Background PER3 is a circadian gene that contains a variable number of tandem repeats (VNTR) which codifies for three genotypes: 4/4; 4/5; and 5/5 and is involved in non-visual response to light, a critical process associated with bipolar disorder onset. Benedetti et al. (Neurosci Lett 445(2):184-7) related this VNTR with bipolar disorder age of onset and linked genotype 5/5 with an earlier onset. In this study, we aimed to investigate these associations of PER3 VNTR genotypes with age of onset in a homogenous sample of German patients with bipolar I disorder through Kaplan-Meier curves. Methods 45 patients were enrolled and divided into three groups according to PER3 VNTR genotypes. Recognizing common biological features, we built a combined group of -5 allele carriers (4/5 + 5/5). As a primary outcome, Kaplan-Meier analysis was conducted to delineate the three genotypes' influence on age of onset. The secondary Kaplan-Meier analysis aimed to evaluate the relation between the 4/4 homozygotes group and the combined group (4/5 + 5/5) with age of onset. Finally, we proceeded to compare groups through a Log Rank Test and performed an analysis of covariance (ANCOVA). Results The Kaplan-Meier analysis with three separate genotypes didn't replicate the findings of Benedetti's study. The analysis comparing genotype 4/4 with the combined group showed the influence of PER3 VNTR variants on the age of onset and relates genotype 4/4 to an earlier onset. ANCOVA between the combined and the 4/4 genotype groups, correlated genotype 4/4 with an increased number of depressive episodes. Conclusion This study showed no significant effect of PER3 VNTR genotypes on the age of onset and in linking genotype 5/5 with an earlier onset age. Contrasting results may arise from intrinsic differences between the two studies but also shed light on hypothetically different levels of functioning of PER3 VNTR genotypes in the context of bipolar pathology. Further studies will require bigger and more homogeneous clinical samples.