The 2002 ASA Preoperative Testing Advisory

In 2002, the ASA published a guide for preoperative preparation of the patient prior to operative and nonoperative procedures.^[78] This publication followed an Internal Health Organization review of preoperative testing and a Medicare statement on preoperative testing. The fact that it was an "advisory" rather than a guideline highlights the point that the data are not robust enough to state recommendations with absolute certainty. Five findings from the ASA Advisory that are echoed in the other reviews bear emphasis:

1. No test, but rather a physician review stating that "this patient is in optimal shape for daily living" is usual for patients undergoing minimally invasive surgery.
2. Evaluation should occur prior to the day of surgery if the patient is not absolutely healthy or the procedure is other than "minimally invasive" (that is, testing was at least partially to be selected based on the invasiveness of the procedure).
3. Certain tests were found to be usual for the most invasive type of surgery.
4. About 17% of anesthesiologists obtained pregnancy testing routinely for all women in the potentially pregnant calendar ages of 11 to 55 years, irrespective of the patient's statement that "there is no way I could potentially be pregnant."

The ASA Advisory's statement that it is appropriate that no test but a physician review saying "this patient is in optimal shape for daily living" for patients undergoing minimally invasive surgery is based on the data presented above buttressed by two randomized controlled studies of minimally invasive procedures.

In the first and largest study, Schein and colleagues^[132] randomly allocated over 19,000 patients at nine centers scheduled to undergo cataract extraction with or without lens implantation to be tested routinely or tested only when indicated. No difference in outcome, hospitalization rate, or any measure of morbidity, mortality, cost, or satisfaction occurred between the "test routinely" and "test only when indicated" groups. The accompanying editorial reiterated the point (I am biased, because I wrote it) that each patient was seen by and under the care of a primary care physician who had already done sufficient testing and management to be able to state that each patient was in optimal shape for daily living.^[133]

In a similar but smaller subsequent study of slightly over 1,000 patients, Lira and colleagues^[134] found the same result: patients who are in ideal shape for daily living as judged by a physician need no additional testing prior to a minimally invasive surgical procedure. Those findings and the results of earlier studies indicate that it is appropriate to segregate patients by invasiveness of procedure to help select tests. A regimen for such segregation by invasiveness of procedure is suggested in Table 25-6 . It is significant that the data and schemata listed in this table require that the patient's condition be judged optimal for daily living. Does this mean that patients should be routinely

941

942

The segregation of patients based on the degree of invasiveness of the planned surgery as well as on the patient's condition brings up certain issues: when to perform a preoperative evaluation in order to determine the condition of the patient, and what to do if you disagree with the primary care physician about the patient's status as optimal for daily living. The judgment about invasiveness depends not just on the CPT (Current Procedural Terminology) code of the American Medical Association (AMA) but also on the surgeon—a laparoscopic cholecystectomy can be minimally invasive in a gifted surgeon's hands, is moderately invasive in most, and is highly invasive in some. The classification of surgical procedures outlined in Table 25-6 is based on the anesthesiologist's evaluation or judgment of usual invasiveness. Such preoperative evaluation almost certainly must of necessity occur prior to the day of surgery if the patient is not absolutely healthy or the procedure is other than "minimally invasive" (that is, testing is at least partially to be selected based on the invasiveness of the procedure).

As to disagreement with the primary care physician's evaluation, or if there is no primary care evaluation, the dialogue of the Advisory committee was and is clear: the final arbitrator of perioperative evaluation is the anesthesiologist providing anesthesia care for the patient. We will review the other two unexpected findings from the Advisory—about tests for highly invasive procedures and pregnancy tests later in this chapter. Let us now review why laboratory tests do not have more benefit for the perioperative patient.

Although laboratory tests can aid in ensuring that a patient's preoperative condition is optimal once a disease is suspected or diagnosed, such tests have several short-comings as screening devices for the discovery of unknown disease. First, they frequently fail to uncover pathologic conditions. Second, they detect abnormalities the discovery of which does not necessarily improve patient care or outcome. Also, laboratory tests are inefficient in screening for asymptomatic diseases. Finally, most abnormalities discovered on preoperative screening, or even on admission screening for nonsurgical purposes, are not recorded (other than on the laboratory report) or pursued appropriately.

By itself, the detection of abnormalities does not justify testing, because most abnormalities in asymptomatic patients do not reflect the presence of disease. For tests reported as continuous results, the distribution of results in a population of patient is gaussian (i.e., normal). The values defining "abnormal" are set arbitrarily, so that test results exceeding those of the highest 2.5% of healthy individuals, or falling below those of the lowest 2.5% of healthy individuals, are said to be abnormal. Test results between these two extremes are "within the reference range." Therefore, 5% of test results from patients without disease will be "outside the hospital reference range." If one were to order 100 hemoglobin determinations for a sample of healthy patients, 5% of the results would be expected to be "abnormal." Ordering multiple preoperative tests increases the chances of at least one abnormal result.

Assuming that results of tests are independent of one another, the more tests ordered, the higher the likelihood of an abnormal result. For example, if two tests are ordered for a patient without disease, the chance of both being normal is 0.95 × 0.95 or 0.90. For 20 tests, the chance that all would be normal would be only 36%. The chance that at least one result will be abnormal is 64%. Thus, if one uses more than 13 tests to screen patients before surgery, one should expect at least one abnormal test result.

AIDS testing provides another example. More than 92% of the population at low risk for HIV infection who have positive (abnormal) results on two enzyme-linked immunosorbent assays (ELISAs) and one Western blot test in reality do not have HIV infection.^[130] Therefore, it is not surprising that the benefit from nonselective testing is so low or that so few abnormal results arising from unwarranted tests are acted on ( Table 25-5 ). Mammography is another example. If both mammograms and breast examinations were performed yearly starting when women are 40 years old, by age 50 over 49 percent would

943

Even for the very elderly, a patient group at higher risk for morbidity and mortality during surgery, the ultimate benefit of routine laboratory screening is doubtful. Domoto and colleagues^[136] examined the yield and benefit of a battery of 19 screening laboratory tests performed routinely in 70 functionally intact elderly patients (average age, 82.6 years) who resided at a chronic care facility. The 70 patients underwent 3,905 screening tests. "New abnormal" results occurred in 5 of the 19 screening tests. Most of these "new abnormalities" were only minimally outside the normal range. Only four discoveries (0.1 percent of all tests ordered) led to a change in patient management, none of which, these workers concluded, benefited any patient in any important way.

Wolf-Klein and associates^[137] retrospectively studied the results of annual laboratory screening of a population of 500 institutionalized and ambulatory elderly patients (average age, 80 years). From the 15,000 tests performed, 756 new abnormalities were discovered, 690 of which were ignored. Sixty-six of the new abnormalities were evaluated; the result was 20 new diagnoses, 12 of which were treated. Two patients of the 500 ultimately may have benefited from eradication of asymptomatic bacteriuria (although eradication of this condition has not been shown to improve the quality of life or to extend life^[138] ).

Studies show that the history and physical examination are the best ways to screen for disease. Delahunt and Turnbull^[139] evaluated 803 patients who were assessed preoperatively for varicose vein stripping or inguinal herniorrhaphy. A total of 1,972 tests produced only 63 abnormalities not indicated by history or physical findings. Furthermore, in no instance did the discovery of these abnormalities influence patient management.

Another study retrospectively evaluated 690 admissions for elective pediatric surgical procedures.^[140] Bates and colleagues ^[141] found that at least 40% of repeat tests carried out in a large teaching hospital were redundant. The history and physical examination indicated the probability of abnormalities in all 12 patients in whom an abnormality was detected by laboratory testing. Clinical diagnosis, and not laboratory testing, was the apparent basis for any change in operative plans.

Narr and co-workers^[23] at the Mayo Clinic found no harm from omitting all laboratory testing for ASA I patients. The sample size of this study was large enough to indicate that more harm than benefit would probably occur by testing ASA I patients. One exception should be noted: the median age of these patients was 21.4 years, making the conclusion about testing of ASA I patients over 40 years of age unclear. Narr and colleagues have acted on their data: no tests are now obtained on ASA I patients undergoing surgery at the Mayo Clinic. Review of a sampling of such patients' records and outcome indicates the validity of such an approach, and others have reported cost savings and operating efficiency resulting from this practice.^{[110] [142]}

Several studies have compared groups of hospitalized patients undergoing routine laboratory screening tests (to supplement the history and physical examination) with groups not undergoing routine screening tests. Wood and Hoekelman ^[143] found that abnormal results from the history, physical examination, or laboratory examination changed the preoperative clinical course for 28 of 1,924 children: surgery was postponed for all 28 children. For only 3 of these children did laboratory tests indicate an abnormality not suggested by the history or physical examination. Thus, the history and physical examination dictated the appropriate laboratory testing for all but 3 of 1,924 patients.

A more specific conclusion is also possible. The abnormalities discovered in these three patients were found on chest radiographs. (These children were part of a study comparing perioperative outcome at two hospitals, one that required chest radiograph as a screening test for elective surgery in children and one that did not.) There were no differences noted in anesthetic or perioperative complications between the two groups. Therefore, Wood and Hoekelman recommended that chest radiographs not be obtained routinely for apparently healthy children.

Even in a referral population, history and physical examination determine more than 90% of the clinical course when a patient is referred for consultation about cardiovascular, neurologic, or respiratory disease.^[144] Other studies also have demonstrated that the history and physical examination accurately indicate all areas in which subsequent laboratory testing proves beneficial to patients. For example, Rabkin and Horne^{[145] [146]} examined the records of 165 patients having a "new" abnormality on the ECG that was "surgically significant" (i.e., a change from a previous tracing that represents a condition possibly affecting perioperative management or outcome). In only two instances were anesthetic or surgical plans altered by the discovery of "new abnormalities" found on the ECG but not indicated by history. Thus, even for these 165 patients, for whom the benefits of a laboratory test should have been maximal because abnormalities were detected before surgery, the history or physical examination determined case management most of the time. Furthermore, in one of the two instances of altered case management (a patient who had atrial fibrillation), physical examination should have indicated the need for an ECG. A history or physical examination was not available for the other patient.

Table 25-5 shows that patients who benefit from testing have risk factors, symptoms, or other conditions in their history that call for testing. In our own study (Apfelbaum JL et al, unpublished data), in patients who were symptomatic with risk factors for disease or who only had risk factors for disease, 606 (5.8%) of 10,419 test results were significantly abnormal. Of these, 124 tests (1.2%) affected care. Of these patients, 6 tests resulted in harm (6 in 10,419, or 0.06%), whereas 91 patients (91 in 10,419, or 0.9%) benefited from a change in care. In contrast, for asymptomatic patients who had no risk factors for disease, only 121 (1.1%) of the 10,899 test results were significantly abnormal. Of these, 10 tests (0.01%) affected care. During the study, every change in care that benefited or harmed a patient stemmed from a single test result. Therefore, the 13 "care-affecting" tests represented 13 "care-affected" patients. Of these 13 patients, 5 (in 10,899, or 0.05%) asymptomatic patients were harmed, whereas only one (in 10,899 test results, or 0.009%) of the patients benefited from a change in care. Neither harm nor benefit was

944

In summary, the studies cited point to the lack of benefit from routine laboratory tests as a method of assessing patients preoperatively. Many of these laboratory tests have been shown to be superfluous to patient care management. History and physical examination are considered the most effective ways of screening for disease. Laboratory tests can be used to screen for disease when the patient has appropriate risk factors and when such tests have proved effective. However, the better use of such tests is to confirm clinical diagnoses or to optimize the patient's condition prior to surgery.