الفهرس 
English SummaryOnly 14 pages are availabe for public view
 |  | from | 77 |  |  |
| | | from | 77 | | |
Abstract
The diagnosis of meningitis in HIV patients is challenging due to altered immune responses. The current laboratory methods remain inadequate or inaccessible in developing countries. Diagnostic scoring systems were recently proposed for use in research settings to help prompt and easy differential diagnosis.
The aim of this study was to create a simple diagnostic rule to predict
meningitis in HIV patients and to address the enigma of differentiating bacterial (BM), tuberculous (TBM), and cryptococcal (CCM) meningitis based on clinical features alone that might be enhanced by easy to obtain laboratory testing.
We retrospectively enrolled a total of 352 HIV patients presenting with
neurological manifestations suggesting meningitis at a tertiary care hospital over the
last two decades (2000-2018). Relevant clinical and laboratory information were retrieved from inpatient records. These patients underwent clinical evaluation and a comprehensive diagnostic workup for meningitis that included LP (CSF cytology,
chemistry, serology and culture), blood (cytology, chemistry, serology and culture)
and imaging investigations. The clinical features independently predicting meningitis
or its different types in microbiologically proven meningitis cases were modelled by
multivariate logistic regression to create diagnostic rules in an exploratory dataset.
The performance of the meningitis prediction rule was assessed, and its applicability
was validated in a confirmatory subset of data.
In total, 234 patients (66.3%) were microbiologically proven to have meningitis
as a final diagnosis [cryptococcal (26.9%), H. influenza (20.9%), pneumococcal
vi
(17.1%), tuberculous (15.0%), meningococcal (14.1%), toxoplasma (4.3%) and viral
(1.7%)], while 118 (33.7) were eventually diagnosed with other neurological disorders
[49 (13.9%) had encephalitis, 31 (13.1%) had HIV encephalopathy, and 38 (19.6%)
were diagnosed with IRIS].
While the classic meningitic symptoms were common, their presence increased
the probability of having meningitis. AIDS clinical stage, jaundice, injecting drug use
(IUD), and CrAg seropositivity were equally important in predicting meningitis among HIV patients. Arthralgia [OR (95% CI)= 101.9 (4.6 – 2255.0)] and elevated
CSF LDH [OR (95% CI)= 5.6 (1.4 – 21.9)] were strong predictors of BM. Patients
with cryptococcal antigenemia had 25 times the odds of having CCM, whereas
neurological deficits were highly suggestive of TBM. The meningitis diagnostic index
had a sensitivity of 78.7%, a specificity of 78.1% and corresponding positive and
negative predictive values of 86.2% and 67.9%, respectively. The model had a
moderate degree of agreement with the initial diagnostic work up [kappa=0.551, p<
0.001]. It accurately predicted meningitis in 81.3% of HIV patients in the confirmatory data set as an external validation with positive and negative predictive values of 79.1% and 52.4%, respectively.In conclusion, the proposed clinical prediction rule has a good diagnostic potential in our population when blood and CSF culture factors were excluded. The results are encouraging with regard to supporting decision‐making in resource-poor settings.