The Mouthpiece Type Factor: Impact of Filter Selection on Secondary Electrospray Ionisation High-Resolution Mass Spectrometry (SESI-HRMS) Breath Profiles.

Molecular breath research using secondary electrospray ionisation high-resolution mass spectrometry (SESI-HRMS) commonly uses single-use mouthpieces to facilitate sampling and reduce contamination risk. Many incorporate bacterial/viral filters, but their influence on untargeted molecular breath profiles is poorly characterised. Objective: To determine whether commercially available mouthpieces alter SESI-HRMS breath profiles, identify molecular classes affected by filtered mouthpieces, and assess whether ComBat batch correction reduces mouthpiece-related variance while preserving inter-subject information. Methods: Three healthy volunteers completed seven testing sets across three measurement days. Each set included all seven mouthpiece types, yielding 147 planned breath measurements, of which 143 were valid. Six mouthpieces incorporated filter membranes; one unfiltered mouthpiece served as reference. Breath was analysed in real time using SESI-HRMS. Data were pre-processed, assessed for mouthpiece effects using guided principal component analysis, corrected using ComBat, and mouthpiece-specific features were putatively annotated against the Human Metabolome Database. Results: Pre-processing yielded 3,718 reproducible breath features. Mouthpiece type produced a strong batch effect (D = 0.9969, p < 0.001), with three clusters: A/B, C/D/E/F, and the unfiltered reference. Mouthpiece type accounted for 42.7% of total feature variance. The unfiltered reference detected 397 features not reproducibly detected with filtered mouthpieces, enriched in low-molecular-weight features putatively assigned to amino-acid-related compounds, short-chain aldehydes, and amines. Filtered mouthpieces contributed mouthpiece-specific features consistent with potential material-related background signals. ComBat reduced mouthpiece-type clustering (D = 0.1407, p = 0.998) while preserving most inter-subject variance. Conclusions: Mouthpiece type is a major determinant of detected SESI-HRMS breath profiles. Filtered mouthpieces may attenuate selected low-molecular-weight polar features and introduce mouthpiece-specific background signals. Batch correction reduces systematic mouthpiece-related variance but cannot recover features not transmitted by the sampling interface. Consistent mouthpiece use, explicit reporting, and mouthpiece-aware harmonisation are essential for reproducible molecular breath research.