Ronald D. Feld, Ph.D., Marian Schwabbauer, Ph.D.*, CLDir(NCA), John D. Olson, M.D., Ph.D.
Upon completion of this section, the reader will be able to:
1) Describe the rationale for the use of laboratory tests in the office laboratory.
2) Discuss pre-analytical factors that affect blood glucose testing.
3) List the criteria for the diagnosis of diabetes.
4) List the pre-analytical factors affecting electrolyte testing.
5) Identify patients who should be screened for hyperlipidemia.
6) Describe the contribution of the various lipid fractions to predicting heart disease.
7) Discuss the analytical factors affecting lipid measurements.
Clinical Chemistry
Introduction
The performance of any test, laboratory or otherwise, should be in response to one of three reasons. The first is screening where a test is performed on a totally asymptomatic or minimally symptomatic patient to detect the presence of occult disease. This is the most difficult application of laboratory testing since most diseases in this population have a low prevalence. If the test decision level is maximized for sensitivity, there will be an excess of false positive results. These will then have to be subjected to further testing to distinguish the true positives from the false positives. An example is the genetic screening that is performed on newborns to detect congenital hypothyroidism or PKU.
The next application of laboratory testing is to confirm or exclude a diagnosis. In this case, the patient is no longer asymptomatic but presents with one or more complaints. One possible diagnosis in a patient presenting with jaundice is viral hepatitis. The testing for the presence of antigen or antibody to hepatitis A, B, or C will help to confirm or exclude this cause of jaundice.
The third application of laboratory testing is for prognosis or to measure the effectiveness of therapy. In this case, a diagnosis has been confirmed and treatment instituted. The return to the euthyroid state in a patient with hypothyroidism by exogenous T4 can be confirmed by normalization of the T4 and TSH.
The ability to perform clinical chemistry procedures in the office laboratory has increased markedly since the introduction of new technology such as dry chemistries and disposable cartridges containing electrodes and pre-packaged reagents. Two parallel uses for this technology, the office laboratory and point-of-care or bedside testing have spurred the rapid introduction of new technology to the easy-to-use testing menu.
While a large menu of clinical chemistry testing is available, the needs of individual practices will differ and help determine which tests are cost effective. Bilirubin testing would have a much larger role in a pediatric practice as opposed to one devoted to cardiology.
The rationale for office testing rests on the need for rapid turnaround of results so therapeutic intervention may be made in real time. It is hoped that prompt intervention will prevent more serious complications that would necessitate treatment at a more complex site at a later date.
Another rationale for office testing is the convenience of the patient who does not have to travel to another site for testing. This must be balanced against the expense of increased regulation, reimbursement rates, and liability issues.
Glucose Testing
The most common chemical test performed in the office laboratory is serum or whole blood glucose. Most methods employ one of two enzymes for measurement, either glucose oxidase or hexokinase. Urine glucose testing, while non-invasive, is not considered helpful because the renal threshold at which blood glucose spills into urine varies considerably from person to person.
Serum and plasma glucose values are about 15% higher than whole blood glucose. Glucose measurements on capillary blood from a fingerstick are about 7% higher than on venous blood because capillary blood more closely resembles arterial whole blood. Plasma glucose can be estimated from whole blood capillary values by the formula:
For example, a capillary glucose of 100 mg/dL with a hematocrit of 42% would equal a plasma glucose of 111 mg/dL.
Pre-analytical variables that can affect glucose measurement are delay in separation of the serum or plasma from the cells, pregnancy, and certain drugs such as steroids. It is estimated that glucose levels are reduced by about 7% per hour when not separated from cells. Patients with very elevated white counts will often display low glucose values due to increased cellular metabolism in the sample tube. Pregnant females usually have lower blood glucose values than non-pregnant females. The criteria for an abnormal glucose tolerance test to diagnose gestational diabetes takes this into account.
The test strips for measuring glucose contain a measuring pad mounted on a solid support strip. The pad contains reagents for measuring glucose in solid form covered by a layer that separates plasma from cells. After a color reaction that is proportional to the glucose present develops, quantitation is accomplished by inserting the strip in a reflectance meter. Technique is very important in performing blood glucose by test strips. The finger must be dry before blood is taken and excessive squeezing should be avoided. The amount of time the blood is left on the pad is critical as is how the blood is washed from the pad.
While most glucose determinations will be performed to monitor known diabetic patients, they may also be used to screen for and diagnose the disease. New criteria for the diagnosis of diabetes were recently adopted. The fasting glucose necessary for the diagnosis was lowered from 140 mg/dL to 126 mg/dL. This was done in an attempt to catch the disease at an earlier stage and to modify the consequences of the disease with treatment. Emphasis is now being placed on the fasting glucose rather than the oral glucose tolerance test which is harder to standardize.
Physicians who treat a large number of diabetics should consider offering fructosamine or hemoglobin A1c determinations. Both monitor average blood glucose levels over a longer time period. Fructosamine measures mostly glycated albumin and reflects past glucose levels for about a one-month time period. Hemoglobin A1c measures a particular species of glycated hemoglobin; and since the life span of the red cell is longer than albumin, it reflects past glucose levels for a period of about three months.
Electrolyte Tests
Electrolytes, especially sodium and potassium, are frequently measured in the office laboratory. The predominant methodology in use for these analytes is that of potentiometry as employed by the ion selective electrode. The potentiometry may be either direct or indirect. In indirect methods, the sample is first mixed with a diluent before measurement. Direct methods usually involve whole blood; the sample is untreated before contacting the electrode.
Indirect methods show interference from factors which change the ratio of plasma water to non-water elements. Elevated lipids and proteins reduce the amount of plasma water in a sample and since the electrolytes are dissolved only in the water phase an effectively smaller sample is mixed without diluent prior to measurement. Normal plasma is about 93% water while samples with elevated lipid or protein may contain only 80-85% water. This can cause pseudohyponatremia since sodium is most notably affected in these samples minimizing pre-analytical variation is an important element in accurate electrolyte measurements. Fist pumping during phlebotomy can significantly increase plasma or serum potassium levels. Increased contact time with cellular elements in the sample tube will also lead to increased potassium levels due to the release of potassium from the intracellular space. This is especially true if the sample has been refrigerated since this reduces glucose metabolism which supplies the ATP needed to run the sodium-potassium pump.
Samples with elevated platelet and white counts can also give falsely elevated potassium levels since potassium is released from these elements during the clotting process. Normally serum and plasma potassium differ by only 0.1 or 0.2 meq/L but in a sample with a platelet count of a million or greater, the serum potassium may be 1-2 meq/L higher than plasma. If this is suspected, the test for potassium should be repeated on a heparinized plasma sample.
Lipid Testing
The National Cholesterol Education Program (NCEP) Expert Panels I and II have strongly advocated screening adults for hyperlipidemia. Children should not be screened unless there is a positive family history. Screening the elderly is controversial. While elderly have a high prevalence of elevated lipids, it is unclear whether treatment of this age group will have a significant effect on morbidity and mortality. Total cholesterol values > 200 mg/dL are considered to be indicative of increased risk for atherosclerotic heart disease and values greater than 240 mg/dL indicate very increased risk. Most cholesterol measurements performed in the office laboratory involve a color reaction utilizing cholesterol oxidase.
NCEP II stressed the necessity of also measuring high-density cholesterol to fully assess risk. Values < 35 are indicative of increased risk even in the face of a normal total cholesterol level. If a fasting sample is used, the low density lipoprotein cholesterol can be calculated using the formula:
The triglyceride must be less than 400 mg/dL for this formula to be valid.
There are many pre-analytical factors to be considered when interpreting lipid values. Position during blood drawing will affect tissue water to blood distribution. Most papers recommend the patient be sitting for 15 minutes before drawing. There is a significant biological variation in lipid values; it is recommended that if lipid values are close to the decision level that two or three measurements are obtained at weekly intervals and averaged.
LDL cholesterol is an acute phase reactant and will be lowered in stress. Lipid status should not be evaluated in subjects with acute illnesses or in pregnant women since pregnancy affects lipid values.
Good precision and accuracy of lipid values is important since lifelong treatment (pdf), momentous lifestyle changes, and effectiveness of treatment decisions will be based on them. The NCEP recommends at least 3% imprecision and no more than 3% deviation from a reference method value.