Interaction of tuberculosis risk factors in San Francisco

Since January 1991 we have been conducting a population-based study of the molecular epidemiology of tuberculosis in San Francisco, and we have carried out a number of analyses that have served to define the dynamics of the disease in the city. Moreover, we have been able to examine the effects of intensified control measures on the incidence and epidemiological patterns of tuberculosis.

Since the study began, organisms from 86.5% of the culture-positive cases have undergone IS6110-based DNA fingerprinting. In addition, 82.2% of the clustered (identical IS6110 fingerprint) cases that had fewer than six IS6110 bands underwent

FIG. 4. Interactions between US-born and foreign-born tuberculosis cases in San Francisco 1993-1994. Data from Chin et al (1998).

secondary genotyping using the polymorphic GC-rich sequence (PGRS) approach. The purpose of secondary PGRS typing is to improve on the specificity of matching by IS6110 in low copy number strains for indicating a transmission link.

Because approximately two-thirds of the incident tuberculosis cases in San Francisco occur in persons born outside the US, we were especially interested in the dynamics of the disease within the foreign-born population and the interactions of the US-born and foreign-born populations. In this analysis we sought to determine, in so far as possible, the amount of transmission of Mycobacterium tuberculosis that occurred in San Francisco (Fig. 4; Chin et al 1998). To develop the study cohort, we excluded all cases from 1991 and 1992. Without prior data, it could not be determined if they had acquired M. tuberculosis during the previous two years. Next, we excluded cases from 1993 and 1994 that had DNA fingerprint matches in 1991 and 1992. Thus, the cohort consisted of persons who had likely developed tuberculosis as a consequence of endogenous reactivation of infection acquired more than two years previously (or possibly acquired from outside San Francisco). Included were all cases from 1995 that had isolates which matched isolates from cases identified from 1993 to 1994, presumably having acquired their infection from the case with the matching isolate. This design, then, enabled us to determine the amount of transmission with progression to disease occurring from endogenously reactivated cases in 1993 to 1995.

This analysis demonstrated that, as we have found in other studies, there is little clustering that involves foreign-born cases, with the possible exception ofpersons born in Mexico (Small et al 1994, Jasmer et al 1998). This indicates a lack of transmission within the foreign-born population and from foreign-born to US-born persons, or vice versa. Thus, presumably, most of the incident cases among the foreign-born persons are the result of reactivation of latent infection.

In the US-born population clustering was more frequent but there was still little evidence of interaction between the two populations. As an indication of the greater frequency of transmission within the US-born population, the ratio of secondary to initial cases in the US-born population was 1 : 5.6 compared with 1 : 121 in the foreign-born population. Of particular note is the finding that all 28 US-born cases in clusters had factors (AIDS, drug use, homelessness) that increased the risk of tuberculosis (Table 1). Neither of the two foreign-born cases in clusters had these risk factors.

This assessment indicates that there are in essence two parallel epidemics of tuberculosis in San Francisco, one (the larger) in the foreign-born population and the second within the indigenous population. These parallel patterns can be characterized in usual sociodemographic terms, but, more importantly, they differ in their pathogenesis. The lack of clustering among the foreign-born population strongly indicates that tuberculosis is resulting from endogenous reactivation of

TABLE 1 Transmission of Mycobacterium tuberculosis in San Francisco

TABLE 1 Transmission of Mycobacterium tuberculosis in San Francisco

Ratio of secondary to

Proportion of source

Number of clustered cases


initial cases***


with riskfactors***


1 :5.6

9/96 (9.4%)



1 :121

2/252 (0.8%)

0 0

Post a comment