Associations were classified as highly suggestive (Class II) if they involved more than 1000 participants, exhibited highly significant summary associations (p value <10 − 6 by random-effects) and had the largest study showing a nominally significant result (p<0.05).
Meta-analyses were deemed free from biases and classified as convincing (Class I) if they satisfied the following criteria: p value <10 − 6 in random-effects meta-analysis; inclusion of more than 1000 participants; low or moderate between-study heterogeneity (I2 <50%); 95% PI excluding the null value; no evidence of small-study effects or excess significance bias; and the largest study showing a nominally significant result (p<0.05).