On 2015 Nov 12, Erick H Turner commented:

To say "no publication bias" overstates the findings. Nonpublication is one aspect of publication bias (also known as reporting bias). Another important type is outcome reporting bias--also known as P-hacking, HARKing (hypothesizing after the results are known), statistical alchemy, or simply spin--but that aspect was unexamined.

Even if nonpublication were the only type of nonpublication, it would overstate the findings to suggest that this is a non-issue for industry-sponsored studies. With well under half of the studies unpublished, nonpublication is a significant problem for studies that were industry-sponsored AND for studies that without such funding. Granted, there is no difference between the two categories, but that's different from suggesting, as the title does, that virtually all of the trials were published.

On 2015 Nov 12, Erick H Turner commented:

To say "no publication bias" overstates the findings. Nonpublication is one aspect of publication bias (also known as reporting bias). Another important type is outcome reporting bias--also known as P-hacking, HARKing (hypothesizing after the results are known), statistical alchemy, or simply spin--but that aspect was unexamined.

Even if nonpublication were the only type of nonpublication, it would overstate the findings to suggest that this is a non-issue for industry-sponsored studies. With well under half of the studies unpublished, nonpublication is a significant problem for studies that were industry-sponsored AND for studies that without such funding. Granted, there is no difference between the two categories, but that's different from suggesting, as the title does, that virtually all of the trials were published.