Tuberculosis and tuberculin quality: best intentions, misleading results.

OBJECTIVE

To compare the performance of three purified protein derivative (PPD) formulations: Tubersol (Connaught); RT23, Statens Serum Institut (SSI); and RT23, Mexico, tested in Mexican populations at low and high risk for tuberculosis (TB).

DESIGN

A double-blinded clinical trial.

SETTING

A university hospital in Mexico City.

PARTICIPANTS

The low-risk population was first or second-year medical students with no patient contact; the high-risk population was healthcare workers at a university hospital.

METHODS

Each of the study subjects received the three different PPD preparations. Risk factors for TB, including age, gender, occupation, bacille Calmette-Guérin (BCG) status, and TB exposure, were recorded. A 0.1-mL aliquot of each preparation was injected in the left and right forearms of volunteers using the Mantoux technique. Blind readings were done 48 to 72 hours later. Sensitivity and specificity were calculated at 10 mm of induration using Tubersol as the reference standard. The SSI tested the potency of the different PPD preparations in previously sensitized guinea pigs.

RESULTS

The low-risk population had a prevalence of positive PPD of 26%. In the low-risk population, RT23 prepared in Mexico, compared to the 5 TU of Tubersol, had a sensitivity of 51%, a specificity of 100%, a positive predictive value of 100%, and a negative predictive value of 86%. The RT23 prepared at the SSI had a sensitivity of 69%, a specificity of 99%, a positive predictive value of 95%, and a negative predictive value of 90%. In the high-risk population, the prevalence of positive PPD was 57%. The RT23 prepared in Mexico had a sensitivity of 33%, a specificity of 100%, and a positive predictive value of 53%; the RT23 prepared at the SSI had a sensitivity of 91%, a specificity of 98%, a positive predictive value of 98%, and a negative predictive value of 89%. RT23 used in Mexico had a potency of only 23% of that of the control. There was no statistical association among those with a positive PPD, irrespective of previous BCG vaccination (relative risk, 0.97; 95% confidence interval, 0.76-1.3;P=.78).

CONCLUSIONS

Healthcare workers had twice the prevalence of positive PPD compared to medical students. RT23 prepared in Mexico had a low sensitivity in both populations compared to 5 TU of Tubersol and RT23 prepared at the SSI. Previous BCG vaccination did not correlate with a positive PPD. Low potency of the RT23 preparation in Mexico was confirmed in guinea pigs. Best intentions in a TB program are not enough if they are not followed by high-quality control.