Session: SUN 624-646-Growth: Clinical Trials & Observational Studies
Poster Board SUN-642
Recently we derived formulas for the conversion of GH doses between mg/m2/wk and mg/kg/wk (1). These were based on the observation that for a given /m2 dose the /kg dose decreases with age and that this decrease becomes greater as the /m2 dose increases. The general formula for conversion from /m2to /kg dose is
In addition to dose and age, in the formula above, gender and diagnosis may also affect dose conversion. Also, the value of 20.5 is a compromise. It is the mean of the ratio of /m2doses to the /kg doses. If the mean is used a value of 20.4 is obtained but a weighted mean yields 20.6. Preliminary analyses suggested that the general formula, using 20.5 and neglecting diagnosis and gender would be adequate in most cases but that more detailed analyses were required. Here we describe the effects of gender, diagnosis, and varying the dose ratio value and identify when more specific formulae may be appropriate.
Over 33,000 doses from the Australian OZGROW database were used in dose ranges of 1 mg/m2/wk from 2-3 mg/m2/wk to 9-10 mg/m2/wk (1). To assess the gender effect this was repeated separately for males and females and formulae derived. For individual diagnoses (GH deficiency-GHD, idiopathic short stature-ISS, Turner syndrome-TS, Prader-Willi-PWS, Chronic Renal Failure-CRF), as numbers were smaller, overlapping ranges 2-4, 3-5,…,7-9 mg/m2/wk were from which specific formulae were derived.
Multiple LR with mg/kg dose as the dependent variable was used to confirm if gender or any specific diagnosis affected dose conversion. Multiple LR was also used as an alternative method of dose conversion.
The general formula and each gender, diagnosis, diagnosis/gender, and LR (general, and gender and diagnosis specific) formula was used to produce mg/kg dose estimates for each mg/m2dose. Residuals were calculated as the absolute difference (mg/kg dose –estimate) and the absolute % difference. The mean residual was calculated for each formula.
Gender and all diagnoses except CRF affected dose conversion (P<0.0001). LR based conversion formulas were always inferior to those based on the original general formula. LR-based: residual means=0.0128 (5.7%) to 0.0181 (7.9%); Original-based: residual means=0.0125 (5.6%) to 0.0169 (7.4%). A ratio of 20.4 was best for ISS but 20.6 for the other diagnoses. A gender and diagnosis specific formula was best for ISS, TS, and PWS. The general formula was best for GHD and CRF. The largest difference between mean residuals of general and specific formulae was 0.0019 (0.9%) for ISS.
Gender and diagnosis specific formulae marginally improve dose conversion estimates over the general formula for some diagnoses. A ratio of 20.4 may be used for ISS and 20.6 for other diagnoses but the improvement from 20.5 is marginal. The general formula proved robust and adequate for most situations.
Nothing to Disclose: IPH, MH, AMC, CSC
*Please take note of The Endocrine Society's News Embargo Policy at http://www.endo-society.org/endo2013/media.cfm
See more of: Abstracts - Orals, Featured Poster Presentations, and Posters