I reproduce the problem. With the above-mentioned code I get result as follows:

# ->
# Found 61444 results
# Fetched 9946 available abstracts
# Read 9946 abstracts

Actually it is well-known PubMed problem of 10,000 results. See discussion here for example.

retmax=61444 sudgested by @Denis Sechin in Entrez.efetch does not solve the problem.

The way I can suggest is to handle entries year by year by changing mindate and maxdate in a cycle as follows:

# mindate = '2013/01/01',  maxdate = '2013/12/31',
# Found 2827 results
# Fetched 2824 available abstracts
# Read 2824 abstracts

# mindate = '2014/01/01',  maxdate = '2014/12/31',
# Found 3102 results
# Fetched 3098 available abstracts
# Read 3098 abstracts