| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Proceedings of the 24th Arnold O. Beckman Conference |
1 Mayo Medical Laboratories, 200 First St., SW, Rochester, MN 55905.
There is a favorite aphorism attributed to Charles Mayo that states, "In the study of some apparently new problems we often make progress by reading the work of great people of the past ... " (1).
Similarly, insight for the future can be gained by reflecting on the example of Dr. Arnold O. Beckman, who celebrated his 101st birthday on April 10, 2001. From his invention of the "acidimeter" (patent no. 2,058,761) to his introduction in 1940 of the DU Spectrophotometer through Beckman Instruments, which redefined "accuracy" in clinical chemistry measurements, Dr. Beckman "walked his talk", having posited that "There is no satisfactory substitute for excellence" (2) and "When you’re faced with the necessity to do something, that’s a stimulus to invention. If (my classmate) hadn’t come in with his lemon juice problem, chances are I never in the world would have thought about making a pH meter"(2).
Preparation
If laboratory professionals are to heed Dr. Beckman’s call to excellence, we must first acknowledge, together with members of the diagnostics industry he helped create, a great obstacle to achieving a level of performance commensurate with patients’ needs. This challenge resides in continuing failure, individually and collectively, to define quality objectives in terms of laboratories’ clinical objectives. Although much needs to be said, has been said, and will be said relative to preanalytical and postanalytical phases of quality assurance, much too little has been said of analytical quality. Little effort has been made toward dispelling the body of collective fiction that suggests that "The quality’s the same everywhere". Dr. N.W. Tietz has expressed concern for the degree to which we have strayed from a laboratorian’s fundamental concern with the need that procedures be accurate, precise, specific, and comparable among laboratories (3).
Rudyard Kipling’s The Village That Voted the Earth Was Flat became, as the victim of a scurrilous hoax, the deserving recipient of barbs, laughs, and ridicule by a world to which such an assertion was ludicrous (4). The potential consequences to patient care of our own gullibility and the naivete of clinicians we serve with regard to inattention to accuracy in testing are grave.
Professional organizations have similarly ascribed to a body of collective fiction affirming that "the quality is the same". A review of recent College of American Pathologists’ Ligand Survey results serves sufficiently to highlight a problem regarding the assumption of quality. Fig. 1
represents the frequency of distributions for results obtained by laboratory participants analyzing prostate-specific antigen (PSA) Specimen K-19 of Survey 2000 (5) as members of peer groups performing the eight most commonly used methods. Action limits relative to PSA concentrations have been widely published and generally accepted, independent of any determination of the same within the community’s practice and relative to that community’s PSA method of choice, local population mix, or specific actions to be taken at the prescribed action limits. Urologists and generalists alike commonly consider a non-age-related PSA value 
4.0 µg/L sufficient to warrant concern and careful follow-up, and a value 10.0 µg/L virtually a mandate for biopsy, a procedure that approaches 20% in false negatives, frequently leads to repeats, and is associated with substantial morbidity and cost. A review of the results of this survey and the breadth of results found "acceptable" shows that it is clear that however popular a particular result is, it is imperative to consider on behalf of the patient, physician, and payer which PSA value is the correct one, ideally represented with appropriate confidence limits.
|
Fig. 2
illustrates similar results for Total Cholesterol Specimen LP-06 on CAP Survey 2000 (6). Decisions made on the basis of such values include insurance ratings and lifetime drug treatment, neither without cost. The specious argument that physicians do not use specific cutoffs for action is easily dismissed. In the absence of data that reliably separate the well from the afflicted, clinicians must resort to "judgment" (a euphemism, in the absence of a history of data-driven decisions, for intuition). Whatever values they might settle on in their practices are undermined in their day-to-day effectiveness by this and other forms of analytic bias. It is this need to depend on intuition that our specialty, more than any other, has hoped to erase. Our focus, however, has been on precision, particularly in the arena of immunoassays, to the exclusion of accuracy; and as these survey results suggest, accepted normative performance targets have become substitutes for traceable and dependable standards.
|
There is evidence of a new awakening to the importance of true standards to the role laboratories are asked to play. Clinical chemists and pathologists will be increasingly expected to participate, not only in diagnosing the sick, but in recommending the appropriate treatment for them. George Klee, long heralding the need for greater focus on analytic accuracy, has published a conceptual model for establishing tolerance for analytic bias and imprecision based on variations in population test distributions (7). With colleagues, he has described the effects analytic bias specifications can have on patient outcomes when coupled to medical guidelines (8). Frederick A. Smith and Steven H. Kroft have prescribed procedures for detecting analytic bias using patient samples (9). First attempts, albeit with some persisting opportunities for improvement, have been made to design reference reagents for PSA, pursuant to establishing an international standard for PSA and free PSA (10). The Centers for Disease Control have promulgated analytic objectives for total cholesterol and related moieties, although far too few laboratories subscribe to programs that maintain their cholesterol results within prescribed variance from traceable standards (11). Myers et al. (12) have described a program for implementing a reference method for cholesterol testing within a network of laboratories. Medical center laboratories around the world must awaken responsible management to the abiding economy of quality and find the resources to readdress the need for accuracy in clinical laboratory testing.
There is a second important opportunity for change in our clinical laboratory practice. Clinical laboratories have been too willing to add to service menus tests that have had little or no clinical value or tests that have such poor performance characteristics over time that they have served more to confuse and compound, rather than to clarify, the clinical evaluation of patients. Clinicians and laboratorians alike have their lists of most useless tests or most unreliable tests. Bolann et al. (13) outlined clinical circumstances in which carefully performed B12 assays can be relied on, as well as situations in which alternative means for identifying cobalamin deficiency should be considered. Magera et al. (14) published in the same issue of Clinical Chemistry an elegant, low-cost procedure for measuring methylmalonic acid in plasma and urine by tandem mass spectrometry. This method of evaluating the possibility of cobalamin deficiency relates values to a traceable standard using a physicochemical method at a very low unit cost. Such emerging technologies invite one to consider newer, clinically meaningful tools in preference to tools designed to meet a price point. An important move in the right direction has been made by the College of American Pathologists Survey Program in discontinuing surveys of anti-single-stranded DNA (15), one of several tests adding no value to patient care.
The issues of test accuracy, test selection, and test method selection have been of great but largely unacknowledged importance to patient care in the past, when laboratories were comfortable in being referred to as "diagnostic laboratories". They will be of trebled significance to a role as the "prescriptive laboratory" of the future.
The Prescriptive Laboratory
Although the majority of objective data in the patient’s medical record originate within the clinical laboratory, those data have been only occasionally pathognomonic or diagnostic of a specific disorder. Laboratory efforts have more typically been corroborative or exclusionary of clinically suspected diagnoses or have been an effective guide to further studies. Out of technical necessity, the tests offered often have been mere surrogates for analytes or systems the laboratory would prefer to measure. The discipline of clinical chemistry, well practiced, has served effectively to encourage among physicians the use of data and reason as preferred substitutes for intuition. The technologic limits in which clinical chemists have practiced have left this medical specialty termed "ancillary". Indeed, many of the challenges that the field has faced have been a direct result of that widely accepted view. The geographic and corporate separation of clinical laboratory practice from the clinical practice of medicine has exacerbated the mindset of the clinical laboratory being seen as an ancillary service and a commodity.
We are in the midst of, not merely on the brink of, a sea change in our role. This role requires, for those who accept it, a degree of accuracy, test selection, and test method selection in which the "diagnostics industry" is scarcely schooled. It mandates a higher degree of communication with clinicians. Effective communication, although facilitated through the use of the most modern tools possible, will rest largely on the clinical chemist’s ability to convey a great amount of knowledge concerning treatment decisions and the effect assays have on those decisions. Our tests have been surrogate in nature, and their role has been corroborative or exclusionary. Today laboratory specialists are asked increasingly to identify pathology at a molecular level with newly available nonsurrogate tests. A more palpable change in daily practice, however, consists of our being asked to decide the patient’s therapy, determine the treatment’s effectiveness, and recommend adjustments in a patient’s treatment. The changes invite a practice pattern far more familiar to the surgical pathologist than to the laboratorian:
Consider that cardiac markers until recently have been divided into indicators of long-term risk for occlusive coronary artery disease (CAD) and those used to evaluate the possibility of occlusive CAD in the acute care setting. The boundaries between these groups of tests are now being blurred, as new studies reveal short-term as well as long-term benefits to the patient receiving therapy directed at immediate lowering of LDL-cholesterol (16). Because hypercholesterolemia is now known to have an immediate effect on endothelial production of prostaglandins that inhibit platelet aggregation, reflections of hypercholesterolemia are increasingly being considered mandates for immediate aggressive therapy (17). Similarly, concern for the short-term effects of hypercholesterolemia, hypertension, and nicotine on endothelial production of prostaglandin I2 and nitric oxide, with resulting vasoconstrictive anomalies, is increasing the vigilance with which abnormal results are addressed pharmacologically (18)(19)(20)(21)(22). High-sensitivity C-reactive protein, widely accepted as a surrogate for inflammation and associated with a higher long-term risk factor for CAD, is increasingly used in the acute care setting and may lead to therapy with broad-spectrum antibiotics or affect the selection of cholesterol-lowering drugs. It is critical to align the accuracy and precision of laboratory tests, as well as the pattern of their delivery, with the practice patterns of clinicians; and we must review with our clinical colleagues how effectively these coordinated efforts are achieving targeted outcomes.
Logical laboratorians are inclined to presume that the treatment of hypertension consists of selecting drugs that lower blood pressure. A review of recommendations from the Sixth Report of the National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure (23) revealed that most therapeutic decisions made in the treatment of increased blood pressure rely on objective search by the clinical laboratory for evidence of comorbidity or target organ damage. Thus, historical or current laboratory evidence of hyperlipidemia, myocardial infarction, congestive heart failure, diabetes mellitus, or kidney malfunction have by convention a very great bearing on (a) whether a patient with high blood pressure will be treated pharmacologically for either the high blood pressure or the accompanying comorbidities and (b) the combination of therapeutic agents with which the patient will be treated. Whether the function of the renal glomerulus is evaluated by creatinine clearance, by glomerular filtration rate (24), or by more sensitive iothalamate clearance studies will have a remarkable effect on the degree to which renal impairment is considered in the selection of therapy. Whether we await gross evidence of congestive failure or use laboratory tools such as atrionaturetic factor, brain naturetic factor, or endothelin plus imaging studies to identify patients with occult or impending congestive heart failure once again will affect the therapeutic decision.
For years then, decisions on risk stratification, further work-up, and treatment have rested largely on laboratory values. Analyses performed within clinical laboratories will increasingly dictate therapy. Evaluation of cell surface markers and their respective associated transduction systems as a guide to therapy further exemplifies this prescriptive role. In an ongoing clinical trial, access by patients with well-differentiated lymphocytic lymphoma to treatment with anti-CD20 rests on the ability to demonstrate expression of CD20 by the neoplasm. This ability to detect such surface markers depends greatly on method and reagent selection as well as the quality of the assay. The medical oncologist assumes that "the laboratory does not miss ... ask them". Similarly, the use of Mylotarg (gemtuzamab ozogamicin for injection), a drug that elegantly ties an anti-CD33 antibody to a chemotherapeutic agent, is limited to use in patients whose neoplasm is demonstrated to express CD33. Clinical trials for the use of anti-CD52 in the treatment of chronic lymphocytic leukemia will be limited to patients with demonstrable CD52 expression on the leukemic cells. Signal transduction inhibitor 571 happens to inhibit phosphorylation of the protein product p210bcr-abl of the Philadelphia translocation t(9;22)(q34;q11), thereby blocking its tyrosine kinase activity (25). Thus, effective treatment of the usual Ph1 chronic myelogenous leukemia may be similarly useful in patients with rare Ph1 acute lymphocytic leukemia. Treatment of psoriasis may in the future depend on the use of biotherapeutics that block the interaction of B7 on antigen-presenting cells with CD28 on resting T cells or on the interaction of lymphocyte-function-associated antigen-3 (LFA-3) with CD2 on the surface of T cells (26). The clinical usefulness of demonstrating overexpression of her-2-neu on the cells of breast ductal carcinoma, and the applicability of anti-her-2-neu (Herceptin), is now the province of the lay press. Largely unaware of the variable sensitivities of methods for quantifying her-2-neu expression and perhaps the variable care with which such tests are performed, the oncologist and patient are for now potential unknowing victims.
Such cellular surface markers add hope for replacing surrogate measurements seemingly destined to last forever. One of the most promising examples is CD64, a surface marker found on polymorphonuclear leukocytes only when they are activated. As techniques for quantifying such markers, particularly measures to properly calibrate and standardize such procedures, are perfected, CD64 and other elements of interferon-
signaling may replace sedimentation rates and C-reactive protein as indicators of active inflammation and/or infection. Studies investigating the usefulness of neutrophil CD64 expression in disorders such as rheumatoid arthritis, coronary angina, occult infections in patients presenting to the emergency room, and the possibility or probability of septicemia in hospitalized patients at risk are currently in progress (27)(28)(29). CD64 is being evaluated as a measure of the effectiveness of a specific antibiotic regimen in the treatment of infection (30).
Similarly, measurement of interleukin-6 has been demonstrated as superior to the erythrocyte sedimentation rate in monitoring the treatment of giant cell arteritis (31).
Diagnosis already is not enough to meet clinical needs and expectations of clinical laboratories. Laboratories will be asked to answer the more tedious question of what shall be done for the patient. Increasingly, our efforts will be directed at prescribing the most promising therapy.
Genomics and Proteomics
The emerging request that the clinical laboratory assist with patient therapy is amplified by the explosion in the ability to evaluate pathology and potential treatments at the level of patient and pathogen genome. Many bioinformatics companies have been founded and funded on a widely accepted premise, that meaningful derivation of clinically useful information from an enormously complex database comprising pathogen genomes, the patient’s genome, and potential genomic targets for effective pharmaceuticals, as well as pathogen and patient genomic expression that can alter the effectiveness of pharmaceuticals, requires exceedingly complex computation support. The quality of such databases may suffer from one great threat, the temptation to oversimplify and underestimate interrelationships among genomic members and the extent to which the permutations of their interactions are dependent on the patient’s own biology, including pharmacologic milieu. Noteworthy attempts to reflect meaningfully the complexity of these permutations in the creation and study of such databases can be found in the work of Atul J. Butte and co-workers (32)(33) in their application of relevance networks to the challenge of grading interrelationships among genes and environmental factors.
Pharmacogenomics promises to revolutionize the sophistication with which therapeutic measures are developed, including temporarily infamous gene therapies. Bonnabry et al. (34) at the Laboratory of Computer-Assisted Therapeutics in Geneva, Switzerland have developed Q-DIPS, a computer-based system for predicting drug metabolism interactions in the context of a specific patient’s cytochrome P450 isoforms. Hersberger et al. (35) have described an effective method for routine genotyping of the majority of CYP2D6 poor metabolizers, detecting CYP2D6*3, CYP2D6*4, and CYP2D6*6 alleles by tetra-primer PCR and the CYP2D6*5 allele by multiplex long PCR. With such technology and computation, it will be possible to anticipate the incapacitating tardive dyskinesia that affects some patients receiving antipsychotics and avoid "idiosyncratic" side effects of narcotics and the resistance to opioid pain relief that afflicts 2–10% of patients homozygous for nonfunctional CYP2D6 alleles. It will also be possible to anticipate the extraordinary sensitivity to warfarin seen in patients with CYP2C9 variant alleles (36). These are but a few examples of opportunities for the clinical laboratory to prevent harm.
Cell surface markers were used above to exemplify the evolving demand for the laboratory’s assistance in designing therapy. The study of gene products in solid tumors and their roles as components of transduction systems has already led to new classes of oncotherapeutics. Ras, the protein product of the ras oncogene, leads to cell division by stimulating signal transduction pathways as well as transcription factors. To activate this cell proliferation pathway, Ras must undergo posttranslational modification mediated by farnesyl transferase (37). Farnesyl transferase inhibitors have therefore attracted much attention. The clinical laboratory is being asked to ascertain which tumors show overexpression of Ras and also evidence an intact transduction system to be blocked.
Rapid bacterial identification and antimicrobial susceptibility testing have already led to substantially fewer laboratory studies, imaging procedures, days of intubation, and days in an intensive or intermediate care area (38). Attempts are being made to augment molecular techniques with physicochemical methods (39), and similar physicochemical techniques are now being used in the study of carcinogenesis by the sequencing of oligonucleotides containing oxidative lesions (40). Such techniques will someday enhance our ability to identify the in vivo effects of putative carcinogens in an individual patient or enable us to measure the effectiveness of an oncotherapeutic drug during a patient’s therapy. Likewise, the ability to detect circulating cancer cells in the peripheral blood of cancer patients continues to evolve. The use of tumor-specific mRNA for this purpose has been reviewed by Pelky et al. (41). More recently, Sorensen et al. (42) quantified melanoma cell-specific mRNA by competitive reverse transcription-PCR and were thereby able to detect a single melanoma cell in 1 mL of blood.
Individualization of Therapy
Listed then below are principles and scientific advances that require the clinical laboratory’s focus if it is to have an optimum effect on the care of sick people:
As laboratory scientists involve themselves more competently in these areas, medical practice will be driven ineluctably toward individualization of patient therapy. The clinical laboratory will continue to evolve from its more contemplative role in patient evaluation to a more decisive role in designing therapy for a specific patient. The treatment of hypertension serves to describe how individualization of therapy might occur (Fig. 3
) (43):
|
Laboratory professionals have enjoyed a privileged familiarity with scientific advances that promise to change patients’ lives. At the same time, the science to which they have had such special access has served to alienate them from their patient. In general, labortorians’ families are scarcely able to comprehend what they do or what they might spend the day thinking. As we work toward the individualization of therapy, the science that has so alienated us from the patient will reunify us and return our focus on the patient as an individual, not merely a member of a population. We will be able to describe our role in terms not merely tangible, but palpable.
William J. Mayo cautioned his colleagues in 1923 that "... the highly scientific development of this mechanistic age had led perhaps to some loss in appreciation of the individuality of the patient and to trusting largely to the laboratories and outside agencies which tended to make the patient not the hub of the wheel, but a spoke" (44). Dr. William J. Mayo’s fears have been realized as laboratory medicine has been increasingly commoditized and extricated from the clinical practice. A simple truth will grow increasingly self-evident, however: The commitment of a clinical laboratory’s professionals to the care of the sick will continue to distinguish those medical teams consistently effective in providing care from those condemned merely to go through the motions.
References
signaling in unstable angina. Circulation 2001;103:1509-1514. receptors on neutrophils and monocytes may reflect ongoing bacterial infection. J Med Microbiol 1996;44:135-140.[Abstract]
therapy. Lab Hematol 1995;1:3-12.
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
