Objectives: To evaluate biocorrosion and eccentric occlusal loading interplay in marginal quality of cervical restorations. Methods: Cervical wedge-shaped cavities were prepared in extracted premolars and restored with a composite. Premolars underwent either an erosive challenge (E: 1% citric acid/10 min), eccentric occlusal loading (EOL: 150 N/2.5 Hz/106 cycles), E before EOL (E + EOL), E intermediate to EOL (EOL/E/EOL), E after EOL (EOL + E), or no E or EOL (C: control). Marginal quality was analyzed based on a series of Optical Coherence Tomography images. Each of the margins was assigned a gap score (0, 1, 2, or 3) and measurement (??m). For each margin, scores data were analyzed with Kruskall Wallis and Dunn tests, and ??m data, with Kruskall Wallis. Overall and for each group, the different margins were compared using Wilcoxon signed-rank test, and the correlation between scores and ??m, Spearman???s correlation coefficient (?? = 0.05). Results: E and EOL, even if associated, did not influence enamel marginal quality. EOL/E/EOL impaired dentin/ cementum marginal quality only in the case of scores and compared to E. E + EOL, EOL + E or EOL and even C, without differences between each other, did not influence results differently from E or EOL/E/EOL. Margins in dentin/cementum always showed lengthier gaps. Except for C, E and EOL + E cervical margin, there was a strong positive correlation between scores and ??m. Conclusions: Eccentric occlusal loading and/or biocorrosion cannot be assumed as causes of marginal failure of cervical restorations in wedge-shaped cavities. A relevant concern may still be the establishment of adhesive interfaces in dentin/cementum. Clinical significance: Although non-carious cervical lesions are strongly being recognized multifactorial and their respective restorations not always behave as expected, biocorrosion and eccentric occlusal loading interplay cannot serve as an explanation for marginal gaps they often present.Funda????o de Amparo ?? Pesquisa do Estado de S??o Paulo (FAPESP)FAPESP: 13/09311-9; 19/20112-4; 14/13444-7