In this Journal feature, information about a real patient is presented in stages (boldface type) to an expert clinician, who responds to the information, sharing his or her reasoning with the reader (regular type). The authors' commentary follows.
A 67-year-old man presented with a 3-month history of fatigue and fever. He had undergone heart transplantation 6 years earlier for idiopathic cardiomyopathy. He reported no weight loss, night sweats, or chills. He also reported no headache, rash, joint swelling, dysuria, or abdominal or respiratory problems.
This patient has a prolonged fever without focal symptoms years after a heart transplant. The risk for particular opportunistic infections in the recipient of a solid-organ transplant depends on the intensity of immunosuppression and the exposure history. Although specific protocols differ among institutions, maintenance immunosuppression is generally reduced during the first year after transplantation. However, lymphocyte-depleting antirejection therapy (e.g., thymoglobulin) might cause a reduction in cell-mediated immunity for several years after use. Given the time course, cytomegalovirus (CMV) is a possibility, as is reactivation of latent disease such as histoplasmosis or infection with Mycobacterium tuberculosis. Although the patient's immunocompromised state increases the likelihood that his fever is infectious in origin, neoplastic and rheumatologic causes should also be considered.
During the previous several weeks, the patient had been seen by his regular physicians and was initially prescribed oral amoxicillin followed by levofloxacin for nasal stuffiness due to presumed sinusitis, with no change in his fever. His temperature elevation occurred daily, generally in the mid-afternoon, and ranged from 38.6 to 38.9°C. Evaluation during this time included a complete blood count, liver-function tests, and routine serum chemical analyses. Results were normal. Serum polymerase-chain-reaction (PCR) assays for CMV and Epstein–Barr virus (EBV) were negative. A serum cryptococcal-antigen test (Cryptococcal Antigen Latex Agglutination System, Meridian Bioscience) was negative. Chest radiography revealed no infiltrate. Computed tomography (CT) of the chest was normal. A CT scan of his abdomen revealed gallstones with mild gallbladder thickening, findings that were confirmed by ultrasonography. He underwent open cholecystectomy 2 weeks before his current presentation; pathological examination revealed gallstones, but the gallbladder was not inflamed. His fever persisted and profound apathy developed while the patient was still hospitalized after his operation. His treating physicians diagnosed depression, but antidepressants were not prescribed. His wife stated that his short-term memory appeared to be worsening; this was also attributed to depression. His sensorium was clear, and there were no seizures or loss of consciousness. He was transferred to an academic medical center for further evaluation.
Even in immunosuppressed patients, sinusitis is usually clinically obvious and would generally respond to antibiotics. Most patients with CMV-related disease will have a positive serum PCR assay, but end-organ disease (particularly in the gastrointestinal tract) can occur without viremia. The significance of a negative plasma PCR assay for EBV is less clear, but in combination with the absence of lymphadenopathy on CT scanning, it lessens the likelihood of a post-transplantation lymphoproliferative disorder. A diagnosis of cryptococcal meningitis or symptomatic extraneural disease is very unlikely after a negative test for serum cryptococcal antigen. The absence of abdominal pain and the normal liver-function tests made cholecystitis unlikely, and the results of the pathological examination ruled it out as a cause of fever.
His central nervous system symptoms, however, require further exploration. A number of viral agents — such as human herpesvirus 6 (HHV-6), lymphocytic choriomeningitis virus, and CMV — can cause encephalitis and fever. Toxoplasma, mycobacteria, and fungal infections should also be considered. Whipple's disease can present with fever and mood or memory disturbances. Noninfectious causes of apathy, such as adverse effects of medication or hypothyroidism, warrant consideration as well.
The patient had otherwise been well since his heart transplantation, without subsequent rejection or infectious episodes. His medications included fluvastatin, mycophenolate, and cyclosporine. He received 5 mg of prednisone daily for the first 3 years, after which the dosage was tapered to 0; he had been off of prednisone treatment for 3 years. The patient was a high-school graduate and lived in the midwestern United States with his wife. He did not smoke, and he drank alcohol only on occasion. He used to enjoy hunting and fishing but had done neither for several months. He fed birds in his garden but had none as pets. He had never traveled outside the country.
The patient is maintained on dual drug therapy and has not had recent treatment for rejection, thus his immunosuppression is relatively light for a recipient of a solid-organ transplant. Endemic fungi are not particularly common in heart-transplant recipients, but in the Midwest, histoplasmosis and blastomycosis should be considered. Zoonoses such as tularemia or Q fever may be acquired in a hunter and may present with fever alone.
On physical examination, his temperature was 38.3°C, his heart rate 95 beats per minute and regular, his respiratory rate 12 breaths per minute, his blood pressure 153/80 mm Hg, and his oxygen saturation 98% while he was breathing ambient air. He was alert and oriented to person, place, and time but appeared apathetic, with a flat affect. He was slow to respond to questions. He scored 22 points on a Mini–Mental State Examination (maximum score, 30), with deficiencies in recall and calculation. His lungs were clear to auscultation bilaterally. Cardiovascular examination revealed distinct heart sounds, with no audible murmur and a well-healed sternotomy scar. His abdomen was soft and nontender, with no palpable organomegaly. The surgical incision in the right upper quadrant appeared to be healing well. The arms and legs were warm and without pitting edema. Neurologic examination showed intact cranial nerves, strength of 5/5 bilaterally, no clonus, and no sensory loss. Funduscopic examination was normal. There was no nuchal rigidity.
His white-cell count was 8800 per cubic millimeter, with 53% neutrophils, 43% lymphocytes, and 4% monocytes. The hemoglobin was 10.8 g per deciliter, and the platelet count 199,000 per cubic millimeter. Liver-function tests were normal. The thyrotropin level was normal. The serum creatinine level was 0.9 mg per deciliter. The results of an electrolyte panel and urinalysis were normal.
Apart from his mental-status abnormalities and temperature, his physical examination is unremarkable. His Mini–Mental State Examination score of 22 indicates mild cognitive impairment (a score below 23 is considered abnormal for a person with education at the high-school level). His laboratory tests reveal anemia, but this is common in patients treated with mycophenolate. In an immunocompromised host, the list of pathogens that could infect the parenchyma of the brain, leptomeninges, or both is long. We should therefore focus on pathogens that may result in prolonged fever followed by changes in mental status. CMV encephalitis may present this way but usually occurs in patients with much greater immunosuppression. Varicella–zoster virus involving the central nervous system can be protean in its presentation but often has a more fulminant cerebral angiitis or strokelike presentation. Encephalitis due to herpes simplex virus or HHV-6 would be expected to progress more rapidly. Pyogenic, fungal, or parasitic (e.g., due to aspergillus species, nocardia species, or toxoplasma) brain abscess may present in this way. The pathogens that cause chronic meningitis (e.g., endemic fungi or mycobacterial pathogens) may cause changes in mental status indirectly (e.g., by causing hydrocephalus). Post-transplantation lymphoproliferative disease and less common entities such as Whipple's disease remain in the differential diagnosis. Magnetic resonance imaging (MRI) of the brain and lumbar puncture are indicated. Surrogate tests (e.g., cerebrospinal fluid tests for cryptococcal antigen, CMV DNA, and EBV PCR; and serum aspergillus galactomannan assay) may provide additional information.
A CT scan of the head showed mild hydrocephalus but no focal lesions, findings that were confirmed on MRI of the brain (Figure 1Figure 1MRI of the Brain.). A lumbar puncture revealed clear cerebrospinal fluid containing 102 nucleated cells comprising 47% lymphocytes and 47% neutrophils, with a protein level of 99 mg per deciliter and a glucose level of 42 mg per deciliter (2.3 mmol per liter). The serum glucose level was 125 mg per deciliter (6.9 mmol per liter). The opening pressure was 22 cm of water. Repeat serum PCR assays for CMV and EBV were negative.
Since the MRI scan reveals no mass or ring-enhancing lesions or areas of cerebritis, brain abscess from any cause can be ruled out. The results of the lumbar puncture are consistent with meningitis; the duration of symptoms probably indicates chronic meningitis, which may lead to hydrocephalus with changes in mental status. The cerebrospinal fluid analysis in this patient is not very helpful in distinguishing among possible causes of his symptoms. In addition to the infections already mentioned, syphilis, Lyme neuroborreliosis, brucella species, acanthamoeba species, actinomyces, and candida species should be mentioned as possible diagnoses. Rarely, enterovirus may cause chronic meningitis; PCR is usually diagnostic. Noninfectious causes of chronic meningitis include various connective tissue, inflammatory, and malignant diseases such as systemic lupus erythematosus, Behçet's disease, and sarcoidosis. His clinical presentation and previous imaging (e.g., chest CT with no changes suggestive of sarcoidosis), however, do not suggest any of these noninfectious diagnoses. If microscopical analysis of the cerebrospinal fluid is negative, further analysis of the specimen is necessary and should include a cryptococcal-antigen test on the cerebrospinal fluid; the Venereal Disease Research Laboratory (VDRL) test for syphilis; a test for histoplasmosis antigen; PCR assays for enterovirus, M. tuberculosis, and the agent of Whipple's disease; and analysis for antibodies to endemic fungi. In general, in the midwestern United States, endemic fungi or M. tuberculosis (if a history of exposure is elicited) are more common causes of chronic meningitis than are the other agents.
Gram's staining and fungal staining of cytospin cerebrospinal fluid were negative for organisms. The cryptococcal-antigen test was negative. A stain for acid-fast bacilli and a cerebrospinal fluid PCR assay for herpes simplex virus were also negative. A VDRL test, a PCR assay for enterovirus, and testing for histoplasmosis were not performed on the cerebrospinal fluid, but some of the specimen was stored for future analyses. A tuberculin skin test revealed no induration. The patient was started on intravenous acyclovir pending the results of the PCR assay for herpes simplex virus. The patient was also started on ceftriaxone, vancomycin, and ampicillin.
The relatively subacute course and lack of nuchal rigidity make bacterial meningitis, including Listeria monocytogenes (the rationale for ampicillin), and viral meningitis, including herpes simplex virus (the rationale for acyclovir), unlikely. A diagnosis of cryptococcal meningitis is very unlikely in light of the negative serum and cerebrospinal fluid tests for antigens. A tuberculin skin test is not a sensitive test for active tuberculosis, particularly in an immunosuppressed patient.
At this point, I would not begin empirical therapy unless the patient's symptoms are rapidly becoming worse. Instead, I would wait for results of these tests as well as of cytologic studies, which may identify lymphoma or carcinomatous meningitis. A parameningeal focus may also result in pleocytosis, but one has not been apparent on imaging studies to this point.
Three days later, cerebrospinal fluid culture grew Cryptococcus neoformans. A PCR assay of the isolate showed it to be C. neoformans var. grubii (serotype A). The prozone phenomenon was ruled out by checking the cerebrospinal fluid for cryptococcal antigen after predilution of cerebrospinal fluid samples; the cryptococcal-antigen test remained negative. The patient was started on liposomal amphotericin B and flucytosine. After 1 week of treatment, his fever resolved and he became less apathetic. A repeat MRI scan 3 weeks later showed no hydrocephalus.
Although human immunodeficiency virus (HIV) infection is the most common condition associated with cryptococcal meningitis, recipients of solid-organ transplants are at risk. The more indolent course observed in this patient is probably related to relatively preserved immune function as compared with a patient with advanced acquired immunodeficiency syndrome. When cryptococcal-antigen tests are performed in both the serum and cerebrospinal fluid of patients with cryptococcal meningitis, false negative tests are unusual. In this case, the false negative result was probably related to low titers rather than the prozone phenomenon that may occur when very high titers of antigen are present.
The patient was treated with liposomal amphotericin B and flucytosine for 6 weeks, followed by 400 mg of oral fluconazole daily indefinitely. Six weeks after the initiation of therapy, he still had problems with short-term recall and concentration but was more animated and interactive. At a follow-up visit 4 months after therapy was started, his problems with short-term memory and concentration had completely resolved; his wife stated, “He is back.”
Commentary
Although the mortality from cryptococcal meningitis has decreased in the United States since the introduction of highly active antiretroviral therapy for HIV,1 cryptococcal meningitis remains a major burden worldwide.2 Epidemiologic reports have also identified other at-risk patients such as organ-transplant recipients; between 20% and 60% of the cases of cryptococcal disease in HIV-negative patients occur in recipients of solid-organ transplants.3,4 The primary risk factor that predisposes to cryptococcal infection in organ-transplant recipients is the extent of immunosuppression; a particularly strong risk factor is the use of corticosteroids, with risk appearing to be directly proportional to the dose. 5-7 Wu and colleagues, however, reported that the incidence of cryptococcal disease was higher in heart-transplant recipients than in recipients of other types of transplants, perhaps because of the greater level of immunosuppression commonly used in this group of patients.3 However, our patient's level of immunosuppression was relatively low.
The taxonomy of cryptococcus has changed in recent years. C. neoformans var. neoformans encompasses serotype D strains and C. neoformans var. grubii includes serotype A strains. C. gattii (serotypes B and C) is now recognized as a separate species, usually infecting patients with normal immune status who live in tropical and subtropical areas. C. neoformans var. grubii accounts for 90% of cases of cryptococcosis seen in the United States, and was the cause of cryptococcal meningitis in our patient.
Since the initial reports, in 1963, of detection of cryptococcal capsular antigen by latex agglutination,8 this assay has become the test of choice for the diagnosis of cryptococcal infections because of the rapid availability of results and the excellent sensitivity and specificity. 9 As with any test, however, false negative and false positive tests may occur10,11; rates of false negative tests of up to 6% have been reported. As for any test, results must be interpreted in the context of pretest probability.
Reasons for a negative cryptococcal-antigen test in a patient with cryptococcal meningitis include low organism burden, early meningitis, poorly encapsulated strains, or the prozone phenomenon,10 whereby a high level of antigen thwarts formation of the antigen–antibody complex. Why did our patient have a negative cryptococcal latex-agglutination test? Our patient's cryptococcal strain was encapsulated, and we ruled out the prozone phenomenon, so the most likely explanation was a low organism burden.
How can clinicians avoid mistakenly ruling out a disease when a highly sensitive test is negative? It is important to remember that no test is 100% sensitive and that technical problems (e.g., operator error or poorly performing reagents) and factors specific to the organism or patient may result in a false negative test. When the diagnosis is elusive, repeating studies or using different diagnostic methods may be necessary. These approaches avoid the cognitive error of “blind obedience,” in which a negative test result inhibits further consideration of a diagnosis despite a high pretest probability of the disease. 12 Diagnostic uncertainty is often greater in immunocompromised hosts, including transplant recipients, in whom typical manifestations of infectious processes are often lacking. This was the case in our patient, who initially had no signs localizing his fever to the central nervous system.
This case highlights another important question: When should we begin treatment for a disease that is suspected but not yet verified? Our discussant appropriately urged ongoing diagnostic testing but cautioned against empirical therapy. The ratio of the benefits of a treatment to its risks determines how sure one has to be of the presence of a disease before administering the treatment, and each of these treatments has a different therapeutic threshold (i.e., the probability of a particular disease above which treatment should be initiated).13 Empirical antifungal treatment with amphotericin, for example, has a relatively high therapeutic threshold because of the nephrotoxicity of the drug and the need for prolonged therapy. With azole antifungal agents, the potential for drug interactions with immunosuppressive medications must be considered.
Our case underscores that a negative test for cryptococcal antigen, like other negative diagnostic tests, does not rule out the disease when the clinical presentation is highly suggestive. Arriving at the correct diagnosis in the present case was especially challenging, since the clinical presentation and cerebrospinal fluid profile of many chronic meningitides are very similar.14,15 Fortunately, because cryptococcus grows readily on selective media, we were able to establish the diagnosis and initiate treatment. Most important, despite our delay in making the diagnosis, our patient recovered fully.
Supported by a K12 institutional training grant (5 K12 AG019247-05) from the National Institutes of Health to the University of Wisconsin School of Medicine and Public Health and a grant from the Society of Critical Care Medicine (to Dr. Safdar), and an Advanced Career Development Award from the Health Services Research and Development Program of the Department of Veterans Affairs (to Dr. Saint).
No potential conflict of interest relevant to this article was reported.
Source Information
From the Department of Medicine, University of Wisconsin–Madison School of Medicine, Madison (N.S., C.L.A., S.N.); and the Department of Internal Medicine, University of Michigan Medical School (D.R.K., S.S.), and the Department of Veterans Affairs Health Services Research and Development Center of Excellence (S.S.) — both in Ann Arbor.
Address reprint requests to Dr. Safdar at the Division of Infectious Diseases, University of Wisconsin–Madison, H4/572 Clinical Sciences Center, 600 Highland Ave., Madison, WI 53792, or at ns2@medicine.wisc.edu.
Subscribe to:
Post Comments (Atom)
0 comments
Post a Comment