Southeast Asia has the highest incidence of tuberculosis worldwide—in 1990, 237 per 100 000 compared with 10.1 per 100 000 in native-born U.S. citizens [4]. In Vietnam, efforts to estimate the prevalence of tuberculosis began in 1990. During a 3-year evaluation period, the International Organization for Migration, in collaboration with the Socialist Republic of Vietnam, examined and processed 278 000 potential immigrants to the United States. The processing included sputum and radiologic examinations. On the basis of this assessment, the national tuberculosis rate was estimated to be 505 per 100 000. Among persons identified as having active tuberculosis in this screening, 30% to 40% of smears were positive (Keane V. Personal communication). Because this high-risk group is one of the fastest growing segments of the immigrant population in the United States, it is important to identify persons with asymptomatic tuberculous infection and treat them appropriately [5-8].

Observational studies of refugee screening programs have noted a variant delayed reaction to tuberculin testing in Indochinese persons [9, 10]. This reaction consists of reactivity (that is, induration) of less than 10 mm on evaluation at a standard reading time of 48 to 72 hours that, when reassessed at 6 days, increases in size to 10 mm or greater. We have referred to this phenomenon as delayed tuberculin reactivity. In addition, in a subset of persons with delayed reactivity, the induration increased 5 mm or more between the two readings but remained less than 10 mm. We have designated this reaction as 5-delayed tuberculin reactivity.

We sought to 1) determine the prevalence of variant reactivity [that is, delayed tuberculin reactivity and 5-delayed tuberculin reactivity] in a cross-sectional group of Indochinese refugees and 2) assess whether initial variant reactivity was a predictor of subsequent booster positivity. We also wanted to ascertain whether variant reactors were immunologically similar to persons with positive purified protein derivative (PPD) results by using in vitro lymphocyte blastogenesis. If so, variant reactivity might be used to identify persons with Mycobacterium tuberculosis infection as candidates for preventive therapy.

Methods

Top Methods Results Discussion Author & Article Info References

Study Group and Study Design

The study group was drawn primarily from a cohort of 402 ethnic Dega (an isolated ethnic group from the central highlands region of Vietnam) who were evacuated from the Vietnam-Kampuchea border region and rapidly resettled in three North Carolina counties in November 1992. Approximately half of the cohort (n = 200) were screened in the health departments of Guilford and Wake Counties.

Participants were eligible if they were 18 years of age or older, were culture negative for tuberculosis, and were not receiving therapy for presumptive tuberculosis. Participants also had to be available for two skin testing evaluations by one of two trained observers at 48 to 72 hours and again at 6 days. Unavailability for two readings was the primary reason for exclusion from the final study group, which consisted of 121 persons (60% of those screened).

One of the authors abstracted health department records using a standardized questionnaire for medical history and current medical examination and demographic information. Each participant had a complete physical examination, a laboratory examination (consisting of a complete blood count, malaria smears, syphilis serologic testing, and testing for hepatitis B surface antigen), and stool examination for parasites.

None of the participants had a history of bacille Calmette–Guérin (BCG) immunization or PPD testing or was receiving immunosuppressive drugs. The lack of previous BCG vaccination was confirmed in all participants by history and physical examination; during the latter, two of the authors looked specifically for BCG scarring. All participants had tested negative for the antibody to the human immunodeficiency virus (HIV) by the enzyme-linked immunosorbent assay 4 weeks before skin testing, and none had known risk factors for HIV infection. No participant had received a live viral vaccine within the previous 8 weeks. All of the above factors are known to affect tuberculin testing and may be associated with false-negative or false-positive tuberculin reactivity [11].

All participants had skin testing with 5 tuberculin units of PPD (Connaught Laboratories, Willowdale, Ontario, Canada), which was placed intradermally on the volar surface of the forearm (Mantoux method). One hundred five participants (87%) were also tested with a candida control. Induration was measured at 72 hours and again at 6 days (144 hours). Persons with a negative PPD result (that is, less than 10 mm of induration with no increase in size on the second reading) and persons with variant reactivity had follow-up booster testing at 10 to 12 weeks.

The "booster effect" is defined as a positive standard tuberculin test result ( 10 mm of induration at 48 to 72 hours) that appears on repeated testing after an initial negative (< 10 mm of induration in our study group) or weakly reactive skin test result. The effect may persist for as long as 1 year after initial testing and is considered to be an enhancement of a remotely established hypersensitivity that has deteriorated.

All participants who had positive booster results were carefully followed to ensure that these reactions did not represent a tuberculin conversion. During the 12 months after the initial testing, no participant developed active tuberculosis. One of two authors measured all skin test reactions (that is, mm of induration) using specialized calipers (Miltex Instrument Co., Inc., Lake Success, New York).

Lymphocyte Blastogenesis

Blastogenesis stimulated by PPD is the standard in vitro correlate of a delayed hypersensitivity skin test response. Donors with positive tuberculin skin test results show an increased lymphocyte blastogenesis response to PPD compared with donors with negative skin test results, and, in general, they retain lymphocyte reactivity for at least 19 years [12]. Boosting of PPd-reactive persons after an initial PPD skin test is further associated with a markedly increased lymphocyte response. This assay was used to determine the significance of variant tuberculin reactivity [12-15].

Lymphocyte blastogenic responses to PPD were obtained from a convenience subset of 57 consecutive persons who presented to the refugee screening programs over 3 days in December 1992. Because blood for lymphocyte studies was drawn when initial tuberculin testing was done, the results and distribution of reactions were unknown in advance. Sixtyfour participants were excluded from lymphocyte studies because they were not available for simultaneous tuberculin and lymphocyte testing when the investigators were on site.

Laboratory Analysis

Ten mL of heparinized blood was obtained from each participant before skin testing. Whole blood was shipped at ambient temperatures to Case Western Reserve University by overnight delivery. Preliminary studies showed no deleterious effects of the shipment on cell viability or function. Peripheral blood mononuclear cells were separated by density sedimentation, and blastogenesis was assayed by standard techniques. Initial studies were done using PPD at 10 µg/mL and 100 µg/mL as stimulus; culture lasted 5 and 7 days, respectively. Seven days of incubation and 10 mu/mL of PPD provided maximum discrimination between known tuberculin-positive and tuberculin-negative reactors. This level was selected as the assay condition for further analysis. These laboratory methods are described in more detail elsewhere [13-15].

Cells were washed and plated (1 x 10⁶/mL in triplicate wells) in round-bottomed microtiter plates (Corning 25850, Corning, New York) and were added at a concentration of 100 or 10 µg of PPD per mL (Lederle Laboratories, Pearl River, New York), and plates were incubated at 37 °C in 5% CO₂ for either 5 or 7 days. Cultures were pulsed with ³H-thymidine (1 µ Ci/well; NCI specific activity, 6.7 Ci/mmol) on days 4 or 6, respectively. At the end of the culture period, cells were harvested onto glass fiber filter paper with PHD cell harvester. Filter papers were counted by liquid scintillation. Counts were expressed as counts per minute (cpm).

Chest Radiography

All chest radiographs were evaluated using a standardized data form by a board-certified thoracic radiologist who was blinded to the results of clinical skin testing. Radiographs assessed were single posterior-anterior views that had been obtained in Kampuchea or Vietnam within 6 weeks before immigration. The diagnostic quality of each radiography was graded as adequate or inadequate. Radiographs were then assessed for the presence of old granulomatous disease, pleural abnormalities, apical abnormalities, and pulmonary nodules [16]. Each radiograph was categorized as consistent or inconsistent with remote or active tuberculous infection.

Data Evaluation and Statistical Analysis

The major dependent variable was skin test reactivity; we also analyzed the following covariates: age, sex, chest radiograph results, presence of anemia, presence of malarial infection, and hepatitis B carrier state. For categorical variables such as sex, we analyzed the association with skin testing status using the chi-square test; the level of significance was 0.05. The Student t-test, one-way analysis of variance, and nonparametric tests were used for continuous variables as appropriate. Where indicated, analysis of variance was followed by Scheffe pairwise comparisons. Linear regression, Pearson correlation coefficients, and Spearman rank-correlation coefficients were used to assess the association between skin test induration and lymphocyte reactivity results. Because lymphocyte blastogenic responses were not normally distributed, we used medians and interquartile ranges (25th and 75th percentiles) to describe these distributions. Similarly, we used box-and-whisker plots to depict these distributions (see the legend of Figure 1 for a specific explanation). Data were managed and analyzed using both EPI Info [17] and Stata statistical software [18].

View larger version (16K): [in this window] [in a new window]   Figure 1. Lymphocyte blastogenic response by reactivity category. The graph is a box-and-whisker plot. The line in the middle of each box represents the median; the box extends from the 25th to 75th percentile (interquartile range); and the lines emerging from the box extend to three quarters of the interquartile range, rolled back to where data are present. The box width is proportional to the number of observations. PPD+ = positive purified protein derivative result; PPDV equals variant purified protein derivative result; PPD– = negative purified protein derivative result.

Results

Top Methods Results Discussion Author & Article Info References

Ninety-five of the 121 study participants (78%) were men, and 26 (22%) were women. The mean age was 35 years (range, 19 to 67 years), and 89 participants (75%) were younger than 40 years of age. At 72 hours, PPD skin test results showed that 35 participants (29%) were tuberculin positive, 54 (45%) were tuberculin negative, and 32 (26%) had variant reactivity (26 delayed tuberculin reactors and six 5-delayed tuberculin reactors). Because no obvious difference was seen between the two variant groups, we combined these groups into one.

One hundred five participants (87%) were simultaneously evaluated for anergy with candida skin testing. All delayed tuberculin reactors had a positive response to candida testing at 48 hours. Two of the PPd-positive reactors, 1 of the PPd-negative participants, and 2 of the 5-delayed tuberculin reactors did not react to candida testing on either reading. As shown in Table 1, no differences were seen among PPd-positive, PPd-negative, and variant reactors in sex (P = 0.19), mean age (P = 0.07), mean hemoglobin level (P > 0.2), malaria status (P = 0.13), or hepatitis B antigen carrier state (P > 0.2).

One hundred nine participants (90%) had chest radiographs available for review. Ninety-seven of these radiographs (80%) were considered to be of satisfactory quality for assessment. The chest radiographs of nine participants were consistent with active tuberculosis, but none of these participants had positive acid-fast bacilli stains or positive cultures. In addition, 16 participants had radiographs consistent with previous tuberculosis that had healed. As shown in Table 1, no statistically significant difference was seen in skin test reactivity between persons whose radiographs were consistent with tuberculous infection and persons without such radiographic evidence (P > 0.2).

Booster testing was done in the group whose initial tuberculin results were negative (n = 54) and in the variant group (n = 32) (on the basis of current standards, the latter group was categorized as tuberculin negative). Follow-up testing was done on 94% of tuberculin-negative and delayed reactor participants. Boosting was completed between 10 and 12 weeks after the initial skin testing, and results were evaluated at 48 to 72 hours and again at 6 days. Twenty of 31 variant reactors (65%) boosted to a positive result at repeat skin testing compared with only 8 of 50 participants with negative tuberculin results (16%) (P < 0.001). One of the variant reactors and four of the persons with negative PPD results were unavailable for retesting.

In vitro lymphocyte studies were done in serum samples from 57 participants (47%). Fifty-six of these participants were available for all skin test readings: Fourteen (25%) were PPD positive, 26 (46%) were PPD negative, and 16 (29%) were variant reactors. No significant difference was seen in the following variables between the group who had lymphocyte studies and the group who did not: skin tuberculin reactions (P > 0.2), mean age (37 and 34 years, respectively; P = 0.07), and chest radiograph findings (chest radiographs were clear in 83% and 67% of participants, respectively; P = 0.07). However, the mean hemoglobin level was lower in participants who had lymphocyte studies than in those who did not (103 g/L compared with 122 g/L; P = 0.01).

Overall, good correlation was seen between the size of the PPD reaction and PPd-induced blastogenesis (Spearman rank-correlation coefficient, 0.48; P < 0.001). In addition, for variant reactors only, induration at 6 days and PPd-induced blastogenesis were not significantly correlated (Spearman rankcorrelation coefficient, 0.24; P > 0.2). The lack of statistical significance may be attributable to the small sample size (n = 16).

As shown in Figure 1, variant reactors had a blastogenic response that was between the responses of tuberculin-positive and tuberculin-negative participants. The median blastogenic responses were the following: 12 667 cpm for tuberculin-negative donors (interquartile range, 4435 to 27 370 cpm), 34 942 cpm for tuberculin-positive donors (interquartile range, 20 515 to 64 314 cpm), and 18 939 cpm for variant reactors (interquartile range, 10 608 to 41 097 cpm). The difference between PPd-negative and PPd-positive reactors was statistically significant (P = 0.003). The differences between variant and PPd-negative reactors and between variant and PPd-positive reactors were not significant (P > 0.2 for both comparisons). However, in a regression model adjusting for baseline blastogenic response, the coefficient for PPd-negative compared with PPd-variant reactors was statistically significant (P = 0.04), whereas the coefficient for PPd-positive compared with PPd-variant reactors was not significant (P = 0.18).

Discussion

Top Methods Results Discussion Author & Article Info References

Many persons (26%) in this group of Indochinese immigrants had delayed tuberculin reactivity. This variant reaction had a strong positive correlation with subsequent booster testing; 65% of variant reactors boosted to a positive response when retested. The degree of induration correlated with reactivity of the lymphocytes to PPD antigen in vitro, confirming previous research findings that the spectrum of skin reactivity is continuous. Our data also confirm the findings of previous clinical studies, which have indicated that a substantial percentage of Indochinese refugees have positive results to PPD retesting [19]. Our study adds to the literature by showing a correlation between delayed skin testing and lymphocyte blastogenesis. An additional strength of our study is the unconfounded nature of the study group, that is, their lack of previous BCG immunization and exposure to western medical care.

Our study has two main limitations. First, not all participants had lymphocyte studies. However, age, tuberculin reactivity, and chest radiography did not significantly differ between persons who had lymphocyte blastogenesis and those who did not. Second, the applicability and prevalence of variant reactivity to non-Asian populations remains to be shown, and differences among Asian subpopulations may exist.

An important component in assessing tuberculin screening is the predictive value of a positive test result. The specificity of the PPD test has been reported to exceed 95% [20], and the probability that a positive reaction represents true infection increases markedly with the prevalence of tuberculosis in the group being studied. Because immigrants from Southeast Asia are at particularly high risk for tuberculous infection [20, 21] and because tuberculin testing has a high specificity, the positive predictive value of tuberculin testing is high.

Tuberculin skin testing with the Mantoux method is the mainstay for identifying asymptomatic persons infected with M. tuberculosis for whom treatment with prophylactic isoniazid may be recommended [22-26]. Evaluation of the results is standardized according to the degree of induration at the site of injection 48 to 72 hours after injection [21-23]. This definition of positivity may need to be reevaluated for high-risk populations. In our study group, variant reactivity [26%] appears to represent true positivity for previous tuberculous infection. The biological explanation for the variant response is not yet known. Given this high rate of tuberculin positivity, refugee screening programs should consider several approaches to assure that high-risk persons are appropriately identified as candidates for preventive therapy. These approaches include 1) reading skin tests at the standard time and again at 6 days and 2) retesting persons with negative skin test results and variant reactions 1 to 2 weeks after initial evaluation.

Variant tuberculin reactivity occurs in a substantial proportion of Indochinese refugees at risk for tuberculous infection. Recognition and treatment of asymptomatic tuberculin reactors are important in controlling the resurgence of tuberculosis and stabilizing the development of drug-resistant strains of tuberculosis [27]. Variant reactivity was associated with subsequent booster positivity and, on the basis of lymphocyte blastogenesis, had an immunologic response between those of standard positive tuberculin reactivity and PPD negativity. This finding supports the hypothesis that the variant response represents true positivity in most persons.

We conclude that Indochinese persons having skin testing should be evaluated at 48 to 72 hours and again at 6 days and that booster testing should routinely be done 1 to 2 weeks after the initial PPD evaluation. Persons whose initial test results show an increase in induration of greater than 5 mm between the 48- 72-hour and 6-day readings require booster testing. If results are positive at the standard reading time, these persons should be considered positive tuberculin reactors and treated accordingly. Further studies are needed of persons whose lymphocyte studies suggest tuberculin reactivity but who do not have subsequent positive skin test results at repeat evaluation.

Ms. Burtt: Wake County Department of Health, 10 Sunnybrook Road, Raleigh, NC 27605.

Ms. Edmonds: Division of Infectious Diseases, Department of Medicine, 10th Floor Biomedical Research Building, Case Western Reserve University School of Medicine, 10900 Euclid Avenue, Cleveland, OH 44106-4984.

Dr. Molina: Department of Radiology, University of North Carolina School of Medicine, Campus Box 7510, Chapel Hill, NC 27599-7510.

Dr. Kiefe: University of Alabama at Birmingham, Division of Preventive Medicine, Department of Medicine, 1717 11th Avenue South, Medical Towers 700, Birmingham, AL 35205-4785.

Dr. Ellner: Division of Infectious Diseases, Department of Medicine, 10th Floor Biomedical Research Building, Case Western Reserve University School of Medicine, 10900 Euclid Avenue, Cleveland, OH 44106-4984.