The Use of a Whole-Blood Benchtop Analyzer (Nova 16) in a Cardiac STAT Intensive Care Unit

Dept. of Pathol. and Lab. Med., The Valley Hospital, Ridgewood, NJ 07450
^a author for correspondence: e-mail IKAMPA{at}Valleyhealth.com

The Nova 16 STAT Chemistry Analyzer is designed to provide quick laboratory results of a chemistry profile consisting of glucose, blood urea nitrogen (BUN), creatinine, sodium, potassium, chloride, carbon dioxide, and hematocrit. The analyzer may be of use in the emergency department, the STAT laboratory, or as a mobile laboratory on a movable cart. We obtained the instrument to provide immediate results to the cardiac STAT intensive care unit for the above chemistry constituents. These analytes were previously performed on serum with a Beckman Instrument CX₃ analyzer with an ~45 min turnaround time. The major components of the 45 min required to provide these laboratory results were the time to allow specimens to clot, time for centrifugation, and internal laboratory transport times. The Nova 16 whole-blood analyzer eliminates the times required for specimen clotting and centrifugation. Bringing the analyzer physically closer to the patient treatment area or locating it in close proximity of a pneumatic tube system provides further reduction in the turnaround times. We evaluated the performance of the Nova 16 to determine the suitability of this instrument for use in the cardiac STAT intensive care unit and to ascertain the effect of the Nova 16 on turnaround times.

The Nova 16 uses biosensors to determine the concentrations of glucose, BUN, and creatinine (1). The glucose sensor consists of glucose oxidase bonded to a polymeric membrane substrate. ß-D-glucose is then oxidized to gluconic acid and hydrogen peroxide. The hydrogen peroxide produced is oxidized at the platinum electrode where the hydrogen peroxide is detected. The BUN assay uses an enzyme, urease, that has been chemically bonded to a membrane. The ammonium ion produced is detected with an ammonium ion-selective electrode. The creatinine sensor consists of three enzymes, creatinine amidohydrolase, creatinine amidinohydrolase, and sarcosine oxidase. Creatinine from the sample diffuses through the membrane where the creatinine is converted to hydrogen peroxide. The hydrogen peroxide is oxidized at the platinum electrode. The total carbon dioxide electrode consists of a gas permeable membrane and a pH electrode. Carbon dioxide, both free and protein-bound, is liberated by acidifying the sample. The liberated carbon dioxide diffuses through the membrane and dissolves in the internal filling solution where the resulting change in pH is quantitated by the pH electrode. Sodium, potassium, and chloride are measured using an ion selective electrode (2).

The correlation studies were performed on whole blood and plasma samples from patients for whom a chemistry profile was requested by the attending physician. The obtained specimens were aliquoted into two samples. One aliquot was centrifuged and then analyzed on a CX3; the remaining aliquot was analyzed on the Nova 16. An EP9 Method (Deming) regression analysis was performed using the EP Evaluator software (David G. Rhoads Associates). The results of the correlation studies were as follows:

Sodium: (y = 0.98x 2.3 mmol/L; r = 0.964; S_yx = 1.6; n = 126)

Potassium: (y = 0.94x 0.06 mmol/L; r = 0.952; S_yx = 0.2; n = 90)

Carbon dioxide: (y = 0.96x 0.3 mmol/L; r = 0.907; S_yx = 2.1; n = 113)

Chloride: (y = 0.98x 6.0 mmol/L; r = 0.956; S_yx = 1.8; n = 159; NOVA 16 values were adjusted by -4.0 mmol/L.)

Creatinine: (y = 1.01x - 1.77 µmol/L; r = 0.975; S_yx = 0.3; n = 93)

(y = 1.01x - 0.2 mg/L)

Glucose: (y = 1.0x 0.09 mmol/L; r = 0.991; S_yx = 8.1; n = 117)

(y = 1.0x 16 mg/L)

BUN: (y = 1.02x - 0.54 µmol/L; r = 0.997; S_yx = 5.7; n = 106)

(y = 1.02x - 15 mg/L)

Two concentrations of commercially available controls (BioRad) were analyzed with each run during the one-month evaluation period. These precision results are shown in Table 1 .

The linearities were evaluated using the Nova Linearity Solutions (Sets F-1 to F-4). The results were analyzed using the EP Evaluator Linearity Module (David G. Rhoads Associates). The following linearity ranges were confirmed: sodium (90–191 mmol/L), potassium (2.0–7.7 mmol/L), chloride (73–149 mmol/L), carbon dioxide (10–38 mmol/L), BUN (2.1–30.7 µmol/L), creatinine (35.4–1768.0 µmol/L), and glucose (1.9–27.8 µmol/L). The manufacturer claims a creatinine linearity from 17.7 to 1768 µmol/L. We found considerable scatter below 35.4 µmol/L. This is also reflected in the creatinine precision data at lower concentrations. We noticed a considerable flattening of the slope during the three-day life span of the membrane. As a result, the reduction in the absorbance-to-concentration ratio contributed to the high CV observed.

Turnaround times were shorter with the NOVA 16. The laboratory time from receiving the specimen to computer verification (mean ± 2 SD) dropped from an average of 45 ± 14 min to 11 ± 8 min. A similar reduction in the turnaround times was recently reported in another study of the NOVA 16 (3).

The analytical performance of the NOVA 16 was evaluated according to guidelines published by NCCLS (4). The data indicated that the performance of the NOVA 16 (except for creatinine) is equivalent to our CX3. This makes the Nova 16 attractive to use in a rapid response laboratory, an intensive care unit, or emergency room setting. We expect the daily maintenance, membrane replacement, and continuous monitoring of the analyzer to require the attention of an experienced laboratory-based technologist. Because of the continuous use of reagents during priming cycles, reagents are consumed rapidly during those cycles. As a result, the cost per reportable patient result is high if the analyzer is used sparingly.

References

1. Durst RA, Siggaard-Anderson O. Electrochemistry. Burtis CA Ashwood ER eds. Tietz textbook of clinical chemistry 2nd ed. 1994:178-182 WB Saunders Philadelphia. .
2. Tietz NW, Pruden EL, Siggaard-Anderson O. Electrolytes. Burtis CA Ashwood ER eds. Tietz textbook on clinical chemistry 2nd ed. 1994:1362-1363 WB Saunders Philadelphia. .
3. Zill M, Peterson V, Karcher R, Sykes E. Evaluation of the Nova 16 for improving turnaround times on emergency center specimens [Abstract]. Clin Chem 1997;43:S163.
4. NCCLS. Method comparison and bias estimation using patient samples: approved guidelines. NCCLS Document EP9-A. Wayne, PA: NCCLS, 1995:1–46..