| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Letters to the Editor |
1 Primary Children’s Medical Center Department of Pathology Salt Lake City, UT
2 Utah Valley Regional Medical Center, Provo, UT
3 Dixie Regional Medical Center St. George, UT
4 Primary Children’s Medical Center Salt Lake City, UT
5 McKay-Dee Hospital Center Ogden, UT
aAddress correspondence to this author at: Primary Children’s Medical Center, Department of Pathology, 100 N. Medical Dr., Salt Lake City, UT 84113-1100. Fax (801) 662-2165; e-mail phil.bach{at}intermountainmail.org.
To the Editor:
C-reactive protein (CRP) is measured on VITROS® Chemistry Systems (Ortho-Clinical Diagnostics) using a "MicroSlide" method. At our hospitals plasma CRP concentrations <10 mg/L are used as an indicator that it is safe to discontinue antibiotic therapy in neonates with known or suspected sepsis (1). After implementing the neonatal sepsis protocol, we found that the VITROS CRP slide gave results up to 33 mg/L higher in neonatal specimens (Fig. 1
, open triangles) than the Behring Nephelometer II (BNII, Dade Behring) high-sensitivity CRP (hsCRP) assay, a well-characterized method for CRP measurements(2).
|
According to the manufacturer’s instructions, hemolysis falsely increases VITROS CRP slide results, but the increases we observed were not limited to hemolyzed specimens. It has been reported (3) that collecting specimens in serum separator tubes increases CRP values measured by the VITROS CRP slide; however, the differences we saw were much greater than those attributed to serum separator tubes. Because the falsely increased results in neonates were at or near our decision point of 10 mg/L, we concluded that the VITROS CRP slide method was not suitable for use with our neonatal sepsis protocol. We began referring neonatal specimens to laboratories able to run CRP on the BNII, the IMMULITE® immunoassay system (Diagnostic Products), or the COBAS Integra® 400 (Roche Diagnostics). We had previously shown (data not included) that the IMMULITE and Integra CRP methods correlated very well with the BNII.
At this point we purchased the VITROS® 5,1 FS Chemistry System, which in addition to analyzing CRP by the MicroSlide method, also determines hsCRP by an immunoturbidimetric MicroTip method. According to the manufacturer, specimens with hsCRP greater than the reportable range of 15 mg/L should not be diluted but retested using the MicroSlide method. Because the MicroSlide method was not suitable for neonates, and we were seeking a faster alternative to sending specimens to another laboratory, we evaluated the VITROS hsCRP MicroTip method, including dilutions, for use with neonates (age <6 months). We tested 249 serum and heparinized plasma specimens left over after clinical testing by one of the other analyzers. This sample set included 143 specimens by the BNII (range 0.2–351 mg/L), 95 by the IMMULITE (range 0.1–139 mg/L), and 11 by the Integra (range 1.0–49 mg/L). Specimens with CRP concentrations >15 mg/L were diluted with saline by the instrument and retested. The institutional review boards of the University of Utah and Intermountain Healthcare gave approval for submission of this data.
Our results are shown by the solid squares in Fig. 1
. The VITROS hsCRP method showed markedly better agreement with the BNII, IMMULITE, and Integra than the MicroSlide CRP method, especially at our decision point of 10 mg/L. The simple linear regression line between VITROS hsCRP MicroTip and the 3 comparative methods, taken together, had a slope of 1.16 (5%–95% CI 1.14–1.17), an intercept of –0.9 mg/L, and R = 0.996. Regression analysis data for individual instruments were as follows: BNII slope 1.16, intercept –0.6 mg/L, R = 0.997; IMMULITE slope 1.12, intercept –0.9 mg/L, R 0.987; and Integra slope 1.14, intercept 0.6 mg/L, and R = 0.992.
Using specimens from both neonates and older patients, we found that the positive bias was introduced by diluting the specimen, manually or by the instrument, with deionized water, saline, or the 0 calibrator. The bias was 16% using a 10-fold dilution and 21% using a 30-fold dilution. Undiluted specimens showed no bias. We speculate that the dilution bias may be the reason that diluting the specimen is not recommended by the manufacturer. Although the hsCRP method is positively biased when the specimen is diluted, the bias is reproducible and relatively small, and does not preclude using the test to manage neonates with sepsis.
Acknowledgments
Grant/funding support: None declared.
Financial disclosures: None declared.
References
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
, VITROS CRP Slide vs Behring Nephelometer II. 
, VITROS hsCRP MicroTip method vs Behring Nephelometer II, DPC IMMULITE, or Roche Integra. Dotted vertical line is the sepsis protocol decision point of 10 mg/L.