TABLE 4.
Multivariate linear mixed-effects regression analyses of determinants affecting log-transformed peak plasma concentration (Cmax) values for first-line antituberculosis drugs in children and adolescents
| Isoniazid | Rifampicin | Pyrazinamide | Ethambutol | |||||
|
β
(95% CI) |
Percentage
change, % (95% CI)# |
β
(95% CI) |
Percentage
change, % (95% CI)# |
β
(95% CI) |
Percentage
change, % (95% CI)# |
β
(95% CI) |
Percentage
change, % (95% CI)# |
|
| (Intercept) | 1.46 (1.27–1.65)*** |
2.21 (2.01–2.41)*** |
3.74 (3.62–3.86)*** |
0.75 (0.49–1.00)*** |
||||
| Dose, mg·kg−1¶ | 0.40 (0.29–0.52)*** |
50 (33–68) |
0.52 (0.33–0.72)*** |
69 (38–106) |
0.16 (0.11–0.22)*** |
18 (11–25) |
0.13 (0.05–0.22)** |
14 (5–24) |
| Age | ||||||||
| <2 years+ | −0.28 (−0.40– −0.16)*** |
−24 (−33– −15) |
−0.42 (−0.57– −0.27)*** |
−34 (−43– −24) |
−0.18 (−0.28– −0.09)*** |
−17 (−24– −8) |
−0.68 (−0.90– −0.46)*** |
−50 (−59– −37) |
| 2–4 years | −0.07 (−0.18–0.04) |
−7 (−16–4) |
−0.18 (−0.32– −0.04)** |
−17 (−28– −4) |
−0.15 (−0.25– −0.06)** |
−14 (−22– −6) |
−0.32 (−0.53– −0.11)** |
−27 (−41– −11) |
| 5–9 years | −0.03 (−0.13–0.06) |
−3 (−12–6) |
−0.09 (−0.22–0.04) |
−8 (−19–4) |
−0.10 (−0.18– −0.02)* |
−9 (−16– −2) |
−0.12 (−0.31–0.06) |
−12 (−26–6) |
| 10–14 years§ | Reference | Reference | Reference | Reference | Reference | Reference | Reference | Reference |
| 15–18 years | −0.03 (−0.31–0.26) |
−3 (−27–29) |
0.06 (−0.31–0.42) |
6 (−27–52) |
−0.02 (−0.26–0.23) |
−2 (−23–25) |
0.10 (−0.51–0.70) |
10 (−40–101) |
| Sex | ||||||||
| Female | Reference | Reference | Reference | Reference | Reference | Reference | Reference | Reference |
| Male | −0.04 (−0.11–0.03) |
−4 (−10–3) |
0.02 (−0.07–0.10) |
2 (−6–11) |
−0.05 (−0.11–0.001)‡ |
−5 (−10– −0.1) |
−0.03 (−0.17–0.10) |
−3 (−15–10) |
| Malnourishedƒ | ||||||||
| No | Reference | Reference | Reference | Reference | Reference | Reference | Reference | Reference |
| Yes, moderate | −0.06 (−0.14–0.03) |
−5 (−13–3) |
−0.03 (−0.14–0.07) |
−3 (−13–8) |
−0.02 (−0.09–0.05) |
−2 (−8–5) |
−0.10 (−0.27–0.07) |
−10 (−24–7) |
| Yes, severe | −0.09 (−0.18– −0.003)* |
−9 (−17– −0.3) |
−0.12 (−0.24– −0.01)* |
−12 (−21– −1) |
−0.10 (−0.18– −0.03)** |
−10 (−16– −3) |
−0.12 (−0.29–0.06) |
−11 (−25–6) |
| Unknown | 0.07 (−0.20–0.34) |
7 (−18–40) |
−0.14 (−0.67–0.39) |
−13 (−49–48) |
0.05 (−0.15–0.26) |
6 (−14–30) |
−0.33 (−0.78–0.12) |
−28 (−54–12) |
| HIV status | ||||||||
| Negative | Reference | Reference | Reference | Reference | Reference | Reference | Reference | Reference |
| Positive | −0.17 (−0.28– −0.06)** |
−16 (−24– −6) |
−0.25 (−0.39– −0.11)*** |
−22 (−32– −10) |
−0.11 (−0.20– −0.03)* |
−11 (−18– −3) |
−0.35 (−0.53– −0.17)*** |
−29 (−41– −15) |
| Unknown | 0.05 (−0.20–0.29) |
5 (−18–33) |
−0.19 (−0.49–0.11) |
−17 (−49–12) |
−0.05 (−0.22–0.12) |
−5 (−20–13) |
0.04 (−0.34–0.43) |
4 (−29–53) |
| Acetylator status, t1/2 phenotype## | ||||||||
| Slow | 0.23 (0.15–0.31)*** |
26 (16–36) |
||||||
| Intermediate | Reference | Reference | ||||||
| Rapid | −0.13 (−0.25– −0.02)* |
−12 (−22– −2) |
||||||
| Unknown | −0.38 (−0.53– −0.23)*** |
−31 (−40– −20) |
||||||
| Random effects | ||||||||
| σ2 (mean±sd) | 0.35±0.59 | 0.49±0.73 | 0.19±0.43 | 0.53±0.73 | ||||
| τ00 studies (mean±sd) | 0.13±0.35 | 0.11±0.36 | 0.03±0.19 | 0.06±0.24 | ||||
| τ11 studies×doses (mean±sd) | 0.05±0.22 | 0.10±0.25 | 0.01±0.09 | |||||
| ρ01 studies | −0.33 | 0.02 | −0.15 | |||||
| ICC | 0.31 | 0.32 | 0.18 | 0.10 | ||||
| Nstudies | 27 | 22 | 23 | 11 | ||||
| Observations | 1292 | 1105 | 1021 | 483 | ||||
| Conditional R2 | 0.51 | 0.55 | 0.30 | 0.23 | ||||
β: fixed-effects regression coefficient; t1/2: elimination half-life; σ2: residual variance; τ00: random intercept variance; τ11: random slope variance; ρ01: random slope–intercept correlation, ICC: interclass correlation estimate; Nstudies: number of included studies (studies or study occasions); conditional R2: proportion of variance explained by both the fixed and random effects. #: percentage change was calculated as: (eβ−1)×100%; ¶: dose was mean-centred by subtracting the mean from each data point, then standardised by dividing each point by the standard deviation; +: among children <2 years of age, Cmax values were not significantly different in patients aged 3–11 months compared with those aged 12–23 months for isoniazid, rifampicin, pyrazinamide and ethambutol (the results were adjusted for drug dose in mg·kg−1, sex, nutritional status and HIV status); §: we used children aged 10–14 years as a reference, assuming that they were the most adult-like among children under <15 years of age, and also to assess the statistical difference with older adolescents aged 15–18 years; ƒ: moderate malnutrition was defined as weight-for-age or height-for-age z-score ≥ −3 but < −2 in children aged <5 years and height-for-age or body mass index-for-age z-score ≥ −3 but < −2 in children aged ≥5 years, and severe malnutrition was defined as weight-for-age or height-for-age z-score < −3 in children aged <5 years, and height-for-age or body mass index-for-age z-score < −3 in children aged ≥5 years; ##: acetylator phenotypes of isoniazid were rapid (t1/2 <1.25 h), intermediate (t1/2 1.25–2 h) and slow (t1/2 >2 h). ***: p<0.001; **: p<0.01; *: p<0.05; ‡: p<0.1.