A systematic meta-analysis of oxygen-to-glucose and oxygen-to-carbohydrate ratios in the resting human brain

Abstract

Glucose is the predominant fuel supporting brain function. If the brain's entire glucose supply is consumed by oxidative phosphorylation, the molar ratio of oxygen to glucose consumption (OGI) is equal to 6. An OGI of less than 6 is evidence of non-oxidative glucose metabolism. Several studies have reported that the OGI in the resting human brain is less than 6.0, but the exact value remains uncertain. Additionally, it is not clear if lactate efflux accounts for the difference between OGI and its theoretical value of 6.0. To address these issues, we conducted a meta-analysis of OGI and oxygen-to-carbohydrate (glucose + 0.5*lactate; OCI) ratios in healthy young and middle-aged adults. We identified 47 studies that measured at least one of these ratios using arterio-venous differences of glucose, lactate, and oxygen. Using a Bayesian random effects model, the population median OGI was 5.46 95% credible interval (5.25-5.66), indicating that approximately 9% of the brain's glucose metabolism is non-oxidative. The population median OCI was 5.60 (5.36-5.84), suggesting that lactate efflux does not account for all non-oxidative glucose consumption. Significant heterogeneity across studies was observed, which implies that further work is needed to characterize how demographic and methodological factors influence measured cerebral metabolic ratios.

Fig 1. Modified PRISMA flow diagram.

Included studies were selected using the indicated selection criteria.

Fig 2. Forest plot for OGI meta-analysis.

Blue squares represent the reported mean OGI for each study. Black lines represent 95% confidence intervals. Numeric values for these quantities are also listed. The blue diamond is the population average from the Bayesian random effects meta-analysis. Error bars/values for the population mean are 95% CIs (n = 40).

Fig 3. Forest plot for OCI meta-analysis.

Same convention as in Fig 2 (n = 37).

Fig 4

Funnel plots for OGI (A) and OCI (B). In each plot, the reported study mean is plotted against its standard error. The population average is the dashed black line, its 95% percent CI is in dark gray, and its 95% prediction interval is in light gray. The lack of any asymmetry is evidence against substantial publication bias. The wide scatter around the population average, however, suggests that there is substantial heterogeneity between studies.