Secondary Logo

Journal Logo

CARDIOVASCULAR ANESTHESIA: Society of Cardiovascular Anesthesiologists: Research Report

The Prevalence and Predictive Value of Abnormal Preoperative Laboratory Tests in Elderly Surgical Patients

Dzankic, Samir, MD; Pastor, Darwin; Gonzalez, Carlos, MD; Leung, Jacqueline M., MD, MPH

Author Information
doi: 10.1213/00000539-200108000-00013
  • Free

Abstract

The population in the United States is aging rapidly. Currently, 13% of people living in the United States are older than 65 yr. As the average life span is increasing, the US Census Bureau projects that in the year 2030, 20% of the population will be ≥65 yr old (1). The use of health care resources by the elderly is disproportionately larger than the resources used by their younger counterparts. According to a report by the National Center for Health Statistics, 35% of procedures are performed in older patients (2). Surgical procedures in older patients are associated with an increased morbidity and mortality rate, possibly because of an increased prevalence of coexisting diseases, decreased physiologic reserve, or a combination of these (3).

Preoperative laboratory testing is performed as part of preoperative assessment. Current recommendations are that healthy older patients (>60 yr old) should be routinely tested for hemoglobin and hematocrit, glucose, blood urea nitrogen and creatinine, 12-lead electrocardiogram, and chest radiograph abnormalities (4,5). However, the usefulness of routine preoperative laboratory testing as a part of preoperative assessment has been scrutinized because of the escalation of health care costs (6–9). It is estimated that between $3 billion and $11 billion are spent in the United States annually on preoperative laboratory testing for all types of surgery (10,11), of which approximately one third is spent on the geriatric surgical patient population. The practice of routine preoperative laboratory testing in healthy older surgical patients is therefore costly, especially because evidence is lacking that such testing may predict or improve perioperative outcomes.

Several studies have investigated the prevalence of abnormal laboratory tests in the geriatric population (12–16), but few have focused specifically on geriatric surgical patients. Accordingly, we conducted a prospective study to evaluate the prevalence and prognostic significance of abnormal preoperative laboratory tests in a cohort of elderly patients undergoing noncardiac surgery.

Methods

This study was conducted during a 1-yr period in 1997 at one of the University of California, San Francisco teaching hospitals (Mount Zion Medical Center). This hospital is typical of a community medical center providing health care to a heterogeneous group of elderly patients living in a San Francisco neighborhood. After approval by the IRB, 544 consecutive patients, aged 70 yr or older, presenting for anesthetic and noncardiac surgery, were included in the study. Patients presenting for surgery requiring only local anesthesia or monitored anesthesia care were excluded.

Data were obtained from medical chart review, anesthesia preoperative records, and daily progress notes. Preoperative risk factors measured included demographic information (such as age, race, and sex), type of surgery, preoperative clinically determined medical conditions (history of angina, dysrhythmia, myocardial infarction, hypertension, congestive heart failure, diabetes mellitus, pulmonary diseases, renal disease, liver dysfunction, neurologic disease, hematologic dysfunction, and history of alcohol consumption and smoking), and ASA physical status classification. The ASA status was determined before surgery, from history and physical examination by the attending anesthesiologists. We also measured surgical risk by the stratification method recommended by the American Heart Association/American College of Cardiology (17). Clinically obtained preoperative laboratory results (hemoglobin and hematocrit, potassium, sodium, glucose, creatinine, and platelet count) were measured. In this cohort, all patients were scheduled to have a complete blood count, including platelet count, before surgery. The performance of other laboratory tests was not controlled by the study. The limits of abnormality chosen for each test were based on the range of normal values determined by our institutional clinical laboratory (Table 1).

Table 1
Table 1:
Prevalence of Abnormal Preoperative Laboratory Results

After surgery, patients were followed prospectively daily while in the hospital by a physician investigator until discharge for the occurrence of predefined in-hospital postoperative adverse outcomes (cardiovascular, pulmonary, renal, hepatic or gastrointestinal, neurologic, infectious, thromboembolic, or surgical difficulties, reoperation, and death). All potential adverse outcomes were validated by a second investigator. Definitions of postoperative adverse outcomes are listed in Table 2.

Table 2
Table 2:
Definition of Postoperative Adverse Outcomes

Categorical variables with two levels, such as history of hypertension, diabetes mellitus, heart disease, or comorbid diseases, were coded as 0 = absent, 1 = present, and unknown (missing) if information could not be determined from medical record review. Nonordered categorical data with more than two levels (such as types of surgery or anesthesia) were entered as K-1 dummy variables (indicator variables). To test for the association of preoperative risk factors with postoperative adverse outcomes, we used a two-step process. First, we tested each preoperative risk factor independently by using univariate analysis, specifically χ2 with Yates correction or Fisher’s exact tests. Variables with continuous data, such as age and laboratory data, were initially examined by logistic regression for their association with postoperative adverse outcomes. In addition, to test the association of abnormal preoperative laboratory tests with postoperative adverse outcomes, laboratory data were stratified into normal and abnormal by using specific cutoff values that were based on the range of normal values determined by our institutional clinical laboratory. These data were further analyzed with similar methods as described previously. Odds ratio (OR), 95% confidence interval (CI), and P values were reported. In a second step, those variables that had significant association with postoperative adverse outcomes on univariate analysis (P ≤ 0.1) were entered in a stepwise multivariate logistic regression model. By using this statistical model, each variable was examined for association with postoperative adverse outcomes while we controlled for all other confounding variables. The contribution of each risk factor was assessed by testing the regression coefficient against zero. Risk factors were removed from the model in a stepwise fashion: the risk factor showing the smallest contribution was deleted at each step. The model-building was stopped when all risk factors remaining in the model had regression coefficients significantly different from zero. P values, OR, and 95% CI were reported. A P value <0.05 was considered statistically significant. Statistical analysis was performed with STATA™ statistical software version 5.0 (Stata Corporation, College Station, TX).

Results

Overall, 544 patients who underwent 575 surgical procedures were studied. For those patients who underwent more than one surgical procedure during the study period, only the first surgical procedure was included in this analysis. The demographic data are presented in Table 3. The mean age was 78 ± 6 yr (range, 70–100 yr). Fifty-six percent of our patients were female, and 44% were male. The majority of patients were Caucasian (82%). The most frequently performed procedures were general surgery (24%), followed by orthopedic surgery (18%). The majority of the patients were classified as ASA II and III (45% and 48%, respectively). The overall in-hospital mortality rate was 3.7%, and 110 patients (20%) developed at least one postoperative adverse outcome (Table 4). The most common postoperative adverse outcomes were cardiac and neurologic complications.

Table 3
Table 3:
Demographic Data
Table 4
Table 4:
Distribution of Postoperative Adverse Outcomes

The prevalence of abnormal laboratory results in our study population is shown in Table 1. A total of 2462 tests were performed, with 170 (6.8%) demonstrating abnormal values. The largest prevalence of abnormal results was found for creatinine (12%), hemoglobin (10%), and glucose (7%) values. Abnormal electrolytes values were infrequent (0.7%–5%). Very few patients had abnormal platelet counts (1.9%). Forty-five percent of all tests were performed in patients with ASA class ≤II, and 55% were performed in patients with ASA class >II. The percentage of abnormal tests in those patients with ASA ≤II was 3.6% (40 of 1112) vs 9.6% (130 of 1350) in those with ASA >II (P < 0.0001) (Table 5).

Table 5
Table 5:
Prevalence of Abnormal Laboratory Results Stratified by ASA Classification

By univariate analysis, high creatinine values, hypernatremia, ASA classification (>II), surgical risk, and age increased the odds of postoperative adverse outcomes (Table 6). Of the coexisting medical conditions, a history of myocardial infarction, congestive heart failure, pulmonary diseases, neurologic diseases, and alcohol consumption (more than two drinks per day) were associated with postoperative adverse outcomes. In contrast, hyperglycemia (glucose values >200 mg/dL), anemia (hemoglobin values <10 g/dL), abnormal potassium levels, hyponatremia, and low platelet count (<115 × 109) were not associated with postoperative adverse outcomes (Table 6).

Table 6
Table 6:
Association of Specific Risk Factors with Adverse Postoperative Outcomes, by Univariatea and Stepwise Multivariate Logistic Regression Analysisb

All variables that were significant after univariate analysis were entered in a stepwise multivariate logistic regression model. In the final model, only ASA classification (>II) (OR, 2.55; 95% CI, 1.56–4.19;P < 0.001) and surgical risk (OR, 3.48; 95% CI, 2.31–5.23;P < 0.001) were significant independent predictors of postoperative adverse outcomes (Table 6).

Discussion

Our study demonstrated that the prevalence of abnormal electrolyte values and thrombocytopenia in elderly surgical patients was small (0.5%–5%). The prevalence of anemia, high creatinine values, and hyper-glycemia was larger (10%, 12%, and 7%, respectively). However, none of the abnormal preoperative tests was associated with postoperative adverse outcomes.

Our results differ from two previous studies that investigated the predictive value of preoperative laboratory tests in general surgical patients (5,18). Velanovich (5) found that an abnormal preoperative glucose value was an independent predictor of postoperative bleeding, whereas McKee and Scott (18) demonstrated the association of postoperative complications with any preoperative laboratory abnormalities. Because both studies demonstrated that the prevalence of preoperative laboratory abnormalities increased with age, they recommended that routine preoperative testing should be performed in patients >60 years old. There are several potential differences between these two studies and ours. First, our population consisted of patients who were ≥70 years of age, whereas the previous two reports studied a much younger group. Second, despite the younger patient population studied by these two studies, a larger prevalence of preoperative laboratory abnormalities was found. The differences in the prevalence of laboratory abnormalities are probably secondary to the narrower range of values defined as normal (except for creatinine) by these studies. Some laboratory test results in these studies (5,18) may be only marginally abnormal and are likely to be of little clinical significance. Therefore, we chose wider ranges to define normal values in order to increase the specificity of the tests and, therefore, their clinical significance. Third, one of the studies (18) performed only univariate analysis to test the association of laboratory abnormalities with postoperative complications without adjusting for the potential confounding effects of covariates.

The range of abnormal laboratory results in the geriatric population reflects the population being studied and the reference range used to define normal values. In general, the healthy geriatric population has a small prevalence of having abnormal laboratory tests (12,14). For example, in a population study of 7196 ambulatory patients, the prevalence of hemoglobin, glucose, and creatinine abnormalities was small (5.5%, 8.3%, and 2.7%, respectively) (12). A similar prevalence was found in a study of geriatric patients presenting for ophthalmic surgery, in which only 5.9% of routinely obtained tests were abnormal (19). In contrast, a substantially larger prevalence of abnormal results was demonstrated in a group of institutionalized elderly patients (11%–33% for hemoglobin, 11%–15% for creatinine, 25%–29% for glucose, 3%–5% for potassium, 9%–10% for sodium, and 4% for platelets) (13,15,16). The range of abnormal laboratory values in our patients fits in between the general and the institutionalized geriatric population. In fact, in a separate analysis for patients classified as ASA I–II, we found the prevalence of laboratory abnormalities to be as small as those found in the general population (3.6%). These results suggest that routine preoperative testing in geriatric surgical patients, particularly in those patients classified as ASA I–II, generally produce few abnormal results.

Our patient population was drawn from a teaching hospital providing care to community-based patients. To determine whether our patients are representative of the general geriatric population, we compared the health status of our patients and their postoperative complication rates with other geriatric populations. The prevalence of chronic medical conditions, such as hypertension and diabetes, in our patients was similar to that of another cohort of high-functioning older people (20). However, more patients in our group had a history of a stroke and angina, most likely reflecting the fact that 35% of our patients were ≥80 years old. The complication rates in our patient population were comparable to those previously reported. In a study of 7306 consecutive patients presenting for noncardiac surgery during 1987 in Denmark (21), the authors found that the rate of cardiovascular complications inpatients ≥70 years was 14.9% (vs 10% in our study), the rate of pulmonary complications was 9.2% (vs 5% in our study), and the mortality rate was 3.8%(22) (vs 3.7% in our study).

Our results in the geriatric surgical patients agree with previous studies that demonstrated the limited value of routine preoperative testing in the general surgical population. These studies found that tests performed without any clinical indications yield a small percentage of abnormal results (7,8,23–26) and rarely cause the change of perioperative management (7,8,23,24,27) or predict the occurrence of postoperative complications (7,8,23–25). If routine preoperative laboratory testing were of limited value in geriatric surgical patients, would the elimination of these tests have a negative effect on a patient’s perioperative care and postoperative outcome? A recent study by Schein et al. (28) with >19,000 elderly patients randomized to undergo cataract surgery, with or without a standard battery of laboratory tests, demonstrated that perioperative morbidity and mortality rates were similar in both groups. The authors recommended that preoperative testing in geriatric patients undergoing cataract surgery or procedures with similar surgical risk should be performed only when clinically indicated by history or physical examinations. The surgical risk associated with cataract surgery is small, and therefore the results of this study may not be directly generalized to all geriatric surgical patients. However, our results, along with those from Schein et al., would suggest that the recommendations to eliminate routine preoperative laboratory testing may be extended to geriatric surgical patients with few comorbidities (e.g., ASA I–II). Our results also suggest that the present guidelines for routine preoperative laboratory testing in older patients should be reevaluated. Specifically, routine preoperative laboratory testing for hemoglobin, creatinine, glucose, platelets, and electrolytes on the basis of age alone may not be indicated. Rather, the performance of these tests should be based on the preoperative history and physical examination, which will determine patient’s comorbidities, type of surgery, and the likelihood that these tests will change perioperative care. Our study suggests that it is the performance of the preoperative history and physical examination that is the critical part of preoperative evaluation, rather than simply testing alone.

The economic implications of our study may be substantial by eliminating many routine preoperative laboratory tests in relatively healthy older surgical patients (ASA class I–II). Nationwide, more than 16 million older patients (>65 years old) undergo surgery annually (2). By using the results from our study, we can estimate that approximately half of the geriatric surgical patients may be classified as ASA I–II. By applying the current reimbursement rates by Medicare of $36.32 per patient for complete blood count with platelets, electrolytes, creatinine, and blood urea nitrogen, and glucose measurements (29), approximately $290 million may be potentially saved annually by eliminating these preoperative tests in geriatric surgical patients who are classified as ASA I–II. In addition, potential cost savings may also be obtained if abnormal test results delay surgery or result in additional follow-up tests.

There are several potential limitations to our study. First, we collected and examined only clinically available data. As shown in Table 1, only half of our patients had preoperative glucose, and two thirds of our patients had preoperative electrolytes and creatinine values measured. However, the multivariate logistic regression analysis included only patients with both preoperative laboratory studies and outcomes. Also, further analysis demonstrated that patients who have no postoperative complications had similar rates of missing data as those with postoperative complications. Similarly, there was no difference between missing laboratory studies and specific surgical procedures. We cannot determine the predictive value of other laboratory tests that are not routinely obtained, such as preoperative albumin level, which was an important predictor of postoperative morbidity and mortality in a Veterans Affairs surgical risk study (30), as well as other tests, such as the six-minute walk test, which assesses exercise capacity.

Second, if physicians treated the conditions related to the preoperative laboratory abnormalities either before or during surgery, postoperative outcomes may have been altered. We examined this possibility by reviewing the surgical cases that were canceled before surgery and found that none was canceled because of preoperative laboratory abnormalities. However, we cannot determine how many patients were under the care of a physician chronically before their preoperative evaluation. In general, most patients included in the study typically had the battery of laboratory tests performed immediately before surgery. Thus, the possibility that abnormal laboratory tests were corrected before preoperative testing was unlikely. We also reviewed the intraoperative management and found that no patients received electrolyte supplementation or platelet transfusion because of preoperative abnormalities. Intraoperative transfusion of red blood cells for minimal blood loss (<200 mL) occurred in only four patients. Although we cannot determine with certainty, it is unlikely that the outcomes for these patients would have been altered simply because of intraoperative red blood cell transfusions.

Third, because of the small prevalence of certain abnormal preoperative tests, we determined the precision of our estimates (ORs) for the laboratory tests by examining the width of the confidence intervals. Our results indicated that the confidence intervals were relatively narrow, reflecting the large sample size of our study, and, therefore, the precision of the estimates. However, we note that for ORs of 1.86 (high creatinine), and 2.5 (sodium abnormalities), these estimates may represent a clinically significant increased risk, although our data suggest that there is no substantial predictive value.

Fourth, it should be noted that our study addressed only preoperative laboratory studies, and this is only one of several components of preoperative evaluation. We did not address other issues of preoperative assessment, such as preoperative history and physical examination, which, as suggested from the results of our study, should still serve as the basis for obtaining tests.

Fifth, it is possible that when assigning the ASA classification, the anesthesiologist considered the laboratory abnormalities in conjunction with the health status of the patient. However, this is unlikely because most anesthesiologists assign the ASA classification primarily on the basis of the physical condition of the patient. Obviously, laboratory tests previously performed for reasons unrelated to surgery as part of a disease work-up might influence the ASA classification. It can be inferred from our results that the majority of the patients may have had regular access to medical care before surgery, because 84% of the patients had at least one of the preoperative clinically determined chronic medical conditions. Whether these patients had laboratory tests performed for reasons unrelated to surgery, thus biasing the ASA classification, cannot be determined.

In conclusion, our results demonstrated that the prevalence of preoperative electrolyte and platelet count abnormalities was small and had a small predictive value. Similarly, preoperative anemia and creatinine and glucose abnormalities, although more prevalent, also had small predictive value. By multivariate analysis that included laboratory data, health status, and patient demographics, only ASA classification and surgical risk were independent predictors of in-hospital adverse postoperative outcomes. Routine preoperative testing for hemoglobin, creatinine, glucose, and electrolytes on the basis of age alone may not be indicated in older patients. Rather, selective laboratory testing, as indicated by history and physical examination, which will determine a patient’s comorbidities and surgical risk, seems to be indicated.

References

1. US Census Bureau. Population Projections Program, Population Division: projections of the total resident population by 5-year age groups, and sex with special age categories—middle series, 2025 to 2045. Washington, DC: US Census Bureau; 2000.
2. National Center for Health Statistics. Number and rate of all-listed surgical and nonsurgical procedures for discharges from short-stay hospitals by selected procedure categories: United States. Hyattsville, MD: US Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Health Statistics, Division of Data Services, 1997.
3. Ergina P, Gold S, Meakins J. Perioperative care of the elderly patient. World J Surg 1993; 17: 192–8.
4. Roizen M, Foss J, Fisher S. Preoperative evaluation. In: Miller RD, ed. Anesthesia. 5th ed. New York: Churchill Livingstone, 2000: 855–6.
5. Velanovich V. The value of routine preoperative laboratory testing in predicting the postoperative complications: a multivariate analysis. Surgery 1991; 109: 236–43.
6. Allison J, Bromley H. Unnecessary preoperative investigations: evaluation and cost analysis. Am Surg 1996; 62: 686–9.
7. Delahunt B, Turnbull P. How cost effective are routine preoperative investigations? N Z Med J 1980; 92: 431–2.
8. Kaplan E, Sheiner L, Boeckmann A. The usefulness of preoperative laboratory screening. JAMA 1985; 253: 3576–81.
9. Vogt A, Henson L. Unindicated preoperative testing: ASA physical status and financial implications. J Clin Anesth 1997; 9: 437–41.
10. Fisher S. Cost-effective preoperative evaluation and testing. Chest 1999; 115: 96S–100S.
11. Johnstone R, Martinec C. Cost of anesthesia. Anesth Analg 1993; 76: 840–8.
12. Collen M, Feldman R, Siegelaub A, Crawford D. Dollar cost per positive test for automated multiphasic screening. N Engl J Med 1970; 283: 459–63.
13. Domoto K, Ben R, Wei J. Yield of routine annual laboratory screening in the institutionalized elderly. Am J Public Health 1985; 75: 243–5.
14. Jernigan J, Gudat J, Blake J. Reference values for blood findings in relatively fit elderly persons. J Am Geriatr Soc 1980; 28: 308–14.
15. Levinstein M, Ouslander J, Rubenstein L, Forsythe S. Yield of routine annual laboratory tests in a skilled nursing home population. JAMA 1987; 258: 1909–15.
16. Wolf-Klein G, Holt T, Silverstone F. Efficacy of routine annual studies in the care of elderly patients. J Am Geriatr Soc 1985; 33: 325–9.
17. Report of the American College of Cardiology American Heart Association Task Force on Practice Guidelines (Committee on Perioperative Cardiovascular Evaluation for Noncardiac Surgery). Executive summary of the ACC/AHA task force report: guidelines for perioperative cardiovascular evaluation for noncardiac surgery. Anesth Analg 1996;82:854–60.
18. McKee R, Scott M. The value of routine preoperative investigations. Ann R Coll Surg Engl 1987; 69: 160–2.
19. McKibbin M. The pre-operative assessment and investigation of ophthalmic patients. Eye 1996; 10: 138–40.
20. Seeman T, Charpentier P, Berkman L, et al. Predicting changes in physical performance in a high functioning elderly cohort: MacArthur studies of successful aging. J Gerontol 1994; 49: M97–108.
21. Pedersen T, Eliasen K, Henriksen E. A prospective study of risk factors and cardiopulmonary complications associated with anaesthesia and surgery: risk indicators of cardiopulmonary morbidity. Acta Anaesthesiol Scand 1990; 34: 144–55.
22. Pedersen T, Eliasen K, Henriksen E. A prospective study of mortality associated with anaesthesia and surgery: risk indicators of mortality in hospital. Acta Anaesthesiol Scand 1990; 34: 176–82.
23. Narr B, Hansen T, Warner M. Preoperative laboratory screening in healthy Mayo patients: cost-effective elimination of tests and unchanged outcomes. Mayo Clin Proc 1991; 66: 155–9.
24. Turnbull J, Caro B. The value of preoperative screening investigations in otherwise healthy individuals. Arch Intern Med 1987; 147: 1101–5.
25. Johnson H, Knee-Ioli S, Butler T. Are routine preoperative laboratory screening tests necessary to evaluate ambulatory surgical patients? Surgery 1988; 104: 639–45.
26. Wattsman T, Davies R. The utility of preoperative laboratory testing in general surgery patients for outpatient procedures. Am Surg 1997; 63: 81–90.
27. Blery C, Szatan M, Fourgeaux B. Evaluation of a protocol for selective ordering of preoperative tests. Lancet 1986; 1: 139–41.
28. Schein O, Katz J, Bass E, et al. The value of routine preoperative medical testing before cataract surgery. N Engl J Med 2000; 342: 168–75.
29. Health Care Financing Administration. Clinical laboratory. Available at: http://www.hcfa.gov/stats/cpt/clfdown.htm. Accessed May 7, 2001.
30. Gibbs J, Cull W, Henderson W, et al. Preoperative serum albumin level as a predictor of operative mortality and morbidity. Arch Surg 1999; 134: 36–42.
© 2001 International Anesthesia Research Society