On top of that, the actual results are behind a paywall and can be very iffy. It sounds like there were only 12 people in the 6-hour group and in each of the other groups. And no indications about other traits like sex, smoking or other habits, and so on. Too small numbers to guarantee against statistical fluctuations. And the "significant" in the abstract may indicate that they used p-values to quantify their results, which are today considered iffy by a large chunk of the statistics community...