Secondary Logo

Journal Logo

Translational Research

Brief Report: Yield and Efficiency of Intensified Tuberculosis Case-Finding Algorithms in 2 High-Risk HIV Subgroups in Uganda

Semitala, Fred C. MBChB, MMed, MPHa,b,c; Cattamanchi, Adithya MD, MASd; Andama, Alfred MSca,c; Atuhumuza, Elly MScc; Katende, Jane SWASAb; Mwebe, Sandra BSN, MPHc; Asege, Lucy BBLTc; Nakaye, Martha BBLT, MScc; Kamya, Moses Robert MMed, MPH, PhDa,c; Yoon, Christina MD, MAS, MPHd

Author Information
JAIDS Journal of Acquired Immune Deficiency Syndromes: December 1, 2019 - Volume 82 - Issue 4 - p 416-420
doi: 10.1097/QAI.0000000000002162
  • Free



Tuberculosis (TB) remains the leading cause of HIV death, accounting for one-third of all HIV deaths worldwide.1 To reduce TB burden, the World Health Organization (WHO) recommends intensified case finding (ICF)—symptom screening, followed by Xpert MTB/RIF (Xpert) confirmatory testing—for all people living with HIV (PLHIV) at every clinic visit.2 However, PLHIV constitute a heterogeneous population with respect to TB risk. Patients new to HIV care represent an HIV subgroup with higher TB prevalence and increased mortality risk, frequently due to undiagnosed TB.3–6 These results suggest that in resource-limited settings, patients new to care should be prioritized for ICF and that ICF should be rigorously implemented to maximize TB case detection. However, rigorous implementation of ICF may be associated with high rates of unnecessary Xpert testing due to the poor specificity of symptom-based TB screening,4,7–9 which may make routine implementation less likely.

The emerging literature suggests that replacing symptom screening with C-reactive protein (CRP) testing may improve the efficiency and reduce the costs of ICF.10–14 Our previous work identified CRP,12–14 which can be measured from capillary blood using a rapid and low-cost point-of-care (POC) assay—as the first and only test thus far to meet the WHO target product profile for an effective TB screening test.15 However, the yield and performance of POC CRP-based ICF among different HIV subgroups is unknown.

To provide routine HIV programs with the evidence needed to efficiently implement ICF, we compared the yield and efficiency of POC CRP-based ICF to the current ICF algorithm among patients new to HIV care and patients already engaged in HIV care. In addition, we also compared the yield and efficiency of symptom- and POC CRP-based ICF between the 2 HIV subgroups.


Study Population

Patient recruitment, study procedures, and the diagnostic accuracy of symptom-based and POC CRP-based TB screening in 1177 patients enrolled from July 2013 to December 2015 have been previously reported.14,16 Here, we present results on the full cohort of 1794 consecutive patients prospectively enrolled through December 2016 and compare the yield and efficiency of symptom- and POC CRP-based ICF in 2 subpopulations of PLHIV initiating antiretroviral therapy (ART) from 2 urban HIV clinics in Kampala, Uganda: patients new to HIV care and patients engaged in care (pre-ART patients with ≥1 previous HIV clinic visit at the time of study entry). During the study period, the Uganda Ministry of Health revised their ART guidelines to adopt “universal test-and-treat” which eliminated the requirement for repeated pre-ART counseling visits [ART initiation decisions were based on CD4 count ≤500 cells/µL or other indications (eg, pregnancy, HIV serodiscordance)].

We enrolled ART-naive adults (age ≥18 years) with a pre-ART CD4 count ≤350 cells/µL within 3 months of study enrollment. We excluded patients taking medication with antimycobacterial activity (anti-TB therapy, TB preventive therapy, and fluoroquinolones) within 3 days of enrollment. All patients provided written informed consent, and the study was approved by the institutional review boards of the University of California, San Francisco, Makerere University, and the Uganda National Council for Science and Technology. This study conforms to the standards for the reporting of Diagnostic Accuracy Studies (STARD) initiative guidelines.17

Study Procedures

Data Collection and TB Screening

At enrollment, we collected standardized demographic and clinical data, including history of HIV clinic attendance; confirmed pre-ART status; administered the WHO symptom screen; and measured CRP concentrations from capillary blood using a standard sensitivity POC assay (BodiTech, South Korea). In accordance with the WHO guidelines, we considered patients to be symptom screen positive if they reported ≥1 TB symptom (current cough, fever, night sweats, and weight loss).2 We defined a POC CRP concentration of ≥8 mg/L (rounding to the nearest whole-number) as screen positive for TB based on receiver-operating characteristics analysis performed in the parent study.14

Sputum Collection and Testing

We collected a spot sputum specimen from all participants at study entry for Xpert MTB/RIF (Cepheid, Sunnyvale, CA) testing. All staff members performing Xpert testing were blinded to clinical and demographic data, including symptom screen status and POC CRP concentrations. We considered patients to have active TB if Xpert results were positive for Mycobacterium tuberculosis. Patients with indeterminate Xpert results underwent repeat Xpert testing with remaining sputum. We considered patients not to have active TB if Xpert results were negative.

Statistical Analysis

We compared categorical and continuous variables using the Wilcoxon rank-sum test, Fisher exact test, or χ2 test, as appropriate. To determine the diagnostic yield of symptom-based and POC CRP-based ICF for each HIV subgroup, we combined Xpert confirmatory testing to either symptom- or POC CRP-based screening. The diagnostic yield of each ICF algorithm is equal to the number of screen-positive patients who were diagnosed with Xpert-positive TB divided by the total number of Xpert-positive TB patients, irrespective of the screening status. We compared differences in diagnostic yield using the unpaired test of 2 proportions. To determine the efficiency of each ICF algorithm, we determined the number needed to test (NNT) using Xpert to detect one TB case for each screening strategy and for each HIV subgroup. We performed all analyses using STATA13 (STATA).


Study Population

We prospectively enrolled 1839 consecutive adults initiating ART and excluded 45 patients for the following reasons: 21 did not meet inclusion criteria, 21 had missing CD4 count results, and 3 had missing POC CRP results. Of the remaining 1794 patients, 1315 (73%) were new to HIV care and 479 (27%) were already engaged in HIV care (Table 1). Compared with patients engaged in care, patients new to care had lower median CD4 cell counts, lower median body mass index, and higher prevalence of Xpert-positive TB (10% vs. 4%, P < 0.001). More patients new to care than engaged in care screened positive for TB by both symptoms (90% vs. 85%, P = 0.01) and POC CRP (41% vs. 30%, P < 0.001).

Demographics and Clinical Characteristics of Patients Engaged and New to Care

Performance of ICF Algorithms Among Patients New to Care

The current ICF algorithm (symptom screening, followed by Xpert testing for all those who screen positive) required 1178/1315 (90%) patients new to care to undergo Xpert testing and identified 124/126 (98%) of all Xpert-positive TB cases (Table 2). Compared with the current ICF algorithm, POC CRP-based ICF reduced by more than half the proportion of patients requiring Xpert testing {90% vs. 41%; difference −48%, [95% confidence interval (CI): −51 to −45]} and missed 7 more TB cases. Thus, POC CRP-based ICF required half as many Xpert assays to detect one TB case (NNT 5 vs. 10) while maintaining similar diagnostic yield [93% (117/126) vs. 98% (124/126), difference −6% (95% CI: −11 to 0)] as the current ICF algorithm.

Performance of TB Screening/Xpert-Based ICF Strategies Among Patients New to HIV Care (N = 1315)

Performance of ICF Algorithms Among Patients Engaged in Care

A similar pattern was observed among patients engaged in care. Compared with the current ICF algorithm, POC CRP-based ICF required half as many Xpert assays to detect one TB case (NNT 8 vs. 19) but had lower diagnostic yield [100% (21/21) vs. 81% (17/21), difference −19% (95% CI: −41 to +3); Table 3], although the difference in yield did not reach statistical significance. Compared with patients new to care, both symptom- and POC CRP-based ICF were less efficient, requiring almost twice as many Xpert assays to detect one TB case.

Performance of TB Screening/Xpert-Based ICF Strategies Among Patients Engaged in HIV Care (N = 479)


Although ICF was introduced in 2011 as the cornerstone of TB control activities for PLHIV,2 there are few published data describing the yield of the current ICF algorithm in reference to Xpert, the confirmatory test used in the vast majority of settings with high TB burden. Therefore, to provide HIV program managers and country-level policymakers the evidence to support ICF scale-up using routine diagnostics, we compared the performance of 2 rigorously implemented ICF algorithms in reference to Xpert and compared ICF yield in 2 pre-ART HIV subgroups. We found that TB prevalence was high among all patients but 2.5-times higher among patients new to care, highlighting the need for increased vigilance in this subgroup and the importance of avoiding unnecessary delays in ART initiation. We also found that the current ICF algorithm, when performed in strict accordance with TB/HIV guidelines,2 detected nearly all (≥98%) TB cases but required most (≥85%) patients to undergo confirmatory testing with Xpert. By contrast, we found that POC CRP-based ICF reduced by half the proportion of patients in both subgroups requiring Xpert confirmatory testing. Among patients engaged in care, the improved efficiency of POC CRP-based ICF came at the expense of lower diagnostic yield, whereas POC CRP-based ICF had similar diagnostic yield as the current ICF algorithm among patients new to care. These results support the use of POC CRP-based ICF for all patients new to HIV care to improve the efficiency and reduce the cost of TB case detection.

A growing body of evidence suggests that using CRP to screen PLHIV for active TB could improve ICF efficiency.10–14 Consistent with these studies, we found that POC CRP-based ICF could substantially reduce the proportion of PLHIV in both subgroups requiring Xpert confirmatory testing, relative to the current ICF algorithm. Among patients new to care, POC CRP-based ICF detected 93% of all TB cases, exceeding the 90% sensitivity target recommended by the WHO for an effective TB screening test. However, among pre-ART patients engaged in care, POC CRP-based ICF detected only 17/21 (81%) of all Xpert-positive TB cases. Although the number of TB cases in this subgroup was small, the reduced yield of POC CRP-based ICF in this HIV subgroup may have important implications for all PLHIV. First, the lower TB prevalence among pre-ART patients engaged in care suggests that before study enrollment, patients with active TB either died, had their TB diagnosed, or transferred or defaulted from care. Multiple studies have shown that patients waiting to initiate ART have higher rates of clinic attrition and mortality (frequently due to undiagnosed TB),18,19 stressing the importance of immediate ART and rigorous ICF when PLHIV are first presenting for care. Second, the reduced yield of POC CRP-based ICF is likely due to the selective loss of the sickest TB/HIV patients (those most likely to die or have their TB diagnosed) while awaiting ART initiation,20 which may have significant implications for serial ICF. Because the accuracy of any test depends on the degree of previous testing in the population, the yield and efficiency of each repeat round of ICF can be expected to be lower than the last. Large well-powered studies are needed to evaluate whether lower POC CRP cut-points could improve the yield of POC CRP-based ICF, when applied to a population that has undergone previous TB screening.

Strengths of this study include (1) a large and well-characterized cohort ie, prototypical of patients initiating ART in high TB/HIV burden countries; (2) implementation of ICF in accordance with a strict protocol and in reference to Xpert, thus providing HIV programs estimates of expected ICF yield in settings where culture is not routinely available; and (3) prospective measurement of CRP concentrations using an US FDA-approved, simple, and low-cost POC assay that is available for immediate scale-up. Our study also has limitations. First, we restricted enrollment to ART-naive patients with advanced HIV because TB risk is highest and the need for ICF is greatest in this population. Future studies of ICF yield are needed in HIV subgroups with lower TB risk. Second, the number of TB cases among pre-ART patients engaged in care is small, which may have impacted measured estimates. Third, although scale-up of HIV test-and-treat will eventually eliminate pre-ART patients engaged in care as an HIV subgroup, the lower TB prevalence and reduced ICF yield in this subgroup suggests that serial ICF can be expected to have lower yield among patients who have previously undergone TB screening. Fourth, some patients new to care may have been misclassified (patients may have received HIV care previously, which may have included previous TB screening); thus, the yield of ICF among patients new to care may be even greater than reported here. Finally, we evaluated the yield of symptom- and POC CRP-based ICF in reference to Xpert (standard cartridge); future studies of ICF yield should combine symptom- and POC CRP-based screening with more sensitive rapid diagnostics (eg, Xpert Ultra) to further expand the evidence base needed to support routine ICF scale-up.

In conclusion, high TB prevalence among patients initiating ART in high TB/HIV burden settings confirms the need for rigorous ICF and immediate ART, when patients first present for HIV care. POC CRP-based ICF at this critical time could improve the efficiency of ICF, thus increasing the likelihood of ICF implementation, while maintaining similarly high diagnostic yield as the current ICF algorithm. POC CRP-based ICF should be prioritized for PLHIV new to HIV care and where resources for Xpert testing are limited.


1. WHO. Global Tuberculosis Report 2018. Geneva, Switzerland: World Health Organization; 2018. Licence: CC BY-NC-SA 3.0 IGO.
2. WHO. Guidelines for Intensified Tuberculosis Case-Finding and Isoniazid Preventive Therapy for People Living With HIV in Resource- Constrained Settings. Geneva, Switzerland: World Health Organization; 2011.
3. Agbor AA, Bigna JJ, Billong SC, et al. Factors associated with death during tuberculosis treatment of patients co-infected with HIV at the Yaounde Central Hospital, Cameroon: an 8-year hospital-based retrospective cohort study (2006-2013). PLoS One. 2014;9:e115211.
4. Henostroza G, Harris JB, Chitambi R, et al. High prevalence of tuberculosis in newly enrolled HIV patients in Zambia: need for enhanced screening approach. Int J Tuberc Lung Dis. 2016;20:1033–1039.
5. Mbu ET, Sauter F, Zoufaly A, et al. Tuberculosis in people newly diagnosed with HIV at a large HIV care and treatment center in Northwest Cameroon: burden, comparative screening and diagnostic yields, and patient outcomes. PLoS One. 2018;13:e0199634.
6. Turinawe K, Vandebriel G, Lowrance DW, et al. Operating characteristics of a tuberculosis screening tool for people living with HIV in out-patient HIV care and treatment services, Rwanda. PLoS One. 2016;11:e0163462.
7. Lawn SD, Brooks SV, Kranzer K, et al. Screening for HIV-associated tuberculosis and rifampicin resistance before antiretroviral therapy using the Xpert MTB/RIF assay: a prospective study. PLoS Med. 2011;8:e1001067.
8. Swindells S, Komarow L, Tripathy S, et al. Screening for pulmonary tuberculosis in HIV-infected individuals: AIDS clinical trials group protocol A5253. Int J Tuberc Lung Dis. 2013;17:532–539.
9. Kufa T, Mngomezulu V, Charalambous S, et al. Undiagnosed tuberculosis among HIV clinic attendees: association with antiretroviral therapy and implications for intensified case finding, isoniazid preventive therapy, and infection control. J Acquired Immune Defic Syndr. 2012;60:e22–8.
10. Lawn SD, Kerkhoff AD, Vogt M, et al. Diagnostic and prognostic value of serum C-reactive protein for screening for HIV-associated tuberculosis. Int J Tuberc Lung Dis. 2013;17:636–643.
11. Shapiro AE, Hong T, Govere S, et al. C-reactive protein as a screening test for HIV-associated pulmonary tuberculosis prior to antiretroviral therapy in South Africa. AIDS. 2018;32:1811–1820.
12. Yoon C, Chaisson LH, Patel SM, et al. Diagnostic accuracy of C-reactive protein for active pulmonary tuberculosis: a meta-analysis. Int J Tuberc Lung Dis. 2017;21:1013–1019.
13. Yoon C, Semitala FC, Asege L, et al. Yield and efficiency of novel intensified tuberculosis case-finding algorithms for people living with HIV. Am J Respir Crit Care Med. 2018;199:643–650.
14. Yoon C, Semitala FC, Atuhumuza E, et al. Point-of-care C-reactive protein-based tuberculosis screening for people living with HIV: a diagnostic accuracy study. Lancet Infect Dis. 2017;17:1285–1292.
15. WHO. High-Priority Target Product Profiles for New Tuberculosis Diagnostics: Report of a Consensus Meeting. Geneva, Switzerland: World Health Organization; 2014.
16. Yoon C, Davis JL, Huang L, et al. Point-of-care C-reactive protein testing to facilitate implementation of isoniazid preventive therapy for people living with HIV. J Acquired Immune Defic Syndr. 2014;65:551–556.
17. Bossuyt PM, Reitsma JB, Bruns DE, et al. Stard 2015: an updated list of essential items for reporting diagnostic accuracy studies. BMJ. 2015;351:h5527.
18. Geng EH, Bwana MB, Muyindike W, et al. Failure to initiate antiretroviral therapy, loss to follow-up and mortality among HIV-infected patients during the pre-ART period in Uganda. J Acquired Immune Defic Syndr. 2013;63:e64–71.
19. Lawn SD, Harries AD, Anglaret X, et al. Early mortality among adults accessing antiretroviral treatment programmes in sub-Saharan Africa. AIDS. 2008;22:1897–1908.
20. Koenig SP, Riviere C, Leger P, et al. High mortality among patients with AIDS who received a diagnosis of tuberculosis in the first 3 months of antiretroviral therapy. Clin Infect Dis. 2009;48:829–831.

point-of-care C-reactive protein; tuberculosis; screening algorithm; HIV subgroups

Copyright © 2019 Wolters Kluwer Health, Inc. All rights reserved.