Affiliations: Department of Laboratory Medicine and Pathology University of Alberta Edmonton, Alberta, Canada,
Mayo Clinic College of Medicine Rochester, MN
Correspondence to: George S. Cembrowski, MD, PhD, Director, Medical Biochemistry, Laboratory Medicine and Pathology, 4B1.24, WMC, Edmonton, Alberta, Canada, T6G 2B7; e-mail: email@example.com
We thank Finkielman et al (May 2005)1 for addressing the important issue of the comparability of whole blood glucose measurements in the ICU setting to glucose levels measured in the clinical laboratory. The authors concluded that while, on average, bedside whole blood glucose measurements (SureStepFlexx; Lifescan; Milpitas, CA) provide an adequate measurement of laboratory-measured plasma glucose, the two measurements (bedside and laboratory measured) are quite different on an individual basis. The noncomparability is due to at least two major factors: susceptibility to interference by many currently available bedside glucose meters, and the authors’ incorrect assumption that venous blood glucose and capillary blood glucose are interchangeable.
The whole blood milieu of the average ICU patient can be very different from that of the average patient with diabetes: ICU hematocrits can be very low, levels of Po2 can be extremely high or low, and many patients have acid-base abnormalities. These three variables can significantly affect the whole blood glucose measurements by currently available analyzers.2–4 In fact, Lifescan cautions in its SureStepFlexx product literature that low hematocrits (< 25%) will render the measured glucose inaccurate. As the SureStepFlexx meter uses glucose oxidase to measure glucose, its operation is dependent on the Po2 of the blood. To prevent this Po2 dependence, at least one meter manufacturer uses glucose dehydrogenase rather than glucose oxidase. Unfortunately, glucose dehydrogenase reacts with maltose, a metabolite of ictodextrin, a substance used in peritoneal dialysis. This reaction will result in artifactually increased glucose in patients receiving this type of dialysis. If the authors have access to the patients’ hematocrit and Po2, can they provide some analysis to demonstrate the glucose-hematocrit and glucose-Po2 dependencies?
The study was comprised of two types of glucose comparisons grouped together in the results and conclusion. In one type of comparison, the same sample, drawn from either an arterial or central vein catheter, was measured at the bedside and in the laboratory. In the second comparison, bedside measurements from a capillary (finger-stick) origin were correlated to laboratory plasma measurements taken from a peripheral vein. Venous-capillary glucoses can vary considerably for up to 4 hours after a meal5 or after a patient has received insulin. Figure 1 in their article, the graph of differences between the laboratory-measured and bedside-measured glucose, does not distinguish the source of the samples compared. Without this information, it may be reasonable to conclude that some of the larger differences represent the nonidentical capillary and venous samples. Can these venous capillary comparisons be excluded and the Figure redrawn?
Dr. Cembrowski has received research support from the following organizations: Abbott Laboratories, HemoCue, Inc, Lifescan, Inc, and Roche Diagnostics.
We thank Drs. Binette and Cembrowski for their important and valid comments about our study (May 2005).1As they correctly pointed out, the whole blood milieu of the average ICU patient can be very different from that of the average patient with diabetes. Although glucose measurements can be affected by several critical care variables, few studies have looked at the performance of point-of care glucose testing in the ICU setting. Although hypotension was not associated with a difference between bedside and plasma glucose values in our study, Atkin et al2found that finger-stick glucose testing does not accurately represent venous glucose in patients with severe hypotension. Maser et al3also noted that the glucose concentration in arterial serum samples was significantly higher than the corresponding capillary whole blood (finger-stick) glucose by 9 mg/mL and 21 mg/dL, nonadjusted and adjusted to the hematocrit, respectively. Additionally, high (> 40%) and low (< 30%) hematocrits negatively and positively biased point-of care glucose testing devices; high arterial oxygen tension (> 250 mm Hg) produced a negative bias, while pH 6.80 to 7.55 did not affect glucose measurements substantially.4 Unfortunately, we are unable to perform the analyses suggested by Drs. Binette and Cembrowski since we did not obtain the hematocrit and arterial oxygen tension values at the time the blood samples were obtained in our study population.
Drs. Binette and Cembrowski speculate that the large differences between the bedside blood and plasma glucose measurements seen in our study were due to nonidentical capillary and venous samples. However, Ray et al5 reported results similar to ours. In a small prospective study5 of 105 arterial blood samples from 10 critically ill adults, they measured glucose at the bedside (One Touch Profile Blood Glucose Monitoring System; LifeScan; Milpitas, CA) and in the core chemistry laboratory and found the mean difference between the two methods to be − 0.72 mg/dL with the 95% limits of agreement between + 41 mg/dL and – 40 mg/dL.5
In spite of the limitations highlighted by Drs. Binette and Cembrowski, we believe that our conclusions are true and supported by our data. Although, on average, bedside glucose measurement provides a reasonable estimation of laboratory-measured plasma glucose, the former method gives an unreliable estimate of the latter in individual patients. This fact has an important clinical implication, particularly in this era of tight glucose control.6 A definitive prospective study on point-of-care glucose testing in the ICU setting seems warranted.
Become a CHEST member and receive a FREE subscription as a benefit of membership.
Individuals can purchase this article on ScienceDirect.
Individuals can purchase a subscription to the journal.
Individuals can purchase a subscription to the journal or buy individual articles.
Learn more about membership or Purchase a Full Subscription.
Institutional access is now available through ScienceDirect and can be purchased at myelsevier.com.
Some tools below are only available to our subscribers or users with an online account.
Download citation file:
Web of Science® Times Cited:
Customize your page view by dragging & repositioning the boxes below.
Enter your username and email address. We'll send you a reminder to the email address on record.
Athens and Shibboleth are access management services that provide single sign-on to protected resources. They replace the multiple user names and passwords necessary to access subscription-based content with a single user name and password that can be entered once per session. It operates independently of a user's location or IP address. If your institution uses Athens or Shibboleth authentication, please contact your site administrator to receive your user name and password.