Efficiency Considerations for Group Sequential Designs with Adaptive Unblinded Sample Size Re-assessment