To the Editors:
Multidrug-resistant tuberculosis (MDR-TB) is defined as TB resistant to at least isoniazid and rifampicin and globally accounts for approximately 500,000 new cases and 150,000 deaths each year [1]. In the latest nationwide anti-TB drug resistance survey carried out in Colombia (2005), MDR-TB accounted for 2.4% of new cases (initial MDR-TB) and 31.4% of previously treated cases [2]. Colombia has 32 departments, 11,000 TB cases per year and three laboratories to perform drug susceptibility testing (DST). Valle del Cauca is the third most populated department and the one with the highest TB caseloads in Colombia, where previous studies had reported 6% of initial MDR-TB [3], suggesting that the problem may be greater than in the rest of the country.
A public–private initiative (PPI) led by the Secretariat of Public Health (SPH) of Valle del Cauca and the Centro Internacional de Entrenamiento e Investigaciones Médicas (CIDEIM) was established to assess the anti-TB drug-resistance situation in this department between January 2007 and December 2008, through operational research. The cities were weighted based on elevated rates of TB: Cali and Buenaventura, which represent 70% of the total caseload in the department (1,700 cases per year). This study was approved by the Ethics Committee of CIDEIM.
Included cases were the following: smear-positive pulmonary TB, previously-untreated and/or previously-treated (including relapse, reinfection and treatment failures or defaulters), residing in the city during the last 12 months, regardless of age. Both public and private health facilities in Cali and Buenaventura were invited to submit fresh sputum samples or isolates of acid-fast bacilli of these cases. In Cali, the local TB programme gave priority to patients less than 16 yrs of age and those from public primary health facilities. In Buenaventura, priority was given to all new cases with positive cultures also from public primary health facilities. 11 other smaller cities were also included; only previously treated cases from public health facilities were sampled. For all patients, TB medication history was obtained from their TB treatment cards. DSTs were performed in the BSL-III laboratory of CIDEIM, using the agar proportion method [4] with Middlebrook 7H10 medium with the following first-line drug (FLD) concentrations: isoniazid (INH) 0.2 and 1.0 μg·mL−1, rifampicin (RMP) 1.0 μg·mL−1, streptomycin (SM) 2.0 μg·mL−1 and ethambutol (EMB) 5.0 μg·mL−1. Pyrazinamide (PZA) resistance was determined by pyrazinamidase assay. External quality control was performed by the Mycobacteriology Laboratory of National Jewish Health (Denver, Colorado, USA) with concordance of 100% for INH, RMP and SM; 83% for PZA and 54% for EMB. MDR isolates were genotyped using spoligotyping [5]. The spoligopatterns obtained were independently analysed by two readers and compared with the SITVIT2 database (Pasteur Institute of Guadeloupe, Les Abymes, Guadeloupe, France). Statistical analysis was conducted with Stata version 9.0 (StataCorp LP, College Station, TX, USA), Chi-squared and Fisher's exact tests were used to determine differences with 5% significance level. A confidence interval of 95% was calculated to proportions.
During this PPI, 299 patients with culture-positive pulmonary TB were included, of which 259 (87%) were from public health facilities. The median age was 34 yrs (range 4–97 yrs), 179 (60.0%) of patients were male. Eight (2.7%) cases were less than 16 yrs of age. Of the 299 TB patients, 182 (60.8%) were new TB cases, the remaining 117 (39.2%) were previously treated cases. Approximately half of the cases (n=142) were from Buenaventura. In this city, 97% of the cases came from public health facilities, of which, 75% belonged to primary level of care.
The patterns of resistance to the different drugs for all previously-untreated and -treated cases are presented in table 1. Of the strains isolated from the 182 new TB cases, 53 (29.1%) were resistant to any FLD, whereas 76 (65%) of the previously-treated cases had resistance to any drug. Among new TB cases, 8.2% (CI 95% 4–12%) were monoresistant to INH. Of those, three isolates were resistant at 0.2 μg·mL−1 but susceptible at 1.0 μg·mL−1 (two cases from Buenaventura and one from Cali). This INH-concentration-related resistance pattern was also found in two treated cases: one monoresistant and one MDR-TB from Buenaventura. The overall resistance to INH was 24.7% (95% CI 17–19), of which 40% were MDR strains. Initial MDR-TB accounted for 9.9% (95% CI 6–14), with Buenaventura alone accounting for 9.1% (95% CI 4.8–15). Seven of the 18 primary MDR-TB cases were resistant to all FLDs. Approximately half (47%) of the previously treated cases were MDR-TB (95% CI 45–63) and this pattern represented around a quarter of all the isolates (24.4%).
From the total of 73 MDR-TB isolates, DNA could be extracted from 68 (93.1%). 10 genotypes could be identified: LAM9, Beijing, H1, U, LAM3, LAM2, X1, S, T2-Uganda, LAM5. Beijing family strain (spoligotype international type (SIT) 190) represented a cluster of 22 cases (20 from Buenaventura) and orphan genotypes accounted for four isolates. The genotype H1 was identified only among previously-treated cases. There was no statistically significant difference between isolates from previously untreated and previously treated cases (Fisher's exact test p=0.15). There was a high frequency of Beijing strains among MDR-TB isolates. Eight of 20 Beijing strains were involved in a cluster of new cases, suggesting recent ongoing transmission.
In this study, the resistance patterns for new TB cases revealed that 40% of INH-resistant isolates were also RMP-resistant and the remaining isolates were at risk for further amplification of drug-resistance given the monotherapy with RMP during the continuation phase of the standard regimen [6]. The rate (9.1%) of initial MDR-TB found in Buenaventura is the highest reported in Colombia. This operational research study used a convenience sample. In Cali, the sample accounted for only 3.5% of the total number of new TB cases in that period and the MDR-TB proportion could have been overestimated. However, in Buenaventura this sample accounted for 28.2% of new TB cases, and these do not necessarily have risk factors other than residing in this very high TB burden setting.
Public–private mix for TB care and control is among the core components of the STOP TB Strategy [7]. Most of these experiences have focused on engagement with the programmatic management of drug-resistant TB and with advocacy, communication and social mobilisation activities [8].
Based on this initiative, the SPH of Valle del Cauca and the public reference hospital of the region established a “Drug-Resistant (DR)-TB Committee” to recommend the proper management for such cases (from both public and private health facilities), which was a significant achievement, given that Colombia does not have universal access to healthcare. To our knowledge, this is the first report of a PPI focusing on case finding of DR-TB in Latin-America. We detected primary MDR-TB cases which otherwise would not have been found, thereby having significant impact on TB control and treatment outcome. Furthermore, national tuberculosis programmes in many high-TB-burden countries do not have the resources for detecting and treating MDR-TB patients as in Colombia, so public–private mix strategies could be efficient ways to build human and laboratory capacity.
In summary, the rate of initial MDR-TB may be greater than 6% in Buenaventura. This city has the highest incidence rate of TB in Colombia (72 cases per 100,000 population in 2008) and high rates of default (9–14% in 2007–2008) and of transmission of MDR-TB Beijing strains. Therefore, it is imperative to reduce default rates and to increase cure rates of both drug-susceptible and drug-resistant TB cases, through the following strategies: 1) building laboratory capacity for performing DST to both FLDs and second-line drugs (SLDs) (including rapid molecular testing for the detection INH- and RMP-resistance); 2) securing uninterrupted access to both FLDs and SLDs; and 3) assuring appropriate medical follow-up for the early detection of treatment failure and for ascertaining definitive cure. “Engaging all healthcare providers” could be an efficient TB control strategy to overcome these challenges.
Acknowledgments
The authors would like to thank E.L. Hernández, M. Pérez and L.M. Nieto, (Centro Internacional de Entrenamiento e Investigaciones Médicas (CIDEIM), Cali, Colombia) and L. Heifets and J. Graham, National Jewish Medical and Research Center (Denver, CO, USA).
- ©ERS 2012
