Long-acting somatostatin analog therapy of acromegaly: A meta- analysis

Context: Although considerable data exist on the use of long-acting somatostatin analogs to treat acromegaly, their reported efficacy differs substantially among trials. Objective: We conducted a meta-analysis to derive definitive estimates of their efficacy for biochemical control and tumor shrinkage. Data Sources: A search of literature was conducted through 2003, primarily via PubMed. Study Selection: Inclusion criteria, met in 44 trials, included at least 3 months of secondary octreotide long-acting release (LAR) or lanreotide slow release (SR) therapy or of primary octreotide LAR, lanreotide SR, or sc octreotide therapy and clearly reported data on biochemical efficacy and/or tumor shrinkage. Fifty other trials screened did not meet analysis inclusion criteria. Data Extraction: Data were extracted by three independent observers. Data Synthesis: Among subjects not selected for somatostatin analog responsiveness before study entry, both GH efficacy criteria and IGF-I normalization were met in a greater proportion of those treated with octreotide LAR vs. lanreotide SR (GH: B = 0.2310, P = 0.016; IGF-I: B = 0.2325, P = 0.007). Prestudy selection for somatostatin analog responsiveness was not a significant predictor of meeting GH efficacy criteria (B = 0.0992; P = 0.12). Preselection was a positive predictor of IGF-I normalization rate (B = 0.1213; P = 0.04), which was greater among preselected than unselected subjects (B = 0.1472; P = 0.0475). IGF- I normalization occurred in a greater proportion of secondary octreotide LAR- vs. primary octreotide-treated subjects (B = 0.2056; P = 0.009). The odds of tumor shrinkage more than 10% were lower in the unselected vs. preselected subjects. However, the effect of drug type was an important predictor of shrinkage; such that regardless of preselection or not, the odds of shrinkage with lanreotide SR were lower than with octreotide LAR (P = 0.003). Shrinkage greater than 10% occurred in a higher percentage of primary octreotide LAR- treated vs. primary octreotide sc- treated subjects (odds ratio = 9.4; P = 0.0001). The overall rate of tumor increase was 1.4%. Conclusions: In this meta-analysis, we have shown that the efficacy of octreotide LAR is greater than lanreotide SR among subjects unselected for prior somatostatin analog responsiveness. Preselection is a significant positive predictor of IGF-I normalization and is associated with increased odds of tumor shrinkage, which is also greatest with octreotide LAR. Biochemical efficacy is similar, but tumor shrinkage is greater when these drugs are given as primary vs. secondary therapy.