Return to CAP Home
Printable Version

  Q and A






November 2007

Q. Preadmission (or on admission) testing for methicillin-resistant Staphylococcus aureus, or MRSA, with molecular (polymerase chain reaction) or conventional culture has decreased hospital infection rates when the detection of MRSA is acted on within a short time after admission (with appropriate isolation). Are patients colonized with MRSA (for example, nasal carriers) simply discharged to return to their nursing homes or other places of residence without treatment, or is some form of therapy, local or otherwise, indicated?

A. The primary purpose of screening for MRSA on admission is to quickly identify colonized patients so they can be placed into contact isolation to stop the spread of MRSA to additional patients. The rationale for decolonizing those who already harbor MRSA in their nose is based on the hypothesis that doing so will prevent future disease in those already colonized by removing MRSA. However, decolonization is of unproven benefit at this time. Thus, which strategy to choose is up to the individual health care organization. Some may elect to perform only admission surveillance. Others may choose, as does Evanston (Ill.) Northwestern Healthcare, to decolonize patients who are found to be nasally MRSA positive by using a simple regimen of mupirocin ointment rubbed into the anterior nares twice a day for five days, accompanied by a chlorhexidine bath or shower on days one, three, and five of the mupirocin application. For ENH, the strategy of active surveillance followed by contact isolation and decolonization of MRSA carriers has led to an 80 percent reduction of MRSA bloodstream infections in the first year of our program.

Lance R. Peterson, MD
Director of Microbiology and Infectious Diseases Research
Clinical Microbiology
Department of Pathology and Laboratory Medicine
Associate Epidemiologist
Evanston Northwestern Healthcare
Evanston, Ill.

Q. A few of our doctors have expressed concern about some of their patients' successive hematocrits, which are drawn several times during a 24-hour period. I know there are a number of preanalytic factors that can influence hematocrit, ranging from the patient's physical state (renal, glucose, and hydration status) to basic phlebotomy technique. For example, one patient had a 25 percent hematocrit result at 8 AM, a 27 percent hematocrit result at 4 PM, and then was back down to a 25 percent hematocrit result at midnight—with no transfusion in between.

Can you provide information that I can use to try to convince two clinicians that they must consider preanalytic factors? We repeated all samples again on our Advia 2120 hematology systems. We also do frequent manual-spun hematocrit checks on the Advia result when we have a CHCM versus MCHC flag. When we run quality control material, it is always within one standard deviation of the target value for the QC product as specified by its manufacturer. We also QC all parameters three times per day, so the analytic part is fine. Our allowable Advia reproducibility is a hematocrit result that is within a range of two percent above or two percent below the first determination. All samples retested were within a one percent hematocrit result of the original test results.

A. Several preanalytic factors can affect hematocrit determination, which could account for the observed slight variation in hematocrit results. One is the state of the intravascular volume of the patient. Hydration changes and fluid shifts can affect intravascular volume and lead to changes in hematocrit as red blood cells are confined to the intravascular space, whereas the water and electrolytes in plasma are not. Changes in the patient's state could also account for the small variation reported. The source of the specimen can influence the result as well. Because fluid shifts normally occur at the capillary level, capillary hematocrits (obtained by finger prick) often yield a hematocrit two or three units higher than a venous hematocrit (obtained by venipuncture). Analytic variation could also account for the observed results. An instrument operating correctly will repeat hematocrit values within three to five percent. At the hematocrit described in the example, this would be about one unit, making the observed variation within the reproducibility of the instrument—for example, the 25 could have been a 26 or the 27 could have been a 26.

One caution should be noted in using spun hematocrits when comparing to instrument-generated hematocrits: An instrument hematocrit is a calculated value (the instrument measures the red blood cell count and mean corpuscular volume), so it is actually the hematocrit with no trapped plasma. Spun hematocrits have some degree of plasma trapping. If the patient has red blood cell alterations that interfere with red blood cell stacking, spun hematocrits can be multiple units higher than the calculated hematocrit obtained by the instrument.

Robert Novak, MD
Department of Pathology
Children's Hospital
Medical Center of Akron
Akron, Ohio

Chair, CAP Hematology and Clinical Microscopy Resource Committee