Original Clinical ScienceçLiver
Cryptococcosis in Patients With Cirrhosis of the Liver and Posttransplant Outcomes Nina Singh, MD,1 Costi D. Sifri, MD,2 Fernanda P. Silveira, MD,1 Rachel Miller, MD,3 Kevin S. Gregg, MD,4 Shirish Huprikar, MD,5 Erika D. Lease, MD,6 Andrea Zimmer, MD,7 J. Stephen Dummer, MD,7 Cedric W. Spak, MD,8 Christine Koval, MD,9 David B. Banach, MD,10 Miloni Shroff, MD,11 Jade Le, MD,11 Darin Ostrander, MD,12 Robin Avery, MD,12 Albert Eid, MD,13 Raymund R. Razonable, MD,14 Jose Montero, MD,15 Emily Blumberg, MD,16 Ahlaam Alynbiawi, MD,17 Michele I. Morris, MD,17 Henry B. Randall, MD,18 George Alangaden, MD,19 Jeffrey Tessier,20 Marilyn M. Wagener,1 and Hsin Yun Sun, MD21 Background. The outcomes and optimal management of cirrhotic patients who develop cryptococcosis before transplantation are not fully known. Methods. We conducted a multicenter study involving consecutive patients with cirrhosis and cryptococcosis between January 2000 and March 2014. Data collected were generated as standard of care. Results. In all, 112 patients were followed until death or up to 9 years. Disseminated disease and fungemia were present in 76.8% (86/112) and 90-day mortality was 57.1% (64/112). Of the 39 patients listed for transplant, 20.5% (8) underwent liver transplantation, including 2 with active but unrecognized disease before transplantation. Median duration of pretransplant antifungal therapy and posttransplant therapy was 43 days (interquartile range, 8-130 days) and 272 days (interquartile range, 180-630 days), respectively. Transplantation was associated with lower mortality (P = 0.002). None of the transplant recipients developed disease progression during the median follow-up of 3.5 years with a survival rate of 87.5%. Conclusions. Cryptococcosis in patients with cirrhosis has grave prognosis. Our findings suggest that transplantation after recent cryptococcal disease may not be a categorical exclusion and may be cautiously undertaken in liver transplant candidates who are otherwise deemed clinically stable.
(Transplantation 2015;99: 2132–2141)
P
atients with end-stage liver disease may develop cryptococcosis as a result of cirrhosis-associated compromised host defenses that include impaired cell-mediated immunity, phagocytic dysfunction, decreased antibody and immunoglobulins, and complement deficiency.1-4 Indeed, cirrhosis is a major risk factor for cryptococcosis and for poor outcomes in those who develop disease.4,5 Cirrhosis-associated cryptococcosis often has high rates of fungemia and mortality.2,6 Increasingly, cryptococcosis has been documented in cirrhotic patients awaiting liver transplantation or in those referred for transplantation.4,7 Anecdotal reports with favorable
outcomes exist for transplants performed inadvertently in patients with unrecognized cryptococcosis or in those treated before transplantation.7,8 A dilemma, however, is whether patients with cryptococcosis during transplant candidacy can safely undergo transplantation. Limited data and guidance are available in this context, and little is known about their outcomes after transplantation.4,7 We conducted a multicenter study to determine the outcome in patients with cirrhosis and cryptococcosis and assess risk factors for death in those who died before transplantation. Additionally, in patients who underwent
Received 4 September 2014. Revision requested 31 October 2014. 15
University of South Florida, Tampa, FL.
16
Department of Medicine, Division of Infectious Diseases, University of Pittsburgh, Pittsburgh, PA.
University of Pennsylvania, Philadelphia, PA.
17
University of Miami, Miller School of Medicine, Miami, FL.
2
University of Virginia, Charlottesville, VA.
18
Department of Surgery, St. Luke's Hospital, Kansas City, MO.
3
University of Iowa, Iowa City, IA.
19
Henry Ford Hospital, Detroit, MI.
4
University of Michigan, Ann Arbor, MI.
20
Virginia Commonwealth University, Richmond, VA.
5
Icahn School of Medicine at Mount Sinai, New York, NY.
21
6
University of Washington, Seattle, WA.
Department of Internal Medicine, National Taiwan University Hospital and National, Taipei, Taiwan.
7
Vanderbilt University, Nashville, TN.
The authors declare no funding or conflicts of interest.
8
Baylor Health System, Dallas, TX.
9
Cleveland Clinic, Cleveland, OH.
N.S. and M.M.W. participated in research design and preparation of the paper. M. M.W. participated in data analysis. All authors participated in performance of the study, review, and editing of the manuscript.
Accepted 8 January 2015. 1
10
Yale School of Medicine, New Haven, CT.
11
University of Texas Southwestern Medical Center, Dallas, TX.
Correspondence: Nina Singh, MD, Infectious Diseases Section, Veterans Affairs Medical Center, University Drive C, Pittsburgh, PA 15240. ([email protected]).
12
Johns Hopkins University, Baltimore, MD.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
13
University of Kansas, Kansas City, KS.
ISSN: 0041-1337/15/9910-2132
14
Mayo Clinic, Rochester, MN.
DOI: 10.1097/TP.0000000000000690
2132
www.transplantjournal.com
Transplantation
■
October 2015
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
■
Volume 99
■
Number 10
Singh et al
© 2015 Wolters Kluwer
transplantation, we assessed the disease status before transplantation, timing of transplantation after diagnosis of cryptocococcosis, posttransplant management, including immunosuppressive regimen, antifungal agent and duration of prophylaxis, disease progression or relapse, and allograft rejection. PATIENTS AND METHODS Participating transplant centers were identified through the American Society of Transplantation Infectious Diseases Community of Practice. Patients were included if they had end-stage liver disease and cryptococcosis regardless of whether they were considered for liver transplant candidacy, referred for liver transplantation, or awaiting liver transplantation. Cryptococcosis was defined based on the European Organization for Research and Treatment in Cancer and the Mycoses Study Group criteria.9 Transplant recipients in whom pretransplant cryptococcosis (positive culture from blood or another clinical specimen, histopathology, or positive antigen) was recognized after transplantation were also included. Patients were excluded if they had developed cryptococcosis after transplantation with no evidence of pretransplant disease or were considered to have donor-derived cryptococcosis. The study involved collection of data generated as standard of care at the study sites between January 2000 and March 2014. Variables collected included demographics, severity of underlying liver disease, comorbid conditions, details of cryptococcal disease, sites of involvement, antifungal therapy, follow-up microbiologic data, and patient outcomes, including after transplantation in those who underwent transplantation. Fungemia, involvement of 2 or more noncontiguous organ sites or the central nervous system (CNS) were regarded as disseminated cryptococcal disease. Immune reconstitution syndrome was defined as previously reported.10 The cause of death was determined by the site investigator. Institutional Review Board approval for this observational study was obtained per local requirements. Statistical Analysis
Stata/SE (version 13.1, Stata Corp, College Station, TX) was used for statistical analyses. Odds ratios and 95% confidence intervals for mortality were estimated using univariate logistic models. A multivariate model was constructed using variables found to be associated with outcome in univariate analysis. The final model was tested using a Pearson goodness of fit test. Similarly, for patients on the transplant list, factors were examined for differences between patients that were transplanted and those that were not. RESULTS In all, 112 cirrhotic patients with cryptococcosis were identified at the participating sites. One of the patients has previously been reported as a case report (7). Patients were followed until death or for up to 9 years; total follow-up in the study population was 80 patient-years. Clinical and demographic characteristics of these patients are shown in Table 1. The median model for end-stage liver disease score of the patients was 22 (interquartile range [IQR] 18-30) and 66% (62/94) were Child-Pugh class C. Comorbid conditions included diabetes mellitus in 29.5% (33/112), malignancy in 15.2% (17/112), and renal replacement therapy
2133
in 13.4% (15/112). Additionally, 20.5% (23/112) were receiving corticosteroids, 2.6% (3/112) had sarcoidosis, and 1 patient had idiopathic CD4+ lymphopenia. None had human immunodeficiency virus infection. Characteristics of Cryptococcal Disease
The diagnosis of cryptococcosis was based on positive culture in 85.7% (96/112) of the patients, including fungemia and positive cerebrospinal fluid (CSF) culture in 47 and 40 cases, respectively. Of patients without culture-based diagnosis, 81.3% (13/16) had positive cryptococcal antigen (CSF and/or serum antigen) and in 3 patients, the diagnosis was based on tissue histopathology. Sites of cryptococcosis included CNS in 48.2% (54/112), pulmonary in 37.5% (42/ 112), peritoneal cavity in 19.6% (22/112), and skin in 2.7% (3/112). Disseminated disease and fungemia were present in 76.8% (86/112) and 46.5% (47/101) of the patients, respectively. Of patients with CNS disease, 83.3% (40/48) had positive culture and 91.1% (41/45) had positive CSF cryptococcal antigen (median titer, 1:128; IQR, 1:16 to 1:1024). Median time from onset of symptoms to diagnosis was 9 days (IQR, 5-21 days). In 15.2% (17/112) of the patients, the diagnosis was established post mortem. The antifungal regimen used in the study patients is shown in Table 1. The most frequently used primary antifungal agent was fluconazole (42.7%, 41/96) followed by liposomal amphotericin B (31.3%, 30/96). Concurrent 5-flucytosine was used in 40.6% of the patients (39/96) (Table 1). Outcome and Risk Factors for Mortality
Overall, mortality at 90 days was 57.1% (64/112). Deaths (n = 64) occurred a median of 11 days (IQR, 3-20) after diagnosis and median of 10.5 days (IQR, 6-22 days) after initiation of antifungal therapy in those treated (n = 46). Death was attributable to cryptococcal disease in 54.6% (35/64) of the patients; other causes of death are listed in Table 2. The median time to death after diagnosis was significantly earlier in patients in whom death was attributable to cryptococcosis compared to other causes (day 4 vs day 21, P = 0.0001). Variables associated with mortality are shown in Table 3. Mortality was significantly higher in patients with more advanced liver disease (Child-Pugh [P = 0.007] and model for end-stage liver disease score [P = 0.002]), renal replacement therapy (P = 0.0240), mechanical ventilation (P = 0.007), hypotension/shock (P = 0.004), impaired mental status (P = 0.023), and fungemia (P < 0.001). On the contrary, headache at presentation (P = 0.014) and transplantation (P = 0.002) were associated with improved outcome (Table 3). Patients with headache had significantly higher CSF white blood cell (median, 70 [IQR, 22-120] vs 6.5 [IQR, 2-98], P = 0.011) and shorter time to initiation of therapy (median, 1 [IQR, 0-2] vs 3 [IQR, 1-6] days, P = 0.012) than those without headache. Transplantation and Posttransplant Outcomes
Of 112 patients, 46 were liver transplant candidates, that is, listed for transplantation at any time during the study. Excluding 7 candidates who died before the diagnosis of cryptococcosis (the clinical cultures yielded Cryptococcus after death), 20.5 % (8/39) of the listed patients underwent liver transplantation that included liver alone in 7 patients and liver-kidney in 1 patient. Of the 8 transplant recipients, 4 had disseminated disease (including 3 with meningitis), 2 had pulmonary only, and 2 had extrapulmonary cryptococcosis
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
2134
Transplantation
■
October 2015
■
Volume 99
■
Number 10
www.transplantjournal.com
TABLE 1.
Demographic and clinical characteristics of the study patients (n = 112) Factor
Value
Age: median (IQR), y Sex: male, % Underlying liver disease (may have more than one) Hepatitis C virus Alcohol Hepatitis B virus Nonalcoholic steatohepatosis Hepatocellular carcinoma Cryptogenic Primary biliary cirrhosis Fulminant hepatic failure Autoimmune hepatitis Primary biliary cirrhosis Hemochromatosis α 1-antitrypsin deficiency Wilson disease Child-Pugh classificationa A B C MELD score, median (IQR)b Comorbid conditions Diabetes mellitus Corticosteroid use Renal replacement therapy Sarcoidosis Malignancy (other than hepatocellular carcinoma)c Idiopathic CD4+ lymphopenia Clinical manifestations Fever Headache Decrease in visual acuity Mental status Alert Nonalert Unknown Hypotension/shock Characteristics of cryptococcal disease Sites of involvement Central nervous system Pulmonary Peritoneal Cutaneous Fungemia Disseminated disease Serum antigen positivity Serum cryptococcal antigen titer, median (IQR) CSFd characteristics Opening pressuree: median (IQR), cm > 20 cm White blood cells, median (IQR) per mm3 Glucose: median (IQR), mg/dL Protein: median (IQR), g/dL Positive CSF antigen CSF antigen titer, median (IQR) Positive culture
58 (51-64) 72.3% (81/112) 41.1% (46/112) 27.7% (31/112) 13.4% (15/112) 12.5% (14/112) 10.7% (12/112) 8.9% (10/112) 5.4% (6/112) 2.7% (3/112) 2.7% (3/112) 1.8% (2/112) 1.8% (2/112) 1.8% (2/112) 1.8% (2/112) 8.5% (8/94) 25.5% (24/94) 66% (62/94) 22 (18-30) 29.5% (33/112) 20.5% (23/112) 13.4% (15/112) 2.7% (3/112) 4.5% (5/112) 0.9% (1/112) 44.6% (49/110) 33.0% (37/112) 7.1% (8/112) 23% (26/112) 74% (83/112) 2.7% (3/112) 22.3% (25/112)
48.2% (54/112) 37.5% (42/112) 19.6% (22/112) 2.7% (3/112) 46.5% (47/101) 76.8% (86/112) 73.3% (66/90) 1:32 (1:4-1:512) 39 (17-140) 70.3% (26/37) 31 (2-107) 47 (25-74) 81 (41-161) 91.1% (41/45) 1:32 (0-1:512) 74.1% (40/54) Continued next page
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
Singh et al
© 2015 Wolters Kluwer
2135
TABLE 1. (Continued) Factor
Value
Primary antifungal therapy Fluconazole Liposomal amphotericin B Amphotericin B deoxycholate Amphotericin B lipid complex Voriconazole
No. receiving the primary agent 42.7% (41/96) 31.3% (30/96) 15.6% (15/96) 9.4% (9/96) 1.0% (1/96)
No. receiving concurrent 5-FC 17% (7/41) 53.3% (16/30) 73% (11/15) 55.5% (5/9) 0/1
a
Child-Pugh classification was available on 94 patients. MELD available on 75 patients. c Malignancies included lymphoma in 3 patients, adenocarcinoma of the colon in one and unspecified in 1 patient. d Values shown for 59 patients who underwent CSF analysis. e Opening pressure performed for 37 patients 5-FC, 5-flucytosine; MELD, model for end-stage liver disease. b
(Table 4). In 2 patients (patients 1 and 7, Table 4), peritoneal fluid and blood cultures, respectively, collected before transplantation yielded Cryptococcus after transplantation. Duration of antifungal therapy before transplantation in patients who were transplanted was median of 42.5 days (IQR, 8130 days, total range, −2 to 260 days; 2 patients were diagnosed after transplantation had already occurred and initiated treatment 1 and 2 days after transplantation, respectively. Detailed information of patients who underwent liver transplantation is shown in Table 5. Of the 31 candidates who did not undergo transplantation, the reasons were death before transplantation in (29%, 9/31), removal from the transplant list after development of cryptococcosis (23%, 7/31), considered too ill for transplantation (9.6%, 3/31), or changed to do not resuscitate code status (3.2%, 1/31). Two patients (6%, 2/31) remained on the active transplant list awaiting organ at the end of the study. Overall, 29% (9/31) of the patients had more than one of the aforementioned reasons for not undergoing transplantation. Clinical and microbiologic characteristics of the 31 patients who did not undergo transplantation are shown in Table 5. Transplanted patients were less likely to be fungemic (P = .063) or have positive serum cryptococcal antigen (P = .065); however, these differences were not statistically significant. Mortality at 90 days in transplant candidates was lower in those who had been transplanted (0/8 vs 61.3%, 9/31, P = 0.003).
TABLE 2.
Causes of death in the study patients Causes of death
No. patients (n = 64) a
Definitely/probably related to cryptococcosis Liver and/or multisystem organ failure Sepsis Renal failure Gastrointestinal bleed Pulmonary hemorrhage Invasive aspergillosis Respiratory failure Unknown a
35 5 2 1 1 1 1 1 17
Death was directly due to cryptococcosis in 6 patients. These include disseminated cryptococcosis and liver failure (n = 4), hypotension with adrenal infiltration (n = 1), and disseminated disease with multiorgan failure (n = 1).
Posttransplant outcomes in patients transplanted are shown in Table 5. All patients received tacrolimus-based immunosuppression and 3 of 8 received induction therapy that included basiliximab in 2 liver transplant recipients and thymoglobulin in 1 liver-kidney transplant recipients. Excluding 2 patients in whom fluconazole therapy was ongoing at last follow up, the median duration of suppressive antifungal therapy (fluconazole in all cases) was 272 days (IQR, 180-630 days) after transplantation. At follow-up of 108 to 3144 days (median, 1137 days), none of the 8 transplant recipients developed progression or recurrence of cryptococcal disease after transplantation. One patient developed immune reconstitution syndrome 6 weeks after diagnosis of cryptococcosis following which allograft rejection developed that was successfully treated. In all, 87.5% (7/8) of the patients were alive at last follow-up of 167 to 3285 days (median, 1895 days). Death in 1 patient who died 249 days after transplantation was related to liver and multiorgan failure; the patient had no evidence of cryptococcosis.
DISCUSSION Observations from this study have clinically significant implications for the management of patients with cirrhosis and cryptococcosis. An estimated 21% to 36% of the cases of cryptococcosis in nonhuman immunodeficiency virus– infected hosts occur in patients with liver cirrhosis.1,11 Our data show that 50% to 77% of these patients had CNS disease, fungemia, or disseminated cryptococcosis, and approximately 60% of the patients died shortly after the diagnosis. Previous studies have also documented unusually high mortality rate, ranging from 81% to 100% in patients with end-stage liver disease and cryptococcosis.1,4 Our patients who succumbed to their illness had more advanced liver disease and had overall greater acuity of illness based on the fact that they were more likely to require renal replacement therapy, mechanical ventilation, and present with shock. Fungemia and impaired mental status were also predictors of poor outcome, whereas headache as presenting symptom correlated with better outcome. The correlation of headache with outcome was likely due the fact that these patients had a greater CSF inflammatory response and initiated therapy earlier than those without headache. Previous studies have shown similar association, including worse outcomes with lack of headache and absence of CSF inflammatory response.12,13
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
2136
Transplantation
■
October 2015
■
Volume 99
■
Number 10
www.transplantjournal.com
TABLE 3.
Factors associated with mortality in patients with cirrhosis and cryptococcosis Factor
Alive (n = 48)
Age: median (IQR), y Female sex Underlying liver disease Hepatitis C virus Alcohol Hepatitis B virus Nonalcoholic steatohepatosis Cryptogenic cirrhosis Hepatocellular carcinoma Child-Pugh scorea (n = 94) A B C MELDa, median (IQR) Comorbid conditions Corticosteroid use Malignancy Diabetes mellitus Renal replacement therapy Mechanical ventilation Idiopathic CD4 lymphopenia Sarcoidosis Clinical and laboratory characteristics of cryptococcosis Time from onset of symptoms to diagnosis: median (IQR), d Headache Visual disturbances Mental status changes Alert Impaired Fever (≥38.3 °F) Hypotension/shock Site of cryptococcal disease Peritoneal Cutaneous Pulmonary Central nervous system Disseminated disease Fungemia CSFb opening pressure, cm Median (IQR) >20 cm Serum antigen titer, median (IQR) Serum titer ≥ 1:512 CSF antigen titer, median (IQR) CSF titer ≥ 1:512 Primary antifungal therapy Disseminated disease Fluconazole based Lipid polyene based AmBd based Concurrent 5-FC use Localized disease Fluconazole based Lipid polyene based AmBd based Concurrent 5-FC use
Died (n = 64)
OR (95% CI) odds of death
P
59 (51-67) 29.2% (14/48)
57 (49-63) 26.6% (17/64)
0.98 (0.94-1.01) 0.88 (0.38-2.02)
0.219 0.761
43.8% (21/48) 25% (12/48) 20.8% (10/48) 8.3% (4/48) 8.3% (4/48) 14.6% (7/48)
39.1% (25/64) 29.7% (19/64) 7.8% (5/64) 15.6% (10/64) 9.4% (6/6) 7.8% (5/64)
0.82 (0.39-1.76) 1.27 (0.54-2.95) 0.32 (0.10-1.02) 2.04 (0.60-6.94) 1.14 (0.30-4.28) 0.50 (0.15-1.67)
0.618 0.584 0.053 0.255 0.848 0.258
19.4% (7/36) 36.1% (13/36) 44.4% (16/36) 18 (13-23)
1.7% (1/58) 19.0% (11/58) 79.3% (46/58) 26 (21-33)
Reference group 5.92 (0.63-55.85) 20.12 (2.29-176.46) 1.10 (1.04-1.18)
— 0.120 0.007 0.002
16.7% (8/48) 20.8% (10/48) 25.0% (12/48) 4.2% (2/48) 2.1% (1/48) 0/48 6.3% (3/48)
23.4% (15/64) 10.9% (7/64) 32.8% (21/64) 20.3% (13/64) 26.6% (17/64) 1.6% (1/64) 0/64
1.53 (0.59-3.97) 0.47 (0.16-1.33) 1.46 (0.63-3.38) 5.86 (1.26-27.38) 17.00 (2.17-132.97) Unable to calculatec 0 (0-0.93)
0.382 0.155 0.371 0.024 0.007 0.571 0.076
14 (5-21) 45.8% (22/48) 12.5% (6/48)
8 (5-20) 23.4% (15/64) 3.1% (2/64)
0.98 (0.96-1.01) 0.36 (0.16-0.81) 0.22 (0.04-1.17)
0.139 0.014 0.077
62.5% (30/48) 37.5% (18/48) 43.8% (21/48) 8.3% (4/48)
40.6% (26/64) 59.4% (38/64) 45.2% (28/62) 32.8% (21/64)
0.41 (0.19-0.88) 2.44 (1.13-5.25) 1.06 (0.50-2.26) 5.37 (1.70-16.94)
0.023 0.023 0.883 0.004
14.6% (7/48) 2.1% (1/48) 39.6% (19/48) 54.2% (26/48) 66.7% (32/48) 24.4% (10/41) 78 (38-233) 81.3% (13/16) 1:16 (0-1:128) 22.0% (9/41) 1:16 (0-1:320) 25.0% (8/32)
23.4% (15/64) 3.1% (2/64) 35.9% (23/64) 43.8% (28/64) 84.4% (54/64) 61.7% (37/60) 32 (16-55) 61.9% (13/21) 1:2 (0-1:256) 24.5% (12/49) 1:128 (1-2:1024) 44.4% (12/27)
1.79 (0.67-4.82) 1.52 (0.13-17.22) 0.86 (0.39-1.85) 0.66 (0.31-1.40) 2.70 (1.09-6.66) 4.99 (2.06-12.05) 0.998 (0.99-1.002) 0.375 (0.08-1.74) 1.00 (0.99-1.001) 1.15 (0.43-3.09) 1.00 (0.99-1.00) 2.4 (0.80-7.23)
0.247 0.737 0.693 0.276 0.031
Cryptococcosis in Patients With Cirrhosis of the Liver and Posttransplant Outcomes Nina Singh, MD,1 Costi D. Sifri, MD,2 Fernanda P. Silveira, MD,1 Rachel Miller, MD,3 Kevin S. Gregg, MD,4 Shirish Huprikar, MD,5 Erika D. Lease, MD,6 Andrea Zimmer, MD,7 J. Stephen Dummer, MD,7 Cedric W. Spak, MD,8 Christine Koval, MD,9 David B. Banach, MD,10 Miloni Shroff, MD,11 Jade Le, MD,11 Darin Ostrander, MD,12 Robin Avery, MD,12 Albert Eid, MD,13 Raymund R. Razonable, MD,14 Jose Montero, MD,15 Emily Blumberg, MD,16 Ahlaam Alynbiawi, MD,17 Michele I. Morris, MD,17 Henry B. Randall, MD,18 George Alangaden, MD,19 Jeffrey Tessier,20 Marilyn M. Wagener,1 and Hsin Yun Sun, MD21 Background. The outcomes and optimal management of cirrhotic patients who develop cryptococcosis before transplantation are not fully known. Methods. We conducted a multicenter study involving consecutive patients with cirrhosis and cryptococcosis between January 2000 and March 2014. Data collected were generated as standard of care. Results. In all, 112 patients were followed until death or up to 9 years. Disseminated disease and fungemia were present in 76.8% (86/112) and 90-day mortality was 57.1% (64/112). Of the 39 patients listed for transplant, 20.5% (8) underwent liver transplantation, including 2 with active but unrecognized disease before transplantation. Median duration of pretransplant antifungal therapy and posttransplant therapy was 43 days (interquartile range, 8-130 days) and 272 days (interquartile range, 180-630 days), respectively. Transplantation was associated with lower mortality (P = 0.002). None of the transplant recipients developed disease progression during the median follow-up of 3.5 years with a survival rate of 87.5%. Conclusions. Cryptococcosis in patients with cirrhosis has grave prognosis. Our findings suggest that transplantation after recent cryptococcal disease may not be a categorical exclusion and may be cautiously undertaken in liver transplant candidates who are otherwise deemed clinically stable.
(Transplantation 2015;99: 2132–2141)
P
atients with end-stage liver disease may develop cryptococcosis as a result of cirrhosis-associated compromised host defenses that include impaired cell-mediated immunity, phagocytic dysfunction, decreased antibody and immunoglobulins, and complement deficiency.1-4 Indeed, cirrhosis is a major risk factor for cryptococcosis and for poor outcomes in those who develop disease.4,5 Cirrhosis-associated cryptococcosis often has high rates of fungemia and mortality.2,6 Increasingly, cryptococcosis has been documented in cirrhotic patients awaiting liver transplantation or in those referred for transplantation.4,7 Anecdotal reports with favorable
outcomes exist for transplants performed inadvertently in patients with unrecognized cryptococcosis or in those treated before transplantation.7,8 A dilemma, however, is whether patients with cryptococcosis during transplant candidacy can safely undergo transplantation. Limited data and guidance are available in this context, and little is known about their outcomes after transplantation.4,7 We conducted a multicenter study to determine the outcome in patients with cirrhosis and cryptococcosis and assess risk factors for death in those who died before transplantation. Additionally, in patients who underwent
Received 4 September 2014. Revision requested 31 October 2014. 15
University of South Florida, Tampa, FL.
16
Department of Medicine, Division of Infectious Diseases, University of Pittsburgh, Pittsburgh, PA.
University of Pennsylvania, Philadelphia, PA.
17
University of Miami, Miller School of Medicine, Miami, FL.
2
University of Virginia, Charlottesville, VA.
18
Department of Surgery, St. Luke's Hospital, Kansas City, MO.
3
University of Iowa, Iowa City, IA.
19
Henry Ford Hospital, Detroit, MI.
4
University of Michigan, Ann Arbor, MI.
20
Virginia Commonwealth University, Richmond, VA.
5
Icahn School of Medicine at Mount Sinai, New York, NY.
21
6
University of Washington, Seattle, WA.
Department of Internal Medicine, National Taiwan University Hospital and National, Taipei, Taiwan.
7
Vanderbilt University, Nashville, TN.
The authors declare no funding or conflicts of interest.
8
Baylor Health System, Dallas, TX.
9
Cleveland Clinic, Cleveland, OH.
N.S. and M.M.W. participated in research design and preparation of the paper. M. M.W. participated in data analysis. All authors participated in performance of the study, review, and editing of the manuscript.
Accepted 8 January 2015. 1
10
Yale School of Medicine, New Haven, CT.
11
University of Texas Southwestern Medical Center, Dallas, TX.
Correspondence: Nina Singh, MD, Infectious Diseases Section, Veterans Affairs Medical Center, University Drive C, Pittsburgh, PA 15240. ([email protected]).
12
Johns Hopkins University, Baltimore, MD.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
13
University of Kansas, Kansas City, KS.
ISSN: 0041-1337/15/9910-2132
14
Mayo Clinic, Rochester, MN.
DOI: 10.1097/TP.0000000000000690
2132
www.transplantjournal.com
Transplantation
■
October 2015
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
■
Volume 99
■
Number 10
Singh et al
© 2015 Wolters Kluwer
transplantation, we assessed the disease status before transplantation, timing of transplantation after diagnosis of cryptocococcosis, posttransplant management, including immunosuppressive regimen, antifungal agent and duration of prophylaxis, disease progression or relapse, and allograft rejection. PATIENTS AND METHODS Participating transplant centers were identified through the American Society of Transplantation Infectious Diseases Community of Practice. Patients were included if they had end-stage liver disease and cryptococcosis regardless of whether they were considered for liver transplant candidacy, referred for liver transplantation, or awaiting liver transplantation. Cryptococcosis was defined based on the European Organization for Research and Treatment in Cancer and the Mycoses Study Group criteria.9 Transplant recipients in whom pretransplant cryptococcosis (positive culture from blood or another clinical specimen, histopathology, or positive antigen) was recognized after transplantation were also included. Patients were excluded if they had developed cryptococcosis after transplantation with no evidence of pretransplant disease or were considered to have donor-derived cryptococcosis. The study involved collection of data generated as standard of care at the study sites between January 2000 and March 2014. Variables collected included demographics, severity of underlying liver disease, comorbid conditions, details of cryptococcal disease, sites of involvement, antifungal therapy, follow-up microbiologic data, and patient outcomes, including after transplantation in those who underwent transplantation. Fungemia, involvement of 2 or more noncontiguous organ sites or the central nervous system (CNS) were regarded as disseminated cryptococcal disease. Immune reconstitution syndrome was defined as previously reported.10 The cause of death was determined by the site investigator. Institutional Review Board approval for this observational study was obtained per local requirements. Statistical Analysis
Stata/SE (version 13.1, Stata Corp, College Station, TX) was used for statistical analyses. Odds ratios and 95% confidence intervals for mortality were estimated using univariate logistic models. A multivariate model was constructed using variables found to be associated with outcome in univariate analysis. The final model was tested using a Pearson goodness of fit test. Similarly, for patients on the transplant list, factors were examined for differences between patients that were transplanted and those that were not. RESULTS In all, 112 cirrhotic patients with cryptococcosis were identified at the participating sites. One of the patients has previously been reported as a case report (7). Patients were followed until death or for up to 9 years; total follow-up in the study population was 80 patient-years. Clinical and demographic characteristics of these patients are shown in Table 1. The median model for end-stage liver disease score of the patients was 22 (interquartile range [IQR] 18-30) and 66% (62/94) were Child-Pugh class C. Comorbid conditions included diabetes mellitus in 29.5% (33/112), malignancy in 15.2% (17/112), and renal replacement therapy
2133
in 13.4% (15/112). Additionally, 20.5% (23/112) were receiving corticosteroids, 2.6% (3/112) had sarcoidosis, and 1 patient had idiopathic CD4+ lymphopenia. None had human immunodeficiency virus infection. Characteristics of Cryptococcal Disease
The diagnosis of cryptococcosis was based on positive culture in 85.7% (96/112) of the patients, including fungemia and positive cerebrospinal fluid (CSF) culture in 47 and 40 cases, respectively. Of patients without culture-based diagnosis, 81.3% (13/16) had positive cryptococcal antigen (CSF and/or serum antigen) and in 3 patients, the diagnosis was based on tissue histopathology. Sites of cryptococcosis included CNS in 48.2% (54/112), pulmonary in 37.5% (42/ 112), peritoneal cavity in 19.6% (22/112), and skin in 2.7% (3/112). Disseminated disease and fungemia were present in 76.8% (86/112) and 46.5% (47/101) of the patients, respectively. Of patients with CNS disease, 83.3% (40/48) had positive culture and 91.1% (41/45) had positive CSF cryptococcal antigen (median titer, 1:128; IQR, 1:16 to 1:1024). Median time from onset of symptoms to diagnosis was 9 days (IQR, 5-21 days). In 15.2% (17/112) of the patients, the diagnosis was established post mortem. The antifungal regimen used in the study patients is shown in Table 1. The most frequently used primary antifungal agent was fluconazole (42.7%, 41/96) followed by liposomal amphotericin B (31.3%, 30/96). Concurrent 5-flucytosine was used in 40.6% of the patients (39/96) (Table 1). Outcome and Risk Factors for Mortality
Overall, mortality at 90 days was 57.1% (64/112). Deaths (n = 64) occurred a median of 11 days (IQR, 3-20) after diagnosis and median of 10.5 days (IQR, 6-22 days) after initiation of antifungal therapy in those treated (n = 46). Death was attributable to cryptococcal disease in 54.6% (35/64) of the patients; other causes of death are listed in Table 2. The median time to death after diagnosis was significantly earlier in patients in whom death was attributable to cryptococcosis compared to other causes (day 4 vs day 21, P = 0.0001). Variables associated with mortality are shown in Table 3. Mortality was significantly higher in patients with more advanced liver disease (Child-Pugh [P = 0.007] and model for end-stage liver disease score [P = 0.002]), renal replacement therapy (P = 0.0240), mechanical ventilation (P = 0.007), hypotension/shock (P = 0.004), impaired mental status (P = 0.023), and fungemia (P < 0.001). On the contrary, headache at presentation (P = 0.014) and transplantation (P = 0.002) were associated with improved outcome (Table 3). Patients with headache had significantly higher CSF white blood cell (median, 70 [IQR, 22-120] vs 6.5 [IQR, 2-98], P = 0.011) and shorter time to initiation of therapy (median, 1 [IQR, 0-2] vs 3 [IQR, 1-6] days, P = 0.012) than those without headache. Transplantation and Posttransplant Outcomes
Of 112 patients, 46 were liver transplant candidates, that is, listed for transplantation at any time during the study. Excluding 7 candidates who died before the diagnosis of cryptococcosis (the clinical cultures yielded Cryptococcus after death), 20.5 % (8/39) of the listed patients underwent liver transplantation that included liver alone in 7 patients and liver-kidney in 1 patient. Of the 8 transplant recipients, 4 had disseminated disease (including 3 with meningitis), 2 had pulmonary only, and 2 had extrapulmonary cryptococcosis
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
2134
Transplantation
■
October 2015
■
Volume 99
■
Number 10
www.transplantjournal.com
TABLE 1.
Demographic and clinical characteristics of the study patients (n = 112) Factor
Value
Age: median (IQR), y Sex: male, % Underlying liver disease (may have more than one) Hepatitis C virus Alcohol Hepatitis B virus Nonalcoholic steatohepatosis Hepatocellular carcinoma Cryptogenic Primary biliary cirrhosis Fulminant hepatic failure Autoimmune hepatitis Primary biliary cirrhosis Hemochromatosis α 1-antitrypsin deficiency Wilson disease Child-Pugh classificationa A B C MELD score, median (IQR)b Comorbid conditions Diabetes mellitus Corticosteroid use Renal replacement therapy Sarcoidosis Malignancy (other than hepatocellular carcinoma)c Idiopathic CD4+ lymphopenia Clinical manifestations Fever Headache Decrease in visual acuity Mental status Alert Nonalert Unknown Hypotension/shock Characteristics of cryptococcal disease Sites of involvement Central nervous system Pulmonary Peritoneal Cutaneous Fungemia Disseminated disease Serum antigen positivity Serum cryptococcal antigen titer, median (IQR) CSFd characteristics Opening pressuree: median (IQR), cm > 20 cm White blood cells, median (IQR) per mm3 Glucose: median (IQR), mg/dL Protein: median (IQR), g/dL Positive CSF antigen CSF antigen titer, median (IQR) Positive culture
58 (51-64) 72.3% (81/112) 41.1% (46/112) 27.7% (31/112) 13.4% (15/112) 12.5% (14/112) 10.7% (12/112) 8.9% (10/112) 5.4% (6/112) 2.7% (3/112) 2.7% (3/112) 1.8% (2/112) 1.8% (2/112) 1.8% (2/112) 1.8% (2/112) 8.5% (8/94) 25.5% (24/94) 66% (62/94) 22 (18-30) 29.5% (33/112) 20.5% (23/112) 13.4% (15/112) 2.7% (3/112) 4.5% (5/112) 0.9% (1/112) 44.6% (49/110) 33.0% (37/112) 7.1% (8/112) 23% (26/112) 74% (83/112) 2.7% (3/112) 22.3% (25/112)
48.2% (54/112) 37.5% (42/112) 19.6% (22/112) 2.7% (3/112) 46.5% (47/101) 76.8% (86/112) 73.3% (66/90) 1:32 (1:4-1:512) 39 (17-140) 70.3% (26/37) 31 (2-107) 47 (25-74) 81 (41-161) 91.1% (41/45) 1:32 (0-1:512) 74.1% (40/54) Continued next page
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
Singh et al
© 2015 Wolters Kluwer
2135
TABLE 1. (Continued) Factor
Value
Primary antifungal therapy Fluconazole Liposomal amphotericin B Amphotericin B deoxycholate Amphotericin B lipid complex Voriconazole
No. receiving the primary agent 42.7% (41/96) 31.3% (30/96) 15.6% (15/96) 9.4% (9/96) 1.0% (1/96)
No. receiving concurrent 5-FC 17% (7/41) 53.3% (16/30) 73% (11/15) 55.5% (5/9) 0/1
a
Child-Pugh classification was available on 94 patients. MELD available on 75 patients. c Malignancies included lymphoma in 3 patients, adenocarcinoma of the colon in one and unspecified in 1 patient. d Values shown for 59 patients who underwent CSF analysis. e Opening pressure performed for 37 patients 5-FC, 5-flucytosine; MELD, model for end-stage liver disease. b
(Table 4). In 2 patients (patients 1 and 7, Table 4), peritoneal fluid and blood cultures, respectively, collected before transplantation yielded Cryptococcus after transplantation. Duration of antifungal therapy before transplantation in patients who were transplanted was median of 42.5 days (IQR, 8130 days, total range, −2 to 260 days; 2 patients were diagnosed after transplantation had already occurred and initiated treatment 1 and 2 days after transplantation, respectively. Detailed information of patients who underwent liver transplantation is shown in Table 5. Of the 31 candidates who did not undergo transplantation, the reasons were death before transplantation in (29%, 9/31), removal from the transplant list after development of cryptococcosis (23%, 7/31), considered too ill for transplantation (9.6%, 3/31), or changed to do not resuscitate code status (3.2%, 1/31). Two patients (6%, 2/31) remained on the active transplant list awaiting organ at the end of the study. Overall, 29% (9/31) of the patients had more than one of the aforementioned reasons for not undergoing transplantation. Clinical and microbiologic characteristics of the 31 patients who did not undergo transplantation are shown in Table 5. Transplanted patients were less likely to be fungemic (P = .063) or have positive serum cryptococcal antigen (P = .065); however, these differences were not statistically significant. Mortality at 90 days in transplant candidates was lower in those who had been transplanted (0/8 vs 61.3%, 9/31, P = 0.003).
TABLE 2.
Causes of death in the study patients Causes of death
No. patients (n = 64) a
Definitely/probably related to cryptococcosis Liver and/or multisystem organ failure Sepsis Renal failure Gastrointestinal bleed Pulmonary hemorrhage Invasive aspergillosis Respiratory failure Unknown a
35 5 2 1 1 1 1 1 17
Death was directly due to cryptococcosis in 6 patients. These include disseminated cryptococcosis and liver failure (n = 4), hypotension with adrenal infiltration (n = 1), and disseminated disease with multiorgan failure (n = 1).
Posttransplant outcomes in patients transplanted are shown in Table 5. All patients received tacrolimus-based immunosuppression and 3 of 8 received induction therapy that included basiliximab in 2 liver transplant recipients and thymoglobulin in 1 liver-kidney transplant recipients. Excluding 2 patients in whom fluconazole therapy was ongoing at last follow up, the median duration of suppressive antifungal therapy (fluconazole in all cases) was 272 days (IQR, 180-630 days) after transplantation. At follow-up of 108 to 3144 days (median, 1137 days), none of the 8 transplant recipients developed progression or recurrence of cryptococcal disease after transplantation. One patient developed immune reconstitution syndrome 6 weeks after diagnosis of cryptococcosis following which allograft rejection developed that was successfully treated. In all, 87.5% (7/8) of the patients were alive at last follow-up of 167 to 3285 days (median, 1895 days). Death in 1 patient who died 249 days after transplantation was related to liver and multiorgan failure; the patient had no evidence of cryptococcosis.
DISCUSSION Observations from this study have clinically significant implications for the management of patients with cirrhosis and cryptococcosis. An estimated 21% to 36% of the cases of cryptococcosis in nonhuman immunodeficiency virus– infected hosts occur in patients with liver cirrhosis.1,11 Our data show that 50% to 77% of these patients had CNS disease, fungemia, or disseminated cryptococcosis, and approximately 60% of the patients died shortly after the diagnosis. Previous studies have also documented unusually high mortality rate, ranging from 81% to 100% in patients with end-stage liver disease and cryptococcosis.1,4 Our patients who succumbed to their illness had more advanced liver disease and had overall greater acuity of illness based on the fact that they were more likely to require renal replacement therapy, mechanical ventilation, and present with shock. Fungemia and impaired mental status were also predictors of poor outcome, whereas headache as presenting symptom correlated with better outcome. The correlation of headache with outcome was likely due the fact that these patients had a greater CSF inflammatory response and initiated therapy earlier than those without headache. Previous studies have shown similar association, including worse outcomes with lack of headache and absence of CSF inflammatory response.12,13
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
2136
Transplantation
■
October 2015
■
Volume 99
■
Number 10
www.transplantjournal.com
TABLE 3.
Factors associated with mortality in patients with cirrhosis and cryptococcosis Factor
Alive (n = 48)
Age: median (IQR), y Female sex Underlying liver disease Hepatitis C virus Alcohol Hepatitis B virus Nonalcoholic steatohepatosis Cryptogenic cirrhosis Hepatocellular carcinoma Child-Pugh scorea (n = 94) A B C MELDa, median (IQR) Comorbid conditions Corticosteroid use Malignancy Diabetes mellitus Renal replacement therapy Mechanical ventilation Idiopathic CD4 lymphopenia Sarcoidosis Clinical and laboratory characteristics of cryptococcosis Time from onset of symptoms to diagnosis: median (IQR), d Headache Visual disturbances Mental status changes Alert Impaired Fever (≥38.3 °F) Hypotension/shock Site of cryptococcal disease Peritoneal Cutaneous Pulmonary Central nervous system Disseminated disease Fungemia CSFb opening pressure, cm Median (IQR) >20 cm Serum antigen titer, median (IQR) Serum titer ≥ 1:512 CSF antigen titer, median (IQR) CSF titer ≥ 1:512 Primary antifungal therapy Disseminated disease Fluconazole based Lipid polyene based AmBd based Concurrent 5-FC use Localized disease Fluconazole based Lipid polyene based AmBd based Concurrent 5-FC use
Died (n = 64)
OR (95% CI) odds of death
P
59 (51-67) 29.2% (14/48)
57 (49-63) 26.6% (17/64)
0.98 (0.94-1.01) 0.88 (0.38-2.02)
0.219 0.761
43.8% (21/48) 25% (12/48) 20.8% (10/48) 8.3% (4/48) 8.3% (4/48) 14.6% (7/48)
39.1% (25/64) 29.7% (19/64) 7.8% (5/64) 15.6% (10/64) 9.4% (6/6) 7.8% (5/64)
0.82 (0.39-1.76) 1.27 (0.54-2.95) 0.32 (0.10-1.02) 2.04 (0.60-6.94) 1.14 (0.30-4.28) 0.50 (0.15-1.67)
0.618 0.584 0.053 0.255 0.848 0.258
19.4% (7/36) 36.1% (13/36) 44.4% (16/36) 18 (13-23)
1.7% (1/58) 19.0% (11/58) 79.3% (46/58) 26 (21-33)
Reference group 5.92 (0.63-55.85) 20.12 (2.29-176.46) 1.10 (1.04-1.18)
— 0.120 0.007 0.002
16.7% (8/48) 20.8% (10/48) 25.0% (12/48) 4.2% (2/48) 2.1% (1/48) 0/48 6.3% (3/48)
23.4% (15/64) 10.9% (7/64) 32.8% (21/64) 20.3% (13/64) 26.6% (17/64) 1.6% (1/64) 0/64
1.53 (0.59-3.97) 0.47 (0.16-1.33) 1.46 (0.63-3.38) 5.86 (1.26-27.38) 17.00 (2.17-132.97) Unable to calculatec 0 (0-0.93)
0.382 0.155 0.371 0.024 0.007 0.571 0.076
14 (5-21) 45.8% (22/48) 12.5% (6/48)
8 (5-20) 23.4% (15/64) 3.1% (2/64)
0.98 (0.96-1.01) 0.36 (0.16-0.81) 0.22 (0.04-1.17)
0.139 0.014 0.077
62.5% (30/48) 37.5% (18/48) 43.8% (21/48) 8.3% (4/48)
40.6% (26/64) 59.4% (38/64) 45.2% (28/62) 32.8% (21/64)
0.41 (0.19-0.88) 2.44 (1.13-5.25) 1.06 (0.50-2.26) 5.37 (1.70-16.94)
0.023 0.023 0.883 0.004
14.6% (7/48) 2.1% (1/48) 39.6% (19/48) 54.2% (26/48) 66.7% (32/48) 24.4% (10/41) 78 (38-233) 81.3% (13/16) 1:16 (0-1:128) 22.0% (9/41) 1:16 (0-1:320) 25.0% (8/32)
23.4% (15/64) 3.1% (2/64) 35.9% (23/64) 43.8% (28/64) 84.4% (54/64) 61.7% (37/60) 32 (16-55) 61.9% (13/21) 1:2 (0-1:256) 24.5% (12/49) 1:128 (1-2:1024) 44.4% (12/27)
1.79 (0.67-4.82) 1.52 (0.13-17.22) 0.86 (0.39-1.85) 0.66 (0.31-1.40) 2.70 (1.09-6.66) 4.99 (2.06-12.05) 0.998 (0.99-1.002) 0.375 (0.08-1.74) 1.00 (0.99-1.001) 1.15 (0.43-3.09) 1.00 (0.99-1.00) 2.4 (0.80-7.23)
0.247 0.737 0.693 0.276 0.031
