Sensitivity and specificity of CRP and symptom screening as tuberculosis screening tools among HIV-positive and negative outpatients at a primary healthcare facility in Lusaka, Zambia: a prospective cross-sectional study

• Published in: BMJ open Vol. 13; no. 4; p. e061907
• Main Authors: Kagujje, Mary, Mwanza, Winnie, Somwe, Paul, Chilukutu, Lophina, Creswell, Jacob, Muyoyeta, Monde
• Format: Journal Article
• Language: English
• Published: England British Medical Journal Publishing Group 18.04.2023; BMJ Publishing Group LTD; BMJ Publishing Group
• Series: Original research
• Subjects: Adult; Biomarkers; C-Reactive Protein - analysis; Cross-Sectional Studies; Data analysis; Health facilities; HIV; HIV & AIDS; HIV Infections - complications; Human immunodeficiency virus; Humans; Laboratories; Mass Screening; Microscopy; Outpatients; Population; Primary Health Care; Prospective Studies; Public Health; Sample size; Sensitivity and Specificity; Tuberculosis; Tuberculosis - diagnosis; Zambia; Zambia; Lusaka Zambia; HIV & AIDS; Public health; Tuberculosis
• ISSN: 2044-6055; 2044-6055
• DOI: 10.1136/bmjopen-2022-061907

ObjectivesTo evaluate the performance of point-of-care C-reactive protein (CRP) as a screening tool for tuberculosis (TB) using a threshold of 10 mg/L in both people living with HIV (PLHIV) and HIV-negative individuals and compare it to symptom screening using a composite reference for bacteriological confirmation of TB.MethodsProspective cross-sectional study.SettingA primary healthcare facility in Lusaka, Zambia.ParticipantsConsecutive adults (≥18 years) presenting for routine outpatient healthcare were enrolled. Of the 816 individuals approached to participate in the study, 804 eligible consenting adults were enrolled into the study, of which 783 were included in the analysis.Primary outcome measuresSensitivity, specificity, positive predictive value and negative predictive value (NPV) of CRP and symptom screening.ResultsOverall, sensitivity of WHO-recommended four-symptom screen (W4SS) and CRP were 87.2% (80.0–92.5) and 86.6% (79.6–91.8) while specificity was 30.3% (26.7–34.1) and 34.8% (31.2–38.6), respectively. Among PLHIV, sensitivity of W4SS and CRP was 92.2% (81.1–97.8) and 94.8% (85.6–98.9) while specificity was 37.0% (31.3–43.0) and 27.5% (22.4–33.1), respectively. Among those with CD4≥350, the NPV for CRP was 100% (92.9–100). In the HIV negative, sensitivity of W4SS and CRP was 83.8% (73.4–91.3) and 80.3% (69.5–88.5) while specificity was 25.4% (20.9–30.2) and 40.5% (35.3–45.6), respectively. Parallel use of CRP and W4SS yielded a sensitivity and NPV of 100% (93.8–100) and 100% (91.6–100) among PLHIV and 93.3% (85.1–97.8) and 90.0% (78.2–96.7) among the HIV negatives, respectively.ConclusionSensitivity and specificity of CRP were similar to symptom screening in HIV-positive outpatients. Independent use of CRP offered limited additional benefit in the HIV negative. CRP can independently accurately rule out TB in PLHIV with CD4≥350. Parallel use of CRP and W4SS improves sensitivity irrespective of HIV status and can accurately rule out TB in PLHIV, irrespective of CD4 count.