Table 1.
Demographics and factors investigated for association with BCG scar formation.
| Factor | Total BCG |
Scar prevalence |
||
|---|---|---|---|---|
| Univariate |
Multivariate |
|||
| (N=3071) | n/N (%) | OR (95% CI) | OR (95% CI) | |
| Sex | ||||
| Male | 822 | 556 (67.6) | 1 (reference) | 1 (reference) |
| Female | 2249 | 1785 (79.4) | 1.84 (1.53–2.20), p < 0.001 | 2.00 (1.65–2.42), p < 0.001 |
| Age | ||||
| 18–49 | 2144 | 1733 (80.8) | 1 (reference) | 1 (reference) |
| ≥50 | 927 | 608 (65.6) | 0.45 (0.38–0.54), p < 0.001 | 0.43 (0.35–0.51), p < 0.001 |
| Nutritional status (BMI) | ||||
| Normal weight (18.5–24.9) | 1266 | 964 (76.1) | 1 (reference) | – |
| Underweight (<18.5) | 35 | 28 (80.0) | 1.25 (0.54–2.90), p = 0.6 | |
| Pre-obesity (25.0–29.9) | 1090 | 820 (75.2) | 0.95 (0.79–1.15), p = 0.6 | |
| Obesity class I (30.0–34.9) | 419 | 328 (78.2) | 1.13 (0.87–1.47), p = 0.4 | |
| Obesity class II (35.0–39.9) | 155 | 121 (78.1) | 1.11 (0.75–1.67), p = 0.6 | |
| Obesity class III (≥40) | 58 | 45 (77.6) | 1.08 (0.58–2.04), p = 0.4 | |
| Unknown | 48 | 35 (72.9) | NA | |
| Smoker | ||||
| No | 2808 | 2131 (75.9) | 1 (reference) | – |
| Yes | 263 | 210 (79.8) | 1.26 (0.92–1.72), p = 0.2 | |
| Diabetes | ||||
| No | 2991 | 2278 (76.2) | 1 (reference) | – |
| Yes | 80 | 63 (78.8) | 1.16 (0.67–1.99), p = 0.6 | |
| Chronic respiratory disease | ||||
| No | 2866 | 2177 (76.0) | 1 (reference) | – |
| Yes | 205 | 164 (80.0) | 1.27 (0.89–1.80), p = 0.2 | |
| Chronic cardiovascular disease | ||||
| No | 2740 | 2107 (76.9) | 1 (reference) | – |
| Yes | 331 | 234 (70.7) | 0.72 (0.56–0.93), p = 0.01 | |
| Study country | ||||
| Australia | 1380 | 1003 (72.7) | 0.70 (0.59–0.83), p < 0.001 | 0.84 (0.70–1.01), p = 0.07 |
| Brazil | 1222 | 1032 (84.4) | 2.24 (1.86–2.69), p < 0.001 | 1.61 (1.29–2.01), p < 0.001 |
| Netherlands | 280 | 187 (66.8) | 0.59 (0.46–0.77), p < 0.001 | 1.01 (0.75–1.36), p = 0.9 |
| Spain | 110 | 52 (47.3) | 0.26 (0.18–0.39), p < 0.001 | 0.31 (0.20–0.46), p < 0.001 |
| UK | 79 | 67 (84.8) | 1.76 (0.95–3.28), p = 0.07 | 1.84 (0.97–3.50), p = 0.06 |
| BCG history | ||||
| 1st BCG | 990 | 677 (68.4) | 1 (reference) | 1 (reference) |
| BCG revaccination | 2081 | 1664 (80.0) | 1.85 (1.55–2.19), p < 0.001 | 1.65 (1.33–2.04), p < 0.001 |
| Previous known LTBI | ||||
| No | 3031 | 2309 (76.2) | 1 (reference) | – |
| Yes | 23 | 16 (69.6) | 0.71 (0.29–1.74), p = 0.5 | |
| Unknown | 17 | 16 (94.1) | NA | |
| Previous TST | ||||
| Negative/None | 2568 | 1961 (76.4) | 1 (reference) | – |
| Positive (>5 mm) | 186 | 149 (80.1) | 1.25 (0.86–1.81), p = 0.2 | |
| Unknown | 317 | 231 (72.9) | NA | |
| BCG batch | ||||
| 118006D | 591 | 431 (72.9) | 0.80 (0.66–0.99), p = 0.04 | – |
| 118017F | 820 | 587 (71.6) | 0.71 (0.60–0.86), p < 0.001 | |
| 118019D | 658 | 536 (81.5) | 1.48 (1.19–1.84), p = 0.001 | |
| 119039B | 82 | 70 (85.4) | 1.84 (0.99–3.42), p = 0.06 | |
| 119053A | 631 | 527 (83.5) | 1.75 (1.39–2.20), p < 0.001 | |
| 200731–014 | 245 | 160 (65.3) | 0.56 (0.42–0.73), p < 0.001 | |
| 200904–017 | 35 | 27 (77.1) | 1.05 (0.48–2.33), p = 0.9 | |
| Unknown | 9 | 3 (33.3) | NA | |
| Co-administered influenza vaccine† | ||||
| No | 1883 | 1490 (79.1) | 1 (reference) | – |
| Yes | 1188 | 851 (71.6) | 0.67 (0.57–0.79), p < 0.001 | |
| Post-injection wheal* | ||||
| Yes | 2898 | 2223 (76.7) | 1 (reference) | 1 (reference) |
| No | 32 | 19 (59.4) | 0.44 (0.22–0.90), p = 0.03 | 0.44 (0.21–0.93), p = 0.03 |
| Unknown | 141 | 99 (70.2) | NA | NA |
| Vaccinator experience | ||||
| ≥20 vaccinees | 2608 | 2007 (77.0) | 1 (reference) | – |
| 0-19 vaccinees | 463 | 334 (72.1) | 0.76 (0.61–0.95), p = 0.02 | |
Abbreviations: BCG, Bacille Calmette-Guérin; BMI, body mass index; OR, odds ratio; LTBI, latent tuberculosis infection; NA, not applicable; TST, tuberculin skin test. *Wheal response (yes/no) analysed for participants who received one BCG dose only. †Stage 1 participants (Australia) were required to receive influenza vaccination on day of randomisation. Significant factors (p-value <0.2) resulting from the univariate logistic regression analysis were included as possible covariates in a multivariate logistic regression model. The model presented in the table was created using backward stepwise exclusion of factors with p-value >0.05, using sequential model testing.