Skip to main content

Table 4 Multivariate analysis of factors associated with development of PR: results from conditional logistic regression model

From: Determinants of treatment-related paradoxical reactions during anti-tuberculosis therapy: a case control study

Variable

Odds ratio (95 % CI)

P value

Age

 

0.98 (0.95–1.02) per increasing year

0.409

Sex

 

0.57 (0.19–1.67) for females

0.302

Site

Chest

1.00

–

Abdominal

8.11 (0.18–356.01)

–

Brain

2.22 (0.22–23.30)

–

Peripheral lymph nodes

64.33 (9.60–431.25)

–

Other or mixed sites

1.23 (0.41–3.74)

<0.001

HIV Status

Negative

1.00

–

Positive

5.05 (1.28–19.85)

–

Not recorded

0.50 (0.02–14.97)

0.028

Immunosuppression

No

1.00

–

Yes

0.01 (0.00–0.27)

–

Not recorded

0.13 (0.00–0.90)

0.002

Tobacco use

No

1.00

–

Yes

0.71 (0.17–2.85)

–

Not recorded

3.36 (0.39–21.04)

0.462

Alcohol use

No

1.00

–

Yes

0.21 (0.04–1.01)

–

Not recorded

0.01 (0.01–0.56)

0.009

ESRa

Mean (mm/h)

Could not fit within model

–

TB Diagnosis

% NAAT negative

1.00

–

% NAAT positive

1.23 (0.11–12.63)

–

% NAAT not performed

0.10 (0.01–1.11)

0.009

% culture negative

1.00

–

% culture positive

6.87 (1.31–36.04)

–

% culture not performed

3.81 (0.66–22.14)

0.045

  1. aESR was only performed on a limited number of patients (n = 139, only 97 of which could be used for case control analysis): it was unable to be included in the model due to small numbers. In univariate analysis a 1-log increase in ESR was associated with a 4.43 increased odds of developing PR (95 % CI 1.30–15.1); this increased to 4.78 (1.29–17.72) when adjusted for HIV status