Graham, M. R., Baker, J. S., Evans, P., Kicman, A., Cowan, D., Hullin, D., Thomas, N. & Davies, B. (2008). Physical effects of short-term recombinant human growth hormone administration in abstinent steroid dependency. Hormone Research, 69, 343-354.

Six days' recombinant human growth hormone (rhGH) administration, in abstinent (at least 12 weeks) anabolic-androgenic steroid dependent males, was compared with controls on a number of factors. A control group (N = 24) and an rhGH-using group (0.058 IU/kg/day; N = 24). Physiological measurements included anthropometry, strength, power, and peak oxygen uptake (VO2peak). Biochemical measurements included haemoglobin, packed cell volume, glucose, sodium, potassium, urea, creatinine, total protein, albumin, thyroid function, testosterone, prolactin, cortisol, growth hormone and insulin-like growth factor-I (IGF-I).

Strength, peak power output, and IGF-I significantly increased and total protein, albumin, and free tetra-iodothyronine significantly decreased compared to controls and within the growth hormone group. Fat-free mass index and VO2peak significantly increased, while body fat and thyroid-stimulating hormone significantly decreased within the growth hormone.

Implication. Short-term recombinant human growth hormone supplementation increased strength and power.

[This study and its discussion includes a number of dubious inferences and assumptions. Only some of those concerning the study's experimental design, strength, and power will be considered along with the general statements regarding the import of what the study showed.

• When study effects are likely to be small, the preferred experimental design is to form matched groups rather than the randomly assigned groups used in this investigation. Matching should have been for physical characteristics, histories of steroid use, and training regimes. The location of smaller but significant effects is facilitated with the multiple-factor matched-groups experimental design even at the cost of group sizes (which seemed to be important to the authors). When randomization is used, the likelihood of a Type I error is substantial. With matching, that likelihood is lowered. In this study, it is possible that the lack of control over potential confounding variables increased the likelihood of Type I errors. A stringent experimental approach is required to investigate drug effects on basic physical measures because of the many factors that affect strength and power, not the least of which is the placebo effect [the rhGH-using Ss knew they were "using" a drug whereas the control Ss did not receive anything.] Other refinements in experimental control should have been used but will not be discussed here.
• Some strength and power measures were statistically significant between the groups but revealed a further justification for asserting the need for a matched-groups experimental design. All measures of peak power, mean power, and strength measures (bench press and squat) were greater in the rhGH-using group than the control group, including the pre-experiment and post-experiment measures [the differences could be an artifact of randomization but it is interesting to note the advantages only occurred in the rhGH-user group]. What is notable is that the highest measures for power came after the cessation of rhGH use, and in strength the improved measures under rhGH-use were maintained across the subsequent period of non-use. The changes in strength and power were independent of rhGH-use because they existed at the termination of supplementation and continued after the cessation of supplementation. Phenomena of this form (in the presence and absence of the independent variable) normally suggest influential variables that were not controlled in the study. The paper does suggest some variables/factors that might lead to such results but they are purely speculative and are no more valid than the assertion of uncontrolled effects. The authors' suggestions for the reasons underlying the significant differences associated with no longer using rhGH are not in concert with the biochemical and hormonal changes reported.
• The paper suggests that "these subjects were experienced former AAS users, in a latent catabolic phase, which may have been ameliorated by the anabolic effect of rhGH administration. This would suggest that the time period [sic] of a 2-year ban, following a doping offence, is too short." The authors essentially claim that the effects of steroid use last longer than two years and that rhGH-use somehow interacts with those sustained effects.

The problem with the above assertion, which is included in the Conclusion of the study, is that there were no data collected to support such assertions and the experimental design did not allow any inference about residual effects of steroid use over any period.

What was needed was a study that involved four matched groups. The four groups should have been 1) rhGH + steroid abstinence, 2) steroid abstinence, 3) rhGH + never-used steroids, and 4) never-used steroids (the actual control). Those groups would have allowed a comparison between steroid abstinence and never-used, which would have suggested, at least in part, what residual effects of steroid use existed (if any, but there are published studies that found steroid use effects are reversible). Without a never-used group, any inference about enduring effects of steroids is invalid in this study. Another benefit of the added groups would have been either the replication or differentiation of the strength and power results (revealed in the study's rhGH-use group) in the never-used-steroids group.

• When studies contain many statistical analyses, the greater the number of those tests, the greater is the likelihood of Type I errors, that is, significant results occur through chance alone because of the assumptions and inexactitudes of statistical analyses. Some of the considerable number of significant results in the study were likely to have been caused by this phenomenon.
• The final consideration offered here is that the strength and power results of this study are contradicted by multiple studies that show no effect of short- and long-term use of rhGH. At best, this investigation suggests that researchers do not know all the factors concerning rhGH use, strength, and power that need to be controlled to truly reveal the relationship between the hormone and the two physical attributes [it is also incorrect to infer strength and power measures reflect sporting performances]. Thus, until this study using non-clinical Ss (e.g., GH deficient patients), published results showed no effects on strength, power, [and VO2max] of rhGH supplementation. The weight of evidence is still on the side of no-effects but future studies need to be cognizant of the equivocal finding of this study.]

Return to Table of Contents for this issue.