Discrimination of Fasting and Fed Manometric Gastric and Small Bowel Contraction Morphology in Healthy Humans Using Principal Components Analysis

Abstract

Introduction: The fasting migrating motor complex (MMC) serves a housekeeper function to clear the stomach and small bowel of undigested debris; the fed motor pattern spreads nutrients along the mucosa to optimize digestion. The MMC phase 3 is a distally migrating complex comprised of regular contractions while phase 2 and fed state contractions have more variable amplitudes suggesting different electrophysiologic mechanisms for induction vs. phase 3. Such manometric interpretations are typically done visually and are too qualitative to be amenable for automated analysis. We employed principal component analysis (PCA) to decompose variations in MMC phase 3 contractions vs. phase 2 and the fed state to quantify how contractile morphology varies in the different phases. Methods: 6 healthy subjects underwent placement of a 3 meter manometry catheter with ports in the antrum, duodenum, and jejunum. Manometry was performed for 4-12 hours. Principal components (PCs) were generated from the 3 seconds before and after each contractile peak: PC1 controlled maximal amplitude; PC2 reflected contractile width; PC3 controlled whether a contractile peak occurred early or late in the waveform. Results: 10,543 contractions were analyzed (5576 MMC phase 3, 2543 phase 2, 2334 fed). PC1 variations were higher during phase 2 and the fed state vs. phase 3, reflecting greater amplitude stability in phase 3 (Table 1). PC2 and PC3 variations were less than for PC1 but were greater in phase 2 and the fed state vs. phase 3, indicating less variability in peak width and timing in phase 3. 69% of phase 3 contractions had early peaks and 31% had late peaks, but PC3 lag autocorrelations between consecutive peaks were weak (r=-0.09) indicating poor temporal contraction shape stability. PC1 scores of phase 3 contraction amplitudes showed temporal variability, being highest 2.5 minutes after the start of the phase 3 complex and then staying constant after 5 minutes (Figure 1). Gastric phase 3 was shorter in duration vs. duodenojejunal complexes.Figure 1Table 1: Principal Component Variability Between Different Manometric PhasesConclusion: Principal component analyses defined differences in contractile amplitude, width, and shape that distinguished MMC phase 3 contractions from those during phase 2 and the fed state. These findings (i) warrant physiologic studies into electrophysiologic mechanisms underlying waveform generation in the different fasting and fed phases and (ii) form a foundation for automated intestinal manometry interpretation soft ware development.