombocytopenia may also be a contributing factor to post-operative haemorrhage and may prevent percutaneous liver biopsy. This is particularly the case when allograft dysfunction occurs early, and in some cases a steroid pulse is given without confirming rejection by liver biopsy. Such practice may influence morbidity and mortality.

Stepwise regression analysis of survival revealed only the percentage fall in platelet count rather than the nadir day or actual nadir count as independent predictors. Recently platelets have been identified as a source of

hepatocyte growth factor (15) and they may play a role in the recovery of the liver from various forms of injury.

However, it is more likely that the percentage fall in platelet count reflects multiple medical problems includ- ing poor graft function, rather than implying a direct

negative influence of the thrombocytopenia on outcome. Mechanisms for the correction of the thrombocyto-

penia are presently not available. However, there is some anecdotal experience that high dose intravenous y-globulin may help (4) and such therapy may be beneficial in autoimmune thrombocytopenia (16). A controlled trial of this therapy is therefore necessary to test its efficacy

in preventing or ameliorating thrombocytopenia post

liver transplant.

Acknowledgements

We would like to thank Prof. P. Bhathal (University of Melbourne) for helpful discussion, Mike Jones (Royal

North Shore Hospital) for statistical advice and Ms. Wendy Macpherson and Rebecca de Jesus for help in

preparing the manuscript.

References

1 Carithers RL1, Fairman P, Mendez-Pilon G, et al. In: Maddrey WC, ed. Post-operative Care in Transplantation of the Liver. New York: Elsevier Science Publishing Co. Inc., 1988; I 11-42.

2 Hutchison DE, Genton E, Porter KA, et al. Platelet changes following clinical and experimental hepatic homotransplantation. Arch Surg 1968; 97: 27.

3 Plevak D J, Halma GA, Forstran LA, et al. Thrombocytopenia after liver transplantation. Trans Proc 1988; 20: Suppl. 630.

4 Munoz S J, Carabasi AR, Moritz M J, Jarrell BE, Maddrey WC. Post operative thrombocytopenia in liver transplant recipients: prognostic implications and treatment with high dose of gammaglobulin. Trans Proc 1989; 21: 3545.

5 Miyata T, Yokoyama I, Todo S, Tzakis A, Selby R, Starzl TE. Endotoxaemia, pulmonary complications and thrombocytopenia in liver transplantation. Lancet 1989; ii: 189.

6 Sheil AGR, Thompson JF, Gallagher ND, et al. Initial report of

the Australian National Pilot Liver Transplantation Programme. Med J Aust 1987; 147: 372.

7 Kalayoglu M, Sollinger HW, Stratta RJ, et al. Extended preservation of the liver for clinical transplantation. Lancet 1988; ii: 617.

8 Castaldo P, Grazi G, Stratta R, et al. Discriminant analysis of risk factors for early thrombocytopenia after orthotopic liver transplantation. First Symposium of the International Society for Peri- operative Care in Liver Transplantation 1990; 119: 37.

9 Corton EG, Plevak D J, O'Connor MK, Forstom CA. Splenic deposition of platelets after liver transplantation. First Symposium of the International Society for Peri-operative Care in Liver Transplantation 1990; 119: 1.

10 Williams R, O'Grady JG. Liver transplantation: Results, advances and problems. J Gastroenterol Hepatol 1990; Suppl 1: 110.

11 Siemensma NP, Bhathal PS, Penington DG. The effect of massive liver resection on platelet kinetics in the rat. J Lab Clin Med 1975; 86: 818.

12 Blumgart LH, Vajvabukka T. Injuries to the liver: analysis of 20 cases. Brit Med J 1972; 1: 158.

22 G.W. McCAUGHAN et al.

13 Ro J, Flatmark A. Haemostatic studies following extensive liver resection in dogs. Scand J Gastroenterol 1973; 8: 615.

14 Pepper H, Lindsay S. Responses to platelets, eosinophils and total leukocytes during and following surgical procedures. Surg Gyn- aecol Obstet 1960; l l0: 319.

15 Nakamura T, Nishizawa T, Hagiya M, et al. Molecular cloning and expression of human hepatocyte growth factor. Nature 1989; 342: 440.

16 Bussel JB, Pham LC. Intravenous treatment with gammaglobulin in adults with immune thrombocytopenic purpura: review of the literature. Vox Sang 1987; 52: 206.

thrombocytopenia post liver transplantation

Documents