The company line is given at 

http://www.thebody.com/content/art13618.html

T20, Fuzeon, shows little acceptance by the patients, because it must be injected twice daily and causes local reactions. This problem is expected to remain for T1249.  Increasing the half-life of C-peptide entry inhibitors by conjugation to larger molecules has most likely so far has not been successful because these peptides loose efficacy in conjugated forms. That is at least what we have observed. 

If people are looking for lessons learned with T20, you might look at the peptide design issues.  T20 was a 36-mer made of all proteinogenic amino acids. There are two immediate issues with this: Firstly, nearly all successful peptide drugs incorporate either D-AAS or otherwise non-proteinogenic units to slow proteolysis.  Secondly, the length put it in a range where fermentative preparation wasn't really a good option, but the all-synthetic method used was extremely expensive, hence a 5-figure per-patient annual price-tag.  I have no information about how quickly the drug loses effectiveness as it's metabolized but it's hard to imagine that this isn't a drawback of the design.    

A lot of optimizaiton has gone into producing T20, T1249, Enfuride and related drugs.  The 30 amino acids separating the pocket binding domain and the lipophilic anchoring region in the HIV sequence the fusion-blocking peptides are based on have been reduced to 23 in T1249, but that still leaves a 40-mer peptide overall.  A group in Portugal showed that cholesterol can serve as the peptide lipid binding region (an 8-mer segment in T1249).  It really seems that the drugs should be chimeras with non-peptide linkers such as PEG bridging the necessary recognition sequences, resulting in minimum peptide lengths and modifications to extend half-lives (and possibly reduce reactions at the injection site?) 

Lets all check back in 5-10 years to see if that;s right.