Advertisement

Xpert MTB/RIF on urine samples to increase diagnosis of TB in people living with HIV in Guinea-Bissau

  • Author Footnotes
    6 These authors contributed equally
    Alexander Dutschke
    Correspondence
    Corresponding author: Alexander Dutschke, Department of Infectious Diseases, Aarhus University Hospital, Palle Juul-Jensens Boulevard 99, 8200 Aarhus, Denmark, Telephone: +45 52178273
    Footnotes
    6 These authors contributed equally
    Affiliations
    Bandim Health Project, Indepth Network, Bissau, Guinea-Bissau

    Department of Infectious Diseases, Aarhus University Hospital, Denmark
    Search for articles by this author
  • Author Footnotes
    6 These authors contributed equally
    Ditte Steiniche
    Footnotes
    6 These authors contributed equally
    Affiliations
    Bandim Health Project, Indepth Network, Bissau, Guinea-Bissau

    Department of Infectious Diseases, Aarhus University Hospital, Denmark
    Search for articles by this author
  • Sanne Jespersen
    Affiliations
    Bandim Health Project, Indepth Network, Bissau, Guinea-Bissau

    Department of Infectious Diseases, Aarhus University Hospital, Denmark
    Search for articles by this author
  • Joao Paulo Nanque
    Affiliations
    Bandim Health Project, Indepth Network, Bissau, Guinea-Bissau

    National HIV Programme, Ministry of Health, Bissau, Guinea-Bissau
    Search for articles by this author
  • Candida Medina
    Affiliations
    National HIV Programme, Ministry of Health, Bissau, Guinea-Bissau
    Search for articles by this author
  • Bo Langhoff Hønge
    Affiliations
    Bandim Health Project, Indepth Network, Bissau, Guinea-Bissau

    Department of Infectious Diseases, Aarhus University Hospital, Denmark

    Department of Clinical Immunology, Aarhus University Hospital, Denmark
    Search for articles by this author
  • Christian Wejse
    Affiliations
    Bandim Health Project, Indepth Network, Bissau, Guinea-Bissau

    Department of Infectious Diseases, Aarhus University Hospital, Denmark

    GloHAU, Center for Global Health, School of Public Health, Aarhus University, Denmark
    Search for articles by this author
  • Author Footnotes
    6 These authors contributed equally
Open AccessPublished:March 24, 2022DOI:https://doi.org/10.1016/j.ijid.2022.03.035

      Highlights

      • HIV patients with TB had more advanced disease with mean CD4 of 108
      • A high Bandim TBscore II was more predictive of TB diagnosis than WHO symptoms
      • Adding Gene Xpert testing on urine samples increased diagnostic yield of TB by 58%
      • 36% of HIV patients with a positive TB test were only found on urine Xpert
      • No specific group of HIV patients was identified to primarily offer Xpert on urine

      Abstract

      Objectives

      We investigated if Xpert MTB/RIF (Xpert) testing on urine samples among newly diagnosed HIV-patients as an adjunctive test to Xpert testing on sputum increases diagnosis. We sought to define subgroups of patients, for whom testing with either test is especially advantageous.

      Methods

      We included patients >15 years, newly diagnosed with HIV, that delivered a urine sample on the day of HIV-diagnosis at the biggest HIV-clinic in Guinea-Bissau between September 5, 2016 and October 13, 2017 into a cross-sectional study. Patients were asked for a sputum sample, which was Xpert tested if returned within 30 days. A questionnaire and physical examination were completed on day of inclusion.

      Results

      We included 390 patients. TB prevalence was 12.6%. Adding Xpert urine test to all newly diagnosed HIV-patients increased diagnostic yield of TB by 58% compared with testing on sputum alone. Patients who tested positive by Xpert on urine samples were clinically similar to those tested with sputum, except that the sputum positives reported more cough (p=0.03).

      Conclusions

      Indiscriminate Xpert urine testing in newly diagnosed HIV-patients with advanced disease increased diagnostic yield. Xpert testing for TB on urine and sputum should be offered as screening in Guinea-Bissau and possibly in similar settings.

      Key words

      Background

      Globally an estimated 38 million people were living with HIV in 2019 and sub-Saharan Africa is home to 66.7% of all people living with HIV (PLWH) (

      UNAIDS 2020. Global HIV and AIDS statistics - 2020 fact sheet https://www.unaids.org/sites/default/files/media_asset/UNAIDS_FactSheet_en.pdf. (accessed 01.07.2021).

      ).
      TB is the leading contributor of death in this group accounting for approximately 1/3 of AIDS-related deaths (

      UNAIDS 2019a. Tuberculosis and HIV- Progress towards the 2020 target, https://www.unaids.org/en/resources/documents/2019/tuberculosis-and-hiv-progress-towards-the-2020-target (accessed 20.07.21).

      ), suggesting difficulties in screening and diagnosis in this group.
      According to WHO in 2019 an estimated 10 million people contracted TB, 1.4 million died of TB and out of those 208,000 had HIV (

      WHO 2020. Global Tuberculosis Report 2020, World Health Organization. https://apps.who.int/iris/bitstream/handle/10665/336069/9789240013131-eng.pdf (accessed 15.07.21).

      ).
      The country Guinea-Bissau has both a high HIV and TB burden in PLWH. UN AIDS describes a HIV prevalence among adults of 3% in the country (), but a recent study reported that the HIV prevalence in the capital is 6.7% (
      • Olesen J.S.
      • Jespersen S.
      • da Silva Z.J.
      • Rodrigues A.
      • Erikstrup C.
      • Aaby P.
      • Wejse C.
      • Hønge B.L.
      HIV-2 continues to decrease, whereas HIV-1 is stabilizing in Guinea-Bissau.
      ). TB prevalence in Guinea-Bissau in newly diagnosed PLWH was recently measured to 13.4% but may be higher with intensified focus of diagnosis (
      • Aunsborg J.W.
      • Hønge B.L.
      • Jespersen S.
      • Rudolf F.
      • Medina C.
      • Correira F.G.
      • Johansen I.S.
      • Wejse C.
      A clinical score has utility in tuberculosis case-finding among patients with HIV: A feasibility study from Bissau.
      ).
      Screening is recommended by WHO to be done via a clinical algorithm (

      WHO 2013a. Systematic screening for active tuberculosis, Principles and recommendation, https://www.who.int/tb/publications/Final_TB_Screening_guidelines.pdf, WHO, 2013 (accessed 15.07.21).

      ). If any HIV-patient presents with one of the following: current cough, fever, weight loss or night sweats, they should be evaluated for active TB. This algorithm has a sensitivity of 79% and a specificity of 50% (

      WHO 2010. Guidelines for intensified tuberculosis case-finding and isoniazid preventive therapy for people living with HIV in resource-constrained settings, http://apps.who.int/iris/bitstream/handle/10665/44472/9789241500708_eng.pdf;jsessionid=9A04B22D2BF378C9987F90B1FE7823AB?sequence=1, WHO Global TB Programme, 2010 (accessed 05.07.21).

      ).
      Culture is the gold standard diagnostic test for TB (
      • Achkar J.M.
      • Lawn S.D.
      • Moosa M.-Y.S.
      • Wright C.A.
      • Kasprowicz V.O.
      Adjunctive Tests for Diagnosis of Tuberculosis: Serology, ELISPOT for Site-Specific Lymphocytes, Urinary Lipoarabinomannan, String Test, and Fine Needle Aspiration.
      ), but is not recommended as an initial diagnostic tool due to resource and waiting requirements (2-3 weeks before answer on a positive test (SSI

      SSI2018. SSI, Tuberkulosebekæmpelse i Danmark et nationalt tuberkuloseprogram 2018, TB guideline, 2018, https://www.ssi.dk/-/media/arkiv/indhold/dk-dansk/sygdomsleksikon/tb-guideline_dk_2018.pdf?la=da (accessed 06.08.2021).

      )) and is often difficult in low-income countries (
      • Lawn S.D.
      • Wood R.
      Tuberculosis in antiretroviral treatment services in resource-limited settings: addressing the challenges of screening and diagnosis.
      ), compared with Xpert MTB/RIF (Xpert) and sputum-smear microscopy, which can provide results within a day (

      WHO 2013a. Systematic screening for active tuberculosis, Principles and recommendation, https://www.who.int/tb/publications/Final_TB_Screening_guidelines.pdf, WHO, 2013 (accessed 15.07.21).

      ).
      Xpert test is used in initial TB diagnosis and is accurate when testing sputum diagnosing 90% of TB in culture-positive patients with a specificity of 99% (
      • Boehme C.C.
      • Nicol M.P.
      • Nabeta P.
      • Michael J.S.
      • Gotuzzo E.
      • Tahirli R.
      • Gler M.T.
      • Blakemore R.
      • Worodria W.
      • Gray C.
      • Huang L.
      • Caceres T.
      • Mehdiyev R.
      • Raymond L.
      • Whitelaw A.
      • Sagadevan K.
      • Alexander H.
      • Albert H.
      • Cobelens F.
      • Cox H.
      • Alland D.
      • Perkins M.D.
      Feasibility, diagnostic accuracy, and effectiveness of decentralised use of the Xpert MTB/RIF test for diagnosis of tuberculosis and multidrug resistance: a multicentre implementation study.
      ), however op to one-third of HIV-TB co-infected patients fail to produce sputum samples within 24 hours (
      • Lawn S.D.
      • Kerkhoff A.D.
      • Burton R.
      • Schutz C.
      • Boulle A.
      • Vogt M.
      • Gupta-Wright A.
      • Nicol M.P.
      • Meintjes G.
      Diagnostic accuracy, incremental yield and prognostic value of Determine TB-LAM for routine diagnostic testing for tuberculosis in HIV-infected patients requiring acute hospital admission in South Africa: a prospective cohort.
      ;
      • Lawn S.D.
      • Kerkhoff A.D.
      • Burton R.
      • Schutz C.
      • van Wyk G.
      • Vogt M.
      • Pahlana P.
      • Nicol M.P.
      • Meintjes G.
      ). Rapid microbiological screening for tuberculosis in HIV-positive patients on the first day of acute hospital admission by systematic testing of urine samples using Xpert MTB/RIF: a prospective cohort in South Africa.
      ). In PLWH with a TB positive blood culture, strongly indicating disseminated disease, Xpert sputum and/or sputum smear mircroscopy detect TB in 19.5% of positive cases (
      • Kerkhoff A.D.
      • Barr D.A.
      • Schutz C.
      • Burton R.
      • Nicol M.P.
      • Lawn S.D.
      • Meintjes G.
      Disseminated tuberculosis among hospitalised HIV patients in South Africa: a common condition that can be rapidly diagnosed using urine-based assays.
      ). HIV-TB co-infected patients are more likely to have atypical findings of TB including: sputum smear negative TB, radiographically non-specific TB, (
      • Dawson R.
      • Masuka P.
      • Edwards D.J.
      • Bateman E.D.
      • Bekker L.G.
      • Wood R.
      • Lawn S.D.
      Chest radiograph reading and recording system: evaluation for tuberculosis screening in patients with advanced HIV.
      ;
      • Getahun H.
      • Harrington M.
      • O'Brien R.
      • Nunn P.
      Diagnosis of smear-negative pulmonary tuberculosis in people with HIV infection or AIDS in resource-constrained settings: informing urgent policy changes.
      ) and uncharacteristic/ lack of symptoms (
      • Sterling T.R.
      • Pham P.A.
      • Chaisson R.E.
      HIV infection-related tuberculosis: clinical manifestations and treatment.
      ). They more often present with disseminated TB, with up to 57% having TB in the kidneys, which often remains undiagnosed resulting in rapid and fatal disease (
      • Crump J.A.
      • Ramadhani H.O.
      • Morrissey A.B.
      • Saganda W.
      • Mwako M.S.
      • Yang L.Y.
      • Chow S.C.
      • Njau B.N.
      • Mushi G.S.
      • Maro V.P.
      • Reller L.B.
      • Bartlett J.A.
      Bacteremic disseminated tuberculosis in sub-saharan Africa: a prospective cohort study.
      ;
      • Gupta R.K.
      • Lucas S.B.
      • Fielding K.L.
      • Lawn S.D.
      Prevalence of tuberculosis in post-mortem studies of HIV-infected adults and children in resource-limited settings: a systematic review and meta-analysis.
      ;
      • Lanjewar D.N.
      The spectrum of clinical and pathological manifestations of AIDS in a consecutive series of 236 autopsied cases in mumbai, India.
      ).
      Xpert testing on urine show promising results to increase diagnostic yield (
      • Lawn S.D.
      • Kerkhoff A.D.
      • Burton R.
      • Schutz C.
      • van Wyk G.
      • Vogt M.
      • Pahlana P.
      • Nicol M.P.
      • Meintjes G.
      ). Rapid microbiological screening for tuberculosis in HIV-positive patients on the first day of acute hospital admission by systematic testing of urine samples using Xpert MTB/RIF: a prospective cohort in South Africa.
      ) as an adjunctive test, since most patients can produce the sample and it is quicker to obtain (
      • Lawn S.D.
      • Kerkhoff A.D.
      • Vogt M.
      • Wood R.
      HIV-associated tuberculosis: relationship between disease severity and the sensitivity of new sputum-based and urine-based diagnostic assays.
      ) compared to Xpert sputum. Xpert urine is especially promising when testing samples from HIV-infected patients with advanced immunodeficiency (
      • Lawn S.D.
      • Kerkhoff A.D.
      • Vogt M.
      • Wood R.
      High diagnostic yield of tuberculosis from screening urine samples from HIV-infected patients with advanced immunodeficiency using the Xpert MTB/RIF assay.
      ;
      • Peter J.G.
      • Theron G.
      • Muchinga T.E.
      • Govender U.
      • Dheda K.
      The diagnostic accuracy of urine-based Xpert MTB/RIF in HIV-infected hospitalized patients who are smear-negative or sputum scarce.
      ).

      Objectives

      This study conducted in the capital Bissau of Guinea-Bissau seeks to contribute to the evidence of Xpert testing on urine samples for TB to increase diagnostic yield of TB in PLHIV and its relevance as an adjunctive test with Xpert testing on sputum. Also, it investigates which screening options that can be used to identify subsets of patients for whom urine or sputum tests should be offered exclusively.

      Methods

      Setting, design and study population

      The outpatient HIV clinic at Hospital National Simão Mendes (HNSM) is the largest anti retro-viral therapy (ART) center in Guinea-Bissau. The clinic is home to a part of the Bissau HIV Cohort, a clinical cohort of HIV-infected patients started in 2007 (
      • Jespersen S.
      • Hønge B.L.
      • Oliveira I.
      • Medina C.
      • da Silva Té D.
      • Correira F.G.
      • Erikstrup C.
      • Laursen A.L.
      • Østergaard L.
      • Wejse C.
      Cohort Profile: The Bissau HIV Cohort-a cohort of HIV-1, HIV-2 and co-infected patients.
      ). A high rate of loss-to-follow-up is seen in the cohort (
      • Hønge B.L.
      • Jespersen S.
      • Nordentoft P.B.
      • Medina C.
      • da Silva D.
      • da Silva Z.J.
      • Ostergaard L.
      • Laursen A.L.
      • Wejse C.
      Loss to follow-up occurs at all stages in the diagnostic and follow-up period among HIV-infected patients in Guinea-Bissau: a 7-year retrospective cohort study.
      ), as well as treatment failure (
      • Jespersen S.
      • Hønge B.L.
      • Medina C.
      • da Silva Té D.
      • Correira F.G.
      • Laursen A.L.
      • Erikstrup C.
      • Østergaard L.
      • Wejse C.
      Lack of awareness of treatment failure among HIV-1-infected patients in Guinea-Bissau - a retrospective cohort study.
      ) and subsequently high mortality (
      • Engell-Sørensen T.
      • Rieckmann A.
      • Medina C.
      • da Silva Té D.
      • Rodrigues A.
      • Fisker A.B.
      • Aaby P.
      • Erikstrup C.
      • Jespersen S.
      • Wejse C.
      • Hønge B.L.
      Life expectancy of HIV-infected patients followed at the largest hospital in Guinea-Bissau is one-fourth of life expectancy of the background population.
      ).
      This cross-sectional study was conducted at the HIV clinic at HNSM between September 5, 2016 and October 13, 2017. Newly diagnosed HIV-positive patients >15 years were asked to deliver a urine sample on the day of HIV-diagnosis at HNSM. All patients that did were included in the study and enrolled into the Bissau HIV Cohort. The included patients were the same day instructed how to perform a sputum sample, given a specimen container and asked to return to HNSM with the sample the following day. Patients were told that time between collecting the sputum and delivering it to the clinic should preferably be done within 24 hours. Sputum samples were included if returned to the clinic within 30 days of HIV-diagnosis. A questionnaire regarding demography, lifestyle and symptoms and a physical examination was completed. Blood samples were taken the day after HIV-diagnosis. All patients were included regardless of symptoms.

      HIV-screening

      Screening for HIV was performed using Determine HIV-1/2 assay (Abbott laboratories, Illinois, USA). The confirmatory HIV test also used for HIV type discrimination varied between SD Bioline HIV 1/2 3.0 (Standard Diagnostics Inc, Kyonggi-do, South Korea), First Response HIV Card 1-2.0 (PMC Medical, Mumbai, India) and Genie III HIV-1/HIV-2 (Bio-Rad, Steenvorde, France (
      • Hønge B.L.
      • Bjarnason Obinah M.P.
      • Jespersen S.
      • Medina C.
      • S T.D.
      • da Silva Z.J.
      • Østergaard L.
      • Laursen A.L.
      • Wejse C.
      • Erikstrup C
      Performance of 3 rapid tests for discrimination between HIV-1 and HIV-2 in Guinea-Bissau, West Africa.
      ). CD4 cell count analyses were performed at the National Public Health Laboratory by Partec CyFlow® SL_3 (Cyflow SL, Partec, Munster, Germany).

      Urine and sputum

      All tests were delivered to the National Laboratory of Public Health (LNSP) for analysis on the same day of testing.
      Sputum samples were analyzed continually as part of the laboratory's routine analysis. Tests were taken early in the morning up to 8 o’ clock at HNSM and arrived at the laboratory at 9. They were analyzed using the Xpert MTB/RIF assay shortly after with answers delivered the same day.
      Urine samples were collected and stored at -20C. They were analyzed using the Xpert MTB/RIF assay, when access to Xpert test equipment was possible and when the running of the laboratory allowed it, usually occurring after a month.
      Any TB positives detected were called on telephone and offered treatment. If unavailable treatment was offered on the next follow-up.

      Symptom screening

      We used two scoring systems to evaluate their potential as screening tools based on symptoms for TB in PLWH.
      Bandim TBscoreII (
      • Rudolf F.
      • Lemvik G.
      • Abate E.
      • Verkuilen J.
      • Schön T.
      • Gomes V.F.
      • Eugen-Olsen J.
      • Østergaard L.
      • Wejse C.
      TBscore II: refining and validating a simple clinical score for treatment monitoring of patients with pulmonary tuberculosis.
      ) consists of 3 patient reported symptoms: cough, dyspnea, chest pain and 3 signs evaluated by a clinician: anemia (pale inferior conjuctivae), low body mass index (BMI) or low mid upper arm circumference (MUAC). Each positive finding gives 1 point, with the exception of BMI and MUAC where values of <16 mg^m2 and <200 mm give 2 points.
      Severity classes for TBscoreII: I, score < 2, II, score 2–3; III, score 4–7; and IV, score > 7 (
      • Rudolf F.
      • Lemvik G.
      • Abate E.
      • Verkuilen J.
      • Schön T.
      • Gomes V.F.
      • Eugen-Olsen J.
      • Østergaard L.
      • Wejse C.
      TBscore II: refining and validating a simple clinical score for treatment monitoring of patients with pulmonary tuberculosis.
      ).
      WHO symptom screening evaluates patients on the presence of one or more of the following symptoms: cough, self-reported fever, weight-loss or night sweats. Presence of ≥1 symptoms result in further testing for TB.

      Statistical analysis

      Data was analyzed using Stata/IC 13.0 (StataCorp LP, College Station, TX, USA). Chi-square test was used to assess categorical valuables and in cases with <5 observations in a group Fiscers Exact test. Continuous variables were compared using ttest and in cases where assumptions failed the non-parametric Wilcoxon rank-sum test. Wilcoxon-Mann-Whitney test was used to compare Xpert testing of urine samples and sputum samples. Missing data was excluded from analysis.

      Ethical considerations

      Written information was given in the official language Portuguese and oral information in the widely spoken Portuguese Creole. Informed written consent, or a fingerprint if illiterate, was provided by all included patients.

      Results

      In total 390 newly HIV-diagnosed patients were included between September 5, 2016 and October 13, 2017. Within 30 days from inclusion, 203 (52.1 %) patients returned with a sputum sample. Median age was 37 years (IQR, 30 to 45) and 33,6% were male. Among the 390 patients included, 260 (66.7%) were HIV-1 positive, 28 (7.2%) HIV-2 positive, 24 (6.2%) HIV 1/2 dually-infeted and 78 (20%) had unknown HIV-types. Median CD4 cell count was 211 cells/uL (IQR=97:423). Median hemoglobin was 9.3 g/dL (IQR=7.9:10.7). The median Bandim TBscoreII was 1 (IQR, 0 to 3) and 345 (96.4%, n=353) showed ≥1 WHO symptom. Twenty-six patients were previous to the study tested positive for TB. Of these six still registered positive, five only on Xpert sputum and one on sputum and urine. Date of previous infection was unknown.
      49 (12.6%) of the included patients tested positive for TB either by sputum or urine sample (Table 1). The positive rate was 10.3% (CI 4.7%-16%) higher for the sputum test versus the urine test, when comparing the two. Five patients who tested negative on the sputum test had a positive urine test and 13 patients who did not return with a sputum sample tested positive for TB in the urine sample. In total 18 patients which equals 36.7% of the total number of TB positive patients had not been found if they had not been asked to deliver a urine sample. This is an increase in diagnostic yield of 49/31 = 58%. Clinically the urine positives reported 100% fever, 94% weight loss, 70.6% night sweat and 35.3% cough (n=17).
      Table 1Number of patients with TB positive and TB negative urine and sputum results.
      Sputum not availableSputum negativeSputum positive
      Urine negative17416726367
      Urine positive135523
      18717231390
      Comparison of TB positives vs TB negatives (Table 2).
      Table 2Characteristics of TB negative and TB positive patients.
      TB negativeTB positive
      n = 341, n (%)n = 49, n (%)p-value
      Age (years) n=390Mean=38.9Mean=35.30.04
      Sex n=3900.41
       Women229 (67.2)30 (61.2)
       Men112 (32.8)19 (38.8)
      HIV type n=3120.29
       HIV-1226 (82.2)34 (91.9)
       HIV-227 (9.8)1 (2.7)
       HIV-1/222 (8)2 (5.4)
      WHO TB symptoms reported
       Cough n=37095 (29.5)27 (56.3)<0.01
       Fever n=373250 (76.9)46 (95.8)<0.01
       Night sweats n=368162 (50.5)32 (68.1)0.02
       Weight loss n=369284 (88.5)45 (93.8)0.27
      Bandim TBscoreII
       Cough n=37095 (29.5)27 (56.3)<0.01
       Dyspnea n=37377 (23.6)17 (36.2)0.06
       Chest pain n=374143 (43.9)26 (54.2)0.18
       Anemia n=34130 (10.1)6 (13.3)0.52
       BMI < 18 kg/m^2 n=35449 (15.8)16 (37.2)<0.01
       BMI <16 kg/m^2 n=35438 (12.2)5 (11.6)0.91
       MUAC < 220 mm N=36944 (13.7)10 (21.3)0.17
       MUAC < 200 mm n=36924 (7.5)4 (8.5)0.79
       Median total Bandim TBscoreII n=3151 (IQR=0:3)3 (IQR=1:4)<0.01
      Bandim TBscoreII >1 n=315126 (45.7)25 (64.1)0.03
      Haemoglobin n=222<0.01
       Mean9.67 (95% CI: 9.28:10.06)7.72 (95% CI: 6.85:8.59)
       Mean difference1.95 (95% CI: 0.91:2.99)
      CD4 n=353<0.01
       Median232 (IQR=109.5:459)108 (IQR=30:208)
      The TB positive patients were significantly younger (mean=35.3 years) than TB negative patients (mean=38.9, p=0.04). TB positive patients suffered more from cough (p<0.01), fever (p<0.01) and night sweats (p=0.02) compared with TB negatives. TB positive patients had a significantly higher Bandim TBscoreII (median=3, p=0.01) than TB negative patients (median=1). CD4 cell count and hemoglobin were significantly lower for TB positive patients compared with TB negative (p<0.01 for both)
      Comparison of patients that solely delivered a urine sample vs patients that also delivered sputum samples (Table 3).
      Table 3Characteristics of patients who solely delivered a urine sample vs. patients who delivered both a urine and sputum sample.
      Solely urine sampleUrine and sputum sample
      n (%), n = 187n (%), n = 203p-value
      Age (years) n=390Mean=38.9Mean=380.43
      Sex n=390
       Women118 (63.1)141 (69.5)0.18
       Men69 (36.9)62 (30.5)
      HIV type n=312
       HIV-1120 (83.9)140 (82.8)0.65
       HIV-214 (9.8)14 (8.3)
       HIV-1/29 (6.3)15 (8.9)
      WHO TB symptoms reported
       Cough n=37047 (27.3)75 (37.9)0.03
       Fever n=373131 (75.3)165 (82.9)0.07
       Night sweats n=36890 (52.3)104 (53.1)0.89
       Weight loss n=369148 (86.1)181 (91.9)0.07
      Bandim TB-II score
       Cough n=37047 (27.3)75 (37.9)0.03
       Dyspnea n=37349 (24.6)45 (25.9)0.78
       Chest pain n=37470 (40.2)99 (49.5)0.07
       Anemia n=34119 (11.8)17 (9.4)0.48
       BMI < 18 kg/m^2 n=35422 (13.3)43 (22.9)0.02
       BMI <16 kg/m^2 n=35425 (15.1)18 (9.6)0.12
       MUAC < 220 mm n=36930 (17.5)24 (12.1)0.14
       MUAC < 200 mm n=36913 (7.6)15 (7.6)0.99
       Median total Bandim TB-II score n=3151 (IQR=0:3)1 (IQR=0:3)0.49
      Haemoglobin n=222<0.05
       Mean9.87 (95% CI: 9.17:10.58)9.11 (95% CI: 8.71:9.51)
       Mean difference-0.77 (95%CI: -1.51:0.02)
      CD4 cell count n=3530.01
       Median253 (IQR=109:493)194 (IQR=83:368)
      These patients were comparable on most variables. Patients that also delivered a sputum sample had lower hemoglobin (p<0.05), CD4 level (p=0.01), more had lower BMI<18 kg/m^2 (p=0.02), more coughed (p=0.03) and they tended to suffer more from fever (p=0.07) and weight loss (p=0.07) compared with patients who only delivered a urine sample. There were no differences between the groups regarding Bandim TBscoreII (median 1, p=0.49).
      Comparison of patients with a positive urine sample vs sputum samples.
      Twenty-six tested positive only with Xpert on sputum and 18 only with Xpert on urine. The 5 positive patients found by both tests were excluded from this analysis. To further assess if urine Xpert should be offered only to specific groups of HIV-infected we tested if patients diagnosed only with a urine-sample were different with regards to any variables compared to those diagnosed via sputum only. The sputum positives experienced more cough (n=43, p=0.03) and more had a BMI<18 mg/m^2 (n=39, p=0.04), otherwise no differences were found. Both groups suffered from severe immunodeficiency with median CD4=138 (IQR=28:248) in sputum positives and 108 (IQR=45:187) in urine positives, with >75% having AIDS defining CD4 levels among urine positives.

      Discordant pairs

      In 26 cases the sputum results were TB positive and urine negative. In 5 cases the urine results were positive and sputum negative. Analysis showed a tendency of higher TBscoreII (n=27, p=0.09), cough (n=31, p=0.06) and BMI<18kg/m^2 (n=30, p=0.052) in the 26 sputum test positive, urine test negative patients.

      Discussion

      We found adding an Xpert test on urine samples on all newly diagnosed HIV-infected patients increased diagnosis of TB by 36.7%. Overall TB prevalence in the study was found to be 12.6% with Xpert testing on urine and 7.9% without, an increase of 59%. Inability to produce sputum and difficulty of follow-up on patients for sputum collection resulted in a large group of 48% of patients that remained sputum untested within <30 days of HIV diagnosis. Of the increase in TB diagnosis by Xpert urine testing 26.5% came from this group. Urine samples are compared with sputum samples easier and quicker to obtain, handle and quality of samples is less of an issue (
      • Lawn S.D.
      • Kerkhoff A.D.
      • Vogt M.
      • Wood R.
      HIV-associated tuberculosis: relationship between disease severity and the sensitivity of new sputum-based and urine-based diagnostic assays.
      ), which is advantageous in a setting where follow-up on patients for sputum collection is limited.
      TB positive and negative patients differed on many variables, but not in ways that allowed us to identify specific subsets to test with Xpert on urine samples. Self-reported cough had the highest difference of 26.8% between TB positives and negatives, but only 56.3% of positives reported cough, so using this as the defining screening tool would lead to underdiagnosis of TB. Using Bandim TBscoreII to screen and identify candidates for Xpert testing on urin is suboptimal in this study since 17.9% of TB positive patients scored 0 and 35.9% scored <2. These TB positive patients would not be found if the screening algorithm discarded patients for further diagnostic tests with low scores. 96.4% of all patients reported ≥1 WHO symptom, so using this as a screening algorithm hardly limits the scope of testing.
      Another way of identifying a subset of patients for whom testing with Xpert urine test is particularly advantageous is to identify those who lack the ability to produce sputum and offer them the urine test instead. Patients that only delivered urine samples were compared with patients that were also able to produce a sputum sample. Overall the groups were very identical. Identification of patients unable to produce sputum cannot help limit scope of Xpert testing on urine in this study.
      We tested if positive patients found only by Xpert urine were different from positive patients found only by Xpert sputum on variables such as CD4, hemoglobin and symptoms of TB. We found no major differences in clinical presentation or biomarkers between the two groups, hence offering different tests based on level of immunodeficiency is not warranted. Both groups were severely immunocompromised, and it is possible that the conclusion could be different in patients with higher CD4 counts, since urine-based tests are more sensitive at low CD4 counts (
      • Bjerrum S.
      • Schiller I.
      • Dendukuri N.
      • Kohli M.
      • Nathavitharana R.R.
      • Zwerling A.A.
      • Denkinger C.M.
      • Steingart K.R.
      • Shah M.
      Lateral flow urine lipoarabinomannan assay for detecting active tuberculosis in people living with HIV.
      ;
      • Lawn S.D.
      • Kerkhoff A.D.
      • Vogt M.
      • Wood R.
      High diagnostic yield of tuberculosis from screening urine samples from HIV-infected patients with advanced immunodeficiency using the Xpert MTB/RIF assay.
      ).
      Discordant pair analysis showed us that Xpert test on sputum samples tended to be better at diagnosing those with more cough (indicative of pulmonal TB).
      Strengths of this study is the recruitment of well-characterized patients in a well-documented cohort (
      • Jespersen S.
      • Hønge B.L.
      • Oliveira I.
      • Medina C.
      • da Silva Té D.
      • Correira F.G.
      • Erikstrup C.
      • Laursen A.L.
      • Østergaard L.
      • Wejse C.
      Cohort Profile: The Bissau HIV Cohort-a cohort of HIV-1, HIV-2 and co-infected patients.
      ). The patients included in the present study are similar on parameters like sex, age at inclusion and median CD4 count compared with the Bissau HIV Cohort and results are likely applicable to the background population. Weaknesses of the study include a lack of a gold-standard of testing, therefore the actual prevalence of TB in the study-population is unknown and consequently sensitivity and specificity analysis are impossible.
      The study population was not an unselected sample of individuals with HIV. Only those who delivered a urine test were included. However, urine tests are easy to obtain from patients (
      • Lawn S.D.
      • Kerkhoff A.D.
      • Burton R.
      • Schutz C.
      • Boulle A.
      • Vogt M.
      • Gupta-Wright A.
      • Nicol M.P.
      • Meintjes G.
      Diagnostic accuracy, incremental yield and prognostic value of Determine TB-LAM for routine diagnostic testing for tuberculosis in HIV-infected patients requiring acute hospital admission in South Africa: a prospective cohort.
      ) and it is unlikely that a significant number of patients were discarded. The study was conducted at a single site in Guinea-Bissau and the extent of generalizability to other similar settings outside of Guinea-Bissau is uncertain, however this study builds upon previous studies that have also demonstrated increased diagnostic yield of Xpert urine testing in PLWH for TB (
      • Lawn S.D.
      • Kerkhoff A.D.
      • Burton R.
      • Schutz C.
      • van Wyk G.
      • Vogt M.
      • Pahlana P.
      • Nicol M.P.
      • Meintjes G.
      ). Rapid microbiological screening for tuberculosis in HIV-positive patients on the first day of acute hospital admission by systematic testing of urine samples using Xpert MTB/RIF: a prospective cohort in South Africa.
      ;
      • Schutz C.
      • Ward A.
      • Burton R.
      • Nicol M.P.
      • Blumenthal L.
      • Meintjes G.
      • Kerkhoff A.D.
      False rifampicin resistant results using Xpert MTB/RIF on urine samples in hospitalised HIV-infected patients.
      ), especially in the severely immunocompromised (
      • Lawn S.D.
      • Kerkhoff A.D.
      • Vogt M.
      • Wood R.
      High diagnostic yield of tuberculosis from screening urine samples from HIV-infected patients with advanced immunodeficiency using the Xpert MTB/RIF assay.
      ;
      • Peter J.G.
      • Theron G.
      • Muchinga T.E.
      • Govender U.
      • Dheda K.
      The diagnostic accuracy of urine-based Xpert MTB/RIF in HIV-infected hospitalized patients who are smear-negative or sputum scarce.
      ). Concentration of urine was seen to significantly increase diagnosis of TB compared with unconcentrated urine (
      • Lawn S.D.
      • Kerkhoff A.D.
      • Burton R.
      • Schutz C.
      • van Wyk G.
      • Vogt M.
      • Pahlana P.
      • Nicol M.P.
      • Meintjes G.
      ). Rapid microbiological screening for tuberculosis in HIV-positive patients on the first day of acute hospital admission by systematic testing of urine samples using Xpert MTB/RIF: a prospective cohort in South Africa.
      ;
      • Peter J.G.
      • Theron G.
      • Muchinga T.E.
      • Govender U.
      • Dheda K.
      The diagnostic accuracy of urine-based Xpert MTB/RIF in HIV-infected hospitalized patients who are smear-negative or sputum scarce.
      ), but this was not possible in our setting due to centrifuge requirements.

      Conclusion

      We found that Xpert testing on urine increased diagnostic yield if given as an adjunctive test to Xpert testing on sputum and we recommend screening all newly diagnosed HIV positive patients in Guinea-Bissau for TB with both Xpert on sputum and urine. Our study found no relevant screening parameters to limit Xpert sputum and/or urine testing for patient subsets. These results may be applicable in similar low-resource settings, where testing capabilities and follow-up on patients are similarly challenged and we recommend further research into this to better diagnosis and outcome.

      Transparency declaration

      This article is part of a supplement entitled Commemorating World Tuberculosis Day March 24th, 2022: “Invest to End TB. Save Lives” published with support from an unrestricted educational grant from QIAGEN Sciences Inc.

      Declaration of interests

      The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

      Authors contributions

      AD, DS, SJ, BLH, JP, CM and CW conceived the study; DS, JP and CM carried out data collection; AD, DS and BLH carried out analysis and interpretation of data. AD and DS drafted the manuscript; all authors critically revised the manuscript for intellectual content. All authors read and approved the final manuscript of the paper. BLH, SJ and CW are guarantors of the paper.

      Acknowledgements

      The authors are grateful to the staff working in the HIV clinic at HNSM and the National Public Health Laboratory for their work effort and their professionalism. We are furthermore grateful for the staff at the Bandim Health Project, for making this study possible.
      The Bissau HIV cohort study group consists of Amabelia Rodrigues, David da Silva, Zacarias da Silva, Candida Medina, Ines Oliviera-Souto, Lars Østergaard, Alex Laursen, Peter Aaby, Anders Fomsgaard, Christian Erikstrup, Christian Wejse, Bo Langhoff Hønge, and Sanne Jespersen (chair).

      Funding

      This study was funded by Aarhus Universities research fund for research at the Institute for Clinical Medicine, Aarhus University Hospital. 120.000 dkk for principal researchers’ salary.

      References

        • Achkar J.M.
        • Lawn S.D.
        • Moosa M.-Y.S.
        • Wright C.A.
        • Kasprowicz V.O.
        Adjunctive Tests for Diagnosis of Tuberculosis: Serology, ELISPOT for Site-Specific Lymphocytes, Urinary Lipoarabinomannan, String Test, and Fine Needle Aspiration.
        The Journal of Infectious Diseases. 2011; 204: S1130-S1141https://doi.org/10.1093/infdis/jir450
        • Aunsborg J.W.
        • Hønge B.L.
        • Jespersen S.
        • Rudolf F.
        • Medina C.
        • Correira F.G.
        • Johansen I.S.
        • Wejse C.
        A clinical score has utility in tuberculosis case-finding among patients with HIV: A feasibility study from Bissau.
        Int J Infect Dis. 2020; 92s: S78-s84https://doi.org/10.1016/j.ijid.2020.03.012
        • Bjerrum S.
        • Schiller I.
        • Dendukuri N.
        • Kohli M.
        • Nathavitharana R.R.
        • Zwerling A.A.
        • Denkinger C.M.
        • Steingart K.R.
        • Shah M.
        Lateral flow urine lipoarabinomannan assay for detecting active tuberculosis in people living with HIV.
        Cochrane Database Syst Rev. 2019; 10Cd011420https://doi.org/10.1002/14651858.CD011420.pub3
        • Boehme C.C.
        • Nicol M.P.
        • Nabeta P.
        • Michael J.S.
        • Gotuzzo E.
        • Tahirli R.
        • Gler M.T.
        • Blakemore R.
        • Worodria W.
        • Gray C.
        • Huang L.
        • Caceres T.
        • Mehdiyev R.
        • Raymond L.
        • Whitelaw A.
        • Sagadevan K.
        • Alexander H.
        • Albert H.
        • Cobelens F.
        • Cox H.
        • Alland D.
        • Perkins M.D.
        Feasibility, diagnostic accuracy, and effectiveness of decentralised use of the Xpert MTB/RIF test for diagnosis of tuberculosis and multidrug resistance: a multicentre implementation study.
        Lancet. 2011; 377: 1495-1505https://doi.org/10.1016/s0140-6736(11)60438-8
        • Crump J.A.
        • Ramadhani H.O.
        • Morrissey A.B.
        • Saganda W.
        • Mwako M.S.
        • Yang L.Y.
        • Chow S.C.
        • Njau B.N.
        • Mushi G.S.
        • Maro V.P.
        • Reller L.B.
        • Bartlett J.A.
        Bacteremic disseminated tuberculosis in sub-saharan Africa: a prospective cohort study.
        Clin Infect Dis. 2012; 55: 242-250https://doi.org/10.1093/cid/cis409
        • Dawson R.
        • Masuka P.
        • Edwards D.J.
        • Bateman E.D.
        • Bekker L.G.
        • Wood R.
        • Lawn S.D.
        Chest radiograph reading and recording system: evaluation for tuberculosis screening in patients with advanced HIV.
        Int J Tuberc Lung Dis. 2010; 14: 52-58
        • Engell-Sørensen T.
        • Rieckmann A.
        • Medina C.
        • da Silva Té D.
        • Rodrigues A.
        • Fisker A.B.
        • Aaby P.
        • Erikstrup C.
        • Jespersen S.
        • Wejse C.
        • Hønge B.L.
        Life expectancy of HIV-infected patients followed at the largest hospital in Guinea-Bissau is one-fourth of life expectancy of the background population.
        Infection. 2021; 49: 631-643https://doi.org/10.1007/s15010-020-01574-6
        • Getahun H.
        • Harrington M.
        • O'Brien R.
        • Nunn P.
        Diagnosis of smear-negative pulmonary tuberculosis in people with HIV infection or AIDS in resource-constrained settings: informing urgent policy changes.
        Lancet. 2007; 369: 2042-2049https://doi.org/10.1016/s0140-6736(07)60284-0
        • Gupta R.K.
        • Lucas S.B.
        • Fielding K.L.
        • Lawn S.D.
        Prevalence of tuberculosis in post-mortem studies of HIV-infected adults and children in resource-limited settings: a systematic review and meta-analysis.
        Aids. 2015; 29: 1987-2002https://doi.org/10.1097/qad.0000000000000802
        • Hønge B.L.
        • Bjarnason Obinah M.P.
        • Jespersen S.
        • Medina C.
        • S T.D.
        • da Silva Z.J.
        • Østergaard L.
        • Laursen A.L.
        • Wejse C.
        • Erikstrup C
        Performance of 3 rapid tests for discrimination between HIV-1 and HIV-2 in Guinea-Bissau, West Africa.
        J Acquir Immune Defic Syndr. 2014; 65: 87-90https://doi.org/10.1097/QAI.0b013e3182a97b81
        • Hønge B.L.
        • Jespersen S.
        • Nordentoft P.B.
        • Medina C.
        • da Silva D.
        • da Silva Z.J.
        • Ostergaard L.
        • Laursen A.L.
        • Wejse C.
        Loss to follow-up occurs at all stages in the diagnostic and follow-up period among HIV-infected patients in Guinea-Bissau: a 7-year retrospective cohort study.
        BMJ Open. 2013; 3e003499https://doi.org/10.1136/bmjopen-2013-003499
        • Jespersen S.
        • Hønge B.L.
        • Medina C.
        • da Silva Té D.
        • Correira F.G.
        • Laursen A.L.
        • Erikstrup C.
        • Østergaard L.
        • Wejse C.
        Lack of awareness of treatment failure among HIV-1-infected patients in Guinea-Bissau - a retrospective cohort study.
        J Int AIDS Soc. 2015; 18: 20243https://doi.org/10.7448/ias.18.1.20243
        • Jespersen S.
        • Hønge B.L.
        • Oliveira I.
        • Medina C.
        • da Silva Té D.
        • Correira F.G.
        • Erikstrup C.
        • Laursen A.L.
        • Østergaard L.
        • Wejse C.
        Cohort Profile: The Bissau HIV Cohort-a cohort of HIV-1, HIV-2 and co-infected patients.
        Int J Epidemiol. 2015; 44: 756-763https://doi.org/10.1093/ije/dyu201
        • Kerkhoff A.D.
        • Barr D.A.
        • Schutz C.
        • Burton R.
        • Nicol M.P.
        • Lawn S.D.
        • Meintjes G.
        Disseminated tuberculosis among hospitalised HIV patients in South Africa: a common condition that can be rapidly diagnosed using urine-based assays.
        Sci Rep. 2017; 7: 10931https://doi.org/10.1038/s41598-017-09895-7
        • Lanjewar D.N.
        The spectrum of clinical and pathological manifestations of AIDS in a consecutive series of 236 autopsied cases in mumbai, India.
        Patholog Res Int. 2011; 2011547618https://doi.org/10.4061/2011/547618
        • Lawn S.D.
        • Kerkhoff A.D.
        • Burton R.
        • Schutz C.
        • Boulle A.
        • Vogt M.
        • Gupta-Wright A.
        • Nicol M.P.
        • Meintjes G.
        Diagnostic accuracy, incremental yield and prognostic value of Determine TB-LAM for routine diagnostic testing for tuberculosis in HIV-infected patients requiring acute hospital admission in South Africa: a prospective cohort.
        BMC Med. 2017; 15: 67https://doi.org/10.1186/s12916-017-0822-8
        • Lawn S.D.
        • Kerkhoff A.D.
        • Burton R.
        • Schutz C.
        • van Wyk G.
        • Vogt M.
        • Pahlana P.
        • Nicol M.P.
        • Meintjes G.
        ). Rapid microbiological screening for tuberculosis in HIV-positive patients on the first day of acute hospital admission by systematic testing of urine samples using Xpert MTB/RIF: a prospective cohort in South Africa.
        BMC Medicine. 2015; 13: 192https://doi.org/10.1186/s12916-015-0432-2
        • Lawn S.D.
        • Kerkhoff A.D.
        • Vogt M.
        • Wood R.
        High diagnostic yield of tuberculosis from screening urine samples from HIV-infected patients with advanced immunodeficiency using the Xpert MTB/RIF assay.
        J Acquir Immune Defic Syndr. 2012; 60: 289-294https://doi.org/10.1097/QAI.0b013e318258c6af
        • Lawn S.D.
        • Kerkhoff A.D.
        • Vogt M.
        • Wood R.
        HIV-associated tuberculosis: relationship between disease severity and the sensitivity of new sputum-based and urine-based diagnostic assays.
        BMC Medicine. 2013; 11: 231https://doi.org/10.1186/1741-7015-11-231
        • Lawn S.D.
        • Wood R.
        Tuberculosis in antiretroviral treatment services in resource-limited settings: addressing the challenges of screening and diagnosis.
        J Infect Dis. 2011; 204 (Suppl 4): S1159-S1167https://doi.org/10.1093/infdis/jir411
        • Olesen J.S.
        • Jespersen S.
        • da Silva Z.J.
        • Rodrigues A.
        • Erikstrup C.
        • Aaby P.
        • Wejse C.
        • Hønge B.L.
        HIV-2 continues to decrease, whereas HIV-1 is stabilizing in Guinea-Bissau.
        Aids. 2018; 32: 1193-1198https://doi.org/10.1097/qad.0000000000001827
        • Peter J.G.
        • Theron G.
        • Muchinga T.E.
        • Govender U.
        • Dheda K.
        The diagnostic accuracy of urine-based Xpert MTB/RIF in HIV-infected hospitalized patients who are smear-negative or sputum scarce.
        PLoS One. 2012; 7: e39966https://doi.org/10.1371/journal.pone.0039966
        • Rudolf F.
        • Lemvik G.
        • Abate E.
        • Verkuilen J.
        • Schön T.
        • Gomes V.F.
        • Eugen-Olsen J.
        • Østergaard L.
        • Wejse C.
        TBscore II: refining and validating a simple clinical score for treatment monitoring of patients with pulmonary tuberculosis.
        Scand J Infect Dis. 2013; 45: 825-836https://doi.org/10.3109/00365548.2013.826876
        • Schutz C.
        • Ward A.
        • Burton R.
        • Nicol M.P.
        • Blumenthal L.
        • Meintjes G.
        • Kerkhoff A.D.
        False rifampicin resistant results using Xpert MTB/RIF on urine samples in hospitalised HIV-infected patients.
        South Afr J HIV Med. 2019; 20: 978https://doi.org/10.4102/sajhivmed.v20i1.978
        • Sterling T.R.
        • Pham P.A.
        • Chaisson R.E.
        HIV infection-related tuberculosis: clinical manifestations and treatment.
        Clin Infect Dis. 2010; 50: S223-S230https://doi.org/10.1086/651495
      1. SSI2018. SSI, Tuberkulosebekæmpelse i Danmark et nationalt tuberkuloseprogram 2018, TB guideline, 2018, https://www.ssi.dk/-/media/arkiv/indhold/dk-dansk/sygdomsleksikon/tb-guideline_dk_2018.pdf?la=da (accessed 06.08.2021).

      2. UNAIDS 2019a. Tuberculosis and HIV- Progress towards the 2020 target, https://www.unaids.org/en/resources/documents/2019/tuberculosis-and-hiv-progress-towards-the-2020-target (accessed 20.07.21).

      3. UNAIDS 2019b. Guinea-Bissau, https://www.unaids.org/en/regionscountries/countries/guinea-bissau (accessed 15.07.2021)

      4. UNAIDS 2020. Global HIV and AIDS statistics - 2020 fact sheet https://www.unaids.org/sites/default/files/media_asset/UNAIDS_FactSheet_en.pdf. (accessed 01.07.2021).

      5. WHO 2010. Guidelines for intensified tuberculosis case-finding and isoniazid preventive therapy for people living with HIV in resource-constrained settings, http://apps.who.int/iris/bitstream/handle/10665/44472/9789241500708_eng.pdf;jsessionid=9A04B22D2BF378C9987F90B1FE7823AB?sequence=1, WHO Global TB Programme, 2010 (accessed 05.07.21).

      6. WHO 2013a. Systematic screening for active tuberculosis, Principles and recommendation, https://www.who.int/tb/publications/Final_TB_Screening_guidelines.pdf, WHO, 2013 (accessed 15.07.21).

      7. WHO 2020. Global Tuberculosis Report 2020, World Health Organization. https://apps.who.int/iris/bitstream/handle/10665/336069/9789240013131-eng.pdf (accessed 15.07.21).