Current issues of ACP Journal Club are published in Annals of Internal Medicine


Venous blood glucose determination was more accurate than fingerstick glucose in hypotension

ACP J Club. 1992 Jan-Feb;116:16. doi:10.7326/ACPJC-1992-116-1-016

Source Citation

Atkin SH, Dasmahapatra A, Jaker MA, Chorost MI, Reddy S. Fingerstick glucose determination in shock. Ann Intern Med. 1991 Jun 15;114:1020-4.



To evaluate the accuracy of fingerstick and venous blood glucose determination by glucose oxidase reagent strip in severely hypotensive patients.


Comparison of fingerstick and venous blood glucose measured by reagent strip with laboratory serum glucose analysis in hypotensive and normotensive patients.


Emergency department and intensive care unit of a university medical center.


41 normotensive patients and 31 severely hypotensive (systolic blood pressure ≤ 80 mm Hg) patients. Excluded were 2 control and 6 hypotensive patients in whom reagent strip or laboratory glucose levels were outside the range (0.56 to 22.76 mmol/L, 10 to 500 mg/dL) of the glucose meter.

Description of test and diagnostic standard

Fingerstick capillary glucose and venous blood glucose levels were determined using a Chemstrip bG reagent strip with an Accu-Chek II Glucose Monitor (Boehringer Mannheim). The venous sample, which was taken within 1 minute of the fingerstick, was also sent to the laboratory for serum glucose analysis using a Beckman Synchron Clinical System CX-3. Results of the diagnostic standard laboratory analysis were available in 1 hour.

Main results

Mean glucose levels in the normotensive control patients were 7.42 ± 0.75mmol/L (133.7 ± 13.5 mg/dL) for fingerstick determination; 7.69 ± 0.73 mmol/L (138.5 ± 13.2 mg/dL) for venous reagent strip; and 7.96 ± 0.73 mmol/L (143.4 ± 13.2 mg/dL) for laboratory testing. For the hypotensive patients, mean glucose was 7.42 ± 1.16 mmol/L (133.6 ± 20.9 mg/dL) for fingerstick determination; 10.60 ± 1.27 mmol/L (191.0 ± 22.9 mg/dL) for venous reagent strip; and 10.62 ± 1.28 mmol/L (191.3 ± 23.1 mg/dL) for the laboratory test. The mean fingerstick glucose value was 91.8% ± 1.6% of the laboratory glucose values in the control group, and 67.5% ± 5.7% in the hypotensive group (P < 0.001 for the difference). The mean venous reagent strip glucose value was 95.8% ± 1.1% of laboratory glucose values for the control group and 99.2% ± 2.5% of the hypotensive group. 90% of the control and 36% of the hypotensive fingerstick glucose values fell within a range of 20% of the laboratory glucose reference values.


Fingerstick glucose determinations did not accurately represent glucose values in severely hypotensive patients. Venous blood glucose testing by reagent strip provide a better, more rapid indication of glucose levels in these patients.

Source of funding: Not stated.

Address for article reprint: Dr. S.H. Atkin, Emergency Department, C384, University Hospital, 150 Bergen Street, Newark, NJ 07103, USA.


Although this investigation does not clarify the mechanisms responsible for the poor correlation between venous and fingerstick glucose measurements in hypotensive patients (which might have been affected by comorbid conditions), it does provide an important caveat for the use of reagent strip glucose determinations. Observations in neonates have determined that heelstick blood samples are less accurate than cord blood for glucose levels (1, 2), depending on the specific type of reagent strip and meter used. As suggested here, venous blood should be used when determining glucose with reagent strips in hypotensive adult patients. Otherwise, hypoglycemia will be overdiagnosed when euglycemia or hyperglycemia actually exists. This may lead to erroneous administration of intravenous glucose, albeit a relatively safe maneuver.

In the ambulatory diabetic patient, several studies attest to the excellent correlation of properly done fingerstick glucose determinations with standard laboratory methods (2, 3). In the outpatient setting, currently used reagent strips approximate capillary glucose within 10% of the venous blood level (3), well within the 20% used in this investigation.

Comparing fingerstick glucose determination with the venous reagent strip and routine plasma glucose in patients with various clinical disorders deserves further study to clarify the mechanism responsible for the present findings and to determine the optimal method of measurement for particular patients.

Robert Bloomfield, MD
Michael Rocco, MD
Carolyn Pedley, MD
Wake Forest University School of Medicine Bowman Gray CampusWinston-Salem, North Carolina, USA

Robert Bloomfield, MD
Wake Forest University School of Medicine Bowman Gray Campus
Winston-Salem, North Carolina, USA

Michael Rocco, MD
Wake Forest University School of Medicine Bowman Gray Campus
Winston-Salem, North Carolina, USA

Carolyn Pedley, MD
Wake Forest University School of Medicine Bowman Gray Campus
Winston-Salem, North Carolina, USA


1. Lin HC, Macguire C, Oh W, Cowott R. Accuracy and reliability of glucose reflectance meters in the high risk neonate. J Pediatr. 1989;115:998-1000.

2. Consensus statement on self-monitoring of blood glucose. Diabetes Care. 1987; 10:95-9.

3. Kelley WN, DeVita VT, Dupont HL, et al., eds. Textbook of Internal Medicine. Philadelphia: J.B. Lippincott; 1989:2316.