Pregnancy after liver transplantation: Report of 8 new cases and review of the literature

Transplant Immunology 15 (2006) 297 – 302 www.elsevier.com/locate/trim

Pregnancy after liver transplantation: Report of 8 new cases and review of the literature Marialuisa Framarino dei Malatesta a,⁎, Massimo Rossi b , Bianca Rocca c , Massimo Iappelli b , Maria Pia Giorno a , Pasquale Berloco b , Raffaello Cortesini b a

Department of Gynecological Sciences and Perinatology, University of Rome “La Sapienza” School of Medicine Rome, Italy b Department of General Surgery “Paride Stefanini” University of Rome “La Sapienza” School of Medicine Rome, Italy c Research Center on Physiopathology of Haemostasis, Catholic University School of Medicine, Rome, Italy Received 1 December 2005; accepted 13 January 2006

Abstract Improved survival and quality of life following liver transplantation are associated with an increased frequency of pregnancies in livertransplanted women. We investigated the outcome, complications, and management of those pregnancies. We have reviewed the literature and report 8 pregnancies in 6 transplant recipients. Seven pregnancies were completed at 38 ± 2 (mean ± standard deviation) weeks. One miscarriage occurred at week 12. Newborns' weight averaged 2938 ± 156 g. Main complications were preeclampsia (n = 1) and reversible cholestasis (n = 1). Among 285 pregnancies reported in literature, 78 ± 20% were successful and the main complications were: preeclampsia (26 ± 19%), hypertension (28 ± 19%), reversible liver dysfunction (27 ± 21%), cesarean delivery (23 ± 10%), preterm birth (31 ± 28%), small for gestational age infants (23 ± 10%), rejection (10 ± 7%). Gestational weeks were 36.7 ± 1.3, perinatal mortality was 4 ± 10%, malformation rate 3%. The rates of both abortions and complications (preeclampsia and/or hypertension) were inversely related to the time interval between transplantation and conception (p b 0.05). Abortions occurred more often in recipients whose underlying disease was autoimmune cirrhosis than in recipients with inherited disorders. Rejection rate was approx. 10%, which appears higher than reported in a non-pregnant population after a comparable time interval from transplant (2–3%). Up to 28 months after delivery, maternal death was 5.5 ± 7%. We conclude that: the time intervals between transplantation and conception as well as the original cause of liver failure influence the outcome and complications of pregnancies in liver recipients. However, neonatal survival is high, while malformations are relatively rare. © 2006 Elsevier B.V. All rights reserved. Keywords: Immunosuppressive therapy; Orthotopic liver transplantation; Pregnancy

1. Introduction Although the first orthotopic liver transplant (OLT) was performed in 1963, low survival rates had precluded pregnancies in transplanted women. Advancement in surgical procedures and in immunosuppressive therapies has resulted in improved survival and quality of life. The first successful pregnancy in a liver transplant recipient was reported by Walcott et al. in 1978 [1] opening perspectives that were first available only to kidney transplant recipients [2]. Since then, an increasing number of pregnancies in liver transplanted women have been reported. However, general guidelines for the management of these preg⁎ Corresponding author. Tel.: +39 06 8848883; fax: +39 06 85301544. E-mail address: [email protected] (M.F. dei Malatesta). 0966-3274/$ - see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.trim.2006.01.001

nancies, which are generally considered at risk, are currently lacking. To draw a profile of post-OLT pregnancy and generate recommendations, which might be useful for patient management, we now report 8 cases and review the literature on complications, management and outcome of pregnancies in OLT recipients. 2. Patients and methods 2.1. Patients Eight pregnancies in 6 liver transplanted patients were studied (Table 1). The interval between transplantation and pregnancy was 45.5 ± 22 months. Immunosuppressive therapy

298

M.F. dei Malatesta et al. / Transplant Immunology 15 (2006) 297–302

Table 1 Patient's characteristics Patient

Age (yrs)

Indication for OLT

Immunosuppressive therapy

Interval, pregnancy (mo)

Gestational age (wks)

Weight at birth (g)

Type of delivery (reason for C/S)

Complications

IA IB II

22 25 26

Cyclosporine Cyclosporine Cyclosporine

36 72 25

39 38 38

2940 2835 2960

C/S (elective) C/S (previous C/S) C/S (breech presentation)

None None None

III IV VA

28 35 27

Cyclosporine Tacrolimus Cyclosporine

38 65 35

37 40 36

2815 3150 2740

Vaginal delivery Vaginal delivery C/S

None None PE

VB

34

Cyclosporine

71

39

3125

C/S (previous C/S)

LD

VI

33

Wilson disease Wilson disease Primary biliary, cirrhosis HCV Autoimmune Primary biliary, cirrhosis Primary biliary, cirrhosis HBV

Cyclosporine

22

12

–

–

Abortion

Abbreviations: PE: preeclampsia; C/S Cesarean section; LD, liver disease.

was maintained at the same level throughout the pregnancy. One patient was treated with tacrolimus (9.6 mg/day) while the remaining 5 patients received cyclosporine (average daily dose: 340 mg bid). The dosage of the administered drug was quite stable during pregnancy and circulating levels never exceeded the therapeutic range. Corticosteroids were never administered.

Patients were under observation from weeks 9 to 10 of amenorrhea. The mean number of prenatal visits was 12 ± 1 and of ultrasonographic examination was 4 ± 1. The following tests were performed monthly: markers for hepatitis B and C, CMV, Toxoplasma, routine blood tests for hepatic and renal functions. Vaginal smear cultures for β-hemolytic Streptococcus, Chlamydia,

Table 2 Data from the literature (series with more than five pregnancies) Reference

Pgns Age at (n) conception (yr)

Scantlebury 20 et al. [5] Ville 19 et al. [4] Radomski 48 et al. [8] Mass 7 et al. [13] Pruvot 7 et al. [6] Patapis 29 et al. [14] Wu 22 et al. [15] Casele and 14 Laifer [11,16] Safadi 5 et al. [10] Raakov 21 et al. [12] Jain 49 et al. [7] Morton 6 et al. [9] Nagy 38 et al. [3] Means ± SD 285

Interval OLT-pgn (mo)

Successful Wks of pgn (%) gestation

Preterm CS (b37 wks) (%) (%)

Weight at birth (g)

% SGA Complications (b2500 g) (% of total pregnancies)

26 ± 5

23.5 ± 30

100

35.6 ± 4

50

65

2238 ± 692

50

30 ± 7

26.4 ± 13

58

38 ± 1.5

0

45

2990 ± 370

10

26 ± 6

35 ± 30

75

37 ± 4

39

47

2604 ± 698

31

H 31; Inf 0; LD 31; 14/0 PE 21; PROM 10; R 5 H 27; Inf 8; LD 27; n.a./0 PROM 0; R 9 H 46; Inf 26; PE 21 R 12 3/0

20–36

10–30

58

na

0

25

na

na

na

25 ± 4

23 ± 19

100

36 ± 28

43

57

2310 ± 655

29

38

na

20

2900

na

H 57; Inf 42; LD 42 PROM 0; PE 33; R 0 LD 46; R 13

27 (median) 24 (median)

51

n.a./0 0

na

43

100

38 ± 2

14

32

2876 ± 589

na

na

20 ± 25

100

na

69

na

na

8

32 ± 6

33.6

60

37.7

na

na

na

na

R 20

0/0

19

46

67

38

19

48

2656

29

0/5

29 ± 6

70 ± 52

100

36.4 ± 3.2 4

47

2697 ± 775

9

26 ± 7

87 ± 67

83

34 ± 2

80

40

2098 ± 512

na

28 ± 5

39 ± 29

63

36 ± 4

29

46

2762 ± 1117 17

26.8 ± 3.4

39 ± 21

78 ± 20

36.7 ± 1.3 31 ± 28

H 43 ; Inf 14; LD 14 ; PE 5; H 2; Inf 0; LD 24; R 2 ; PROM 0 H 20 ; Inf 0; LD 60 PE 60 ; PROM 0 H 21 ; Inf 0; LD 8 ; PE 21; P 9; R 17 H 28 ± 19; Inf 20 ± 14; LD 27 ± 21; PE 26 ± 19; PROM 7 ± 13; R 10 ± 7;

43 ± 12 2612 ± 276

23 ± 10

H 14 ; Inf 14 LD 5 ; PE 14; R 5; H 15; Inf 0; PE 31; PROM 31;

Maternal/ neonatal deaths (%)

0/0 0/33

10/6 n.a./0 17/0 5.5 ± 7/4 ± 10

Abbreviations: pgn: pregnancies; SGA: small for gestational age; H: hypertension; Inf: infections; LD: liver dysfunction (reversible); PROM: premature rupture of membranes; PE: preeclampsia; R: rejection; na: not available. Data are means ± standard deviations, unless specified.

M.F. dei Malatesta et al. / Transplant Immunology 15 (2006) 297–302

2.2. Literature review We performed a MEDLINE search from 1978 to 2005 using the terms “liver transplantation” and “pregnancy”, and we considered only series of more than 5 pregnancies. Detailed data from available reports were included in the analysis (see Table 2 for details).

Statistical analysis was carried out using the SPSS software (SPSS for Windows, version 6.0). Data were evaluated by means of nonparametric tests for independent samples (Mann– Whittney U and Kolmogorov–Smirnov tests with comparable results). The association between different parameters was also assessed by multiple linear regression. Data are expressed as Mean ± Standard Deviation (SD). The level of significance was set at p b 0.05. 3. Results In our series, the mean age at conception was 29 ± 5 years (Table 1). Seven pregnancies were successfully completed and the mean gestational age at delivery was 38 ± 2 weeks. One miscarriage occurred at week 12. One pregnancy was complicated by preeclampsia and another by intrahepatic cholestasis at week 30, which spontaneously regressed by week 35. No episodes of rejection or infection were observed. All deliveries were uneventful, i.e. without obstetrical or liver complications. Two vaginal deliveries occurred at weeks 37 and 40. Five Cesarean sections (C-sections) were performed: in all but one patient Csections were performed for obstetrical indications, one patient required C-section for psychological problems, and one was a preterm delivery (36 weeks) (Table 2). All C-sections were performed under epidural anesthesia. Mean weight of newborns was 2938 ± 155 g, and newborns were all normal for their gestational age. The Apgar score was 8 ± 1. All newborns survived and no malformations or congenital abnormalities were observed. The follow-up of these children and their mothers ranged

* interval OLT-pgn (months)

50 40 30 20 10 0 Cyclosporinbased therapy

2.3. Statistical analysis

80

60

% of total population

Mycoplasma, Ureaplasma, Candida and E. coli were performed at weeks 22, 32 and 36. The fetal heart rate was weekly assessed from week 28 until delivery. All patients gave their informed consent.

p<0.01

60 40 20 0

Fig. 1. Interval between OLT and conception in successful pregnancies vs. abortions. The interval between OLT and conceptions (months) of successful pregnancies (white bar, n = 140) vs. abortions (black bar, n = 24) is shown. Data are means ± standard deviations. Data are from the present series and from the literature.

299

Tacrolimusbased therapy

Others

Fig. 2. Distribution of immunosuppressive therapy in abortions vs. successful pregnancies. The plot represents the distribution of the immunosuppressive therapy in abortions (black bars), vs. successful pregnancies (white bars). Cyclosporin-based therapy included Cyclosporine A, with or without corticosteroids and/or azathioprine; Tacrolimus-based therapy included tacrolimus, with or without corticosteroids. Data are from the present series and from the literature.

from 1 to 10 years and no deaths, physical or mental problems were observed. We also reviewed 285 pregnancies in OLT recipients from the literature [3–15]. Some of the considered parameters are reported in Table 2. On average, 78 ± 20% of pregnancies were successful in the 13 different studies included in the analysis [3–15]. The group of unsuccessful pregnancies, which comprises spontaneous abortions and therapeutic termination of pregnancies, was characterized by a significantly shorter interval between OLT and conception compared to the successful pregnancy group (24 ± 20 vs. 48 ± 42 months, respectively, p b 0.001), as shown in Fig. 1. However, cyclosporine- or tacrolimusbased immunosuppressive treatments were similarly distributed in the 2 groups (Fig. 2). It also appears that abortions were more frequent in patients who have had autoimmune cirrhosis (Fig. 3), although this association was at the limit of significance (χ2 = 3.83, p = 0.05). Notably, congenital inheritable disorders such as Alagille syndrome, α1antithripsin deficiency, Caroli syndrome were not observed in the abortion group. In successful pregnancies, weeks of gestations were on average below 37 (36.7 ± 1.3), and a high percentage of C-sections (43 ± 12%) was reported. There was a high incidence of preterm (31 ± 28%) or small for gestational age (SGA) (23 ± 10%) infants, and weight at birth was in the low range of percentile (2612 ± 272 g, 25–30th percentile). The main reason for C-section was fetal distress (26%). Other causes are shown in Fig. 4. Preeclampsia, hypertension and reversible liver dysfunction were the most common complications, with a rate of 26%, 28% and 27% respectively. Acute rejection averaged 10% (Table 2). The group of pregnancies complicated by hypertension and/or preeclampsia was characterized by a significantly shorter interval between transplantation and conception compared to uncomplicated pregnancies (20 ± 15 vs. 59 ± 20 months, respectively, p b 0.01). However, the duration of pregnancy was comparable in the hypertension/preeclampsia and uncomplicated pregnancy groups (35.7 ± 2 vs. 36.6 ± 3 weeks, respectively, p: n.s.) and showed no direct correlation to its earliness following transplantation (r = 0.12, n = 122, p: n.s.). The incidence of perinatal and neonatal deaths was approx. 4% (Table 2). Nine malformation were reported in 229 successful pregnancies. However, one case may have been caused by the alcoholic habits of the mother and another to an autosomic dominant disorder inherited from the mother. Therefore a likely estimate is 3% incidence of

300

M.F. dei Malatesta et al. / Transplant Immunology 15 (2006) 297–302

#

% of total population

35 30 25 20 15 10 5

d

lis W

ce du

on Al co Bi ho lia l ry at re Cr sia yp to ge ne Vi ra tic lh O ep th at er itis co ng en ita Ch l ol an g itis Bu dd Ch M al ia fo ri rm at io ns O th er s

Dr

ug

-in

os rrh

Ci ry

Pr im

ar

y

Bi lia

Au to i

m

m

un

e

is

0

Fig. 3. Causes for OLT in abortions vs. successful pregnancies. The plot represents the distribution of different diseases leading to OLT in abortions (black bars), vs. successful pregnancies (white bars). Data are from the present series and from the literature. #: p = 0.05 between abortions and successful pregnancies.

30 25 20 15 10 5

ive pr ev Ch ia ol es ta sis PR O M

ac en t

a

er

ec t

El

Pl

el iv

at

ld

gi na

va

y

io n

ss re

en t le d

Fa i

Br

ee c

h

pr

es

PE

ld ist ta Fe

io

us

CS

0

ev

The present study reports 8 pregnancies in OLT recipients and is the most-updated review on this issue. On the basis of present observations it is possible to identify the main problems of these pregnancies, depict possible risk factors and propose future therapeutic/prophylactic interventions. One of the main findings emerging from our review is that the interval between OLT and conception has an effect both on the rate of abortions and on preeclamptic/hypertensive complications occurring in successful pregnancies. The inverse association between interval from OLT and abortions has been observed before although on a small series of patients. More convincing evidence has been recently provided by Nagy et al. [3] based on the retrospective analysis of 38 pregnancies. Our analysis of a larger number of patients (n = 293 in total) confirms the conclusions by Nagy et al. Thus, the time interval between OLT and conception in the abortion/terminations group was 24 ± 20 months in our analysis and 24 ± 24 months in the analysis of Nagy et al. Taken together, these data suggest that physicians should advice women to consider conception only after at least 24–36 months following OLT. Casele and Laifer, in 1998 [16], recommended a shorter interval (6 months from OLT), yet this conclusion is not supported by the above data. At present, there is no clear explanation for this phenomenon. It is possible that a longer interval from OLT is required to stabilize liver function, lower immunosuppressive therapy at maintenance levels and control the related medical conditions (hypertension, diabetes). On the other hand, cytokine levels, such as interleukin-10, might be higher soon after transplant, impairing the angiogenetic process of placentation and favoring abortions/terminations [17]. In fact, interleukin-10 is more actively produced in the proximity of OLT, can mediate acute

Pr

4. Discussion

rejections [18], and is known as anti-angiogenetic factor [17]. Prospective studies should investigate the relationship between cytokine levels and pregnancy outcomes in OLT recipients. The other novel observation emerging from the present study is that a shorter interval between pregnancy and OLT is also associated with a higher incidence of pregnancy-related complications such as preeclampsia and/or hypertension. Preeclampsia and hypertension involve vascular dysfunction and their etiopathogenesis is still debated in this setting. It has been hypothesized that cyclosporin and tacrolimus cause vasoconstriction of renal vessels and reduce endogenous production of nitric oxide [19], and that prolonged exposure to cyclosporin can cause hypertension [16]. However, our analysis does not support this hypothesis, because if this was true, hypertension and preeclampsia should have been more frequent in patients transplanted long before. It is more likely that the immunologic/ inflammatory cytokine environment which is present soon after OLT can interfere with normal physiological changes occurring

% of total population

malformations in infants from 235 pregnancies (including our cases). Maternal death, over the first 22–28 months after delivery, was 5.5% in a population of 206 patients. Some studies (4 of 13) did not provide long-term follow-up of mothers.

Fig. 4. Causes for C-sections in OLT-recipient pregnancies. The plot represents causes for C-sections in 68 cases as % of the whole population. Abbreviations: PE: preeclampsia; PROM: premature rupture of membranes. Data are from the present series and from the literature.

M.F. dei Malatesta et al. / Transplant Immunology 15 (2006) 297–302

during normal pregnancy. Indeed, experimental data have demonstrated that in rats endothelial dysfunction and loss of vasodilatory properties are present very early after transplantation and decrease with time [20]. Transient endothelial dysfunction might also be an early feature of OLT in humans influencing placental circulation. Our review shows that approx. 70% of transplanted women have stable graft function during pregnancy and post-partum. In the remaining 30% of pregnancies after OLT, a mild, spontaneously reversible increase in bilirubin and transaminases has been observed. Since this required no increased immunosuppression, differential diagnosis with rejection is mandatory. The cause of this transient increase is unclear. The incidence of rejection over 285 pregnancies was around 10%, with a high variability among different studies (from 0% to 20%). Considering the average interval from OLT in these studies (39 months, Table 2), this rate of late rejection appears much higher than the one described in the non-pregnant population, 3–4 yrs after OLT, which is approx. 2–3% [21–24]. Nevertheless such discrepancy might be biased by heterogeneous criteria used for diagnosis (clinical, histopathological, exjuvantibus), and incomplete follow-up. Therefore, uniform criteria should be used in future study to establish whether pregnancy or puerperium influences rejection thus requiring a different therapeutic approach. Preterm delivery and low birth weight appear as major problems in transplanted pregnant patients and their babies. Fetal distress is the major cause for C-section in these women. The etiopathogenesis of low birth weight may be linked to placental vascular damage caused by hypertension, immunosuppressive regimen or hepatic dysfunction. The review of the literature, including our own series, shows a 3% incidence of malformations in children of OLT-transplanted mothers, which is comparable to the incidence in the normal population. Therefore, immunosuppressive therapy in utero seems to minimally affect organogenesis. However, longterm follow-up should be performed to determine whether infants of OLT-recipient mothers are at risk to develop certain immunological, oncological or metabolic disturbances during adolescence or adulthood. The overall death rate of OLT patients after delivery (5.5 ± 7%) was similar to that observed in the general OLT population (range 3–12.6%) at a comparable time interval after OLT, i.e. approx. 3 yrs [21–24]. Although pregnancy does not appear to worsen survival, prospective long-term studies are still needed to clarify this issue. Finally, our review suggests that the disease causing OLT might play some role in the outcome of pregnancies. Abortions were more frequent in patients who have had autoimmune cirrhosis. Disease-related mechanisms, such as the presence of lupus anticoagulant, which is often present in autoimmune hepatitis and is known to be involved in abortions in nontransplanted women, may play some role [25]. For this reason, we propose to screen OLT-recipients with previous autoimmune diseases for lupus anticoagulant. Although we acknowledge the limitations of any retrospective analysis and the need to perform prospective studies based

301

on the present study we can conclude the following: 1) post-OLT pregnancies display a high-risk profile and therefore require careful management to prevent complications; conception should be avoided in the first 2–3 yrs after OLT; 2) the disease causing OLT should be carefully considered by the physician because it might influence the outcome of pregnancy; 3) detailed information on the maternal and fetal risks of pregnancy should be provided. Randomized prospective studies are needed to assess the role of cytokines in pregnancy and determine whether primary prophylactic strategies may be effective. There is now good evidence that antiplatelet agents (principally low dose aspirin) prevent preeclampsia [26]. Low-dose aspirin (in patients with normal platelet counts) might ameliorate the outcome of postOLT pregnancies as well, considering their high-risk profile in terms of hypertension, eclampsia and abortions. Acknowledgements Supported by the ‘Consorzio Interuniversitario Trapianti d'Organo’, University of Rome School of Medicine and a grant from the Italian Ministry of Research and Education FIRB (n. RBNE01A882_005) to B.R.

References [1] Walcott WO, Derick DE, Joller JJ, Snyder DL. Successful pregnancy in a liver transplant patient. Am J Obstet Gynecol 1978;132(3):340–1. [2] Framarino dei Malatesta ML, Poli L, Pierucci F, Paolucci A, Pretagostini R, Di Nicuolo A, et al. Pregnancy and kidney transplantation: clinical problems and experience. Transplant Proc 1993;25(4):2188–9. [3] Nagy S, Bush MC, Berkowitz R, Fishbein TM, Gomez-lobo V. Pregnancy outcome in liver transplant recipients. Obstet Gynecol 2003;102(1):121–8. [4] Ville Y, Fernandez H, Samuel D, Bismuth H, Frydman R. Pregnancy in liver transplant recipients: course and outcome in 19 cases. Am J Obstet Gynecol 1993;168(3):896–902. [5] Scantlebury VP, Gordon R, Tzakis A, Koneru B, Bowman J, Mazzaferro V, et al. Childbearing after liver transplantation. Transplantation 1990;49 (2):317–21. [6] Pruvot FR, Declerck N, Valat-Rigot AS, Gambiez L, Labalette M, Gottrand F, et al. Grossese après transplantation hépatique: quel est le risque maternel? Gastroenterol Clin Biol 1996;20(5):461–75. [7] Jain AJ, Reyes J, Marcos A, Mazariegos C, Eghtesad B, Fontes PA, et al. Pregnancy after liver transplantation with tacrolimus immunosuppression: a single center's experience update at 13 years. Transplantation 2003;76 (5):827–32. [8] Radomski JS, Ahlswede BA, Jarrel BE, Mannion J, Cater J, Moritz MJ, et al. Outcomes of 500 pregnancies in 335 female kidney, liver and heart transplant recipients. Transplant Proc 1995;27(6):1089–90. [9] Morton A. Liver transplantation and pregnancy. Aust N Z J Obstet Gynaecol 2003;43(3):236–8. [10] Safadi R, Ilan Y, Jurim O, Schenker J, Galun E, Ashur Y, et al. Pregnancy and liver transplantation. Transplant Proc 1999;31(4):1899–900. [11] Casele HL, Laifer SA. Pregnancy after liver transplantation. Semin Perinatol 1998;22(2):149–55. [12] Raakow R, Neuhaus R, Buscher U, Schmidt S, Rayes N, Glanemann M, et al. Parenthood following liver transplantation. Transplant Proc 2001;33 (1–2):1450–2. [13] Mass K, Quint EH, Punch MR, Merion RM. Gynecological and reproductive function after liver transplantation. Transplantation 1996;62 (4):476–9.

302

M.F. dei Malatesta et al. / Transplant Immunology 15 (2006) 297–302

[14] Patapis P, Irani S, Mirza DF, Gunson BK, Lupo L, Mayer AD, et al. Outcome of graft function and pregnancy following liver transplantation. Transplant Proc 1997;29(1–2):1565–6. [15] Wu A, Nashan B, Messner U, Schmidt HH, Guenther HH, Niesert M, et al. Outcome of 22 successful pregnancies after liver transplantation. Clin Transplant 1998;12(5):454–64. [16] Casele HL, Laifer SA. Association of pregnancy complications and choice of immunosuppressant in liver transplant patients. Transplantation 1998;65 (4):581–3. [17] Kohno T, Mizukami H, Suzuki M, Saga Y, Takei Y, Shimpo M, et al. Interleukin-10-mediated inhibition of angiogenesis and tumor growth in mice bearing VEGF-producing ovarian cancer. Cancer Res 2003;63 (16):5091–4. [18] Warle MC, Metselaar HJ, Hop WC, Tilanus HW. Cytokine gene polymorphisms and acute liver graft rejection a meta-analysis. Liver Transplant 2005;11(9):19–26. [19] Gardiner SM, March JE, Kemp PA, Fallgren B, Bennet T. Regional haemodynamic effects of cyclosporine A, tacrolimus and sirolimus in conscious rats. Br J Pharmacol 2004;141(4):634–43. [20] Andriambeloson E, Bigaud M, Schraa EO, Kobel T, Lobstein V, Pally C, et al. Endothelial dysfunction and denudation in rat aortic allografts. Arterioscler Thromb Vasc Biol 2001;21(1):67–73.

[21] Rabkin JM, de La Melena V, Orloff SL, Corless CL, Rosen HR, Olyaei AJ. Late mortality after orthotopic liver transplantation. Am J Surg 2001;181 (5):475–9. [22] Pruthi J, Medkiff KA, Esrason KT, Donovan JA, Yoshida EM, Erb SR, et al. Analysis of causes of death in liver transplant recipients who survived more than 3 years. Liver Transplant 2001;7(9):811–5. [23] Levy MF, Somasundar PS, Jennings LW, Jung GJ, Molmenti EP, Fasola CG, et al. The elderly liver transplant recipient: a call for caution. Ann Surg 2001;233(1):107–13. [24] Garcia CE, Garcia RF, Mayer AD, Neuberger J. Liver transplantation in patients over sixty years of age. Transplantation 2001;72(4):679–84. [25] Galli M, Luciani D, Bertolini G, Barbui T. Lupus anticoagulants are stronger risk factors for thrombosis than anticardiolipin antibodies in the antiphospholipid syndrome: a systematic review of the literature. Blood 2003;102(8):1827–32. [26] Duley L, Henderson-Smart D, Kinight M, King J. Antiplatelet drugs for prevention of pre-eclampsia and its consequences: systematic review. Br Med J 2001;322(7282):329–33.