Study design
We conducted a hospital-based cross-sectional study from April to July 2017.
Study site and population
The study was conducted at Mbagathi Hospital, a referral facility located in Nairobi County. The evaluation targeted patients ≥18 years scheduled for lumbar puncture (LP) and routine blood sample collection for CM diagnosis.
Inclusion criteria
LP requested by health care provider, availability of remnant serum and CSF (≥ 500µl) after routine Cryptococcus LA or culture was performed on HIV-positive patients ≥18 years with the ability and willingness to provide informed consent.
Exclusion criteria
Patients involuntarily incarcerated in the hospital for psychiatric or physical illness, any patient/patient with a guardian who was deemed mentally unstable or unable to provide informed consent. Patients on any antifungal treatment and patients who were having repeat LPs.
Sample size assumptions and calculation
The sample size of 125 participants was calculated using Fisher’s formulae, assuming the expected proportion of agreement between LFA and other methodologies (p) was 97.7 % and the precision (P) of 3% (16).
Sampling methods
We reviewed records for 6 months at Mbagathi Hospital and obtained an average of 95 cases of suspected CM in a month. To achieve a sample size of 125 patients using the estimated sampling frame of 285 within a period of 3 months, every second HIV-positive patient ≥18 years suspected of CM and scheduled for routine LP and blood collection for CM diagnosis was enrolled after giving written consent to obtain an additional capillary blood and use of the remnant CSF and serum for evaluation of LFA.
Data collection
The attending laboratory technologist collected blood samples from patients as part of routine testing requested by the clinician. Serum was centrifuged and separated for LA and LFA assays. CSF samples were collected by LP and centrifuged. The supernatant was used for LFA and LA assays and the pellet for culture. Sera were tested by LA, and CSF (where available) was tested by India ink microscopy and culture for clinical management of the patient. Leftover sera and CSF samples were used for the laboratory evaluation. A minimum of 500 µl of the remaining sample was aliquoted into 1.8 ml cryogenic vials. The samples were stored at 4°C for a maximum of 72h or at -20°C awaiting transportation to the Central Microbiology Reference Laboratory (CMRL). Additionally, a non-routine finger prick capillary blood sample was requested from all enrolled patients. Using standardized lancets and micro capillary tubes, ~50 µl of blood was transferred into micro centrifuge tubes containing LFA specimen diluent. The test was performed at the sample collection site per manufacturer instructions. Sera and CSF sample processing were done at CMRL per laboratory standard operating procedures and manufacturer instructions (Figure 1).
Data management and analysis
Data on the test type and test results in different sample types were entered and cleaned in MS-Excel version 2013. Statistics were calculated using GraphPad QuickCalcs 2017 (GraphPad Software, Inc., La Jolla, CA) with categorical data analysis to assess sensitivity, specificity, predictive values, confidence intervals (CIs) of proportion, overall percent agreement, and kappa (k) coefficients of India ink, LA, LFA and culture in sera, CSF and capillary blood. Interpretation of k was per standard guidelines (17).