A developmental origin for the craniofacial endotype of obstructive sleep apnea: the premaxilla and the reinterpretation of growth rotation

PREPRINT - submitted version, not peer reviewed. Manuscript as submitted to Sleep and Breathing (Springer) on 3 November 2026; currently under peer review. Content may change; the version of record will be linked here. Abstract Background The craniofacial endotype of obstructive sleep apnea—maxillary constriction, posterior mandibular rotation, increased anterior facial height and a caudal hyoid—is well characterised; its origin is not. The spatial matrix hypothesis attributes it to genetic and epigenetic modulation of gene expression. This account, like growth-modification roadmaps, inherits an orthodontic premise settled on the Copenhagen implant studies, which recorded neither functional nor genetic variables, rested on five to twenty-one subjects, and never measured the premaxillary–maxillary suture. Main body This review connects four published hypotheses—hyoid descent, perinatal thiamine insufficiency, the premaxilla as a nasopalatine sensory gateway, and vitamin D endocrine reserve—into one developmental model. Perinatal metabolic reserve and oxygenation shape premaxillary and palatal form; premaxillary geometry and nasopalatine afference calibrate brainstem setpoints for tongue posture and breathing mode; the resulting loading determines vertical rotation, hyoid position and airway collapsibility, whose intermittent hypoxia suppresses condylar growth and closes the loop. Primate and clinical studies show rotation reversing with breathing mode; twin heritability is reinterpreted as gene–environment interaction, since co-twins share the perinatal window; ankyloglossia and clefts serve as natural experiments. The model differs from the spatial matrix hypothesis at the level of mechanism. Conclusions Growth rotation, and with it the craniofacial endotype, is proposed to be the outcome of a definable perinatal and sensory window, not a genetic programme. Discriminating tests are specified: exposure-discordant sibling and twin designs, premaxillary sensory phenotyping, perinatal cohorts, interceptive trials and re-analysis of the implant films. If supported, the window for influencing the endotype closes earlier than either discipline assumes. Context. This review is part of the PremaxProject and synthesises four previously published component hypotheses.