Aims

Liver cirrhosis increases the risk of developing dysglycaemia (pre-diabetes and diabetes), thus people with cirrhosis should undergo regular screening for dysglycaemia. The utility of screening using the laboratory glycated haemoglobin (HbA1c) test has been questioned in this setting. This study examines the relationship between different potential screening modalities: 75 g oral glucose tolerance test (OGTT) and HbA1c, using continuous glucose monitoring (CGM) as a comparator.

Methods

Participants ≥18 years with no known diabetes, were recruited from a gastroenterology cirrhosis surveillance register. Study measurements included a 75 g OGTT, laboratory HbA1c and two weeks of ‘blinded’ CGM (Freestyle Libre Pro). The possibility of intravascular haemolysis affecting HbA1c interpretation was also assessed.

Results

All 20 participants had compensated cirrhosis. OGTT tended to diagnose more dysglycaemia (N = 7) than did HbA1c (N = 4). Bland-Altman analysis showed laboratory and CGM-estimated HbA1c were broadly comparable, with a difference of 4mmol/mol (95% CI −3 to 12), or 0.4% (95% CI −0.3 to 1.1). Laboratory HbA1c tended to be higher than the CGM-estimated HbA1c, perhaps reflecting positive lifestyle changes in participants during their two weeks of wearing ‘blinded’ CGM (Hawthorne effect). In the population studied, there was no evidence that haemolysis affected interpretation of HbA1c results.

Conclusions

In the setting of compensated cirrhosis, the OGTT and HbA1c remain standard screening test for diabetes, but multiple studies show the OGTT diagnoses more people with dysglycaemia than does the HbA1c. Blinded CGM in an ambulatory, real world setting provides additional insights into glycaemic excursions but cannot be used to diagnose dysglycaemia.