In this issue of Diabetes Care, we read with interest the letter of Jungheim and Koschinsky (1) comparing glucose measurements using blood extracted from the finger versus blood extracted from the forearm. The phenomenon they discuss is not a simple function of measurement technology, but a complex function of circulatory physiology. Our cognizance and study of the phenomenon resulted in the explicit instruction to the users of the TheraSense FreeStyle blood glucose monitoring system to rub the test site before drawing blood. The increased perfusion from rubbing significantly reduces the difference in fingertip and forearm blood glucose measurements (see discussion below). It is significant that Jungheim and Koschinsky did not rub the test site. In addition, their protocol, which involved a glucose tolerance test followed by intravenous insulin, created physiological extremes and influenced the observed differences in study subjects.

An initial TheraSense study involved 100 subjects with both type 1 and type 2 diabetes, in which blood glucose measurements from the arm versus the finger were taken at random times throughout the day. Blood from the forearm (obtained without rubbing) and blood from the fingertip was measured for glucose concentration with the FreeStyle meter. The elevated intercept and low slope were unexpected (Table 1). This was further examined in a study where finger and arm blood samples (obtained without rubbing) were measured over several hours in patients with type 1 diabetes. The study indicated that changes in blood glucose are first detected in finger blood and lag in forearm blood (Fig. 1). A similar study, which included blood from the forearm before and after rubbing, indicated that arm and finger differences are reduced by rubbing the test site (Fig. 2). This led to a final study of 120 subjects with type 1 and type 2 diabetes, in which blood obtained from the forearm (after rubbing) and blood obtained from the finger was measured for glucose concentration using the FreeStyle meter. The comparison yielded a correlation that was nearly ideal (Table 1). The final study also assessed the accuracy and clinical utility of the FreeStyle meter by comparing venous finger and arm (with rubbing) blood to YSI-plasma readings (Table 1).

The current state of our research indicates that there would be very little difference in therapeutic decisions when the arm (following rubbing) rather than the finger is used as the test site. However, when blood glucose concentration is falling rapidly, the lag in glucose change could cause a delay in the detection of hypoglycemia. Accordingly, when testing with the express purpose of detecting hypoglycemia (such as when symptoms of hypoglycemia are present or when a meal has been delayed after taking insulin), the finger may be the preferred test site. Our studies indicate that a delay in the detection of hypoglycemia is not a common occurrence in routine testing on the arm. However, this phenomenon must be seriously considered and thoroughly understood. TheraSense has undertaken additional studies under a variety of circumstances to better understand this complex question of circulatory physiology. These studies will be the subject of future publications.

We feel that the obvious benefits of new technologies should not be overshadowed by the manageable risks. The introduction of fingertip blood glucose testing enabled the aggressive management of glucose levels in people with diabetes, yet it also increased the frequency of hypoglycemic events. Clearly, the benefits of intensive insulin therapy outweighed the risks of hypoglycemia. Similarly, greatly reducing the pain associated with blood glucose testing by permitting testing on the forearm and other sites is likely to have a significant positive impact on compliance with monitoring regimens.

Figure 1 —

Glucose profile of arm (no rubbing) and finger. —□—, average finger; —⧫—, average arm; —▴—, difference. For average finger and average arm, the data points are the average of duplicate measurements.

Figure 1 —

Glucose profile of arm (no rubbing) and finger. —□—, average finger; —⧫—, average arm; —▴—, difference. For average finger and average arm, the data points are the average of duplicate measurements.

Close modal
Figure 2 —

Glucose profile of arm (no rubbing), arm (with rubbing), and finger. —□—, Average finger; —⧫—, arm unrubbed; —▴—, arm rubbed; —◊—, difference unrubbed; —▵—, difference rubbed. For average finger, the data points are the average of duplicate measurements.

Figure 2 —

Glucose profile of arm (no rubbing), arm (with rubbing), and finger. —□—, Average finger; —⧫—, arm unrubbed; —▴—, arm rubbed; —◊—, difference unrubbed; —▵—, difference rubbed. For average finger, the data points are the average of duplicate measurements.

Close modal
Table 1 —

Linear regression statistics and Clarke Error Grid Analysis

ComparisonIntercept (mg/dl)SloperClarke Error Grid Zones (% of readings in the zone)
ABCD
Arm (no rubbing) vs. finger 19.4 0.913 0.956 NA NA NA NA 
Arm (rubbing) vs. finger −0.5 1.027 0.971 NA NA NA NA 
Venous vs. venous YSI 7.1 0.923 0.992 99.6 0.4 
Finger vs. finger YSI 6.6 0.934 0.982 98.3 1.7 
Arm (rubbing) vs. finger YSI 9.0 0.945 0.967 87.7 11.4 0.8 
ComparisonIntercept (mg/dl)SloperClarke Error Grid Zones (% of readings in the zone)
ABCD
Arm (no rubbing) vs. finger 19.4 0.913 0.956 NA NA NA NA 
Arm (rubbing) vs. finger −0.5 1.027 0.971 NA NA NA NA 
Venous vs. venous YSI 7.1 0.923 0.992 99.6 0.4 
Finger vs. finger YSI 6.6 0.934 0.982 98.3 1.7 
Arm (rubbing) vs. finger YSI 9.0 0.945 0.967 87.7 11.4 0.8 

NA, not applicable.

1
Koschinsky T, Jungheim K: Risky delay of hypoglycemia detection by glucose monitoring at the arm.
Diabetes Care
24
:
1303
–1304,
2001

Address correspondence to Geoff McGarraugh, TheraSense Inc. 1360 South Loop Rd., Alameda, CA 94502. E-mail: [email protected].

G.M. is employed by and holds stock in TheraSense, which manufacturers and markets the FreeStyle blood glucose monitoring system.