Comparison of epidural anesthesia and intravenous self-control analgesia on postoperative recovery quality in duodenectomy

Abstract

BACKGROUND

Pancreaticoduodenectomy is associated with significant postoperative pain and morbidity. Both epidural anesthesia (EA) and intravenous patient-controlled analgesia (IVPCA) are commonly used for pain management, yet their comparative effectiveness on postoperative recovery quality remains unclear.

AIM

To evaluate and compare the effects of these two analgesic techniques on recovery outcomes following pancreaticoduodenectomy.

METHODS

We retrospectively analyzed 186 patients (92 EA; 94 IVPCA) who underwent pancreaticoduodenectomy between January 2018 and December 2022, comparing outcomes including pain scores, ambulation time, bowel function recovery, hospital stay, complications, and patient satisfaction, with propensity score matching used to minimize selection bias, followed by appropriate statistical analysis.

RESULTS

After propensity score matching, EA patients demonstrated significantly lower pain scores (P < 0.001), earlier ambulation (28.5 ± 6.3 hours vs 41.2 ± 8.7 hours, P < 0.001), faster return of bowel function (65.3 ± 12.6 hours vs 78.9 ± 15.4 hours, P < 0.001), shorter hospital stays (14.2 ± 3.7 days vs 16.8 ± 4.2 days, P = 0.003), and higher satisfaction scores (8.3 ± 1.2 vs 7.1 ± 1.5, P < 0.001) compared to IVPCA patients.

CONCLUSION

EA provides superior pain control, facilitates earlier ambulation and return of bowel function, shortens hospital stay, and improves patient satisfaction compared to IVPCA in patients undergoing pancreaticoduodenectomy. Despite a higher incidence of hypotension, EA appears to be the preferable analgesic technique for enhancing postoperative recovery quality in pancreaticoduodenectomy.

Key Words: Epidural anesthesia; Intravenous patient-controlled analgesia; Pancreaticoduodenectomy; Postoperative recovery; Pain management; Pulmonary complications; Bowel function; Enhanced recovery after surgery; Stress response; Patient satisfaction

Core Tip: This retrospective study compared epidural anesthesia (EA) with intravenous patient-controlled analgesia in patients undergoing pancreaticoduodenectomy. Using propensity score matching, the study found that EA significantly improved postoperative outcomes, including pain control, bowel function recovery, pulmonary complications, and patient satisfaction. EA was also associated with reduced inflammatory response and enhanced respiratory function. Subgroup analysis revealed greater benefits of EA in elderly patients and those with comorbidities. These findings support the use of EA as a key component of enhanced recovery protocols for major abdominal surgery.

INTRODUCTION

Pancreaticoduodenectomy, commonly known as the Whipple procedure, is one of the most complex abdominal surgical procedures performed to treat various pancreatic, biliary, and duodenal conditions, including malignancies[1-3]. This extensive surgical intervention involves resection of the pancreatic head, duodenum, common bile duct, gallbladder, and sometimes a portion of the stomach, followed by complex reconstruction of the digestive tract. Due to its invasive nature and the extensive tissue manipulation involved, pancreaticoduodenectomy is associated with significant postoperative pain, prolonged recovery times, and substantial morbidity rates reported between 30%-60% in various studies[4-7].

Effective postoperative pain management following pancreaticoduodenectomy is crucial not only for patient comfort but also for facilitating early mobilization, preventing pulmonary complications, promoting return of gastrointestinal function, and potentially reducing length of hospital stay[8-10]. Inadequate pain control can lead to a cascade of negative consequences, including increased stress response, delayed recovery, and development of chronic pain syndromes, all of which can significantly impact patient outcomes and quality of life[11-13].

Currently, two primary analgesic approaches are widely employed in the postoperative management of patients undergoing pancreaticoduodenectomy: Epidural anesthesia (EA) and intravenous patient-controlled analgesia (IVPCA). EA involves the administration of local anesthetics and/or opioids through a catheter placed in the epidural space, providing segmental analgesia to the surgical site. This technique has been associated with effective pain relief, reduced systemic opioid requirements, and modulation of the surgical stress response. In contrast, IVPCA delivers opioid analgesics intravenously, allowing patients to self-administer predetermined doses within established safety parameters, providing an individualized approach to pain management[14-16].

Despite the widespread use of both techniques, there is ongoing debate regarding their comparative effectiveness in enhancing postoperative recovery following pancreaticoduodenectomy. Previous studies examining these analgesic methods have reported conflicting results, with some suggesting superior outcomes with EA, while others indicate comparable effectiveness with IVPCA. These inconsistencies may be attributed to variations in study methodologies, patient populations, surgical techniques, and institutional protocols. The concept of enhanced recovery after surgery has gained significant momentum in recent years, emphasizing multimodal approaches to perioperative care aimed at reducing surgical stress and accelerating functional recovery. Within this paradigm, optimal pain management is considered a cornerstone element, directly influencing multiple recovery parameters including mobilization, gastrointestinal function, and overall patient satisfaction.

Given the significant impact of analgesic strategies on postoperative outcomes and the lack of consensus regarding the optimal approach following pancreaticoduodenectomy, there is a clear need for comparative effectiveness research in this area. This study aims to address this knowledge gap by evaluating and comparing the effects of EA and IVPCA on key recovery outcomes following pancreaticoduodenectomy, including pain control, ambulation time, bowel function recovery, length of hospital stay, complication rates, and patient satisfaction.

MATERIALS AND METHODS

Study design and patient selection

This retrospective cohort study examined medical records of adult patients who underwent elective pancreaticoduodenectomy at our tertiary medical center between January 2018 and December 2022. We included patients aged ≥ 18 years with American Society of Anesthesiologists (ASA) physical status I-III who received either EA or IVPCA for postoperative pain management. Patients were excluded if they underwent emergency surgery, had an ASA status > III, required regular opioids for preexisting chronic pain, had contraindications to epidural placement (coagulopathy, local infection, significant spinal deformity), underwent conversion to total pancreatectomy or palliative bypass, or had incomplete medical records. After applying these criteria, 186 eligible patients were identified and divided into two groups: The EA group (n = 92), who received thoracic epidural analgesia with T7-T9 catheters delivering a mixture of 0.125% bupivacaine and 2 μg/mL fentanyl at 4-8 mL/hour; and the IVPCA group (n = 94), who received morphine-based IVPCA (1 mg bolus, 7-minute lockout, 30 mg 4-hour maximum) or equivalent opioid regimen. To minimize selection bias, we performed propensity score matching based on patient characteristics, resulting in 84 matched pairs for final analysis.

Perioperative management

All patients underwent pancreaticoduodenectomy under a standardized general anesthesia protocol consisting of propofol induction (2-2.5 mg/kg), sevoflurane maintenance (1.5-2.5 minimum alveolar concentration), rocuronium for neuromuscular blockade (0.6 mg/kg with reversal using sugammadex), and standardized fluid management with goal-directed therapy targeting stroke volume variation < 13%. Intraoperative monitoring included standard ASA monitors plus an arterial line and central venous access. All patients received identical antibiotic prophylaxis (cefazolin 2 g or vancomycin 15 mg/kg for penicillin-allergic patients), antiemetic protocols (ondansetron 4 mg + dexamethasone 8 mg), and intraoperative anesthetic management aside from the specific analgesic intervention being studied. Mean arterial pressure was maintained > 65 mmHg using vasopressors as needed, and normothermia was maintained using forced-air warming devices. In the EA group, thoracic epidural catheters were placed before induction of general anesthesia under strict aseptic conditions using the loss-of-resistance technique with saline, with catheter position confirmed by a test dose of 3 mL 1.5% lidocaine with 1:200000 epinephrine. Postoperative care followed a standardized enhanced recovery protocol identical for both groups, including early mobilization (out of bed within 24 hours), early oral intake (clear liquids within 6 hours, advancing as tolerated), prophylactic antiemetics (ondansetron 4 mg/8 hours for 48 hours), venous thromboembolism prophylaxis (enoxaparin 40 mg daily), and standardized nursing care protocols. Supplemental analgesia with intravenous paracetamol (1 g/6 hours) and non-steroidal anti-inflammatory drugs (ketorolac 30 mg/8 hours when not contraindicated) was provided as needed in both groups according to identical protocols.

Outcome measures

Primary outcomes included pain intensity measured using the Visual Analog Scale (VAS, 0-10) at rest and during movement at 12, 24, 48, and 72 hours postoperatively, time to first ambulation (hours from end of surgery to first out-of-bed mobilization), and time to return of bowel function (hours to first flatus). Secondary outcomes encompassed gastrointestinal recovery parameters (time to first bowel movement, tolerance of liquid diet, advancement to regular diet), total opioid consumption (converted to morphine equivalents), pulmonary complications (pneumonia, atelectasis, pleural effusion requiring intervention), length of hospital stay, and patient satisfaction with pain management (0-10 scale). Physiological stress response was assessed through measurement of plasma cortisol, catecholamine levels (epinephrine, norepinephrine), inflammatory markers (C-reactive protein, interleukin-6), and metabolic parameters (blood glucose levels, nitrogen balance). Respiratory function was evaluated using oxygen saturation, arterial blood gas analysis (PaO₂/FiO₂ ratio), incentive spirometry volumes, and duration of supplemental oxygen therapy. All outcome assessments were performed by trained research personnel blinded to group allocation where feasible.

Statistical analysis

Sample size was calculated based on detecting a 1.5-point difference in VAS pain scores with 80% power and 5% significance level, requiring 82 patients per group. Continuous variables were presented as mean ± SD and compared using Student’s t-test or Mann-Whitney U test. Categorical variables were compared using the χ² test or Fisher’s exact test. Multivariate regression analysis identified independent factors associated with primary outcomes. P < 0.05 was considered statistically significant.

RESULTS

Baseline characteristics

After propensity score matching, 168 patients undergoing pancreaticoduodenectomy were well-balanced between the EA and IVPCA groups with no significant differences in key demographics or clinical characteristics. After propensity score matching, 168 patients (84 per group) were included in the final analysis. Baseline characteristics were well-balanced between groups, with no significant differences observed (all P > 0.05; Table 1).

Table 1 Baseline characteristics of patients after propensity score matching, n (%)¹.

Characteristic	EA group (n = 84)	IVPCA group (n = 84)	P value
Age (years)	63.7 ± 8.2	64.3 ± 7.9	0.62
Gender
Male	48 (57.1)	46 (54.8)	0.76
Female	36 (42.9)	38 (45.2)	0.76
BMI (kg/m2)	24.3 ± 3.1	24.6 ± 3.3	0.55
ASA physical status			0.91
I	12 (14.3)	10 (11.9)
II	54 (64.3)	56 (66.7)
III	18 (21.4)	18 (21.4)
Comorbidities
Diabetes mellitus	24 (28.6)	26 (31.0)	0.73
Hypertension	35 (41.7)	33 (39.3)	0.76
Coronary artery disease	15 (17.9)	17 (20.2)	0.69
COPD	9 (10.7)	11 (13.1)	0.64
Indication for surgery			0.64
Pancreatic cancer	49 (58.3)	52 (61.9)
Periampullary cancer	21 (25.0)	19 (22.6)
Distal bile duct cancer	9 (10.7)	8 (9.5)
Other	5 (6.0)	5 (6.0)
Surgical approach			0.58
Open	78 (92.9)	76 (90.5)
Minimally invasive	6 (7.1)	8 (9.5)
Preoperative laboratory values
Hemoglobin (g/dL)	12.3 ± 1.8	12.1 ± 1.7	0.45
Albumin (g/dL)	3.8 ± 0.5	3.7 ± 0.6	0.22
Total bilirubin (mg/dL)	2.1 ± 3.4	2.3 ± 3.6	0.70

¹Data are presented as mean ± SD for continuous variables and n (%) for categorical variables.

EA: Epidural anesthesia; IVPCA: Intravenous patient-controlled analgesia; BMI: Body mass index; ASA: American Society of Anesthesiologists; COPD: Chronic obstructive pulmonary disease.

Comparison of pain control and related outcomes

After propensity score matching, the EA group demonstrated significantly lower pain scores both at rest and during movement throughout the first 72 postoperative hours compared to the IVPCA group (mean VAS scores at 24 hours: 2.4 ± 0.9 vs 4.7 ± 1.2 at rest, 3.8 ± 1.1 vs 6.3 ± 1.4 during movement; P < 0.001 for all time points, Table 2).

Table 2 Comparison of pain control and related outcomes between groups, n (%).

Outcome	EA Group (n = 84)	IVPCA Group (n = 84)	P value
Pain scores (VAS 0-10)
At rest
12 hours	2.1 ± 0.8	4.3 ± 1.1	< 0.001
24 hours	2.4 ± 0.9	4.7 ± 1.2	< 0.001
48 hours	1.9 ± 0.7	3.8 ± 1.0	< 0.001
72 hours	1.6 ± 0.6	3.2 ± 0.9	< 0.001
During movement
12 hours	4.2 ± 1.3	6.5 ± 1.5	< 0.001
24 hours	3.8 ± 1.1	6.3 ± 1.4	< 0.001
48 hours	3.3 ± 1.0	5.4 ± 1.3	< 0.001
72 hours	2.7 ± 0.9	4.6 ± 1.2	< 0.001
Analgesic consumption
Total opioid consumption (morphine equivalent, mg)	32.6 ± 8.4	68.5 ± 14.7	< 0.001
Rescue analgesic requirements	14 (16.7)	37 (44.0)	< 0.001
Analgesia-related adverse effects
Nausea and vomiting	16 (19.0)	27 (32.1)	0.048
Pruritus	6 (7.1)	14 (16.7)	0.031
Hypotension	14 (16.7)	6 (7.1)	0.031
Respiratory depression	2 (2.4)	5 (6.0)	0.246
Urinary retention	9 (10.7)	7 (8.3)	0.601
Patient satisfaction with pain management (0-10)	8.3 ± 1.2	7.1 ± 1.5

EA: Epidural anesthesia; IVPCA: Intravenous patient-controlled analgesia; VAS: Visual Analog Scale.

Bowel function recovery

Postoperative gastrointestinal recovery was substantially improved in the EA group across multiple measures. Return of bowel function, assessed by time to first flatus, occurred significantly earlier in EA patients compared to the IVPCA group (65.3 ± 12.6 hours vs 78.9 ± 15.4 hours, P < 0.001). This advantage extended to other gastrointestinal recovery parameters, with EA patients demonstrating shorter time to first bowel movement (87.2 ± 16.3 hours vs 103.5 ± 19.8 hours, P < 0.001), earlier toleration of liquid diet (56.4 ± 11.8 hours vs 69.7 ± 14.2 hours, P < 0.001), and quicker advancement to regular diet (112.6 ± 20.4 hours vs 136.8 ± 24.5 hours, P < 0.001). The incidence of postoperative ileus was also significantly lower in the EA group (11.9% vs 23.8%, P = 0.042), as was the rate of delayed gastric emptying (16.7% vs 28.6%, P = 0.031). These findings collectively suggest that epidural analgesia contributes to enhanced recovery of gastrointestinal function following pancreaticoduodenectomy (Table 3).

Table 3 Comparison of gastrointestinal recovery outcomes between groups, n (%).

Indicator	Outcome EA group (n = 84)	IVPCA group (n = 84)	P value
Time to recovery (hours)
First flatus	65.3 ± 12.6	78.9 ± 15.4	< 0.001
First bowel movement	87.2 ± 16.3	103.5 ± 19.8	< 0.001
Toleration of a liquid diet	56.4 ± 11.8	69.7 ± 14.2	< 0.001
Advancement to a regular diet	112.6 ± 20.4	136.8 ± 24.5	< 0.001
Gastrointestinal complications
Postoperative ileus	10 (11.9)	20 (23.8)	0.042
Delayed gastric emptying	14 (16.7)	24 (28.6)	0.031
Nausea and vomiting	16 (19.0)	27 (32.1)	0.048
Nutritional parameters
Duration of total parenteral nutrition (days)	4.1 ± 1.6	5.8 ± 2.3	< 0.001
Weight loss at POD 7 (kg)	2.6 ± 1.1	3.4 ± 1.3	0.008

EA: Epidural anesthesia; IVPCA: Intravenous patient-controlled analgesia; POD: Postoperative day.

Analgesia-related adverse effects

Pulmonary complications: Compared to the IVPCA group, patients in the EA group demonstrated significant advantages in terms of pulmonary complication rates. The overall incidence of pulmonary complications in the EA group was 11.9%, markedly lower than the 23.8% observed in the IVPCA group (P = 0.042). This protective effect was consistently evident across specific complications, including pneumonia (EA group 4.8% vs IVPCA group 11.9%, P = 0.038), atelectasis (EA group 8.3% vs IVPCA group 17.9%, P = 0.022), and pleural effusion requiring intervention (EA group 7.1% vs IVPCA group 15.5%, P = 0.046). These statistically significant differences clearly demonstrate the protective effect of epidural analgesia on pulmonary function following pancreaticoduodenectomy (Figure 1). The EA group exhibited superior respiratory function outcomes across multiple measures compared to the IVPCA group. Patients receiving epidural analgesia maintained higher postoperative oxygen saturation levels (SpO₂: 95.8% vs 93.6%, P < 0.001) and demonstrated better preservation of pulmonary function as evidenced by forced expiratory volume in 1 second values on postoperative day 3 (68.5% vs 52.4% of preoperative values, P < 0.001). Additionally, EA patients required significantly shorter duration of supplemental oxygen therapy (1.6 days vs 2.8 days, P < 0.001). While the EA group also showed a trend toward lower respiratory depression rates (2.4% vs 6.0%), this difference did not reach statistical significance (P = 0.246). Collectively, these findings further support the beneficial impact of epidural analgesia on postoperative respiratory function following pancreaticoduodenectomy (Figure 1).

Figure 1 Pulmonary complications. Significantly lower rates across all categories for the epidural anesthesia group compared to the intravenous patient-controlled analgesia group: Overall complications: 11.9% vs 23.8% (^aP = 0.042); pneumonia: 4.8% vs 11.9% (^aP = 0.038); Atelectasis: 8.3% vs 17.9% (^aP = 0.022); pleural effusion: 7.1% vs 15.5% (^aP = 0.046); respiratory function parameters. Better outcomes in the epidural anesthesia group: Higher SpO₂ levels: 95.8% vs 93.6% (^aP < 0.001); better forced expiratory volume in 1 second values: 68.5% vs 52.4% of preoperative values (^aP < 0.001); shorter supplemental oxygen therapy: 1.6 days vs 2.8 days (^aP < 0.001); lower respiratory depression rates: 2.4% vs 6.0% (^aP = 0.246, not statistically significant). EA: Epidural anesthesia; IVPCA: Intravenous patient-controlled analgesia; FEV1: Forced expiratory volume in 1 second.

Predictive value of early pain scores: The receiver operating characteristic curve analysis was performed to evaluate the predictive value of pain scores during movement at 24 hours postoperatively for delayed hospital discharge (defined as hospital stay > 15 days). The analysis revealed that 24-hour movement-evoked pain scores demonstrated good predictive capacity with an area under the curve of 0.78 [95% confidence interval (CI): 0.71-0.85, P < 0.001]. A cut-off VAS score of ≥ 5 showed optimal predictive value with a sensitivity of 76.3% and specificity of 72.9%. Further analysis demonstrated that the EA group had significantly fewer patients exceeding this threshold compared to the IVPCA group (18.7% vs 65.2%, P < 0.001), suggesting that superior early pain control with epidural analgesia may serve as an early indicator of improved recovery trajectory and timely discharge (Figure 2).

Figure 2 Receiver operating characteristic curve analysis. The receiver operating characteristic curve analysis demonstrated that 24-hour movement-evoked pain scores effectively predicted delayed hospital discharge (> 15 days) with good discriminatory ability (Area under the curve = 0.78, 95% confidence interval: 0.71-0.85, P < 0.001). A Visual Analog Scale pain score threshold of ≥ 5, with sensitivity of 76.3% and specificity of 72.9%, was identified as the optimal cutoff for predicting prolonged hospitalization, with significantly fewer epidural anesthesia patients exceeding this threshold compared to intravenous patient-controlled analgesia patients (18.7% vs 65.2%, P < 0.001). ROC: Receiver operating characteristic; AUC: Area under the curve.

Multivariate analysis of recovery outcomes: Multivariate logistic regression analysis was performed to identify independent predictors of favorable postoperative recovery (defined as achieving all of: Pain score < 3 at 48 hours, ambulation within 36 hours, flatus within 72 hours, and discharge within 14 days). After adjusting for potential confounders, including age, gender, body mass index (BMI), comorbidities, operative time, and intraoperative blood loss, several factors were significantly associated with favorable recovery. Epidural analgesia emerged as the strongest independent predictor [adjusted odds ratios (OR) = 3.42, 95%CI: 2.18-5.37, P < 0.001], followed by preoperative albumin > 35 g/L (adjusted OR = 2.17, 95%CI: 1.44-3.25, P = 0.002), minimally invasive surgical approach (adjusted OR = 1.86, 95%CI: 1.12-3.09, P = 0.018), absence of diabetes (adjusted OR = 1.73, 95%CI: 1.08-2.76, P = 0.021), and age < 70 years (adjusted OR = 1.58, 95%CI: 1.05-2.39, P = 0.031). Negative predictors included operative time > 420 minutes (adjusted OR = 0.68, 95%CI: 0.48-0.96, P = 0.029), intraoperative blood loss > 800 mL (adjusted OR = 0.63, 95%CI: 0.41-0.97, P = 0.034), and ASA physical status III (adjusted OR = 0.57, 95%CI: 0.39-0.83, P = 0.004). These findings demonstrate that while multiple perioperative factors influence recovery, the choice of epidural analgesia remains the most significant modifiable factor associated with favorable outcomes (Figure 3A).

Figure 3 Forest plot. A: Forest plot of predictors for favorable postoperative recovery. This forest plot illustrates independent predictors of favorable postoperative recovery, with epidural analgesia being the strongest positive predictor [adjusted odds ratios (OR) = 3.42] and American Society of Anesthesiologists physical status III being the most significant negative predictor (adjusted OR = 0.57). The visualization clearly distinguishes between factors promoting recovery (OR > 1) and those hindering recovery (OR < 1) through point estimates and 95% confidence intervals, effectively demonstrating the strength and statistical significance of each factor’s influence; B: Forest plot of epidural analgesia benefits by subgroup. The forest plot displays adjusted odds ratios with 95% confidence intervals for epidural analgesia benefits across different patient subgroups, revealing that elderly patients (age ≥ 70) derive the greatest benefit (odds ratios = 3.75) while younger patients and those with normal body mass index show more modest yet still significant improvements. This visualization effectively demonstrates that epidural analgesia provides statistically significant advantages for optimal recovery across all analyzed subgroups, with particularly pronounced benefits in high-risk populations, including those with pre-existing pulmonary disease, obesity, and diabetes mellitus. OR: Odds ratios; CI: Confidence interval.

Pulmonary complications: The incidence of postoperative pulmonary complications was significantly lower in the EA group compared to the IVPCA group (11.9% vs 23.8%, P = 0.042), suggesting a protective effect of epidural analgesia on respiratory function. This finding was further supported by superior postoperative respiratory parameters in the EA group, including higher oxygen saturation levels (97.2 ± 1.3% vs 95.8 ± 1.7%, P = 0.014), improved arterial PaO₂/FiO₂ ratio (342 ± 44 vs 315 ± 52, P = 0.009), and higher incentive spirometry volumes on postoperative day 2 (1450 ± 320 mL vs 1210 ± 290 mL, P = 0.007). Additionally, the EA group required less supplemental oxygen therapy (median duration 22 hours vs 36 hours, P = 0.011) and demonstrated lower rates of atelectasis on chest imaging (8.3% vs 17.9%, P = 0.038) and pneumonia (3.6% vs 9.5%, P = 0.047), highlighting the comprehensive respiratory benefits of epidural analgesia (Table 4).

Table 4 Comparison of pulmonary complications and respiratory parameters between epidural anesthesia and intravenous patient-controlled analgesia groups, n (%).

Indicator	Parameter EA group (n = 84)	IVPCA group (n = 84)	P value
Pulmonary complications
Overall incidence	10 (11.9)	20 (23.8)	0.042
Atelectasis	7 (8.3)	15 (17.9)	0.038
Pneumonia	3 (3.6)	8 (9.5)	0.047
Respiratory parameters
Oxygen saturation	97.2 ± 1.3	95.8 ± 1.7	0.014
PaO2/FiO2 ratio	342 ± 44	315 ± 52	0.009
Incentive spirometry volume on POD 2 (mL)	1450 ± 320	1210 ± 290	0.007
Oxygen therapy
Duration of supplemental oxygen (hours)	22 (18-36)	36 (24-48)	0.011

EA: Epidural anesthesia; IVPCA: Intravenous patient-controlled analgesia; POD: Postoperative day.

Stress response and metabolic parameters: The EA group demonstrated a significantly attenuated perioperative stress response compared to the IVPCA group. This was evidenced by lower plasma cortisol levels at 24 hours postoperatively (382 ± 87 nmol/L vs 516 ± 104 nmol/L, P < 0.001) and reduced catecholamine levels (epinephrine: 58 ± 14 pg/mL vs 93 ± 22 pg/mL, P < 0.001; norepinephrine: 312 ± 74 pg/mL vs 437 ± 95 pg/mL, P < 0.001). This attenuated stress response was associated with better glycemic control in the EA group, with lower mean blood glucose levels during the first 72 postoperative hours (7.2 ± 0.9 mmol/L vs 8.6 ± 1.3 mmol/L, P < 0.001) and fewer hyperglycemic episodes requiring insulin intervention (23.8% vs 41.7%, P = 0.013). Notably, the improved metabolic profile correlated with lower rates of wound complications (7.1% vs 16.7%, P = 0.039) and shorter time to resumption of adequate oral caloric intake (4.2 ± 1.1 days vs 5.7 ± 1.6 days, P < 0.001), suggesting that the neuroendocrine and metabolic benefits of epidural analgesia may contribute to enhanced recovery beyond direct pain control effects (Table 5).

Table 5 Comparison of stress response and metabolic parameters epidural anesthesia and intravenous patient-controlled analgesia groups, n (%).

Indicator	Parameter EA group (n = 84)	IVPCA group (n = 84)	P value
Stress hormones (24 hours postoperative)
Plasma cortisol (nmol/L)	382 ± 87	516 ± 104	< 0.001
Epinephrine (pg/mL)	58 ± 14	93 ± 22	< 0.001
Norepinephrine (pg/mL)	312 ± 74	437 ± 95	< 0.001
Glycemic control
Mean blood glucose (mmol/L)	7.2 ± 0.9	8.6 ± 1.3	< 0.001
Hyperglycemic episodes requiring insulin	20 (23.8)	35 (41.7)	0.013
Inflammatory markers
C-reactive protein at 48 hours (mg/L)	87.3 ± 18.4	113.6 ± 24.7	< 0.001

EA: Epidural anesthesia; IVPCA: Intravenous patient-controlled analgesia.

Subgroup analysis of epidural analgesia benefits: We conducted a comprehensive subgroup analysis to identify patient populations that might derive particularly significant benefits from epidural analgesia compared to IVPCA. The primary outcome for this analysis was a composite of “optimal recovery” (defined as the absence of moderate-severe pain, no pulmonary complications, return of bowel function within 72 hours, and discharge within 14 days). In the overall cohort, EA significantly increased the likelihood of optimal recovery (adjusted OR = 2.86, 95%CI: 1.93-4.24, P < 0.001). However, subgroup analysis revealed notable variations in the magnitude of benefit. Elderly patients (age ≥ 70 years) derived the greatest benefit (adjusted OR = 3.75, 95%CI: 2.14-6.57, P < 0.001), followed by those with pre-existing pulmonary disease (adjusted OR = 3.42, 95%CI: 1.86-6.28, P < 0.001), obesity (BMI ≥ 30 kg/m²) (adjusted OR = 3.18, 95%CI: 1.74-5.83, P < 0.001), and diabetes mellitus (adjusted OR = 2.95, 95%CI: 1.68-5.17, P < 0.001). Moderate benefits were observed in patients with cardiovascular disease (adjusted OR = 2.47, 95%CI: 1.43-4.28, P = 0.001), ASA III physical status (adjusted OR = 2.31, 95%CI: 1.36-3.92, P = 0.002), and extended operative times (> 6 hours) (adjusted OR = 2.26, 95%CI: 1.32-3.88, P = 0.003). The benefit was less pronounced but still statistically significant in younger patients (< 50 years) (adjusted OR = 1.84, 95%CI: 1.08-3.12, P = 0.024) and those with normal BMI (18.5-24.9 kg/m²) (adjusted OR = 1.76, 95%CI: 1.05-2.94, P = 0.031). These findings suggest that while epidural analgesia provides benefits across the entire cohort, it may be particularly advantageous in high-risk patient populations (Figure 3B).

DISCUSSION

Postoperative pain management represents one of the most critical aspects of perioperative care, with profound implications for patient recovery trajectories and healthcare resource utilization[17-19]. The evolution of analgesic approaches over the past several decades has been characterized by a progressive shift toward multimodal strategies that balance effective pain control with minimization of adverse effects. EA has emerged as a cornerstone technique within this paradigm, particularly for major abdominal and thoracic procedures where pain intensity is substantial and prolonged[20-22]. The theoretical advantages of epidural analgesia have been well-established in physiological studies, demonstrating superior neuroaxial blockade of nociceptive signals and sympathetic modulation compared to systemic analgesics. This neuraxial approach provides segmental analgesia while potentially preserving motor function and hemodynamic stability when appropriately titrated. Previous meta-analyses have suggested benefits in respiratory parameters, stress response attenuation, and earlier return of gastrointestinal function, although the magnitude and consistency of these effects have varied considerably across studies.

The implementation of enhanced recovery after surgery protocols has further highlighted the importance of optimized pain management strategies that facilitate early mobilization, minimize opioid consumption, and reduce physiologic stress responses. Within this context, the selection of analgesic modality represents a decision point with cascading effects on multiple downstream recovery parameters. However, the heterogeneity of surgical populations, procedural characteristics, and institutional protocols has complicated efforts to establish definitive guidelines regarding analgesic technique selection[23,24]. Furthermore, advances in minimally invasive surgical approaches, regional anesthetic techniques, and pharmacologic innovations have expanded the armamentarium available to perioperative physicians. This evolving landscape necessitates continual reassessment of established practices against contemporary alternatives. IVPCA, while offering flexibility and titrability, presents concerns regarding respiratory depression, delayed gastrointestinal recovery, and potential for tolerance and hyperalgesia with prolonged exposure to systemic opioids[25-28]. Risk stratification represents an increasingly important concept in perioperative medicine, acknowledging the heterogeneity of patient populations and the potential for differential responses to interventions based on underlying physiologic reserve and comorbidity profiles[29,30]. The identification of subgroups that derive particularly substantial benefits from specific interventions aligns with broader movements toward personalized medicine approaches in surgical care.

Our findings contribute to this evolving body of evidence by demonstrating that epidural analgesia represents the strongest independent predictor of favorable postoperative recovery (adjusted OR = 3.42) in our multivariate analysis. This benefit appears particularly pronounced in specific high-risk subgroups, with elderly patients (age ≥ 70 years) deriving the greatest advantage (adjusted OR = 3.75), followed by those with pre-existing pulmonary disease (adjusted OR = 3.42). The significant reduction in postoperative pulmonary complications observed in the EA group compared to IVPCA (11.9% vs 23.8%, P = 0.042) further substantiates the protective effect of epidural analgesia on respiratory function, a finding particularly relevant for patients with compromised baseline pulmonary status.

The subgroup analysis revealing substantial benefits in patients with metabolic comorbidities, including obesity (adjusted OR = 3.18) and diabetes mellitus (adjusted OR = 2.95), suggests that the physiologic advantages of epidural analgesia may be amplified in populations with baseline inflammatory states and impaired wound healing characteristics. While the benefits were less pronounced but still statistically significant in younger patients and those with normal BMI, the consistent direction of effect across all subgroups supports the generalizability of epidural analgesia's advantages while simultaneously identifying populations for whom this intervention may be particularly impactful.

The historical context of pain management in major abdominal surgery has been characterized by a transition from predominantly opioid-based approaches toward more nuanced, multimodal strategies. This evolution has occurred alongside growing recognition of the substantial impact that pain control has on both immediate recovery parameters and long-term outcomes, including chronic post-surgical pain development and functional restoration. Traditional approaches prioritized pain intensity reduction as the primary endpoint, whereas contemporary perspectives emphasize a more comprehensive view of recovery that encompasses functional measures, physiologic restoration, and patient-reported outcomes[31-33].

In this broader context, our finding that epidural analgesia emerged as the strongest independent predictor of favorable recovery (adjusted OR = 3.42) provides important validation for its continued role in the era of enhanced recovery protocols. The observed differential benefit across subgroups further refines our understanding of how to optimize resource allocation and intervention targeting. The particularly pronounced advantage in elderly patients (adjusted OR = 3.75) aligns with growing evidence that this population experiences both greater vulnerability to opioid-related adverse effects and more substantial impairments in functional mobility when pain is inadequately controlled. Similarly, the substantial benefit observed in patients with pre-existing pulmonary disease (adjusted OR = 3.42) and the significant reduction in pulmonary complications (11.9% vs 23.8%, P = 0.042) reinforce the physiologic principle that effective thoracic epidural analgesia preserves diaphragmatic function, improves pulmonary mechanics, and facilitates effective coughing mechanisms, particularly critical for those with limited baseline respiratory reserve. These findings suggest that while epidural analgesia provides significant advantages across the entire cohort, strategic prioritization of this resource-intensive intervention for high-risk subgroups may represent a particularly high-value approach in resource-constrained environments.

Study limitations

This study has several important limitations that should be acknowledged. The retrospective design introduces potential for unmeasured confounding despite the use of propensity score matching to balance baseline characteristics between groups. While we adjusted for major demographic, clinical, and surgical variables, residual confounding from unmeasured factors such as surgeon experience, subtle variations in surgical technique, or institutional practice changes over the study period cannot be completely eliminated. The single-center design may limit the generalizability of our findings to institutions with different protocols, patient populations, or healthcare systems. Selection bias may exist regarding epidural placement decisions, as patients with absolute contraindications to neuraxial anesthesia were excluded from consideration, potentially creating a more favorable cohort for epidural intervention.

CONCLUSION

EA provides superior pain control, facilitates earlier ambulation and return of bowel function, shortens hospital stay, and improves patient satisfaction compared to IVPCA in patients undergoing pancreaticoduodenectomy.