| JOURNAL HOME | CME HOME | THIS MONTH | PAST ISSUES | ETOC | COLLECTIONS |
| AUTHORS | REVIEWERS | EDITORIAL BOARD | FEEDBACK | RSS | HELP |
 
|
|
|
||
|
|
Anesth Analg 2008; 107:325-332 © 2008 International Anesthesia Research Society doi: 10.1213/ane.0b013e3181770f55
Section Editor: |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
Abstract |
|---|
 Top Abstract IntroductionMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
BACKGROUND: A previously published clinical trial of epidural-supplemented versus general anesthesia, Veterans Affairs Cooperative Study No. 345, showed no difference in 30-day mortality and morbidity rates between the two treatments. We hypothesized that long-term postoperative survival would be increased by epidural anesthesia/analgesia supplementation during colon cancer resection.
METHODS: We studied long-term survival after resection of colon cancer in a trial of general anesthesia with and without epidural anesthesia and analgesia supplementation for resection of colon cancer in Veterans Affairs Cooperative Study No. 345. Cox and log-normal survival models were used to test the effects of pathological stage, type of anesthesia and other covariates on survival in 177 patients.
RESULTS: The presence of distant metastases had the greatest effect on survival. Thus, analyses were performed separately for patients with and without metastases. For those without metastasis, the hazard ratio for the treatment effects changed at 1.46 years. Before 1.46 years, epidural supplementation was associated with improved survival (P = 0.012), while later, the type of anesthesia did not appear to affect survival (P = 0.27). Hypertension was associated with poorer survival (P = 0.029), as was alcoholism in patients who received epidural anesthesia (P = 0.014). Survival of patients with metastases was unaffected by type of anesthesia. There was a significant age by hypertension interaction (P = 0.002). Patients survived longer if they were hypertensive, but had reduced survival if they were older than 66 years and hypertensive.
CONCLUSION: Epidural supplementation was associated with enhanced survival among patients without metastases before 1.46 years. Epidural anesthesia had no effect on survival of patients with metastases. Additional studies to confirm or refute these findings are warranted.
|
Introduction |
|---|
|
TopAbstractIntroductionMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Epidural anesthesia may theoretically affect long-term survival after colon cancer surgery due to changes in visceral blood flow and immune function. To explore this possibility, we examined the effect of epidural-supplemented general anesthesia versus general anesthesia without epidural supplementation on long-term survival of patients having resection of colon cancer who had been enrolled in Veterans Affairs (VA) Cooperative Study 345, Effect of Epidural Anesthesia and Analgesia on Perioperative Outcome.1 That study was initially designed to compare the short-term effects of general anesthesia with and without epidural anesthesia and analgesia supplementation in patients undergoing abdominal surgery.
|
METHODS |
|---|
|
TopAbstractIntroductionMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
The Original Randomized Study
Cooperative Study Number 345 (CSP 345) was a multicenter, prospective, randomized study to determine whether epidural anesthesia and postoperative epidural analgesia decreased the incidence of death and major complications during and after four types of intraabdominal surgical procedures. The results of the study have been published previously.1 It enrolled 1021 patients who required anesthesia for intraabdominal aortic, gastric, biliary, or colon operations between March 1992 and August 1994. All patients were male, admitted to 1 of 15 participating Department of VA Medical Centers with ASA physical status I-III, and having elective surgery. Patients were excluded if they had had a myocardial infarction within 6 months of screening, abdominal surgery within 3 months of screening, history of chemotherapy or immune suppressive therapy other than steroids, or had contraindications to epidural anesthesia. Enrolled patients were assigned randomly to receive either general anesthesia with postoperative analgesia of parenteral opioids, or general anesthesia supplemented with epidural anesthesia and postoperative epidural analgesia. The randomization was stratified within each site, balancing surgical type, age, and the Goldman index of cardiac risk.2 The primary end-point was death or morbidity, including myocardial infarction, congestive heart failure, ventricular tachycardia, third degree heart block, severe hypotension, pulmonary embolus, respiratory failure, thrombosis, hemorrhage, cerebral hypoxia or renal failure, within 30 days after surgery. Patients were also monitored for postoperative pain, time of ambulation, and length of hospital stay. The study found no significant difference in the incidence of death and major complications between the two treatment groups in the 30-day postoperative period.1
In CSP 345, patients randomized to epidural supplementation received 0.5% bupivacaine with epinephrine 1:200,000 via a lumbar or thoracic epidural catheter. A level of anesthesia of at least T-6 was established before inducing general anesthesia. General anesthesia was maintained in both groups with isoflurane, nitrous oxide, vecuronium, and fentanyl. In the group randomized to epidural supplementation, boluses of 5–10 mL of 0.5% bupivacaine with epinephrine were given as needed. The catheter was used as long as judged clinically appropriate after surgery for pain management, without a specific protocol. Epidural morphine and other drugs were used at the discretion of the local clinicians. Patients randomized to unsupplemented general anesthesia received IV opioids for postoperative pain.1
Long-Term Survival
We obtained approval from the IRBs of all the study centers willing and able to participate. This approval included authorization to perform the study without recontacting the original study patients to obtain their consents, to collect the names, social security numbers, and pathologic tumor stage of the patients enrolled in the trial, and to use this information to determine how long patients survived after surgery.
Long-term survival of CSP 345 patients who had undergone surgery for colon cancer was obtained by querying the VA Beneficiary Information and Records Locator System (BIRLS).3,4 This database is considered to have a very high degree of accuracy because it is based on the payment of burial benefits when a veteran dies. Names, social security numbers, and birth dates, used as identifiers for the BIRLS system, were obtained from patients’ original consent forms and records. Survival follow-up ended in December 2002. Survival was calculated as the number of days from the date of randomization in the study to the date of death or last contact with the patient and was converted to years by dividing by 365.25.
The electronic database from the original study was used to determine baseline variables and risk factors for survival, type of anesthesia, and date of surgery. Tumor-Nodes-Metastasis staging was obtained from the tumor registries at the participating hospitals. Tumor-Nodes-Metastasis staging was translated into stages 0–IV using the pathologic staging system that was in use at the time the patients had their surgery.5
Statistical Analysis
Student's t-test and 
2 contingency tables were used to compare baseline characteristics between the two treatments. P values were not adjusted for multiple comparisons among these characteristics. Univariate survival analyses were performed using the Kaplan-Meier survival procedure.6 The Tarone-Ware procedure7 was used to compare the effects of treatment on survival stratified by metastasis status. Multivariate survival analyses were performed for the nonmetastasis subgroup using Cox regression.8 The log-normal regression model was used to analyze the data from the metastasis subgroup.
By way of explanation for the nonmetastasis subgroup, in survival analysis, the hazard rate, commonly called the risk and designated as h(t), is the probability of dying per unit time given survival to a given time point. This is also called the force of mortality or the time-specific death rate. A critical assumption for the Cox regression model is the proportionality of the hazard (PH) rates over time. We used the Grambsch and Therneau procedure to test this assumption.9 If the assumption is true, the survival curves will not cross over time. If the curves cross, the PH assumption is violated, as is the case for the nonmetastasis subgroup in this study. In order to analyze survival data that violate the PH assumption, a more generalized extended Cox model was used. This procedure, described in Klein and Moeschberger,10 analyzes the survival curves over two time intervals in which the required PH assumption is satisfied over each interval. The mathematical model for the hazard function at time t, excluding the other covariates that satisfy the proportional hazards assumption, is:
|
|
where h0(t) is the unspecified baseline hazard function; ET1 = TRT if t < t0, 0 otherwise, and ET2 = TRT
if t 
t0, 0 otherwise; and 
1 and 2 are the regression coefficients for treatment (TRT) over each of the two time intervals. Visually, t0 is the point at which the slopes of the survival curves differ because of the change in the hazard ratios (HR).
The HR is now expressed as
|
|
where TRT = 1 for unsupplemented and TRT = 0 for epidural-supplemented general anesthesia.
Estimates of 1, 2, and the regression coefficients for the covariates are obtained by maximum likelihood procedures.11 To determine the optimal change point t0 a profile likelihood procedure12,13 was used as follows: Iterate over each of the n1 plus n2 observed uncensored survival times as possible values of t0. For each observed value, compute the value of the likelihood function with the maximum likelihood estimators of 1, 2, and the regression coefficients for the covariates at that time point from the extended Cox model. Repeating this procedure for each uncensored survival time provides a graph of the likelihood function against time. The point where the graph is at the maximum is the optimal change point t0.
Since the treatment variable in this study was found to violate the PH assumption, model building was performed by initially stratifying on type of anesthesia and considering the following covariates with their two-way interactions: pathological cancer staging/ metastasis, age at randomization, presence or absence of hypertension (as indicated in the medical record), systolic blood pressure, diastolic blood pressure, hematocrit, race, presence of chronic obstructive pulmonary disease, alcoholism, hypertension, diabetes, and whether the patient had ever smoked. These baseline characteristics were taken from the CSP 345 database. Akaike's information criterion (AIC) was used to select the variables.14 This selection procedure balances the model fit against the number of covariates entered. The effect of the change in the hazard ratios at t0 is observed subsequently when the survival curves cross. The extended Cox model procedure produces estimates of the treatment effects represented by the two regression coefficients, 1 and2, and tests for the significance of anesthesia effects, cancer staging, and the other covariate effects. A plot of the log ratio of the survival probabilities was used to visually demonstrate the difference between epidural-supplemented and unsupplemented general anesthesia over time. Pointwise 95% confidence intervals (CI) were constructed using the bootstrap resampling procedure.15
The analysis of the survival data in the metastasis subgroup consisted of fitting the log of the survival time against the candidate covariates. The procedure maximizes the likelihood function, which considers the right-censored survival times. The S-PLUS16 statistical analysis package was used to perform the statistical calculations and plot the survival graphs for the nonmetastasis and metastasis subgroups.
|
RESULTS |
|---|
|
TopAbstractIntroductionMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Patient Population
There were 247 patients in CSP 345 who had surgery for colon cancer. Pathologic staging data were obtained for 177 of these patients and the results are reported here. Pathological staging data were unobtainable for 70 patients because the tumor registries at some sites were not recording data at the time the original study was conducted, or because we were unable to locate an investigator who would be responsible for collecting the pathological data. Many of the original investigators were no longer at the participating medical centers when we attempted to obtain the staging data, approximately 10 years after the original trial. The survival experience for these 70 patients was similar to that for the 177 patients for which staging data were available.
Baseline Characteristics and Pathological Staging
Baseline characteristics of patients who received epidural anesthesia and those who did not were comparable (Table 1). The distributions across treatment groups were comparable for Stages 0–II, but a larger proportion of patients with Stage III or IV disease were randomized to unsupplemented general anesthesia (Table 2; P = 0.029, Kruskal-Wallis test for ordered categories17). Forty-one (44.5%) of the 92 patients in the unsupplemented group had stage III or IV disease whereas 24 (28.2%) of the 85 patients in the epidural-supplemented group had stage III or IV disease. The staging categories were collapsed into nonmetastasis (stages 0–II, n = 112) and metastasis (stages III or IV, n = 65) for the survival analysis.
|
|
Kaplan-Meier Survival Analyses
Kaplan-Meier survival analysis showed a highly significant difference between patients without pathological evidence of metastasis (Stages 0–II) versus those with metastasis (Stages III–IV), combined across the two treatment arms. As expected, patients without evidence of metastasis survived significantly longer than those with metastasis; the median survival time was 6.14 years with 95% CI [5.22, 7.99] for patients without metastasis versus 2.01 years with 95% CI [1.55, 2.47] for patients with metastases (Tarone-Ware: P < 0.0001, Fig. 1).
|
Survival was further examined by treatment within the two metastasis subgroups using the Kaplan-Meier procedure stratified on metastasis status. Patients without metastasis assigned to receive epidural anesthesia exhibited better survival initially. However, the survival curves for the 2 treatments merge at approximately 2.5 years and are coincident until 4 years. At that point, the curves separate, with the group without epidural anesthesia/analgesia trending towards improved long-term survival (Fig. 1). In patients with metastasis, the survival curves for the two treatments coincide until about 1 year, after which they separate, and patients without epidural anesthesia trend towards better long-term survival. Further analyses were performed within the metastasis subgroups to determine whether the survival differences between the treatments were statistically significant.
Survival Analyses by Metastasis Subgroup
Nonmetastasis
The crossing of the survival curves in the nonmetastasis subgroup violates the PH assumption. In order to compensate for this violation and to assess the effects of a set of prognostic factors (covariates) on long-term survival, the extended Cox model described in the Methods section was used.
From the 11 covariates and their two-way interactions, the AIC stepwise procedure, performed on a Cox PH model stratified on treatment, selected hypertension and the interaction of alcoholism with type of anesthesia. These covariates were included as predictors of survival in the extended Cox model in addition to the treatment effects for the first (t < t0) and second (t > t0) time periods. The estimates and P values for the coefficients are reported in Table 3. The overall Cox model is highly significant (likelihood ratio P = 0.002) and the PH assumption is well satisfied over the two intervals, as determined by the Grambsch-Therneau test and plots of the Schoenfeld residuals (not shown). Figure 2 shows the fitted extended Cox model with the hypertension covariate set to 0.64 and 0.47, the proportions of patients with hypertension in the epidural-supplemented and unsupplemented groups, respectively. The treatment by alcoholism interaction for the group with epidural anesthesia was set to its mean of 0.26 (the proportion of these patients classified as alcoholics). Since the treatment by alcoholism interaction for the group without epidural anesthesia/analgesia was not significant and the estimated coefficients and standard errors for the other parameters were essentially the same with or without this interaction, it was removed from the model. The resulting extended Cox model fits the Kaplan-Meier survival curves.
|
|
The slope change point t0, determined by where the profile log-likelihood is maximized, is 1.46 years. This results in later crossing of the survival curves at about 3.2 years. For the reduced model (without the treatment by alcoholism interaction for patients without epidural supplementation; table not shown) the risk of dying in the first 1.46 years for patients who did not receive epidural supplementation is 4.65 times the risk for those who did (P = 0.012, 95% CI [1.40, 15.42]). For patients who survived beyond 1.46 years, not receiving epidural anesthesia appears to provide a lower risk of mortality, 0.71, but the benefit is not statistically significant (P = 0.260, 95% CI [0.39, 1.29]). Ninety-six (86%) of the 112 patients without metastasis survived beyond 1.46 years. Over both time periods, the risk of dying for patients having hypertension was 1.79 times higher than the risk for those who did not have it (P = 0.029, 95% CI [1.08, 2.96]). The risk of mortality for alcoholics who received epidurals was 2.33 times the risk for nonalcoholics receiving epidural anesthesia (P = 0.014, 95% CI [1.19, 4.56]) over both time periods.
Figure 3 shows the plot of the natural log of the ratio of the survival probabilities (equivalently, cumulative hazard rate of epidural-supplemented anesthesia minus the cumulative hazard rate of unsupplemented anesthesia) with pointwise 95% bootstrap confidence limits. The upper 95% confidence limit is below zero before 1.5 years. It is interesting to note that the time point at which this 95% CI crosses zero is essentially the same as the cut point, 1.48 years, obtained from the extended Cox model. Thus, patients who received epidural supplementation had better early survival. However the benefit was lost after 1.5–2.0 years.
|
Metastasis
The survival curves for the metastasis subgroup show satisfactory compliance with the proportional hazards assumption. Several survival models were fit, but the log-normal model fit the Kaplan-Meier curves most closely, as it predicted the natural log of the survival time, t, by a linear combination of the treatment and other covariates. In addition to epidural supplementation, the stepwise AIC procedure selected age (P = 0.059), hypertension (P = 0.003), and the age by hypertension interaction (P = 0.002) (Table 4). The resulting summary survival curves from the log-normal model, with age set to 69 (mean age) for the epidural anesthesia/analgesia group, to 68 (mean age) for the unsupplemented group, and the hypertension covariate set to 0.54 and 0.46 (the proportion of hypertensive patients for the epidural-supplemented and unsupplemented groups, respectively), fit the Kaplan-Meier survival curves very well (Fig. 4).
|
|
Type of anesthesia did not have a statistically significant effect (P = 0.22), and when it was removed from the model the coefficients and the P values for the remaining covariates change very little. The full model shown in Table 4 indicates that advanced age and the diagnosis of hypertension are associated with increased survival time among patients with metastases. However, if a patient is age 66 years or more and hypertensive, the survival time is decreased because of the negative coefficient for the age by hypertension interaction (–0.119). For example, from the log-linear model, a patient receiving epidural anesthesia/ analgesia who was 68 years old at the time of randomization without hypertension would be expected to live 2.48 years whereas the same patient with hypertension would be expected to live 2.01 years.
|
DISCUSSION |
|---|
|
TopAbstractIntroductionMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
The analysis of the data showed metastatic disease to have a large adverse effect on survival, as expected. The early survival benefit for patients without metastases who had epidural-supplemented general anesthesia may have been due to the reductions in thrombotic events,18–20 perioperative infection,19,21 and/or pulmonary complications found among high-risk patients.18,21 Even within the 30-day follow-up period of the original trial, the subgroup having abdominal aortic surgery had a reduced incidence of death and major complications associated with epidural anesthesia. This was attributed to reduced rates of new myocardial infarction, stroke, and respiratory failure compared to patients randomized to general anesthesia without epidural supplementation.1 It is possible that, if the original study had had a 1 to 2 year follow-up period, other subgroups randomized to epidural supplementation might have been found to have improved survival or reduced morbidity rates.
Among patients without metastases, hypertension had an adverse effect on survival, potentially due to the well-defined adverse effects of hypertension on the cardiovascular system. Alcoholism had an adverse effect on survival among patients without metastases who received epidural anesthesia, but no effect on survival among those who received unsupplemented general anesthesia. This may have been due to the presence of alcoholic liver disease, including cirrhosis, in some alcoholic patients. Likewise, cell-mediated immunity, specifically lymphocyte function, particularly CD8+ cells and natural killer cells, are believed to be important in preventing cancer metastasis but are reduced in humans after epidural-supplemented general anesthesia compared to patients receiving general anesthesia.22,23 The suppression of the natural immune response added to the hemodynamic response to epidural anesthesia could compound the effects on the liver, which may already be compromised in alcoholics.
Among patients with lymphatic or more distant metastases, the early survival benefit attributable to epidural anesthesia was not found, possibly because it was masked by the high overall death rate of these patients (Fig. 1). Hypertension was associated with improved survival regardless of type of anesthesia (P = 0.006). The most frail patients with metastatic cancer may have had low blood pressures and poor survival. The interaction of age and hypertension (P = 0.004), such that elderly hypertensive patients did not survive as long, may have been due simply to the fact that, over time, hypertension has a relentless adverse effect on the cardiovascular system. Among patients who were not hypertensive, there was a trend (P = 0.087) for older patients with metastatic disease to survive longer than younger patients. There is controversy as to whether younger patients may have more aggressive tumors and/or poorer survival than older patients.24,25
Our investigation should be viewed as a preliminary exploration of the effect of epidural anesthesia on long-term survival after cancer surgery. It is unfortunate that we do not have data on the cause of death of our patients. Cause of death might have been different in the early postoperative period, during which epidural anesthesia was associated with a significant increase in survival, from causes of death in later years, when there was no benefit associated with epidural anesthesia. Knowledge of the cause of death may have shed some light on the reduced survival probability in all time periods of alcoholic patients without metastases who received epidural anesthesia. Further studies are needed to confirm these findings. Data on other types of cancer, other patient populations, adjuvant chemotherapy or radiation, and time and location of recurrence would also be important. Accordingly, we offer our findings as an impetus for future research in this area.
|
ACKNOWLEDGMENTS |
|---|
We acknowledge the contributions and thoughtful suggestions of Kelvin Lee, PhD, Assistant Director, VA Cooperative Studies Program Coordinating Center, Menlo Park, California, and Molly Kok, Business Officer, Portland VA Medical Center, who obtained the survival times from the BIRLS registry. We also acknowledge the contributions of the following investigators for obtaining IRB approval at their respective VA Medical Centers and/or obtaining the tumor staging data from their tumor registries: John Allison, MD, Charleston, SC; Stephen Baird, MD, San Diego, CA; Eric DuBois, Portland, OR; Stephen Ewing, MD, Minneapolis, MN; Margaret Garrison, Ashville, NC; William Holmes, Boston, MA; Raymond Joehl, MD, Hines, IL; Frank Johnson, MD, St. Louis, MO; Lynell Klassen, MD, Omaha, NE; Steven Krasnow, MD, Washington, DC; William Schmeling, MD, Milwaukee, WI; and David Wong, MD, Long Beach, CA.
|
Footnotes |
|---|
Accepted for publication March 20, 2008.
Supported by a Department of Veterans Affairs Merit Review Epidemiology Grant and by funds from the Department of Anesthesiology & Perioperative Medicine, OR Health & Science University.
None of the authors has financial interests that would be affected by the outcome of this study.
Reprints will not be available from the authors.
|
REFERENCES |
|---|
|
TopAbstractIntroductionMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
- Park WY, Thompson JS, Lee KK. Effect of epidural anesthesia and analgesia on perioperative outcome: a randomized, controlled Veterans Affairs Cooperative Study. Ann Surg 2001; 234:560–71[Web of Science][Medline]
- Goldman L, Caldera DL, Nussbaum SR, Southwick FS, Krogstad D, Murray B, Burke DS, O'Malley TA, Goroll AH, Caplan CH, Nolan J, Carabello B, Slater EE. Multifactorial index of cardiac risk in noncardiac surgical procedures. N Engl J Med 1977;297:845–50[Abstract]
- Fisher SG, Weber L, Goldberg J, Davis F. Mortality ascertainment in the veteran population: alternatives to the national death index. Am J Epidemiol 1995;131:242–50
- Cowper DC, Hynes DM, Kubal JD, Murphy PA. Using administrative databases for outcomes research: select examples from VA Health Services Research and Development. J Med Syst 1999;23:249–59[Medline]
- Beahrs OH, Henson DE, Hutter RVP, Kennedy BJ; Ed's. Am Joint Committee on Cancer manual for staging of cancer. 4th ed. Philadelphia: J.B. Lippincott Co, 1992
- Kaplan E, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc 1958;53:457–81[Web of Science]
- Tarone R, Ware J. On distribution-free tests for equality of survival distributions. Biometrika 1977;64:156–60
[Abstract/Free Full Text] - Cox DR. Regression models and life tables. J Royal Stat Soc Ser B 1972;34:187–202
- Grambsch P, Therneau TM. Proportional hazards test and diagnostics based on weighted residuals. Biometrika 1994; 81:515–23
[Abstract/Free Full Text] - Klein JP, Moeschberger ML. Survival Analysis: Techniques for Censored and Truncated Data. New York: Springer 1997
- DeGroot MH. Probability and statistics. 2nd ed. New York: Adison-Wesley, 1986
- Shao J. Mathematical statistics. 2nd ed. New York: Springer-Velag, 2003
- Tableman M, Kim JS. Survival analysis using S: analysis of time-to-event data. Boca Raton: Chapman & Hall/CRC, 2004
- Akaike H. A new look at the statistical model identification. IEEE Trans Automat Contr 1974;AC-19:716–23
- Efron B, Tibshirani R. The bootstrap (with discussion). Statistical Science 1986;1:54–77
- S-PLUS 7.0 guide to statistics. Vol. 2. Seattle: Insightful Corporation, 2005
- Siegle S, Castellan NJ. Nonparametric statistics for the behavioral sciences. 2nd ed. New York: McGraw-Hill, 1988
- Rodgers A, Walker N, Schug S, McKee A, Kehlet H, van Zundert A, Sage D, Futter M, Saville G, Clark T, MacMahon S. Reduction of postoperative mortality and morbidity with epidural or spinal anaesthesia: results from overview of randomized trials. BMJ 2000;321:1493
[Abstract/Free Full Text] - Tuman KJ, McCarthy RJ, March RJ, DeLaria GA, Patel RV, Ivankovich AD. Effects of epidural anaesthesia and analgesia on coagulation and outcome after major vascular surgery. Anesth Analg 1991;73:696–704
[Abstract/Free Full Text] - Christopherson R, Beattie C, Frank S, Norris EJ, Meinert CL, Gottlieb SO, Yates H, Rock P, Parker S, Perler BA, Williams GM. Perioperative morbidity in patients randomized to epidural or general anesthesia for lower extremity vascular surgery. Anesthesiology 1993;79:422–34[Web of Science][Medline]
- Yeager M, Glass D, Neff R, Brinck-Johnsen. Epidural anesthesia and analgesia in high-risk surgical patients. Anesthesiology 1987;66:729–36[Web of Science][Medline]
- Gu Chen-yi, Lou Yi, Cai Yun-biao, Hu J. Effects of different anesthesia methods on the T lymphocyte subsets and hemodynamics of patients with abdominal tumor in peri-operative period. Chinese J Traditional Western Med 2004;24:973–75
- Volk T, Schenk M, Voigt K, Tohtz S, Putzier M, Kox WJ. Postoperative epidural anesthesia preserves lymphocyte, but not monocyte, immune function after major spine surgery. Anesth Analg 2004;98:1086–92
[Abstract/Free Full Text] - O'Connell JB, Maggard MA, Liu JH, Etzioni DA, Livingston EH, Ko CY. Do young colon cancer patients have worse outcomes? World J Surg 2004;28:258–62[Web of Science][Medline]
- Mostafa G, Matthews BD, Norton HJ, Kercher KW, Sing RF, Heniford BT. Influence of demographics on colorectal cancer. The Am Surgeon 2004;70:259–64
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
