Our main outcome was glycosylated haemoglobin A (HbA1c), a measure of the average blood glucose levels for the last 8–12 weeks.10 (link) Potassium EDTA venous blood was analysed for HbA1c at the examination centre using the HemoCue HbA1c 501 system (HemoCue, Ängelholm, Sweden) according to the manufacturer’s instructions.11 More than 90% of samples were analysed for HbA1c within 2 hours.
As a secondary outcome, participants who came to the examination in the morning after a 12-hour fast were also tested for capillary blood fasting plasma glucose (FPG) using the HemoCue Glucose 201 RT (HemoCue). FPG was analysed immediately after sampling. Because of logistic constraints, participants were not randomised for FPG testing; we tested participants who were able and willing to come fasting in the morning.
The device used to measure HBA1c had a lower limit of quantitation (LOQ) of 4% NGSP=20 mmol/mol. All results1c (see online supplementary appendix 1 ).
As a secondary outcome, participants who came to the examination in the morning after a 12-hour fast were also tested for capillary blood fasting plasma glucose (FPG) using the HemoCue Glucose 201 RT (HemoCue). FPG was analysed immediately after sampling. Because of logistic constraints, participants were not randomised for FPG testing; we tested participants who were able and willing to come fasting in the morning.
The device used to measure HBA1c had a lower limit of quantitation (LOQ) of 4% NGSP=20 mmol/mol. All results