The Effect of Blood Sample Volume on 11 Glucose Monitoring Systems Free

The effect of variable blood sample volume on the accuracy of 11 glucose meters was studied to verify the reliability of self-monitoring of blood glucose. A total of 11 meters were assessed: OneTouch FastTake, OneTouch Basic, OneTouch Profile, and SureStep (LifeScan Canada, Burnaby, B.C., Canada); AccuSoft Advantage and AccuSoft Manager (Roche, Hoffman-LaRoche, Laval, P.Q., Canada); Precision Pen and Precision QID (MediSense Canada, Mississauga, ON, Canada); and Glucometer Elite, Glucometer Elite XL, and Glucometer DEX (Bayer, Toronto, ON, Canada). Venous blood collected from 16 fasting patients with diabetes was used to test each meter brand in triplicate. Sample volumes tested were of 1, 2, 3, 4, 5, 10, and 20 μl. Each patient contributed to the 5-μl sample plus two other sample volumes. The 5-μl volume, which is the usual volume required by the manufacturer of most meters, was considered the reference for comparison with other volumes tested, thus excluding the confounding effects of hematocrit, humidity, hypotension, and hypoxia. Several replicates of each volume size were tested. The number of times meters gave no result or an error message was recorded. Results were then calculated as percentages of the reference value and considered accurate if within 20% of the reference, as recommended by the Food and Drug Administration–ational Committee for Clinical Laboratory Standards (1). It has been recognized that most current meters do not comply with the 5% accuracy recommended by the American Diabetes Association (2–4).

All meters gave mostly nonmisleading results (accurate result or error message) at the 3-μl volume or above. However, below the 3-μl volume, most meters gave misleading results (Table 1). At the 1-μl volume, the mean calculated as a percentage from the 5-μl reference volume varied between 40 and 68%. OneTouch Basic and FastTake were the most likely to give a misleading result for a smaller blood volume as shown by Somers’ d statistic. Only four meters were reliable at all volumes showing no association with the Jonckheere tests: Precision QID, Precision Pen, AccuSoft Advantage, and AccuSoft Manager (data not shown).

Our results confirm that insufficient blood sample volume remains potentially a major source of user error with many of the current meters, including those with built-in detection of insufficient blood sample. Inaccurate results underestimated the real glucose value. Patients should be aware of this phenomenon since falsely low readings may result in unnecessary treatment of hypoglycemia and weight gain.