Absence of nocturnal elevation of plasma prolactin concentrations in Cushing's disease.

The mean prolactin concentration during sleep and the per cent change (increase) and increment in the nocturnal elevation of plasma prolactin concentrations were significantly reduced (P less than .01, less than .01, and less than .001, respectively) in 6 untreated patients with clinically active Cushing's disease, compared with concentrations seen in normal subjects. There was no significant difference between mean-awake and mean-sleep plasma prolactin concentration in these patients, in contrast to that seen in normal subjects. In contrast, patients with hypercorticism secondary to adrenal adenoma (n = 2) or receiving long-term, high-dose corticosteroid therapy (n = 6) did not differ significantly from normal subjects in the percentage change seen in the nocturnal elevation of their plasma prolactin concentrations. Although mean prolactin concentration during sleep was reduced in these two categories of hypercorticism, a significant difference was still present between mean-awake and mean-sleep plasma prolactin concentration. (The nocturnal increment in plasma prolactin concentrations in these patients with hypercorticism compared with that of normal subjects was reduced, P less than .02). The difference between the percentage change and nocturnal increment observed in patients with Cushing's disease versus that seen in these two categories of hypercorticism was significant at P less than .02 and P less than .01, respectively. Four untreated patients with localized hypothalamic tumors also showed a significant reduction in the percentage change and increment in the nocturnal elevation of plasma prolactin concentrations (P less than .05 and P less than .02). In contrast to these findings, the prolactin responses to thyrotropin-releasing hormone (indicative of pituitary responsiveness) in patients with Cushing's disease, patients with adrenal adenoma, or those receiving chronic corticosteroid therapy, were all within the normal range seen in sex-matched control subjects. These findings provide additional evidence for the suggestion that there is altered hypothalamic function in patients with Cushing's disease.

Influence of specimen treatment on nonreactive HTLV-III sera.

We investigated effects of specimen treatment on HTLV-III antibody assay results. Freezing and thawing specimens 10 times did not increase specimen mean absorbance (A) values measured by the Abbott immunoenzymometric assay (IEMA). Heating blood donor serum specimens at 56 degrees C for at least 10 min inactivates virus, but converted 7.1% of nonreactive specimens to positive when measured by this IEMA. Also heating at 56 degrees C for 30 min increased A values of most specimens; however, the conversion to positive within assay ranged from 0 to 75%, with a mean of 17%. When 56 nonreactive patient specimens were heated, 55% became Abbott IEMA reactive, although ENI IEMA values did not increase and Western blot results remained nonreactive. Binding of immunoglobulins G and A to Abbott IEMA beads did not cause this conversion after heat treatment. We postulate that heating specimens causes alteration of some serum matrices and introduces nonspecific binding to assay bead antigen or antibody reagents. We recommend that for the Abbott IEMA, specimens may be stored frozen, and must not be heated to avoid altered reactivity.

Laboratory results. Timeliness as a quality attribute and strategy.

Although timeliness of results reporting has not been a major focus in clinical laboratories, there is increasing pressure from clinicians to report results rapidly. Even though there are only sparse data, timeliness in reporting of laboratory results undoubtedly affects clinician and patient satisfaction as well as length of hospital stay. Improving turnaround time (TAT) is a complex task involving education, equipment acquisition, and planning. All the steps from test ordering to results reporting should be monitored and steps taken to improve the processes. Various strategies to improve TAT at each step in the testing process are discussed.

Quality improvement practices in clinical and anatomic pathology services. A College of American Pathologists Q-probes study of the program characteristics and performance in 580 institutions.

Participants of the College of American Pathologists Q-Probes program described their quality improvement practices for clinical and anatomic pathology. In 580 institutions, the median time required for a median of 12 indicators of quality was 40 hours/month, with the number of indicators and the time spent directly dependent on bed size (P = .0001). The overwhelming majority of participants reported benefit from their quality improvement programs in terms of patient outcomes, as a management tool, and for risk management. Six indicators in clinical pathology and four indicators in anatomic pathology were used in more than 75% of laboratories, whereas an additional seven indicators in clinical pathology and five in anatomic pathology were used in more than 50% of laboratories. The authors conclude that quality improvement practices are similar among laboratories, and irrespective of increasing regulatory requirements, pathologists and senior laboratory personnel spend large amounts of time for activities that they believe improve the quality of services rendered.

Protein method influences on calculation of tissue receptor concentration.

Many protein methods are used for estimation of tissue receptor concentration. The authors compared performance, analytic variability, and accuracy of six protein methods used in these calculations. They found the Lowry protein procedure standardized with bovine serum albumin (BSA), usually considered the reference method, to be the most imprecise and most time consuming method. When the BSA standards from the Lowry procedure were assayed with the other methods, results ranged from 74 to 141% of expected. For three other protein standards, reactivity among the six methods varied almost twofold. Comparison of Lowry protein concentrations in cytosols from 46 tumors biopsies with other methods indicated best agreement was with an automated turbidometric (TCA) or a Coomassie dyebinding procedure. Use of protein standardization for the two direct spectrophotometric procedures decreased overestimation of receptor protein concentrations. Because receptor concentration is the quotient of receptor quantity and protein concentrations, tissue receptor results are dependent in part on standardization and choice of protein method.

Prevalence of non-A,non-B hepatitis/hepatitis C virus antibody in human immunoglobulins.

Human intravenous immunoglobulins prepared by the cold ethanol fractionation technique of Cohn are considered safe with respect to infectivity. However, there have been several instances of transmission of both hepatitis B and non-A,non-B hepatitis viruses after administration of intravenous immunoglobulins. To determine the prevalence of hepatitis C virus antibody in intravenous immunoglobulins and protein preparations, 30 commercially available products were tested. Using the Abbott enzyme immunoassay for hepatitis C virus antibody, 27 of 30 (90%) immunoglobulins tested positive. The Ortho immunoassay showed that 28 of 30 (93%) were positive, with one discordant result between the Ortho and Abbott assays. An antigen-blocking or neutralization test (Abbott) confirmed the results of the Ortho assay. Bovine, sheep, goat, and horse sera also were tested before and after isolation of animal immunoglobulins. All results on the animal sera were negative, indicating that the fractionation process did not produce false-positive results. The high prevalence rate of hepatitis C virus antibody in intravenous immunoglobulins has important implications for follow-up of recipients, selection of serum donors, and implementation of anti-hepatitis C virus testing.

Analytical biases with liquid quality control material.

Biases were compared for eight analytes determined using common chemistry instruments. Data for three levels of liquid control sera were obtained from over 350 laboratories participating in a quality control program. These data were contrasted with comparable results for lyophilized material and reported in a CAP proficiency survey. In general, biases appeared to be proportional to analyte level. For some analytes, biases with liquid material were less than with lyophilized material, but for most analytes, biases were greater with liquid than lyophilized material. However, when liquid material is used for monitoring interlaboratory precision, the consequences of analytic biases are minimized because of large pool sizes.

Presence of HTLV-III antibodies in immune serum globulin preparations.

The authors tested 15 immune serum globulin pharmaceutical preparations for antibody reactivity to human T cell lymphotrophic virus type III (HTLV-III) by the Abbott immunoenzymometric assay (IEMA). Although no evidence of HTLV-III infectivity has appeared after injection of similar preparations into humans, the authors found all samples IEMA reactive. Upon dilution, the authors demonstrated parallel decreases of antibody reactivity among two samples of gamma globulin, the Abbott-positive control, and a markedly reactive patient specimen. Gamma globulin isolated from sera of six animal species was nonreactive in the Abbott assay. All samples were nonreactive with the H-9 cell line antigen. Antibody reactivity to HTLV-III was confirmed in 13 of 15 gamma globulin samples when tested by the Electro-Nucleonics IEMA, and 14 samples contained at least the p24 band on Western blot analysis. Although false positivity occurs in IEMA assays possibly because of elevated protein concentrations and nonspecific binding, the authors, results show that in most circumstances immune serum globulin preparations tested do contain true reactivity to HTLV-III.

Bacterial interference with urine osmolality measurements.

We report bacterial interference with urine osmolality measurements using an instrument based on the principle of freezing point depression. Although the exact nature of the interfering activity has not been defined, the phenomenon is associated with a bacterium, identified as Pseudomonas putida, and is removed from the specimens by filtration at 0.45 micron. The bacteria led to osmometer dysfunction presumably by acting as a nidus for crystallization and preventing proper supercooling of specimens.

Circulating human antibodies as a cause of falsely depressed ferritin values.

We have previously reported spuriously elevated values for serum thyrotropin (TSH) measured by immunoradiometric assay (IRMA). We tested those sera showing interference with the TSH assay in a ferritin IRMA which used 125I-rabbit antiferritin antibody and solid phase goat antiferritin antibody as reagents and spleen ferritin as standard. In this assay, we obtained falsely low ferritin values that were corrected by addition of 0.2% non-immune rabbit serum to the labelled antibody. Two other radioimmunoassays gave results with these sera that were not falsely lowered. The interference was shown to be due to human IgG reactive with rabbit serum, the specificity of the interfering antibody being similar in all affected sera. Human antibodies directed against immunoassay reagents may lead to spuriously increased or decreased immunoassay results depending on the specific reagents involved.

Timeliness of automated routine laboratory tests: a College of American Pathologists Q-Probes study of 653 institutions.

Benchmarks for timeliness of early morning routine clinical laboratory tests were developed from over 17,000 urea nitrogen and 16,000 white blood cell count measurements made for inpatients in 653 institutions participating in the College of American Pathologists Q-Probes program. Urea nitrogen and white blood cell counts were considered surrogates for routine chemistry and hematology tests. Laboratories at the 50th percentile reported median urea nitrogen and white blood cell counts by 09.04 and 08.51 h, respectively, whereas those at the 10th percentile reported these median measurements by 11.30 and 11.18 h, respectively. Results were available sooner in non-teaching than teaching institutions, and in smaller rather than larger institutions, with the degree of computerization affecting test availability. Timeliness also was affected by instrument type and mode of operation, but was unaffected by the percentage of stat testing. Based on modeling by regression analysis, there was little evidence that longer routine test turnaround times affect patient length of stay.

Clinical laboratory quality control: a costly process now out of control.

We studied laboratory internal quality control (QC) processes using the College of American Pathologists Q-Probes program. Over 500 institutions participated, providing practices based on approximately 710,000 cholesterol, 880,000 calcium, 400,000 digoxin, and 1,180,000 hemoglobin QC results. The costs of QC included participant median control samples rates comprising 9.1, 9.4, 37.0, and 6.8% for the four analytes respectively, repeat patient test rates of 0.36% for hemoglobin to 0.65% for digoxin, and median delays in reporting results when QC exceptions occurred of 15.8 min for calcium to 24.7 min for hemoglobin. Quality control practices were complex and highly variable among participants and frequently differed from internal laboratory protocols and from long-established quality guidelines. We conclude that QC is costly, and laboratorians frequently do not follow established QC practices, in part because they are complex. To improve compliance, we believe QC practices must be simplified.

Quality control for the clinical laboratory.

Quality control is a management discipline by which laboratories assure the validity of their results for bedside management of patients. The thrust of quality control has been the laboratory's efforts at improving analytic variability associated with the measurement process. Through increasing emphasis on automation and participation in laboratory improvement programs, laboratory imprecision and inaccuracy have decreased over the past 15 years. Recent advances in the practice of quality control include use of new types of quality control materials, changing concepts of what is out of control, and the introduction of microcomputer applications.

Pituitary tumors.

The most frequent pituitary neoplasm is the adenoma, which can produce signs and symptoms owing to mass effects or secretion of pituitary hormones. The authors discuss these mass effects of pituitary tumors on the endocrine system and excess hormone secretion by pituitary adenomas.

Hypercortisolism.

The diagnostic evaluation of a patient with hypercortisolism should progress through screening and then confirmatory hormonal measurements. However, tests of the hypothalamic-pituitary-adrenal axis to predict the cause of hypercortisolism always should be viewed as providing supportive data, as exceptions that occur limit the usefulness of these tests. Familiarity with the diagnostic errors of the procedures is essential.

Controversies in thyroid function testing.

Thyroid function tests can be abnormal in certain patients in the absence of thyroid dysfunction. All thyroid function tests can be affected in these patients, including the free thyroxine assays and free thyroxine index. Familiarity with the conditions affecting thyroid function tests is necessary to avoid incorrect diagnosis and unnecessary testing.

Quality assurance measurements in departments of pathology and laboratory medicine.

Previously, characteristics of the quality assurance program in each department of pathology and laboratory medicine were the uniqueness of the monitors selected, the thresholds set for activity, and definitions of good performance. In 1989, members of the College of American Pathologists developed and implemented the first nationwide quality assurance program for pathology and laboratory medicine. This program, Q-Probes, has been expanded for 1990; after completion of planning under way for 1991, indicators will be in place for all steps in the total testing process. More than 1100 institutions had subscribed by the first quarter of 1990. Input forms for measurements of clinical indicators of quality are distributed to participants and, after completion, submitted to a central computer facility where keypunching and data manipulations occur. The interpretation of results is enhanced by peer comparisons of the participant data collected. In many circumstances studied, a large proportion of the data are aggregated at extreme ends of the distribution. Because data distributions are nongaussian for the indicators studied, it is recommended that peer comparisons use the 10% to 90% range and that the central tendency be described as the median. Despite large databases for some indicators that exceed by more than 10- to 100-fold those previously reported, extensive categorization may preclude percentile ranking of participants. Critiques are distributed describing practices and suggestions for improvement. Steady improvement of each participant and of nationwide practice patterns are documented by remeasurements.

Use of proficiency test performance to determine clinical laboratory director qualifications.

Many activities and policies influence laboratory test quality. Proficiency test results are one measure of laboratory quality, and during the past 25 years, five studies have examined the relationship of laboratory director educational requirements to proficiency test results. Data from three studies support the association between director qualifications and quality as measured by proficiency test performance, whereas no relationship was found in the other two studies. Possible reasons for conflicting results include differences in database size and demographics; in addition, proficiency test results may be inappropriate, although widely used, as the sole measure of laboratory director performance.

Immunoassay. Development and directions in antibody technology.

Improvements and new applications for the technique of immunoassay have been extensive, owing in part to improvements in specific binding reagents used. Early reports advocated application of naturally occurring antibodies, specific serum-binding proteins, and receptors as binding reagents in immunoassays. Each is still used today, although applications are limited because affinities are generally low. Polyclonal antibodies produced in laboratory animals in response to injection of specific antigens have been the cornerstone of immunoassay because of their high affinities and wide range of specificities. Recently, monoclonal antibodies have begun to replace polyclonal antibodies because of improved specificity, ease of use, production of an unlimited supply, and advantages of worldwide assay standardization. Although only 50 analytes are currently measured by Food and Drug Administration--approved reagents, widespread use in the future is predicted.

College of American Pathologists Conference XXVIII on alternate site testing. What must we now do?

Implementation of the conference consensus that alternative site testing should be under the direction of the laboratory director supported by an institutional structure of users, laboratorians, and possibly administrators will optimize success. Information is needed if, when, and where alternative site testing improves the care of patients, and if faster results are associated with better and less expensive patient care. Costs and the timeliness of both central clinical laboratory and alternative site testing must be estimated in an identical fashion, then both must be compared with the costs and timeliness attributed to all steps of the total testing process included. As has been done with glucose, accuracy and precision goals must be developed for analytes measured at alternative test sites, then agreement reached on what these goals are and whether speed of measurement can influence these goals. Consideration of successful models is essential by those implementing or reevaluating alternative site testing programs. Required scientific information cannot be obtained in a timely manner unless improved funding occurs.