Household contact tracing reveals extensive community spread of MDR and XDR TB

Approximately half of adult household contacts of drug resistant TB cases had resistance profiles that differed from the index case of TB, according to a presentation given by Dr. Tony Moll at the 40th Union World Conference on Lung Health in Cancún, Mexico.

“This discrepancy between drug resistance in index cases and their household contacts suggests community spread of MDR- and XDR-TB,” Dr. Tony Moll of the Church of Scotland Hospital in Tugela Ferry, South Africa, told the conference.

The household contact tracing study was conducted in Msinga, a rural sub-district of KwaZulu Natal, where the annual TB case rate is over 1,000/100,000 population and about 75% of TB patients are also infected with HIV.

Since 2005, 852 drug resistant TB cases have been identified. Of these, 43% of these have been multidrug resistant TB (MDR-TB), and 57% extensively resistant (or XDR-TB).

Methods

In a four-year period from 2005-2008, the homes of each index case of multi- and extensively resistant TB cases were visited an average of 2-3 times, and every adult contact within the household was screened for TB (the study excluded children under the age of 13 due to the difficulty in diagnosing pulmonary TB in young children).

A TB symptom history was conducted; sputa were collected on all sputum producers, and a chest X-ray was obtained for each adult contacts with productive cough or other signs and symptoms of tuberculosis. A physician evaluated each TB suspect.

Results

There were 711 index cases. Of these, 306 persons were identified as having MDR-TB: household contact tracing was possible in 255 (83%) of these cases but only 221 (72%) were included in the final analysis (some were excluded because there were no adults available for evaluation or missing data).

The remaining 405 index cases had XDR-TB: 333 (82%) of their households were traced, with 287 (71%) were included in the analysis. In all, 508 households were included in the study.

Nearly two-thirds (64%) of index cases were sputum smear-positive but only 50% had a previous history of TB, indicating transmission of already resistant strains in the 50% presenting with their first episode of TB. Eighty-one percent of index cases died with a median survival of only 32 days.

There were 1059 adult household contacts among the MDR-TB cases: 793 (75%) were screened for respiratory symptoms and 773 (97%) provided sputum sample for culture and drug sensitivity testing.

Among the 1372 household contacts of XDR-TB cases, 973 (71%) were available for screening and 940 (97%) provided sputum sample for culture and drug sensitivity testing (DST). In all, complete data was available for 1713 adult household contacts identified.

A median of 79 days passed from sputum collection of the index case to identification of household contacts, although the time for the actual household contact tracing was usually within a week of the diagnosis of drug resistant TB in the index case. The delay in susceptibility testing prolonged the period of time during which household contacts were exposed to the drug-resistant TB case.

Contact tracing identified cases of TB in 55 (11%) households. Of these, 47 households had only one TB case, 14 had two cases and 1 household had 3 cases.

Although survival was better for household contacts compared to index cases, there was still significant mortality (14% and 52% of MDR-TB XDR and XDR TB household cases, respectively) within the median 506 day follow-up period.

Notably about half of the household contact TB cases had DST results that were discordant from the household’s index case, suggesting possible transmission in other community settings. The spread of resistant TB within the community needs further investigation.

Limitations of the study included unknown HIV status on most household contacts, so there was no control for HIV infection in the comparison of outcomes in the survivors and index patients. The study considered only household contacts and not other casual or close contacts. The investigation provides a minimum estimate of the household contacts, as they were not able to find each household contact.

Discussion

Further studies are needed to examine prevention control at the household level. “This study underlies the need for earlier diagnosis, particularly in this setting where the mortality is so high in the index cases,” stated Dr. Tony Moll, the study investigator.

Today, the TB cure rate is 83% in Msinga and the default rate 0%. There are dedicated tracing teams to investigate households of resistant cases. “The strong commitment by the district managers of the TB programme is a key part of its success,” said Dr. Moll. The study team hopes to extend the household tracing to children and report on those results in a future meeting.