Abstract
The aim of this study was to assess the nationwide use of epidural analgesia (EA) and the incidence of postoperative complications in patients undergoing major liver resections (MLR) with and without EA in the United States. The 2001 to 2010 Nationwide Inpatient Sample was queried to identify adult patients undergoing MLR. A 1:1 matched cohort of patients having MLR with and without EA was assembled using propensity-score matching techniques. Differences in the rate of postoperative complications were compared between the matched groups. We identified 68,028 MLR. Overall, 5.9% of patients in the database had procedural codes for postoperative EA. A matched cohort of 802 patients per group was derived from the propensity-matching algorithm. Although use of EA was associated with more blood transfusions (relative risk, 1.36; 95% confidence interval, 1.12–1.65; P = 0.001) and longer hospital stay (median [interquartile range], 6 [5–8] vs 6 [4–8] days), the use of coagulation factors and the incidence of postoperative hemorrhage/hematomas or other postoperative complications were not higher in patients receiving EA. In conclusion, the use of EA for MLR is low, and EA does not seem to influence the incidence of postoperative complications. EA, however, was associated with an increased use of blood transfusions and a longer hospital stay.
Liver resection is a major abdominal surgical procedure with a high risk of postoperative morbidity and mortality (1). Pain after liver resection can be intense and prolonged (2, 3). Inadequate pain management can lead to increased postoperative morbidity and delayed recovery (4). Epidural analgesia has been shown to provide excellent dynamic pain relief as well as improve postoperative pulmonary, cardiovascular, and gastrointestinal function (2, 5–9). Epidural analgesia can enhance rehabilitation and reduce hospital length of stay after major abdominal surgical procedures, presumably due to superior pain relief as well as reduced opioid use and reduced opioid-related adverse effects (10, 11). Recent studies have shown that the use of epidural analgesia as part of a fast track protocol–enhanced recovery can reduce hospital stay after liver resection (12, 13). However, the use of epidural analgesia in patients undergoing liver resection remains controversial (3), probably due to concerns for postoperative coagulation disturbances and subsequent catastrophic neurologic injuries resulting from epidural hematoma (14, 15). In addition, routine use of epidural analgesia is being increasingly questioned due to its several potential adverse effects (16–18). Current patterns of use of epidural analgesia for liver resection in the US are unknown. Furthermore, data on the benefits and incidence of complications related to the use of epidural analgesia for liver surgery are scant (2, 3). The purpose of this study was to examine the utilization and associated complications of epidural analgesia in patients undergoing open liver resection surgery in the US. We hypothesized that use of epidural analgesia would improve perioperative outcomes after major liver resection surgery. The Nationwide Inpatient Sample (NIS), the largest all-payer inpatient database in the US, was used for this purpose.
METHODS
The population for this study consisted of adult patients undergoing major liver resections (excluding liver transplants) in the US. Data were obtained from the 2001 to 2010 NIS datasets from the Healthcare Cost and Utilization Project of the Agency for Health Care Research and Quality (19). The NIS is a stratified probability sample representing 20% of the universe of US community nonrehabilitation hospitals. To ensure nationwide representativeness, the NIS sampling strategy stratifies hospitals according to five characteristics: geographic region, control (public vs private), urban or rural location, teaching status, and bed size. Once a hospital is selected for the NIS in a specific year, all of its discharge data are included in the survey in that year. Approximately 8 million hospital discharges from about 1000 hospitals are available in the database each year. The number of states contributing to the NIS has been increasing over time, with 33 states contributing in 2001 and 45 states contributing in 2010. Given the deidentified and publicly available nature of the NIS data, the study was determined to be exempt from review by the University of Texas Southwestern Medical Center institutional review board.
Hospital discharges with International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) procedure codes for open major liver resections, including partial hepatectomy (50.22) and total hepatic lobectomy (50.3), were identified from the NIS datasets. Records for patients younger than 18 years or with ICD-9-CM procedure codes for total hepatectomy (50.4), liver transplantation (50.5), liver donors (v59.6), and other nonmajor liver procedures (50.21, 50.23–50.29), as well as liver resections because of trauma (diagnosis ICD-9-CM codes 864–869) and records with codes for complications of transplanted liver (996.82), were excluded from the analyses. Laparoscopic liver resections (ICD-9-CM procedure code 50.25) were also excluded. Patients who received epidural analgesia for postoperative pain management were identified using the ICD-9-CM procedure codes 03.90 and 03.91.
The Agency for Health Care Research and Quality Comorbidity Software, a family of tools developed as part of the Healthcare Cost and Utilization Project, was used to create 26 comorbidity variables (including hepatitis, alcoholism, and coagulation disorders) from the up to 25 ICD-9-CM diagnosis codes available in each hospital discharge record (20). Furthermore, the Deyo adaptation of the Charlson comorbidity index was calculated for each patient based on ICD-9-CM diagnosis codes available from the database. The Charlson comorbidity index is a validated measure for use with administrative data that correlates with in-hospital morbidity and mortality after surgical procedures (21). Charlson scores were further collapsed into three categories: 0; 1 to 2; and ≥ 3. Geographic region was defined according to the hospital's census region in Northeast, Midwest or North Central, South, and West. Hospital characteristics including teaching status of hospital (teaching vs nonteaching), location of hospital (urban vs rural), and bed size (small, medium, large) are provided as separate variables in the NIS.
Based on ICD-9-CM codes assigned to the principal diagnosis, cases were categorized as primary malignant neoplasm of the liver or bile ducts (155.0–156.9), secondary malignant neoplasm of the liver (197.7), and other benign diseases of the liver (211.5, 572–576). Cirrhosis and other chronic liver diseases were identified by the codes 571.0–571.9 and chronic viral hepatitis B and C using codes 070.2, 070.3, and 070.7. Variables were created to adjust for the effect of the type of principal diagnosis on outcomes.
The outcomes of interest for the study included any complication related to the use of epidural analgesia (e.g., spinal hemorrhage/hematoma or abscess, spinal ischemia, spinal decompression, or procedures that may have been performed when a complication of epidural analgesia was suspected, such as spinal magnetic resonance imaging [MRI]/computed tomography [CT] scans and transfusion of coagulation factors), as well as the incidence of any postoperative adverse events including in-hospital death, respiratory failure, pneumonia, ileus, pulmonary embolism or deep vein thrombosis, urinary retention, myocardial infarction, and acute renal failure. In addition, hospital length of stay was compared between the groups. All the endpoints were selected a priori based on current literature on use of epidural analgesia for major abdominal surgery (5, 7, 22, 23). In-hospital death was determined directly from a variable present in the database. In-hospital postoperative adverse events were determined from the diagnostic and procedure ICD-9-CM codes.
Baseline characteristics of patients undergoing liver resections with and without epidural analgesia were described using univariate analyses of the weighted NIS data. Weighted analyses on the nonmatched sample were conducted using the SURVEY FREQ, SURVEY REG, and SURVEY MEANS procedures of the SAS software, to account for the NIS survey design. Continuous variables are summarized as means ± standard deviations, except for heavily skewed distributions, which are reported as medians and interquartile ranges. Discrete variables are presented as frequencies and group percentages. Trends in the use of epidural analgesia for liver resection across the study period were assessed with the Cochran-Armitage trend test.
A 1:1 matched cohort of patients receiving or not receiving epidural analgesia was created based on propensity scores derived from a logistic regression model (constructed to estimate the conditional probability for receiving epidural analgesia). The independent variables included in the regression model for propensity scores consisted of demographic characteristics, comorbidity score, type of principal diagnosis, comorbidities such as cirrhosis and chronic viral hepatitis, type of health care insurance, and hospital characteristics. Propensity matching was done using a greedy 8 to 1 digit-matching algorithm technique. Differences in the incidence of postoperative adverse events were assessed between the matched groups using McNemar's tests. Relative risks with 95% confidence intervals were calculated for each outcome. Due to its positively skewed distribution, hospital length of stay was described as medians and interquartile ranges and compared between the matched groups using Wilcoxon signed-rank tests. All statistical tests were two-tailed, and a P value of 0.05 was considered statistically significant. SAS 9.2 software (Cary, NC) was used for all the analyses.
RESULTS
We found 68,028 major liver resections recorded in NIS between 2001 and 2010. This may represent 340,140 such procedures having been performed in the US during that time. The number of liver resections increased from 3937 in 2001 to 9836 in 2010 (Figure 1a). Most liver resections were performed for treatment of cancer: 51.3% for metastatic liver disease and 18.2% for primary carcinomas of the liver or bile ducts. About one-third of liver resections (33.7%) were associated with a principal diagnosis of benign neoplasms or other benign diseases of the liver.
Figure 1.
(a) Trends in number of major liver resections performed in the United States from 2001 to 2010 and (b) trends in proportion of patients undergoing major liver resections with the use of postoperative epidural analgesia. Dotted lines represent 95% confidence limits for the estimated weighted frequency or percentages.
Figure 1b displays the percentage of patients receiving epidural analgesia for major liver resections in the US across the study period. Overall, epidural analgesia was administered in 5.9% (n = 4044) of the patients. In 2001, epidural analgesia was used in 7.2% of the cases, while in 2010 it was used in 6.7% of the cases. However, there was not a statistically significant linear trend towards decreasing use of epidural analgesia across the study period (P for trend = 0.108; Cochran-Armitage trend test).
Table 1 describes the baseline clinical, demographic, and hospital characteristics of patients undergoing major liver resections with and without the use of epidural analgesia in 2001 to 2010 in the nonmatched cohort. Although there were no significant differences in the age distributions, patients receiving epidural analgesia had higher comorbidity scores than those without epidural analgesia. Also, metastatic liver disease was more common among patients receiving epidural analgesia (60.3% vs 51.3%; P < 0.0001). In contrast, the prevalence of hepatic cirrhosis was lower among patients having epidural analgesia (8.0% vs 11.5%, P = 0.006). Hospital characteristics, such as teaching status, bed size, or urban/rural location of the hospital, were not associated with differences in the use of epidural analgesia for liver resection.
Table 1.
Baseline characteristics of patients undergoing major liver resections with and without epidural analgesia, United States, 2001–2010, nonmatched sample
| Epidural analgesia |
|||||
|---|---|---|---|---|---|
| Characteristics | No (n = 63,876) | Yes (n = 4044) | P value | ||
| Age categories (years) | 0.337 | ||||
| 18 to 39 | 7,708 | (12.0%) | 406 | (10.0%) | |
| 40 to 64 | 33,456 | (52.3%) | 2097 | (51.9%) | |
| 65 to 74 | 15,144 | (23.7%) | 1021 | (25.2%) | |
| 75+ | 7,675 | (12.0%) | 520 | (12.9%) | |
| Female sex | 32,865 | (51.4%) | 2068 | (51.1%) | 0.891 |
| Charlson comorbidity index | 0.031 | ||||
| 0 | 7,878 | (12.3%) | 387 | (9.6%) | |
| 1–2 | 10,548 | (16.5%) | 568 | (14.0%) | |
| ≥3 | 45,557 | (71.2%) | 3089 | (76.4%) | |
| Geographic region of hospital | 0.075 | ||||
| Northeast | 17,202 | (26.9%) | 791 | (19.6%) | |
| Midwest | 12,530 | (19.6%) | 1442 | (35.7%) | |
| South | 19,342 | (30.2%) | 1114 | (27.5%) | |
| West | 14,909 | (23.3%) | 697 | (17.2%) | |
| Hospital bed size | 0.498 | ||||
| Small | 3,548 | (5.6%) | 420 | (10.4%) | |
| Medium | 7,257 | (11.4%) | 505 | (12.5%) | |
| Large | 53,077 | (83.1%) | 3119 | (77.1%) | |
| Urban location of hospital | 62,709 | (98.2%) | 3924 | (97.0%) | 0.435 |
| Procedure in teaching hospital | 56,397 | (88.3%) | 3394 | (83.9%) | 0.339 |
| Patient comorbidities | |||||
| Hypertension | 24,934 | (38.9%) | 1576 | (38.9%) | 0.997 |
| Chronic lung disease | 5,881 | (9.2%) | 396 | (9.8%) | 0.584 |
| Heart failure | 1,352 | (2.1%) | 87 | (2.2%) | 0.938 |
| Diabetes mellitus | 9,179 | (14.3%) | 572 | (14.1%) | 0.889 |
| Chronic renal failure | 1,171 | (1.8%) | 76 | (1.9%) | 0.915 |
| Obesity* | 3,406 | (5.3%) | 199 | (4.9%) | 0.706 |
| Chronic hepatitis | 1,160 | (1.8%) | 77 | (1.9%) | 0.851 |
| Alcoholism | 1,319 | (2.1%) | 63 | (1.5%) | 0.305 |
| Cirrhosis | 7,383 | (11.5%) | 324 | (8.0%) | 0.006 |
| Surgical diagnosis | |||||
| Primary carcinoma of liver or bile ducts | 11,634 | (18.2%) | 677 | (16.7%) | 0.354 |
| Metastatic liver disease | 32,830 | (51.3%) | 2438 | (60.3%) | <0.001 |
| Other benign liver disease | 21,577 | (33.7%) | 1218 | (30.1%) | 0.171 |
Obesity was defined as body mass index >30 kg/m2 using International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM) codes 278.00 and 278.01.
None of the patients in either group experienced any complication directly related to the use of epidural catheters (including spinal ischemia, abscess, or hematoma) or any event indicating that any of these complications was suspected or treated (including use of spinal MRIs or CT scans or procedures for decompression of the spinal cord). Univariate analyses of perioperative outcomes in the unmatched sample are displayed in Table 2. The unadjusted rate of perioperative blood transfusions and postoperative atelectasis was higher in the group receiving epidural analgesia. However, the incidence of other adverse events was not statistically different between the groups.
Table 2.
Outcomes in patients undergoing major liver resections with and without epidural analgesia, United States, 2001–2010, unmatched sample
| Epidural analgesia |
|||||
|---|---|---|---|---|---|
| Postoperative outcome | No (n = 63,984) | Yes (n = 4044) | P value | ||
| In-hospital death | 1550 | (2.4%) | 88 | (2.2%) | 0.655 |
| Hemorrhage/hematoma | 2132 | (3.3%) | 115 | (2.8%) | 0.459 |
| Wound dehiscence | 440 | (0.7%) | 13 | (0.3%) | 0.196 |
| Respiratory failure | 2408 | (3.8%) | 186 | (4.6%) | 0.301 |
| Pulmonary embolism or DVT | 1090 | (1.7%) | 64 | (1.6%) | 0.805 |
| Sepsis | 1434 | (2.2%) | 83 | (2.0%) | 0.715 |
| Blood transfusion | 11859 | (18.5%) | 977 | (24.1%) | 0.047 |
| Transfusion of coagulation factors | 4042 | (6.3%) | 249 | (6.1%) | 0.906 |
| Acute myocardial infarction | 410 | (0.6%) | 42 | (1.0%) | 0.330 |
| Acute renal failure | 2243 | (3.5%) | 121 | (2.9%) | 0.555 |
| Cardiac complications | 1607 | (2.5%) | 98 | (2.4%) | 0.890 |
| Ileus | 5996 | (9.4%) | 480 | (11.9%) | 0.055 |
| Atelectasis | 4839 | (7.6%) | 427 | (10.6%) | 0.033 |
| Pneumonia | 1390 | (2.2%) | 116 | (2.9%) | 0.226 |
| Urinary tract infection | 2147 | (3.3%) | 112 | (2.8%) | 0.302 |
| Urinary retention | 1136 | (1.8%) | 48 | (1.2%) | 0.247 |
| Any infection | 1882 | (2.9%) | 132 | (3.2%) | 0.550 |
| LOS (days, median, IQR) | 6 | 5–8 | 6 | 5–8 | <0.001 |
DVT indicates deep vein thrombosis; IQR, interquartile range; LOS, length of stay.
Table 3 describes the baseline characteristics of patients in the propensity-matched sample. A cohort of 802 patients not receiving epidural analgesia and 802 patients receiving epidural analgesia for liver resections, well balanced on baseline characteristics, was derived from the propensity-matching algorithm. The rate of in-hospital mortality was the same in both groups (2.1%). The matched analyses confirmed that patients receiving epidurals were significantly more likely to have transfusion of blood products during the hospitalization (24.3% vs 17.8%; relative risk = 1.36; 95% confidence interval = 1.12 to 1.65; P = 0.001). However, the use of transfusion of coagulation factors (6.2% vs 6.2%, P = 1.000) and the incidence of postoperative hemorrhage or hematomas (2.6% vs 3.4%, P = 0.379) was similar between the groups. In the propensity-matched cohort, the use of epidural analgesia was not associated with differences in the incidence of postoperative respiratory complications (respiratory failure, pneumonia, atelectasis), cardiac complications or myocardial infarction, thrombotic events, acute renal failure, ileus, sepsis, or urinary complications (Table 4). Finally, the length of hospital stay (median, [interquartile range]) was 6 [5–8] days vs 6 [4–8] days for patients with and without epidurals, respectively.
Table 3.
Baseline characteristics of patients undergoing major liver resections with and without epidural analgesia, United States, 2001–2010, propensity-matched sample
| Epidural analgesia |
|||||
|---|---|---|---|---|---|
| Characteristics | No (n = 802) | Yes (n = 802) | P value | ||
| Age categories (years) | 0.947 | ||||
| 18 to 39 | 85 | (10.6%) | 81 | (10.1%) | |
| 40 to 64 | 406 | (50.6%) | 415 | (51.8%) | |
| 65 to 74 | 202 | (25.2%) | 203 | (25.3%) | |
| 75+ | 109 | (13.6%) | 103 | (12.8%) | |
| Female sex | 401 | (50.0%) | 411 | (51.2%) | 0.600 |
| Charlson comorbidity index | 0.981 | ||||
| 0 | 79 | (9.9%) | 77 | (9.6%) | |
| 1–2 | 115 | (14.3%) | 114 | (14.2%) | |
| ≥3 | 608 | (75.8%) | 611 | (76.2%) | |
| Geographic region of hospital | 0.726 | ||||
| Northeast | 136 | (17.0%) | 149 | (18.6%) | |
| Midwest | 289 | (36.0%) | 283 | (35.3%) | |
| South | 224 | (27.9%) | 230 | (28.7%) | |
| West | 153 | (19.1%) | 140 | (17.5%) | |
| Hospital bed size | 0.717 | ||||
| Small | 95 | (11.9%) | 86 | (10.7%) | |
| Medium | 104 | (13.0%) | 100 | (12.5%) | |
| Large | 603 | (75.2%) | 616 | (76.8%) | |
| Urban location of hospital | 783 | (97.6%) | 777 | (96.9%) | 0.359 |
| Procedure in teaching hospital | 665 | (82.9%) | 669 | (83.4%) | 0.789 |
| Patient comorbidities | |||||
| Hypertension | 327 | (40.8%) | 317 | (39.5%) | 0.610 |
| Chronic lung disease | 70 | (8.7%) | 77 | (9.6%) | 0.544 |
| Heart failure | 17 | (2.1%) | 18 | (2.2%) | 0.864 |
| Diabetes mellitus | 104 | (13.0%) | 115 | (14.3%) | 0.423 |
| Chronic renal failure | 25 | (3.1%) | 15 | (1.9%) | 0.109 |
| Obesity | 48 | (5.9%) | 41 | (5.1%) | 0.445 |
| Chronic hepatitis | 18 | (2.2%) | 15 | (1.9%) | 0.597 |
| Alcoholism | 10 | (1.2%) | 13 | (1.6%) | 0.528 |
| Cirrhosis | 72 | (8.9%) | 64 | (7.9%) | 0.473 |
| Surgical diagnosis | |||||
| Primary carcinoma of liver or bile ducts | 125 | (15.6%) | 132 | (16.4%) | 0.633 |
| Metastatic liver disease | 481 | (60.0%) | 484 | (60.3%) | 0.878 |
| Other benign liver disease | 247 | (30.8%) | 240 | (29.9%) | 0.704 |
Table 4.
Outcomes in patients undergoing major liver resections with and without epidural analgesia, United States, 2001–2010, propensity-matched sample
| Epidural analgesia |
||||||
|---|---|---|---|---|---|---|
| Postoperative outcome | No (n = 802) | Yes (n = 802) | Relative risk (95% CI) | P value | ||
| In-hospital death | 17 | (2.1%) | 17 | (2.1%) | 1.00 (0.51–1.94) | 1.000 |
| Hemorrhage/hematoma | 27 | (3.4%) | 21 | (2.6%) | 0.77 (0.44–1.36) | 0.379 |
| Wound dehiscence | 7 | (0.9%) | 3 | (0.4%) | 0.43 (0.11–1.65) | 0.204 |
| Respiratory failure | 32 | (4.0%) | 37 | (4.6%) | 1.15 (0.72–1.83) | 0.538 |
| Pulmonary embolism or DVT | 16 | (2.0%) | 13 | (1.6%) | 0.81 (0.39–1.67) | 0.574 |
| Sepsis | 19 | (2.4%) | 16 | (2.0%) | 0.84 (0.43–1.62) | 0.608 |
| Blood transfusion | 143 | (17.8%) | 195 | (24.3%) | 1.36 (1.12–1.65) | 0.001 |
| Transfusion of coagulation factors | 50 | (6.2%) | 50 | (6.2%) | 1.00 (0.68–1.46) | 1.000 |
| Acute myocardial infarction | 4 | (0.5%) | 8 | (1.0%) | 2.00 (0.60–6.61) | 0.246 |
| Acute renal failure | 23 | (2.9%) | 24 | (3.0%) | 1.04 (0.59–1.83) | 0.882 |
| Cardiac complications | 20 | (2.5%) | 20 | (2.5%) | 1.00 (0.54–1.84) | 1.000 |
| Ileus | 84 | (10.5%) | 94 | (11.7%) | 1.11 (0.84–1.47) | 0.426 |
| Atelectasis | 72 | (9.0%) | 81 | (10.1%) | 1.12 (0.83–1.52) | 0.444 |
| Pneumonia | 19 | (2.4%) | 23 | (2.9%) | 1.21 (0.66–2.20) | 0.531 |
| Urinary tract infection | 21 | (2.6%) | 22 | (2.7%) | 1.05 (0.58–1.89) | 0.877 |
| Urinary retention | 17 | (2.1%) | 10 | (1.2%) | 0.58 (0.27–1.27) | 0.174 |
| Any infection | 29 | (3.6%) | 26 | (3.2%) | 0.89 (0.53–1.50) | 0.680 |
| LOS (days, median, IQR) | 6 | 4–8 | 6 | 5–8 | <0.001 | |
CI indicates confidence interval; DVT, deep vein thrombosis; IQR, interquartile range; LOS, length of stay.
DISCUSSION
This study of a large cohort of patients undergoing major liver resection shows that epidural analgesia is not widely used in this patient population. Despite the increase in the number of states contributing data to the NIS between 2001 and 2010, the sampling methodology of the database has not changed during that period of time. Therefore, our findings suggest that major liver resections are increasing in the US (Figure 1a). Although there was not a significant linear trend in the use of epidural analgesia for major liver resections during the study period, a notable decrease was observed in 2007 (Figure 1b). However, with the data available, we cannot determine if this dip is a true decrease in the use of epidural analgesia or the effect of undercoding the procedure.
The propensity-matching technique allowed for a more robust comparison between the patients who received epidural analgesia and those who did not, because all the observable variables such as demographics, comorbidities, type of facility, and type of surgical procedure were well balanced between the groups. Except for an increased incidence of blood transfusions in the epidural group, the propensity matching analyses revealed similar rates of postoperative complications despite the use of epidural analgesia.
One would assume that epidural analgesia would be used more often in teaching hospitals, particularly in larger hospitals with acute pain services. However, we found that hospital characteristics (e.g., teaching status, size, and location) did not influence the use of epidural analgesia. It is possible that the reports of lack of benefits of epidural analgesia (25, 26) combined with the concerns of potential complications may have resulted in the reduced use of epidural analgesia (27–29). In addition, the use of epidural analgesia may have been influenced by the reports of the high failure rate of epidural analgesia (30). Also, use of epidural analgesia may have been further reduced due to the reports of similar postoperative outcomes with the use of rational multimodal analgesia techniques (31).
There are several observations that are worth noting. The prevalence of hepatic cirrhosis was lower among the patients receiving epidural analgesia, probably due to concerns of coagulopathy and epidural hematoma in patients with cirrhosis. Patients receiving epidural analgesia had higher comorbidity scores and metastatic liver disease. This suggests a preferential use of epidural analgesia in sicker patients, indicating that epidural analgesia was considered an appropriate analgesic technique for pain management in the sicker patients undergoing liver resection. Despite the higher comorbidity burden in the patients receiving epidural analgesia, the incidence of complications, including mortality, was similar. This may suggest that epidural analgesia may offer some protection against postoperative complications in the high-risk population.
Patients receiving epidural analgesia were more likely to receive transfusion during their hospital stay. Although the reasons for this observation are not clear, it is possible that the patients in the epidural group received larger amounts of crystalloids (3), probably due to vasodilation caused by sympathetic blockade from epidural analgesia. The resulting hemodilution, therefore, may have triggered blood transfusion. Also, because the patients in the epidural group were sicker, higher hematocrit levels may have been maintained in this patient group. Of note, the use of coagulation factors and the incidence of hemorrhage and hematoma formation were similar with or without epidural analgesia.
Interestingly, we found that the patients receiving epidural analgesia had a longer hospital stay. Similar observations have been reported in patients undergoing colonic resection (24). It is possible that the delayed discharge was due to the higher comorbidities in the group that received epidural analgesia. Also, we can speculate that the longer hospital stay may be related to unplanned delays in epidural catheter removal because of concerns of epidural hematoma related to inadequate postoperative coagulation (32). However, we could not confirm this observation because the NIS datasets do not provide information on laboratory test results or timing of removal of the epidural catheters.
Although we studied a large nationwide sample, this study has several limitations related to the use of administrative datasets. Retrospective analysis prohibits examination and incorporation of factors other than those provided in the dataset. There is a lack of clinical information, including the details of analgesic regimens used in the nonepidural analgesia population as well as the details regarding the epidural analgesia regimens. In addition, there is no information on the degree of pain relief as well as some outcome measures such as time to ambulation. Unfortunately, the restricted use of epidural analgesia (i.e., limited sample size afforded from this database) limits our ability to assess differences in outcomes, particularly in complications with a very low incidence such as epidural hematoma or abscess formation.
References
- 1.Simons JP, Ng SC, Hill JS, Shah SA, Zhou Z, Tseng JF. In-hospital mortality from liver resection for hepatocellular carcinoma: a simple risk score. Cancer. 2010;116(7):1733–1738. doi: 10.1002/cncr.24904. [DOI] [PubMed] [Google Scholar]
- 2.Mondor ME, Massicotte L, Beaulieu D, Roy JD, Lapointe R, Dagenais M, Roy A. Long-lasting analgesic effects of intraoperative thoracic epidural with bupivacaine for liver resection. Reg Anesth Pain Med. 2010;35(1):51–56. doi: 10.1097/AAP.0b013e3181c6f8f2. [DOI] [PubMed] [Google Scholar]
- 3.Tzimas P, Prout J, Papadopoulos G, Mallett SV. Epidural anaesthesia and analgesia for liver resection. Anaesthesia. 2013;68(6):628–635. doi: 10.1111/anae.12191. [DOI] [PubMed] [Google Scholar]
- 4.Joshi GP. Multimodal analgesia techniques and postoperative rehabilitation. Anesthesiol Clin North America. 2005;23(1):185–202. doi: 10.1016/j.atc.2004.11.010. [DOI] [PubMed] [Google Scholar]
- 5.Amini A, Patanwala AE, Maegawa FB, Skrepnek GH, Jie T, Gruessner RW, Ong ES. Effect of epidural analgesia on postoperative complications following pancreaticoduodenectomy. Am J Surg. 2012;204(6):1000–1004. doi: 10.1016/j.amjsurg.2012.05.022. discussion 1004–1006. [DOI] [PubMed] [Google Scholar]
- 6.Freise H, Van Aken HK. Risks and benefits of thoracic epidural anaesthesia. Br J Anaesth. 2011;107(6):859–868. doi: 10.1093/bja/aer339. [DOI] [PubMed] [Google Scholar]
- 7.Pöpping DM, Elia N, Marret E, Remy C, Tramèr MR. Protective effects of epidural analgesia on pulmonary complications after abdominal and thoracic surgery: a meta-analysis. Arch Surg. 2008;143(10):990–999. doi: 10.1001/archsurg.143.10.990. discussion 1000. [DOI] [PubMed] [Google Scholar]
- 8.van Lier F, van der Geest PJ, Hoeks SE, van Gestel YR, Hol JW, Sin DD, Stolker RJ, Poldermans D. Epidural analgesia is associated with improved health outcomes of surgical patients with chronic obstructive pulmonary disease. Anesthesiology. 2011;115(2):315–321. doi: 10.1097/ALN.0b013e318224cc5c. [DOI] [PubMed] [Google Scholar]
- 9.Zhu Z, Wang C, Xu C, Cai Q. Influence of patient-controlled epidural analgesia versus patient-controlled intravenous analgesia on postoperative pain control and recovery after gastrectomy for gastric cancer: a prospective randomized trial. Gastric Cancer. 2013;16(2):193–200. doi: 10.1007/s10120-012-0168-z. [DOI] [PubMed] [Google Scholar]
- 10.Ali M, Winter DC, Hanly AM, O’Hagan C, Keaveny J, Broe P. Prospective, randomized, controlled trial of thoracic epidural or patient-controlled opiate analgesia on perioperative quality of life. Br J Anaesth. 2010;104(3):292–297. doi: 10.1093/bja/aeq006. [DOI] [PubMed] [Google Scholar]
- 11.Nishimori M, Low JH, Zheng H, Ballantyne JC. Epidural pain relief versus systemic opioid-based pain relief for abdominal aortic surgery. Cochrane Database Syst Rev. 2012;7:CD005059. doi: 10.1002/14651858.CD005059.pub3. [DOI] [PubMed] [Google Scholar]
- 12.Lin DX, Li X, Ye QW, Lin F, Li LL, Zhang QY. Implementation of a fast-track clinical pathway decreases postoperative length of stay and hospital charges for liver resection. Cell Biochem Biophys. 2011;61(2):413–419. doi: 10.1007/s12013-011-9203-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.van Dam RM, Hendry PO, Coolsen MM, Bemelmans MH, Lassen K, Revhaug A, Fearon KC, Garden OJ, Dejong CH Enhanced Recovery After Surgery (ERAS) Group. Initial experience with a multimodal enhanced recovery programme in patients undergoing liver resection. Br J Surg. 2008;95(8):969–975. doi: 10.1002/bjs.6227. [DOI] [PubMed] [Google Scholar]
- 14.Hase T, Takita K, Hashimoto T, Morimoto Y. Coagulation profiles following donor hepatectomy and implications for the risk of epidural hematoma associated with epidural anesthesia. Masui. 2011;60(7):840–845. [PubMed] [Google Scholar]
- 15.Stamenkovic DM, Jankovic ZB, Toogood GJ, Lodge JP, Bellamy MC. Epidural analgesia and liver resection: postoperative coagulation disorders and epidural catheter removal. Minerva Anestesiol. 2011;77(7):671–679. [PubMed] [Google Scholar]
- 16.Cook TM, Counsell D, Wildsmith JA Royal College of Anaesthetists Third National Audit Project. Major complications of central neuraxial block: report on the Third National Audit Project of the Royal College of Anaesthetists. Br J Anaesth. 2009;102(2):179–190. doi: 10.1093/bja/aen360. [DOI] [PubMed] [Google Scholar]
- 17.Rawal N. Epidural technique for postoperative pain: gold standard no more? Reg Anesth Pain Med. 2012;37(3):310–317. doi: 10.1097/AAP.0b013e31825735c6. [DOI] [PubMed] [Google Scholar]
- 18.Wildsmith JA. Continuous thoracic epidural block for surgery: gold standard or debased currency? Br J Anaesth. 2012;109(1):9–12. doi: 10.1093/bja/aes177. [DOI] [PubMed] [Google Scholar]
- 19.Bauer AJ, Boeckxstaens GE. Mechanisms of postoperative ileus. Neurogastroenterol. Motil. 2004;16(Suppl 2):54–60. doi: 10.1111/j.1743-3150.2004.00558.x. [DOI] [PubMed] [Google Scholar]
- 20.Healthcare Cost and Utilization Project (HCUP) Comorbidity Software, Version 3.7. Available at http://www.hcup-us.ahrq.gov/toolssoftware/comorbidity/comorbidity.jsp; accessed December 15, 2012.
- 21.Deyo RA, Cherkin DC, Ciol MA. Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. J Clin Epidemiol. 1992;45(6):613–619. doi: 10.1016/0895-4356(92)90133-8. [DOI] [PubMed] [Google Scholar]
- 22.Revie EJ, McKeown DW, Wilson JA, Garden OJ, Wigmore SJ. Randomized clinical trial of local infiltration plus patient-controlled opiate analgesia vs. epidural analgesia following liver resection surgery. HPB (Oxford) 2012;14(9):611–618. doi: 10.1111/j.1477-2574.2012.00490.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Wu CL, Rowlingson AJ, Herbert R, Richman JM, Andrews RA, Fleisher LA. Correlation of postoperative epidural analgesia on morbidity and mortality after colectomy in Medicare patients. J Clin Anesth. 2006;18(8):594–599. doi: 10.1016/j.jclinane.2006.03.020. [DOI] [PubMed] [Google Scholar]
- 24.Halabi WJ, Jafari MD, Nguyen VQ, Carmichael JC, Mills S, Stamos MJ, Pigazzi A. A nationwide analysis of the use and outcomes of epidural analgesia in open colorectal surgery. J Gastrointest Surg. 2013;17(6):1130–1137. doi: 10.1007/s11605-013-2195-4. [DOI] [PubMed] [Google Scholar]
- 25.Leslie K, Myles P, Devereaux P, Williamson E, Rao-Melancini P, Forbes A, Xu S, Foex P, Pogue J, Arrieta M, Bryson G, Paul J, Paech M, Merchant R, Choi P, Badner N, Peyton P, Sear J, Yang H. Neuraxial block, death and serious cardiovascular morbidity in the POISE trial. Br J Anaesth. 2013;111(3):382–390. doi: 10.1093/bja/aet120. [DOI] [PubMed] [Google Scholar]
- 26.Rigg JR, Jamrozik K, Myles PS, Silbert BS, Peyton PJ, Parsons RW, Collins KS. MASTER Anaethesia Trial Study Group. Epidural anaesthesia and analgesia and outcome of major surgery: a randomised trial. Lancet. 2002;359(9314):1276–1282. doi: 10.1016/S0140-6736(02)08266-1. [DOI] [PubMed] [Google Scholar]
- 27.Cameron CM, Scott DA, McDonald WM, Davies MJ. A review of neuraxial epidural morbidity: experience of more than 8,000 cases at a single teaching hospital. Anesthesiology. 2007;106(5):997–1002. doi: 10.1097/01.anes.0000265160.32309.10. [DOI] [PubMed] [Google Scholar]
- 28.Christie IW, McCabe S. Major complications of epidural analgesia after surgery: results of a six-year survey. Anaesthesia. 2007;62(4):335–341. doi: 10.1111/j.1365-2044.2007.04992.x. [DOI] [PubMed] [Google Scholar]
- 29.Power GE, Warden B, Cooke K. Changing patterns in the acute pain service: epidural versus patient-controlled analgesia. Anaesth Intensive Care. 2005;33(4):501–505. doi: 10.1177/0310057X0503300413. [DOI] [PubMed] [Google Scholar]
- 30.McLeod G, Davies H, Munnoch N, Bannister J, MacRae W. Postoperative pain relief using thoracic epidural analgesia: outstanding success and disappointing failures. Anaesthesia. 2001;56(1):75–81. doi: 10.1046/j.1365-2044.2001.01763-7.x. [DOI] [PubMed] [Google Scholar]
- 31.Joshi GP, Bonnet F, Kehlet H. PROSPECT collaboration. Evidence-based postoperative pain management after laparoscopic colorectal surgery. Colorectal Dis. 2013;15(2):146–155. doi: 10.1111/j.1463-1318.2012.03062.x. [DOI] [PubMed] [Google Scholar]
- 32.Revie EJ, Massie LJ, McNally SJ, McKeown DW, Garden OJ, Wigmore SJ. Effectiveness of epidural analgesia following open liver resection. HPB (Oxford) 2011;13(3):206–211. doi: 10.1111/j.1477-2574.2010.00274.x. [DOI] [PMC free article] [PubMed] [Google Scholar]