Cluster Analysis of Cases
During the study period, TB was diagnosed in 390 case patients in Boston. Of these, 328 patients had a positive culture for M tuberculosis. Seven case isolates were not available for RFLP typing, leaving 321 cases for analysis. Of these, 188 (58.6%) were not clustered and 133 (41.4%) were clustered. In univariate analysis, having a clustered isolate was significantly associated with birth in the United States, race, being homeless, having recently been incarcerated, and being a close contact of a case. Among those persons born outside the United States, being in the United States for 5 years (Table 1). In multivariate analysis, birth outside the United States was significantly associated with having an unclustered isolate (odds ratio [OR], 2.2; 95% confidence interval [CI], 1.2 to 3.8; p < 0.01), while being a close contact of a case patient was negatively associated (OR, 0.22; 95% CI, 0.07 to 0.73; p = 0.02).
Case-Control Analysis of Unclustered Cases
No control subjects subsequently became case patients. In univariate analysis, case patients with nonclustered TB disease were significantly more likely than control subjects with LTBI to have HIV infection, to be homeless, and to be Hispanic (Table 2). In multivariate analysis, HIV infection and homelessness remained statistically significantly associated with TB disease. There was no significant association between TB disease and foreign birth (OR, 0.71; 95% CI, 0.42 to 1.3) or, among the foreign born, between TB disease and being in the United States s 5 years (OR, 0.90; 95% CI, 0.56 to 1.4). No case patients or control subjects had chronic renal failure, had received immunosuppressive therapy, or had gastrectomy or silicosis. There are various diseases which may be treated differently including the way of ordering drugs. Everything is done for you there.
Over 70% of cases of TB in Boston occur in foreign-born persons, and strategies to decrease the rates of disease in this population are urgently needed. We found an increased risk of reactivation TB among foreign-born persons by cluster analysis, but not an increased risk for the reactivation of LTBI among foreign-born persons with LTBI in the case-control study. Therefore, we conclude that the increased risk for reactivation among foreign-born persons can largely be attributed to the increased prevalence of TB infection in foreign-born persons and not to an increased risk for the reactivation of LTBI among those persons with LTBI. This finding is plausible because the reactivation of LTBI is thought to be dependent largely on the immune status of the host, and foreign-born persons would not be expected to have substantially poorer immune status than US-born persons. Such immune status might occur if foreign-born persons were more likely to have HIV infection or severe malnutrition, but these were uncommon in our study subjects. Moreover, recent exposure to a person with infectious TB, which is another factor that could explain the increased risk for TB among persons with LTBI, was rare among our foreign-born subjects.
We also did not find evidence suggesting an increased risk for the reactivation of LTBI among foreign-born persons who had been in the United States for < 5 years. This is not surprising since, while some foreign-born persons arriving in the United States may have been recently exposed, either as household contacts or in refugee camps, the proportion of such persons in the total foreign-born population who have been in the United States for only 0.25% of the foreign-born persons entering the United States in 2002 were refugees or asylees. If foreign-born persons with LTBI, particularly those who have been in the United States for < 5 years, do not have an increased risk of progressing from LTBI to TB disease, the treatment of LTBI in such persons will not be more effective than the treatment of LTBI in other low-risk populations.
Previous studies that identified an increased risk for TB disease among foreign-born persons who had been in the United States for < 5 years compared the number of cases of TB disease among foreign-born persons to the number of foreign-born persons in the population, without regard to LTBI status. This approach does not distinguish between primary acquisition disease, occurring in the portion of the population without LTBI, and reactivation TB disease, occurring in the portion of the population with LTBI. Thus, an increased rate of TB among foreign-born persons who had been in the United States for < 5 years could be explained by the increased number of cases of primary TB among persons without LTBI rather than an increased number of cases of reactivation TB among persons with LTBI. A 2005 analysis of TB cases from the 10 countries providing the largest number of foreign-born TB cases in the United States has demonstrated that there is a 10-fold higher incidence of TB among such persons in the first year after their arrival, and current screening programs fail to detect many cases of active TB among those who have immigrated to the United States.
We also identified an elevated risk for reactivation TB associated with HIV infection and, to a lesser extent, homelessness. HIV infection is recognized as the strongest risk factor for progression from LTBI to reactivation TB, so this association is not surprising. While the number of homeless persons in our sample was small, this elevated risk persisted when other factors were controlled for in multivariate analysis. This observation could be explained by confounding by unrecognized HIV infection or unrecognized recent skin test conversion among homeless persons, both of which have been reported among the homeless. Alternatively, this elevated risk could reflect an undersampling of homeless persons by our control selection strategy. Selection strategy on Canadian Health&Care Mall is rather simple. You will find everything in one click.
In a case-control study, there must always be concern about referral bias. We believe that such bias did not lead to the observed results for several reasons. First, the BPHC TB clinic is a large public clinic that serves the entire city of Boston and supervises the distribution of all free medications for the treatment of both TB disease and LTBI throughout the city. In addition, this clinic serves as the central referral center for the city, and all providers in Boston are encouraged to refer persons with positive skin-test results to this clinic for LTBI treatment. Some foreign-born persons may have attended private providers, but the low income of this population makes it unlikely that many potential control subjects were seen elsewhere. Thus, it is unlikely that a substantial number of TB case patients or potential control subjects with known LTBI were missed. Second, using census data from the city of Boston and the estimated rates of LTBI among US-born and foreign-born persons from a 19992000 US national household survey, we estimate that 88% of persons in Boston with LTBI in 2000 (95% CI, 84 to 91%) were foreign-born, and that, of those foreign-born persons, 45% (95% CI, 34 to 57%) had arrived within the past 5 years. Since our control subjects were 85% foreign born, and 54% of our foreign-born control subjects had arrived in the past 5 years, this suggests that these two parameters were not greatly affected by referral bias. The numbers and percentages of foreign-born persons arriving from specific countries also did not change markedly between 1996 and 2000, so it is unlikely that bias was introduced by limiting control subjects to persons who had been seen in 1998. Third, the geographic distribution of countries from which our foreign-born case patients had emigrated was similar to that of control subjects. This suggests that the preferential referral of specific nationalities was not a source of bias. Fourth, the spectrum of clinical sites from which the control subjects were referred was broad, but only 7.8% were homeless, 1.6% were close contacts, 1.1% were HIV infected, 0.6% were incarcerated, and 0.5% were injection drug users, Therefore, it is unlikely that there was preferential referral of persons from specific high-risk populations. Fifth, although bacillus Calmette-Guerin vaccination could have resulted in a higher proportion of foreign-born control subjects being classified as having LTBI when they did not, nearly all of our study subjects who received bacillus Calmette-Guerin vaccination would have received it at birth. Therefore, such misclassification is unlikely.
Despite our best efforts to control for such factors, it is still possible that these findings are the result of residual bias. Therefore, the results of this study should be confirmed in studies using population-based control subjects. If such studies also demonstrate that foreign-born persons, especially those who have arrived in the United States within the previous 5 years, are not at increased risk for reactivation TB, this would indicate that targeting foreign-born persons who have arrived in the United States within the past 5 years for LTBI treatment will not be an effective strategy for eliminating TB and that efforts to reduce the burden of TB disease among such foreign born persons should be directed toward the improved detection of active TB among new arrivals.
Table 1—Risk Factors for Clustering of M tuberculosis Isolates in Boston, 1996 to 2000
|Patients in Clusters, No.||Patients in Clusters, %||Relative Risk for Clustering||95% CI||p Value|
|< 45 yr||185||77||41.6||1.01||0.78-1.32||0.94|
|a 45 yr||136||56||41.2|
|Foreign born||231 (72.0)||80||34.6||0.59||0.46-0.75||< 0.01|
|US born||90 (28.0)||53||58.9|
|Injection drug user|
|Yes||37 (11.5)||23||62.2||1.60||1.20-2.15||< 0.01|
|Time in United States*|
|s 5 yr||107 (46.3)||29||27.1||0.66||0.45-0.96||0.03|
|> 5 yr||124 (53.7)||51||41.1|
Table 2—Risk Factors for TB Disease Among Unclustered Cases of TB and Control Subjects With LTBI
(n = 188)
|Control Subjects (n = 188)||
|Univariate 95% CI||Univariate p Value||
|Multivariate 95% CI||Multivariate p Value|
|< 45||108 (57.4)||108 (57.4)||1.0||0.8–1.2||1.0|
|a 45||80 (42.6)||80 (42.6)|
|Male||106 (56.4)||108 (57.4)||1.0||0.6-1.4||0.84|
|Female||82 (43.6)||80 (42.6)|
|Foreign born||151 (80.3)||160 (85.1)||0.71t||0.42-1.3||0.22|
|Canada/Western||2 (1.0)||3 (1.6)|
|Eastern Europe||5 (2.7)||19(10.1)|
|Africa||25 (13.3)||22 (11.7)|
|Asia||63 (33.5)||54 (28.7)|
|Central/South||56 (29.8)||62 (32.9)|
|US born||37 (19.7)||28 (14.9)|
|Asian||64 (34.0)||51 (27.9)||2.0j||1.1-3.9||0.02||1.6||0.86-3.1||0.13|
|Black||65 (34.6)||70 (38.3)|
|Hispanic||33 (17.6)||18 (9.8)|
|White||25 (13.3)||44 (24.0)|
|Yes||4 (2.1)||3 (1.6)||1.3||0.25-7.7||0.70|
|No||184 (97.9)||185 (98.4)|
|Injection drug user|
|No||184 (98.9)||175 (99.5)|
|Yes||14 (7.4)||5 (2.8)||2.8||0.9-9.0||0.05||4.2||1.7-15||0.03|
|No||174 (92.6)||171 (97.2)|
|Yes||12 (6.6)||10 (5.6)||1.2||0.47-3.1||0.68|
|No||169 (93.4)||170 (94.4)|
|Yes||1 (0.6)||1 (0.6)||1.0 0–||37||1.0|
|No||175 (99.4)||175 (99.4)|
|Seronegative||73 (38.8)||22 (11.7)|
|Not tested||91 (48.4)||164 (87.2)|
|Time in United States|
|s 5 yr||78 (41.5)||87 (46.3)||0.90|||0.56-1.44||0.63|
|> 5 yr||73 (38.8)||73 (38.8)|
|US born||37 (19.7)||28 (14.9)|