Commentary: Can capture–recapture analysis of epidemiological and molecular data help us understand recent tuberculosis transmission?

Andrew A Vernon and M Scott McNabb

Division of TB Elimination and National Center for HIV, STD and TB Prevention, Centers for Disease Control and Prevention, Atlanta GA, USA.

The report by Iñigo et al.1 describes two interesting approaches, one epidemiological and the other molecular, to elucidate the proportion of tuberculosis (TB) cases attributable to recent transmission. Considering each approach critically helps us recognize what these two approaches can contribute to our traditional understanding of TB transmission.

Traditional approaches have shown that certain groups are at higher risk for TB (e.g. males, prisoners, the homeless, injecting drug users) and that non-traditional settings in which these groups congregate (e.g. prisons, homeless shelters) are important in facilitating recent TB transmission. However, using traditional approaches, the epidemiological linkages suspected in these non-traditional settings have often been difficult to establish with certainty, perhaps because they involved unrecognized or brief (‘casual’) contact. TB transmission resulting from casual contact may be especially important among high-risk groups having immunodeficiency, where rapid progression to disease is facilitated by host susceptibility. Prior literature suggests also that prisons play a major role in the transmission of TB in the greater community,2 and increased human immunodeficiency virus (HIV) prevalence among prisoners may amplify this role.3 These traditional approaches of TB case finding and contact investigation remain the staple of TB prevention and control programmes throughout the world, but may not help to further understand transmission.

Capture–recapture analysis has been used in epidemiology for over 50 years,4 gaining favour (as evidenced by its frequent use) during the past 20 years.5,6 The technique makes four underlying assumptions, which are delineated by Iñigo et al.: (1) cases must have equal ‘catchability’ in each source; (2) cases must be uniquely matched among the various data sources; (3) case ascertainment by the sources must be independent; and (4) the population under study must be ‘closed’. In fact, assumption (1) subsumes assumption (4); assumption (2) must be met if the method is to allow proper matching of cases; and assumption (3) is almost always violated to some degree, since most epidemiological data sources have some positive dependence. The consequence of this is a bias which is conservative, however, since it leads to some degree of underestimation of the proportion in question (e.g. proportion of cases of TB attributable to recent transmission).

Mycobacterium tuberculosis (Mtb) genotyping techniques offer TB prevention and control a new molecular tool,7 which demonstrates hitherto unrecognized linkages among TB cases, especially among those who frequent non-traditional settings.8 But the ability of Mtb genotyping to estimate recent TB transmission has been both asserted9–11 and challenged.12–15 Several population-based factors influence genotypic strain clustering, especially when it is used as a surrogate estimate for recent TB transmission.

These include (1) the period of observation, (2) the characteristics of the population under observation, (3) the background TB prevalence and diversity rates, (4) the completeness of the sampling (or reporting), (5) the diversity of Mtb strains in the population, and (6) the stability of the IS6110 molecular marker (the so-called ‘molecular clock’).16,17

The sampling frame used in this study was 3 years, which is slightly less than the reported 3.2 year half-life of IS6110.16 Thus, the sampling period and IS6110 stability are not likely to have affected these results. The effects of population characteristics may vary. While genotyping may establish the relatedness of strains, it does not definitively establish the timing of the relation. A strain may have undergone several transmissions between two related cases, or many years may have passed between disease onset in the two related cases.17 Moreover, while recent transmission appears often to have occurred when matching genotype patterns are found in urban centres, this may not be the case in rural settings.18 Recent transmission may occur more frequently in domestic populations than in foreign-born populations,19,20 and clustering appears to indicate recent transmission more often in younger than in older persons.14 Thus, there are limits to the generalizability of results obtained in any particular setting, which may be a function of the characteristics of the local population (e.g. proportion foreign-born or migrant, proportion immunodeficient, age of the cases, etc.). The properties and interpretation of clustering statistics may differ substantially between settings. The relationship between the proportion of matched Mtb genotypes and the proportion of TB morbidity due to recent TB transmission is not always 1:1.

Because genotyping requires the availability of a viable isolate of Mtb, the population for which genotyping is performed must per force be a subset of all cases with positive cultures. In turn, cases with positive cultures must be a subset of all cases (culture positive, negative, or not done). The authors do not comment on the completeness of reporting in the study area: if this were very high, then patients for whom genotyping was performed represent a subset of all reported culture positive cases—increasing the positive dependence among the data sources. Patients whose isolate was available for genotyping may be a biased subset of all culture positive patients, and these may in turn be a biased subset of all cases. For example, cases which are more advanced (e.g. smear positive and cavitary) are more likely to be diagnosed, more likely to be culture positive, and more likely to transmit to others. Such cases may provide an overestimate of the proportion of all cases associated with recent transmission. Similarly, patients with a positive culture are more likely not to be extrapulmonary cases, and extrapulmonary patients are less likely to transmit infection to others—and Iñigo and colleagues did find that extrapulmonary disease was significantly less frequent in their molecular dataset.

The genotyping methodology used (restriction fragment length polymorphism [RFLP], with addition of spoligotyping for those strains with fewer than six bands) is well established and specific. Recent developments of a newer polymerase chain reaction (PCR)-based methodology (mycobacterial interspersed repetitive units [MIRU]) suggests that RFLP does have a low rate of false positivity (i.e. the isolates are not really identical).21,22 Spoligotyping, by itself, is less specific, but, when applied as a supplemental test to low-copy number isolates, it does enhance specificity.23

Taken together, these concerns should qualify our uncritical acceptance of molecular strain clustering as an indication of recent TB transmission. At the same time, they help to demonstrate that these new techniques do expand our understanding of TB transmission in the modern era. The paper by Iñigo et al. represents a creative extension of these modern approaches to the epidemiology of TB. It usefully serves to emphasize the importance of the interruption of transmission, which lies at the heart of the control and eventual elimination of TB.


    References
 Top
 References
 
1 Iñigo J, Arce A, Martín-Moreno JM, Herruzo R, Palenque E, Chaves F. Recent transmission of tuberculosis in Madrid: application of capture– recapture analysis to conventional and molecular epidemiology. Int J Epidemiol 2003;32:763–69.[Abstract/Free Full Text]

2 Stead WW. Undetected tuberculosis in prison. Source of infection for community at large. JAMA 1978;240:2544–47.[Abstract]

3 Braithwaite RL, Arriola KR. Male prisoners and HIV prevention: a call for action ignored. Am J Public Health 2003;93:759–63.[Abstract/Free Full Text]

4 Hook EB, Regal RR. Capture–recapture methods in epidemiology: methods and limitations. Epidemiol Rev 1995;17:243–64.[ISI][Medline]

5 Verstraeten T, Baughman AL, Cadwell B, Zanardi L, Haber P, Chen RT. Vaccine Adverse Event Reporting System Team. Enhancing vaccine safety surveillance: a capture–recapture analysis of intussusception after rotavirus vaccination. Am J Epidemiol 2001;154:1006–12.[Abstract/Free Full Text]

6 Tocque K, Bellis MA, Beeching NJ, Davies PD. Capture recapture as a method of determining the completeness of tuberculosis notifications. Commun Dis Public Health 2001;4:141–43.[Medline]

7 Edlin BR, Tokars JI, Grieco MH et al. An outbreak of multidrug resistant tuberculosis among hospitalized patients with the acquired immunodeficiency syndrome. N Engl J Med 1992;326: 1514–21.[Abstract]

8 Cronin W, Golub J, Lathan M et al. Molecular epidemiology of tuberculosis in a low- to moderate-incidence state: are contact investigations enough? EID 2002;8:1271–79.

9 Alland D, Kalkut GE, Moss AR et al. Transmission of tuberculosis in New York City. An analysis by DNA fingerprinting and conventional epidemiologic methods. N Engl J Med 1994;330:1710–16.[Abstract/Free Full Text]

10 Small PM, Hopewell PC, Singh SP et al. The epidemiology of tuberculosis in San Francisco. A population-based study using conventional and molecular methods. N Engl J Med 1994;330:1703–09.[Abstract/Free Full Text]

11 Weis SE, Pogoda JM, Yang Z et al. Transmission dynamics of tuberculosis in Tarrant county, Texas. Am J Respir Crit Care Med 2002;166: 36–42.[Abstract/Free Full Text]

12 Braden CR, Templeton GL, Cave MD et al. Interpretation of restriction fragment length polymorphism analysis of Mycobacterium tuberculosis isolates from a state with a large rural population. J Infect Dis 1997; 175:1446–52.[ISI][Medline]

13 Vynnycky E, Borgdorff MW, van Soolingen D, Fine PE. Annual Mycobacterium tuberculosis infection risk and interpretation of clustering statistics. Emerg Infect Dis 2003;9:176–83.[ISI][Medline]

14 Vynnycky E, Nagelkerke N, Borgdorff MW, van Soolingen D, van Embden JD, Fine PE. The effect of age and study duration on the relationship between ‘clustering’ of DNA fingerprint patterns and the proportion of tuberculosis disease attributable to recent transmission. Epidemiol Infect 2001;126:43–62.[ISI][Medline]

15 Glynn JR, Vynnycky E, Fine PEM. Influence of sampling on estimates of clustering and recent transmission of Mycobacterium tuberculosis derived from DNA fingerprinting techniques. Am J Epidemiol 1999; 149:366–71.[Abstract]

16 de Boer AS, Borgdorff MW, de Haas PE, Nagelkerke NJ, van Embden JD, van Soolingen D. Analysis of rate of change of IS6110 RFLP patterns of Mycobacterium tuberculosis based on serial patient isolates. J Infect Dis 1999;180:1238–44.[CrossRef][ISI][Medline]

17 Lillebaek T. Molecular evidence of endogenous reactivation of M. tuberculosis. (abstract). Int J Tuberc Lung Dis 2001;5(Suppl.1):S35.

18 Braden CR, Templeton GL, Cave MD et al. Interpretation of restriction fragment length polymorphism analysis of Mycobacterium tuberculosis isolates from a state with a large rural population. J Infect Dis 1997; 175:1446–52.[ISI][Medline]

19 Chin DP, DeRiemer K, Small PM et al. Differences in contributing factors to tuberculosis incidence in US-born and foreign-born persons. Am J Respir Crit Care Med 1998;158:1797–803.[Abstract/Free Full Text]

20 Borgdorff MW, Behr MA, Nagelkerke NJ, Hopewell PC, Small PM. Transmission of tuberculosis in San Francisco and its association with immigration and ethnicity. Int J Tuberc Lung Dis 2000;4:287–94.[ISI][Medline]

21 Cowan LS, Mosher L, Diem L, Massey JP, Crawford JT. Variable-number tandem repeat typing of Mycobacterium tuberculosis isolates with low copy numbers of IS6110 by using mycobacterial interspersed repetitive units. J Clin Microbiol 2002;40:1592–602.[Abstract/Free Full Text]

22 Kwara A, Schiro R, Cowan LS et al. Evaluation of the epidemiologic utility of secondary typing methods for differentiation of Mycobacterium tuberculosis isolates. J Clin Microbiol 2003; 41:2683–85.[Abstract/Free Full Text]

23 Cronin WA, Golub JE, Magder LS et al. Epidemiologic usefulness of spoligotyping for secondary typing of Mycobacterium tuberculosis isolates with low copy numbers of IS6110. J Clin Microbiol 2001;39:3709–11.[Abstract/Free Full Text]





This Article
Extract
FREE Full Text (PDF)
Alert me when this article is cited
Alert me if a correction is posted
Services
Email this article to a friend
Similar articles in this journal
Similar articles in ISI Web of Science
Similar articles in PubMed
Alert me to new issues of the journal
Add to My Personal Archive
Download to citation manager
Request Permissions
Google Scholar
Articles by Vernon, A. A
Articles by McNabb, M S.
PubMed
PubMed Citation
Articles by Vernon, A. A
Articles by McNabb, M S.