Review articleThe case against routine preoperative laboratory testing
Section snippets
Normal and abnormal test results
It is important for physicians to understand how laboratories define normal and abnormal test results. For many blood test results, a continuous distribution of results is possible. For example, a hemoglobin level may theoretically be measured from 0 through infinity with any value in between. Other tests may have only ordinal results. For example, the degree of proteinuria reported on a dipstick determination may be reported only in a small number of discrete values (1+, 2+, etc.). Finally,
Rationale for preoperative testing
There are several theoretic reasons why clinicians might order routine preoperative tests. These include: (1) to detect unsuspected abnormalities that might influence the risk of perioperative morbidity or mortality; (2) to establish a baseline value for a test that has a high likelihood of being monitored and changing after the surgical procedure is complete; and (3) for medical-legal reasons.
For the first rationale, three actions are possible regarding an abnormal test result. First,
Hemoglobin
Blood loss as a result of surgery is common and, in many types of surgery, perioperative transfusion is necessary. Severe anemia during surgery risks tissue hypoxia from impaired oxygen delivery. Thus, unsuspected severe anemia theoretically may predispose the patient to tissue hypoxia in the perioperative period if not corrected preoperatively. For this reason, a case can be made on the basis of physiology for routine hemoglobin (or hematocrit) measurements for patients in whom significant
White blood cell count
The prevalence of unanticipated elevations of white blood cell counts is very low. In 4 of the 5 available studies, the prevalence was less than 1% (Table 2). In the two studies in which authors assessed management changes related to abnormal white blood cell counts, no patient underwent a management change [11], [12]. In the two studies that allow the calculation of likelihood ratios for abnormal white blood cell counts, the findings suggest that an unanticipated elevated white blood cell
Platelet count
Ten investigations have been published regarding the usefulness of platelet counts (Table 3). The incidence of abnormalities of platelet counts is 0.9%. One study reported a yield of 8%, but all patients in this investigation had slightly elevated counts rather than thrombocytopenia [11]. Management changes on the basis of unanticipated abnormalities in platelet count are rare. In our article, only 0.02% of all measurements of platelet count were abnormal and influenced surgical management. For
Coagulation tests
Many different coagulation tests are available to the clinician. Theoretically, these tests might detect a predisposition to perioperative hemorrhage. In this article, we discuss only the prothrombin time and partial thromboplastin time, as they are the most commonly ordered laboratory tests to screen for coagulation disorders. We have chosen not to discuss the bleeding time as it is infrequently used as a routine screening test before surgery. Furthermore, the studies that have investigated
Electrolytes
A theoretic rationale for measuring serum electrolytes routinely before surgery would be to identify patients at risk for adverse postoperative events including arrhythmia. Previous studies have primarily evaluated abnormalities of serum sodium and potassium. Clinical lore suggests that hypokalemia would be the most clinically important preoperative electrolyte abnormality to detect. The evidence does not, however, suggest a relationship between this laboratory abnormality and postoperative
Renal function tests
Over the past decade, a growing body of literature suggests that preoperative renal insufficiency is one of the most important risk factors for postoperative complications in both cardiac and noncardiac surgery. For example, in the revised cardiac risk index of Lee et al, preoperative serum creatinine >2.0 mg/dL was 1 of 6 predictors of risk for postoperative cardiac complications in a validated multifactorial analysis of patients undergoing noncardiac surgery [22]. The recently revised
Serum glucose
Established diabetes requiring treatment is a risk factor for postoperative cardiac complications. In the recent revised cardiac risk index by Lee et al, diabetes requiring insulin therapy was 1 of 6 independent risk factors for postoperative cardiac complications; the relative risk associated with this factor was 3.0 [22]. Diabetes also increases morbidity and mortality among patients undergoing coronary artery bypass surgery [24], [27]. In addition, diabetes increases the risk of sternal
Hepatic tests
Patients with advanced liver disease and cirrhosis have a marked increase in the risk of postoperative complications and death. The risk increases with increasing severity of liver disease, as classified by the Child-Pugh criteria. For example, among patients with cirrhosis undergoing abdominal surgery, mortality rates for Child-Pugh class A, B, and C are 10%, 31%, and 76%, respectively [30]. But few data exist to suggest that asymptomatic elevations of transaminases or alkaline phosphatase in
Urinalysis
The rationale for routine preoperative testing would be to identify asymptomatic abnormalities that would modify preoperative care or predict an increased risk for postoperative complications. Potential relationships between abnormalities and adverse outcomes include pyuria and wound infection, and glucosuria and hyperglycemic complications. Table 10 summarizes the primary data sources on preoperative urinalysis. The incidence of abnormal urinalyses was 19.1%. Among the subset of studies that
Electrocardiogram
The potential value of a routine preoperative ECG would be to detect abnormalities that would increase the risk of postoperative cardiac complications or to serve as a baseline in the event that a postoperative ECG is required. Baseline findings that may modify risk include the finding of Q waves that confer risk according to the original Goldman cardiac risk index [35]. If one uses, however, the more recent revised risk index, which outperformed the original Goldman index, any clinical
Chest radiograph
Routine preoperative chest radiographs are more likely to be abnormal than are most other preoperative tests. For example, in a study of 3959 patients who received a preoperative chest radiograph, 23% of studies were abnormal [40]. Abnormalities included those of lung parenchyma (13%), heart (7%), pleura (2%), mediastinum (3%), and other abnormalities (2%). In our survey of 15 studies of routine preoperative chest radiographs (Table 12), 21.2% of all such studies were abnormal. Two potential
Outcomes of patients with no routine perioperative testing
Our review establishes the low predictive value of most commonly obtained preoperative laboratory tests. Several recent reports have reached the same conclusion through evaluation of the outcomes of patients subjected to surgery with no routine preoperative testing. Narr et al performed a retrospective review of 1044 patients who underwent elective surgery or diagnostic procedures [45]. The patients were disproportionately young and healthy. The median age was 21 years, and 97% of patients were
Recommendations
We have shown that routine preoperative testing before elective surgery, without regard to patient-related factors that increase the likelihood of abnormal test results, leads to a low incidence of abnormal results—most of which are either ignored by clinicians or are false-positive results that do not predict an increase in the risk of postoperative complications. Several tests, in particular, BUN or creatinine, and hemoglobin measurements, are more likely to identify a subset of patients at
Using protocols and policy to influence testing
Most institutions have established minimal criteria for routine preoperative testing. Thus, facility policies often require the physician caring for the patient to obtain a minimal set of tests before surgery. In many settings, assurance that the tests are complete is the responsibility of support staff for the surgeon or anesthesiologist. If tests are not obtained before surgery, the patient's surgery may be delayed; hence, most physicians are compliant with these policies.
The establishment of
Summary
In this article, we have shown that almost all “routine” laboratory tests before surgery have limited clinical value. Clinicians should order only a small number of routine tests based on age as noted in Table 13. Selective use of other preoperative tests should be based on history and physical examination findings that identify subgroups of patients who are more likely to have abnormal results. In general, clinicians should order tests only if the outcome of an abnormal test will influence
References (83)
Preoperative laboratory testing: should any tests be “routine” before surgery?
Med Clin N Amer
(1993)- et al.
Supraventricular tachycardia after coronary artery bypass grafting surgery and fluid and electrolyte variables
Heart Lung
(1996) - et al.
ACC/AHA guideline update for perioperative cardiovascular evaluation for noncardiac surgery: executive summary. A report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Committee to Update the 1996 Guidelines on Perioperative Cardiovascular Evaluation for Noncardiac Surgery)
J Am Coll Cardiol
(2002) - et al.
Mild renal failure is associated with adverse outcome after cardiac valve surgery
Am J Kidney Dis
(2000) - et al.
Perioperative factors influencing the outcome of elective abdominal aorta aneurysm repair
Eur J Vasc Endovasc Surg
(2000) - et al.
Preoperative prediction of postoperative morbidity in coronary artery bypass grafting
Ann Thorac Surg
(1996) - et al.
Preoperative laboratory screening in healthy Mayo patients: cost-effective elimination of tests and unchanged outcomes
Mayo Clin Proc
(1991) Surgery in the patient with liver disease
Mayo Clin Proc
(1999)- et al.
Efficacy of the routine admission urinalysis
Am J Med
(1987) - et al.
The unproven utility of the preoperative urinalysis: economic evaluation
J Clin Epidemiol
(1989)