INT J TUBERC LUNG DIS 18(6):676–681 Q 2014 The Union http://dx.doi.org/10.5588/ijtld.13.0839
Microbiological investigation for tuberculosis among HIV-infected children in Soweto, South Africa L. Fairlie,* E. Muchiri,* C. N. Beylis,† T. Meyers,‡ H. Moultrie* *Wits Reproductive Health & HIV Institute, School of Clinical Medicine, †National Health Laboratory Service, Mycobacteriology Referral Laboratory & Department of Clinical Microbiology & Infectious Diseases, and ‡Harriet Shezi Children’s Clinic, Department of Paediatrics, Chris Hani Baragwanath Hospital, Faculty of Health Sciences, University of the Witwatersrand, Cape Town, South Africa SUMMARY S E T T I N G : A paediatric human immunodeficiency virus (HIV) clinic in an academic hospital in Soweto, South Africa. O B J E C T I V E S : 1) To describe and compare the clinical, immunological and virological characteristics of HIVinfected children co-treated for tuberculosis (TB), and 2) to compare those investigated microbiologically with those who were not, with a description of the results of the microbiological TB investigation. D E S I G N : Retrospective analysis of TB-HIV-infected children aged ,15 years treated for TB between 1 October 2007 and 15 March 2009. R E S U LT S : Anti-tuberculosis treatment was initiated in 616/3358 (18%) children during the study period. Microbiological TB investigation results were available for 399/616 (65%), among whom culture-confirmed TB was diagnosed in 49 (12%). Drug susceptibility testing
was performed in 29/49 (59%) children: 5/29 (17%) were isoniazid-resistant, and 3 had multidrug-resistant TB. Children aged .8 years and those between 3 and 8 years were more likely to have culture-confirmed TB than those aged ,3 years (aOR 9.4, 95%CI 2.26–39.08 vs. aOR 6.7, 95%CI 1.60–27.69), as were those with CD4 count ,200 cells/mm3 compared to those with .500 cells/mm3 (aOR 3.95, 95%CI 1.23–12.72). C O N C L U S I O N : Our study in HIV-infected children showed a high TB case rate, a low rate of definite TB and a high rate of drug-resistant TB based on World Health Organization case definitions. Increased uptake of available TB tests and availability of new diagnostic tests remains a priority in high TB-HIV burden settings. K E Y W O R D S : paediatric TB-HIV; microbiological TB investigations
TUBERCULOSIS (TB) is the most common opportunistic infection in human immunodeficiency virus (HIV) infected children, with up to 40% receiving anti-tuberculosis treatment at the time of commencing combination antiretroviral therapy (cART).1 In South Africa, an estimated 460 000 (range 410 000– 520 000) children are HIV-infected;2 the TB incidence rate is 993 per 100 000 population,3 and up to 52% of children with culture-confirmed TB are HIVinfected.4 In HIV-infected children, the risk of TB disease is up to four times greater than in those with a CD4% ,15%.5 Potential for both over- and underdiagnosing TB in HIV-infected children is high, with overlapping clinical features between both diseases further complicated by additional co-existing comorbidities common in this population.6 The World Health Organization (WHO) case definitions for TB disease include clinical features and microbiological investigations; radiological features are more important in those cases with negative
microbiological results.7 In children, the tuberculin skin test (TST) and chest X-ray (CXR) may assume greater diagnostic importance than in adults.8,9 However, the TST cannot distinguish tuberculous infection from TB disease, and is frequently negative in immunocompromised and malnourished individuals.10 Wide variability in CXR interpretation between experienced clinicians and poor correlation with culture-confirmed TB suggests that CXR is an unreliable adjunct to TB diagnosis in children.11,12 Microbiological confirmation of TB remains the most accurate method of diagnosis, and two Cape Townbased studies have reported successful microbiological TB investigation using gastric washing and induced sputum procedures in primary health clinics.13,14 These tests are not performed routinely in most clinical settings due to limited staff capacity, skill shortage, low yield and difficulty in conducting them among small children. Co-treatment with anti-tuberculosis treatment and
Correspondence to: Lee Fairlie, Wits Reproductive Health and HIV Institute, Corner Esselen and Klein Streets, Hillbrow 2001, South Africa. Tel: (þ27) 82 780 9997. Fax: (þ27) 11 358 5400. e-mail: [email protected] Article submitted 21 November 2013. Final version accepted 24 January 2014.
Microbiological TB investigation in HIV-infected children
cART in HIV-infected children remains complicated, making accurate TB diagnosis important. The drugdrug interactions between rifampicin (RMP) based anti-tuberculosis treatment and cART, particularly lopinavir/ritonavir, that is inadequately boosted may result in reduced virological suppression rates.1,15 Polypharmacy with anti-tuberculosis treatment and cART may affect adherence and increase the possibility of both ART-associated and anti-tuberculosis drug-associated side effects. Data describing the clinical, immunological and virological associations with TB diagnosis in HIVinfected children are scarce, as are reports on microbiological TB investigation results. This retrospective cross-sectional review sought to describe and compare clinical, immunological and virological characteristics of HIV-infected children co-treated for TB, distinguished by whether or not the children underwent microbiological TB investigations, and to describe the results of the microbiological TB investigations conducted.
METHODS Study subjects and design We conducted a retrospective cross-sectional review of microbiological TB investigations performed among HIV-infected children aged ,15 years attending Harriet Shezi Children’s Clinic (HSCC), Soweto, South Africa, who were treated for TB between 1 October 2007 and 15 March 2009. Standard of care The HSCC is a public out-patient paediatric HIV clinic at the Chris Hani Baragwanath Academic Hospital (CHBAH) in Soweto, South Africa. HIV was diagnosed according to the guidelines of the South African National Department of Health (SANDoH).16 TB screening was conducted at the initial and subsequent clinic visits using symptomatic enquiry and clinical assessment, with further investigations in TB suspects. WHO case definitions were applied as follows: TB suspects were defined as those presenting with signs and symptoms suggestive of TB; TB cases were defined as a definite TB case or where a health care worker has diagnosed TB and has initiated a full course of anti-tuberculosis treatment; and a definite case of TB was defined as a confirmed diagnosis of Mycobacterium tuberculosis using culture or newer methods such as molecular line-probe assay or, in countries where this is not possible, two positive specimens for acid-fast bacilli (AFB).7 Investigations included CXR, TST and microbiological examinations. All microbiological specimens were obtained from the respiratory tract and extrapulmonary sites, including blood, pleural fluid and cerebrospinal fluid (CSF), as clinically indicated. Children diagnosed with TB were prescribed a
677
standard three-drug regimen of RMP, isoniazid (INH) and pyrazinamide, with ethambutol/ethionamide (EMB/ETH) added as a fourth drug for complicated TB, including TB meningitis or disseminated TB.17 cART was initiated according to SANDoH guidelines at the time, which recommended starting cART 2 months after TB treatment initiation or completing anti-tuberculosis treatment before starting cART, if possible, in stable children.16 Laboratory investigations Microbiological investigations for TB included smear microscopy, culture and drug susceptibility testing (DST). Specimens were processed according to WHO guidelines using the N-acetyl-L-cysteine-sodium hydroxide decontamination process: a 0.5 ml portion of the processed, centrifuged pellet was inoculated into BACTECe MGITe (BD, Sparks, MD, USA) tubes and incubated for a total of 6 weeks. Growth detected in automated MGIT instruments was confirmed as AFB-positive using Ziehl-Neelsen stained smear. M. tuberculosis complex was confirmed using the GenoTypew Mycobacterium CM assay (Hain Lifesciences, Nehren, Germany). DST was only performed on clinician request, in accordance with the National TB Programme Guidelines, and was not routine performed during the study period.17 MTC isolates were initially tested against INH and RMP using the MGITe960 proportion method (BD). If resistance was found against either or both drugs, the isolate underwent further DST against streptomycin (SM), EMB, ofloxacin (OFX), kanamycin (KM) and ETH using the MGIT proportion method. Definitions of drug-susceptible and drug-resistant TB were as follows: susceptible M. tuberculosis, susceptibility to both INH and RMP; INH-monoresistant and RMP-monoresistant, resistance to only INH or RMP, respectively; multidrug-resistant TB (MDR-TB), resistance to both INH and RMP; and extensively drug-resistant TB (XDR-TB) if a MDRTB isolate was also resistant to OFX and KM. Data collection Demographic, clinical and laboratory data at the time of initiating anti-tuberculosis treatment, with a 3month window period, were extracted from three databases: demographic and HIV-related data from the HSCC clinical database; microbiological TB data from the National Health Laboratory Service (NHLS) database; and TB treatment data from the CHBAH TB Care Centre (TBCC) ETR.net database, a SANDoH TB data system. The TBCC and HSCC databases were merged using the hospital’s unique patient identifier to include only children treated for TB. Microbiological M. tuberculosis investigation results for children included in the merged data set of children who attended HSCC and were treated for TB were then imported from the NHLS Mycobacteriol-
678
The International Journal of Tuberculosis and Lung Disease
ogy Referral Laboratory data set. The final data set included all children treated for TB, regardless of whether or not they were investigated microbiologically for TB. Manual file reviews and NHLS database searches were conducted by study staff for missing values. Data extracted included age (calculated from date of birth), weight-for-age Z-scores (WAZ scores), height-for-age Z-scores (HAZ scores), WHO clinical staging, HIV log10 viral load and CD4 count. WAZ and HAZ scores were generated using WHO reference ranges. Children’s characteristics at TB treatment initiation were stratified according to whether or not the children were investigated microbiologically for TB and by TB investigation results. Statistical analysis Statistical analysis and data management were performed in STATA version 12.0 (STATA Corporation, College Station, TX, USA). Medians and interquartile ranges (IQRs) were used to summarise the data. v2 tests were used to explore associations between grouped categorical variables of age, anthropometrical measures, CD4 count and percentage, HIV log 10 viral load and the outcome of TB investigations. Age was categorised as ,3, 73–,8 and 78 years, in keeping with differential TB disease risk.8 The Wilcoxon rank sum test and the v2 test were used to compare continuous variables and categorical variables, respectively. Univariable and multivariable logistic models were fitted to identify factors associated with a positive TB investigation outcome among HIV-infected children investigated microbiologically for TB. Multivariate models were developed using step-wise backward elimination using likelihood ratio tests to identify the model with the best fit. Ethics approval This study was approved by the University of the Witwatersrand Human Sciences Research Ethics Committee, Cape Town, South Africa.
RESULTS Of the 3358 HIV-infected children who attended the HSCC at least once during the study period, 616 (18%) were co-treated for TB and were included in the analysis (Figure); 129 were undergoing cART at the time (Table 1). At the initiation of anti-tuberculosis treatment, the median age was 4.2 years (IQR 1.1 to 8.4); 67% of the children had WHO clinical Stage IV disease, the median WAZ score was 2.0 (IQR 1.1 to 3.3) and the median HAZ score 2.4 (IQR 1.5 to 3.3). The median CD4 count was 425 cells/mm3 (IQR 180 to 847), median CD4 percentage 14.4% (IQR 8.4 to 21) and the median log10 HIV viral load 4.6 copies/ml (IQR 3.5 to 5.3). A higher
Figure Description of human immunodeficiency virus infected children attending HSCC between 1 October 2007 and 15 March 2009, stratified by anti-tuberculosis treatment and microbiological investigation. HSCC ¼ Harriet Shezi Children’s Clinic; TB ¼ tuberculosis; DST ¼ drug susceptibility testing; MDRTB ¼ multidrug-resistant TB.
than expected proportion among those investigated microbiologically for TB were male (P ¼ 0.01). There were no other clinically or statistically significant differences in characteristics of children who were and those who were not investigated microbiologically (Table 1). Microbiological investigations were conducted in 399 (65%) of the children co-treated for TB (Figure). Twenty-five (6%) were AFB-positive using ZiehlNeelsen staining. TB was confirmed in 49 (12%) children using mycobacterial culture. The median age of children with culture-confirmed TB was 8.2 years (IQR 5.7 to 10.1) compared to 3.3 years (IQR 0.9 to 7.9) in those investigated microbiologically with negative TB culture results (P 6 0.001) (Table 2). Children with culture-confirmed TB (definite TB) had a median CD4 count of 157 cells/mm3 (IQR 41 to 299) and CD4 percentage of 10% (IQR 3 to 19) compared to 491 cells/mm3 (IQR 239 to 924) and 14% (IQR 8 to 20) in culture-negative children (P ¼ 0.02 and ,0.001, respectively) (Table 2). DST was performed in 29 (59%) children with cultureconfirmed TB. Drug-resistant TB was diagnosed in 5 (17%) children, of whom 3 were MDR-TB. All three MDR-TB strains were resistant to EMB and SM, and a third child’s M. tuberculosis isolate was also resistant to ETH. Two children were INHmonoresistant (Figure).
Microbiological TB investigation in HIV-infected children
Table 1
679
Characteristics of children attending the HSCC treated for TB, 1 October 2007–15 March 2009
Male sex, n (%) Age, years CD4% CD4 count, cells/mm3 Log10 HIV viral load, copies/ml WHO Stage 4, % Weight-for-age Z-scores Height-for-age Z-scores Receiving cART, n (%)
All children treated for TB (n ¼ 616) Median [IQR]
Not investigated microbiologically for TB (n ¼ 217) Median [IQR]
Investigated microbiologically for TB (n ¼ 399) Median [IQR]
P value
339 (55) 4.2 [1.1 to 8.4] 14.4 [8.4 to 21.0] 425 [180 to 847] 4.6 [3.5 to 5.3] 67 2.0 [–1.1 to 3.3], (n ¼ 440) 2.4 [–1.5 to 3.3], (n ¼ 521) 129 (21)
105 (48) 4.2 [1.2 to 8.6] 15.2 [10.0 to 22.0] 389 [164 to 747] 4.6 [3.2 to 5.1] 65 1.9 [–0.9 to 3.2], (n ¼ 146) 2.4 [–1.5 to 3.1], (n ¼ 176) 50 (23)
234 (58) 4.2 [1.0 to 8.3] 13.8 [7.9 to 20.3] 445 [200 to 859] 4.6 [3.7 to 5.4] 68 2.1 [–1.1 to 3.4], (n ¼ 294) 2.4 [–1.5 to 3.3], (n ¼ 345) 79 (20)
0.01 0.23 0.07 0.33 0.21 0.42 0.25 0.43 0.35
HSCC ¼Harriet Shezi Children’s Clinic; TB ¼tuberculosis; IQR ¼interquartile range; HIV ¼ human immunodeficiency virus; WHO ¼ World Health Organization; cART ¼ combined antiretroviral therapy.
Multivariate logistic regression confirmed that confirmation of TB using culture was independently associated with both older age and a lower CD4 count, after adjusting for sex, WAZ and HAZ. Children aged .8 years and those aged between 3 and 8 years were more likely to have cultureconfirmed TB compared to those aged ,3 years (adjusted odds ratio [aOR] 9.4, 95% confidence interval [CI] 2.26–39.08 and aOR 6.7, 95%CI 1.60– 27.69, respectively). Children with a CD4 count ,200 cells/mm3 had increased odds of cultureconfirmed TB compared with those with a CD4 count .500 cells/mm3 (aOR 3.95, 95%CI 1.23– 12.72).
DISCUSSION Our study showed a high TB case rate (18%) in HIVinfected children attending a public HIV treatment clinic, with a low microbiological TB investigation rate and low rate of definite TB that was confirmed using culture or AFB in those investigated microbiologically. Similar high TB-HIV co-treatment rates have been described in other sub-Saharan settings. In Uganda, 17% of HIV-infected children (median age 6.6 years) were co-treated for TB,18 while in a study from four South African hospitals in Johannesburg
and Cape Town, 21% of HIV-infected children (median age 6.3 years) were co-treated for TB; in 23% of the children, a definite TB diagnosis was made (AFB smear n ¼ 30, TB culture n ¼ 24, biopsy n ¼ 11), although the number with culture-confirmed TB was not described.19 TB case rates of 40% by 1 year of follow-up on cART have been described in South African HIV-infected children aged ,2 years.20 Although only 12% of TB cases had cultureconfirmed TB (definite TB cases), it is of concern that only 65% of children classified as a TB case were investigated microbiologically. Apart from male children undergoing more microbiological investigations, there was no other clinical or immunological difference between children who underwent microbiological TB investigations and those who did not. In our study, children investigated microbiologically using culture-confirmed TB were older and more immunologically compromised than those who were culture-negative. The correlation between immunological suppression and confirmed TB in children is likely due to a number of factors: although we do not have data on the source of the specimens, 80% of TB specimens processed at the NHLS are respiratory in origin, and the same is likely in this study. Older children are more likely to be able to expectorate and produce adequate sputum samples. We also expect
Table 2 Characteristics of HIV-infected children co-treated for TB at HSCC from 1 October 2007 to 15 March 2009 stratified by microbiological TB investigation results TB confirmed on culture (n ¼ 49) Median [IQR]
Negative TB culture (n ¼ 350) Median [IQR]
P value
Male sex, n (%) 25 (51) 209 (60) 0.25 Age, years 8.2 [5.7 to 10.1] 3.3 [0.9 to 7.9] ,0.001 CD4% 9.8 [3.0 to 18.7] 14.4 [8.4 to 20.5] 0.02 157 [41 to 299] 491 [239 to 924] ,0.001 CD4, cells/mm3 4.6 [3.4 to 5.2] 4.7 [3.7 to 5.4] 0.35 Log10 HIV viral load, copies/ml WHO Stage 4, % 76 67 0.22 Weight-for-age Z-scores 2.1 [–1.2 to 3.3], (n ¼ 30) 2.1 [–1.1 to 3.4], (n ¼ 264) 0.85 Height-for-age Z-scores 2.4 [–1.4 to 3.2], (n ¼ 40) 2.5 [–1.5 to 3.3], (n ¼ 305) 0.48 HIV ¼ human immunodeficiency virus; TB ¼ tuberculosis; HSCC ¼ Harriet Shezi Children’s Clinic; IQR ¼ interquartile range; WHO ¼ World Health Organization.
680
The International Journal of Tuberculosis and Lung Disease
Table 3 Logistic regression models for factors associated with TB confirmation in HIV-infected children co-treated for TB Characteristic at start of anti-tuberculosis treatment
Unadjusted OR (95%CI)
Unadjusted P value
aOR (95%CI)
Adjusted P value
1 1.42 (0.72–2.21)
0.25
1 1.89 (0.70–5.10)
0.21
1 4.01 (1.49–10.79) 9.39 (3.75–23.47)
,0.001 ,0.001
1 6.65 (1.60–27.69) 9.40 (2.26–39.08)
0.009 0.002
CD4 count, cells/mm3 7500 200–500 ,200
1 2.69 (1.02–7.10) 6.81 (2.75–16.8)
0.04 ,0.001
1 0.65 (0.16–2.54) 3.95 (1.23–12.72)
0.53 0.02
CD4 % 725 ,25
1 1.35 (0.46–4.00)
0.59
1 0.79 (0.33–1.90)
0.60
1 1.51 (0.75–3.00)
0.24
Sex Male Female Age, years ,3 3–8 .8
Viral load, copies/ml ,400 7400 WHO Stage Stage 3 Stage 4
On cART at start of anti-tuberculosis treatment Yes 1 No 0.95 (0.45–2.01)
0.91
Weight-for-age Z-score
1.04 (0.84–1.29)
0.69
0.62 (0.38–1.02)
0.06
Height-for-age Z-scores
1.12 (0.93–1.35)
0.25
1.78 (1.09–2.91)
0.02
TB ¼ tuberculosis; HIV ¼ human immunodeficiency virus; OR ¼ odds ratio; CI ¼ confidence interval; aOR ¼ adjusted OR; WHO ¼ World Health Organization; cART ¼combined antiretroviral therapy.
declining immune function with advancing age and disease progression in HIV-infected children. Adult studies, however, have shown lower rates of microbiological TB diagnosis with severe immunodeficiency.21 We also found that children who had cultureconfirmed TB and underwent DST had high rates of INH-resistant TB (17%), with 10% having MDRTB. These findings are consistent with studies in Cape Town and Johannesburg that have shown INHresistant TB in up to 15% and MDR-TB in 9% of children with culture-confirmed TB.4,22 In the light of these findings, it is worrying that more children were not investigated microbiologically for TB before commencing anti-tuberculosis treatment. TB culture is labour-intensive and time-consuming, leading to long turnaround times for results and low sensitivity in children.10 Recent WHO policy updates recommend the use of the Xpertw MTB/RIF assay (Cepheid, Sunnyvale, CA, USA) rather than conventional microscopy, culture and DST as the initial diagnostic test in children presumed to have MDR-TB or HIV-associated TB. However, rollout of the Xpert test has been limited in routine clinic settings.23 The Xpert test relies on the collection of respiratory specimens, which remains a barrier to microbiological investigations in children. Two Cape Town-based studies have shown that
gastric washings and induced sputum for microbiological TB testing are feasible in a primary clinic setting and that gastric washings on two consecutive days may have a diagnostic yield similar to induced sputum. This technique is particularly useful in infants and young children.13,14 Ongoing training of health care workers in specimen collection techniques is required to ensure that these rapid technologies are increasingly utilised in community settings. The study has some limitations. Data were collected retrospectively and were dependent on accurate recording by the attending clinician and those performing data capture. As TB treatment data were not accurately collected in the HSCC database, only children registered for anti-tuberculosis treatment at the CHBAH TBCC were included. This may have led to the exclusion of children initiated on antituberculosis treatment outside the CHBAH, with a possible underestimation of TB cases. Inconsistent recording of patient names or identification of numbers between data sets may have resulted in some children who were not linked in the NHLS data set undergoing microbiological investigations for TB. We were unable to break down the source of microbiological specimens due to the data extraction methods used at the time; however, as 80% of TB specimens received by NHLS are pulmonary specimens, we assumed that the same proportion would
Microbiological TB investigation in HIV-infected children
apply to this study. Due to the large size of the cohort and the quality of the CXR and TST data in the files, data were not confirmed in a file review. Studies have found poor correlation between culture-confirmed TB and CXR findings and wide variability in interpretation between clinicians.12 We did not have sufficiently accurate data to describe TB suspects who were not treated for TB.
9
10
11
CONCLUSIONS Our study showed a high TB case rate in HIV-infected children, with a low rate of microbiological TB investigation and microbiological confirmation. It is worrying to note the high rate of drug-resistant TB among those investigated microbiologically. These findings suggest that efforts to increase the use of currently available tests should be encouraged through staff education and training. There is also a need for new simple, cheap, rapid and reliable TB tests, particularly in resource-limited settings with a high endemic TB burden. Acknowledgements The authors would like to thank the children and their families attending HSCC. Conflict of interest: none declared.
References 1 Frohoff C, Moodley M, Fairlie L, et al. Treatment outcomes among HIV-infected infants and young children following modifications to protease inhibitor-based therapy due to tuberculosis treatment. PLOS ONE 2011; 6: e17273. 2 Joint United Nations Programme on HIV/AIDS. HIV and AIDS estimates 2011. Geneva, Switzerland: UNAIDS, 2011. http:// www.unaidsorg/en/regionscountries/countries/southafrica/. Accessed March 2014. 3 World Health Organization. Global tuberculosis report, 2012. WHO/HTM/TB/2012.6. Geneva, Switzerland: WHO, 2012. 4 Fairlie L, Beylis N C, Reubenson G, Moore D P, Madhi S A. High prevalence of childhood multidrug-resistant tuberculosis in Johannesburg, South Africa: a cross sectional study. BMC Infect Dis 2011; 11: 1. 5 Elenga N, Kouakoussui K A, Bonard D, et al. Diagnosed tuberculosis during the follow-up of a cohort of human immunodeficiency virus-infected children in Abidjan, Cote ˆ d’Ivoire: ANRS 1278 Study. Pediatr Infect Dis J 2005; 24: 1077. 6 Edwards D J, Kitetele F, Rie A V. Agreement between clinical scoring systems used for the diagnosis of pediatric tuberculosis in the HIV era. Int J Tuberc Lung Dis 2007; 11: 263–269. 7 World Health Organization. Treatment of tuberculosis: guidelines. 4th ed. WHO/HTM/TB/2009.420. Geneva, Switzerland: WHO, 2009. 8 Marais B J, Graham S M, Cotton M F, Beyers N. Diagnostic and
12
13
14
15
16
17
18
19
20
21
22
23
681
management challenges for childhood tuberculosis in the era of HIV. J Infect Dis 2007; 196 (Suppl 1): S76-S85. Marais B J, Gie R P, Schaaf H S, et al. The natural history of childhood intra-thoracic tuberculosis: a critical review of literature from the pre-chemotherapy era. Int J Tuberc Lung Dis 2004; 8: 392. Eamranond P, Jaramillo E. Tuberculosis in children: reassessing the need for improved diagnosis in global control strategies. Int J Tuberc Lung Dis 2001; 5: 594–603. Hatherill M, Hanslo M, Hawkridge T, et al. Structured approaches for the screening and diagnosis of childhood tuberculosis in a high prevalence region of South Africa. Bull World Health Organ 2010; 88: 312–320. Sawry S, Moultrie H, Mahomed N, van Rie A. Diagnostic performance of CXRs for the diagnosis of TB in children initiating HAART. Third South African TB Conference, Durban, South Africa, 12–15 June 2012: Abstract #175. Hatherill M, Hawkridge T, Zar H J, et al. Induced sputum or gastric lavage for community based diagnosis of childhood pulmonary tuberculosis? Arch Dis Child 2009; 94: 195–201. Moore H A, Apolles P, Villiers de P J, Zar H J. Sputum induction for microbiological diagnosis of childhood pulmonary tuberculosis in a community setting. Int J Tuberc Lung Dis 2011; 15: 1185–1190. McIlleron H, Ren Y, Nuttall J, et al. Lopinavir exposure is insufficient in children given double doses of lopinavir/ritonavir during rifampicin-based treatment for tuberculosis. Antiviral Ther 2011; 16: 417–421. South African National Department of Health. Guidelines for the management of HIV-infected children. Pretoria, South Africa: DoH, 2008. South African National Department of Health. South African National Tuberculosis Guidelines. Pretoria, South Africa: DoH, 2008. Bakeera-Kitaka S, Conesa-Botella A, Dhabangi A, et al. Tuberculosis in human immunodefi ciency virus infected Ugandan children starting on antiretroviral therapy. Int J Tuberc Lung Dis 2011; 15: 1082–1086. Martinson N A, Moultrie H, Niekerk R V, et al. HAART and risk of tuberculosis in HIV-infected South African children: a multi-site retrospective cohort. Int J Tuberc Lung Dis 2009; 13: 862–867. Reitz C, Ashraf Coovadia, Ko S, et al. Initial response to protease-inhibitor-based antiretroviral therapy among children less than 2 years of age in South Africa: effect of cotreatment for tuberculosis. J InfectDis 2010; 201: 1121–1131. Kenyon T A, Creek T, Laserson K, et al. Risk factors for transmission of Mycobacterium tuberculosis from HIV-infected tuberculosis patients, Botswana. Int J Tuberc Lung Dis 2002; 6: 843–850. Seddon J A, Hesseling A C, Marais B J, Jordaan A, Victor T, Schaaf H S. The evolving epidemic of drug-resistant tuberculosis among children in Cape Town, South Africa. Int J Tuberc Lung Dis 2012; 16: 928–933. World Health Organization. Automated real-time nucleic acid amplification technology for rapid and simultaneous detection of tuberculosis and rifampicin resistance: Xpert MTB/RIF system for the diagnosis of pulmonary and extrapulmonary TB in adults and children. Policy update. WHO/HTM/TB/2013. 14. Geneva, Switzerland: WHO, 2013.
Microbiological TB investigation in HIV-infected children
i
RESUME C O N T E X T E : Service pe´ diatrique pour le virus de l’immunod´eficience humaine (VIH) dans un hopital ˆ universitaire de Soweto, Afrique du Sud. O B J E C T I F : 1) D´ecrire et comparer les caract´eristiques cliniques, immunologiques et virologiques d’enfants infect´es par le VIH et trait´es pour tuberculose (TB) ; et 2) comparer ceux qui ont eu une recherche microbiologique aux autres et d´ecrire les r´esultats de cette recherche. S C H E´ M A : Analyse r´etrospective d’enfants infect´es par le VIH et la TB ag´ ˆ es de ,15 ans du 1er octobre 2007 au 15 mars 2009. R E´ S U LT A T S : Le traitement anti-tuberculeux a e´ t´e initi´e chez 616/3358 (18%) enfants pendant la p´eriode d’´etude. Les r´esultats des recherches microbiologiques e´ taient disponibles pour 399/616 (65%) enfants et un diagnostic de TB confirm´e par culture pour 49/399 (12%). Le test de sensibilit´e aux m´edicaments a e´ t´e r´ealis´e chez 29/49 (59%) enfants, avec 5/29 (17%)
ayant une r´esistance a` l’isoniazide et trois ayant une TB multir´esistante. Les enfants ag´ ˆ es de .8 ans et de 3–8 ans avaient plus souvent une TB confirm´ee par culture que les enfants ag´ ˆ es de ,3 ans (ORa 9,4 ; IC95% 2,26–39,08 et ORa 6,7 ; IC95% 1,6–27,69, respectivement). Il en allait de mˆeme pour les enfants dont les CD4 e´ taient ,200 cellules/mm3 par rapport a` ceux dont les CD4 e´ taient .500 cellules/mm3 (ORa 3,95 ; IC95% 1,23– 12,72). C O N C L U S I O N : Notre e´ tude d’enfants infect´es par le VIH a montr´e une incidence e´ lev´ee de TB, un faible nombre de TB confirm´ees et un taux e´ lev´e de TB pharmaco-r e´ sistante selon les de´ finitions de l’Organisation Mondiale de la Sante´ . Une plus grande utilisation de tests de TB disponibles et la disponibilit´e de nouveaux tests diagnostiques restent une priorit´e dans les r´egions qui supportent un lourd fardeau de TB et de VIH.
RESUMEN M A R C O D E R E F E R E N C I A: Un consultorio de atencion ´ pedia´trica de la infeccion ´ por el virus de la inmunodeficiencia humana (VIH) en un hospital universitario de Soweto, en Sura´frica. O B J E T I V O: 1) Describir y comparar las caracter´ısticas cl´ınicas, inmunitarias y v´ıricas de los ninos ˜ infectados por el VIH que recib´ıan de manera simulta´nea tratamiento antituberculoso, en quienes se hab´ıa practicado o no un examen microbiologico ´ y 2) analizar los resultados de esta investigacion ´ de la tuberculosis (TB). M E´ T O D O: Se practico ´ un ana´lisis retrospectivo de los ninos ˜ de edad coinfectados por el VIH y la ˜ de ,15 anos TB entre el 18 de octubre del 2007 y el 15 de marzo del 2009. R E S U LT A D O S: Durante el per´ıodo del estudio se inicio ´ el tratamiento antituberculoso en 616/3358 (18%) ninos ˜ con diagnostico ´ de infeccion ´ por el VIH. Se contaba con investigaciones microbiologicas ´ de la TB en 399/616 (65%) casos y se confirmo´ el diagnostico ´ de TB por cultivo en 49/399 (12%) examinados. Se practicaron
pruebas de sensibilidad a los medicamentos en 29/49 (59%) casos: cinco ninos presentaron resistencia a ˜ isoniazida (17%) y en tres ninos se encontro´ TB ˜ multidrogorresistente. Fue ma´s frecuente confirmar el diagnostico ´ de TB por cultivo en los ninos ˜ de ˜ de .8 anos edad (ORa 9,4; IC95% 2,26–39,08) y en los ninos ˜ del grupo entre 3 y 8 anos ˜ (ORa 6,7; IC95% 1,60–27,69) que en los ninos ˜ de edad y tambi´en en los ˜ de ,3 anos ninos con un recuento de c´elulas CD4 ,200 c´elulas/ ˜ que presentaban mm3, al compararlos con los ninos ˜ .500 c´elulas/mm3 (ORa 3,95; IC95% 1,23–12,72). ´ N: En el presente estudio, los ni nos CONCLUSIO ˜ infectados por el VIH exhibieron una alta tasa de TB, con baja proporcion ´ de confirmacion ´ del diagnostico ´ y un alto ´ındice de TB farmacorresistente, segun ´ las definiciones de la Organizacion ´ Mundial de la Salud. Obtener una mayor utilizaci on ´ de las pruebas diagnosticas ´ de la TB que existen y la puesta al alcance de nuevas pruebas diagnosticas ´ sigue siendo una prioridad en los entornos con una alta carga de morbilidad por coinfeccion ´ por el VIH y la TB.
Microbiological investigation for tuberculosis among HIV-infected children in Soweto, South Africa L. Fairlie,* E. Muchiri,* C. N. Beylis,† T. Meyers,‡ H. Moultrie* *Wits Reproductive Health & HIV Institute, School of Clinical Medicine, †National Health Laboratory Service, Mycobacteriology Referral Laboratory & Department of Clinical Microbiology & Infectious Diseases, and ‡Harriet Shezi Children’s Clinic, Department of Paediatrics, Chris Hani Baragwanath Hospital, Faculty of Health Sciences, University of the Witwatersrand, Cape Town, South Africa SUMMARY S E T T I N G : A paediatric human immunodeficiency virus (HIV) clinic in an academic hospital in Soweto, South Africa. O B J E C T I V E S : 1) To describe and compare the clinical, immunological and virological characteristics of HIVinfected children co-treated for tuberculosis (TB), and 2) to compare those investigated microbiologically with those who were not, with a description of the results of the microbiological TB investigation. D E S I G N : Retrospective analysis of TB-HIV-infected children aged ,15 years treated for TB between 1 October 2007 and 15 March 2009. R E S U LT S : Anti-tuberculosis treatment was initiated in 616/3358 (18%) children during the study period. Microbiological TB investigation results were available for 399/616 (65%), among whom culture-confirmed TB was diagnosed in 49 (12%). Drug susceptibility testing
was performed in 29/49 (59%) children: 5/29 (17%) were isoniazid-resistant, and 3 had multidrug-resistant TB. Children aged .8 years and those between 3 and 8 years were more likely to have culture-confirmed TB than those aged ,3 years (aOR 9.4, 95%CI 2.26–39.08 vs. aOR 6.7, 95%CI 1.60–27.69), as were those with CD4 count ,200 cells/mm3 compared to those with .500 cells/mm3 (aOR 3.95, 95%CI 1.23–12.72). C O N C L U S I O N : Our study in HIV-infected children showed a high TB case rate, a low rate of definite TB and a high rate of drug-resistant TB based on World Health Organization case definitions. Increased uptake of available TB tests and availability of new diagnostic tests remains a priority in high TB-HIV burden settings. K E Y W O R D S : paediatric TB-HIV; microbiological TB investigations
TUBERCULOSIS (TB) is the most common opportunistic infection in human immunodeficiency virus (HIV) infected children, with up to 40% receiving anti-tuberculosis treatment at the time of commencing combination antiretroviral therapy (cART).1 In South Africa, an estimated 460 000 (range 410 000– 520 000) children are HIV-infected;2 the TB incidence rate is 993 per 100 000 population,3 and up to 52% of children with culture-confirmed TB are HIVinfected.4 In HIV-infected children, the risk of TB disease is up to four times greater than in those with a CD4% ,15%.5 Potential for both over- and underdiagnosing TB in HIV-infected children is high, with overlapping clinical features between both diseases further complicated by additional co-existing comorbidities common in this population.6 The World Health Organization (WHO) case definitions for TB disease include clinical features and microbiological investigations; radiological features are more important in those cases with negative
microbiological results.7 In children, the tuberculin skin test (TST) and chest X-ray (CXR) may assume greater diagnostic importance than in adults.8,9 However, the TST cannot distinguish tuberculous infection from TB disease, and is frequently negative in immunocompromised and malnourished individuals.10 Wide variability in CXR interpretation between experienced clinicians and poor correlation with culture-confirmed TB suggests that CXR is an unreliable adjunct to TB diagnosis in children.11,12 Microbiological confirmation of TB remains the most accurate method of diagnosis, and two Cape Townbased studies have reported successful microbiological TB investigation using gastric washing and induced sputum procedures in primary health clinics.13,14 These tests are not performed routinely in most clinical settings due to limited staff capacity, skill shortage, low yield and difficulty in conducting them among small children. Co-treatment with anti-tuberculosis treatment and
Correspondence to: Lee Fairlie, Wits Reproductive Health and HIV Institute, Corner Esselen and Klein Streets, Hillbrow 2001, South Africa. Tel: (þ27) 82 780 9997. Fax: (þ27) 11 358 5400. e-mail: [email protected] Article submitted 21 November 2013. Final version accepted 24 January 2014.
Microbiological TB investigation in HIV-infected children
cART in HIV-infected children remains complicated, making accurate TB diagnosis important. The drugdrug interactions between rifampicin (RMP) based anti-tuberculosis treatment and cART, particularly lopinavir/ritonavir, that is inadequately boosted may result in reduced virological suppression rates.1,15 Polypharmacy with anti-tuberculosis treatment and cART may affect adherence and increase the possibility of both ART-associated and anti-tuberculosis drug-associated side effects. Data describing the clinical, immunological and virological associations with TB diagnosis in HIVinfected children are scarce, as are reports on microbiological TB investigation results. This retrospective cross-sectional review sought to describe and compare clinical, immunological and virological characteristics of HIV-infected children co-treated for TB, distinguished by whether or not the children underwent microbiological TB investigations, and to describe the results of the microbiological TB investigations conducted.
METHODS Study subjects and design We conducted a retrospective cross-sectional review of microbiological TB investigations performed among HIV-infected children aged ,15 years attending Harriet Shezi Children’s Clinic (HSCC), Soweto, South Africa, who were treated for TB between 1 October 2007 and 15 March 2009. Standard of care The HSCC is a public out-patient paediatric HIV clinic at the Chris Hani Baragwanath Academic Hospital (CHBAH) in Soweto, South Africa. HIV was diagnosed according to the guidelines of the South African National Department of Health (SANDoH).16 TB screening was conducted at the initial and subsequent clinic visits using symptomatic enquiry and clinical assessment, with further investigations in TB suspects. WHO case definitions were applied as follows: TB suspects were defined as those presenting with signs and symptoms suggestive of TB; TB cases were defined as a definite TB case or where a health care worker has diagnosed TB and has initiated a full course of anti-tuberculosis treatment; and a definite case of TB was defined as a confirmed diagnosis of Mycobacterium tuberculosis using culture or newer methods such as molecular line-probe assay or, in countries where this is not possible, two positive specimens for acid-fast bacilli (AFB).7 Investigations included CXR, TST and microbiological examinations. All microbiological specimens were obtained from the respiratory tract and extrapulmonary sites, including blood, pleural fluid and cerebrospinal fluid (CSF), as clinically indicated. Children diagnosed with TB were prescribed a
677
standard three-drug regimen of RMP, isoniazid (INH) and pyrazinamide, with ethambutol/ethionamide (EMB/ETH) added as a fourth drug for complicated TB, including TB meningitis or disseminated TB.17 cART was initiated according to SANDoH guidelines at the time, which recommended starting cART 2 months after TB treatment initiation or completing anti-tuberculosis treatment before starting cART, if possible, in stable children.16 Laboratory investigations Microbiological investigations for TB included smear microscopy, culture and drug susceptibility testing (DST). Specimens were processed according to WHO guidelines using the N-acetyl-L-cysteine-sodium hydroxide decontamination process: a 0.5 ml portion of the processed, centrifuged pellet was inoculated into BACTECe MGITe (BD, Sparks, MD, USA) tubes and incubated for a total of 6 weeks. Growth detected in automated MGIT instruments was confirmed as AFB-positive using Ziehl-Neelsen stained smear. M. tuberculosis complex was confirmed using the GenoTypew Mycobacterium CM assay (Hain Lifesciences, Nehren, Germany). DST was only performed on clinician request, in accordance with the National TB Programme Guidelines, and was not routine performed during the study period.17 MTC isolates were initially tested against INH and RMP using the MGITe960 proportion method (BD). If resistance was found against either or both drugs, the isolate underwent further DST against streptomycin (SM), EMB, ofloxacin (OFX), kanamycin (KM) and ETH using the MGIT proportion method. Definitions of drug-susceptible and drug-resistant TB were as follows: susceptible M. tuberculosis, susceptibility to both INH and RMP; INH-monoresistant and RMP-monoresistant, resistance to only INH or RMP, respectively; multidrug-resistant TB (MDR-TB), resistance to both INH and RMP; and extensively drug-resistant TB (XDR-TB) if a MDRTB isolate was also resistant to OFX and KM. Data collection Demographic, clinical and laboratory data at the time of initiating anti-tuberculosis treatment, with a 3month window period, were extracted from three databases: demographic and HIV-related data from the HSCC clinical database; microbiological TB data from the National Health Laboratory Service (NHLS) database; and TB treatment data from the CHBAH TB Care Centre (TBCC) ETR.net database, a SANDoH TB data system. The TBCC and HSCC databases were merged using the hospital’s unique patient identifier to include only children treated for TB. Microbiological M. tuberculosis investigation results for children included in the merged data set of children who attended HSCC and were treated for TB were then imported from the NHLS Mycobacteriol-
678
The International Journal of Tuberculosis and Lung Disease
ogy Referral Laboratory data set. The final data set included all children treated for TB, regardless of whether or not they were investigated microbiologically for TB. Manual file reviews and NHLS database searches were conducted by study staff for missing values. Data extracted included age (calculated from date of birth), weight-for-age Z-scores (WAZ scores), height-for-age Z-scores (HAZ scores), WHO clinical staging, HIV log10 viral load and CD4 count. WAZ and HAZ scores were generated using WHO reference ranges. Children’s characteristics at TB treatment initiation were stratified according to whether or not the children were investigated microbiologically for TB and by TB investigation results. Statistical analysis Statistical analysis and data management were performed in STATA version 12.0 (STATA Corporation, College Station, TX, USA). Medians and interquartile ranges (IQRs) were used to summarise the data. v2 tests were used to explore associations between grouped categorical variables of age, anthropometrical measures, CD4 count and percentage, HIV log 10 viral load and the outcome of TB investigations. Age was categorised as ,3, 73–,8 and 78 years, in keeping with differential TB disease risk.8 The Wilcoxon rank sum test and the v2 test were used to compare continuous variables and categorical variables, respectively. Univariable and multivariable logistic models were fitted to identify factors associated with a positive TB investigation outcome among HIV-infected children investigated microbiologically for TB. Multivariate models were developed using step-wise backward elimination using likelihood ratio tests to identify the model with the best fit. Ethics approval This study was approved by the University of the Witwatersrand Human Sciences Research Ethics Committee, Cape Town, South Africa.
RESULTS Of the 3358 HIV-infected children who attended the HSCC at least once during the study period, 616 (18%) were co-treated for TB and were included in the analysis (Figure); 129 were undergoing cART at the time (Table 1). At the initiation of anti-tuberculosis treatment, the median age was 4.2 years (IQR 1.1 to 8.4); 67% of the children had WHO clinical Stage IV disease, the median WAZ score was 2.0 (IQR 1.1 to 3.3) and the median HAZ score 2.4 (IQR 1.5 to 3.3). The median CD4 count was 425 cells/mm3 (IQR 180 to 847), median CD4 percentage 14.4% (IQR 8.4 to 21) and the median log10 HIV viral load 4.6 copies/ml (IQR 3.5 to 5.3). A higher
Figure Description of human immunodeficiency virus infected children attending HSCC between 1 October 2007 and 15 March 2009, stratified by anti-tuberculosis treatment and microbiological investigation. HSCC ¼ Harriet Shezi Children’s Clinic; TB ¼ tuberculosis; DST ¼ drug susceptibility testing; MDRTB ¼ multidrug-resistant TB.
than expected proportion among those investigated microbiologically for TB were male (P ¼ 0.01). There were no other clinically or statistically significant differences in characteristics of children who were and those who were not investigated microbiologically (Table 1). Microbiological investigations were conducted in 399 (65%) of the children co-treated for TB (Figure). Twenty-five (6%) were AFB-positive using ZiehlNeelsen staining. TB was confirmed in 49 (12%) children using mycobacterial culture. The median age of children with culture-confirmed TB was 8.2 years (IQR 5.7 to 10.1) compared to 3.3 years (IQR 0.9 to 7.9) in those investigated microbiologically with negative TB culture results (P 6 0.001) (Table 2). Children with culture-confirmed TB (definite TB) had a median CD4 count of 157 cells/mm3 (IQR 41 to 299) and CD4 percentage of 10% (IQR 3 to 19) compared to 491 cells/mm3 (IQR 239 to 924) and 14% (IQR 8 to 20) in culture-negative children (P ¼ 0.02 and ,0.001, respectively) (Table 2). DST was performed in 29 (59%) children with cultureconfirmed TB. Drug-resistant TB was diagnosed in 5 (17%) children, of whom 3 were MDR-TB. All three MDR-TB strains were resistant to EMB and SM, and a third child’s M. tuberculosis isolate was also resistant to ETH. Two children were INHmonoresistant (Figure).
Microbiological TB investigation in HIV-infected children
Table 1
679
Characteristics of children attending the HSCC treated for TB, 1 October 2007–15 March 2009
Male sex, n (%) Age, years CD4% CD4 count, cells/mm3 Log10 HIV viral load, copies/ml WHO Stage 4, % Weight-for-age Z-scores Height-for-age Z-scores Receiving cART, n (%)
All children treated for TB (n ¼ 616) Median [IQR]
Not investigated microbiologically for TB (n ¼ 217) Median [IQR]
Investigated microbiologically for TB (n ¼ 399) Median [IQR]
P value
339 (55) 4.2 [1.1 to 8.4] 14.4 [8.4 to 21.0] 425 [180 to 847] 4.6 [3.5 to 5.3] 67 2.0 [–1.1 to 3.3], (n ¼ 440) 2.4 [–1.5 to 3.3], (n ¼ 521) 129 (21)
105 (48) 4.2 [1.2 to 8.6] 15.2 [10.0 to 22.0] 389 [164 to 747] 4.6 [3.2 to 5.1] 65 1.9 [–0.9 to 3.2], (n ¼ 146) 2.4 [–1.5 to 3.1], (n ¼ 176) 50 (23)
234 (58) 4.2 [1.0 to 8.3] 13.8 [7.9 to 20.3] 445 [200 to 859] 4.6 [3.7 to 5.4] 68 2.1 [–1.1 to 3.4], (n ¼ 294) 2.4 [–1.5 to 3.3], (n ¼ 345) 79 (20)
0.01 0.23 0.07 0.33 0.21 0.42 0.25 0.43 0.35
HSCC ¼Harriet Shezi Children’s Clinic; TB ¼tuberculosis; IQR ¼interquartile range; HIV ¼ human immunodeficiency virus; WHO ¼ World Health Organization; cART ¼ combined antiretroviral therapy.
Multivariate logistic regression confirmed that confirmation of TB using culture was independently associated with both older age and a lower CD4 count, after adjusting for sex, WAZ and HAZ. Children aged .8 years and those aged between 3 and 8 years were more likely to have cultureconfirmed TB compared to those aged ,3 years (adjusted odds ratio [aOR] 9.4, 95% confidence interval [CI] 2.26–39.08 and aOR 6.7, 95%CI 1.60– 27.69, respectively). Children with a CD4 count ,200 cells/mm3 had increased odds of cultureconfirmed TB compared with those with a CD4 count .500 cells/mm3 (aOR 3.95, 95%CI 1.23– 12.72).
DISCUSSION Our study showed a high TB case rate (18%) in HIVinfected children attending a public HIV treatment clinic, with a low microbiological TB investigation rate and low rate of definite TB that was confirmed using culture or AFB in those investigated microbiologically. Similar high TB-HIV co-treatment rates have been described in other sub-Saharan settings. In Uganda, 17% of HIV-infected children (median age 6.6 years) were co-treated for TB,18 while in a study from four South African hospitals in Johannesburg
and Cape Town, 21% of HIV-infected children (median age 6.3 years) were co-treated for TB; in 23% of the children, a definite TB diagnosis was made (AFB smear n ¼ 30, TB culture n ¼ 24, biopsy n ¼ 11), although the number with culture-confirmed TB was not described.19 TB case rates of 40% by 1 year of follow-up on cART have been described in South African HIV-infected children aged ,2 years.20 Although only 12% of TB cases had cultureconfirmed TB (definite TB cases), it is of concern that only 65% of children classified as a TB case were investigated microbiologically. Apart from male children undergoing more microbiological investigations, there was no other clinical or immunological difference between children who underwent microbiological TB investigations and those who did not. In our study, children investigated microbiologically using culture-confirmed TB were older and more immunologically compromised than those who were culture-negative. The correlation between immunological suppression and confirmed TB in children is likely due to a number of factors: although we do not have data on the source of the specimens, 80% of TB specimens processed at the NHLS are respiratory in origin, and the same is likely in this study. Older children are more likely to be able to expectorate and produce adequate sputum samples. We also expect
Table 2 Characteristics of HIV-infected children co-treated for TB at HSCC from 1 October 2007 to 15 March 2009 stratified by microbiological TB investigation results TB confirmed on culture (n ¼ 49) Median [IQR]
Negative TB culture (n ¼ 350) Median [IQR]
P value
Male sex, n (%) 25 (51) 209 (60) 0.25 Age, years 8.2 [5.7 to 10.1] 3.3 [0.9 to 7.9] ,0.001 CD4% 9.8 [3.0 to 18.7] 14.4 [8.4 to 20.5] 0.02 157 [41 to 299] 491 [239 to 924] ,0.001 CD4, cells/mm3 4.6 [3.4 to 5.2] 4.7 [3.7 to 5.4] 0.35 Log10 HIV viral load, copies/ml WHO Stage 4, % 76 67 0.22 Weight-for-age Z-scores 2.1 [–1.2 to 3.3], (n ¼ 30) 2.1 [–1.1 to 3.4], (n ¼ 264) 0.85 Height-for-age Z-scores 2.4 [–1.4 to 3.2], (n ¼ 40) 2.5 [–1.5 to 3.3], (n ¼ 305) 0.48 HIV ¼ human immunodeficiency virus; TB ¼ tuberculosis; HSCC ¼ Harriet Shezi Children’s Clinic; IQR ¼ interquartile range; WHO ¼ World Health Organization.
680
The International Journal of Tuberculosis and Lung Disease
Table 3 Logistic regression models for factors associated with TB confirmation in HIV-infected children co-treated for TB Characteristic at start of anti-tuberculosis treatment
Unadjusted OR (95%CI)
Unadjusted P value
aOR (95%CI)
Adjusted P value
1 1.42 (0.72–2.21)
0.25
1 1.89 (0.70–5.10)
0.21
1 4.01 (1.49–10.79) 9.39 (3.75–23.47)
,0.001 ,0.001
1 6.65 (1.60–27.69) 9.40 (2.26–39.08)
0.009 0.002
CD4 count, cells/mm3 7500 200–500 ,200
1 2.69 (1.02–7.10) 6.81 (2.75–16.8)
0.04 ,0.001
1 0.65 (0.16–2.54) 3.95 (1.23–12.72)
0.53 0.02
CD4 % 725 ,25
1 1.35 (0.46–4.00)
0.59
1 0.79 (0.33–1.90)
0.60
1 1.51 (0.75–3.00)
0.24
Sex Male Female Age, years ,3 3–8 .8
Viral load, copies/ml ,400 7400 WHO Stage Stage 3 Stage 4
On cART at start of anti-tuberculosis treatment Yes 1 No 0.95 (0.45–2.01)
0.91
Weight-for-age Z-score
1.04 (0.84–1.29)
0.69
0.62 (0.38–1.02)
0.06
Height-for-age Z-scores
1.12 (0.93–1.35)
0.25
1.78 (1.09–2.91)
0.02
TB ¼ tuberculosis; HIV ¼ human immunodeficiency virus; OR ¼ odds ratio; CI ¼ confidence interval; aOR ¼ adjusted OR; WHO ¼ World Health Organization; cART ¼combined antiretroviral therapy.
declining immune function with advancing age and disease progression in HIV-infected children. Adult studies, however, have shown lower rates of microbiological TB diagnosis with severe immunodeficiency.21 We also found that children who had cultureconfirmed TB and underwent DST had high rates of INH-resistant TB (17%), with 10% having MDRTB. These findings are consistent with studies in Cape Town and Johannesburg that have shown INHresistant TB in up to 15% and MDR-TB in 9% of children with culture-confirmed TB.4,22 In the light of these findings, it is worrying that more children were not investigated microbiologically for TB before commencing anti-tuberculosis treatment. TB culture is labour-intensive and time-consuming, leading to long turnaround times for results and low sensitivity in children.10 Recent WHO policy updates recommend the use of the Xpertw MTB/RIF assay (Cepheid, Sunnyvale, CA, USA) rather than conventional microscopy, culture and DST as the initial diagnostic test in children presumed to have MDR-TB or HIV-associated TB. However, rollout of the Xpert test has been limited in routine clinic settings.23 The Xpert test relies on the collection of respiratory specimens, which remains a barrier to microbiological investigations in children. Two Cape Town-based studies have shown that
gastric washings and induced sputum for microbiological TB testing are feasible in a primary clinic setting and that gastric washings on two consecutive days may have a diagnostic yield similar to induced sputum. This technique is particularly useful in infants and young children.13,14 Ongoing training of health care workers in specimen collection techniques is required to ensure that these rapid technologies are increasingly utilised in community settings. The study has some limitations. Data were collected retrospectively and were dependent on accurate recording by the attending clinician and those performing data capture. As TB treatment data were not accurately collected in the HSCC database, only children registered for anti-tuberculosis treatment at the CHBAH TBCC were included. This may have led to the exclusion of children initiated on antituberculosis treatment outside the CHBAH, with a possible underestimation of TB cases. Inconsistent recording of patient names or identification of numbers between data sets may have resulted in some children who were not linked in the NHLS data set undergoing microbiological investigations for TB. We were unable to break down the source of microbiological specimens due to the data extraction methods used at the time; however, as 80% of TB specimens received by NHLS are pulmonary specimens, we assumed that the same proportion would
Microbiological TB investigation in HIV-infected children
apply to this study. Due to the large size of the cohort and the quality of the CXR and TST data in the files, data were not confirmed in a file review. Studies have found poor correlation between culture-confirmed TB and CXR findings and wide variability in interpretation between clinicians.12 We did not have sufficiently accurate data to describe TB suspects who were not treated for TB.
9
10
11
CONCLUSIONS Our study showed a high TB case rate in HIV-infected children, with a low rate of microbiological TB investigation and microbiological confirmation. It is worrying to note the high rate of drug-resistant TB among those investigated microbiologically. These findings suggest that efforts to increase the use of currently available tests should be encouraged through staff education and training. There is also a need for new simple, cheap, rapid and reliable TB tests, particularly in resource-limited settings with a high endemic TB burden. Acknowledgements The authors would like to thank the children and their families attending HSCC. Conflict of interest: none declared.
References 1 Frohoff C, Moodley M, Fairlie L, et al. Treatment outcomes among HIV-infected infants and young children following modifications to protease inhibitor-based therapy due to tuberculosis treatment. PLOS ONE 2011; 6: e17273. 2 Joint United Nations Programme on HIV/AIDS. HIV and AIDS estimates 2011. Geneva, Switzerland: UNAIDS, 2011. http:// www.unaidsorg/en/regionscountries/countries/southafrica/. Accessed March 2014. 3 World Health Organization. Global tuberculosis report, 2012. WHO/HTM/TB/2012.6. Geneva, Switzerland: WHO, 2012. 4 Fairlie L, Beylis N C, Reubenson G, Moore D P, Madhi S A. High prevalence of childhood multidrug-resistant tuberculosis in Johannesburg, South Africa: a cross sectional study. BMC Infect Dis 2011; 11: 1. 5 Elenga N, Kouakoussui K A, Bonard D, et al. Diagnosed tuberculosis during the follow-up of a cohort of human immunodeficiency virus-infected children in Abidjan, Cote ˆ d’Ivoire: ANRS 1278 Study. Pediatr Infect Dis J 2005; 24: 1077. 6 Edwards D J, Kitetele F, Rie A V. Agreement between clinical scoring systems used for the diagnosis of pediatric tuberculosis in the HIV era. Int J Tuberc Lung Dis 2007; 11: 263–269. 7 World Health Organization. Treatment of tuberculosis: guidelines. 4th ed. WHO/HTM/TB/2009.420. Geneva, Switzerland: WHO, 2009. 8 Marais B J, Graham S M, Cotton M F, Beyers N. Diagnostic and
12
13
14
15
16
17
18
19
20
21
22
23
681
management challenges for childhood tuberculosis in the era of HIV. J Infect Dis 2007; 196 (Suppl 1): S76-S85. Marais B J, Gie R P, Schaaf H S, et al. The natural history of childhood intra-thoracic tuberculosis: a critical review of literature from the pre-chemotherapy era. Int J Tuberc Lung Dis 2004; 8: 392. Eamranond P, Jaramillo E. Tuberculosis in children: reassessing the need for improved diagnosis in global control strategies. Int J Tuberc Lung Dis 2001; 5: 594–603. Hatherill M, Hanslo M, Hawkridge T, et al. Structured approaches for the screening and diagnosis of childhood tuberculosis in a high prevalence region of South Africa. Bull World Health Organ 2010; 88: 312–320. Sawry S, Moultrie H, Mahomed N, van Rie A. Diagnostic performance of CXRs for the diagnosis of TB in children initiating HAART. Third South African TB Conference, Durban, South Africa, 12–15 June 2012: Abstract #175. Hatherill M, Hawkridge T, Zar H J, et al. Induced sputum or gastric lavage for community based diagnosis of childhood pulmonary tuberculosis? Arch Dis Child 2009; 94: 195–201. Moore H A, Apolles P, Villiers de P J, Zar H J. Sputum induction for microbiological diagnosis of childhood pulmonary tuberculosis in a community setting. Int J Tuberc Lung Dis 2011; 15: 1185–1190. McIlleron H, Ren Y, Nuttall J, et al. Lopinavir exposure is insufficient in children given double doses of lopinavir/ritonavir during rifampicin-based treatment for tuberculosis. Antiviral Ther 2011; 16: 417–421. South African National Department of Health. Guidelines for the management of HIV-infected children. Pretoria, South Africa: DoH, 2008. South African National Department of Health. South African National Tuberculosis Guidelines. Pretoria, South Africa: DoH, 2008. Bakeera-Kitaka S, Conesa-Botella A, Dhabangi A, et al. Tuberculosis in human immunodefi ciency virus infected Ugandan children starting on antiretroviral therapy. Int J Tuberc Lung Dis 2011; 15: 1082–1086. Martinson N A, Moultrie H, Niekerk R V, et al. HAART and risk of tuberculosis in HIV-infected South African children: a multi-site retrospective cohort. Int J Tuberc Lung Dis 2009; 13: 862–867. Reitz C, Ashraf Coovadia, Ko S, et al. Initial response to protease-inhibitor-based antiretroviral therapy among children less than 2 years of age in South Africa: effect of cotreatment for tuberculosis. J InfectDis 2010; 201: 1121–1131. Kenyon T A, Creek T, Laserson K, et al. Risk factors for transmission of Mycobacterium tuberculosis from HIV-infected tuberculosis patients, Botswana. Int J Tuberc Lung Dis 2002; 6: 843–850. Seddon J A, Hesseling A C, Marais B J, Jordaan A, Victor T, Schaaf H S. The evolving epidemic of drug-resistant tuberculosis among children in Cape Town, South Africa. Int J Tuberc Lung Dis 2012; 16: 928–933. World Health Organization. Automated real-time nucleic acid amplification technology for rapid and simultaneous detection of tuberculosis and rifampicin resistance: Xpert MTB/RIF system for the diagnosis of pulmonary and extrapulmonary TB in adults and children. Policy update. WHO/HTM/TB/2013. 14. Geneva, Switzerland: WHO, 2013.
Microbiological TB investigation in HIV-infected children
i
RESUME C O N T E X T E : Service pe´ diatrique pour le virus de l’immunod´eficience humaine (VIH) dans un hopital ˆ universitaire de Soweto, Afrique du Sud. O B J E C T I F : 1) D´ecrire et comparer les caract´eristiques cliniques, immunologiques et virologiques d’enfants infect´es par le VIH et trait´es pour tuberculose (TB) ; et 2) comparer ceux qui ont eu une recherche microbiologique aux autres et d´ecrire les r´esultats de cette recherche. S C H E´ M A : Analyse r´etrospective d’enfants infect´es par le VIH et la TB ag´ ˆ es de ,15 ans du 1er octobre 2007 au 15 mars 2009. R E´ S U LT A T S : Le traitement anti-tuberculeux a e´ t´e initi´e chez 616/3358 (18%) enfants pendant la p´eriode d’´etude. Les r´esultats des recherches microbiologiques e´ taient disponibles pour 399/616 (65%) enfants et un diagnostic de TB confirm´e par culture pour 49/399 (12%). Le test de sensibilit´e aux m´edicaments a e´ t´e r´ealis´e chez 29/49 (59%) enfants, avec 5/29 (17%)
ayant une r´esistance a` l’isoniazide et trois ayant une TB multir´esistante. Les enfants ag´ ˆ es de .8 ans et de 3–8 ans avaient plus souvent une TB confirm´ee par culture que les enfants ag´ ˆ es de ,3 ans (ORa 9,4 ; IC95% 2,26–39,08 et ORa 6,7 ; IC95% 1,6–27,69, respectivement). Il en allait de mˆeme pour les enfants dont les CD4 e´ taient ,200 cellules/mm3 par rapport a` ceux dont les CD4 e´ taient .500 cellules/mm3 (ORa 3,95 ; IC95% 1,23– 12,72). C O N C L U S I O N : Notre e´ tude d’enfants infect´es par le VIH a montr´e une incidence e´ lev´ee de TB, un faible nombre de TB confirm´ees et un taux e´ lev´e de TB pharmaco-r e´ sistante selon les de´ finitions de l’Organisation Mondiale de la Sante´ . Une plus grande utilisation de tests de TB disponibles et la disponibilit´e de nouveaux tests diagnostiques restent une priorit´e dans les r´egions qui supportent un lourd fardeau de TB et de VIH.
RESUMEN M A R C O D E R E F E R E N C I A: Un consultorio de atencion ´ pedia´trica de la infeccion ´ por el virus de la inmunodeficiencia humana (VIH) en un hospital universitario de Soweto, en Sura´frica. O B J E T I V O: 1) Describir y comparar las caracter´ısticas cl´ınicas, inmunitarias y v´ıricas de los ninos ˜ infectados por el VIH que recib´ıan de manera simulta´nea tratamiento antituberculoso, en quienes se hab´ıa practicado o no un examen microbiologico ´ y 2) analizar los resultados de esta investigacion ´ de la tuberculosis (TB). M E´ T O D O: Se practico ´ un ana´lisis retrospectivo de los ninos ˜ de edad coinfectados por el VIH y la ˜ de ,15 anos TB entre el 18 de octubre del 2007 y el 15 de marzo del 2009. R E S U LT A D O S: Durante el per´ıodo del estudio se inicio ´ el tratamiento antituberculoso en 616/3358 (18%) ninos ˜ con diagnostico ´ de infeccion ´ por el VIH. Se contaba con investigaciones microbiologicas ´ de la TB en 399/616 (65%) casos y se confirmo´ el diagnostico ´ de TB por cultivo en 49/399 (12%) examinados. Se practicaron
pruebas de sensibilidad a los medicamentos en 29/49 (59%) casos: cinco ninos presentaron resistencia a ˜ isoniazida (17%) y en tres ninos se encontro´ TB ˜ multidrogorresistente. Fue ma´s frecuente confirmar el diagnostico ´ de TB por cultivo en los ninos ˜ de ˜ de .8 anos edad (ORa 9,4; IC95% 2,26–39,08) y en los ninos ˜ del grupo entre 3 y 8 anos ˜ (ORa 6,7; IC95% 1,60–27,69) que en los ninos ˜ de edad y tambi´en en los ˜ de ,3 anos ninos con un recuento de c´elulas CD4 ,200 c´elulas/ ˜ que presentaban mm3, al compararlos con los ninos ˜ .500 c´elulas/mm3 (ORa 3,95; IC95% 1,23–12,72). ´ N: En el presente estudio, los ni nos CONCLUSIO ˜ infectados por el VIH exhibieron una alta tasa de TB, con baja proporcion ´ de confirmacion ´ del diagnostico ´ y un alto ´ındice de TB farmacorresistente, segun ´ las definiciones de la Organizacion ´ Mundial de la Salud. Obtener una mayor utilizaci on ´ de las pruebas diagnosticas ´ de la TB que existen y la puesta al alcance de nuevas pruebas diagnosticas ´ sigue siendo una prioridad en los entornos con una alta carga de morbilidad por coinfeccion ´ por el VIH y la TB.
