Abstract
Home monitoring — why look beyond the ears?
An important part of the long-term management of diabetic cats is the measurement of blood glucose at home by the owner to generate blood glucose curves. 1 The main advantage of home monitoring is the avoidance of stress hyperglycaemia associated with veterinary consultation and hospitalisation. Blood glucose curves enable assessment of insulin efficacy, the time point of the glucose nadir (peak insulin effect), the duration of insulin effects and fluctuations in blood glucose. They help to identify any required adjustments in the insulin dosage, as well as potentially life-threatening hypoglycemia or associated Somogyi phenomenon. 2 Tight blood glucose control with multiple daily measurements and insulin adjustments offers very good prospects of remission. 3
Blood glucose testing is traditionally performed with hand-held glucometers using the marginal ear vein nick technique or the vacuum lancet (‘Vaculance’) method (Fig l). 4,5
Traditional method of blood glucose testing. (a) Capillary blood sampling at the pinna using the Vaculance method; (b) glucose measurement with a hand-held glucometer. No restraint of this cat was necessary for either procedure. Courtesy of Robert Polster
However, in one study, 8/26 cat owners were not able to perform glucose measurements. 1 One of the most common reasons for discontinuation of home monitoring is the inability to obtain an adequate blood volume. The development of aural haematomas is an exception. 6
Anecdotally, we know that the metacarpal and metatarsal pads have been used as an alternative sampling site, which seems reasonable as the paws are accessible and have a good blood supply. We therefore set out to investigate the clinical value and accuracy of metacarpal/tarsal pad blood glucose measurements as an alternative to sampling from the inner surface of the pinna.
Clinical study
Seventy-five client-owned hospitalised cats of various breeds, presented over a period of 2 months, were included in this cross-sectional study. The cats ranged in age from 5 months to 15 years (mean 8.2 years ± SD 4.6 years); 49 were male (46 castrated) and 26 were female (all spayed). The mean weight was 4.65 kg (SD 1.39 kg, range 2.00–7.70 kg) and 9.3% had been diagnosed with diabetes mellitus.
Samples were taken from unfasted patients and measured using the FreeStyle Freedom glucometer (Abbott Diabetes Care, Alameda, CA, USA), which has been validated for use in cats. 7 Its useful range is 1.1–27.8 mmol/l glucose and the sample size is 0.3 μl. Results are displayed within 5 s. The intra-assay precision is 0.998 and the coefficient of variation is 3.6%. In cats, results are 16% lower on average than with the reference method. 7 Lancing was performed either with a lancing device (Microlet Vaculance; Bayer HealthCare, Mishawaka, USA) at the ear pinna, after vigorously rubbing the ears, or with a standard lancet (Glucoject Dual; Menarini Diagnostics, Wokingham, UK) at the paws. Due to the pulvinate and small surface area of the large pads, the application of a vacuum lancet was not possible at this site.
In the first 50 cats, ear testing was compared with metacarpal pad testing (Fig 2), whereas in the subsequent 25 patients it was compared with metatarsal pad testing. The order of site testing (ear/pad or pad/ear) was randomised by flipping a coin. Measurements were performed in sternal recumbency on a table in the intensive care unit. Cats were minimally restrained by a nurse and measurements were performed by one of the authors. The limbs were flexed caudolaterally with one hand and lancing was performed with the other hand. The second measurement was always performed within a minute of the first. One minute after lancing, the paws were checked for secondary bleeding.
Alternative-site testing at the metacarpal pad. (a) Lancing with a standard lancing device; (b and inset) glucose measurement with a hand-held glucometer. Courtesy of Robert Polster
A maximum of three attempts to obtain a blood sample was made per cat and sampling site. The number of successful first attempts was recorded, as were any vocalisations, twitching of any part of the body and attempts to escape or attack. Highly fractious, dehydrated, icteric or febrile cats, as well as patients with obvious orthopaedic or neurological problems, were excluded. No cat had received analgesics or sedatives. Blood pressure was not recorded. The study was approved by the veterinary ethics committee of Austria (GZ 68.205/0216-II/10b/2009).
Normal distribution of data was tested using the Kolmogorov-Smirnov test. If normal distribution was not certain, statistical comparisons were undertaken with the Wilcoxon rank sum test. The χ 2 test was applied for categorical variables. Pearson's coefficient of correlation was used to describe the relationship between ear and alternative-site glucose concentrations. P values <0.05 were considered statistically significant.
Results
Blood samples were easy to obtain from the metacarpal/tarsal paw pads using a standard lancet and the cats tolerated this sampling method better than expected. Cats that tolerated blood collection from their ears also tolerated sampling from their paws, and vice versa. Although the initial plan had been to test the metatarsal pad in 50 cats, this part of the study was stopped after 25 cats as the extraction of the hind limb from beneath the body was actively resisted by nearly one-third of patients. During restraint, attempts to attack were not observed. One cat struggled vigorously and measurements were not possible. Twitching was observed in 21% of cats (16/75) during ear sampling and 33% of cats (25/75) during paw sampling. This difference was not significant (χ 2 = 2.719, P = 0.099). Vocalisation was observed in 6.7% of cats (5/75) in both groups. If the initial blood drop was too small, gently squeezing the pad almost always yielded an adequate volume. There was a trend for larger drops to be formed on the paws than on the ear.
During the study, a total of 235 punctures was performed (131 on ears, 68 on front paws, 36 on hind paws). No difference in first-attempt success rate was observed between the sites (ears [47%] vs metacarpal pads [70%], χ 2 = 1.825, P = 0.177; ears vs metatarsal pads [64%], χ 2 = 0.693, P = 0.405; ears vs both pads [68%], χ 2 = 1.901, P = 0.168; metacarpal vs metatarsal pads, χ 2 = 0.505, P = 0.477). On average, 1.6 attempts were needed to achieve a successful measurement from the ear site, compared with 1.3 attempts for the paw sites. It was not possible to obtain a sample from the ear in 9.3% of cases (7/75), from the metacarpal pad in 4% of cases (2/50) and from the metatarsal pad in 8% of cases (2/25). The major reason was an inability to generate a sufficient blood volume. Bleeding stopped within 1 min in all cats.
Scatter plot of the differences in blood glucose concentrations (mmol/l) in capillary samples obtained from the ears and paws of 66 cats. The dashed line represents the regression line
Glucose measurements on samples from the ear ranged from 2.39–27.03 mmol/l (mean 7.23 mmol/l, SD ± 4.41 mmol/l). Glucose differences between sites were small (ear/metacarpal pad: mean 0.05 mmol/l, SD ± 0.63 mmol/l, range −1.05 to 1.44 mmol/l; ear/metatarsal pad: mean −0.02 mmol/l, SD ± 0.78 mmol/l, range −1.78 to 1.39 mmol/l) and insignificant (metacarpal pad/ear: Z = −0.428, P = 0.669; metatarsal pad/ear: Z = −0.348, P = 0.728; see Fig 3). The correlation between the ratios metacarpal pad/pinna and metatarsal pad/pinna was r = 0.993 (P <0.001) and r = 0.989 (P <0.001), respectively.
Discussion
The feasibility of capillary blood glucose measurements in cats has been demonstrated in several studies. 1,3,6,8 Most owners of diabetic cats are willing and able to perform home monitoring. 9 Although glucometers are in regular use, so far only the pinna of the ear has been validated as a sampling site. This is in contrast to human patients where samples for glucose measurement are often taken from the palm, forearm, thigh and ear lobe, as well as from the fingertips. 10,11 In this study, the paws, especially the metacarpal pads, were an excellent alternative sampling site in cats, with a comparable first-attempt success rate. An important advantage of this alternative site is that the use of standard lancets is possible. The drop size can easily be increased by gently squeezing the pad. Although we had expected significant resistance when puncturing the pads, the procedure was tolerated surprisingly well. Only the positioning of the hind limb proved to be a problem — causing resistance in many cats.
No difference to ear sampling was observed in respect of either the cat's reactions to pricking or the measured glucose concentrations. Of course, this finding only applies to a single measurement under steady-state glucose concentrations. Inequalities might be found after repeated lancing, when skin soreness might become relevant, and/or during rapid changes in blood glucose concentrations. In human diabetics, delays in observed blood glucose kinetics between different sampling sites can be observed if the blood glucose is changing dynamically. 12,13 This might also apply to cats and glucose concentrations obtained by capillary blood glucose measurement should be interpreted cautiously. As blood sampling was performed by an experienced veterinarian, less than 10% of the patients were diabetics and health status may have affected tolerance in this study, further work is necessary to assess the utility of this technique, especially in the home environment.