Low sensitivity of a urine LAM-ELISA in the diagnosis of pulmonary tuberculosis

BMC Infectious Diseases

Table 3 Univariate and multivariate influence of HIV status and Sex on Lam Sensitivity

			Univariate Results			Multivariate Results
Covariate	Stratum	n	Prevalence ratio for specificity	(95% CI)	p-value	Prevalence ratio for specificity	(95% CI)	p-value
HIV status
	Negative	19	1			1
	Positive	50	2.945	(1.193–7.272)	0.019	2.619	(1.045 – 6.561)	0.040
Sex
	Female	30	1			1
	Male	39	0.577	(0.359 – 0.927)	0.023	0.694	(0.440 – 1.097)	0.118

The results of a Poisson regression included only gold standard positive participants. The outcome variable (sensitivity) was coded as 1 for true positive LAM results and as 0 for false negative LAM results. Apart from age and sex no other predictors were included into the multivariate model.
Prevalence ratio for specificity = probability to receive a (correct) positive LAM result for TB infected patients who are HIV+ve or male, divided by probability to receive a (correct) positive LAM result for TB infected patients who are HIV-ve or female.