- Email: [email protected]
Effect of race on outcome of orthotopic liver transplantation: a cohort study
ARTICLES
Effect of race on outcome of orthotopic liver transplantation: a cohort study
Satheesh Nair, Joseph Eustace, Paul J Thuluvath
Summary Background There is significant evidence to suggest that long-term survival after renal transplantation is significantly lower in African Americans than in other races. We aimed to establish whether there was a difference in survival in African Americans compared with other races after orthotopic liver transplantation (OLT) and whether race was an independent predictor of survival. Methods We collected data from the United Network of Organ Sharing transplant registry for all liver transplants done between 1988 and 1996 in the USA. We also recorded information on age, sex, race, blood group, and cause of death for the donors and recipients. Findings 2-year graft survival was significantly lower for African Americans (601 of 884, 68%) and Asians (266 of 416, 64%) compared with white Americans (8703 of 11 762, 74%) and Hispanics (878 of 1220, 72%). 2-year and 5-year patient survival were significantly lower for African Americans (654 of 884 [74%], 270 of 565 [48%]) and Asians (287 of 416 [69%], 92 of 252 [37%]) compared with white Americans (9786 of 11 762 [83%], 4357 of 7514 [58%]) and Hispanics (964 of 1220 [79%], 341 of 657 [52%]). Compared with white Americans, African American (hazard ratio 1·36, 95% CI 1·16–1·60, p<0·0001) and Asian (1·25, 1·01–1·56, p=0·03) race were independent predictors of poor survival at 2 years. Interpretation African Americans and Asians have a worse outcome after OLT compared with white Americans and Hispanics. The higher rate of chronic rejection in African Americans and a relatively worse outcome in other minority races merits further examination. Lancet 2002; 359: 287–93
Division of Gastroenterology and Hepatology, Ochsner Medical Foundation, New Orleans, LA, USA (S Nair MD); and Department of Medicine, Johns Hopkins University School of Medicine, Baltimore, MD (J Eustace MD, P J Thuluvath FRCP) Correspondence to: Dr Paul J Thuluvath, Johns Hopkins University School of Medicine, Room 429, 1830 Bldg, 1830 E Monument Street, Baltimore, MD 21205, USA (e-mail: [email protected])
THE LANCET • Vol 359 • January 26, 2002 • www.thelancet.com
Introduction Due to an overall increase in the number of liver transplantations, a large number of African Americans, Hispanics, and Asians have undergone orthotopic liver transplantation (OLT) in the past decade. From published reports, it is not clear whether there is a difference in survival among the different races following OLT. There is significant evidence to suggest that the long-term survival in African Americans is lower compared with white Americans after renal transplantation.1–7 Inferior histocompatibility matching, poor compliance, and a lower socioeconomic status have been implicated for the poor survival among African Americans.1–8 However, it is unclear whether these observations might be extrapolated to patients undergoing OLT. A few studies have analysed the survival among African Americans after OLT, with conflicting results.9–14 Some studies reported a significantly lower survival for African Americans, whereas other studies reported a better or equivalent rate of survival.9–14 Several factors could explain these conflicting results; most of these studies had small number of African Americans and, moreover, these studies did not control for confounding factors such as United Network of Organ Sharing (UNOS) status, the cause of liver disease, renal function, diabetes, and obesity.15–22 In this study, we have analysed UNOS data for 1988 to 1996 to establish whether race was an independent predictor of survival after OLT and to identify the possible reasons behind such a difference, if it existed.
Patients and methods We extracted data from the UNOS transplant registry for all liver transplantations done in the USA between 1988 and 1996. Complete follow-up data were available on 14 282 patients at 2 years, 8988 patients at 5 years, and 600 patients at 10 years. The following information was available on both donor and recipients: age, sex, race, blood group, and cause of death. In addition, we collected data on recipients’ bodyweight and height at the time of listing, UNOS status, serum creatinine concentration at the time of transplantation, presence of diabetes, cause of liver disease, incidence of primary graft non-function and chronic rejection, 2-year graft survival, and 2-year patient survival. The rate of retransplantation and the interval between the initial transplantation and retransplantation were also recorded. We excluded patients whose race was not reported or was listed as other. Liver transplantation in the paediatric age-group (less than 18 years) was also excluded. UNOS status at the time of transplantation was recorded as status 1, 2, 3, and 4; these numbers do not correspond to the current listing criteria and status (currently UNOS status is recorded as 1, 2a, 2b, and 3).
287
For personal use. Only reproduce with permission from The Lancet Publishing Group.
ARTICLES
However, even under the older system, status 1 was applied to sicker patients requiring intensive care, although under the current system many of these patients would be categorised as 2a. The donor and recipient matching based on blood type was classified into three groups: ABO matched when donor and recipient were of same blood; ABO mismatched when donor blood type was O and recipient was of other blood groups; and ABO mismatched when donor blood group was A, B, or AB and recipient of another blood group. We classified the cause of liver disease as alcoholic cirrhosis, hepatitis C, hepatitis B, alcoholic and hepatitis C, primary biliary cirrhosis, primary sclerosing cholangitis, autoimmune liver disease, cryptogenic cirrhosis, and acute liver disease (fulminant liver failure), and other. Patients with a diagnosis of non-A or non-B hepatitis (1988–90 data) were included as hepatitis C. The cause of acute liver disease was not consistently included in the UNOS data and hence not available for analysis. We calculated body-mass index for all patients from weight and height.2 Renal function was arbitrarily classified into normal (serum creatinine ⭐88·4 mol/L [1 mg/dL]), mild renal insufficiency (serum creatinine 88·4–133 mol/L [1–1·5 mg/dL]), moderate renal failure (serum creatinine 141·4–221 mol/L [1·6–2·5 mg/dL]), and severe renal failure (serum creatinine ⭓229·8 mol/L [⭓2·6 mg/dL]). A serum creatinine of 88·4 mol/L or less was used to define normal to avoid underestimation of renal insufficiency since loss of muscle mass was common in patients with cirrhosis.23 Statistical analysis We used the 2 test for categorical variables and ANOVA for continuous variables to compare various outcome measures and pretransplantation variables between the groups. Mann-Whitney U test was used for comparison of continuous data that were not normally distributed. Patient survival at 2 years was used as the primary endpoint. The effect of race of the recipient, age (recipient and donor), sex, blood group, UNOS status, cold ischaemia time, renal function, body-mass index, and cause of liver disease on patient and graft survival was determined by univariate analysis. We examined the effect of race on graft and patient survival with the Cox Demographics and clinical factors Age (mean [SD], years) Sex (female/male) Blood group O ABO matched Body-mass index >35 (kg/m2) Serum creatinine (mol/L)‡ Normal creatinine (%) Diabetes (%) Donor age (years) Donor >50 years (%) UNOS status 1 (%) Cold ischaemia time (h)‡ Cause of liver disease Alcoholic liver disease Hepatitis C Hepatitis B Primary biliary cirrhosis Pimary sclerosing cholangitis Autoimmune hepatitis Cryptogenic cirrhosis Others§
proportional hazard model. The model was then further adjusted for the various confounders for which information was available. Overall survival was analysed by Kaplan-Meier survival analysis, and Cox-Mantel test was used to determine whether there were any differences between the groups. For this analysis, the patients who had retransplantation were excluded. For all analyses, a type I error of 0·05 was used. We used SPSS (version 9) for data analysis.
Results During the study period, 23 884 (3207 paediatric) patients underwent OLT. 3394 patients had incomplete UNOS status data. Race was not reported for 92 patients and race was reported as other in another 522 patients; these patients were excluded from the study. A total of 16 669 patients who had OLT during this period were available for final analysis. Of the 16 669 patients who had OLT, 13 599 (82%) were white Americans, 1042 (6%) were African Americans, 1534 (9%) were Hispanics, and 494 (3%) were Asians. Among the donors, there were 80% white Americans, 10% African Americans, 8% Hispanics, and 1% Asians. During the study period, there was a steady increase in the total number of African American donors, but there was no proportional increase in the number of African American recipients. The proportion of African Americans who received OLT declined steadily while there was a three-fold increase in the proportion of Hispanic and Asian recipients. African Americans were significantly younger than the other groups (table 1). More African Americans had UNOS status 1 at the time of transplantation than white Americans. Data on diabetes were available only for 65% of patients; the proportions of patients with diabetes and severe obesity (a body-mass index of more than 35 kg/m2) were similar among the four races. Similarly, there were no differences in cold ischaemia time, donor age, and serum creatinine concentration among the groups. African Americans and Asians had a significantly lower proportion of ABO-matched transplantation compared with white Americans or Hispanics. The most common cause of chronic liver disease was hepatitis C for African
White (n=13 599)
African American (n=1042)
Hispanics (n=1534)
Asians (n=494)
48·9 (11·0) 42/58 5711 (42%) 12 239 (90%) 896 (14%) 118 (101) 7629 (57%) 709 (5%) 32·7 (15·4) 2201 (16%) 1931 (14%) 10·6 (5·3)
41·3 (12·2)* 58/42 498 (48%) 906 (87%) 65 (12%) 121 (117) 588 (56%) 60 (6%) 32·2 (15·8) 174 (17%) 210 (20%)* 10·5 (5·6)
47·5 (11·7) 48/52 767 (50%) 1396 (91%) 99 (14%) 114 (108) 864 (57%) 121 (7%) 32·8 (16·5) 295 (19%) 229 (15%) 10·4 (5·8)
50·4 (11·1) 41/59 159 (32%)† 425 (86%) 11 (4%)* 158 (97) 276 (55%) 34 (6%) 31·5 (16·4) 93 (19%) 88 (18%) 10·5 (5·4)
18% 19% 4% 10% 10% 4% 10% 25%
11% 17% 3% 3%* 11% 11% 8% 36%
14% 25% 3% 9% 3% 5% 12% 29%
4% 19% 28%* 2% 2% 2% 5% 38%
*p<0·05 compared with white people; †Most common blood group among Asians was A (34%); ‡Not normally distributed, compared with Mann-Whitney U test; §Others included combination of diseases and other rare disorders.
Table 1: Demographic and clinical characteristics of recipients
288
THE LANCET • Vol 359 • January 26, 2002 • www.thelancet.com
For personal use. Only reproduce with permission from The Lancet Publishing Group.
ARTICLES
White (n=13 599) Outcome measures Primary graft non-function 30-day mortality Chronic graft failure rejection 2-year graft survival Retransplantation† 2-year patient survival 5-year patient survival Cause of death‡ Infections Cardiovascular Cerebrovascular accident Renal failure Malignancy§ Lymphoproliferative diseases Operative death Respiratory failure
African American (n=1042)
Hispanics (n=1534)
Asians (n=494)
5% 6% 6% 74% 28% 83% 58%
7% 9%* 12%* 68%* 22% 74%* 48%
6% 7% 6% 72% 32% 79% 52%
5% 8% 7% 64%* 18% 69%* 37%
38% 19% 8% 3% 19% 2% 6% 5%
43% 17% 7% 4% 13% 1% 9% 7%
46% 16% 9% 2% 12% 1% 10% 4%
38% 19% 10% 3% 20% 1% 5% 2%
*p<0·05 compared with white people; †% of those who developed graft failure within 2 years; ‡In 2762 patients, miscellaneous excluded; §Excludes lymphoproliferative diseases.
Table 2: Outcome measurements among the races
Americans, white Americans, and Hispanics, whereas hepatitis B was the most common cause among Asian recipients. Acute liver failure was the indication for OLT in 20% of African American patients. Table 2 shows the outcome measurements among the races. The overall mean graft and patient survivals during this period were 73% and 79%, respectively. There was a modest decrease in graft survival during the early half of the study period, but this stabilised during the later part. However, there was a modest increase in patient survival over the later half of the study period. This difference had resulted in a widening of the gap between patient survival and graft survival (figure 1). The primary graft non-function was marginally higher in African-Americans than in white Americans. 2-year graft survival was significantly lower in African Americans and Asians than in white Americans and Hispanics. During the study period, more white Americans than people from any other race had retransplantation. Among patients who lost the graft within 2 years, the retransplantation rate was 20–30%; the retransplantation rate was lowest among Asians (18%) followed by African Americans (22%). The rate of graft failure due to chronic rejection was almost twice as high among African Americans compared with other races. 30-day mortality was also higher in African Americans than the other races. 2-year patient survival was 90 2-year patient survival 2-year graft survival
Survival (%)
84 78 72 66 60 1988 1989 1990 1991 1992 1993 1994 1995 1996 Year of transplantation Figure 1: 2-year graft and patient survival according to year of liver transplantation
THE LANCET • Vol 359 • January 26, 2002 • www.thelancet.com
significantly lower for African Americans and Asians compared with white Americans and Hispanics. Information on 5-year survival was available only for 8988 patients (7514 white, 565 African American, 657 Hispanics, and 252 Asians). We believe that the data entry was perhaps incomplete at 5 years and therefore patients with incomplete data were censored and were considered “lost for follow up”. This reduced the number of patients at risk and might have under-estimated survival marginally (as there could be a bias in entering death). 5-year survival was lower for Asians and African Americans than for white Americans and Hispanics. Cause of death was available for 2762 patients (table 2). Infection was the leading cause of death, followed by cardiovascular diseases. Malignancy (excluding lymphoproliferative disease) was the second commonest cause of death in Asians and was mainly due to the recurrence of hepatocellular carcinoma. There were no significant differences in the cause of death among different races. Table 3 lists the graft and patient survival among the four races for various liver diseases. African Americans with hepatitis C had a lower survival rate than white Americans and Hispanics. Likewise, survival among African Americans was lower for acute liver failure. The 2-year survival for hepatitis B was similar with other diseases for white Americans, African Americans, and Hispanics. However, Asians with hepatitis B had a much lower survival rate than other races after OLT. The Hispanics who Cause of liver disease White African (n=13 599) American (n=1042) Demographics Alcoholic liver disease Hepatitis C Hepatitis B Autoimmune PSC PBC Acute liver failure Cryptogenic cirrhosis
74/80 72/80 70/74 78/83 81/88 82/87 68/70 74/79
70/74 67/72† 71/80 70/78 83/86 81/88 59†/64† 62/67†
Hispanics Asians (n=1534) (n=494)
78/82 70/78 77/77 62†/70† 76/89‡ 83/88 70/82 70/75
80/80 69/77 56†/60† 68/80 64/70‡ 66/87‡ 73/84‡ 68/68‡
Results are presented as graft survival/patient survival. PSC=primary sclerosing cholangitis; PBC=primary biliary cirrhosis. *2-year graft and patient survival in % adjusted to nearest full number; †p<0·05 compared with white people; ‡Number of patients less than 50.
Table 3: 2-year graft and 2-year patient survival* among races based on the cause of liver disease
289
For personal use. Only reproduce with permission from The Lancet Publishing Group.
ARTICLES
Variable
Risk ratio (95% CI)
p
Variable
Patient survival Hazard ratio* (95% CI) p
Races (reference: white Americans) African Americans Hispanics Asians
1·34 (1·17–1·52) 0·98 (0·86–1·11) 1·78 (1·51–2·1)
0·0001 0·77 0·0001
Age
1·02 (1·01–1·02)
Sex (reference: male)
0·90 (0·84–0·96)
0·0001
Race (reference: white Americans) African Americans Hispanics Asians
1·36 (1·16–1·60) 1·01 (0·86–1·18) 1·25 (1·01–1·56)
0.007 0·89 0·03
0·004
Age
1·01 (1·01–1·02)
0·0001
Body-mass index
1·00 (0·99–1·01)
0·63
Sex (reference: male)
0·99 ( (0·90–1·09)
0·87
Serum creatinine*
1·17 (1·15–1·20)
0·0001
Donor age
1·01 (1·01–1·02)
0·0001
Donor age
1·01 (1·01–1·02)
0·0001
Creatinine
1·03 (1·10–1·16)
0·0001
ABO mismatch (donor type O)
1·32 (1·14–1·53)
0·0002
Cold ischaemic time
1·01 (0.99–1·01)
0·39
ABO mismatch (donor type A or B or AB) 1·29 (1·05–1·59)
0·0157
ABO mismatch (donor type O)
1·06 (0·91–1·2)
0·41 0·09
Cold ischaemia time
1·01 (1·00–1·02)
0·0009
ABO mismatch (donor type A, B or AB) 1·20 (0·96–1·50)
UNOS status (reference: status 4) UNOS status 1 UNOS status 2 UNOS status 3
2·14 (1·78–2·59) 1·52 (1·2–1·8) 0·94 (0·8–1·13)
0·0001 0·001 0·57
UNOS status (reference: status 4) UNOS status 1 UNOS status 2 UNOS status 3
1·66 (1·35–2·04) 1·23 (1·01–1·50) 0·90 (0·75–1·09)
0·0001 0·04 0·3
Cause of liver disease (reference: PSC)† Alcoholic liver disease 1·67 (1·41–1·99) Hepatitis C virus 1·79 (1·49–2·15) Hepatitis B virus 2·5 ( 2·02–3.02) Autoimmune 1·66 (1·46–2·13) PBC 1.08 (0·88–1·37) Acute liver failure 2·69 (2·10–3·30) Cryptogenic cirrhosis 1·85 (1·51–2·26)
0·0001 0·0001 0·0001 0·0001 0·40 0·0001 0·001
Cause of liver disease (reference: PSC) Alcoholic liver disease 1·50 (1·21–1·85) Hepatitis C 1·63 (1·32–2·02) Hepatitis B 2.38 (1·85–3·0) Autoimmune 1·95 (1·49–2·5) Primary biliary cirrhosis 1·00 (0·78–1·3) Acute liver failure 1·84 (1·42–2·39) Cryptogenic cirrhosis 1·63 (1·30–2·0)
0·0001 0·0001 0·0001 0·0001 0·9 0·0001 0·0001
Table 4: Results of univariate analysis for patient survival at 2 years
had OLT for autoimmune hepatitis had a lower survival rate than did African Americans and white Americans. Table 4 shows the results of univariate analysis of various pretransplant characteristics as a determinant of patient survival at 2 years. Univariate analysis showed that race, age, sex, donor age, presence of diabetes, serum creatinine, UNOS status, cold ischaemia time, and cause of liver disease were significant predictors of patient survival at 2 years. In univariate analysis, when compared with primary sclerosing cholangitis, hepatitis B was a significant predictor of poor survival (hazard ratio 2·5, 95% CI 2·02–3·02) as were alcoholic cirrhosis, hepatitis C, and autoimmune liver diseases. All variables that were reported to be significant by univariate analysis were entered into the Cox regression model, and the results are summarised in table 5. As compared with white Americans, African American race was an independent predictor of poor survival at 2 years. Other significant independent predictors of poor survival were UNOS status 1, serum creatinine, and donor and recipient age. As expected, cause of liver disease had a significant effect on 2-year survival. Since there were more African Americans with UNOS status 1, a strong independent predictor of survival, we analysed the data for only UNOS status 1 and also by excluding UNOS status 1. Although multivariate model simultaneously adjusts for other confounding variables, we did a subgroup analysis to confirm the robustness of our findings. When only UNOS status 1 patients were analysed, African American race was an independent predictor (hazard ratio 1·5, 95% CI 1·14–2·0, p=0·004). Similar results were obtained when UNOS status 1 was excluded (1·3,
290
Patients with no reported UNOS status, or those with cause of liver disease other than the one used in the univariate analysis were excluded. PSC=primary sclerosing cholangitis. *Adjusted hazard ratio.
Table 5: Results of multivariate Cox regression model for patient survival at 2 years
1·1–1·6, p=0·003). We also analysed the data at 5 years with Cox regression model. African American (risk ratio 1·5, 95% CI 1·3–1·7, p<0·001), Asian (hazard ratio 1·4, 95% CI 1·2–1·7, p<0·001), and Hispanic (hazard ratio 1·2, 95% CI 1·09–1·4, p=0·002) race were all independent predictors of survival at 5 years. Figure 2 shows the Kaplan-Meier curve for patient survival. Survival rate was significantly lower for African Americans and Asians, but not for Hispanics, when compared with white Americans. The lower survival of African Americans and Asians 1·0 0·9 Cumulative survival
PSC=primary sclerosing cholangitis; PBC=primary biliary cirrhosis. *Serum creatinine at the time of OLT; †Patients with more than one disease (acute liver disease+hepatitis C, hepatitis B+C, hepatitis B+D), Budd Chiari syndrome, metabolic disorders, and other rare disorders are not included.
0·8 0·7 0·6 0·5 0·4
White Americans Hispanics African Americans Asians
0·3 0·2 0·1 0 0
Number of patients White Americans 13599 Hispanics 1534 African Americans 1042 494 Asians
1000
2000 3000 Survival (days)
8523 861 570 240
3627 271 213 70
1006 68 58 11
4000
5000
109 2 5 0
0 0 0 0
Figure 2: Kaplan-Meier survival curve for different races Survival was significantly lower for African Americans and Asians compared with white Americans.
THE LANCET • Vol 359 • January 26, 2002 • www.thelancet.com
For personal use. Only reproduce with permission from The Lancet Publishing Group.
ARTICLES
Variable
Hazard ratio* (95% CI) p
Race (reference: white Americans) African Americans Hispanics Asians
1·16 (1·01–1·33) 1·02 (0·89–1·15) 1·18 (0·98–1·43)
0·03 0·79 0·07
Age
1·01 (1·00–1·01)
0·0001
Sex (reference: male)
1·10 (0·96–1·27)
0·15
Donor age
1·02 (1·01–1·02)
0·0001
Creatinine
1·09 (1·07–1·12)
0·0001
Cold ischaemia time
1·01 (0·99–1·01)
0·08
ABO mismatch (donor type O)
1·22 (1·02–1·46)
0·03
ABO mismatch (donor type A, B, or AB) 1·13 (0·85–1·51)
0·97
UNOS status (reference: status 4) UNOS status 1 UNOS status 2 UNOS status 3
1·73 (1·48–2·07) 1·23 (1·04–1·47) 1·03 (0·87–1·06)
0·0001 0·015 0·72
Cause of liver disease (reference: PSC) Alcoholic liver disease 1·16 (0·98–1·37) Hepatitis C 1·40 (1·19–1·65) Hepatitis B 1·69 (1·37–2·08) Autoimmune 1·47 (1·18–1·83) PBC 0·89 (0·53–1·46) Acute liver failure 1·48 (1·20–1·82) Cryptogenic cirrhosis 1·29 (1·09–1·55)
0·07 0·0001 0·0001 0·0001 0·24 0·0001 0·004
PSC=primary sclerosing cholangitis; PBC=primary biliary cirrhosis. Patients with no reported UNOS status, or those with causes other than the one used in univariate analysis were excluded. *Adjusted hazard ratio.
Table 6: Results of multivariate Cox regression model for graft survival at 2 years
was evident within a year of transplantation, but the difference became more evident with time. Table 6 shows the Cox model for graft survival at 2 years. Kaplan-Meier analysis with Cox-Mantel test for comparison revealed a significantly lower graft survival for African Americans (p=0·03), with a trend toward an increased graft loss in Asians (p=0·07). The mean graft survival for African Americans was at least 300 days less than for white Americans and Hispanics. Graft loss was due to rejection, primary non-function, recurrent disease, vascular complication, or patient death. Race was an independent predictor of graft survival at 2 years. The other independent predictors of survival were cold ischaemic time, creatinine, donor age, and cause of liver disease. In patients with diabetes for whom we had data (n=3575), presence of diabetes was an independent predictor of poor 2-year graft (hazard ratio 1·29, 95% CI 1·08–1·54, p=0·003) and patient (1·34, 1·09–1·66, p=0·0001) survival.
Discussion In this study, we have shown that African American race was an independent predictor of patient survival after OLT. Other independent predictors of survival were UNOS status 1, donor and recipient age, serum creatinine concentration, diabetes, and cause of liver disease. Several reasons have been proposed to explain the lower survival in African Americans after renal transplantation, such as poor HLA matching between the donor and recipient, which leads to an increase in the number of rejection episodes and graft loss, poor socioeconomic status, and lack of insurance benefits resulting in inadequate post-transplant care.1–8 Some of these explanations, especially poor compliance, although they remain unsubstantiated, might also be applicable to liver transplantation recipients. In our study, only two pretransplant
THE LANCET • Vol 359 • January 26, 2002 • www.thelancet.com
variables, recipient age and UNOS status, were significantly different in the African Americans. The African Americans were at least 7 years younger than white Americans, a factor that should have favourably affected survival. By contrast, there was a higher proportion of African Americans with UNOS status 1, which indicated that these patients were probably very sick at the time of transplantation. Moreover, acute fulminant failure was a common indication for OLT among African Americans. Fulminant hepatic failure and UNOS status 1 are generally associated with a lower patient survival after OLT.24,25 The higher 30-day mortality in African Americans could be explained by these observations. The poor survival among African Americans could also be explained by the higher frequency of primary graft non-function in African Americans than in white Americans. In addition, the rate of chronic rejection was higher among African Americans. It is possible that a combination of these factors might have resulted in a 2-year survival that is 10% less in African Americans than in white Americans. Although poor compliance and inadequate insurance among African Americans have been blamed for the higher rate of chronic rejection, there is no convincing evidence to support these suggestions. It is more likely that immunological factors, yet unidentified, might be contributing to chronic rejection. Moreover, most of the currently available immunosuppressive drugs were tested in predominantly white Americans and there might be a need to test these drugs more rigorously in minorities. Although the survival of Asians was lower compared with other races, when patients with hepatitis B were excluded, the survival rate at 2 years was similar to African Americans. The second common cause of death among Asians was malignancy, which reflects a higher frequency of recurrent hepatocellular carcinoma (data not shown). The disease-specific survival analysis among races yielded some interesting findings. The survival of Hispanics and whites were similar for most causes, but African Americans showed a lower survival with hepatitis C and acute failure. Many small studies have shown that African Americans were less likely to have sustained hepatitis C virus clearance with medical treatment compared with white Americans.26 Our analysis showed that African Americans had a poorer 2-year survival after OLT for hepatitis C, an observation that had not been reported before. It is possible that African Americans have a relatively higher viral load when they undergo OLT and hence have a more severe recurrent hepatitis C. Another possibility is that these patients had more episodes of rejection that required steroid boluses and monoclonal antibody use, factors implicated in a more aggressive disease recurrence. Survival for patients with hepatitis B was similar to that for those with other liver diseases for all races except in Asians who had a relatively poor 2-year survival. It is possible, although we have no data, that more Asians had DNA for hepatitis B virus or were HBeAg positive (replicative phase) at the time of OLT than other groups. Recurrent hepatocellular carcinoma could also partly explain the lower survival.27 Primary sclerosing cholangitis and primary biliary cirrhosis had the best survival for all races.
291
For personal use. Only reproduce with permission from The Lancet Publishing Group.
ARTICLES
We identified a number of predictors of survival in our univariate analysis and our observations are consistent with previous reports.28–35 In regression analysis, most of the factors that were predictive in univariate analysis were also significantly associated with lower survival except sex of the patient. Diabetes was also reported to be a significant predictive factor when data were analysed including only those patients with that information in the database; most of these patients were transplanted between 1994 and 1996. In this subgroup, hepatitis B was not an independent predictor of survival. This finding, not totally expected, was probably due to careful selection of patients with hepatitis B and hepatocellular carcinoma, better treatment of hepatitis B with antiviral agents, which reduced the pretransplant viral load, and use of posttransplant immunoglobulin during this period and thereafter.36-39 In summary, African Americans had a lower survival than white Americans and Hispanics after OLT and African American race was an independent predictor of a lower patient survival. Higher postoperative mortality (30-day mortality) due to a higher number of seriously ill patients (UNOS status 1 and fulminant hepatic failure), and higher rates of chronic rejection and primary graft non-function could partly explain some of these observations. However, we do not believe that any of these observations could explain a 5-year survival that is 17% less in African Americans compared with white Americans. Our study suggests that there is a clear need for prospective studies to examine our observations further. Until then, the reasons for poor survival in African Americans will remain speculative and will probably be dismissed as being due to poor compliance with therapy. Moreover, the higher rate of chronic rejection in African Americans suggests that there should be more rigorous drug trials in this patient population.
7
8
9
10
11
12 13
14
15
16
17
18
19
20
21
Contributors S Nair was involved in the collection, analysis, and interpretation of the data, and the drafting of the manuscript. J Eustace analysed and interpreted the data, provided statistical expertise, and helped revise the manuscript. P J Thuluvath contributed to the concept and design of the study, collected, interpreted, and analysed the data, and helped revise the manuscript.
Conflict of interest statement None declared.
22 23
24
25
Acknowledgments We thank UNOS for providing us with the data.
26
References
27
1
2 3
4
5
6
Isaacs RB, Nock SL, Spencer CE, et al. Racial disparities in renal transplant outcomes. Am J Kidney Dis 1999; 34: 706–12. Hariharan S, Schroeder TJ, First MR. Effect of race on renal transplant outcome. Clin Transplant 1993; 7: 235–39. Valente JF, Hariharan S, Peddi VR, et al. Causes of renal allograft loss in black vs white transplant recipients in the cyclosporine era. Clin Transplant 1997; 11: 231–36. Flechner SM, Modlin CS, Serrano DP, et al. Determinants of chronic renal allograft rejection in cyclosporine-treated recipients. Transplantation 1996; 62: 1235–41. Gjertson DW. A multi-factor analysis of kidney graft outcomes at one and five years post-transplantation: 1996 UNOS Update. Clin Transpl 1996; 343–60. Kasiske BL, Neylan JF III, Riggio RR, et al. The effect of race on access and outcome in transplantation. N Engl J Med 1991; 324: 302–07.
292
28
29
30
31
32
Gallichio MH, Hudson S, Young CJ, Diethelm AG, Deicrhoi MH. Renal retransplantation at the University of Alabama at Birmingham: incidence and outcome. Clin Transpl 1998; 169–75. Ojo AO, Port FK, Held PI, et al. Inferior outcome of two-haplotype matched renal transplants in blacks: role of early rejection. Kidney Int 1995; 48: 1592–99. Dawson S III, Imagawa DK, Johnson C, et al. UCLA liver transplantation: analysis of immunological factors affecting outcome. Artif Organs 1996; 20: 1063–72. Eckhoff DE, McGuire BM, Young CJ, et al. Race: a critical factor in organ donation, patient referral and selection, and orthotopic liver transplantation? Liver Transpl Surg 1998; 4: 499–505. Teperman L, Scantlebury V, Tzakis A, Staschak S, Todo S, Starzl TE. Liver transplantation in Black recipients: Pittsburgh. Transplantation Proc 1989; 21: 3963–65. Belle SH, Detre KM. Report from Pitt-UNOS liver transplantation registry. Transplantation Proc 1993; 25: 1137–42. Gonwa TA, Morris CA, Mai ML, Husberg BS, Goldstein RM, Klintmalm GB. Race and liver transplantation. Arch Surg 1991; 126: 1141–43. Pillay P, Van Thiel DH, Gavaler JS, Starzl TE. Effect of race upon organ donation and recipient survival in liver transplantation. Dig Dis Sci 1990; 35: 1391–96. Ratziu V, Samuel D, Sebagh M, et al. Long-term follow-up after liver transplantation for autoimmune hepatitis: evidence of recurrence of primary disease. J Hepatol 1999; 30: 131–41. Charlton M, Seaberg E, Wiesner R, et al. Predictors of patient and graft survival following liver transplantation for hepatitis C. Hepatology 1998; 28: 823–30. Graziadei IW, Wiesner RH, Marotta PJ, et al. Long-term results of patients undergoing liver transplantation for primary sclerosing cholangitis. Hepatology 1999; 30: 1121–27. Shields PL, Tang H, Neuberger JM, Gunson BK, McMaster P, Pirenne J. Poor outcome in patients with diabetes mellitus undergoing liver transplantation. Transplantation 1999; 68: 530–35. Fraley DS, Burr R, Bernardini J, Angus D, Kramer DJ, Johnson JP. Impact of acute renal failure on mortality in end-stage liver disease with or without transplantation. Kidney Int 1998; 54: 518–24. Sawyer RG, Pelletier SJ, Prett TL. Increased early morbidity and mortality with acceptable long-term function in severely obese patients undergoing liver transplantation. Clin Transplant 1999; 13: 126–30. Braunfeld MY, Chan S, Pregler J, et al. Liver transplantation in the morbidly obese. J Clin Anesth 1996; 8: 585–90. Keeffe E, Gettys C, Esquivel CO. Liver transplantation inpatients with severe obesity. Transplantation 1994; 57: 309–11. Caregaro L, Menon F, Angeli P, et al. Limitations of serum creatinine level and creatinine clearance as filtration markers in cirrhosis. Arch Intern Med 1994; 154: 201–05. Bismuth H, Samuel D, Castaing D, et al. Orthotopic liver transplantation in fulminant and subfulminant hepatitis. The Paul Brousse experience. Ann Surg 1995; 222: 109–19. Devlin J, Wendon J, Heaton N, Tan KC, Williams R. Pretransplantation clinical status and outcome of emergency transplantation for acute liver failure. Hepatology 1995; 21: 1018–24. Reddy KR, Hoofnagle JH, Tong MJ, et al. Racial differences in responses to therapy with interferon in chronic hepatitis C. Consensus Interferon Study Group. Hepatology 1999; 30: 787–93. Beasley RP. Hepatitis B virus: the major etiology of hepatocellular carcinoma. Cancer 1988; 61: 1942–56. Grande L, Matus D, Manyalic M, et al. Effect of donor age on graft outcome after liver transplantation. Transplant Proc 1999; 31: 2482–83. Wiesner RH, Demetris AJ, Belle SH, et al. Acute hepatic allograft rejection: incidence, risk factors, and impact on outcome. Hepatology 1998; 28: 638–45. Tisone G, Romagnoli J, Vennarecci G, et al. Is donor age a risk factor for poor graft function after orthotopic liver transplantation? Transplant Proc 1997; 29: 2825–26. Detre KM, Lombardero M, Belle S, et al. Influence of donor age on graft survival after liver transplantation—United Network for Organ Sharing Registry. Liver Transpl Surg 1995; 1: 311–19. Piratvisuth T, Tredger JM, Hayllar KA, Williams R. Contribution of true cold and rewarming ischemia times to factors determining outcome after orthotopic liver transplantation. Liver Transpl Surg 1995; 1: 296–301.
THE LANCET • Vol 359 • January 26, 2002 • www.thelancet.com
For personal use. Only reproduce with permission from The Lancet Publishing Group.
ARTICLES
33 Hoofnagle JH, Lombardero M, Zetterman RK, et al. Donor age and outcome of liver transplantation. Hepatology 1996; 24: 89–96. 34 Marino IR, Doyle HR, Aldrighetti L, et al. Effect of donor age and sex on the outcome of liver transplantation. Hepatology 1995; 22: 1754–62. 35 Ploeg RJ, D’Alessandro AM, Knechtle SJ, et al. Risk factors for primary dysfunction after liver transplantation—a multivariate analysis. Transplantation 1993; 55: 807–13. 36 Chen PJ, Chen DS. Hepatitis B virus infection and hepatocellular carcinoma: molecular genetics and clinical perspectives. Semin Liver
Dis 1999; 19: 253–62. 37 Burbach GJ, Bienzle U, Neuhaus R, et al. Intravenous or intramuscular anti-HBs immunoglobulin for the prevention of hepatitis B reinfection after orthotopic liver transplantation. Transplantation 1997; 63: 478–80. 38 Ho BM, So SK, Esquivel CO, Keeffe EB. Liver transplantation in Asian patients with chronic hepatitis B. Hepatology 1997; 25: 223–25. 39 Terrault NA, Zhou S, Combs C, et al. Prophylaxis in liver transplant recipients using a fixed dosing schedule of hepatitis B immunoglobulin. Hepatology 1996; 24: 1327–33.
Clinical picture Metastatic calcification Ronald C W Ma, Clive S Cockram
A 72-year-old woman had obstructive jaundice from an inoperable carcinoma of the head of the pancreas. 6 days after bypass surgery, she developed generalised tonic clonic convulsions. Her adjusted serum calcium was 1·36 mmol/L, phosphate 1·25 mmol/L, albumin 24 g/L, and glucose 10·7 mmol/L. Serum parathyroid hormone was <0·1 pmol/L, and magnesium was 0·4 nmol/l. Cerebral computed tomography showed intracranial calcification in the subcortical white matter, basal ganglia, cerebellum, corona radiata, and centrum semiovale, consistent with metastatic calcification from long-standing hypoparathyroidism (figure). We gave her intravenous calcium
and magnesium replacement, and she had no further convulsions. We found a thyroidectomy scar on careful examination, and the patient said that she had had a total thyroidectomy many years previously, and had been taking thyroxine. She had not had regular serum calcium measurements, and did have occasional parasthesiae. She also took a thiazide diuretic, which had been withheld on admission to hospital. The combination of obstructive jaundice and consequent alteration of vitamin D metabolism, thiazide withdrawal, recent bypass surgery, and intravenous fluid administration probably exacerbated her long-standing iatrogenic hypoparathyroidism.
Department of Medicine and Therapeutics, Prince of Wales Hospital, Shatin, Hong Kong (RCW Ma MRCP, CS Cockram FRCP)
THE LANCET • Vol 359 • January 26, 2002 • www.thelancet.com
293
For personal use. Only reproduce with permission from The Lancet Publishing Group.
Effect of race on outcome of orthotopic liver transplantation: a cohort study
Satheesh Nair, Joseph Eustace, Paul J Thuluvath
Summary Background There is significant evidence to suggest that long-term survival after renal transplantation is significantly lower in African Americans than in other races. We aimed to establish whether there was a difference in survival in African Americans compared with other races after orthotopic liver transplantation (OLT) and whether race was an independent predictor of survival. Methods We collected data from the United Network of Organ Sharing transplant registry for all liver transplants done between 1988 and 1996 in the USA. We also recorded information on age, sex, race, blood group, and cause of death for the donors and recipients. Findings 2-year graft survival was significantly lower for African Americans (601 of 884, 68%) and Asians (266 of 416, 64%) compared with white Americans (8703 of 11 762, 74%) and Hispanics (878 of 1220, 72%). 2-year and 5-year patient survival were significantly lower for African Americans (654 of 884 [74%], 270 of 565 [48%]) and Asians (287 of 416 [69%], 92 of 252 [37%]) compared with white Americans (9786 of 11 762 [83%], 4357 of 7514 [58%]) and Hispanics (964 of 1220 [79%], 341 of 657 [52%]). Compared with white Americans, African American (hazard ratio 1·36, 95% CI 1·16–1·60, p<0·0001) and Asian (1·25, 1·01–1·56, p=0·03) race were independent predictors of poor survival at 2 years. Interpretation African Americans and Asians have a worse outcome after OLT compared with white Americans and Hispanics. The higher rate of chronic rejection in African Americans and a relatively worse outcome in other minority races merits further examination. Lancet 2002; 359: 287–93
Division of Gastroenterology and Hepatology, Ochsner Medical Foundation, New Orleans, LA, USA (S Nair MD); and Department of Medicine, Johns Hopkins University School of Medicine, Baltimore, MD (J Eustace MD, P J Thuluvath FRCP) Correspondence to: Dr Paul J Thuluvath, Johns Hopkins University School of Medicine, Room 429, 1830 Bldg, 1830 E Monument Street, Baltimore, MD 21205, USA (e-mail: [email protected])
THE LANCET • Vol 359 • January 26, 2002 • www.thelancet.com
Introduction Due to an overall increase in the number of liver transplantations, a large number of African Americans, Hispanics, and Asians have undergone orthotopic liver transplantation (OLT) in the past decade. From published reports, it is not clear whether there is a difference in survival among the different races following OLT. There is significant evidence to suggest that the long-term survival in African Americans is lower compared with white Americans after renal transplantation.1–7 Inferior histocompatibility matching, poor compliance, and a lower socioeconomic status have been implicated for the poor survival among African Americans.1–8 However, it is unclear whether these observations might be extrapolated to patients undergoing OLT. A few studies have analysed the survival among African Americans after OLT, with conflicting results.9–14 Some studies reported a significantly lower survival for African Americans, whereas other studies reported a better or equivalent rate of survival.9–14 Several factors could explain these conflicting results; most of these studies had small number of African Americans and, moreover, these studies did not control for confounding factors such as United Network of Organ Sharing (UNOS) status, the cause of liver disease, renal function, diabetes, and obesity.15–22 In this study, we have analysed UNOS data for 1988 to 1996 to establish whether race was an independent predictor of survival after OLT and to identify the possible reasons behind such a difference, if it existed.
Patients and methods We extracted data from the UNOS transplant registry for all liver transplantations done in the USA between 1988 and 1996. Complete follow-up data were available on 14 282 patients at 2 years, 8988 patients at 5 years, and 600 patients at 10 years. The following information was available on both donor and recipients: age, sex, race, blood group, and cause of death. In addition, we collected data on recipients’ bodyweight and height at the time of listing, UNOS status, serum creatinine concentration at the time of transplantation, presence of diabetes, cause of liver disease, incidence of primary graft non-function and chronic rejection, 2-year graft survival, and 2-year patient survival. The rate of retransplantation and the interval between the initial transplantation and retransplantation were also recorded. We excluded patients whose race was not reported or was listed as other. Liver transplantation in the paediatric age-group (less than 18 years) was also excluded. UNOS status at the time of transplantation was recorded as status 1, 2, 3, and 4; these numbers do not correspond to the current listing criteria and status (currently UNOS status is recorded as 1, 2a, 2b, and 3).
287
For personal use. Only reproduce with permission from The Lancet Publishing Group.
ARTICLES
However, even under the older system, status 1 was applied to sicker patients requiring intensive care, although under the current system many of these patients would be categorised as 2a. The donor and recipient matching based on blood type was classified into three groups: ABO matched when donor and recipient were of same blood; ABO mismatched when donor blood type was O and recipient was of other blood groups; and ABO mismatched when donor blood group was A, B, or AB and recipient of another blood group. We classified the cause of liver disease as alcoholic cirrhosis, hepatitis C, hepatitis B, alcoholic and hepatitis C, primary biliary cirrhosis, primary sclerosing cholangitis, autoimmune liver disease, cryptogenic cirrhosis, and acute liver disease (fulminant liver failure), and other. Patients with a diagnosis of non-A or non-B hepatitis (1988–90 data) were included as hepatitis C. The cause of acute liver disease was not consistently included in the UNOS data and hence not available for analysis. We calculated body-mass index for all patients from weight and height.2 Renal function was arbitrarily classified into normal (serum creatinine ⭐88·4 mol/L [1 mg/dL]), mild renal insufficiency (serum creatinine 88·4–133 mol/L [1–1·5 mg/dL]), moderate renal failure (serum creatinine 141·4–221 mol/L [1·6–2·5 mg/dL]), and severe renal failure (serum creatinine ⭓229·8 mol/L [⭓2·6 mg/dL]). A serum creatinine of 88·4 mol/L or less was used to define normal to avoid underestimation of renal insufficiency since loss of muscle mass was common in patients with cirrhosis.23 Statistical analysis We used the 2 test for categorical variables and ANOVA for continuous variables to compare various outcome measures and pretransplantation variables between the groups. Mann-Whitney U test was used for comparison of continuous data that were not normally distributed. Patient survival at 2 years was used as the primary endpoint. The effect of race of the recipient, age (recipient and donor), sex, blood group, UNOS status, cold ischaemia time, renal function, body-mass index, and cause of liver disease on patient and graft survival was determined by univariate analysis. We examined the effect of race on graft and patient survival with the Cox Demographics and clinical factors Age (mean [SD], years) Sex (female/male) Blood group O ABO matched Body-mass index >35 (kg/m2) Serum creatinine (mol/L)‡ Normal creatinine (%) Diabetes (%) Donor age (years) Donor >50 years (%) UNOS status 1 (%) Cold ischaemia time (h)‡ Cause of liver disease Alcoholic liver disease Hepatitis C Hepatitis B Primary biliary cirrhosis Pimary sclerosing cholangitis Autoimmune hepatitis Cryptogenic cirrhosis Others§
proportional hazard model. The model was then further adjusted for the various confounders for which information was available. Overall survival was analysed by Kaplan-Meier survival analysis, and Cox-Mantel test was used to determine whether there were any differences between the groups. For this analysis, the patients who had retransplantation were excluded. For all analyses, a type I error of 0·05 was used. We used SPSS (version 9) for data analysis.
Results During the study period, 23 884 (3207 paediatric) patients underwent OLT. 3394 patients had incomplete UNOS status data. Race was not reported for 92 patients and race was reported as other in another 522 patients; these patients were excluded from the study. A total of 16 669 patients who had OLT during this period were available for final analysis. Of the 16 669 patients who had OLT, 13 599 (82%) were white Americans, 1042 (6%) were African Americans, 1534 (9%) were Hispanics, and 494 (3%) were Asians. Among the donors, there were 80% white Americans, 10% African Americans, 8% Hispanics, and 1% Asians. During the study period, there was a steady increase in the total number of African American donors, but there was no proportional increase in the number of African American recipients. The proportion of African Americans who received OLT declined steadily while there was a three-fold increase in the proportion of Hispanic and Asian recipients. African Americans were significantly younger than the other groups (table 1). More African Americans had UNOS status 1 at the time of transplantation than white Americans. Data on diabetes were available only for 65% of patients; the proportions of patients with diabetes and severe obesity (a body-mass index of more than 35 kg/m2) were similar among the four races. Similarly, there were no differences in cold ischaemia time, donor age, and serum creatinine concentration among the groups. African Americans and Asians had a significantly lower proportion of ABO-matched transplantation compared with white Americans or Hispanics. The most common cause of chronic liver disease was hepatitis C for African
White (n=13 599)
African American (n=1042)
Hispanics (n=1534)
Asians (n=494)
48·9 (11·0) 42/58 5711 (42%) 12 239 (90%) 896 (14%) 118 (101) 7629 (57%) 709 (5%) 32·7 (15·4) 2201 (16%) 1931 (14%) 10·6 (5·3)
41·3 (12·2)* 58/42 498 (48%) 906 (87%) 65 (12%) 121 (117) 588 (56%) 60 (6%) 32·2 (15·8) 174 (17%) 210 (20%)* 10·5 (5·6)
47·5 (11·7) 48/52 767 (50%) 1396 (91%) 99 (14%) 114 (108) 864 (57%) 121 (7%) 32·8 (16·5) 295 (19%) 229 (15%) 10·4 (5·8)
50·4 (11·1) 41/59 159 (32%)† 425 (86%) 11 (4%)* 158 (97) 276 (55%) 34 (6%) 31·5 (16·4) 93 (19%) 88 (18%) 10·5 (5·4)
18% 19% 4% 10% 10% 4% 10% 25%
11% 17% 3% 3%* 11% 11% 8% 36%
14% 25% 3% 9% 3% 5% 12% 29%
4% 19% 28%* 2% 2% 2% 5% 38%
*p<0·05 compared with white people; †Most common blood group among Asians was A (34%); ‡Not normally distributed, compared with Mann-Whitney U test; §Others included combination of diseases and other rare disorders.
Table 1: Demographic and clinical characteristics of recipients
288
THE LANCET • Vol 359 • January 26, 2002 • www.thelancet.com
For personal use. Only reproduce with permission from The Lancet Publishing Group.
ARTICLES
White (n=13 599) Outcome measures Primary graft non-function 30-day mortality Chronic graft failure rejection 2-year graft survival Retransplantation† 2-year patient survival 5-year patient survival Cause of death‡ Infections Cardiovascular Cerebrovascular accident Renal failure Malignancy§ Lymphoproliferative diseases Operative death Respiratory failure
African American (n=1042)
Hispanics (n=1534)
Asians (n=494)
5% 6% 6% 74% 28% 83% 58%
7% 9%* 12%* 68%* 22% 74%* 48%
6% 7% 6% 72% 32% 79% 52%
5% 8% 7% 64%* 18% 69%* 37%
38% 19% 8% 3% 19% 2% 6% 5%
43% 17% 7% 4% 13% 1% 9% 7%
46% 16% 9% 2% 12% 1% 10% 4%
38% 19% 10% 3% 20% 1% 5% 2%
*p<0·05 compared with white people; †% of those who developed graft failure within 2 years; ‡In 2762 patients, miscellaneous excluded; §Excludes lymphoproliferative diseases.
Table 2: Outcome measurements among the races
Americans, white Americans, and Hispanics, whereas hepatitis B was the most common cause among Asian recipients. Acute liver failure was the indication for OLT in 20% of African American patients. Table 2 shows the outcome measurements among the races. The overall mean graft and patient survivals during this period were 73% and 79%, respectively. There was a modest decrease in graft survival during the early half of the study period, but this stabilised during the later part. However, there was a modest increase in patient survival over the later half of the study period. This difference had resulted in a widening of the gap between patient survival and graft survival (figure 1). The primary graft non-function was marginally higher in African-Americans than in white Americans. 2-year graft survival was significantly lower in African Americans and Asians than in white Americans and Hispanics. During the study period, more white Americans than people from any other race had retransplantation. Among patients who lost the graft within 2 years, the retransplantation rate was 20–30%; the retransplantation rate was lowest among Asians (18%) followed by African Americans (22%). The rate of graft failure due to chronic rejection was almost twice as high among African Americans compared with other races. 30-day mortality was also higher in African Americans than the other races. 2-year patient survival was 90 2-year patient survival 2-year graft survival
Survival (%)
84 78 72 66 60 1988 1989 1990 1991 1992 1993 1994 1995 1996 Year of transplantation Figure 1: 2-year graft and patient survival according to year of liver transplantation
THE LANCET • Vol 359 • January 26, 2002 • www.thelancet.com
significantly lower for African Americans and Asians compared with white Americans and Hispanics. Information on 5-year survival was available only for 8988 patients (7514 white, 565 African American, 657 Hispanics, and 252 Asians). We believe that the data entry was perhaps incomplete at 5 years and therefore patients with incomplete data were censored and were considered “lost for follow up”. This reduced the number of patients at risk and might have under-estimated survival marginally (as there could be a bias in entering death). 5-year survival was lower for Asians and African Americans than for white Americans and Hispanics. Cause of death was available for 2762 patients (table 2). Infection was the leading cause of death, followed by cardiovascular diseases. Malignancy (excluding lymphoproliferative disease) was the second commonest cause of death in Asians and was mainly due to the recurrence of hepatocellular carcinoma. There were no significant differences in the cause of death among different races. Table 3 lists the graft and patient survival among the four races for various liver diseases. African Americans with hepatitis C had a lower survival rate than white Americans and Hispanics. Likewise, survival among African Americans was lower for acute liver failure. The 2-year survival for hepatitis B was similar with other diseases for white Americans, African Americans, and Hispanics. However, Asians with hepatitis B had a much lower survival rate than other races after OLT. The Hispanics who Cause of liver disease White African (n=13 599) American (n=1042) Demographics Alcoholic liver disease Hepatitis C Hepatitis B Autoimmune PSC PBC Acute liver failure Cryptogenic cirrhosis
74/80 72/80 70/74 78/83 81/88 82/87 68/70 74/79
70/74 67/72† 71/80 70/78 83/86 81/88 59†/64† 62/67†
Hispanics Asians (n=1534) (n=494)
78/82 70/78 77/77 62†/70† 76/89‡ 83/88 70/82 70/75
80/80 69/77 56†/60† 68/80 64/70‡ 66/87‡ 73/84‡ 68/68‡
Results are presented as graft survival/patient survival. PSC=primary sclerosing cholangitis; PBC=primary biliary cirrhosis. *2-year graft and patient survival in % adjusted to nearest full number; †p<0·05 compared with white people; ‡Number of patients less than 50.
Table 3: 2-year graft and 2-year patient survival* among races based on the cause of liver disease
289
For personal use. Only reproduce with permission from The Lancet Publishing Group.
ARTICLES
Variable
Risk ratio (95% CI)
p
Variable
Patient survival Hazard ratio* (95% CI) p
Races (reference: white Americans) African Americans Hispanics Asians
1·34 (1·17–1·52) 0·98 (0·86–1·11) 1·78 (1·51–2·1)
0·0001 0·77 0·0001
Age
1·02 (1·01–1·02)
Sex (reference: male)
0·90 (0·84–0·96)
0·0001
Race (reference: white Americans) African Americans Hispanics Asians
1·36 (1·16–1·60) 1·01 (0·86–1·18) 1·25 (1·01–1·56)
0.007 0·89 0·03
0·004
Age
1·01 (1·01–1·02)
0·0001
Body-mass index
1·00 (0·99–1·01)
0·63
Sex (reference: male)
0·99 ( (0·90–1·09)
0·87
Serum creatinine*
1·17 (1·15–1·20)
0·0001
Donor age
1·01 (1·01–1·02)
0·0001
Donor age
1·01 (1·01–1·02)
0·0001
Creatinine
1·03 (1·10–1·16)
0·0001
ABO mismatch (donor type O)
1·32 (1·14–1·53)
0·0002
Cold ischaemic time
1·01 (0.99–1·01)
0·39
ABO mismatch (donor type A or B or AB) 1·29 (1·05–1·59)
0·0157
ABO mismatch (donor type O)
1·06 (0·91–1·2)
0·41 0·09
Cold ischaemia time
1·01 (1·00–1·02)
0·0009
ABO mismatch (donor type A, B or AB) 1·20 (0·96–1·50)
UNOS status (reference: status 4) UNOS status 1 UNOS status 2 UNOS status 3
2·14 (1·78–2·59) 1·52 (1·2–1·8) 0·94 (0·8–1·13)
0·0001 0·001 0·57
UNOS status (reference: status 4) UNOS status 1 UNOS status 2 UNOS status 3
1·66 (1·35–2·04) 1·23 (1·01–1·50) 0·90 (0·75–1·09)
0·0001 0·04 0·3
Cause of liver disease (reference: PSC)† Alcoholic liver disease 1·67 (1·41–1·99) Hepatitis C virus 1·79 (1·49–2·15) Hepatitis B virus 2·5 ( 2·02–3.02) Autoimmune 1·66 (1·46–2·13) PBC 1.08 (0·88–1·37) Acute liver failure 2·69 (2·10–3·30) Cryptogenic cirrhosis 1·85 (1·51–2·26)
0·0001 0·0001 0·0001 0·0001 0·40 0·0001 0·001
Cause of liver disease (reference: PSC) Alcoholic liver disease 1·50 (1·21–1·85) Hepatitis C 1·63 (1·32–2·02) Hepatitis B 2.38 (1·85–3·0) Autoimmune 1·95 (1·49–2·5) Primary biliary cirrhosis 1·00 (0·78–1·3) Acute liver failure 1·84 (1·42–2·39) Cryptogenic cirrhosis 1·63 (1·30–2·0)
0·0001 0·0001 0·0001 0·0001 0·9 0·0001 0·0001
Table 4: Results of univariate analysis for patient survival at 2 years
had OLT for autoimmune hepatitis had a lower survival rate than did African Americans and white Americans. Table 4 shows the results of univariate analysis of various pretransplant characteristics as a determinant of patient survival at 2 years. Univariate analysis showed that race, age, sex, donor age, presence of diabetes, serum creatinine, UNOS status, cold ischaemia time, and cause of liver disease were significant predictors of patient survival at 2 years. In univariate analysis, when compared with primary sclerosing cholangitis, hepatitis B was a significant predictor of poor survival (hazard ratio 2·5, 95% CI 2·02–3·02) as were alcoholic cirrhosis, hepatitis C, and autoimmune liver diseases. All variables that were reported to be significant by univariate analysis were entered into the Cox regression model, and the results are summarised in table 5. As compared with white Americans, African American race was an independent predictor of poor survival at 2 years. Other significant independent predictors of poor survival were UNOS status 1, serum creatinine, and donor and recipient age. As expected, cause of liver disease had a significant effect on 2-year survival. Since there were more African Americans with UNOS status 1, a strong independent predictor of survival, we analysed the data for only UNOS status 1 and also by excluding UNOS status 1. Although multivariate model simultaneously adjusts for other confounding variables, we did a subgroup analysis to confirm the robustness of our findings. When only UNOS status 1 patients were analysed, African American race was an independent predictor (hazard ratio 1·5, 95% CI 1·14–2·0, p=0·004). Similar results were obtained when UNOS status 1 was excluded (1·3,
290
Patients with no reported UNOS status, or those with cause of liver disease other than the one used in the univariate analysis were excluded. PSC=primary sclerosing cholangitis. *Adjusted hazard ratio.
Table 5: Results of multivariate Cox regression model for patient survival at 2 years
1·1–1·6, p=0·003). We also analysed the data at 5 years with Cox regression model. African American (risk ratio 1·5, 95% CI 1·3–1·7, p<0·001), Asian (hazard ratio 1·4, 95% CI 1·2–1·7, p<0·001), and Hispanic (hazard ratio 1·2, 95% CI 1·09–1·4, p=0·002) race were all independent predictors of survival at 5 years. Figure 2 shows the Kaplan-Meier curve for patient survival. Survival rate was significantly lower for African Americans and Asians, but not for Hispanics, when compared with white Americans. The lower survival of African Americans and Asians 1·0 0·9 Cumulative survival
PSC=primary sclerosing cholangitis; PBC=primary biliary cirrhosis. *Serum creatinine at the time of OLT; †Patients with more than one disease (acute liver disease+hepatitis C, hepatitis B+C, hepatitis B+D), Budd Chiari syndrome, metabolic disorders, and other rare disorders are not included.
0·8 0·7 0·6 0·5 0·4
White Americans Hispanics African Americans Asians
0·3 0·2 0·1 0 0
Number of patients White Americans 13599 Hispanics 1534 African Americans 1042 494 Asians
1000
2000 3000 Survival (days)
8523 861 570 240
3627 271 213 70
1006 68 58 11
4000
5000
109 2 5 0
0 0 0 0
Figure 2: Kaplan-Meier survival curve for different races Survival was significantly lower for African Americans and Asians compared with white Americans.
THE LANCET • Vol 359 • January 26, 2002 • www.thelancet.com
For personal use. Only reproduce with permission from The Lancet Publishing Group.
ARTICLES
Variable
Hazard ratio* (95% CI) p
Race (reference: white Americans) African Americans Hispanics Asians
1·16 (1·01–1·33) 1·02 (0·89–1·15) 1·18 (0·98–1·43)
0·03 0·79 0·07
Age
1·01 (1·00–1·01)
0·0001
Sex (reference: male)
1·10 (0·96–1·27)
0·15
Donor age
1·02 (1·01–1·02)
0·0001
Creatinine
1·09 (1·07–1·12)
0·0001
Cold ischaemia time
1·01 (0·99–1·01)
0·08
ABO mismatch (donor type O)
1·22 (1·02–1·46)
0·03
ABO mismatch (donor type A, B, or AB) 1·13 (0·85–1·51)
0·97
UNOS status (reference: status 4) UNOS status 1 UNOS status 2 UNOS status 3
1·73 (1·48–2·07) 1·23 (1·04–1·47) 1·03 (0·87–1·06)
0·0001 0·015 0·72
Cause of liver disease (reference: PSC) Alcoholic liver disease 1·16 (0·98–1·37) Hepatitis C 1·40 (1·19–1·65) Hepatitis B 1·69 (1·37–2·08) Autoimmune 1·47 (1·18–1·83) PBC 0·89 (0·53–1·46) Acute liver failure 1·48 (1·20–1·82) Cryptogenic cirrhosis 1·29 (1·09–1·55)
0·07 0·0001 0·0001 0·0001 0·24 0·0001 0·004
PSC=primary sclerosing cholangitis; PBC=primary biliary cirrhosis. Patients with no reported UNOS status, or those with causes other than the one used in univariate analysis were excluded. *Adjusted hazard ratio.
Table 6: Results of multivariate Cox regression model for graft survival at 2 years
was evident within a year of transplantation, but the difference became more evident with time. Table 6 shows the Cox model for graft survival at 2 years. Kaplan-Meier analysis with Cox-Mantel test for comparison revealed a significantly lower graft survival for African Americans (p=0·03), with a trend toward an increased graft loss in Asians (p=0·07). The mean graft survival for African Americans was at least 300 days less than for white Americans and Hispanics. Graft loss was due to rejection, primary non-function, recurrent disease, vascular complication, or patient death. Race was an independent predictor of graft survival at 2 years. The other independent predictors of survival were cold ischaemic time, creatinine, donor age, and cause of liver disease. In patients with diabetes for whom we had data (n=3575), presence of diabetes was an independent predictor of poor 2-year graft (hazard ratio 1·29, 95% CI 1·08–1·54, p=0·003) and patient (1·34, 1·09–1·66, p=0·0001) survival.
Discussion In this study, we have shown that African American race was an independent predictor of patient survival after OLT. Other independent predictors of survival were UNOS status 1, donor and recipient age, serum creatinine concentration, diabetes, and cause of liver disease. Several reasons have been proposed to explain the lower survival in African Americans after renal transplantation, such as poor HLA matching between the donor and recipient, which leads to an increase in the number of rejection episodes and graft loss, poor socioeconomic status, and lack of insurance benefits resulting in inadequate post-transplant care.1–8 Some of these explanations, especially poor compliance, although they remain unsubstantiated, might also be applicable to liver transplantation recipients. In our study, only two pretransplant
THE LANCET • Vol 359 • January 26, 2002 • www.thelancet.com
variables, recipient age and UNOS status, were significantly different in the African Americans. The African Americans were at least 7 years younger than white Americans, a factor that should have favourably affected survival. By contrast, there was a higher proportion of African Americans with UNOS status 1, which indicated that these patients were probably very sick at the time of transplantation. Moreover, acute fulminant failure was a common indication for OLT among African Americans. Fulminant hepatic failure and UNOS status 1 are generally associated with a lower patient survival after OLT.24,25 The higher 30-day mortality in African Americans could be explained by these observations. The poor survival among African Americans could also be explained by the higher frequency of primary graft non-function in African Americans than in white Americans. In addition, the rate of chronic rejection was higher among African Americans. It is possible that a combination of these factors might have resulted in a 2-year survival that is 10% less in African Americans than in white Americans. Although poor compliance and inadequate insurance among African Americans have been blamed for the higher rate of chronic rejection, there is no convincing evidence to support these suggestions. It is more likely that immunological factors, yet unidentified, might be contributing to chronic rejection. Moreover, most of the currently available immunosuppressive drugs were tested in predominantly white Americans and there might be a need to test these drugs more rigorously in minorities. Although the survival of Asians was lower compared with other races, when patients with hepatitis B were excluded, the survival rate at 2 years was similar to African Americans. The second common cause of death among Asians was malignancy, which reflects a higher frequency of recurrent hepatocellular carcinoma (data not shown). The disease-specific survival analysis among races yielded some interesting findings. The survival of Hispanics and whites were similar for most causes, but African Americans showed a lower survival with hepatitis C and acute failure. Many small studies have shown that African Americans were less likely to have sustained hepatitis C virus clearance with medical treatment compared with white Americans.26 Our analysis showed that African Americans had a poorer 2-year survival after OLT for hepatitis C, an observation that had not been reported before. It is possible that African Americans have a relatively higher viral load when they undergo OLT and hence have a more severe recurrent hepatitis C. Another possibility is that these patients had more episodes of rejection that required steroid boluses and monoclonal antibody use, factors implicated in a more aggressive disease recurrence. Survival for patients with hepatitis B was similar to that for those with other liver diseases for all races except in Asians who had a relatively poor 2-year survival. It is possible, although we have no data, that more Asians had DNA for hepatitis B virus or were HBeAg positive (replicative phase) at the time of OLT than other groups. Recurrent hepatocellular carcinoma could also partly explain the lower survival.27 Primary sclerosing cholangitis and primary biliary cirrhosis had the best survival for all races.
291
For personal use. Only reproduce with permission from The Lancet Publishing Group.
ARTICLES
We identified a number of predictors of survival in our univariate analysis and our observations are consistent with previous reports.28–35 In regression analysis, most of the factors that were predictive in univariate analysis were also significantly associated with lower survival except sex of the patient. Diabetes was also reported to be a significant predictive factor when data were analysed including only those patients with that information in the database; most of these patients were transplanted between 1994 and 1996. In this subgroup, hepatitis B was not an independent predictor of survival. This finding, not totally expected, was probably due to careful selection of patients with hepatitis B and hepatocellular carcinoma, better treatment of hepatitis B with antiviral agents, which reduced the pretransplant viral load, and use of posttransplant immunoglobulin during this period and thereafter.36-39 In summary, African Americans had a lower survival than white Americans and Hispanics after OLT and African American race was an independent predictor of a lower patient survival. Higher postoperative mortality (30-day mortality) due to a higher number of seriously ill patients (UNOS status 1 and fulminant hepatic failure), and higher rates of chronic rejection and primary graft non-function could partly explain some of these observations. However, we do not believe that any of these observations could explain a 5-year survival that is 17% less in African Americans compared with white Americans. Our study suggests that there is a clear need for prospective studies to examine our observations further. Until then, the reasons for poor survival in African Americans will remain speculative and will probably be dismissed as being due to poor compliance with therapy. Moreover, the higher rate of chronic rejection in African Americans suggests that there should be more rigorous drug trials in this patient population.
7
8
9
10
11
12 13
14
15
16
17
18
19
20
21
Contributors S Nair was involved in the collection, analysis, and interpretation of the data, and the drafting of the manuscript. J Eustace analysed and interpreted the data, provided statistical expertise, and helped revise the manuscript. P J Thuluvath contributed to the concept and design of the study, collected, interpreted, and analysed the data, and helped revise the manuscript.
Conflict of interest statement None declared.
22 23
24
25
Acknowledgments We thank UNOS for providing us with the data.
26
References
27
1
2 3
4
5
6
Isaacs RB, Nock SL, Spencer CE, et al. Racial disparities in renal transplant outcomes. Am J Kidney Dis 1999; 34: 706–12. Hariharan S, Schroeder TJ, First MR. Effect of race on renal transplant outcome. Clin Transplant 1993; 7: 235–39. Valente JF, Hariharan S, Peddi VR, et al. Causes of renal allograft loss in black vs white transplant recipients in the cyclosporine era. Clin Transplant 1997; 11: 231–36. Flechner SM, Modlin CS, Serrano DP, et al. Determinants of chronic renal allograft rejection in cyclosporine-treated recipients. Transplantation 1996; 62: 1235–41. Gjertson DW. A multi-factor analysis of kidney graft outcomes at one and five years post-transplantation: 1996 UNOS Update. Clin Transpl 1996; 343–60. Kasiske BL, Neylan JF III, Riggio RR, et al. The effect of race on access and outcome in transplantation. N Engl J Med 1991; 324: 302–07.
292
28
29
30
31
32
Gallichio MH, Hudson S, Young CJ, Diethelm AG, Deicrhoi MH. Renal retransplantation at the University of Alabama at Birmingham: incidence and outcome. Clin Transpl 1998; 169–75. Ojo AO, Port FK, Held PI, et al. Inferior outcome of two-haplotype matched renal transplants in blacks: role of early rejection. Kidney Int 1995; 48: 1592–99. Dawson S III, Imagawa DK, Johnson C, et al. UCLA liver transplantation: analysis of immunological factors affecting outcome. Artif Organs 1996; 20: 1063–72. Eckhoff DE, McGuire BM, Young CJ, et al. Race: a critical factor in organ donation, patient referral and selection, and orthotopic liver transplantation? Liver Transpl Surg 1998; 4: 499–505. Teperman L, Scantlebury V, Tzakis A, Staschak S, Todo S, Starzl TE. Liver transplantation in Black recipients: Pittsburgh. Transplantation Proc 1989; 21: 3963–65. Belle SH, Detre KM. Report from Pitt-UNOS liver transplantation registry. Transplantation Proc 1993; 25: 1137–42. Gonwa TA, Morris CA, Mai ML, Husberg BS, Goldstein RM, Klintmalm GB. Race and liver transplantation. Arch Surg 1991; 126: 1141–43. Pillay P, Van Thiel DH, Gavaler JS, Starzl TE. Effect of race upon organ donation and recipient survival in liver transplantation. Dig Dis Sci 1990; 35: 1391–96. Ratziu V, Samuel D, Sebagh M, et al. Long-term follow-up after liver transplantation for autoimmune hepatitis: evidence of recurrence of primary disease. J Hepatol 1999; 30: 131–41. Charlton M, Seaberg E, Wiesner R, et al. Predictors of patient and graft survival following liver transplantation for hepatitis C. Hepatology 1998; 28: 823–30. Graziadei IW, Wiesner RH, Marotta PJ, et al. Long-term results of patients undergoing liver transplantation for primary sclerosing cholangitis. Hepatology 1999; 30: 1121–27. Shields PL, Tang H, Neuberger JM, Gunson BK, McMaster P, Pirenne J. Poor outcome in patients with diabetes mellitus undergoing liver transplantation. Transplantation 1999; 68: 530–35. Fraley DS, Burr R, Bernardini J, Angus D, Kramer DJ, Johnson JP. Impact of acute renal failure on mortality in end-stage liver disease with or without transplantation. Kidney Int 1998; 54: 518–24. Sawyer RG, Pelletier SJ, Prett TL. Increased early morbidity and mortality with acceptable long-term function in severely obese patients undergoing liver transplantation. Clin Transplant 1999; 13: 126–30. Braunfeld MY, Chan S, Pregler J, et al. Liver transplantation in the morbidly obese. J Clin Anesth 1996; 8: 585–90. Keeffe E, Gettys C, Esquivel CO. Liver transplantation inpatients with severe obesity. Transplantation 1994; 57: 309–11. Caregaro L, Menon F, Angeli P, et al. Limitations of serum creatinine level and creatinine clearance as filtration markers in cirrhosis. Arch Intern Med 1994; 154: 201–05. Bismuth H, Samuel D, Castaing D, et al. Orthotopic liver transplantation in fulminant and subfulminant hepatitis. The Paul Brousse experience. Ann Surg 1995; 222: 109–19. Devlin J, Wendon J, Heaton N, Tan KC, Williams R. Pretransplantation clinical status and outcome of emergency transplantation for acute liver failure. Hepatology 1995; 21: 1018–24. Reddy KR, Hoofnagle JH, Tong MJ, et al. Racial differences in responses to therapy with interferon in chronic hepatitis C. Consensus Interferon Study Group. Hepatology 1999; 30: 787–93. Beasley RP. Hepatitis B virus: the major etiology of hepatocellular carcinoma. Cancer 1988; 61: 1942–56. Grande L, Matus D, Manyalic M, et al. Effect of donor age on graft outcome after liver transplantation. Transplant Proc 1999; 31: 2482–83. Wiesner RH, Demetris AJ, Belle SH, et al. Acute hepatic allograft rejection: incidence, risk factors, and impact on outcome. Hepatology 1998; 28: 638–45. Tisone G, Romagnoli J, Vennarecci G, et al. Is donor age a risk factor for poor graft function after orthotopic liver transplantation? Transplant Proc 1997; 29: 2825–26. Detre KM, Lombardero M, Belle S, et al. Influence of donor age on graft survival after liver transplantation—United Network for Organ Sharing Registry. Liver Transpl Surg 1995; 1: 311–19. Piratvisuth T, Tredger JM, Hayllar KA, Williams R. Contribution of true cold and rewarming ischemia times to factors determining outcome after orthotopic liver transplantation. Liver Transpl Surg 1995; 1: 296–301.
THE LANCET • Vol 359 • January 26, 2002 • www.thelancet.com
For personal use. Only reproduce with permission from The Lancet Publishing Group.
ARTICLES
33 Hoofnagle JH, Lombardero M, Zetterman RK, et al. Donor age and outcome of liver transplantation. Hepatology 1996; 24: 89–96. 34 Marino IR, Doyle HR, Aldrighetti L, et al. Effect of donor age and sex on the outcome of liver transplantation. Hepatology 1995; 22: 1754–62. 35 Ploeg RJ, D’Alessandro AM, Knechtle SJ, et al. Risk factors for primary dysfunction after liver transplantation—a multivariate analysis. Transplantation 1993; 55: 807–13. 36 Chen PJ, Chen DS. Hepatitis B virus infection and hepatocellular carcinoma: molecular genetics and clinical perspectives. Semin Liver
Dis 1999; 19: 253–62. 37 Burbach GJ, Bienzle U, Neuhaus R, et al. Intravenous or intramuscular anti-HBs immunoglobulin for the prevention of hepatitis B reinfection after orthotopic liver transplantation. Transplantation 1997; 63: 478–80. 38 Ho BM, So SK, Esquivel CO, Keeffe EB. Liver transplantation in Asian patients with chronic hepatitis B. Hepatology 1997; 25: 223–25. 39 Terrault NA, Zhou S, Combs C, et al. Prophylaxis in liver transplant recipients using a fixed dosing schedule of hepatitis B immunoglobulin. Hepatology 1996; 24: 1327–33.
Clinical picture Metastatic calcification Ronald C W Ma, Clive S Cockram
A 72-year-old woman had obstructive jaundice from an inoperable carcinoma of the head of the pancreas. 6 days after bypass surgery, she developed generalised tonic clonic convulsions. Her adjusted serum calcium was 1·36 mmol/L, phosphate 1·25 mmol/L, albumin 24 g/L, and glucose 10·7 mmol/L. Serum parathyroid hormone was <0·1 pmol/L, and magnesium was 0·4 nmol/l. Cerebral computed tomography showed intracranial calcification in the subcortical white matter, basal ganglia, cerebellum, corona radiata, and centrum semiovale, consistent with metastatic calcification from long-standing hypoparathyroidism (figure). We gave her intravenous calcium
and magnesium replacement, and she had no further convulsions. We found a thyroidectomy scar on careful examination, and the patient said that she had had a total thyroidectomy many years previously, and had been taking thyroxine. She had not had regular serum calcium measurements, and did have occasional parasthesiae. She also took a thiazide diuretic, which had been withheld on admission to hospital. The combination of obstructive jaundice and consequent alteration of vitamin D metabolism, thiazide withdrawal, recent bypass surgery, and intravenous fluid administration probably exacerbated her long-standing iatrogenic hypoparathyroidism.
Department of Medicine and Therapeutics, Prince of Wales Hospital, Shatin, Hong Kong (RCW Ma MRCP, CS Cockram FRCP)
THE LANCET • Vol 359 • January 26, 2002 • www.thelancet.com
293
For personal use. Only reproduce with permission from The Lancet Publishing Group.