Abstract
Background
Optimizing postoperative recovery for elderly gastric cancer patients is of great clinical significance. This retrospective cohort study aimed to explore the association between preoperative use of dexamethasone and postoperative outcomes in elderly patients undergoing gastric cancer resection.
Methods
In this retrospective cohort study, elderly patients who underwent gastric cancer resection were enrolled. Baseline and perioperative data including use of dexamethasone for antiemetic purpose were collected. Patients were divided into control group (no dexamethasone), low-dose group (< 10 mg) and high-dose group (≥ 10 mg) based on the dose of dexamethasone administered preoperatively on the discretion of anesthesiologists. The primary endpoint was postoperative length of hospital stay (pLOS), defined as the duration from surgery to initial discharge (prolonged pLOS referred to pLOS ≥ median pLOS of all patients). The association between perioperative dexamethasone use and prolonged pLOS were analyzed with multivariable logistic regression models. Dose-response trend test and sensitivity analysis were conducted to confirm the reliability of the findings.
Results
The high-dose group had significantly shorter pLOS and lower overall complication rate than the control group (10 [9–13] vs 13 [10–17] days, P = 0.010; 20.9% vs 41.9%, P = 0.011). After adjusting for multiple confounders, preoperative dexamethasone ≥10 mg was independently associated with shorter pLOS (OR 0.355, 95% CI 0.166–0.756, P = 0.007). The high-dose group showed lower risks of pulmonary infection and surgical site infection. Both dexamethasone groups had reduced incidence of acute coronary syndrome.
Conclusion
Preoperative dexamethasone ≥10 mg was associated with improved short-term postoperative outcomes in elderly gastric cancer patients, while the dose <10 mg only correlated with lower acute coronary syndrome risk. These observational associations need further prospective validation.
Keywords: preoperative dexamethasone, elderly patients, gastric cancer, postoperative prognosis
Introduction
Gastric cancer remains a prevalent malignancy globally, despite a gradual decline in its incidence over the past few decades.1,2 With the rapid aging of the global population, the number of elderly patients undergoing gastric cancer surgery has been steadily increasing.3 Elderly patients often have multiple comorbidities and age-related organ function decline, increasing postoperative morbidity and mortality.4–6 Therefore, identifying effective interventions to improve postoperative prognosis in elderly gastric cancer patients is of great clinical significance.7
Surgical inflammatory responses are essential for tissue healing,8 but excessive inflammatory response can impair organ function and delay recovery.9 Preoperative dexamethasone, used prior to anesthetic induction to prevent postoperative nausea and vomiting,10 has been integrated into many “enhanced recovery after surgery” (ERAS) protocols due to its potent anti-inflammatory effect.11 Studies have suggested that preoperative administration of dexamethasone is associated with a decrease in systemic inflammatory response (SIR) and complications after elective cancer surgery.12,13 Another study has also found that compared to lower doses, higher doses of perioperative dexamethasone were associated with a reduction in systemic inflammation and complications after colon cancer surgery, without having a negative impact on survival.14
Currently, most existing studies focused on general surgical populations, and high-quality evidence targeting elderly gastric cancer patients is still limited. What’ more, prior literature mainly confirmed the anti-inflammatory association of dexamethasone, rather than definitive evidence for improving postoperative prognosis in this specific elderly population, and the potential risks of dexamethasone, including hyperglycemia, infection and immunosuppression also should not be ignored. Based on the findings and limitations of previous studies, we hypothesized that preoperative dexamethasone use is associated with different postoperative recovery outcomes among dose groups in elderly gastric cancer patients. The present retrospective cohort study, with 10 mg as the cut-off value according to our institutional routine practice, aimed to explore this relationship to provide reference for perioperative management.
Materials and Methods
Study Population
Elderly patients (≥ 65 years) who underwent gastric cancer resection at our institute between April 2015 and November 2018 were enrolled in this retrospective cohort study. Exclusion criteria: (1) concurrent presence of other primary tumors; (2) reoperation due to postoperative recurrence of gastric cancer; (3) missing key clinical data or reliance on manually recorded data that could not be verified.
Patient data were retrieved from the hospital’s electronic medical record system and analyzed retrospectively. All patient medical data were fully anonymized and only used for this academic research to ensure patient data confidentiality. The Clinical Research Ethics Committee of Peking University Cancer Hospital approved our study protocol (approval number: 2018YJZ71) and waived the written informed consent due to the observational nature of the study. This study was performed in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki and its later amendments.
Anesthesia, Surgery, and Perioperative Management
All patients received general anesthesia with endotracheal intubation. Anesthesia induction was performed with intravenous anesthetics (propofol and/or etomidate), opioids (fentanyl or sufentanil), and muscle relaxants. Anesthesia maintenance was achieved with inhalation anesthetics (sevoflurane or isoflurane) and opioids (fentanyl and/or sufentanil and/or oxycodone and/or dezocine and/or pentazocine and/or remifentanil). Dexamethasone (mainly 5 or 10 mg, also with other doses) was administered to prevent postoperative nausea and vomiting depending on the discretion of anesthesiologists.15 No other specific indications for glucocorticoids administration. Some patients received preoperative neoadjuvant chemotherapy and the type of surgery was determined based on tumor location and extent. All other perioperative treatments were conducted in accordance with routine clinical practice.
Data Collection
Relevant data were collected from the hospital’s electronic medical record system, including: (1) Baseline data: gender, age, body mass index (BMI), smoking history, alcohol consumption history, preoperative comorbidities, preoperative laboratory test results, American Society of Anesthesiologists (ASA) classification, and preoperative neoadjuvant chemotherapy; (2) Anesthetic and surgical information: preoperative dexamethasone dose, type and dose of anesthetics, intraoperative transfusion and fluid volume, perioperative opioids (converted to sufentanil equivalent),16–18 operation type, duration and intraoperative blood loss; (3) Tumor and postoperative data: pathological results, tumor differentiation, pTNM stage, postoperative complications and pLOS.
Group Stratification Based on Preoperative Dexamethasone Dose
Preoperative dexamethasone doses for antiemesis were extracted from anesthesia records. Patients were stratified into three groups: control group (no dexamethasone), low-dose group (group L, preoperative dexamethasone < 10 mg), and high-dose group (group H, preoperative dexamethasone ≥ 10 mg). In our clinical practice, the conventional antiemetic doses are mainly 5 mg and 10 mg; doses such as 2 mg, 8 mg, 15 mg are extremely rare, but still exist.
Study Endpoints
Primary endpoint: postoperative length of hospital stay (pLOS), defined as the interval between surgery date and first discharge. Secondary endpoint: postoperative complications (Clavien-Dindo grades II–V). All complications were evaluated by professional surgeons, and the follow-up was limited to in-hospital period.
Statistical Analysis
All statistical analyses were performed using SPSS 25.0. Continuous data are presented as median (IQR), categorical data as counts (percentages). Univariable logistic regression was used to screen factors associated with prolonged pLOS. Variables with P < 0.20 in univariable analysis and clinically important indicators were included in multivariable models. The overall median pLOS was 12 days, so we defined prolonged pLOS as ≥12 days. Postoperative complications were excluded from multivariable covariates, because complications belong to intermediate variables on the causal pathway and will cause over-adjustment bias. We conducted dose-response trend test and sensitivity analysis to confirm the reliability of the findings. There is no missing data in the final analysis of the present study. All statistical tests were two-tailed, and P < 0.05 was considered statistically significant.
Considering the retrospective design, we also performed a post-hoc power analysis to evaluate the statistical power of the study. Using the G*Power 3.1.9.7 software, we calculated the power for the primary outcome based on the observed data: with a two-sided type I error rate of 0.05, the control group had a 62.16% rate of prolonged postoperative length of stay, the observed odds ratio for high-dose dexamethasone was 0.355, the high-dose group accounted for 24.1% of the total sample, and we adjusted for 11 confounding variables in the model. The results showed that the present study had an achieved power of 92.7%, which far exceeded the conventional threshold of 80% for adequate statistical power, indicating that the primary findings of this study are unlikely to be affected by insufficient sample size or type II error.
Results
Patient Enrollment and Baseline Characteristics
Of 296 initially enrolled patients, 18 were excluded, including 8 with gastric cancer recurrence and reoperation, 6 with concurrent other primary tumors, 1 with no electronic record data, and 3 with missing critical clinical data (1 missing tumor pathological stage and 2 missing glucocorticoid-related information). The remaining 278 patients (median age 69 [interquartile range, IQR 66–73] years) were included in the final analysis, with 74 patients in the control group (dexamethasone dose, 0 mg), 137 patients in group L (median dexamethasone dose, 5 [IQR 5–5] mg), and 67 patients in group H (10 [10–10] mg) (Figure 1, Table 1 and Supplementary Table 1).
Figure 1.
Flowchart of the study.
Table 1.
Baseline and Perioperative Data of Different Groups
| Variable | All Patients (n=278) | Control (n=74) | Group L (n=137) | Group H (n=67) | P value |
|---|---|---|---|---|---|
| Age (year) | 69 (66–73) | 69 (67–73) | 70 (67–73) | 68 (66–73) | 0.838 |
| Male sex | 213 (76.6) | 53 (71.6) | 110 (80.3) | 50 (74.6) | 0.337 |
| Body mass index (kg/m2) | 23.3 (20.8–25.2) | 22.6 (20.4–25.0) | 23.6 (20.9–25.5) | 22.8 (21.2–24.8) | 0.290 |
| Smoking history | 120 (43.2) | 30 (40.5) | 57 (41.6) | 33 (49.3) | 0.515 |
| Drinking history | 78 (28.1) | 21 (28.4) | 31 (22.6) | 26 (38.8) | 0.055 |
| Preoperative comorbidities | |||||
| Previous stroke | 27 (9.7) | 6 (8.1) | 15 (10.9) | 6 (9.0) | 0.836 |
| HBP | 106 (38.1) | 23 (31.1) | 55 (40.1) | 28 (41.8) | 0.333 |
| Coronary heart disease | 25 (9.0) | 7 (9.5) | 14 (10.2) | 4 (6.0) | 0.632 |
| COPDa | 12 (4.3) | 1 (1.4) | 8 (5.8) | 3 (4.5) | 0.297 |
| Diabetes | 46 (16.5) | 9 (12.2) | 21 (15.3) | 16 (23.9) | 0.149 |
| Renal dysfunctionb | 8 (2.9) | 1 (1.4) | 3 (2.2) | 4 (6.0) | 0.215 |
| Hepatic dysfunctionc | 20 (7.2) | 3 (4.1) | 14 (10.2) | 3 (4.5) | 0.163 |
| ASA grade | 0.004 | ||||
| I | 11 (4.0) | 7 (9.5) | 2 (1.5) | 2 (3.0) | |
| II | 236 (84.9) | 65 (87.8) | 114 (83.2) | 57 (85.1) | |
| III | 31 (11.2) | 2 (2.7) | 21 (15.3) | 8 (11.9) | |
| Preoperative lab test | |||||
| Preoperative WBC count >10×109/L | 48 (17.3) | 17 (23.0) | 22 (16.1) | 9 (13.4) | 0.285 |
| Preoperative albumin <35 g/L | 26 (9.4) | 12 (16.2) | 10 (7.3) | 4 (6.0) | 0.057 |
| Preoperative anemiad | 2 (0.7) | 0 (0) | 2 (1.5) | 0 (0) | 0.370 |
| Preoperative chemotherapy | 105 (37.8) | 32 (43.2) | 53 (38.7) | 20 (29.9) | 0.263 |
| Pathological diagnosis | 0.775 | ||||
| Adenocarcinoma | 252 (90.6) | 67 (90.5) | 125 (91.2) | 60 (89.6) | |
| Signet-ring cell carcinoma | 9 (3.2) | 1 (1.4) | 4 (2.9) | 4 (6.0) | |
| Adenocarcinoma combined signet-ring cell carcinoma | 6 (2.2) | 3 (4.1) | 2 (1.5) | 1 (1.5) | |
| Neuroendocrine tumor | 9 (3.2) | 2 (2.7) | 5 (3.6) | 2 (3.0) | |
| Squamous cell carcinoma | 2 (0.7) | 1 (1.4) | 1 (0.7) | 0 (0) | |
| pTNM stagee | 0.581 | ||||
| IA | 64 (23.0) | 11 (14.9) | 33 (24.1) | 20 (29.9) | |
| IB | 28 (10.1) | 6 (8.1) | 16 (11.7) | 6 (9.0) | |
| IIA | 38 (13.7) | 13 (17.6) | 17 (12.4) | 8 (11.9) | |
| IIB | 34 (12.2) | 13 (17.6) | 16 (11.7) | 5 (7.5) | |
| III | 102 (36.7) | 28 (37.8) | 49 (35.8) | 25 (37.3) | |
| IV | 12 (4.3) | 3 (4.1) | 6 (4.4) | 3 (4.5) | |
| Differentiation grade | 0.161 | ||||
| NA | 107 (38.2) | 28 (37.8) | 58 (42.3) | 21 (31.3) | |
| Low | 161 (58.2) | 41 (55.4) | 75 (54.7) | 45 (67.2) | |
| Moderate | 8 (2.9) | 3 (4.1) | 4 (2.9) | 1 (1.5) | |
| High | 2 (0.7) | 2 (2.7) | 0 (0) | 0 (0) | |
| Range of gastrectomy | 0.145 | ||||
| Total gastrectomy | 137 (49.3) | 40 (54.1) | 72 (52.6) | 25 (37.3) | |
| Distal gastrectomy | 128 (46.0) | 31 (41.9) | 60 (43.8) | 37 (55.2) | |
| Proximal gastrectomy | 9 (3.2) | 1 (1.4) | 5 (3.6) | 3 (4.5) | |
| Palliative resection | 4 (1.4) | 2 (2.7) | 0 (0) | 2 (3.0) | |
| Type of procedure | 0.289 | ||||
| Laparoscopic surgery | 63 (22.7) | 21 (28.4) | 26 (19.0) | 16 (23.9) | |
| Open surgery | 215 (77.3) | 53 (71.6) | 111 (81.0) | 51 (76.1) | |
| Duration of surgery (min) | 215 (172–251) | 217 (182–259) | 210 (169–250) | 214 (172–252) | 0.197 |
| Inhalational anesthetics | 0.012 | ||||
| Sevoflurane | 261 (93.9) | 65 (87.8) | 134 (97.8) | 62 (92.5) | |
| Isoflurane | 17 (6.1) | 9 (12.2) | 3 (2.2) | 5 (7.5) | |
| Intraoperative fluid | |||||
| Crystalloids (mL) | 1550 (1200–1713) | 1500 (1200–1800) | 1600 (1100–1700) | 1500 (1200–1700) | 0.596 |
| Artificial colloid (mL)f | 1000 (500–1000) | 700 (500–1000) | 1000 (500–1000) | 500 (500–1000) | 0.009 |
| Blood transfusion | 6 (2.2) | 3 (4.1) | 3 (2.2) | 0 (0) | 0.321 |
| Sufentanil equivalent (µg)g,h | 256 (165–401) | 210 (150–375) | 330 (198–435) | 210 (152–305) | <0.001 |
| Preoperative DXM (mg) | 5 (0–5) | 0 (0–0) | 5 (5–5) | 10 (10–10) | – |
Notes: Data are median (interquartile range) or number (%). P value in bold indicate those <0.05. aIncludes COPD (chronic bronchitis and emphysema), asthma, and old tuberculosis. bSerum creatinine >133µmol/L. cAlanine transaminase, aspartate transaminase and/or total bilirubin 2 times higher than the upper normal limit. dFor men, hemoglobin <120 g/L; for women, hemoglobin <110 g/L. eAccording to the 8th Edition of the American Joint Committee on Cancer/Union for International Cancer Control Staging System. fIncludes hydroxyethyl starch 130/0.4, hydroxyethyl starch 200/0.5, and succinylated gelatin. gIncludes morphine, fentanyl, sufentanil, oxycodone, and dezocine. h1 mg of morphine (iv)=15 µg of fentanyl=1.5 µg of sufentanil=1 mg of oxycodone=1 mg of dezocine.16–18.
Abbreviations: ASA, American Society of Anesthesiologists; COPD, chronic obstructive pulmonary disease; pTNM, pathological tumor-node-metastasis; WBC, white blood cell; DMX, dexamethasone.
Comparison of Baseline and Perioperative Variables
Comparison of baseline and perioperative data among the three groups revealed significant differences in ASA grade III proportion (2 [2.7%] vs 21 [15.3%] vs 8 [11.9%], P = 0.004), sevoflurane use (65 [87.8%] vs 134 [97.8%] vs 62 [92.5%], P = 0.012), perioperative opioid dose (median 210 [IQR 150–375] vs 330 [198–435] vs 210 [152–305] μg sufentanil equivalent, P < 0.001) and artificial colloid infusion volume (700 [500–1000] vs 1000 [500–1000] vs 500 [500–1000] mL, P = 0.009). No other significant differences were noted (Table 1).
Postoperative Outcomes
Compared with the control group, patients in group H had a significantly shorter pLOS (median 10 [IQR 9–13] vs 13 [10–17] days, P = 0.010) and a significantly lower total incidence of postoperative complications (14 [20.9%] vs 31 [41.9%], P = 0.011). With regard to specific postoperative complications, patients in group H had a significantly lower incidence of postoperative pulmonary infection (2 [3.0%] vs 12 [16.2%] in control, P = 0.010) and surgical site infection (1 [1.5%] vs 17 [23.0%] in control, P<0.001; 1 [1.5%] vs 18 [13.1%] in group L, P = 0.008). Both group L and H had lower acute coronary syndrome incidence than the control group (3 [2.2%] vs 7 [9.5%], P = 0.018; 0 [0%] vs 7 [9.5%], P = 0.014, respectively). No significant differences were observed in the incidence of other postoperative complications among the three groups (Table 2).
Table 2.
Postoperative Outcomes of Different Groups
| Variables | All Patients (n=278) | Control (n=74) | Group L (n=137) | Group H (n=67) | P1 value | P2 value | P3 value |
|---|---|---|---|---|---|---|---|
| Primary outcome | |||||||
| Length of hospital stay (day)a | 12 (9–15) | 13 (10–17) | 13 (10–15) | 10 (9–13) | 0.108 | 0.010 | 0.171 |
| Secondary outcome | |||||||
| Postoperative complicationsb | 90 (32.4) | 31 (41.9) | 45 (32.8) | 14 (20.9) | 0.192 | 0.011 | 0.077 |
| Postoperative strokec | 5 (1.8) | 1 (1.4) | 2 (1.5) | 2 (3.0) | 0.949 | 0.604 | 0.461 |
| Acute coronary syndromed | 10 (3.6) | 7 (9.5) | 3 (2.2) | 0 (0) | 0.018 | 0.014 | 0.552 |
| Pulmonary infectione | 30 (10.8) | 12 (16.2) | 16 (11.7) | 2 (3.0) | 0.354 | 0.010 | 0.062 |
| Delayed gastric emptyingf | 13 (4.7) | 3 (4.1) | 7 (5.1) | 3 (4.5) | 0.731 | 0.901 | 0.844 |
| Hepatic insufficiencyg | 11 (4.0) | 4 (5.4) | 5 (3.6) | 2 (3.0) | 0.547 | 0.683 | 0.807 |
| Acute kidney injuryh | 7 (2.5) | 2 (2.7) | 5 (3.6) | 0 (0) | 0.714 | 0.498 | 0.113 |
| Anastomotic fistulai | 26 (9.4) | 8 (10.8) | 16 (11.7) | 2 (3.0) | 0.850 | 0.101 | 0.062 |
| Postoperative bleedingj | 12 (4.3) | 5 (6.8) | 5 (3.6) | 2 (3.0) | 0.311 | 0.445 | 0.807 |
| Deep venous thrombosisk | 3 (1.1) | 1 (1.4) | 2 (1.5) | 0 (0) | 0.949 | >0.999 | 0.320 |
| Surgical site infectionl | 36 (12.9) | 17 (23.0) | 18 (13.1) | 1 (1.5) | 0.067 | <0.001 | 0.008 |
Notes: Results are median (interquartile range) or number (%). P values in bold indicate those of <0.05. P1, control vs group L; P2, control vs group H; P3, group L vs group H. aTime between the day of surgery and the first discharge date. bClass II–V on Clavien-Dindo classification.19 cConfirmed by clinical symptoms and medical imaging examination. dConfirmed by clinical symptoms, electrocardiographic changes, and elevated cardiac troponin I levels. eNew infiltration on chest X-ray with fever and treated with intravenous antibiotics. fGastric tube drainage >800 mL/d, with no other obvious reasons. gProlonged prothrombin time (>18 s) and elevated serum total bilirubin concentration (>30 μmol/L or 1.7 mg/dL). hIncreased serum creatinine ≥0.3 mg/dL (≥26.5 μmol/L) within 7 days after surgery. iGastric juice–like fluid in the abdominal drainage fluid with a volume >100 mL per day after surgery or gastric juice–like fluid extracted from the local peritoneal effusion under ultrasound and computed tomography positioning. jRequired second surgery for hemostasis. kConfirmed by lower limbs venous ultrasonography and treated with anticoagulants. lPus expressed from the incision and bacteria cultured from the pus or purulent puncture extract.
Association Between Preoperative Dexamethasone and Postoperative Length of Hospital Stay
The median pLOS in the entire study cohort was 12 days, so prolonged pLOS was defined as ≥12 days. Multivariable logistic regression analysis was performed to evaluate the relationship between potential risk factors and prolonged pLOS. 9 variables with P < 0.20 in the univariable analysis (preoperative dexamethasone dose, preoperative coronary heart disease [CHD], ASA grade, pTNM stage, type of surgery, gastrectomy scope, surgical duration, intraoperative artificial colloid infusion volume, and intraoperative blood transfusion) and 2 clinically important variables (age and tumor differentiation grade) were included in the multivariable logistic regression model. Postoperative complications were excluded from multivariable covariates, because complications belong to intermediate variables on the causal pathway and will cause over-adjustment bias. After adjusting for confounding factors, preoperative dexamethasone ≥10 mg (median 10 [IQR 10–10] mg) was independently associated with a shorter pLOS (odds ratio [OR] 0.355, 95% confidence interval [CI] 0.166–0.756, P = 0.007). Among other factors, total gastrectomy (compared with distal and proximal gastrectomy), open surgery (compared with laparoscopic surgery), and prolonged duration of surgery were independently associated with a longer pLOS (Table 3).
Table 3.
Risk Factors for Postoperative Length of Hospital Stay After Gastric Cancer Surgery: Univariable and Multivariable Logistic Regression Analyses
| Variable | Univariable Analysis | Multivariable Analysis | ||
|---|---|---|---|---|
| OR (95% CI) | P value | OR (95% CI) | P value | |
| Preoperative dexamethasone | ||||
| No use | ||||
| <10mg | 0.880 (0.493 to 1.573) | 0.667 | 0.728 (0.379 to1.399) | 0.341 |
| ≥10 mg | 0.340 (0.171 to 0.674) | 0.002 | 0.355 (0.166 to 0.756) | 0.007 |
| Age, years | 1.003 (0.953 to1.054) | 0.919 | 1.018 (0.961 to 1.078) | 0.545 |
| Male sex | 0.871 (0.498 to 1.525) | 0.630 | ||
| Body mass index, kg/m2 | 0.999 (0.924 to 1.079) | 0.971 | ||
| Smoking history | 0.989 (0.614 to 1.593) | 0.965 | ||
| Drinking history | 0.907 (0.537 to 1.532) | 0.714 | ||
| History of stroke | 0.896 (0.405 to 1.984) | 0.787 | ||
| Coronary heart disease | 1.887 (0.786 to 4.531) | 0.155 | 1.106 (0.381 to 3.212) | 0.853 |
| Hypertension | 0.803 (0.494 to 1.305) | 0.376 | ||
| COPDa | 1.720 (0.506 to 5.851) | 0.385 | ||
| Diabetes | 1.002 (0.531 to 1.890) | 0.996 | ||
| Hepatic dysfunctionb | 0.830 (0.334 to 2.062) | 0.688 | ||
| Renal dysfunctionc | 2.585 (0.513 to 13.043) | 0.250 | ||
| ASA grade (III vs I and II) | 2.236 (0.990 to 5.050) | 0.053 | 2.474 (0.897 to 6.826) | 0.080 |
| Preoperative WBC count >10×109/L | 1.351 (0.717 to 2.547) | 0.352 | ||
| Preoperative albumin <35 g/L | 1.663 (0.715 to 3.872) | 0.238 | ||
| Preoperative anemiad | 0.840 (0.052 to 13.566) | 0.902 | ||
| Preoperative chemotherapy | 1.062 (0.652 to 1.728) | 0.810 | ||
| Pathological diagnosis (adenocarcinoma vs others)e | 0.955 (0.574 to 1.589) | 0.860 | ||
| pTNM stage (IIB, III, IV vs IA, IB, IIA)f | 1.432 (0.891 to 2.300) | 0.138 | 0.979 (0.564 to 1.701) | 0.979 |
| Differentiation grade (low or NA vs moderate or high) | 1.261 (0.776 to 2.048) | 0.350 | 1.366 (0.790 to 2.361) | 0.264 |
| Scope of gastrectomy (total vs distal or proximal gastrectomy) | 3.433 (2.091 to 5.636) | <0.001 | 2.721 (1.591 to 4.654) | <0.001 |
| Type of surgery (laparoscopic vs open gastrectomy) | 0.426 (0.240 to 0.759) | 0.004 | 0.471 (0.243 to 0.912) | 0.025 |
| Duration of surgery | 1.006 (1.002 to 1.010) | 0.004 | 1.005 (1.000 to 1.010) | 0.035 |
| Inhalational anesthetics (isoflurane vs sevoflurane) | 1.585 (0.569 to 4.412) | 0.378 | ||
| Sufentanil equivalent (mg)g | 1.001 (0.999 to 1.003) | 0.318 | ||
| Blood transfusionh | 4.315 (0.498 to 37.425) | 0.185 | 1.982 (0.160 to 24.489) | 0.594 |
| Artificial colloids (mL)i | 1.001 (1.000 to 1.002) | 0.022 | 1.001 (1.000 to 1.002) | 0.108 |
| Crystalloids (mL) | 1.000 (1.000 to 1.001) | 0.738 | ||
Notes: P values in bold indicate P<0.20 in univariable analysis and P<0.05 in multivariable analysis. aIncludes COPD (chronic bronchitis and emphysema), asthma, and old tuberculosis. bAlanine transaminase, aspartate transaminase and/or total bilirubin 2 times higher than the upper normal limit. cSerum creatinine >133µmol/L. dFor adult men, hemoglobin <120 g/L; for adult women, hemoglobin <110 g/L. eIncludes signet-ring cell carcinoma, adenocarcinoma combined signet-ring cell carcinoma, neuroendocrine tumor, lymphoid stromal carcinoma, melanoma, and squamous cell carcinoma. fAccording to the American Joint Committee on Cancer 8th Edition Cancer Staging System. g1 mg of morphine (iv)=15 µg of fentanyl=1.5 µg of sufentanil=1 mg of oxycodone=1 mg of dezocine.16–18 hPacked red blood cells were transfused. iIncludes hydroxyethyl starch 130/0.4, hydroxyethyl starch 200/0.5, and succinylated gelatin.
Abbreviations: ASA, American Society of Anesthesiologists; CI, confidence interval; COPD, chronic obstructive pulmonary disease; OR, odds ratio; pTNM, pathological tumor-node-metastasis; WBC, white blood cell.
Dose-Response Trend Test and Sensitivity Analysis
Trend χ2 test showed a significant linear trend between preoperative dexamethasone dose and the risk of prolonged postoperative length of stay: with the increase of dexamethasone dose, the risk of prolonged hospitalization decreased gradually (Pearson R = −0.185, P = 0.002). However, stratified analysis confirmed that only high-dose dexamethasone (≥ 10mg) was associated with significantly reduced risk, suggesting that the association was characterized by a threshold effect rather than a continuous linear dose-response relationship.
To verify the robustness of our primary findings, we also conducted a series of sensitivity analyses. First, we altered the cutoff value for defining prolonged postoperative length of stay (ie, from the primary 12 days to 10 days) to test the stability of the association. Second, we excluded 19 patients with extremely long postoperative hospitalization (>30 days) to eliminate the interference of outliers. Third, we further adjusted for additional preoperative indicators (ie, preoperative hypertension, diabetes) in the multivariate model to test for potential residual confounding. All sensitivity analyses yielded consistent results: the protective effect of high-dose dexamethasone (≥ 10mg) against prolonged hospitalization remained statistically significant, with the odds ratio only fluctuating between 0.340 and 0.371, and all P values < 0.05. These results confirmed that our primary conclusion was robust and unlikely to be affected by outcome definition, outliers, or residual confounding.
Discussion
Optimizing postoperative recovery in elderly cancer patients has become a major clinical focus.7 This retrospective study found that preoperative high-dose dexamethasone (≥ 10 mg) was associated with reduced risk of prolonged postoperative length of stay (pLOS) and lower postoperative complication risk in elderly gastric cancer patients, and this association remained significant after multivariate adjustment (OR = 0.355, 95% CI = 0.166–0.756, P = 0.007).
To confirm the reliability of this finding, we conducted a series of validation analyses: First, we performed a trend χ2 test test, which confirmed a significant linear trend between dexamethasone dose and the risk of prolonged hospitalization (Pearson R = −0.185, P=0.002), while stratified analysis verified that only the high-dose group had significant benefit, supporting the threshold effect of dexamethasone rather than continuous dose-response correlation. Second, we conducted multiple sensitivity analyses by altering the outcome cutoff, excluding outliers, and adjusting for additional confounders, all of which yielded consistent results: the odds ratio of high-dose dexamethasone only fluctuated between 0.340 and 0.371, with all P values < 0.05, confirming the robustness of our primary conclusion. Third, we performed post-hoc power analysis, which showed that our study had an achieved power of 92.7%, far exceeding the conventional 80% threshold, indicating that the primary findings were unlikely to be affected by insufficient sample size or type II error.
Previous studies have suggested that dexamethasone may mitigate excessive postoperative inflammatory responses and restore surgical-induced immune impairment, which could potentially contribute to improved postoperative recovery.20–24 A study of 556 patients undergoing elective cancer surgery at a single center between 2008 and 2016 showed that preoperative dexamethasone administration was associated with reduced SIR and postoperative complications.12 Another study of 480 patients who underwent potentially curative cancer resection between 2008 and 2017 also reported that perioperative dexamethasone was associated with reduced systemic inflammation and complications after colon cancer surgery.14 Consistently, our findings were in line with these previous reports: preoperative dexamethasone was associated with better postoperative outcomes. However, we did not measure perioperative inflammatory or immune indicators in this study, thus we cannot draw direct causal inference for the biological pathway, so the underlying mechanism of the observed protective effect remains to be fully elucidated.
Of note, we also evaluated the safety of high-dose dexamethasone, and found no significant increase in the risk of postoperative complications or anastomotic leakage in the high-dose group, confirming that a single preoperative dose of 10mg dexamethasone was safe for this cohort. What’ more, the present study also demonstrated that patients who received preoperative dexamethasone at ≥ 10 mg had a significantly lower incidence of postoperative pulmonary infection and surgical site infection compared with the control group, which was consistent with previous evidence that dexamethasone could reduce the risk of various infectious complications.25–28
Regarding the dose-effect relationship, a previous study compared the effects of different doses of dexamethasone (0 mg, 4 mg, or 8 mg) on postoperative recovery and found that higher doses of perioperative dexamethasone were associated with reduced systemic inflammation and complications after colon cancer surgery.14 Another meta-analysis explored the dose-effect relationship of dexamethasone by calculating milligram per kilogram doses and stratifying studies into low (<0.10 mg·kg−1), intermediate (0.11–0.20 mg·kg−1), or high (>0.21 mg·kg−1) dose groups, and found that high doses, but not low doses, of dexamethasone reduced 24-hour opioid consumption compared with placebo.29 Consistently, in the present study, we also identified a clear threshold effect of dexamethasone: only patients receiving ≥ 10 mg dexamethasone had significant improvements in pLOS and overall complications, while low-dose dexamethasone (<10mg) only reduced the incidence of postoperative acute coronary syndrome.
In the present study, we defined prolonged postoperative length of stay (pLOS) as a binary outcome based on the following considerations. First, in the clinical practice of our center, for elderly gastric cancer patients undergoing gastrectomy, the expected postoperative hospital stay is 10–12 days (median 12 days for the present cohort). A pLOS exceeding 12 days indicates delayed postoperative recovery, which is a clinically meaningful endpoint for perioperative management. Second, the distribution of continuous pLOS was right-skewed in our cohort, with several extreme outliers caused by non-recovery-related factors (eg, family care arrangement, waiting for rehabilitation transfer). Binary transformation could reduce the interference of these outliers on the effect estimation, making the result more robust. Third, compared with the continuous pLOS, the binary endpoint can more intuitively reflect the clinical prognosis of patients, and the estimated odds ratio is easier to interpret for clinical clinicians. Fourth, to further confirm that our findings were not affected by the selection of cutoff value, we conducted a series of sensitivity analyses including altering the cutoff to 10 days, and all results remained consistent, confirming that the protective effect of high-dose dexamethasone was robust regardless of the definition of prolonged pLOS.
This study has several limitations. First, obvious baseline imbalances existed in ASA grade, anesthetic type, opioid and colloid dosage, which may cause residual confounding. Second, this is a single-center retrospective study; dexamethasone was used at anesthesiologist’s discretion, leading to selection bias and indication bias. Third, a small number of patients may receive postoperative dexamethasone, bringing potential bias. Fourth, the sample size is limited, which may affect statistical power. Fifth, the study period (2015–2018) may bring temporal variation of clinical practice. So the results need further verification by large-scale, multi-center prospective trials. Despite the above limitations, this study provides valuable observational evidence for elderly gastric cancer perioperative management.
Conclusions
This retrospective observational study revealed a threshold effect of preoperative dexamethasone: preoperative dexamethasone ≥ 10 mg was associated with shorter hospital stay and lower complication risk, with confirmed safety, while low-dose dexamethasone had limited benefit. These findings provide valuable reference for perioperative management of elderly gastric cancer patients, while all findings are observational associations and cannot be regarded as definite clinical recommendations. Further prospective randomized studies are needed for validation.
Acknowledgments
The authors gratefully acknowledge Fei Shan, Professor, the First Department of Gastrointestinal Surgery, Peking University Cancer Hospital, Beijing, China for his help with data collection.
Data Sharing Statement
The data sets used and analyzed in the present study are available from the corresponding author on reasonable request.
Ethics Approval
The study protocol was approved by the clinical research ethics committee of Peking University Cancer Hospital (Approval No. 2018YJZ71).
Author Contributions
All authors made a significant contribution to the work reported, whether that is in the conception, study design, execution, acquisition of data, analysis and interpretation, or in all these areas; took part in drafting, revising or critically reviewing the article; gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work.
Disclosure
All authors declare no conflicts of interest in this work.
References
- 1.Zhu AL, Sonnenberg A. Is gastric cancer rising again? J Clin Gastroenterol. 2012;46(9):804–10. doi: 10.1097/MCG.0b013e3182604254 [DOI] [PubMed] [Google Scholar]
- 2.López MJ, Carbajal J, Alfaro AL, et al. Characteristics of gastric cancer around the world. Crit Rev Oncol Hematol. 2023;181:103841. doi: 10.1016/j.critrevonc.2022.103841 [DOI] [PubMed] [Google Scholar]
- 3.Amemiya T, Oda K, Ando M, et al. Activities of daily living and quality of life of elderly patients after elective surgery for gastric and colorectal cancers. Ann Surg. 2007;246(2):222–228. doi: 10.1097/SLA.0b013e3180caa3fb [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Yancik R, Wesley MN, Ries LA, et al. Comorbidity and age as predictors of risk for early mortality of male and female colon carcinoma patients. Cancer. 1998;82(11):2123–2134. doi: 10.1002/(SICI)1097-0142(19980601)82:11<2123::AID-CNCR6>3.0.CO;2-W [DOI] [PubMed] [Google Scholar]
- 5.Yancik R, Ries LA. Cancer in older persons. Magnitude of the problem. Cancer. 1994;74(7 Suppl):1995–2003. [DOI] [PubMed] [Google Scholar]
- 6.Sehl M, Sawhney R, Naeim A. Physiologic aspects of aging: impact on cancer management and decision making, part II. Cancer J. 2005;11(6):461–473. doi: 10.1097/00130404-200511000-00005 [DOI] [PubMed] [Google Scholar]
- 7.Otowa Y, Okamoto S, Fujinaka R, et al. Feasibility and effectiveness of gastrectomy for elderly gastric cancer patients. In Vivo. 2019;33(4):1307–1311. doi: 10.21873/invivo.11604 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.van Berge Henegouwen MI, van der Poll T, van Deventer SJ, et al. Peritoneal cytokine release after elective gastrointestinal surgery and postoperative complications. Am J Surg. 1998;175(4):311–316. doi: 10.1016/S0002-9610(98)00010-5 [DOI] [PubMed] [Google Scholar]
- 9.Lin E, Calvano SE, Lowry SF. Inflammatory cytokines and cell response in surgery. Surgery. 2000;127(2):117–126. doi: 10.1067/msy.2000.101584 [DOI] [PubMed] [Google Scholar]
- 10.Karanicolas PJ, Smith SE, Kanbur B, et al. The impact of prophylactic dexamethasone on nausea and vomiting after laparoscopic cholecystectomy. Ann Surg. 2008;248(5):751–762. doi: 10.1097/SLA.0b013e3181856024 [DOI] [PubMed] [Google Scholar]
- 11.watt DG, McSorley ST, Horgan PG, et al. Enhanced recovery after surgery: which components impact on the systemic inflammatory response following colorectal surgery? Medicine. 2015;94(36):e1286. doi: 10.1097/MD.0000000000001286 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.McSorley ST, Roxburgh CSD, Horgan PG, et al. The impact of preoperative dexamethasone on the magnitude of the postoperative systemic inflammatory response and complications following surgery for colorectal cancer. Ann Surg Oncol. 2017;24(8):2104–2112. doi: 10.1245/s10434-017-5817-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Blank M, Katsiampoura A, Wachtendorf LJ, et al. Association between intraoperative dexamethasone and postoperative mortality in patients undergoing oncologic surgery. Ann Surg. 2023;278(1):e105–e114. doi: 10.1097/SLA.0000000000005526 [DOI] [PubMed] [Google Scholar]
- 14.McSorley ST, Dolan RD, Roxburgh CS, et al. Possible dose dependent effect of perioperative dexamethasone and laparoscopic surgery on the postoperative systemic inflammatory response and complications following surgery for colon cancer. Eur J Surg Oncol. 2019;45(9):1613–1618. doi: 10.1016/j.ejso.2019.05.020 [DOI] [PubMed] [Google Scholar]
- 15.Gan TJ, Belani KG, Bergese S, et al. Fourth consensus guidelines for the management of postoperative nausea and vomiting. Anesth Analg. 2020;131(2):411–448. [DOI] [PubMed] [Google Scholar]
- 16.Treillet E, Laurent S, Hadji Y. Practical management of opioid rotation and equianalgesia. J Pain Res. 2018;11:2587–2601. doi: 10.2147/JPR.S170269 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Clotz MA, Nahata MC. Clinical uses of fentanyl, sufentanil, and alfentanil. Clin Pharm. 1991;10(8):581–593. [PubMed] [Google Scholar]
- 18.Silvasti M, Rosenberg P, Seppälä T, et al. Comparison of analgesic efficacy of oxycodone and morphine in postoperative intravenous patient-controlled analgesia. Acta Anaesthesiol Scand. 1998;42(5):576–580. doi: 10.1111/j.1399-6576.1998.tb05169.x [DOI] [PubMed] [Google Scholar]
- 19.Dindo D, Demartines N, Clavien PA. Classification of surgical complications: a new proposal with evaluation in a cohort of 6336 patients. Ann Surg. 2004;240(2):205–213. doi: 10.1097/01.sla.0000133083.54934.ae [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Asehnoune K, Le Moal C, Lebuffe G, et al. Effect of dexamethasone on complications or all cause mortality after major non-cardiac surgery. BMJ. 2021;373:n1162. doi: 10.1136/bmj.n1162 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Bouras M, Asehnoune K, Roquilly A. Contribution of dendritic cell responses to sepsis-induced immunosuppression and to susceptibility to secondary pneumonia. Front Immunol. 2018;9:2590. doi: 10.3389/fimmu.2018.02590 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Keh D, Boehnke T, Weber-Cartens, et al. Immunologic and hemodynamic effects of “low-dose” hydrocortisone in septic shock. Am J Respir Crit Care Med. 2003;167(4):512–520. doi: 10.1164/rccm.200205-446OC [DOI] [PubMed] [Google Scholar]
- 23.Roquilly A, Broquet A, Jacqueline C, et al. Hydrocortisone prevents immunosuppression by interleukin-10+ natural killer cells after trauma-hemorrhage. Crit Care Med. 2014;42(12):e752–e761. doi: 10.1097/CCM.0000000000000658 [DOI] [PubMed] [Google Scholar]
- 24.Kaufmann I, Briegel J, Schliephake F, et al. Stress doses of hydrocortisone in septic shock. Intensive Care Med. 2008;34(4):544–549. doi: 10.1007/s00134-007-0868-8 [DOI] [PubMed] [Google Scholar]
- 25.Assante J, Collins S, Hewer I. Infection associated with single-dose dexamethasone for prevention of postoperative nausea and vomiting: a literature review. AANA J. 2015;83(4):281–288. [PubMed] [Google Scholar]
- 26.Annane D, Bellissant E, Bollaert PE, et al. Corticosteroids for treating sepsis. Cochrane Database Syst Rev. 2019;12:CD002243. doi: 10.1002/14651858.CD002243.pub4 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Meijvis SC, Hardeman H, Remmelts HH, et al. Dexamethasone and length of hospital stay in patients with community-acquired pneumonia. Lancet. 2011;377(9782):2023–2030. doi: 10.1016/S0140-6736(11)60607-7 [DOI] [PubMed] [Google Scholar]
- 28.Horby P, Lim WS, Emberson JR, et al. Dexamethasone in hospitalized patients with Covid-19. N Engl J Med. 2021;384(8):693–704. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.De Oliveira GS, Almeida MD, Benzon HT, et al. Perioperative single dose systemic dexamethasone for postoperative pain. Anesthesiology. 2011;115(3):575–588. doi: 10.1097/ALN.0b013e31822a24c2 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The data sets used and analyzed in the present study are available from the corresponding author on reasonable request.