Suspected Central Nervous System Infections in HIV-Infected Adults

Objectives: To study the differential diagnosis of HIV-infected patients with suspected central nervous system (CNS) infections and the association of CD4 counts with the final diagnosis. Methods: We analyzed HIV-infected patients from a prospective cohort study on the diagnostic accuracy of clinical and laboratory characteristics in adults with suspected CNS infections in an academic hospital in Amsterdam, the Netherlands, who underwent cerebrospinal fluid (CSF) examination. Results: Thirty-four (9.4%) out of 363 patients with suspected CNS infections were HIV-positive of whom 18 (53%) were diagnosed to have CNS infection, with median CD4 counts of 255 cells/μl. The spectrum of CNS infections consisted of progressive multifocal leukoencephalopathy in three patients (17%); cryptococcal meningoencephalitis, toxoplasma encephalitis, angiostrongylus eosinophilic meningitis, and HIV encephalitis each in two (11%); and cytomegalovirus encephalitis, neurosyphilis, tuberculous meningoencephalitis, histoplasma encephalitis, and varicella-zoster virus meningitis each in one (6%). Clinical characteristics and blood parameters did not differ between HIV-infected patients with CNS infections and other diagnoses. The best predictor for CNS infections was the CSF leukocyte count (AUC = 0.77, 95 CI% 0.61–0.94). The diagnosis of CNS infection was not associated with the CD4 count (median 205 vs. 370, p = 0.21). Two patients (11%) with CNS infections died and two (11%) had neurological sequelae. Conclusions: Half of the patients with suspected CNS infections are diagnosed with a CNS infection, and this was not related to CD4 counts. The best predictor for CNS infections was the CSF leukocyte count.


INTRODUCTION
Human immunodeficiency virus (HIV) is a neurotrophic, neuroinvasive, and neurovirulent pathogen (1), which can cause direct infection of the central nervous system (CNS) but also predisposes to a variety of other neuroinfections through impaired T-cell mediated immunity (2)(3)(4). It has been estimated that in the era before the introduction of combination antiretroviral therapy (cART), around 10% of HIV-infected patients initially presented with neurological disorders, and 30-50% of them developed neurological complications during the progression of their disease (5). Although recent large observational cohorts of HIV-infected patients have reported a significant decrease in the overall incidence of the most frequent HIV-associated neurological disorders in the cART era (6,7), CNS infections have remained an important cause of morbidity and mortality (8).
Diagnosis of CNS infections in HIV-infected patients can be challenging as a variety of conditions should be considered in the differential diagnosis, and it may be caused by a broad spectrum of pathogens with diverse and overlapping clinical manifestations (9). The cause of CNS infections in HIV have been related to the CD4 lymphocyte count, of which a value below 200 has been described to predispose to cerebral toxoplasmosis, progressive multifocal leukoencephalopathy (PML), and cryptococcal meningitis (10,11). HIV patients with higher CD4 counts, on the other hand, are expected to be infected with non-opportunistic pathogens (9). A number of studies have reported the spectrum of CNS infections in patients with a low range of CD4 count (7,(12)(13)(14), but the distribution of various CNS infections has not addressed in HIV-infected subgroups with a higher range of CD4 count.
We describe a population of HIV-infected patients included in a prospective study on the diagnostic accuracy of clinical and laboratory characteristics in patients with suspected CNS infections. We analyzed the causes of CNS infections, the differential diagnosis, and the association of CD4 counts with the final diagnosis.

METHODS
We studied HIV-positive patients over 16 years included in a cohort of patients with suspected CNS infections in an academic hospital in Amsterdam, the Netherlands, who underwent CSF examination. The methods have been described in detail previously (15). Data collection regarding past medical history, clinical characteristics on presentation, and results of ancillary investigations including blood and CSF analysis, microbiological examinations, and imaging studies was done using online case record forms.
The final diagnosis was classified by two researchers based on all available clinical data, including microbiological examination of CSF. The classification was performed sorting each diagnosis into one of the following five categories: (1) CNS infection, (2) nervous system inflammation without infection, (3) noninfectious non-inflammatory neurological disorder, (4) nonneurological infection, and (5) other systemic disorder. Outcome was scored using the Glasgow outcome score (GOS). A favorable outcome was defined as a score of 5, and an unfavorable outcome was defined as a score of 1-4.
Continuous data were described with medians and interquartile range and categorical variables with frequency and percentage. The Mann-Whitney U-test was used for continuous variables, whereas Fisher's exact test and chi-square tests were used for categorical variables, as appropriate. The area under the curve (AUC) was measured using the receiver operating characteristic curve (ROC). A p-value < 0.05 was considered as statistically significant. The study was approved by the medical ethical committee of the Academic Medical Centre, Amsterdam, the Netherlands.
Out of 34 HIV-infected patients, 18 (53%) were categorized as having a CNS infection, 7 (21%) with non-infectious non-inflammatory neurological disorders, 6 (18%) with nonneurological infections, 2 (6%) with non-infectious nonneurological disorders, and 1 (3%) with non-infectious neuroinflammatory disorders. Of those who were finally diagnosed with CNS infection, 13 (72%) were male and 5 (28%) females with a median age of 43 years. Eight (44%) patients with CNS infections were on cART, and one discontinued treatment 1 month before. Their median CD4 was 205 (IQR 55-455). Out  The spectrum of CNS infections included PML in three (17%) cases; cryptococcal meningoencephalitis, toxoplasma encephalitis, angiostrongylus eosinophilic meningitis, and HIV encephalitis each in two (11%); and cytomegalovirus encephalitis, syphilitic meningitis, tuberculous (TB) meningoencephalitis, histoplasma encephalitis, and varicella-zoster virus (VZV) meningitis each in one (6%). Final diagnoses were microbiologically confirmed in all but two patients, which were classified diagnosed as probable viral CNS infection by the assessors. Both of them were on cART and had a CD4 count of more than 200/µl. Two patients who were diagnosed with Angiostrongylus cantonensis eosinophilic meningitis were young, male, and on cART and had relatively high CD4 counts (430 and 1,020/µl) and marked CSF pleocytosis (441 and 814/µl). In three (16%) patients who were diagnosed with cerebral toxoplasmosis, cryptococcal meningitis, and neurosyphilis, the CNS infection was their first manifestation of HIV infection (Tables 2, 3).
Overall, 8 of 34 (24%) patients had an unfavorable outcome, 4 in each group of CNS infection and other diagnoses (22 and 25%, respectively). Of these, two (11%) patients with CNS infections died. The causes of death were nosocomial sepsis and post-anoxic encephalopathy following resuscitation, and were deemed to be not related to the CNS infection.

DISCUSSION
Our study shows that half of HIV-infected patients with suspected CNS infections are indeed diagnosed with a CNS infection. Our population included patients with a relatively high median CD4 count (255 cells/µl), and presence of CNS infection was not associated to the CD4 count. Most of the previous studies, even after the introduction of cART, reported the spectrum of neurologic complications or CNS infections in HIV-infected patients with a median CD4 count in a range between 30 and 90 cells/µl (7,13,14,(16)(17)(18)(19)(20). A higher median CD4 count in HIV-infected patients in our study might be related to high access to HIV-related health services in the Netherlands with 89% of HIV-infected patients linked to care and 92% treated with cART in 2017 (21). We found that CNS infections in HIV-infected patients had many different causes, of which most were opportunistic pathogens, even though the CD4 count was relatively high. This finding, despite a relatively high median CD4 count in our patients, could be partly attributed to unmasking of occult opportunistic infections as a result of immune reconstitution inflammatory syndrome (IRIS) in HIV-infected patients. Unmasking IRIS is manifested by a robust, inflammatory immune response against the infective pathogen while deteriorating the clinical condition of the patient who recently started taking cART and was previously unable to respond to the opportunistic infection because of immunosuppression (22).
The best predictor for CNS infections was the CSF leukocyte count. This is comparable with findings in HIV-negative patients with suspected CNS infections (15). Clinical characteristics had previously been found to have a low predictive value in patients with suspected CNS infections (15). Typical clinical characteristics for meningitis, such as neck stiffness, headache, fever, and an altered mental status were low among HIV patients with CNS infections (<50%). Therefore, there should be a low threshold for CSF examination in HIV-positive patients with neurological symptoms. Because of the risk of space occupying lesions, cranial imaging should be performed in these patients prior to the lumbar puncture (23). If patients have suspected bacterial meningitis, antibiotics and dexamethasone if indicated should be administered before sending them to the CT scanner (24).
We identified a broad range of causative pathogens. PML, cryptococcal meningoencephalitis, and cerebral toxoplasmosis were the most common opportunistic neurological infections. Studies from other European countries, including Spain (16), France (10), and the United Kingdom (25), also reported similar findings and identified these three opportunistic infections among the four most common causes of CNS involvement in HIV-infected patients. However, the population of HIV-infected individuals in these studies had lower CD4 counts, as compared with our cohort. Although tuberculous meningoencephalitis has been reported as one of the most frequent causes of neurological syndromes in HIV-infected patients in the countries endemic for tuberculosis (7,12,19,20,(26)(27)(28)(29), it was diagnosed in only one of our patients.
Although about one third of all patients with CNS infection in the whole cohort were diagnosed to have bacterial meningitis, none of the HIV-infected patients were diagnosed with bacterial meningitis. Previous studies on bacterial meningitis show that the incidence is 8-19 times increased in HIV-infected patients, but the absolute risk is still small (30,31). In resource-rich countries, other causes of CNS infections have been reported to be more prevalent in HIV-infected patients (32).
Our study had several limitations. First, the number of HIV-infected patients in this study is relatively small, which makes it difficult to draw firm conclusions about the spectrum of HIV-infected patients with CNS complications. Second, only patients who underwent lumbar puncture were included in the study. As a result, we did not study HIV-infected patients whose diagnoses were confirmed by stereotactic or open surgical biopsy or were not candidate for lumbar puncture. This may influence the frequency distribution of the diagnoses because neurological disorders that manifest as single or multiple CNS mass lesions may require brain biopsy for confirmation of the diagnosis. However, recent advances in neuroimaging and molecular-based laboratory tests on CSF have increased the yield of diagnostic tests in patients with CNS infections, decreasing the need for more invasive diagnostic procedures (33).

CONCLUSION
In conclusion, half of the patients with suspected CNS infections are diagnosed with a CNS infection. While the CD4 count is not a good predictor for diagnosis of CNS infection vs. other diagnosis in HIV-infected patients, CSF white cell counts are essential in making the diagnosis. Therefore, if a CNS infection is suspected, cerebrospinal fluid examination is indicated in all patients in whom lumbar puncture can be performed safely (34).

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by the Medical Ethical Committee of the Academic Medical Centre, Amsterdam, the Netherlands. The patients/participants provided their written informed consent to participate in this study.

AUTHOR CONTRIBUTIONS
FS: performed the research, analyzed the data, and wrote the first draft of the paper. FS, DB, and MB: designed the study and have read and approved the final manuscript. DB and MB: revised the paper. All authors contributed to the article and approved the submitted version.