Predictors of Immune Reconstitution Syndrome in Organ Transplant Recipients With Cryptococcosis: Implications for the Management of Immunosuppression

(See the Editorial Commentary by Fishman on pages 45–7.)

Cryptococcus neoformans is a significant opportunistic pathogen in solid organ transplant (SOT) recipients. An estimated 2.8% and up to 5% of transplant recipients develop cryptococcosis after transplant [1–5]. Central nervous system (CNS) involvement has been documented in 54%–72% of patients with cryptococcal disease. Mortality rates in organ transplant recipients with cryptococcosis range from 15% to 20% [1, 2].

CD4⁺ T-helper (Th) responses are critical mediators of immunity against Cryptococcus. Whereas cells of Th1 and Th17 [6–11] lineage confer protection, T-regulatory cells and Th2 responses compromise host resistance and facilitate disease progression [9, 11, 12]. Cryptococcus per se has immunosuppressive capabilities and preferentially inhibits proinflammatory responses while inducing Th2 [13]. Reversal of pathogen-mediated immunosuppression with effective antifungal therapy restores inflammatory responses [14].

Iatrogenic immunosuppression in transplant recipients is associated with anti-inflammatory responses that are vital for graft tolerance, whereas reduction of immunosuppression has the potential to promote an inflammatory milieu [10]. In transplant patients with serious opportunistic mycoses such as cryptococcosis, the discontinuation or reduction of immunosuppressive therapy upon diagnosis is a common practice [15]. The cumulative effect of reversal of pathogen-induced immunosuppression, withdrawal or reduction of immunosuppression, and administration of effective antifungal therapy is associated with a shift in the host immune repertoire toward physiologic (but sometimes pathologic) inflammatory responses that can lead to immune reconstitution syndrome (IRS) [16]. Thus, restoration of host immunity, although pivotal for the eradication of this yeast, has the potential to contribute to unwanted inflammation. A growing body of literature shows that any clinical condition associated with a rapid reversal of immunosuppression is permissive to the development of IRS [16–18].

This study was undertaken to better characterize the clinical phenomenon of IRS in transplant recipients with cryptococcosis, with the goal to systematically determine variables associated with the occurrence of IRS, and to assess the clinical outcomes in transplant recipients with Cryptococcus-associated IRS.

PATIENTS AND METHODS

The study population consisted of SOT recipients with cryptococcal disease at the participating centers. Cryptococcosis was defined based on criteria proposed by the European Organization for Research and Treatment in Cancer/Mycoses Study Group [19]. This was an observational study with no specific study interventions and procedures. The management of each case was according to standard of care at the transplant centers. Data collected for this study included demographic characteristics, type of organ transplant, prior transplant, cytomegalovirus infection and disease, rejection 6 months prior to cryptococcosis, renal failure (serum creatinine ≥2 mg/dL at the time of diagnosis), immunosuppressive regimen at the time of diagnosis, characteristics of cryptococcal disease, antifungal therapy, management of immunosuppressive therapy after the diagnosis of cryptococcal disease, and outcomes (graft and patient survival) 12 months after the diagnosis.

Patients were considered to have CNS cryptococcosis if they had a positive cerebrospinal fluid (CSF) culture or positive CSF cryptococcal antigen [1, 20]. Involvement of 2 or more noncontiguous organ sites or of the CNS was regarded as disseminated cryptococcal disease. IRS was defined based on case definitions proposed by Singh and Perfect [16] and previously used in the transplant setting [21] (Table 1). Institutional review board approval for this observational study was obtained per local requirements.

Statistical Analysis

Stata/SE software, version 13.1 (StataCorp, College Station, Texas), was used for all statistical comparisons. Logistic regression models were used to evaluate risk factors for IRS. The dependent variable was the binary: IRS/not IRS. The odds ratios (ORs) with 95% confidence intervals (CIs) were determined for each risk factor. A multivariate model was constructed from factors found to be associated with IRS at the 0.10 level in the univariate models. A χ² test for trend was used to calculate the odds of IRS associated with the number of risk factor present, by comparing the odds of developing IRS with 1 or more risk factors compared with no risk factors. A similar analysis, limited to patients with CNS disease, was employed. The dependent variable was IRS, and ORs with 95% CIs were estimated for each risk factor. The presence of 1 or more statistically significant risk factors were compared to the odds associated with no risk factors in the χ² for trend test.

RESULTS

Eighty-nine SOT recipients with cryptococcosis were studied, including 35 (39%) kidney, 24 (27%) liver, 12 (13%) heart, 9 (10%) kidney-pancreas, 7 (8%) lung, and 2 (2%) pancreas transplant recipients. None of the patients had human immunodeficiency virus infection. The immunosuppressive regimen was calcineurin inhibitor–based in 83 (93%) of patients, including tacrolimus in 78 and cyclosporine in 5 patients. At the time of diagnosis, 82% of cases were receiving prednisone (median dose, 5 mg) and 34% (30/89) had baseline renal dysfunction

Cryptococcosis developed a median of 17 months (interquartile range [IQR], 6.3–45.4 months) after transplant; 41% of cases occurred within 1 year of transplant. Overall, 65% (58/89) of patients had pulmonary disease with cryptococcosis limited to the lungs in 39.3% (35/89). CNS disease was observed in 46.1% (41/89) and cutaneous disease in 12.4% (11/89) of patients. Disseminated cryptococcosis was observed in 57.3% (51/89), and 36.8% (25/68) grew Cryptococcus from blood cultures. Serum cryptococcal antigen was positive in 82.5% (66/80), with titers ≥1:64 in 62% of patients.

Characteristics of Patients With IRS

Thirteen (14.6%) of 89 transplant recipients with cryptococcosis developed IRS; these included 3 liver (including 1 liver-kidney), 4 kidney, 3 pancreas, and 2 heart transplant recipients. Detailed information of the patients with IRS is outlined in Table 2. IRS developed a median of 45 days (IQR, 15–76 days) after initiation of antifungal therapy

Risk Factors for IRS

Demographics and clinical characteristics of cases with and without IRS are shown in Table 2. Age, sex, type of transplant, renal failure at baseline, prior rejection, time to onset of cryptococcosis, fungemia, and serum cryptococcal antigen titer did not correlate with IRS. None of the following modifications in immunosuppression were associated with the development of IRS: any dose reduction (without discontinuation) of calcineurin inhibitor, >50% reduction of calcineurin inhibitor, discontinuation of prednisone, any reduction of prednisone, >50% reduction of prednisone, and discontinuation of azathioprine/mycophenolate mofetil (Table 3).

However, compared with SOT recipients without IRS, cases with IRS were significantly more likely to have CNS disease (84.6% [11/13] vs 40.0% [30/76]; P = .008), disseminated cryptococcosis at baseline (92.3% [12/13] vs 51.3% [39/76]; P = .022), and discontinuation of the calcineurin inhibitor (50% [6/12] vs 13.8 [9/65]; P = .008). Of the patients with cryptococcal disease limited to the lungs, only 2.9% developed IRS compared with 22.2% (P = .012) of those with extrapulmonary involvement (Table 2). Logistic regression analyses were performed to determine the variables that predicted risk of IRS. Disseminated and extrapulmonary disease correlated highly with CNS involvement (28/41 patients with CNS disease by definition had disseminated cryptococcosis, and all had extrapulmonary disease). Multivariate analyses therefore included CNS disease and discontinuation of calcineurin inhibitor in the model. CNS disease (adjusted OR, 6.23 [95% CI, 1.22–31.75]; P = .03) and discontinuation of calcineurin inhibitor (adjusted OR, 5.11 [95% CI, 1.33–19.69]; P = .02) were independently associated with IRS. IRS occurred in 2.6% (1/13) of patients with none of the aforementioned factors, 18.8% (6/32) with 1 factor, and 50% (6/12) of the patients with both factors present (χ² for trend, P = .0001).

IRS in Patients With CNS Disease

Given that IRS developed in 26.8% (11/41) of the patients with CNS cryptococcosis, and 84.6% (11/13) of the patients with IRS had CNS disease, we assessed potential risk factors for IRS specifically in patients with CNS cryptococcosis. Compared with transplant recipients without CNS disease–associated IRS, CNS cases with IRS were more likely to have abnormal CNS imaging at baseline (81.8% [9/11] vs 40.0% [10/25]; P = .03) and to have calcineurin-inhibitor agent discontinued at the onset of disease (54.6% [6/11] vs 21.4% [6/28]; P = .051) (Table 4). None (0/11) of the cases without aforementioned factors, 35% (7/20) with 1 factor, and all patients (4/4) with both abnormal imaging and discontinuation of calcineurin-inhibitor agent (χ² for trend, P = .0004) developed IRS.

Outcomes

Graft rejection after the diagnosis of cryptococcosis was documented in 15.4% (2/13) of patients with IRS vs 2.6% (2/76) in those without IRS (P = .07). Calcineurin inhibitor was discontinued within 1 week of diagnosis for 1 patient with IRS and rejection, and the other 3 patients with rejection episodes had 50%–75% reduction in CNI within 1 week of diagnosis; this was not significantly different from those not experiencing rejection (P = .289). The mortality rate was 13.5% (12/89) for the entire cohort; 15.3% (2/13) in patients with IRS; and 13.2% (10/76) in patients without IRS (P = .828). None of the deaths were attributable to IRS (Table 5).

DISCUSSION

Observations in this study have clinically relevant implications for the management of SOT recipients with cryptococcosis. IRS developed in 14% of the transplant recipients with cryptococcosis. IRS has been observed as case reports and in 5%–11% of transplant recipients in previous case series, typically between 4 and 6 weeks after antifungal therapy [17, 18, 21]. Prior studies, however, were not specifically designed to assess IRS. To our knowledge, the current study is the first in-depth assessment of the characteristics, risk factors (including the impact of changes in immunosuppression), and outcomes using previously proposed criteria for IRS in SOT recipients.

Meningitis is the most common manifestation of CNS cryptococcosis. It is commonly associated with infection of the subarachnoid space and involvement of the underlying parenchyma to a variable extent [22]. Overt neuroimaging abnormalities such as CNS mass lesions, leptomeningitis, or hydrocephalus have been documented in 11%–25% of cases with cryptococcal CNS disease [2, 22]. Such lesions occurred in 29% of the SOT recipients with cryptococcosis in a previous study [22]. Whereas most neuroimaging abnormalities were present at baseline, approximately 20% of the lesions developed after the initiation of antifungal therapy and these were attributed to IRS [22]. Outcome varied depending on the type of lesions. Mortality was 50% in patients with CNS mass lesions; those with leptomeningitis (despite higher CSF cryptococcal antigen titers) had lower mortality (12.5%) [22].

In our study, patients with CNS disease or disseminated cryptococcosis were significantly more likely to develop IRS. Conversely, those with disease limited to the lungs were less likely to have IRS. Disseminated disease is indicative of greater severity of disease or organism burden and conferred a higher risk of IRS in other hosts and SOT recipients with opportunistic infections, including cryptococcosis [21, 23]. Among patients with CNS disease in our study, those with neuroimaging abnormalities, particularly mass lesions, were more likely to develop IRS, irrespective of serum or CSF cryptococcal antigen titers and fungemia. These data suggest that host immune response is a significant contributor to pathogenesis and disease expression. Along these lines, a previous study has also suggested that host immune function is a key determinant of morbidity and outcomes in patients with CNS cryptococcosis [24].

A major challenge confronting the transplant clinician is the management of immunosuppression in transplant recipients diagnosed with cryptococcosis. Reduction or even the discontinuation of immunosuppressive drug therapies, whenever possible, appears to be intuitively logical in transplant recipients with serious infections. Expert opinion and guidelines of the Infectious Diseases Society of America and the American Society of Transplantation recommend reduction in immunosuppression, with consideration of initially lowering the corticosteroid dose [25, 26]. However, no formal studies or evidence-based data correlating the risk of IRS with modification in immunosuppression exist. Our study shows that discontinuation of calcineurin inhibitors was the only modification of immunosuppression that influenced the development of IRS. Patients in whom the calcineurin inhibitor was discontinued had a 5-fold higher risk of IRS. Neither discontinuation nor reduction of other agents affected this risk. The biologic basis of these observations may be the selective effect of specific immunosuppressive agents on immune function with greater suppression of Th1 and Th17 with calcineurin inhibitors compared with other components of the immunosuppressive regimen [27]. Given that calcineurin inhibitors have synergistic interactions with antifungal agents, continuation at reduced dosages as opposed to stopping these agents appears rational [28].

Although not statistically significant, patients with IRS had a 7-fold higher risk of allograft rejection than those without IRS. Similar observations have also been reported in a previous study [29]. It is plausible that the immune milieu in patients with IRS is also conducive to inflammatory pathology that underlies allograft injury. Reduction or discontinuation of calcineurin inhibitors could also have been a contributory factor. The precise basis of these observations needs to be determined, however.

Several limitations of this study deserve to be acknowledged. Reliable diagnostic criteria or laboratory assays that can unequivocally diagnose IRS do not exist, and there are no markers for IRS specifically in SOT recipients. In our patients, this entity occurred in the setting of microbiologic response and there was no documented evidence of disease progression or alternative etiologies to account for the presentation. Nevertheless, it is plausible that some cases may have been misdiagnosed or had disease progression. Indeed, experts believe that IRS and microbiologic disease progression can appear simultaneously [25] and that these phenomena may not be mutually exclusive. It is also possible that patients with baseline imaging abnormalities may have been more likely to have follow-up studies performed. In addition, the study was based on data generated by routine clinical care, and protocol imaging studies were not performed. Finally, formal assessments of the net state of immunosuppression with immunologic assays that measure the degree of immune impairment were not done. Strengths of our study include systematic collection of data using a uniform data collection tool and the fact that this was a multicenter study, which renders our findings generalizable.

IRS is a clinical entity that is gaining recognition as a significant complication in transplant recipients with cryptococcosis. We have determined factors that can be used to identify patients at risk for this entity and have discussed the potential biologic plausibility for their association with IRS. Most importantly, we have shown that the manner by which the immunosuppression is modified may independently influence the risk of IRS. Future studies are warranted to further delineate this entity and define its biologic basis.

Note

Potential conflicts of interest. The following authors have received research funding through their institutions: B. D. A. has received support from Pfizer, Astellas, Charles River Laboratories, Synexis, and Viamet; G. M. L. has received support from Astellas, Qiagen, and T2 Biotechnologies; N. S. has received investigator-initiated grant support from Pfizer. All other authors report no potential conflicts.

All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.

References

Author notes