Associations of plasma glycerophospholipid profile with modifiable lifestyles and incident diabetes in middle-aged and older Chinese

Aims/hypothesis Glycerophospholipid (GPL) perturbance was linked to the pathogenesis of diabetes in animal studies but prospective studies in humans are rare, particularly in Asians. We aimed to investigate the associations between plasma GPLs and incident diabetes and to explore effects of lifestyle on the associations in a Chinese population. Methods The study included 1877 community-dwelling Chinese individuals aged 50-70 years (751 men and 1126 women), free of diabetes at baseline and followed for 6 years. A total of 160 GPL species were quantified in plasma at baseline by using high-throughput targeted lipidomics. Log-Poisson regression was used to assess the associations between GPLs and incidence of diabetes. Results Over the 6 years of follow-up, 499 participants (26.6%) developed diabetes. After multivariable adjustment, eight GPLs were positively associated with incident diabetes (RRper SD 1.13-1.25; all false-discovery rate [FDR]-corrected p < 0.05), including five novel GLPs, namely phosphatidylcholines (PCs; 16:0/18:1, 18:0/16:1, 18:1/20:3), lysophosphatidylcholine (LPC; 20:3) and phosphatidylethanolamine (PE; 16:0/16:1), and three reported GPLs (PCs 16:0/16:1, 16:0/20:3 and 18:0/20:3). In network analysis, a PC-containing module was positively associated with incident diabetes (RRper SD 1.16 [95% CI 1.06, 1.26]; FDR-corrected p < 0.05). Notably, three of the diabetes-associated PCs (16:0/16:1, 16:0/18:1 and 18:0/16:1) and PE (16:0/16:1) were associated not only with fatty acids in the de novo lipogenesis (DNL) pathway, especially 16:1n-7 (Spearman correlation coefficients = 0.35-0.62, p < 0.001), but also with an unhealthy dietary pattern high in refined grains and low in fish, dairy and soy products (|factor loadings| >= 0.2). When stratified by physical activity levels, the associations of the eight GPLs and the PC module with incident diabetes were stronger in participants with lower physical activity (RRper SD 1.24-1.49, FDR-corrected p < 0.05) than in those with the median and higher physical activity levels (RRper SD 1.03-1.12, FDR-corrected p >= 0.05; FDR-corrected p(interaction) < 0.05). Conclusions/interpretation Eight GPLs, especially PCs associated with the DNL pathway, were positively associated with incident diabetes in a cohort of Chinese men and women. The associations were most prominent in participants with a low level of physical activity.