Page 74 - JCTR-11-1

Journal of Clinical and
Translational Research Pediatric dosing of antituberculosis medicines

clinical dose(s) of these medicines were used as the of the body weight range was used. For example, for body
observed dose(s) to compare with the predicted dose(s) by weight ranging from 10 to 20 kg, a mid-point of 15 kg was
allometry or Salisbury Rule. The allometric method used used in equation 3.
in this study was previously developed and validated using
external data. 12,15,26,32 2.3. Statistical analysis
In the literature, a 2-fold prediction error (0.5 – 2-fold) is
2.1. Method 1: Salisbury rule considered acceptable. However, a 2-fold prediction error
The following two methods known as Salisbury Rule were appears to be too high and may be of little practical value
used for the prediction of antituberculosis medicines even for the first-time-pediatric dose selection. Therefore,
proposed by Lack and Stuart-Taylor : in this study, a prediction error of 0.5 – 1.5 (a 50%
30
For children weighing less than 30 kilograms: prediction error on either side of 1) in place of 0.5-2-fold
error was considered acceptable. A more stringent criteria
2 × Weight in kilograms = % of adult dose (1) in terms of 0.7-1.3 (a 30% prediction error on either side
For children weighing greater than ≥30 kilograms but less of 1) was also used. The fold-error between predicted and
than 70 kg: observed was calculated as follows:
Weight in kilograms + 30 = % of adult dose (2) Fold error = predicted dose/observed dose (4)
The predicted dose in this study was a single value
2.2. Method 2: Allometric scaling
and was compared with a single recommended dose. In
Generally, pediatric dose is recommended based on per kg clinical practice, due to the differences in response across
body weight (derived from adult dose and body weight). the patient population, the recommended dose used in this
This approach assumes that there is a linear relationship study may differ from the administered dose.
(exponent 1.0) between body weight and dose, irrespective
of age. Considering that, body weight based dosing across 3. Results
the age groups may not be a linear process, a theoretical The results of this study are summarized below and in
exponent 0.75 has been proposed. This exponent, though Table 1 and in the supplementary Tables S1-S7. There
12
useful, is not universally applicable, particularly for younger were 7 antituberculosis medicines with 62 observations
children (typically those aged 2 years or younger). (different weight groups for each drug) for allometry or
12
Therefore, in this study, a middle ground strategy was taken the Salisbury Rule.
to choose an allometric exponent to predict pediatric dose
across all ages. The mid-point between 0.75 and 1.0 is 0.87 For allometric scaling, 96.8%, 90.3%, and 88.7%
and it was rounded to 0.90. Hence, exponent 0.90 was used observations were within 0.5 – 2-fold, 0.5 – 1.5-fold,
to predict pediatric dosse using body weight in equation 3. and 0.7 – 1.3-fold prediction error, respectively. For the
This approach was taken in other studies for pediatric dose Salisbury Rule, 100%, 98.4%, and 80.6% observations
prediction. 12,15,26,32 The pediatric doses of antituberculosis were within 0.5 – 2-fold, 0.5 – 1.5-fold, and 0.7 – 1.3-fold
medicines across different age groups were predicted by prediction error, respectively (Table 1). Overall, predicted
equation 3: dose of antituberculosis medicines in children by the two
proposed methods reconciled very well with the observed
Pediatric Dose = Adult Dose × (weight of the child/weight or recommended clinical dose. The robustness of the
of the adults) 0.9 (3) methods can be gauged from the fact that more than 80%
Where, the ‘adult dose’ is the adult dose of a given
drug. Generally, in equation 3, an adult body weight of 70 Table 1. Prediction of pediatric antituberculosis dose based
kg is used. However, it was noted that the recommended on the Salisbury rule or allometry
antituberculosis dose in most instances is equal to an Fold error Allometry (n=62) Salisbury Rule (n=62)
adult dose starting as low as 34 kg body weight. Therefore, Number (%) Range Number (%) Range
an adult body weight was varied according to the lowest 0.5–2.0 60 (96.8) 0.70 – 1.80 62 (100) 0.55 – 1.68
starting dose for adults. For example, if the recommended
dose in adults starts from 30 kg body weight then the adult 0.5–1.5 56 (90.3) 0.70 – 1.44 61 (98.4) 0.55 – 1.38
body weight used in equation 3 was 50 kg (a midpoint 0.7–1.3 55 (88.7) 0.70 – 1.27 50 (80.6) 0.72 – 1.28
between 30 to 70 kg). Similarly, an adult body weight >2 2 (3.2) 2.28 – 2.40 0 0
of 60 kg as a mid-point was used between 50 and 70 kg <0.5 0 0 0 0
in equation 3. For children, the recommended dose was Note: The analysis included seven antituberculosis medicines, with
based on body weight range or band therefore, a midpoint 62 total observations (different weight groups for each drug).

Volume 11 Issue 1 (2025) 68 doi: 10.36922/jctr.24.00070