Advertisement

Prevalence of drug-resistant tuberculosis in Zimbabwe: A health facility-based cross-sectional survey

Open AccessPublished:July 27, 2019DOI:https://doi.org/10.1016/j.ijid.2019.07.021

      Highlights

      • The prevalence of RR-TB was [4.0% (95% CI, 2.9, 5.4%) and 14.2% (95% CI, 8.9, 21.1%)] among new and retreatment patients, respectively.
      • The prevalence of MDR-TB was 2.0% (95% CI, 1.3, 3.1%) and 6.4% (95% CI, 2.4, 10.3%) among new and retreatment TB patients, respectively.
      • Weighted prevalence of RR-TB and MDR-TB among new TB patients were 4.6% (95% CI, 3.0, 6.2%) and 1.8% (95% CI, 1.0, 2.5%), respectively.
      • Risk factors for RR-TB included prior TB treatment, self-reported HIV infection, travel outside Zimbabwe for >one month (univariate), and age <15 years.
      • The prevalence of MDR-TB has remained stable since the subnational survey in 1994.

      Abstract

      Objective

      To determine the prevalence of resistance to rifampicin alone; rifampicin and isoniazid, and second-line anti-TB drugs among sputum smear-positive tuberculosis patients in Zimbabwe.

      Design

      A health facility-based cross-sectional survey.

      Results

      In total, 1114 (87.6%) new and 158 (12.4%) retreatment TB patients were enrolled. MTB was confirmed by Xpert MTB/RIF among 1184 (93%) smear-positive sputum samples. There were 64 samples with Xpert MTB/RIF-determined rifampicin resistance. However, two were rifampicin susceptible on phenotypic drug susceptibility testing. The prevalence of RR-TB was [4.0% (95% CI, 2.9, 5.4%), n = 42/1043) and 14.2% (95% CI, 8.9, 21.1%; n = 20/141) among new and retreatment patients, respectively. The prevalence of MDR-TB was 2.0% (95% CI, 1.3, 3.1%) and 6.4% (95% CI, 2.4, 10.3%) among new and retreatment TB patients, respectively. Risk factors for RR-TB included prior TB treatment, self-reported HIV infection, travel outside Zimbabwe for ≥one month (univariate), and age <15 years. Having at least a secondary education was protective against RR-TB.

      Conclusion

      The prevalence of MDR-TB in Zimbabwe has remained stable since the 1994 subnational survey. However, the prevalence of rifampicin mono-resistance was double that of MDR-TB.

      Keywords

      Introduction

      In the modern era, Mycobacterium tuberculosis (MTB) drug resistance is among the key challenges in ending TB (
      • Mariandyshev Andrei
      • Eliseev Platon
      Drug-resistant tuberculosis threatens WHO’s end-TB strategy.
      ). In 2016, there were 600 000 new cases globally of multi-drug resistant tuberculosis (MDR-TB), defined as resistance to at least isoniazid and rifampicin (RIF), resulting in an estimated quarter-million annual deaths (
      • World Health Organisation
      Global TB report 2016.
      ). An estimated 92 629 MDR-TB cases (approximately 16% of the global burden) occurred on the African continent. However, 70% of these were not notified to health authorities, and only one-half of the countries have completed a formal drug resistance survey (DRS) (
      • World Health Organisation
      Global tuberculosis report 2017.
      ).
      Although neighboring South Africa reports the second highest absolute number of notified rifampicin-resistant cases in the world (second only to India) (
      • World Health Organisation
      Global tuberculosis report 2015.
      ), and studies from the north of Zimbabwe have indicated a possible increase in MDR-TB prevalence among retreatment cases (
      • Metcalfe J.Z.
      • Makumbirofa S.
      • Makamure B.
      • Sandy Charles
      • Bara W.
      • Mungofa S.
      • et al.
      Drug-resistant tuberculosis in high-risk groups, Zimbabwe.
      ), no nationally representative DRS has been performed in Zimbabwe. In 1994 a sub-national DRS was conducted in Zimbabwe and the prevalence of MDR-TB was 1.9% (95% CI, 1.1, 3.2) and 8.3% (95% CI, 2.9, 21.8) among new and retreatment TB patients, respectively (
      • Mwinga Alwyn
      Drug-resistant tuberculosis in Africa.
      ). At this time in the mid-1990s, HIV was rapidly becoming hyperendemic in Zimbabwe (
      • Harries Anthony D.
      • Hargreaves James
      • Kemp J.R.
      • Jindani Amina
      • Enarson D.A.
      • Maher Dermot
      • et al.
      Deaths from tuberculosis in Sub-Saharan African countries with a high prevalence of HIV-1.
      ), and standard “short-course” 6-month regimens including RIF were yet to be adopted (this was done in 1994), antiretroviral drugs (ARVs) were unavailable, and it was questioned whether the societal costs of treating MDR-TB was worth control of the relatively small risk it presented (
      • Schaaf H.S.
      • Botha P.
      • Beyers N.
      • Gie R.P.
      • Vermulen H.A.
      • Groenwald P.
      • et al.
      The 5-year outcome of multidrug resistant tuberculosis patients in the Cape Province of South Africa.
      ). Since that time, there have been dramatic changes to the TB diagnostic and MDR-TB treatment landscapes; substantial increases in movement of economic and political migrants across borders in the Southern African region; and rapid and sustained scale-up of ARVs.
      The best estimates of the burden of drug resistant TB in Africa require well-performed, population-based survey findings (
      • Ismail Nazir Ahmed
      • Mvusi Lindiwe
      • Nanoo Ananta
      • Dreyer Andries
      • Omar Shaheed V.
      • Babatunde Sanni
      • et al.
      Prevalence of drug-resistant tuberculosis and imputed burden in South Africa: a national and sub-national cross-sectional survey.
      ). We undertook a cross-sectional survey in 2015–2016 to determine the prevalence among new and retreatment TB cases of Rifampicin-mono-resistance, MDR-TB, and resistance to second-line agents among those with MDR-TB. We also sought to assess the risk factors for rifampicin-resistant TB and to compare the MDR-TB estimate to that obtained in 1994 as a measure of the burden of drug-resistant TB.

      Methods

      Study design

      A population-based cross-sectional study. Initially, patients in sampled health facilities were screened and diagnosed for TB using smear microscopy. Those who were smear-positive were asked to enroll in the survey as per WHO guidelines (
      • World Health Organisation
      Guidelines for surveillance of drug resistance in tuberculosis.
      ). The survey was conducted from August 2015-September 2016 on sputum-positive new and retreatment TB patients, regardless of age or HIV status, and not already on anti-TB therapy.
      Since rifampicin resistance has conventionally been considered a proxy for MDR, and due to resource constraints, only Xpert MTB/RIF-determined rifampicin-resistant (RR) specimens proceeded to solid culture and first- and second-line DST. All patients whose samples had RR-TB strains on Xpert were re-interviewed to verify history of TB treatment.

      Survey procedures

      A survey questionnaire eliciting socio-demographic and clinical information (e.g. self-reported HIV status and history of TB) was administered to all consenting participants at enrolment. Two spot-sputum specimens were collected from consenting patients within two days of a smear-positive TB diagnosis. Sputum collection was done under the supervision of trained nurses. About 5 mL of spot-sputum specimens were collected in two 50 mL screw-capped falcon tubes, each containing 5 mL of Cetyl-Pyridium-Chloride. This was done to maintain the integrity of the sample in case of delays (of up to 30 days) in sample transportation to the National Reference Laboratory. Each tube was labelled with a unique patient identification number (PIN). The specimens were triple-packaged in zip-lock bags to minimize spillage and contamination and were stored at room temperature. A private courier transported the specimens to the National TB Reference Laboratory (NTBRL).
      At the NTBRL, both specimens were vortexed for 15 s, pooled, and then split again. One specimen was tested using the Xpert MTB/RIF assay and the other was archived. A barcode reader was used to minimize transcription errors when inputting PIN numbers. In case of errors, assays were repeated using the remaining specimens from first specimens. Subsequent procedures were based upon Xpert MTB/RIF results: if RR-TB was not detected or MTB was not detected, no further procedures were performed. If RR-TB was detected, archived specimens were retrieved, decontaminated and the resultant sputum deposits were inoculated on LJ and pyruvate agar media according to standard operating procedures (
      • Stop TB Partnership
      Mycobacteriology laboratory manual.
      ). The media were incubated at 37 °C and growth of MTB was observed weekly for up to 6 weeks. Part of the deposit was inoculated on LJ agar slants in 5 mL cryo-vials for shipment to a Supranational TB Reference Laboratory (SRL) in Antwerp, Belgium for external quality assurance. At the NTBRL, phenotypic culture and drug susceptibility testing (CDST) was done on LJ on all MTB positive isolates using the proportion method (
      • Stop TB Partnership
      Mycobacteriology laboratory manual.
      ). First-line DST was done for the drugs streptomycin, isoniazid, rifampicin and ethambutol (SIRE), and second-line DST for kanamycin, amikacin, ofloxacin, moxifloxacin and capreomycin. All the isolates were stored at −20 °C in cryo-vials with 10% glycerol. Hain Line Probe assay (LPA) (Hain LifeSciences, Nehren, Germany) was carried out on all cultures that failed to grow. Discordances between Xpert MTB/RIF RR-TB results and first-line phenotypic DST were resolved by conventional Sanger DNA sequencing of rpoB at the SRL. There was a 100% concordance in sensitivity and specificity between NTBRL and the SRL on the drugs kanamycin, capreomycin, ofloxacin and rifampicin. Sensitivity and specificity of isoniazid were 90% and 89%, respectively.

      Sampling

      Sampling was done as per WHO guidelines (
      • World Health Organisation
      Guidelines for surveillance of drug resistance in tuberculosis.
      ). First, probability proportional to size sampling was used to select 63 of 146 national TB diagnostic sites that were functional in 2012, and 20 of 56 national TB diagnostic sites that became functional between 2012 and 2014. Within each selected diagnostic site, consecutive eligible patients were enrolled until the required number of new cases for that site was reached, or the end of the survey period was reached.
      As per WHO recommendations, sample size was calculated based on new patients only; retreatment patients were sampled on convenience. For new patients, a sample size of 677 was based on the following assumptions: (i) a total national notification of 12,405, based on 2012 programme data; (ii) an absolute precision of 1% at 95% confidence interval (CI); (iii) a priori estimated prevalence of MDR-TB of 1.9%, based on the 1994 sub-national survey. After factoring in a design effect of two and accounting for possible losses of up to 20%, a minimum sample of 1625 new smear-positive patients was estimated.

      Survey and data management

      A survey management team and a steering committee were established to ensure smooth implementation of the survey. A pilot survey was conducted in 10% of the sites. Three teams from the national office were trained and they later provided on-site trainings to survey teams (TB nurses and laboratory staff) in different provinces starting with low-volume sites.
      Each recruiting facility maintained a survey register which captured patient demographic and clinical data. Each patient had a PIN which was linked to all the survey tools (survey register, laboratory request form and NTBRL laboratory register). Xpert MTB/RIF and CDST results were reported to facilities to inform clinical management of patients. Quality of data was ensured through training of survey teams, cross-checking original forms during support visits by local teams and during data monitoring missions supported by staff from WHO and KNCV.
      De-anonymised data were sent to the central level by a courier for double-data entry into the Census and Surveys Processing System (CSPro) database by Zimbabwe National Statistics Agency staff. Electronic data were stored in a password-protected computer and backed-up on CDs stored in a locked-file cabinet. Source documents were stored in locked-file cabinets (Table 1).
      Table 1Number and (proportion) of participants who were enrolled in the Zimbabwe Drug Resistant survey by province, 2015–16.
      ProvinceTotal number notified during the survey periodNumber of new patients(%)Expected number of new patients
      Total52791114(65.5)1700


      align="right"
      Manicaland298135(79.4)170
      Mashonaland Central250138(73.8)187
      Mashonaland East30478(65.5)119
      Mashonaland West282133(65.2)204
      Matabeleland North15271(54.9)153
      Matabeleland South32883(46.4)119
      Midlands300136(88.9)153
      Masvingo25893(60.8)153
      Harare2879166(46.5)357
      Bulawayo22881(95.3)85

      Data analysis

      Data were exported to SPSS version 20 (Chicago, Illinois, USA) for analysis. Categorical variables were summarized using frequencies. Continuous variables were summarized using means and medians as appropriate. Weighted analysis of prevalence of RR-TB and MDR-TB were done using exact sampling probabilities to adjust for sampling error due to combining two sampling methods and the capping of patient recruitment at 12 months. Odds ratios and their 95% CI for factors associated with RR-TB were calculated using the stepwise logistic regression. Level of significance was set at p < 0.05.

      Ethics

      This survey was approved by the Medical Research Council of Zimbabwe and the Research Council of Zimbabwe. All the participants provided written informed consent/assent prior to enrolment and collection of sputum specimens.

      Results

      A total of 5279 sputum smear-positive patients were notified during the survey period. Of these, 1301 (24.6%) were initially enrolled and tested using Xpert MTB/RIF (Figure 1). Twenty-nine patients (2%) were excluded due to lack of survey forms and/or barcoding. The analysis population was 1272 patients: 1114 (87.6%) new and 158 (12.4%) retreatment (Figure 1). Of these, 766 (60.2%) were male, the median age was 34 years [(interquartile range (IQR), 27–42 years)], 699 (55.0%) self-reported a history of HIV infection, and 765 (60.1%) were recruited from urban clusters (Table 2). A total of 293 (23%) participants had a history of travel outside Zimbabwe of ≥one month’s duration.
      Figure 1
      Figure 1Flow of participants who were enrolled in the Zimbabwe DRS 2015–2016.
      MDR-TB = multi-drug resistant TB; FQ = Fluoroquinolone; XDR-TB = extensively drug resistant TB; RMR = rifampicin mono-resistant TB; Poly-resistant = resistance to >one first-line anti-TB drug, other than both isoniazid and rifampicin.
      Table 2Socio-demographic and clinical characteristics of patients enrolled in the Zimbabwe TB drug resistant survey, 2015–2016.
      Demographic characteristicsTB patientsTotal
      NewRetreatment
      n(%)n(%)n(%)
      Column percentages.
      Total1114(87.6)158(12.4)1272
      Sex
       Male668(60.0)98(62.0)766(60.2)
       Female446(40.0)60(38.0)506(39.8)
      Age group
       <1518(1.6)1(0.6)19(1.5)
       15–24171(15.4)13(8.2)184(14.5)
       25–34415(37.3)43(27.2)458(36.0)
       35–44315(28.3)50(31.6)365(28.7)
       45–54116(10.4)33(20.9)149(11.7)
       55–6446(4.1)7(4.4)53(4.2)
       ≥6531(2.8)11(7.0)42(3.3)
       Unknown2(0.2)0(0)2(0.2)
      HIV status
       Positive580(52.1)119(75.3)699(55.0)
       Negative492(44.2)34(21.5)526(41.4)
       Unknown42(3.7)5(3.2)47(3.6)
      History of any travel outside Zimbabwe
       For ≥1 month243(21.8)50(31.6)293(23.0)
       To South Africa166(14.9)32(20.3)198(15.6)
       Other SADC countries62(5.6)17(10.8)79(6.2)
       To other SADC countries8(0.7)1(0.6)9(0.7)
       Unknown7(0.6)0(0.0)7(0.6)
      Marital status
       Never married229(20.6)22(13.9)251(19.7)
       Married600(53.9)75(47.5)675(53.1)
       Divorced177(15.9)33(20.9)210(16.5)
       Widowed89(8.0)22(13.9)111(8.7)
       Unknown19(1.7)6(3.8)25(2.0)
      Level of education
       None39(3.5)4(2.5)43(3.4)
       Primary312(28.0)46(29.1)358(28.1)
       Secondary700(62.8)94(59.5)794(62.4)
       Tertiary55(4.9)13(8.2)68(5.3)
       Missing8(0.7)1(0.6)9(0.7)
      Cluster location
       Urban671(60.2)94(59.5)765(60.1)
       Rural443(39.8)64(40.5)507(39.9)
      SADC = Southern Africa Development Community; TB = tuberculosis.
      a Column percentages.

      Bacteriologic results

      Of the 1272 valid Xpert MTB/RIF assays, 1184 (93.1%) detected MTB. There were 44 (3.5%) new and 20 (1.6%) retreatment TB patients who had Xpert-determined RR-TB. Of these 64, 50 (78.1%) successfully grew on culture at the NTBRL. First and second-line phenotypic DST confirmed RR-TB in 48 (96%), while two cultures (4%) were susceptible to all the first-line drugs (SIRE) according to phenotypic CDST, Hain LPA (at the NTBRL), and rpoB gene sequencing at the SRL. Twenty-five cultures [(52.1%) (95% CI, 38.3, 65.5)] had MDR-TB; 20 demonstrated rifampicin mono-resistance (RMR); three had poly-resistance and two were rifampicin susceptible. Of the 25 MDR-TB cultures, one (4.0%) demonstrated fluoroquinolone and aminoglycoside resistance in addition to MDR (XDR-TB).
      The crude prevalence of RR-TB was [4.0% (95% CI, 2.9, 5.4%), n = 42/1043] and [14.2% (95% CI, 8.9, 21.1%), n = 20/141] among new and retreatment patients, respectively. The crude prevalence of MDR-TB was 2.0% [(95% CI, 1.3, 3.1%)] and [6.4% (95% CI, 2.4, 10.3%)] among new and retreatment TB patients, respectively. Among new patients, the weighted prevalence of RR-TB and MDR-TB were [4.6% (95% CI, 3.0, 6.2)] and [1.8% (95% CI, 1.0, 2.5)] respectively.

      Risk factors for RR-TB

      In univariate analysis, a history of travel outside Zimbabwe for ≥one month [(odds ratio [(OR = 1.74, 95% CI, 1.02, 2.97)] had increased odds of RR-TB. In multivariate analysis, HIV-positivity [adjusted odds ratio (aOR) = 2.12 (95% CI, 1.09, 4.05)], age <15 years [aOR = 6.37 (95% CI, 1.51, 26.87)], and a previous history of TB treatment [aOR = 3.53 (95% CI, 1.86, 6.25)] were associated with RR-TB, while having at least a secondary education was protective [(aOR = 0.52; 95% CI, 0.29, 0.97)] (Table 3). After stratifying by type of TB patient, a positive HIV status [aOR = 2.19; 95% CI, 1.07, 4.46)] and history of travel outside Zimbabwe [aOR = 2.05; 95% CI, 1.05, 4.03)] were significantly associated with RR-TB among new patients (Table 4).
      Table 3Risk factors for rifampicin resistance among patients diagnosed with smear-positive sputum during the TB drug resistant survey, Zimbabwe (2015–2016).
      VariableTotalRR-TB detectedOR (95% CI)aOR 95% CI
      n(%)
      Row percentages.
      118462(5.2)
      Sex
       Female46620(4.3)RefRef
       Male71842(5.8)1.38 (0.78, 2.52)1.43 (0.69, 2.46)
      Age group
       <15184(22.2)6.90 (1.80, 26.45)
      Significant.
      6.37 (1.51, 26.87)
      Significant.
       15–241767(4.0)RefRef
       25–3443119(4.4)1.11 (0.46, 2.70)0.96 (0.38, 2.42)
       35-4433722(6.5)1.68 (0.68, 4.77)1.25 (0.46, 3.27)
       45–541285(3.6)0.91 (0.28, 2.92)0.52 (0.16, 1.75)
       55–64483(6.2)1.61 (0.40, 6.47)1.04 (0.24, 4.42)
       ≥65342(5.9)0.51 (0.30, 7.60)0.90 (0.16, 4.98)
      Level of education
       Primary and less36327(7.4)RefRef
       Secondary and above81334(4.2)0.54 (0.31, 0.95)
      Significant.
      0.52 (0.29, 0.97)
      Significant.
       Unknown81(12.5)1.78 (0.21, 14.99)2.83 (0.30, 27.08)
      Cluster location
       Urban71437(5.2)Ref
       Rural47025(5.3)1.03 (0.58, 1.78)0.90 (0.54, 1.71)
      HIV status
       Negative50814(2.8)RefRef
       Positive63246(7.3)2.77 (1.46, 5.52)
      Significant.
      2.12 (1.09, 4.05)
      Significant.
       Unknown442(4.5)1.68 (0.18, 7.70)1.34 (0.29, 6.24)
      History of any travel outside Zimbabwe
       for ≥1 month28122(7.8)1.74 (1.02, 2.97)
      Significant.
      1.69 (0.95, 2.99)
       To South Africa19017(8.9)1.55 (0.55, 4.36)1.49 (0.57, 4.39)
       To other SADC countries27021(7.8)0.84 (0.10, 6.91)0.87 (0.15, 6.42)
      Treatment history
       New104342(4.0)RefRef
       Retreatment14120(14.2)3.94 (2.11, 7.11)
      Significant.
      3.53 (1.86, 6.25)
      Significant.
      OR = odds ratio; HIV = human immune-deficiency virus; aOR = adjusted odds ratio; SADC = Southern Africa Development Community; Ref = reference.
      a Row percentages.
      b Significant.
      Table 4Factors associated with rifampicin resistance, among patients diagnosed with smear-positive sputum during the TB drug resistant survey, Zimbabwe (2015–2016), disaggregated by type of TB patient.
      Risk factorsType of TB case
      New (n = 1043)Retreatment (n = 141)Total (n = 1184)
      aOR95% CIaOR95% CIaOR95% CI
      Sex
       Female1.080.55, 2.122.160.68, 6.821.430.69, 2.46
       MaleReferenceReferenceReference
      Age group
       <158.591.47, 50.04
      Significant.
      6.371.51, 26.87
      Significant.
       15–24ReferenceReferenceReference
       25–341.620.45, 5.840.280.05, 1.480.960.38, 2.42
       35–442.440.68, 8.770.290.06, 1.401.250.46, 3.27
       45–540.640.10, 4.060.190.03, 1.110.520.16, 1.75
       55–642.430.45,13.27N/A1.040.24, 4.42
       ≥651.910.19,19.800.200.02, 2.560.900.16, 4.98
      Level of education
       ≤PrimaryReferenceReferenceReference
       ≥Secondary0.520.27, 1.020.750.21, 2.670.520.29, 0.97
      Significant.
       Unknown5.110.51, 51.252.830.30, 27.08
      HIV status
       NegativeReferenceReferenceReference
       Positive2.191.07,4.46
      Significant.
      1.760.44, 7.092.121.09, 4.05
      Significant.
       Unknown0.890.11, 7.252.720.21,34.711.340.29, 6.24
      History of travel outside Zimbabwe
       for ≥ 1 month2.051.05,4.03
      Significant.
      1.090.37, 3.161.690.95, 2.99
      Treatment history
       NewReference
       Re-treatment3.531.86, 6.25
      Significant.
      OR = odds ratio; HIV = human immune-deficiency virus; aOR = adjusted odds ratio. – = undefined.
      a Significant.

      Discussion

      This first nationally representative TB-DRS survey for Zimbabwe was conducted following significant socio-political and epidemiological changes in the country. We demonstrated that the prevalence of MDR-TB has remained stable since 1994 though the prevalence of RR-TB is now double that of MDR-TB. We also observed that the factors associated with RR-TB were a previous history of TB, HIV positivity, age <15 years, lower than secondary education and a stay outside Zimbabwe for ≥a month.
      The observed prevalence of MDR-TB was consistent with prevalence reported from South Africa [(2.1% (95% CI, 1.5, 2.7) and 4.6% (CI, 95%: 3.2, 6.0) among new and previously treated patients, respectively in the 2012–2014 survey] and Botswana [(2.5%, 95% CI, 1.6, 3.7) and 6.6%, 95% CI, 3.3, 11.7) among new and previously treated patients during 2007–2008)] (
      • National Institute for Communicable Diseases
      South African tuberculosis drug resistance survey 2012–14.
      ,
      • Menzies H.J.
      • Moalosi G.
      • Anisimova V.
      • Gammino V.
      • Sentle C.
      • Bachhuber M.A.
      • et al.
      Increase in anti-tuberculosis drug resistance in Botswana: results from the fourth national drug resistance survey.
      ), but was lower than prevalence recorded in both Lesotho [(3.1% (95% CI, 2.1, 4.3) and (12.8% (95% CI, 8.8, 18.2)] among previously treated patients and Namibia [(3.8% (95% CI, 2.8, 5.1) and (16.4% (95% CI, 12.9, 20.6)] among previously treated patients for the survey carried out in 2008–2009 (
      • Maama-Maime Llang B.
      • Mareka Mathabo
      • Ershova Julia V.
      • Tlali Thabong E.
      • Kao Kekeletso
      • Phalatse Mamakhetha
      • et al.
      Antituberculosis drug resistance survey in Lesotho, 2008–2009: lessons learned.
      ,
      • Ministry of Health and Social Services: National Tuberculosis and Leprosy Programme
      Namibia national tuberculosis drug resistance survey (2008–2009).
      ). The increase in the prevalence of RR-TB could be attributed to a high proportion of RMR observed in this study. This may be due to acquired resistance to rifampicin. Studies have shown associations between RMR and factors like a positive HIV status, previous histories of TB, use of antifungals and use of rifabutin (
      • Meyssonnier Vanina
      • Van Bui Thuy
      • Veziris Nicolas
      • Jarlier Vincent
      • Robert Jérôme
      Rifampicin mono-resistant tuberculosis in France: a 2005–2010 retrospective cohort analysis.
      ,
      • Ridzon R.
      • Whitney C.G.
      • McKenna M.T.
      • Taylor J.P.
      • Ashkar S.H.
      • Nitta A.T.
      • et al.
      Risk factors for rifampin mono-resistant tuberculosis.
      ,
      • Villegas Leonela
      • Otero Larissa
      • Sterling Timothy R.
      • Huaman Moises A.
      • Van der Stuyft Patrick
      • Gotuzzo Eduardo
      • et al.
      Prevalence, risk factors, and treatment outcomes of isoniazid- and rifampicin-mono-resistant pulmonary tuberculosis in Lima, Peru.
      ). The association between a positive HIV-status and RR-TB could be attributed to acquired drug resistance resulting from “preferential adherence” to antiretroviral drugs at the expense of anti-TB drugs among TB co-infected patients as reported in a previous qualitative study (
      • Daftary Amrita
      • Padayatchib Nesri
      • Max O’Donnellb C.
      Preferential adherence to antiretroviral therapy over tuberculosis (TB) treatment: a qualitative study of drug–resistant TB/HIV co-infected patients in South Africa.
      ).
      The findings that the previous histories of TB and stay outside Zimbabwe for ≥1 month (bivariate analysis) were associated with RR-TB were not surprising. The latter may stem from the fact that most Zimbabweans go to neighbouring, high TB-burden countries as economic emigrants and living conditions there may foster acquisition of RR-TB. Neither DRS surveys in southern Africa nor studies, including systematic reviews, found any association between MDR-TB and HIV infection (
      • Lukoye Deus
      • Ssengooba Willy
      • Musisi Kenneth
      • Kasule George W.
      • Cobelens Frank G.J.
      • Joloba Moses
      • et al.
      Variation and risk factors of drug resistant tuberculosis in Sub-Saharan Africa: a systematic review and meta-analysis.
      ,
      • National Institute for Communicable Diseases
      South African tuberculosis drug resistance survey 2012–14.
      ,
      • Maama-Maime Llang B.
      • Mareka Mathabo
      • Ershova Julia V.
      • Tlali Thabong E.
      • Kao Kekeletso
      • Phalatse Mamakhetha
      • et al.
      Antituberculosis drug resistance survey in Lesotho, 2008–2009: lessons learned.
      ,
      • Ministry of Health and Social Services: National Tuberculosis and Leprosy Programme
      Namibia national tuberculosis drug resistance survey (2008–2009).
      ,
      • Menzies H.J.
      • Moalosi G.
      • Anisimova V.
      • Gammino V.
      • Sentle C.
      • Bachhuber M.A.
      • et al.
      Increase in anti-tuberculosis drug resistance in Botswana: results from the fourth national drug resistance survey.
      ).
      We do not know the reasons why attainment of secondary education was protective against RR-TB. Perhaps, attainment of secondary education is associated with better socio-economic status and positive health behavioral traits. By contrast, a study done in China showed that attainment of high school was a risk factor for RR-TB (
      • Yang Xiujun
      • Yuan Yanli
      • Pang Yu
      • Wang Bo
      • Bai Yunlong
      • Wang Yanhua
      • et al.
      The burden of MDR/XDR tuberculosis in coastal plains population of China.
      ). Drug resistant surveys from southern Africa did not assess the relationship between education and risk of RR-TB. Studies from high-income countries failed to show the relationship between education and risk of RR-TB. Thus more research is needed to investigate this relationship.
      This study enrolled more patients from urban than from rural areas, a consistent finding with program data in which high notifications are recorded in urban centers. This exposes geographical health inequities despite a similar burden of MTB between rural and urban centers as evidenced in this study and the 2014 Zimbabwe National TB prevalence survey (
      • Ministry of Health and Child Care
      Zimbabwe national population-based tuberculosis prevalence survey.
      ).
      This survey had some limitations. First, in adopting sputum smear-positivity as our starting point, we invariably underreported RR-TB in this high HIV-burden setting given that children and HIV-positive patients often produce pauci-bacillary specimens. Second, all sputum specimens in which RR-TB was not detected did not undergo CDST. Thus, the prevalence of isoniazid mono-resistance is unknown, though it is known to be rising elsewhere in the southern Africa Development Community (SADC) (
      • Variava Ebrahim
      • Martinson Neil
      Drug-resistant tuberculosis: the rise of the monos.
      ,
      • National Institute for Communicable Diseases
      South African tuberculosis drug resistance survey 2012–14.
      ). Future DRS surveys should determine the prevalence of isoniazid-mono resistance. Third, 6% of new and 11% of retreatment sputum smear-positive specimens tested Xpert MTB/RIF-negative, raising the possibility of non-tuberculosis mycobacteria (NTM). Indeed the previn adopting sputum smear-positivity as our starting point, we invariably underreported RR-TB in this high HIV-burden setting given that children and HIV-positive patients often produce pauci-bacillary specimens. Second, all sputum specimens in which RR-TB was not detected did not undergo CDST. Thus, the prevalence of isoniazid mono-resistance is unknown, though it is known to be rising elsewhere in the southern Africa Development Community (SADC) (
      • Variava Ebrahim
      • Martinson Neil
      Drug-resistant tuberculosis: the rise of the monos.
      ,
      • National Institute for Communicable Diseases
      South African tuberculosis drug resistance survey 2012–14.
      ). Future DRS surveys should determine the prevalence of isoniazid-mono resistance. Third, 6% of new and 11% of retreatment sputum smear-positive specimens tested Xpert MTB/RIF-negative, raising the possibility of non-tuberculosis mycobacteria (NTM). Indeed the prevalence of NTM is very high in Zimbabwe. During the 2014 TB prevalence survey the prevalence of NTM was estimated to be 16.9% (964/5705) of all the survey presumptive TB cases. Of the NTM isolates obtained in a convenient sample of specimens collected during the 2014 TB prevalence survey, the prevalence of clinically significant NTM such as Mycobacteria Avium complex (MAC) was 51.5% (41/81) (
      • Chin’ombe Nyasha
      • Muzividzi Boniface
      • Munemo Ellen
      • Nziramasanga Pasipanodya
      Molecular identification of nontuberculous mycobacteria in humans in Zimbabwe Using 16S ribosequencing.
      ,
      • Ministry of Health and Child Care
      Zimbabwe national population-based tuberculosis prevalence survey.
      ). Fourth, HIV status was obtained by self-report. Incorporation of HIV testing in TB-DRS surveys could have provided crucial information for the NTP on the relationship between HIV and drug-resistant TB (
      • World Health Organization
      Guidelines for surveillance of drug resistance in tuberculosis.
      ). Lastly, we did not do multiple imputation to control for bias on the results for 14 samples that did not have DST results (no culture growths or contaminated cultures) since our data were robust. However, it would have been a useful exercise to compare potential differences in results between our models and the imputation models.
      Several programmatic implications arise from this study. First, there is need to improve early and universal access to DST (in Zimbabwe and elsewhere in SADC) for at least rifampicin, in line with the WHO End TB strategy (
      • STOP TB Partnership
      The paradigm shift 2016–2020: global plan to end TB.
      ). Second, since isoniazid prophylactic therapy (IPT) has been scaled up in Zimbabwe with ≥20,000 PLHIV having been started on IPT by December 2015 and IPT completion rates of ≥89% have been attained (
      • Takarinda K.C.
      • Harries A.D.
      • Choto R.C.
      • Mutasa-Apollo T.
      • Chakanyuka-Musanhu C.
      Routine implementation of isoniazid preventive therapy in HIV-infected patients in seven pilot sites in Zimbabwe.
      ,
      • Takarinda K.C.
      • Choto R.C.
      • Mutasa-Apollo T.
      • Chakanyuka-Musanhu C.
      • Timire C.
      • Harries A.D.
      Scaling up isoniazid preventive therapy in Zimbabwe: has operational research influenced policy and practice?.
      ), and within the context of South African isoniazid-mono resistance of [4.9%, 95% CI: 4.1%–5.8%)],(
      • National Institute for Communicable Diseases
      South African tuberculosis drug resistance survey 2012–14.
      ) estimating the prevalence of and continued monitoring of isoniazid mono-resistance should be prioritized in Zimbabwe. Third, although sample size was small and should be interpreted as hypothesis-generating, we noted an increased risk of RR-TB among older children and adolescents, and warrants additional studies examining the determinants of childhood RR-TB in Zimbabwe. Lastly, we could not follow up on the treatment outcomes of this group.
      In conclusion, the prevalence of MDR-TB in Zimbabwe has remained stable since the 1994 subnational survey.

      Author contributions

      CS, RM, MN, JNS, JC, HM, KC, NK, BMM designed the study. JNS, JC, KC, CT, RM, MN, CS, NK analysed the data. CT, JZM and CS drafted the manuscript. BMM processed the specimens. All authors read and approved the manuscript for intellectual content. All the authors read and approved the final manuscript.

      Conflict of interest statement

      None declared.

      Funding

      This survey was funded by the USAID’s Challenge TB through the World Health Organisation . The funders had no role in the design and interpretation of the findings.

      Acknowledgements

      A special thank you goes to the participants who volunteered to take part in the drug resistant survey. Thanks are also due to all staff who worked tirelessly to make this survey a success.

      References

        • Chin’ombe Nyasha
        • Muzividzi Boniface
        • Munemo Ellen
        • Nziramasanga Pasipanodya
        Molecular identification of nontuberculous mycobacteria in humans in Zimbabwe Using 16S ribosequencing.
        Open Microbiol J. 2016; 10: 113-123https://doi.org/10.2174/1874285801610010113
        • Daftary Amrita
        • Padayatchib Nesri
        • Max O’Donnellb C.
        Preferential adherence to antiretroviral therapy over tuberculosis (TB) treatment: a qualitative study of drug–resistant TB/HIV co-infected patients in South Africa.
        Glob Public Health. 2014; 9: 1107-1116https://doi.org/10.1080/17441692.2014.934266
        • Harries Anthony D.
        • Hargreaves James
        • Kemp J.R.
        • Jindani Amina
        • Enarson D.A.
        • Maher Dermot
        • et al.
        Deaths from tuberculosis in Sub-Saharan African countries with a high prevalence of HIV-1.
        Lancet. 2001; 357: 1519-1523https://doi.org/10.1016/S0140-6736(00)04639-0
        • Ismail Nazir Ahmed
        • Mvusi Lindiwe
        • Nanoo Ananta
        • Dreyer Andries
        • Omar Shaheed V.
        • Babatunde Sanni
        • et al.
        Prevalence of drug-resistant tuberculosis and imputed burden in South Africa: a national and sub-national cross-sectional survey.
        Lancet Infect Dis. 2018; 18: 779-787https://doi.org/10.1016/S1473-3099(18)30222-6
        • Lukoye Deus
        • Ssengooba Willy
        • Musisi Kenneth
        • Kasule George W.
        • Cobelens Frank G.J.
        • Joloba Moses
        • et al.
        Variation and risk factors of drug resistant tuberculosis in Sub-Saharan Africa: a systematic review and meta-analysis.
        BMC Public Health. 2015; 15: 1-13https://doi.org/10.1186/s12889-015-1614-8
        • Maama-Maime Llang B.
        • Mareka Mathabo
        • Ershova Julia V.
        • Tlali Thabong E.
        • Kao Kekeletso
        • Phalatse Mamakhetha
        • et al.
        Antituberculosis drug resistance survey in Lesotho, 2008–2009: lessons learned.
        PLoS One. 2015; 10: 2008-2009https://doi.org/10.1371/journal.pone.0133808
        • Mariandyshev Andrei
        • Eliseev Platon
        Drug-resistant tuberculosis threatens WHO’s end-TB strategy.
        Lancet Infect Dis. 2017; 17: 674-675https://doi.org/10.1016/S14733099(17)30246-3
        • Menzies H.J.
        • Moalosi G.
        • Anisimova V.
        • Gammino V.
        • Sentle C.
        • Bachhuber M.A.
        • et al.
        Increase in anti-tuberculosis drug resistance in Botswana: results from the fourth national drug resistance survey.
        Int J Tuberc Lung Dis. 2014; 18: 1026-1033https://doi.org/10.5588/ijtld.13.0749
        • Metcalfe J.Z.
        • Makumbirofa S.
        • Makamure B.
        • Sandy Charles
        • Bara W.
        • Mungofa S.
        • et al.
        Drug-resistant tuberculosis in high-risk groups, Zimbabwe.
        Emerg Infect Dis. 2014; 20: 135-137
        • Meyssonnier Vanina
        • Van Bui Thuy
        • Veziris Nicolas
        • Jarlier Vincent
        • Robert Jérôme
        Rifampicin mono-resistant tuberculosis in France: a 2005–2010 retrospective cohort analysis.
        BMC Infect Dis. 2014; 14: 18https://doi.org/10.1186/1471-2334-14-18
        • Ministry of Health and Child Care
        Zimbabwe national population-based tuberculosis prevalence survey.
        (Harare, Zimbabwe)2014
        • Ministry of Health and Social Services: National Tuberculosis and Leprosy Programme
        Namibia national tuberculosis drug resistance survey (2008–2009).
        (Windhoek, Namibia)2012
        • Mwinga Alwyn
        Drug-resistant tuberculosis in Africa.
        Ann N Y Acad Sci. 2006; : 106-112https://doi.org/10.1111/j.1749-6632.2001.tb11366.x
        • National Institute for Communicable Diseases
        South African tuberculosis drug resistance survey 2012–14.
        (Pretoria, South Africa)2014
        • Ridzon R.
        • Whitney C.G.
        • McKenna M.T.
        • Taylor J.P.
        • Ashkar S.H.
        • Nitta A.T.
        • et al.
        Risk factors for rifampin mono-resistant tuberculosis.
        Am J Respir Crit Care Med. 2013; 157: 1881-1884https://doi.org/10.1164/ajrccm.157.6.9712009
        • Schaaf H.S.
        • Botha P.
        • Beyers N.
        • Gie R.P.
        • Vermulen H.A.
        • Groenwald P.
        • et al.
        The 5-year outcome of multidrug resistant tuberculosis patients in the Cape Province of South Africa.
        Trop Med Int Health. 1996; 1: 718-722
        • Stop TB Partnership
        Mycobacteriology laboratory manual.
        (Geneva, Switzerland)2014
        • STOP TB Partnership
        The paradigm shift 2016–2020: global plan to end TB.
        (Geneva, Switzerland)2015
        • Takarinda K.C.
        • Choto R.C.
        • Mutasa-Apollo T.
        • Chakanyuka-Musanhu C.
        • Timire C.
        • Harries A.D.
        Scaling up isoniazid preventive therapy in Zimbabwe: has operational research influenced policy and practice?.
        Public Health Action. 2019; 8: 218-224https://doi.org/10.5588/pha.18.0051
        • Takarinda K.C.
        • Harries A.D.
        • Choto R.C.
        • Mutasa-Apollo T.
        • Chakanyuka-Musanhu C.
        Routine implementation of isoniazid preventive therapy in HIV-infected patients in seven pilot sites in Zimbabwe.
        Public Health Action. 2017; 7: 55-60https://doi.org/10.5588/pha.16.0102
        • Variava Ebrahim
        • Martinson Neil
        Drug-resistant tuberculosis: the rise of the monos.
        Lancet Infect Dis. 2018; 18: 705-706https://doi.org/10.1016/S1473-3099(18)30247-0
        • Villegas Leonela
        • Otero Larissa
        • Sterling Timothy R.
        • Huaman Moises A.
        • Van der Stuyft Patrick
        • Gotuzzo Eduardo
        • et al.
        Prevalence, risk factors, and treatment outcomes of isoniazid- and rifampicin-mono-resistant pulmonary tuberculosis in Lima, Peru.
        PLoS One. 2016; 11e0152933https://doi.org/10.1371/journal.pone.0152933
        • World Health Organisation
        Guidelines for surveillance of drug resistance in tuberculosis.
        (Geneva, Switzerland)5th ed. 2015
        • World Health Organisation
        Global tuberculosis report 2015.
        (Geneva, Switzerland)2015
        • World Health Organisation
        Global TB report 2016.
        (Geneva, Switzerland)2016
        • World Health Organisation
        Global tuberculosis report 2017.
        (Geneva, Switzerland)2017
        • World Health Organization
        Guidelines for surveillance of drug resistance in tuberculosis.
        5th edition. 2015 (Geneva, Switzerland)
        • Yang Xiujun
        • Yuan Yanli
        • Pang Yu
        • Wang Bo
        • Bai Yunlong
        • Wang Yanhua
        • et al.
        The burden of MDR/XDR tuberculosis in coastal plains population of China.
        PLoS One. 2015; 10: 1-12https://doi.org/10.1371/journal.pone.0117361