Quality control in the new environment: QC materials; an ideal control closely simulates patient specimens, is stable, and comes in large homogeneous lots.
Process control is the cornerstone of a total quality assurance program. Intenral process control consists of activities undertaken during the analytic process and is directed principally toward assessing the acceptability of patient data. External process control, through surveys arising from outside the laboratory, tests a lab's ability to obtain the correct result on an unknown specimen.
In internal process control, specimens that have known concentrations are inserted into the analytic run. These control specimens serve as tools for assessing the degree to which a given analytic run represents previous runs in control. The database obtained from internal quality control testing is used to evaluate patient data during analytic runs in progress; to quantitate analytic variation at clinically useful concentration levels; and to facilitate interlaboratory comparison of performance for accuracy and precision, through a regional database.
The proficiency surveys in external process control are typically sponsored by professional organizations or governmental agencies. They are used both for quality control and as an evaluation mechanism for accreditation and/or licensure.
Extended internal (or regional) quality control programs also draw widespread participation in the United States. Large numbers of labs work on common pools of control materials. Comparison of results gives participants information on their accuracy, as measured by distance from target means, and relative precision. These measurements aid in lab self-evaluation and peer review.
Successful programs for internal and external process control rest on such components as analytic processes, control materials, and data handling resources. The central role played by control materials is critical to the success of the total quality assurance effort. Although internal quality control may be practiced without control materials--through patient-derived statistics, for example--by far the most widely used and effective methods are based on the control sample technique.
External, internal, and regional quality control programs place important demands on quality control materials. These materials should mimic the unknown specimens being analyzed as much as possible. Ideally, they will be homogeneous and stable, and contain matrix and analytes in a concentration similar to that of patient specimens. They should also be available in large lots with minimal vial-to-vial variability.
An ideal material for chemistry would be fresh stable liquid human serum with no additives, in unlimited quantities. Since this material cannot exist, it is evident that we must make certain compromises in selecting materials for quality control.
Procedures such as lyophilization, freezing, and osmotic manipulation are necessary to render control materials stable. A high degree of homogeneity between vials is critical for validating quantitation of statistical quality control parameters, both within and between laboratories. Excessive inhomogeneity increases the apparent imprecision of an individual laboratory's data as well as grouped interlaboratory performance data in surveys and regional programs.
The concentration of analytes in the control material should be at or near the concentration of the unknown specimens. Since this is not practical for all specimens, controls are generally employed at or near clinically relevant decision levels.
In chemistry, two levels of controls, containing a wide variety of analytes, are most commonly used. Typical levels, for example, might be 8.5 and 10.5 mg/dl for calcium and 130 and 150 mEq/L for sodium. With enzymes, such as CK, LD, and transaminases, values at the upper limit of normal and at two to three times the upper limit of normal are frequently employed.
Control materials today are almost all prepared commercially, most commonly using plasma obtained by plasmapheresis. When these materials are available in large lot sizes, such as chemistry pools of 1,000 to 2,000 liters, there is enough for many laboratories to share common pools and an interlaboratory database for expanded internal (regional) quality control.
Two to four hundred laboratories may share regional pools for a year or more, with the average laboratory using approximately 4 to 8 liters per level annually. In external quality control programs, controls are usually circulated by mail and analyzed over a short period of time--for example, one or two days. Thousands of labs may share these pools, with each requiring a maximum volume of only a few ml per challenge.
Homogeneity in control materials is necessary to attribute variation in analytic data correctly to the analytic system. A lack of homogeneity may result from incomplete mixing of pools or from differential stability of unstable analytes within portions of a pool. Inhomogeneity is not considered to be a significant problem with pools produced within the last decade.
Variability in lyophilized controls can also be traced to lyophilization and reconstitution procedures. This variability may derive from differences in vial fill volumes, in the amounts of residual moisture in lyophilized dplugs, and in the quantities of reconstitution fluid employed at the bench. If the manufacturer and the analyst are careful, the contribution of these factors to total analytic variation will be negligible.
In addition, if variation in analyte concentration, such as drift, is to be attributed corectly to the analytic process, the constituents in control materials must remain stable for as long as the pools are in use. Considerable work has been performed to characterize the stability of analytes in quality control materials. Most analytes are stable in the common matrices employed, but in various lots, instability of several analytes has been documented.
For practical purposes, it is wise to recalculate quality control target values and limits during the life of the quality control pools, rather than rely on fixed limits determined at an early stage of use. This minimizes the number of false error signals wrongly attributed to the analytic process instead of stability-related changes in analyte concentration.
An often overlooked feature of quality control materials is their commutability, or interchangeability with patient specimens--specifically, the applicability of data derived from controls to patient specimens. This is affected to a great degree by the composition of the background matrix in which analytes are dissolved and suspended. Since all quality control materials are subjected to stabilization procedures, as well as various other manufacturing processes, they are rendered dissimilar to fresh human specimens in varying degrees.
The applicability of data from quality control procedures to clinically relevant measurements varies by analyte, method, and matrix. There are documented differences in the behavior of nonhuman and liquid controls relative to lyophilized human serum. Analysts must understand the extent to which controls fail to detect analytically and/or clinically significant errors.
Since the mid-1950s, clinical labs have benefited from a major technologic improvement: Manufacturers began producing large lots of homogeneous control materials in lyophilized or freeze-dried form. This freed laboratories from the need to prepare their own control materials from banked blood or pooled leftover serum. These procedures had yielded controls with significant instability, frequent hepatitis B contamination, and variable target values.
Although early pools of lyophilized quality control materials reflected advanced technology, instability was found among certain analytes. Significant deterioration in glucose concentration and decreased activity of various enzymes, particularly CK, were noticeable problems. Major improvements in stabilization, mass freeze drying, and final filtration have essentially resolved these problems.
The control sample method is widely applied to internal quality control in clinical chemistry, cellular hematology, and coagulation. In chemistry, quality control materials are available for routine analytes, enzymes, blood gas studies, ligand assay, and therapeutic drugs. Commonly used chemistry QC products are derived from human or animal (primarily bovine) serum. These materials are stabilized either by freeze-dry or powder-dry lyophilization or by osmotic manipulation using ethylene glycol additive.
Cellular hematology products largely use a human red cell base, supplemented by a variety of particulate additives--for example, avian cells or synthetics simulating leukocytes or platelets. Coagulation controls are produced from donor plasma manipulated to adjust the concentration of coagulation factors.
Lyophilization accomplishes stabilization by reducing the residual aqueous medium to less than 1 per cent and by allowing low temperature storage with a neglible effect on analytes. In traditional freeze-dry lyophilization, vials containing aliquotted serum are frozen and subsequently subjected to vacuum dehydration. Gradual heating removes the rest of the moisture.
A relatively recent process utilizes powder drying. Control materials are flash frozen by spraying fine particles into cold gas at subzero temperatures. Vacuum drying follows, and stirring keeps the mixture of fine dry particles or "beads" homogeneous. The resulting product has higher optical clarity than traditionally lyophilized materials because there is less denaturation of lipoprotein. The College of American Pathologists' survey program for clinical chemistry uses these products.
A variety of agents have been employed to reconstitute lyophilized materials. These include distilled water and buffers with ammonium or TRIS cation and carbonate or bicarbonate anion. Sodium bicarbonate, a common reconstituting agent, offers the advantages of containing a noninterfering physiologic cation and being adjustable to provide two clinically useful bicarbonate levels.
Osmotic manipulation can also achieve stabilization. When freezing points are lowered, controls can be stored as a liquid at freezer temperatures. One widely used control contains approximately 30 per cent ethylene glycol as an osmotic agent.
To obtain clinically suitable analyte concentrations, manufacturers of control material must dilute base materials and/or add various constituents during preparation. With inorganic analytes and most organic analytes, pure materials are easily obtained for addition. With protein analytes, peptides, and enzymes, however, the process becomes more complex because additive materials of human origin are not always readily available.
Since controls should simulate patient specimens as much as possible, it is desirable that animal additives also react as closely as possible to their human counterparts. Particular problems are posed when antigenic, immunochemical, or binding properties vary between human and nonhuman protein or peptide additives.
Moreover, differences in enzyme activity related to isoenzyme and species differences may be significant. Enzyme additives are generally most satisfactory when they are of human origin and derived from the isoenzymes predominant in human serum. Practical considerations preclude obtaining human additives for many enzymes, but such additives are available and widely used for aspartate aminotransferase, acid phosphatase, and amylase.
Several quality control products have been employed for blood gas analysis. They are commercially prepared by adjusting acidity to simulate acidotic, alkalotic, and normal specimens. The pO.sub.2 and pCO.sub.2 are adjusted as well to include clinically representative values. The matrices used include aqueous buffer, hemolyzed and whole blood, and fluorocarbon.
Significant vial-to-vial variation in these measurements occurs when the temperature is not controlled carefully, or when opened vials are not assayed rapidly, allowing equilibration of specimens with room air. Unlike blood- or protein-based products, aqueousbased blood gas controls lack the sensitivity to detect membrane coating by protein. In addition, protein-based controls are sensitive to temperature changes that do not noticeably affect aqueous materials.
A variety of assays have been shown to be superior when human control material is used. These include procedures to measure protein and peptide analytes when protein binding is an important factor. Human material is demonstrably better than nonhuman for quality control in numerous methods of measuring total iron binding capacity, bilirubin, thyroxine, and albumin.
Among the cited disadvantages of human control material are its cost (generally, it is slightly more expensive than nonhuman products), its limited availability (one must rely on donor centers rather than abattoirs), and the potential transmission of such infectious agents as hepatitis. Most products are screened for hepatitis B reactivity and HIV and appropriately labeled following manufacture.
The advantages frequently cited for controls of animal source are increased clarity over human materials when lyophilized, wider availability, and lower price.
With respect to liquid verus lyophilized matrix, liquid materials are relatively simple to use, possess good stability, and produce less waste--open vials can be recooled for use on subsequent days. Major disadvantages of ethylene-glycol-based material are its significant lack of commutability for various procedures and limitations on its use with certain analyzers.
For instance, the material is not suitable for quality control of osmolality because of the high ethylene glycol content. In addition, the relatively high viscosity causes liquid controls to behave differently from fresh or lyophilized materials in analyzers where fine tubing conducts the sample. Futhermore, these controls have limited use with film-based reagent systems because of the interaction of ethylene glycol with support material.
Lyophilized control materials vary in vial-to-vial filling volume by less than 1 per cent. Products are membrane-filtered, resulting in negligible bacterial counts. They maintain a moisture content of less than 1 per cent by weight. Disadvantages are the necessary reconstitution and the alteration of analyte concentrations during lyophilization. Alterations include loss of CO.sub.2; an increase in pH; denaturation of very low-density lipoprotein (VLDL), which causes increased turbidity; and slight changes in viscosity and osmolatiy.
Turbidity in lyophilized materials is attributable to denaturated lipoprotein. Lyophilization by the quick freeze-powder fill techniques, as an alternative to routine lyophilization, results in less denaturation of VLDL and a product that has greater optical clarity. With this material, however, electrostatic charges may cause some particles or "beads" to adhere to the vial neck or cap at the time of reconstitution. The loss of particles may lead to an increase in measured inaccuracy and imprecision.
There is widespread use of both assayed and unassayed controls in clinical laboratories. The assayed approach is valuable in the case of labile control materials with a relatively short shelf life, where it is impractical to obtain serial data for determining a laboratory's own assay values. For example, control products in cellular hematology are often used for as little as one or two months. Here, the assay data provide provisional target values.
Assayed controls are also helpful for infrequently measured or esoteric analytes. In such cases, there is a low likelihood of obtaining a usable regional database for interlaboratory comparison, similar to that employed with more commonly measured analytes.
Unassayed material is preferred when it possesses good long-term stability--one year or more. Expected values for various methods can be obtained from the interlaboratory data summary provided by a regional quality control program. The clinical laboratory should derive target values and decision limits for internal quality control from its own measurements, rather than use the target values and ranges supplied on commerical assay sheets. A limited number of observations at the commencement of the pool--approximately 20--is enough to furnish provisional target values and limits. Lot-to-date cumulative values are most commonly used for this pupose.
When unassayed materials are employed in conjunction with a regional quality control program, the laboratory has access to assay values of other labs analyzing the same material by the same or similar methods. the summarized group data, tabulated by analyte and method, form a document that is superior to a commercial assay sheet. Unassayed controls are priced substantially below assayed controls, most typically about 1/2 to 1/3 the cost. Laboratories can thus save hundreds or thousands of dollars each year, depending on usage rate.
Chemistry control are increasingly available in complex formulations so that a great variety of tests can be performed on the same product. Widely used materials now contain routine analytes in addition to multiple enzymes, ligand constituents, and therapeutic drugs. Since it is possible to perform a broad spectrum of testing on material from a single vial, laboratories are less dependent on relatively expensive specialty controls for lipid, enzyme, ligand, and drug analyses. this translates into significant cost savings for large labs. However, smaller laboratories with limited testing programs may not achieve comparable savings.
Table I indicates the analytes available in lots of lyophilized human controls that are currently being used in the Great Lakes, Southeast, and New England regional Quality control Programs.
To sum up, quality control in the clinical laboratory includes control of the analytic process. Internal process control usually involves analysis of control samples along with patient specimens. Ideally, these samples consist of materials that closely simulate patient specimens, are stable, come in large homogeneous lots, and are characterized by a high degree of commutability.
Departures from the ideal--in the form of nonhuman matrix and additives, the presence of osmotically active stabilizers, and/or lyophilization--may affect the degree to which control data reflect changes occurring in the analysis of patient specimens. These effects should be thoroughly understood before introduction of any material as a control in an assay system.
Commercial controls of high quality are widely available for chemistry, including enzyme analysis, ligand assay, and therapeutic drug measurements; blood gas procedures; cellular hematology; and coagulation. Unassayed controls used with regional quality control programs' databases offer substantial potential for cost reduction.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Part 5; internal and external process control|
|Author:||Lawson, Noel S.|
|Publication:||Medical Laboratory Observer|
|Date:||Jan 1, 1987|
|Previous Article:||How rotation can help boost productivity; developing well-rounded laboratory generalists through a hands-on rotation system is one solution to less...|
|Next Article:||Can we keep our blood supply safe from AIDS?|