Background: XpertMTB/RIF

Transcription

1 Xpert MTB/RIF vs microscopy as the first line TB test in South Africa: mortality, yield,initial loss to follow up and proportion treated. The XTEND study GJ Churchyard On behalf of the XTEND team (Xpert for TB - Evaluating a New Diagnostic) Prof Churchyard has no financial relationships with commercial entities to disclose

2 Background: XpertMTB/RIF Improved diagnosis of TB is a global priority for TB control Xpertis a rapid, molecular, cartridge-based TB test with greater sensitivity than sputum smear microscopy an immediate rifampicin resistance result WHO conditional recommendation (203): XpertMTB/RIF recommended initial diagnostic test for adults investigated for TB Impact of XpertMTB/RIF will depend on the health system in which it is used Evidence concerning patient, program and costeffectiveness outcomes are required to inform policy

3 Background: South Africa (SA) SA has the third largest number of TB cases globally Since 20, SA replaced sputum smear microscopy with XpertMTB/RIF as the first-line test for TB The SA XpertMTB/RIF program is the largest in the world accounts for >50% of all cartridges procured globally A pragmatic, cluster-randomized trial was embedded within the XpertMTB/RIF national roll out to evaluate patient and programme relevant outcomes

4 Background: South Africa (SA) SA has the third largest number of TB cases globally Since 20, SA replaced sputum smear microscopy with XpertMTB/RIF as the first-line test for TB The SA XpertMTB/RIF program is the largest in the world accounts for >50% of all cartridges procured globally A pragmatic, cluster-randomized trial was embedded within the XpertMTB/RIF national roll out to evaluate patient and programme relevant outcomes

5 XTEND trial: primary objective To evaluate the effect of XpertMTB/RIF implementation on mortality among adults investigated for TB

6 XTEND trial: secondary objectives To compare, in the XTEND study, by study arm: Among clinic attendees investigated for TB Proportion index test positive Proportion starting TB treatment by 6 months Amongthose with apositive index test result Proportion initially lost to follow-up

7 Study clusters Cluster defined as a lab and 2 primary care clinics (PHC) served by, but not co-located with, that lab 20 laboratories in medium-burden districts in 4 provinces were identified 2 PHCs served by each lab were selected Xpert arm (0 clusters) Microscopy arm (0 clusters) laboratory laboratory Clinic Clinic 2 Clinic Clinic 2

8 XpertMTB/RIF implementation in participating laboratories GX 6-module instruments were introduced into Xpertarm laboratories Laboratory staff received training on XpertMTB/RIF use from the NHLS Microscopy arm laboratories continued using smear microscopy during enrolment of participants, implementing Xpert MTB/RIF ~6 months later Participant enrolment in the Xpertarm was initiated a median of3.6 (range -7.4) months after starting routine use of GeneXpert

9 Cohort of persons investigated for TB Clinic clients TB symptoms Sputum requested Cohort of persons investigated for TB Contacted by phone or home visit Enrolled week Month Month 2 Month 4 6 month review of participants or nominated contacts Alive/Dead Treated for TB: Yes/No

10 Cohort of persons investigated for TB Cohort of persons investigated for TB 6 month review of participants or nominated contacts Alive/Dead Contacted by phone or home visit Clinic clients TB symptoms Sputum requested Enrolled week Month Month 2 Month 4 Treated for TB: Yes/No Persons investigated for TB by clinic staff Xpertarm: sputum specimen collected for Xpert MTB/RIF microscopy arm:2 sputum specimens collected for fluorescence microscopy CXR/ sputum for culture if requested by clinic staff, guided by relevant algorithm

11 Cohort of persons investigated for TB Clinic clients Cohort of persons investigated for TB Contacted by phone or home visit TB symptoms Sputum requested Enrolled week Month Month 2 Month 4 6 month review of participants or nominated contacts A systematic sample of adults, not on TB treatment, who had had a sputum specimen requested by clinic staff to investigate for possible TB were recruited by study staff Alive/Dead Treated for TB: Yes/No

12 Cohort of persons investigated for TB Cohort of persons investigated for TB Contacted by phone or home visit Clinic clients TB symptoms Sputum requested Enrolled week Month Month 2 Month 4 6 month review of participants or nominated contacts At enrolment, personal identifiers, including SA identity number, locator information, demographic and clinical data relevant to TB and mortality risk were collected No extra tests for study purposes Alive/Dead Treated for TB: Yes/No

13 Cohort of persons investigated for TB Clinic clients Cohort of persons investigated for TB Contacted by phone or home visit TB symptoms Sputum requested Enrolled week Month Month 2 Month 4 6 month review of participants or nominated contacts Participants were contacted telephonically by study staff in order to maintain contact and update locator information Alive/Dead Treated for TB: Yes/No

14 Cohort of persons investigated for TB Clinic clients Cohort of persons investigated for TB Contacted by phone or home visit TB symptoms Sputum requested Enrolled week Month Month 2 Month 4 6 month review of participants or nominated contacts Prior to the 6-month interview, study staff reviewed patients clinic records for information such as results of index and subsequent tests and TB treatment and/or ART start dates Alive/Dead Treated for TB: Yes/No

15 Cohort of persons investigated for TB Cohort of persons investigated for TB Contacted by phone or home visit Clinic clients TB symptoms Sputum requested Enrolled week Month Month 2 Month 4 6 month review of participants or nominated contacts 6 months after enrolment, study staff interviewed participants telephonically, or by home visit, to ascertain whether they had started TB treatment or ART and treatment start dates Alive/Dead Treated for TB: Yes/No

16 Primary outcome Deaths were ascertained through Reports from participant-nominated contacts Clinic staff National vital statistics database using SA ID numbers

17 Secondaryoutcomes Among clinic attendees investigated for TB Proportion index test positive Index specimen sputum taken at enrolment Result obtained from National Health Laboratory Service or record review Proportion starting TB treatment within 6 months Based on participant self-report and/or record review

18 Secondaryoutcomes Among those index test positive Initial loss to follow-updefined as notstarting TB treatment within 28 days from enrolment

19 Analysis Analysis was conducted using methods appropriate to the trial design, with a small number of clusters Adjustment for individual-level baseline factors showing imbalance by study arm

20 Xpert 20 clusters N=4,972 people investigated for TB Microscopy 0 clusters 2,54 participants 0 clusters 2,344 participants Screened Not eligible/did not consent: Xpert: N=97 Microscopy: N=63 Enrolled 0 clusters 2,43 participants 0 clusters 2,368 participants Withdrawn: Xpert: N=20 Microscopy: N=36 0 clusters 2,324 participants In analysis N=4,656 people investigated for TB 0 clusters 2,332 participants

21 Baseline characteristics Xpert(n=2324) Microscopy (n=2332) Age (median(iqr)) 35 (28-45) 37 (29-48) Female (%) 64.2% 59.9% HIV status, self-report (%) Known HIV positive 72.8% 62.2% 79.4% 62.3% HIV+, ART ever (%) 33.5% 32.7% CD4 count (median(iqr)) 303 (7-457) 35 (92-480) Body mass index (%) < >25 # TB symptoms (%) % 45.7% 45.6% 9.8% 23.5% 32.4% 34.3% 2.4% 46.2% 4.5% 6.0% 9.6% 27.0% 47.5%

22 Baseline characteristics Xpert(n=2324) Microscopy (n=2332) Age (median(iqr)) 35 (28-45) 37 (29-48) Female (%) 64.2% 59.9% HIV status, self-report (%) Known HIV positive 72.8% 62.2% 79.4% 62.3% HIV+, ART ever (%) 33.5% 32.7% CD4 count (median(iqr)) 303 (7-457) 35 (92-480) Body mass index (%) < >25 # TB symptoms (%) % 45.7% 45.6% 9.8% 23.5% 32.4% 34.3% 2.4% 46.2% 4.5% 6.0% 9.6% 27.0% 47.5%

23 Ascertainment of vital status at 6 months Xpert (n=2324) Microscopy (n=2332) % n % n Vital status known 98.9% % 2309 Lost to follow-up.% 25.0% 23

24 Effect of XpertMTB/RIF on mortality risk over 6 months Xpert Microscopy Risk ratio (95% CI) Deaths/N % Deaths/N % Unadjusted Adjusted 2 9/ % 6/ % 0.86 ( ).0 ( ) summary ignores cluster, 2 adjusted for age group, sex, body mass index group, number of TB symptoms and HIV status

25 Effect of XpertMTB/RIF on mortality risk over 6 months Xpert Microscopy Risk ratio (95% CI) Deaths/N % Deaths/N % Unadjusted Adjusted 2 9/ % 6/ % 0.86 ( ).0 ( ) summary ignores cluster, 2 adjusted for age group, sex, body mass index group, number of TB symptoms and HIV status Cumulative proportion dying Months since enrolment Number at risk (deaths) Microscopy 293 (63) 22 (3) 2083 (20) 0 Xpert 2097 (45) 204 (30) 2008 (6) 0. Microscopy Xpert Kaplan-Meier failure curves for mortality among all study participants (N=4656), by study arm

26 Risk factor analysis for mortality (six months from enrolment) Positive index test No Yes OR aor 95% CI P-value Gender Female Male Age <

27 Risk factor analysis for mortality (six months from enrolment) Positive index test No Yes OR aor 95% CI P-value Gender Female Male Age <

28 Risk factor analysis for mortality (six months from enrolment) Positive index test No Yes OR aor 95% CI P-value Gender Female Male Age <

29 Risk factor analysis for mortality (six months from enrolment) # TB symptoms Body mass index (kg/m 2 ) HIV status, self report < HIV- HIV+. ART- HIV+, ART+ Unknown OR aor 95% CI P-value < <0.00

30 Risk factor analysis for mortality (six months from enrolment) # TB symptoms Body mass index (kg/m 2 ) HIV status, self report < HIV- HIV+. ART- HIV+, ART+ Unknown OR aor 95% CI P-value < <0.00

31 Risk factor analysis for mortality (six months from enrolment) # TB symptoms Body mass index (kg/m 2 ) HIV status, self report < HIV- HIV+. ART- HIV+, ART+ Unknown OR aor 95% CI P-value < <0.00

32 Proportion index test positive 97% (44/4656) had index specimen result available Xpert Microscopy Prevalence ratio (95% CI) positive % positive % Unadjusted Adjusted /N /N 200/ % 74/ %.27( ).49 ( ) adjusted for age group, sex, body mass index group and number of tuberculosis symptoms

33 Proportion index test positive 97% (44/4656) had index specimen result available Xpert Microscopy Prevalence ratio (95% CI) positive % positive % Unadjusted Adjusted /N /N 200/ % 74/ %.27( ).49 ( ) adjusted for age group, sex, body mass index group and number of tuberculosis symptoms

34 Initial loss to follow-up (No evidence of starting TB treatment within 28 days from enrolment) Among 374 with a positive index test result Xpert Microscopy Risk ratio (95% CI) notb Rx % no TB Rx % Unadjusted Adjusted /N /N 34/ % 26/74 4.9% 0.97 ( ) 0.96 ( ) adjusted for body mass index group and number of tuberculosis symptoms Cumulative proportion starting TB treatment Days from enrolment Kaplan-Meier curves - time to starting TB treatment, among those test positive Microscopy Xpert

35 Proportion rifampicin resistant In the Xpertarm, of the 200 participants with a positive index result, 8 (4%) had a rifampicin resistance signal

36 Proportion treated for TB Overall.6 % (54/4656) were treated for TB over the 6 month follow-up period Xpert Microscopy Risk ratio (95% CI) TBrx/N % TBrx/N % Unadjusted Adjusted 250/ % 29/ % 0.88 ( ).04 ( ) adjusted for age group, sex, body mass index group and number of tuberculosis symptoms Cumulative proportion starting TB treatment Months since enrolment Number at risk (TB) Microscopy 2332 (246) 202 (33) 96 (2) 0 Xpert 2324 (26) 2058 (23) 2003 () 0. Microscopy Xpert Kaplan-Meier curves - time to starting TB treatment, among clinic attendees investigated for TB

37 Proportion with microbiological confirmation among those treated for TB Microbiological confirmation: as any positive Xpert, microscopy or culture Overall 7. % (385/54) of individuals treated for TB had microbiological confirmation Xpert Microscopy Prevalence ratio (95% CI) micro+/n % micro+/n % Unadjusted Adjusted 96/ % 89/ %.2 ( ).20 ( ) adjusted for age group, sex, body mass index group and number of tuberculosis symptoms

38 Proportion with microbiological confirmation among those treated for TB Microbiological confirmation: as any positive Xpert, microscopy or culture Overall 7. % (385/54) of individuals treated for TB had microbiological confirmation Xpert Microscopy Prevalence ratio (95% CI) micro+/n % micro+/n % Unadjusted Adjusted 96/ % 89/ %.2 ( ).20 ( ) adjusted for age group, sex, body mass index group and number of tuberculosis symptoms

39 Limitations Small numbers of rifampicin resistant samples we cannot comment on the effect of Xpert MTB/RIF on outcomes for drug resistant TB Analysis of costing data is in progress Population level impact will be determined through mathematical modeling

40 Conclusion Mortality among persons investigated for TB was high Mortality was not reduced after implementation of Xpert MTB/RIF Unknown HIV status, or positive status but not taking ART, were important determinants of mortality

41 Conclusion People being investigated for TB should know their HIV status, and linkage to HIV care improved The relatively constant death rate over 6 months of follow-up suggests opportunities to intervene Those at highest risk of mortality are Older age Multiple TB symptoms HIV status: Unknown HIV positive not taking ART

42 Conclusion Xpert, compared with microscopy increased the proportion test positive by 50%: may facilitate rapid initiation of TB treatment by nurses; potential to result in cost savings both for patients and for TB programs did not reduce initial loss to follow-up did not increase proportion starting TB treatment increased the proportion of bacteriologically-confirmed TB cases among those starting treatment by 20%

43 Key message Scale-up of a more sensitive diagnostic test requires strengthened health systems, particularly ensuring that people know their HIV status and those eligible, start ART promptly people with confirmed TB start treatment quickly

44 XTEND investigators Aurum Institute Gavin Churchyard Kerrigan McCarthy Violet Chihota National Department of Health Lerole D. Mametja Norbert Ndjeka Lindiwe Mvusi National Health Laboratory Service Wendy Stevens Linda Erasmus University of Cape Town Mark Nicol Helen Cox Edina Sinanovic London School of Hygiene & Tropical Medicine Anna Vassall Katherine Fielding Alison Grant World Health Organization Christopher Dye

45 The thousand of participants in the XTEND trial XTEND study team: Acknowledgements Jo burg:sibuseginindza, Sarah Yates, Joseph Makhura, Bandile Ndlazi, SiphokaziMngcomzelo, Nolundi Msheshwe, Bongiwe Nxumalo, Samantha Naicker, Dumisane Mlotshwa, Puleng Marokane, Flora Popane Eastern Cape: Lungile Vezi, Asanda Mgoli, Akhona Mseleni, Nomtu Gladys Qabaka, Nombuso Nofelitho, Makiwe Mbelu, NtandokaziNgoma, MandisaMatanzima, SimpiweMemela, NobendibaEuriamGqola, TumseOriental Ngqokwe, NondumisoMtshwelo, NonhlanhlaTetiMpakama, BusiswaDyariwe, NtsikaMahlutshana, LizwilombuleloMtwecu, SaneleMkhutshwa, Americana BullieCamba, BonganiNkaqa, Nicholas Ziqubu, NtombizikhonaMkalipi, OlwethuSogoni, SomikaziMgwebi, SilulamiZweni, LindiweMnyukana, NwabisaNgqube, Khayakazi Biko, Nomava Tembisa Ncandana Free State: TombomziMotsoeneng,Julia Motahane, Emily Lekitlane, ModikengMapeka, PulaneMatsime, Maria Mopeli, ItumelengMoloi, ElizahMakhaengHlasa, MoetiLeseba, Gauteng: Zanele Nthebe, Ntokozo Ndlovu, Belinda Dambuza, Nombuso Mokone, Tshegofatso Kelefetswe, Nonhlanhla Uwazureke, Nondumiso Khanyile, Gugulethu Kunene, Kanyesile Radebe, Muzi Mkhwanazi, Anna Mpoelang, Malusi Tshitsha, Agnes Koena, Xolani Nkosi, Soneni Maphosa, Tankiso Molefe, Rosinah Maponyane, Sannah Mokgobo Motau, France Kgoale, DuduzileRadebe, Macy Mavundza, MiyelaniNgobeni, XoliswaMbanjwa, MmoniMothupi, Gloria Ndlovu, Pauline Maswagane, MatomeNtjana, KhutsoPako, MakhananaMawila, TshepisoMonakale, Lucille Motsapi, Lebo Dlamini Mpumalanga: Edith Becka, Fikile Khoza,Bongumusa Dlamini,Mary-Faith Mabitsela, Morabane Malesoena, Ntoana Mbete, Solani Mthimunye

46 Partners and funders Funders Bill and Melinda Gates Foundation Partners Eastern Cape Department of Health: Ms Miyakazi Nokwe, Ms Nomvume Ndinisa Free State Department of Health: Ms Tshidi Morigihlane Gauteng Department of Health: Dr Dimakatso Moli City of JHB: Ms Antonia Barnard City of Tshwane: Ms Julia Mokale Ekurhuleni: Ms Sibongile Mokoena Mpumalanga Department of Health: Ms Duduzile Mbambo

47 Summary of XTEND results Primary outcome n Xpert Microscopy Adjusted Risk ratio % % (95% CI) Mortality risk over 6 months % 5.0%.0 ( ) Secondary outcomes n Xpert % Microscopy % Adjusted Effect measure (95% CI) Index test positive % 7.8%.49 (.00, 2.23) Initial loss to follow-up, over 28 days % treated for TB over 6 months %with microbiological confirmation, among those treated for TB % 4.9% 0.96 ( ) % 2.5%.04 ( ) % 65.0%.20 ( )