Epidemiology and diagnosis of MDR-TB in children H Simon Schaaf Desmond Tutu TB Centre Department of Paediatrics and Child Health, Stellenbosch University, and Tygerberg Children s Hospital (TCH)
Definitions in DR-TB Mono-drug resistance: Resistance to single drug e.g. HMR or RMR-TB Poly-drug resistance: Resistance to 2 or more drugs, but not to both H and R, e.g. H&S or H&E resistance MDR-TB: Resistance to H & R (+/- other drugs) Pre-XDR-TB: MDR plus a fluoroquinolone (FQN) or 2 nd -line injectable drug (SLID) resistance XDR-TB: MDR & resistance to a FQN & a SLID
Epidemiology of DR-TB in children Estimated MDR-TB in children is 25,000-32,000 new cases per year world-wide, but there is little surveillance data on MDR-TB (i.e. resistance to INH and rifampicin) in children More important: 2 million children are infected with MDR M.tb strains a possible future risk
An estimated 850 000 children developed TB in 2014. HMR-TB = 58 000; MDR-TB = 25 000; XDR-TB = 1200 Lancet Infect Dis 2016 An estimated 67 million children were infected with M. tb HMR infection = 5 million; MDR infection = 2 million; XDR infection = 100 000 Africa and southeast Asia have the highest numbers of children with TB, but the WHO Eastern Mediterranean region, European region, and Western Pacific region also contribute substantially to the burden of DR-TB because of their much higher proportions of resistance. Interpretation Far more DR-TB occurs in children than is diagnosed, and there is a large pool of DR-TB infection.
Proportion of incident TB in children by drug-resistance status, 2014 A median 6 9% (IQR 6 6 7 1) of incident TB in children was HMR and 2 9% (2 7 3 1) was MDR. Of MDR-TB in children a median 4 7% (IQR 4 3 5 1) was XDR. Patterns of drug resistance varied strongly both between and within WHO regions. In the European region, by contrast with all other regions, the proportion of cases that was MDR was higher than the proportion that was HMR. The proportion of children with MDR-TB who had 2nd-line drug resistance seemed lowest in the African and Western Pacific region Global burden of drug-resistant tuberculosis in children: a mathematical modelling study Dodd PJ, et al. Lancet Infect Dis 2016
Trends of DR in childhood TB: WC % Years & Total number of children DST done
How does TB in children differ from adults? May have acute or chronic onset of TB infants often have acute onset with rapid progression of disease; this may lead to diagnosis not being considered Primary TB is often paucibacillary disease (rarely have cavities) therefore more difficult to bacteriologically confirm TB diagnosis and drug resistance Extrapulmonary TB is more common up to 30% Young children are at high risk of disseminated TB including TB meningitis and miliary TB. Penetration of drugs into CSF and various tissues is important Almost all new/transmitted drug resistance
Diagnosis : MDR-TB in children DR-TB is a microbiological diagnosis In children often difficult (paucibacillary TB): Confirmed if DR M. tuberculosis strain is isolated from a child Probable (or presumed) DR-TB if known contact with an infectious adult DR-TB case (>78-90% concordance in several studies of children) Suspect DR-TB if: a child gets worse on Rx, failing adherent Rx TB in a child contact where the adult source case is a treatment failure with unknown DST, a retreatment case or source case died of TB during adherent Rx
Conventional approach to TB diagnosis A constellation of the following: History of chronic symptoms History of TB contact with an infectious TB case Clinical examination (incl. growth assessment) Tuberculin skin testing (or IGRAs) Chest radiography Bacteriological confirmation Histology (especially EPTB) HIV testing (high prevalence areas or patients at risk) (Scoring systems and diagnostic algorithms) Response to anti-tb treatment
What is NOT helpful in DR-TB diagnosis? Clinical symptoms and signs, chest radiography or other imaging these are similar/same in DS- TB and DR-TB HIV test - although MDR-TB may be more common, it does not indicate resistance Histology without microbiological tests Initial response to anti-tb treatment even MDR-TB cases may respond to first-line therapy but treatment failure and relapse common
What may be helpful in DR-TB diagnosis? History: TB source case with DR-TB? - Ask/look for source case s microbiology/dst results! - >1 source case? May have different DST results Previous TB episode? - MDR-TB more common also in children usually missed diagnosis because DST was not known - Relapse or reinfection? - Treatment adherence in recent episode?
What may be helpful in DR-TB diagnosis? TB treatment failure or poor response to treatment (clinically or microbiologically) - adherence to treatment? - paradoxical reaction - reconstitution of immune system (not only in HIV-infected patients)? - alternative diagnosis? - additional pathology (superimposed infection)
Case 1 Young boy with severe kyphosis after failing 2 years of adherent first-line anti-tb treatment for spinal TB. All cultures negative (even on biopsy on follow-up AFB positive, but culture-negative). What would you do?
Case 2 18-month-old girl presented with cough >2 weeks, fever and loss of weight (LOW) Mother denies any contact with TB, but TST 12mm Started on first-line anti-tb treatment May 2016
Case 2 (follow-up) Bacteriology 6/52 later: Induced sputum culturepositive M. tuberculosis - INH and RIF resistant Recalled: aunt brings child denies TB contact, but when asked where mother is - had nervous breakdown after death of sister (caregiver of child) had Pre-XDR-TB (Rx adapted child also pre-xdr- TB) July 2016
MDR-TB in children Is mainly new (transmitted) drug resistance this has been confirmed with DST and DNAfingerprinting Is more difficult to acquire because of the paucibacillary nature of primary disease, but it is possible with cavitary pulmonary disease In our experience DR-TB is not less infectious but may cause less disease than DS-TB Disease in children usually (>90%) develops within 12 months of infection
History of contact A close contact = living in the same household (or in frequent contact) with source case. Sputum smear-positive TB case > infectious than smear-negative/culture-positive source cases, but still infectious! May have >1 source case with different DST results determine MOST LIKELY source case Close contact with source case is found in only 40-70% of cases. May be infected outside of household Often undiagnosed or other TB cases in the family - in infants, may be worthwhile to screen mother Find out about DST of adult source cases! All above is part of RISK assessment for DR-TB
Why do we need microbiological confirmation? In developing countries, the majority of child TB cases are diagnosed at primary care level without microbiological confirmation, as specimens from children are difficult to obtain, children mainly have paucibacillary disease and TB microbiology is often not available In our experience >80% of culture-confirmed cases seen in hospital had the diagnosis of TB made before the culture result was available If DST of the adult source case is known, child contact should be treated according to adult isolate s DST result, as concordance between source case and child s isolates is between 78-90% in different studies
Why do we need bacteriological confirmation? To confirm TB in difficult cases, e.g. uncertain lung pathology, HIV-infected children, extrapulmonary TB To confirm drug resistance if a source case with DR- TB is identified To determine DST in children with unknown source cases, especially if they have poor response to first-line treatment In my view: ALL children who have a positive culture for M. tuberculosis should have a DST done. This may sound logical in developed countries, but it is often not done in developing countries
Which samples for M. tuberculosis culture? Cultures positive in 30-40% of hospital-based cases; lower in community-based studies. Respiratory samples in children: Induced sputum ~ gastric aspirate Tracheal aspirates or BAL lastly NPA FNA biopsies are useful for diagnosis of EPTB Any other body fluid/biopsy of tissue suspected of TB (e.g. CSF, bone/sinovial biopsy)
For optimal microbiological/culture yield Try to collect samples BEFORE starting ANY anti- TB treatment especially if DR-TB suspected. (exception TB meningitis urgent treatment) Collect GOOD QUALITY samples Optimal management of samples after collection and until arrival in the laboratory (TRANSPORT) is extremely important, as this will influence final yield Example: If Xpert MTB/RIF is often positive, but majority of cultures are negative even before starting anti-tb treatment, it could indicate sample management or laboratory problem
Xpert MTB/RIF (GeneXpert) Xpert MTB/RIF replaces smear microscopy; it should not replace mycobacterial culture! Detection threshold 130-150 CFU/ml Xpert MTB/RIF provides only DST for RIF With increasing numbers of RIF-monoresistant TB cases in adults and Xpert MTB/RIF results only, managing child contacts is a problem Xpert can be done on sputum, GAs, FNA of LN, CSF and other! Xpert MTB/RIF Ultra? >10-100 CFU/ml?
2013 WHO Policy Update: Xpert MTB/RIF for children 13 paediatric studies, 2603 children: Sensitivity of Xpert vs. culture for M. tuberculosis detection from respiratory samples: 66% (Specificity 98%) Sensitivity of Xpert vs. culture for detection of rifampicin resistance: 86% (Specificity: 98%) For pulmonary TB in children: Xpert MTB/RIF should be used as the initial diagnostic test in children suspected of having MDR- TB or HIV-associated TB (strong recommendation, very low quality evidence) Courtesy Liz Walters
EPTB in children: Xpert vs. culture Specimen Sensitivity Specificity CSF 80 99 Pleural fluid 44 98 Gastric fluid 84 98 Xpert should be used as the initial diagnostic test for CSF specimens (strong recommendation, very low quality of evidence) Xpert may be used as a replacement test for specific non-respiratory specimens [for] extrapulmonary TB (conditional recommendation, very low quality of evidence WHO 2013
Culture vs. Xpert MTB/RIF Characteristic Culture and DST Xpert MTB/RIF Speed Slow (2-6 weeks) Rapid (2 hours capacity limit) Positive yield 30-40% 18-25% (less than culture) Detection threshold 10-100 cfu/ml 130-150 cfu/ml (Xpert-ULTRA 10-100?) DST Any drug DST/LPA Only RIF DST Specimens Any type Resp specimens, increasing number of other specimens
Automated liquid culture and DST in children Culture & DST slow, but provides best yield (30-70% depending on extent of disease) Detection threshold culture: 10-100 CFU/mL Phenotypic or genotypic DST can be done on cultured isolates, the latter providing more rapid results. If available, culture and DST should be done, with/without Xpert MTB/RIF - more sensitive than Xpert - opportunity to do more DST than RIF only
Line Probe Assays DNA strip-based tests that use PCR and reverse hybridization methods for the rapid detection of mutations associated with drug resistance (commercial kits). Detection threshold version 2 of GenoType MTBDRplus is same as with Xpert directly on samples Confirms M. tuberculosis complex and provides DST results for both INH and RIF (with mutations) Problems: Need laboratory set-up, more skilled staff and has risk of cross contamination Advantage: identifies the mutation conferring resistance to INH. This could assist with chosing the correct drugs, e.g. high-dose INH vs. ethionamide
Line Probe Assay result
Algorithm to manage specimens for drug susceptibility testing in children with suspected tuberculosis according to available laboratory infrastructure. Schaaf et al. Curr Opin Infect Dis 2014, 27:211 219
Second-line DST Pre-XDR and XDR-TB increasing Phenotypic testing (culture-based) still gold standard GenoType MTBDRsl (now version 2) DST for second-line drugs: second-line injectables, fluoroquinolones and ethambutol now WHO approved, but sensitivity lower than culturebased DST therefore still need phenotypic testing if susceptible DST for other second-line drugs rarely routinely available in developing settings
Diagnostic challenges Certainty of MDR-TB in child contacts diagnosis if bacteriologically negative? (80-90% concordance) Interpretation of discordant DST results genotypic (Xpert vs LPA) vs phenotypic (culture-based) DST which is correct? (Rx implications; Low-level RIF resistance?) Gene sequencing identifying mutations and deletions that actually confer resistance to drugs. Also other mechanisms of drug resistance? Developing rapid tests for second-line and new drugs in order to initiate appropriate treatment from the start Could Whole Genome Sequencing, which is rapidly becoming more affordable, be the solution to identifying all resistance?
WHOLE GENOME SEQUENCING AND DRUG RESISTANCE GENE DISCOVERY there has been significant interest and effort in using WGS to characterise drug-resistant M. tuberculosis isolates to shed light on drug resistance. An interesting observation is that there are over 100 genetic loci that seem to be associated with drug resistance. This has led to the suggestion that drug resistance may be more complex than previously realised. Besides the conventional drug resistance gene mutations, there are many other mutations that may collectively contribute to the drug resistance phenotype. Zhang Y, Yew WW. Mechanisms of drug resistance in Mycobacterium tuberculosis: update 2015. Int J Tuberc Lung Dis 2015:19:1276 1289
Rapid Whole-Genome Sequencing of Mycobacterium tuberculosis Isolates Directly from Clinical Samples. Brown AC, et al. J Clin Microbiol 2015;53:2230-7 WGS provides comprehensive data on resistance mutations and strain typing for monitoring transmission, but this has previously been achievable only from cultures of M. tuberculosis. Here we describe a method to capture full M. tuberculosis genomes directly from infected sputum samples, allowing WGS without the requirement of culture. Identification of known resistance mutations within a week of sample receipt offers the prospect for personalized rather than empirical treatment of drugresistant tuberculosis, including the use of antimicrobialsparing regimens, leading to improved outcomes.
A. Trauner et al. Drugs (2014) 74:1063 1072
What if: cultures negative/no cultures done? Recent studies showed the importance of empirical MDR- TB treatment if child has clinical disease combined with exposure to an MDR-TB source case Of child MDR-TB contacts who develop TB, most have MDR-TB, although complete concordance of DST patterns is sometimes low (<50%). Genotypic results, however, showed >75% concordance in source case contact pairs Recent household contact-tracing studies document high rates of TB infection (27 50%) and disease (3 8%) in contacts. These data confirm the importance of contact tracing for case-finding and provide supportive evidence for the supposition that MDR-TB contacts should be managed according to the DST result of their source case Schaaf et al. Curr Opin Infect Dis 2014, 27:211 219
Case 3 11yr-old boy presented with chronic cough, loss of weight and night sweats CXR: adult-type cavitary TB Xpert POSITIVE resistant RIF (x 5 samples!) Culture-positive x 6; LPA INH and RIF susceptible (confirmed on 2 nd specimen)? Phenotypic DST: INH resistant; RIF susceptible Vague history of friend s father having MDR-TB What are you going to do?
Mixed infection? Case 3 Xpert more sensitive for RIF DST? Phenotypic testing incorrectly done? Gene-sequencing: - rpob gene mutation mut L511P (low-level RIF resistance not included in LPA and below RIF DST threshold for phenotypic testing); however, LPA equally sensitive for RIF resistance to Xpert - INH on rechecking LPA katg mutation (vague band mixed infection) but could also be INH resistance due to non katg or inha mutation