Population pharmacodynamic modelling of lanreotide (Lan) autogel (ATG) in patients with acromegaly

PDF of poster

Objective: To establish the concentration of Lan vs growth hormone (GH) response in patients with acromegaly over time.

Methods: A phase II, multicentre, randomised study was conducted in 108 patients with acromegaly who may or may not have been previously treated with a somatostatin analogue. Patients were randomised to receive placebo or Lan ATG 60, 90 or 120 mg once every 28 days by deep subcutaneous injection. The 52-week study had four phases: i) washout (weeks -12–0), for those previously treated; ii) double-blind, placebo-controlled (weeks 0–4), single dose of Lan ATG or placebo; iii) single-blind, fixed-dose (weeks 4–20), four injections; and iv) open-label dose titration (weeks 20–52), eight injections and two dose adjustments were allowed. Concentrations of Lan and mean serial measurements of GH in serum used for pharmacodynamic analysis were measured at washout and weeks 4, 13–16 and 52. Observed serum concentrations of Lan were used to build the population pharmacodynamic model for GH using NONMEM version V software programme.

Results: A total of 691 concentration–response pair data were used to develop the population model. GH levels were described as a function of observed concentrations of Lan using an inhibitory sigmoidal E_MAX model. The estimate of GH at baseline (GH₀) was higher for patients not treated with somatostatin analogues in the last 3 months (15.9 vs 8.74 ng/mL); the data supported the existence of two different sub-populations (responders and non-responders) with regard to IC₅₀ [concentration of Lan eliciting a half-maximal decrease in GH (E_MAX)]. Typical estimates were: 100 ng/mL (IC₅₀ non-responders), 0.612 ng/mL (IC₅₀ responders), 0.822 ng/mL (E_MAX) and 2.63 (sigmoidicity factor).  No differences in E_MAX were found by pre-treatment status with somatostatin analogues in the previous 3 months.

The fraction of responders within the population was 0.863. The lower sensitivity found in the small proportion of non-responders indicates that these patients do not respond to Lan treatment even at high doses.

The degree of inter-patient variability in GH₀, IC₅₀, and E_MAX was 85%, 43% and 17%, respectively. The model was internally (using the posterior predictive check) and externally (with data from another clinical trial) validated.

Conclusions: E_MAX was estimated as 82% demonstrating the efficacy of Lan, and 86% of patients responded to treatment. In those patients a Lan concentration of 0.612 ng/mL resulted in a 50% decrease in GH levels.