Abstract
Aim
To evaluate the potential benefit of external mirabilite dressing combined with abdominal drainage for treating severe acute pancreatitis (SAP).
Methods
This retrospective study included SAP patients from two centres who were divided into three groups: combined therapy (A, n = 24), mirabilite alone (B, n = 31), and drainage alone (C, n = 25). The outcomes included symptom recovery, laboratory parameters, intra-abdominal pressure (IAP), imaging findings, complications, and costs. Quantile regression adjusted for the Ranson score was performed.
Results
Despite the higher baseline Ranson score in Group A (4.00 vs. 3.00 in B and 4.00 in C; p = 0.026), Group A had a faster time to abdominal pain relief (7.50 vs. 11.00 and 9.00 days; P = 0.016), a faster time to flatus (2.00 vs. 3.00 and 2.00 days; P = 0.042), and faster white blood cell count normalization (3.00 vs. 6.00 and 5.00 days; P = 0.049). The IAP at discharge was lowest in Group A (7.35 vs. 8.46 and 8.82 mmHg, P = 0.032). Exudate area reduction (741.06 vs. 164.03 and 494.20 mm², P = 0.001) and pseudocyst area (323.94 vs. 484.30 and 453.38 mm², P = 0.038) were also both better in Group A. No differences were observed in the incidence of complications or health economic indicators (all p > 0.05).
Conclusion
In this retrospective, non-randomized study, combining external mirabilite dressing with abdominal drainage in early-stage SAP patients was associated with short-term clinical benefits without increasing total hospitalization costs. These findings are hypothesis-generating and require prospective validation.
Keywords: Severe acute pancreatitis, Mirabilite dressing, Abdominal drainage, Inflammatory response reduction, Combination therapy
Introduction
Acute pancreatitis is a common inflammatory disorder of the exocrine pancreas, whose aetiology includes gallstones, hypertriglyceridaemia, alcohol use, and other less common factors [1]. Gallstone disease is the most common aetiology. Some studies have suggested that the small, sharp projections of spiculated pigment stones predispose them to migrate and become impacted at the sphincter of Oddi, which can increase pressure in the common bile duct, leading to reflux into the pancreatic duct and subsequent activation of pancreatic enzymes [2]. While most cases are self‑limiting, a subset of patients progress to a severe form.
Severe acute pancreatitis (SAP), the most critical form of acute pancreatitis, is characterized by an aggressive disease course, systemic inflammatory response syndrome (SIRS), and frequent progression to multiple organ dysfunction syndrome (MODS). Despite advances in critical care, SAP mortality remains high, ranging from 20% to 40% [3, 4]. Therapeutic strategies that directly modify early disease progression remain a critical clinical need.
A pathophysiological hallmark of SAP is the rapid accumulation of protease- and cytokine-rich ascitic fluid within the peritoneal cavity. This pancreatitis-associated ascitic fluid (PAAF) acts as a potent driver of intra-abdominal hypertension (IAH), gut barrier disruption, and secondary systemic infections [5]. Although peritoneal drainage is routinely employed to evacuate PAAF and reduce intra-abdominal pressure (IAP), its efficacy as a monotherapy is often inconsistent and may be associated with risks such as haemorrhage or incomplete fluid clearance [6]. Conversely, conservative management alone is often insufficient to control rapid systemic inflammation. Thus, neither approach alone is fully satisfactory, and a combined strategy warrants investigation.
Mirabilite (Glauber’s salt, Na₂SO₄·10 H₂O) is a naturally occurring hydrated form of sodium sulfate. It has been used in traditional East Asian medicine for centuries to treat abdominal distension and inflammatory conditions [7]. From a modern biomedical perspective, mirabilite acts as a hyperosmotic and hygroscopic agent. When applied externally, it creates an osmotic gradient that promotes interstitial fluid redistribution. Preliminary clinical observations suggest that mirabilite may also reduce systemic inflammatory markers and lower the IAP in patients with acute pancreatitis [8, 9].
Given the complementary mechanisms of peritoneal drainage and mirabilite, a combined approach could provide potential synergistic benefits. However, few studies have directly compared combined therapy with monotherapy in patients with SAP. Therefore, this study aimed to evaluate the therapeutic effects of the combined treatment versus either intervention alone.
Materials and methods
Study design and patient selection
This retrospective cohort study was conducted at two medical centres: Changzhou Second Peopleʼs Hospital of Nanjing Medical University and Ma’anshan People’s Hospital. We reviewed the medical records of patients with a primary diagnosis of SAP who were admitted between January 2020 and January 2025. Patients were included if they met all of the following criteria: (1) confirmed diagnosis of SAP according to the revised Atlanta classification; (2) age between 18 and 70 years; (3) admission within 72 h of symptom onset; and (4) sonographic evidence of intra-abdominal fluid collection (defined as a fluid depth ≥ 1 cm or an estimated volume ≥ 50 mL). Patients were excluded if any of the following conditions were present: (1) preexisting severe chronic diseases affecting other major organ systems, such as heart, lung or kidney failure; (2) known contraindications to abdominal paracentesis, such as severe coagulopathy; (3) surgical intervention for SAP, such as laparoscopic or open surgery, within the first 7 days of admission; or (4) allergy to mirabilite or its components (Fig. 1).
Fig. 1.
Flowchart of study subject enrolment
Treatment allocation and interventions
Treatment allocation was nonrandomized and based on the clinical judgement of the attending physician, including imaging findings (extent of peripancreatic fluid collection, presence of infected collections), Ranson score at admission, and patient preference.
Group assignment
Patients were divided into three groups on the basis of the treatment they received:
Group A (combined therapy, n = 24): external mirabilite dressing plus peritoneal drainage
Group B (mirabilite alone, n = 31): external mirabilite dressing only
Group C (drainage alone, n = 25): peritoneal drainage only
Standard medical treatment
All enrolled patients received standard medical treatment upon admission, including fluid resuscitation, fasting, proton pump inhibitors for acid suppression, octreotide for the inhibition of pancreatic secretion, prophylactic or therapeutic antibiotics as indicated, and comprehensive nutritional support.
External application of mirabilite
Mirabilite (hydrated sodium sulfate, Na₂SO₄·10 H₂O; also known as Glauber’s salt) was applied as an external dressing. A cotton bag containing 1000 g of crystalline mirabilite powder (40 cm × 20 cm) was placed over the epigastric and periumbilical area, covering the pancreatic region. The dressing was replaced every 6 h. Treatment continued until the patient achieved any two of the following: (1) resolution of abdominal pain, (2) passage of flatus or stool, and (3) reduction in the IAP to < 12 mmHg.
Peritoneal drainage
Peritoneal drainage was performed when patients met any of the following criteria: (1) suspected infected peripancreatic fluid collection (fever > 38.5 °C, procalcitonin > 2.0 ng/mL, or positive Gram stain on diagnostic aspirate); (2) intra-abdominal hypertension refractory to medical management (IAP > 20 mmHg despite nasogastric decompression and prokinetics); or (3) symptomatic large fluid collection (> 5 cm) causing respiratory compromise or persistent pain.
The procedure was performed under ultrasound guidance using an 8–12 Fr pigtail catheter inserted in the periumbilical region or bilateral flanks connected to a closed drainage bag. Drainage was removed when the output was < 50 mL/day for two consecutive days and when the patient was clinically stable.
Outcomes
The primary outcome was the time to abdominal pain relief. All other outcomes were defined as secondary.
Clinical scores
Ranson and APACHE-II scores were both calculated on the basis of vital signs and laboratory test results at admission, with the APACHE-II score additionally assessed at discharge.
Symptom recovery
The time to flatus (days) and duration of abdominal pain (days) were recorded.
Laboratory parameters
The time to significant reduction (to two-thirds of peak values) and time to normalization were recorded for white blood cell (WBC) count, C-reactive protein (CRP) level, serum calcium level, and serum amylase (AMS) level.
IAP
IAP was measured at admission and at discharge using a bladder catheter technique.
Imaging findings
Computed tomography severity index (CTSI) scores were calculated on the basis of abdominal CT scans performed at admission and discharge. The area of the peripancreatic exudate (mm²) was measured, and the reduction in the exudate area was calculated as the admission value minus the discharge value. The pancreatic pseudocyst area (mm²) was also documented.
Complications
The following complications were recorded: intra-abdominal infection, pulmonary infection, septic shock, single or multiple organ failure, acidosis, pancreatic pseudocysts, pancreatic abscess, and abdominal compartment syndrome (ACS). Systemic inflammatory response syndrome (SIRS) duration was recorded on the basis of the following established criteria: (1) temperature > 38 °C or < 36 °C; (2) heart rate > 90 bpm; (3) respiratory rate > 20 breaths/min or PaCO₂ <4.3 kPa; and (4) WBC > 12 × 10⁹/L or < 4 × 10⁹/L or band neutrophils > 10%.
Health economic indicators
The length of hospital stay (days), length of ICU stay (days), and total hospitalization cost (CNY) were recorded.
Statistical analysis
Data analysis was conducted using SPSS 27.0. Continuous variables are presented as the median (interquartile range, IQR), and categorical variables are presented as n (%). Comparisons among the three groups were performed using the Kruskal‒Wallis test. When the overall test results were significant (p < 0.05), post hoc pairwise comparisons were conducted using Dunn’s test with Bonferroni correction, and the results were analysed using nonparametric tests. Categorical data were analysed using the chi-square test or Fisher’s exact test, as appropriate. To adjust for baseline confounding (notably the higher Ranson score in Group A), outcomes (time to abdominal pain relief, WBC normalization, IAP at discharge, exudate area reduction, pseudocyst area, SIRS duration) were analysed using quantile regression (median, τ = 0.5) adjusted for the admission Ranson score. Group A served as the reference category. P < 0.05 was considered to indicate statistical significance.
Results
General characteristics
A total of 80 patients diagnosed with SAP were enrolled. The baseline demographic and clinical characteristics, including sex, age, BMI, and aetiology, were compared among the three groups. No statistically significant differences were observed in any of the baseline variables (all P > 0.05; Table 1).
Table 1.
Baseline patient characteristics
| Variables | A (n = 24) | B (n = 31) | C (n = 25) | P |
|---|---|---|---|---|
| Age, (years) | 46.00 (37.50,58.25) | 44.00 (34.50,54.00) | 44.00 (33.00,49.00) | 0.312 |
| BMI, (kg/m²) | 25.68 (24.13,27.68) | 25.07 (22.77,27.67) | 27.64 (25.54,28.89) | 0.099 |
| Gender, n (%) | 0.080 | |||
| male | 17 (70.83) | 19 (61.29) | 10 (40.00) | |
| female | 7 (29.17) | 12 (38.71) | 15 (60.00) | |
| Aetiology, n (%) | 0.178 | |||
| Biliary | 10 (41.67) | 9 (29.03) | 9 (36.00) | |
| Lipoprotein disorders | 6 (25.00) | 12 (38.71) | 14 (56.00) | |
| Alcoholic | 3 (12.50) | 4 (12.90) | 0 (0.00) | |
| Other | 5 (20.83) | 6 (19.35) | 2 (8.00) |
Kruskal‒Wallis test, Chi-Square test, Monte Carlo simulation
M Median, Q₁ 1st Quartile, Q₃ 3rd Quartile
Clinical scores
The Ranson score differed significantly among the three groups (P = 0.026). Post hoc analysis with Bonferroni correction revealed that Group A had a significantly higher Ranson score than Group B did (adjusted P = 0.044), indicating a more severe initial disease state in Group A. No significant difference was observed between Group A and Group C (adjusted P = 0.503) or between Group B and Group C (adjusted P = 1.000; Table 2).
Table 2.
Comparison of clinical scores
| Variable | A (n = 24) | B (n = 31) | C (n = 25) | P | adjusted P |
|---|---|---|---|---|---|
| Ranson score (points) | 4.00 (3.00, 5.25) | 3.00 (2.00, 4.00) | 4.00 (2.00, 4.00) | 0.026 | A vs. B: P = 0.044 |
| A vs. C: P = 0.503 | |||||
| B vs. C: P = 1.000 | |||||
| APACHE-II (admission) (points) | 8.00 (5.00, 10.50) | 8.00 (6.50, 10.00) | 8.00 (5.00, 11.00) | 0.841 | |
| APACHE-II (discharge) (points) | 4.00 (3.00, 6.00) | 4.00 (2.00, 5.00) | 3.00 (2.00, 5.00) | 0.490 |
Kruskal‒Wallis test, Dunn’s post hoc test with Bonferroni correction
M Median, Q₁ 1st Quartile, Q₃ 3rd Quartile
Pairwise comparisons were performed only when the overall Kruskal‒Wallis test indicated significance (p < 0.05). Adjusted P < 0.05 was considered statistically significant
With respect to the APACHE-II scores, no significant differences were found among the three groups at either admission (P = 0.841) or discharge (P = 0.490).
Symptom recovery
As shown in Table 3, the primary outcome of time to abdominal pain relief was significantly shorter in Group A than in Group B (median: 7.50 vs. 11.00 days, p = 0.016). For the secondary outcome of time to flatus, a difference was observed between Group A and Group B (median: 2.00 vs. 3.00 days, p = 0.042).
Table 3.
Comparison of symptom recovery time
| Variable | A (n = 24) | B (n = 31) | C (n = 25) | P |
|---|---|---|---|---|
| Time to flatus (days) | 2.00 (1.00,2.25) | 3.00 (2.00,4.00) | 2.00 (1.00,3.00) | 0.042 |
| Time to abdominal pain relief (days) | 7.50 (4.75,11.25) | 11.00 (9.00,13.50) | 9.00 (6.00,10.00) | 0.016 |
Kruskal‒Wallis test
M Median, Q₁ 1st Quartile, Q₃ 3rd Quartile
To further adjust for potential confounding by baseline disease severity, quantile regression (median, τ = 0.5) was performed with Group A as the reference category (Table 4). After controlling for the Ranson score, compared with Group A, Group B remained associated with a significantly longer time to abdominal pain relief (coefficient = 4.667, 95% CI 2.208–7.125; P < 0.001). Compared with Group A, Group C also had a significantly longer time to abdominal pain relief (coefficient = 2.667, 95% CI 0.134–5.199; P = 0.039). A higher Ranson score was independently associated with prolonged pain relief (coefficient = 0.667; p = 0.026).
Table 4.
Quantile regression (median, τ = 0.5) for symptom recovery time
| Parameter | Coefficient | 95% CI | P |
|---|---|---|---|
| Time to flatus (days)1 | |||
| Intercept (Group A) | 2.000 | 0.705–3.295 | 0.003 |
| Ranson score | 0.000 | -0.250–0.250 | 1.000 |
| Group B vs. Group A | 1.000 | -0.054–2.054 | 0.063 |
| Group C vs. Group A | 0.000 | -1.085–1.085 | 1.000 |
| Time to abdominal pain relief (days)2 | |||
| Intercept (Group A) | 4.333 | 1.311–7.355 | 0.006 |
| Ranson score | 0.667 | 0.083–1.251 | 0.026 |
| Group B vs. Group A | 4.667 | 2.208–7.125 | < 0.001 |
| Group C vs. Group A | 2.667 | 0.134–5.199 | 0.039 |
Group A served as the reference category
CI Confidence interval
1Pseudo R² = 0.071, MAE = 1.150
2Pseudo R² = 0.081, MAE = 3.021. Both model adjusted for Ranson score
In contrast, the time to flatus did not differ significantly between groups after adjustment (Group B vs. A: coefficient = 1.000, p = 0.063; Group C vs. A: coefficient = 0.000, p = 1.000), suggesting that the unadjusted difference was largely explained by baseline severity.
Laboratory parameters
The time to WBC normalization differed among the three groups (P = 0.049), with Group A showing a shorter median recovery time (3.00 days) than Group B (6.00 days) and Group C (5.00 days). No significant differences were observed among the three groups for the other laboratory parameters (all P > 0.05; Table 5).
Table 5.
Comparison of improvement times for laboratory parameters
| Variables | A (n = 24) | B (n = 31) | C (n = 25) | P |
|---|---|---|---|---|
| Time to WBC significant decrease (days) | 2.50 (1.75,4.00) | 3.00 (2.00,4.50) | 3.00 (2.00,5.00) | 0.618 |
| Time to WBC normalization (days) | 3.00 (2.00,6.00) | 6.00 (3.00,8.50) | 5.00 (3.00,9.00) | 0.049 |
| Time to CRP significant decrease (days) |
3.50 (2.75, 6.25) |
3.00 (2.00,4.00) | 4.00 (2.00,6.00) | 0.149 |
| Time to CRP normalization (days) |
7.50 (5.75, 11.75) |
8.00 (5.00,14.00) | 10.00 (6.00,13.00) | 0.700 |
| Time to calcium significant recovery (days) | 4.00 (2.00,6.00) | 3.00 (1.00,4.00) | 3.00 (2.00,3.00) | 0.062 |
| Time to calcium normalization (days) | 5.50 (4.00,9.25) | 7.00 (4.50,11.00) | 4.00 (3.00,6.00) | 0.080 |
| Time to AMS significant decrease (days) | 2.00 (1.00,2.25) | 2.00 (1.00,2.50) | 2.00 (1.00,2.00) | 0.889 |
| Time to AMS normalization (days) | 3.00 (2.00,4.00) | 3.00 (1.50,4.00) | 3.00 (1.00,4.00) | 0.536 |
Kruskal‒Wallis test
M Median, Q₁ 1st Quartile, Q₃ 3rd Quartile
To further adjust for baseline disease severity, quantile regression (median, τ = 0.5) was performed for time to WBC normalization, with Group A as the reference category (Table 6). After controlling for the Ranson score, compared with Group A, Group B remained associated with a significantly longer time to WBC normalization (coefficient = 3.000, 95% CI 0.366–5.634; P = 0.026). The difference between Group C and Group A did not reach statistical significance (coefficient = 2.000, 95% CI -0.713–4.713, P = 0.146). The Ranson score was not significantly associated with time to WBC normalization (coefficient = 0.000, 95% CI -0.626–0.626; P = 1.000).
Table 6.
Quantile regression (median, τ = 0.5) for time to WBC normalization
| Parameter | Coefficient | 95% CI | P |
|---|---|---|---|
| Intercept (Group A) | 3.000 | -0.238–6.238 | 0.069 |
| Ranson score | 0.000 | -0.626–0.626 | 1.000 |
| Group B vs. Group A | 3.000 | 0.366–5.634 | 0.026 |
| Group C vs. Group A | 2.000 | -0.713–4.713 | 0.146 |
Group A served as the reference category
CI Confidence interval
Pseudo R² = 0.064, MAE = 2.725. Model adjusted for Ranson score
IAP
Baseline IAP at admission was comparable among the three groups (P = 0.227). Compared with that at admission, the IAP decreased significantly in all three groups (all P < 0.05). A suggestive significant difference in the IAP at discharge was observed among the groups (P = 0.032), with Group A showing a lower median IAP (7.35 mmHg) than Group B (8.46 mmHg) and Group C (8.82 mmHg) did (Table 7). Importantly, this lower discharge IAP in Group A occurred despite its higher baseline Ranson score (indicating more severe disease).
Table 7.
Comparison of intra-abdominal pressure
| Variables | A (n = 24) | B (n = 31) | C (n = 25) | P |
|---|---|---|---|---|
| IAP (admission) (mmHg) | 13.24 (11.03,18.02) | 11.76 (10.29,13.79) | 13.24 (11.03,15.44) | 0.227 |
| IAP (discharge) (mmHg) | 7.35 (5.88,8.82) | 8.46 (7.35,9.75) | 8.82 (6.62,10.29) | 0.032 |
Kruskal‒Wallis test
M Median, Q₁ 1st Quartile, Q₃ 3rd Quartile
To adjust for baseline severity, quantile regression (median, τ = 0.5) was performed for the IAP at discharge, with Group A as the reference category (Table 8). After controlling for the Ranson score, compared with Group A, Group C remained associated with a higher IAP at discharge (coefficient = 1.957, 95% CI 0.361–3.552; P = 0.017). The difference between Group B and Group A did not reach statistical significance (coefficient = 1.227, 95% CI -0.322–2.775; P = 0.119). The Ranson score was not significantly associated with the IAP at discharge (coefficient = 0.243, 95% CI -0.125–0.611; P = 0.192).
Table 8.
Quantile regression (median, τ = 0.5) for the IAP at discharge
| Parameter | Coefficient | 95% CI | P |
|---|---|---|---|
| Intercept (Group A) | 6.377 | 4.473–8.281 | < 0.001 |
| Ranson score | 0.243 | -0.125–0.611 | 0.192 |
| Group B vs. Group A | 1.227 | -0.322–2.775 | 0.119 |
| Group C vs. Group A | 1.957 | 0.361–3.552 | 0.017 |
Group A served as the reference category
CI Confidence interval
Pseudo R² = 0.086, MAE = 1.533. Model adjusted for Ranson score
Imaging findings
As shown in Table 9, the baseline exudate area differed significantly among the three groups (p = 0.010), with Group A having a larger baseline area than Group B did. No significant differences were observed in the baseline CTSI scores (p = 0.164), discharge CTSI scores (p = 0.572), or discharge exudate area (p = 0.168).
Table 9.
Comparison of imaging findings
| Variables | A (n = 24) | B (n = 31) | C (n = 25) | P (between-group) |
|---|---|---|---|---|
| CTSI score (admission) (points) | 6.00 (4.00,6.50) | 4.00 (4.00,6.00) | 4.00 (4.00,4.00) | 0.164 |
| CTSI score (discharge) (points) | 4.00 (2.00,4.50) | 4.00 (2.00,4.00) | 4.00 (3.00,6.00) | 0.572 |
| P (within-group) | 0.029 | 0.037 | 0.041 | |
| Exudate area (admission) (mm2) | 1686.92 (1255.87,2615.02) | 1226.59 (1082.11,1520.88) | 1860.22 (1105.27,2209.64) | 0.010 |
| Exudate area (discharge) (mm2) | 1036.60 (828.05,1424.49) | 1126.98 (761.43,1402.18) | 1215.63 (922.54,1776.87) | 0.168 |
| P (within-group) | < 0.001 | < 0.001 | 0.008 | |
| Exudate area reduction (mm2) | 741.06 (363.80,1333.21) | 164.03 (59.66,420.24) | 494.20 (7.85,752.16) | 0.001 |
Kruskal‒Wallis test, Wilcoxon signed-rank test for paired comparisons. Reduction = admission value - discharge value
M Median, Q₁ 1st Quartile, Q₃ 3rd Quartile
Importantly, despite having a larger baseline exudate area, Group A showed the greatest reduction in exudate area (admission minus discharge), with a statistically suggestive difference among the groups (p = 0.001): the median reduction was 741.06 mm² in Group A vs. 164.03 mm² in Group B and 494.20 mm² in Group C.
Within-group analyses revealed significant reductions in the exudate area from admission to discharge in all three groups (all P < 0.01; Table 10). With respect to CTSI scores, decreases from admission to discharge were observed in all three groups (Group A: P = 0.029; Group B: P = 0.037; Group C: P = 0.041).
Table 10.
Quantile regression (median, τ = 0.5) for the Exudate area
| Parameter | Coefficient | 95% CI | P |
|---|---|---|---|
| Exudate area (admission) (mm2)1 | |||
| Intercept (Group A) | 1585.696 | 1071.548–2099.844 | < 0.001 |
| Ranson score | 36.582 | -62.760–135.924 | 0.466 |
| Group B vs. Group A | -406.776 | -824.993–11.441 | 0.056 |
| Group C vs. Group A | 194.345 | -236.461–625.150 | 0.372 |
| Exudate area reduction (mm2)2 | |||
| Intercept (Group A) | 496.526 | 45.824–947.228 | 0.031 |
| Ranson score | 35.394 | -51.688–122.477 | 0.421 |
| Group B vs. Group A | -427.186 | -793.795 – -60.576 | 0.023 |
| Group C vs. Group A | -214.688 | -592.332–162.957 | 0.261 |
Group A served as the reference category
CI Confidence interval
1Pseudo R² = 0.114, MAE = 536.999. Models were adjusted for Ranson score
2Pseudo R² = 0.106, MAE = 410.667. Models were adjusted for Ranson score
To further adjust for baseline disease severity, quantile regression (median, τ = 0.5) was performed for the exudate area parameters, with Group A as the reference category (Table 10). After controlling for the Ranson score, compared with Group A, Group B demonstrated a smaller reduction in exudate area (coefficient = -427.186, 95% CI -793.795 to -60.576; P = 0.023). No significant difference was observed between Group C and Group A in terms of reduced exudate area (coefficient = -214.688, 95% CI -592.332 to 162.957, P = 0.261). The Ranson score was not significantly associated with any of the exudate area parameters (all P > 0.05).
Complications
The incidence of major local and systemic complications was monitored throughout hospitalization and during a 2-month follow-up period. No statistically significant differences were observed among the three groups for most complications (all P > 0.05; Table 11). However, for ACS, a trend toward lower incidence was observed in Group C (16.00%) compared to Group B (45.16%), although the difference did not reach statistical significance (P = 0.065).
Table 11.
Comparison of complication rates
| Variables | A (n = 24) | B (n = 31) | C (n = 25) | P |
|---|---|---|---|---|
| Systemic complications, n (%) | ||||
| Intra-abdominal infection | 5 (20.83) | 8 (25.81) | 6 (24.00) | 0.949 |
| Pulmonary infection | 4 (16.67) | 4 (12.90) | 3 (12.00) | 0.848 |
| Septic shock | 1 (4.17) | 3 (9.68) | 1 (4.00) | 0.628 |
| Single organ failure | 4 (16.67) | 7 (22.58) | 5 (20.00) | 0.939 |
| Multiple organ failure | 5 (20.83) | 4 (12.90) | 4 (16.00) | 0.701 |
| Acidosis | 3 (12.50) | 3 (9.68) | 6 (24.00) | 0.388 |
| Local complications, n (%) | ||||
| Pancreatic pseudocysts | 21 (87.50) | 28 (90.32) | 22 (88.00) | 1.000 |
| Pancreatic abscess | 2 (8.33) | 1 (3.23) | 1 (4.00) | 0.680 |
| ACS | 7 (29.17) | 14 (45.16) | 4 (16.00) | 0.065 |
Chi-square test or Fisher’s exact test. For variables with expected cell frequencies below 5, Monte Carlo simulation (10,000 samples) was used
The high incidence of pancreatic pseudocysts (87.5%–90.3%) in this cohort likely reflects the inclusion criteria, which required sonographic evidence of fluid collection at admission. This finding indicates that a selected patient population with a substantial baseline peripancreatic exudate burden is predisposed to pseudocyst formation.
Among patients who developed pancreatic pseudocysts (21 in Group A, 28 in Group B, and 22 in Group C), a difference in the pseudocyst area was observed among the three groups (P = 0.019). Compared with Group B (498.10 mm²) and Group C (488.55 mm²), Group A had the smallest median cyst area (343.39 mm²) (Table 12).
Table 12.
Comparison of pseudocyst areas
| Variables | A (n = 21) | B (n = 28) | C (n = 22) | P |
|---|---|---|---|---|
| Pancreatic pseudocyst area (mm2) | 343.39 (214.41,450.38) | 498.10 (390.36,840.21) | 488.55 (366.52,576.51) | 0.019 |
Kruskal–Wallis test
M Median, Q₁ 1st Quartile, Q₃ 3rd Quartile
Only patients who developed pseudocysts were included in this analysis. Actual numbers: A: n = 21 (87.5% of 24), B: n = 28 (90.3% of 31), C: n = 22 (88.0% of 25)
After quantile regression adjustment for the Ranson score (Table 13), compared with Group A, both Group B and Group C remained associated with larger pseudocyst areas (Group B vs. A: coefficient = 215.567, 95% CI 42.999–388.135, p = 0.015; Group C vs. A: coefficient = 178.664, 95% CI 2.709–354.619, p = 0.047). The Ranson score was not significantly associated with the pseudocyst area (p > 0.05).
Table 13.
Quantile regression (median, τ = 0.5) for pseudocyst areas
| Parameter | Coefficient | 95% CI | P |
|---|---|---|---|
| Intercept (Group A) | 226.812 | 6.038–447.586 | 0.044 |
| Ranson score | 23.951 | -18.786–66.688 | 0.267 |
| Group B vs. Group A | 215.567 | 42.999–388.135 | 0.015 |
| Group C vs. Group A | 178.664 | 2.709–354.619 | 0.047 |
Group A served as the reference category. Only patients who developed pseudocysts were included in this analysis
CI Confidence interval
Pseudo R² = 0.027, MAE = 561.461. Model adjusted for Ranson score
All the enrolled patients developed SIRS during hospitalization. In an exploratory analysis, although the results of the Kruskal‒Wallis test revealed no significant difference in SIRS duration among the three groups (P = 0.073; Table 14), quantile regression adjusted for Ranson scores revealed that compared with Group A, Group B (coefficient = 2.600; 95% CI 0.493–4.707; P = 0.016) and Group C (coefficient = 2.400; 95% CI 0.230–4.570; P = 0.031) had longer SIRS durations (Table 15). However, given that the unadjusted analysis was non-significant (P = 0.073), these findings should be considered exploratory and interpreted with caution. Additionally, a higher Ranson score was associated with a longer duration of SIRS (P = 0.019).
Table 14.
Comparison of SIRS duration
| Variables | A (n = 24) | B (n = 31) | C (n = 25) | P |
|---|---|---|---|---|
| SIRS duration (days) | 5.00 (4.00,7.25) | 7.00 (5.50,8.50) | 7.00 (4.00,8.00) | 0.073 |
Kruskal‒Wallis test
M Median, Q₁ 1st Quartile, Q₃ 3rd Quartile
Table 15.
Quantile regression (median, τ = 0.5) for SIRS duration (exploratory analysis)
| Parameter | Coefficient | 95% CI | P |
|---|---|---|---|
| Intercept (Group A) | 2.400 | -0.190–4.990 | 0.069 |
| Ranson score | 0.600 | 0.099–1.101 | 0.019 |
| Group B vs. Group A | 2.600 | 0.493–4.707 | 0.016 |
| Group C vs. Group A | 2.400 | 0.230–4.570 | 0.031 |
Group A served as the reference category
CI Confidence interval
Pseudo R² = 0.065, MAE = 2.828. Model adjusted for Ranson score
Health economic indicators
No statistically significant differences were found among the three treatment groups regarding the length of hospital stay, duration of ICU stay, or total hospitalization costs (all P > 0.05; Table 16).
Table 16.
Comparison of health economic indicators
| Variables | A (n = 24) | B (n = 31) | C (n = 25) | P |
|---|---|---|---|---|
| Length of hospital stay (days) | 15.50 (9.75,21.50) | 11.00 (9.00,16.50) | 13.00 (10.00,18.00) | 0.208 |
| Length of ICU stay (days) | 0.00 (0.00,3.75) | 0.00 (0.00,4.00) | 0.00 (0.00,7.00) | 0.843 |
| Total cost (CNY) | 31443.16 (12971.25,62354.06) | 18475.51 (13679.97,47610.15) | 38039.39 (21691.28, 56630.69) | 0.318 |
Kruskal‒Wallis test. Costs are reported in CNY. Using the average exchange rate of 1 USD ≈ 7.2 CNY during the study period, the median total costs were approximately 4,367 USD (Group A), 2,566 USD (Group B), and 5,283 USD (Group C). These figures are specific to the Chinese healthcare setting
M Median, Q₁ 1st Quartile, Q₃ 3rd Quartile
Discussion
In this retrospective cohort study of SAP patients, compared with either intervention alone, combining external mirabilite dressing with abdominal drainage was associated with faster symptom recovery, lower discharge IAP, and a greater reduction in peripancreatic exudate (Tables 3, 4, 5, 6, 7, 8, 9 and 10). However, no significant differences were observed in major complications, hospital stay, ICU stay, or total costs (Tables 11 and 16). These findings suggest that combination therapy may provide short-term benefits, especially for controlling symptoms and reducing inflammation, but its effect on overall outcomes is still unknown and needs confirmation in future studies.
Our findings are largely consistent with those of multiple independent studies investigating combination strategies involving mirabilite. The use of mirabilite (Glauber’s salt, hydrous sodium sulfate) as a topical hyperosmotic agent has been documented in the gastroenterology and surgery literature. Prospective studies have demonstrated that the external application of Glauber’s salt significantly decreases the IAP, improves the APACHE II score, and restores intestinal motility by promoting fluid redistribution [9]. Furthermore, recent surgical evidence has indicated that topical mirabilite directly accelerates the absorption of abdominal fluid collections by creating a strong transdermal osmotic gradient [10, 11]. These biophysical properties strongly support our observation of a significantly greater reduction in peripancreatic exudate in the combined group.
Zeng et al. conducted a randomized trial of 234 paediatric patients who underwent ERCP and reported that external mirabilite application significantly reduced the incidence of post-ERCP pancreatitis and alleviated postprocedural pain. Notably, compared with the blank group, the mirabilite group had significantly lower TNF-α expression and significantly higher IL-10 expression at 24 h post-procedure. Although this study focused on PEP prophylaxis rather than SAP treatment, it provides robust evidence from a large multicentre sample that external mirabilite application effectively downregulates proinflammatory cytokines (TNF‑α) while upregulating anti‑inflammatory cytokines (IL‑10) [12]. This mechanistic profile directly supports our observation of accelerated normalization of inflammatory markers (Table 5).
More directly relevant to our study, a prospective study by Li et al. evaluated early ultrasound-guided percutaneous catheter drainage in patients with SAP accompanied by acute fluid collection. Compared with the conventional group, the observation group demonstrated a significantly shorter duration for the resolution of abdominal pain and faster improvement in clinical symptoms [13]. Although this study revealed these short-term clinical benefits, it aligns with our findings that surgical interventions in the early phase primarily improve physiological parameters and symptom control without necessarily altering extended clinical endpoints, thus reinforcing the value of combined adjunctive strategies in the early management of SAP.
The observed pattern of improvement is consistent with that of complementary mechanisms, although formal synergy was not tested. As a hyperosmotic agent composed of hydrated sodium sulfate (Na₂SO₄·10 H₂O), an external mirabilite dressing reduces tissue oedema by drawing fluid from inflamed tissues, which improves splanchnic microcirculation and downregulates inflammatory mediators such as TNF-α and IL-6 [14, 15]. Wu et al. demonstrated that mirabilite significantly reduced the serum levels of diamine oxidase and D-lactate, markers of intestinal permeability, after 7 days of treatment. The treatment also lowered IAP. These findings suggest that the primary benefit of mirabilite may come from preserving gut barrier function rather than from direct systemic anti-inflammatory effects [16].
Moreover, peritoneal drainage directly removes the PAAF. This fluid is a potent driver of IAH and systemic inflammation. It is rich in cytokines and activated pancreatic enzymes [17–20]. By reducing the IAP, drainage improves organ perfusion and prevents abdominal compartment syndrome [21]. Our finding that drainage alone was sufficient to lower the IAP (Table 7) but not to shorten the duration of SIRS (Table 14) supports this mechanistic distinction: drainage addresses the mechanical consequences of fluid accumulation, whereas mirabilite modulates local tissue inflammation and gut barrier function.
In general, the combination strategy removes the biochemical trigger (PAAF) and relieves IAH via drainage, whereas mirabilite reduces residual tissue oedema and protects gut barrier function. This two-pronged mechanism may explain why the combined treatment group outperformed both monotherapies across multiple short-term outcomes, including exudate reduction and abdominal pain relief (Table 3).
The optimal timing of percutaneous drainage in SAP remains debated. Current guidelines generally recommend deferring intervention until ≥ 4 weeks after onset to allow the encapsulation of necrotic collections, a strategy supported by evidence that delayed drainage is associated with lower rates of adverse events. However, some patients with suspected infected pancreatic necrosis and worsening organ dysfunction despite antibiotic therapy may benefit from early drainage [22, 23]. Our study involved early drainage in patients with substantial ascitic fluid, and our findings suggest that the combination of early drainage with mirabilite may offer clinical benefits in patients.
Beyond the interpretation of individual outcomes, our study has several potential strengths. First, the combined therapy showed consistent benefits across multiple domains, suggesting a potential multidimensional effect. Second, these advantages remained observable after adjustment for baseline severity, supporting the robustness of our findings despite the retrospective design. Third, the integration of external mirabilite with drainage offers a pragmatic strategy that warrants investigation in larger prospective studies.
A notable observation across our study and the literature is the consistent dissociation between improvements in short-term surrogate outcomes and the absence of demonstrable benefits in hard clinical endpoints. In the prospective cohort study by Huang et al., despite significant improvements in multiple short-term clinical indicators, no mortality benefit was demonstrated, as the study did not report differences in mortality between groups [6]. Similarly, a multicentre RCT by Ke et al. involving 120 patients with necrotizing pancreatitis and early persistent organ failure revealed that early catheter drainage for acute necrotic collections did not improve the composite primary outcome of major complications and death compared with standard delayed care. We hypothesize that several factors may explain this difference. First, our study (n = 80) and most existing studies are underpowered to detect differences in relatively rare events such as mortality or major organ failure. Second, the pathophysiology of SAP-related organ failure involves complex, multifactorial mechanisms beyond IAH and local inflammation, including microcirculatory dysfunction, mitochondrial injury, and immune paralysis. Third, the short-term nature of our follow-up may lead to overlooking later differences in complication rates or quality of life outcomes.
Other limitations should be acknowledged. This was a retrospective, nonrandomized study, which may introduce selection bias despite our efforts to adjust for baseline differences using quantile regression. Specifically, we adjusted only for the Ranson score, the only variable with a significant baseline imbalance (P = 0.026). Other baseline variables, including the APACHE-II score (P = 0.841), BMI (P = 0.099), sex (P = 0.080), and aetiology (P = 0.178), did not significantly differ. The baseline exudate area was numerically greater in Group A than in Group B (P = 0.056) and comparable between Group A and Group C (P = 0.372). These non-significant differences may reflect limited statistical power rather than true equivalence. Given our modest sample size (n = 80), including multiple covariates would risk overfitting. Although we adjusted for the Ranson score using quantile regression, residual confounding and confounding by indication cannot be fully excluded. Moreover, although we applied the Bonferroni correction for pairwise comparisons, the analysis of multiple secondary outcomes carries a potential risk of type I error inflation.
In conclusion, compared with either intervention alone, the combination of external mirabilite dressing and abdominal drainage in early-stage SAP was associated with short-term clinical benefits without increasing total hospitalization costs. However, no significant differences were observed in terms of major complications. Given the retrospective design and modest sample size, these findings are hypothesis-generating and require validation in prospective, multicentre randomized controlled trials.
Acknowledgements
None.
Abbreviations
- SAP
Severe acute pancreatitis
- SIRS
Systemic inflammatory response syndrome
- MODS
Multiple organ dysfunction syndrome
- PAAF
Pancreatitis-associated ascitic fluid
- IAH
Intra-abdominal hypertension
- IAP
Intra-abdominal pressure
- APACHE-II
Acute Physiology and Chronic Health Evaluation
- WBC
White blood cell
- CRP
C-reactive protein
- AMS
Amylase
- CTSI
Computed tomography severity index
- BMI
Body mass index
- ACS
Abdominal compartment syndrome
- TNF
Tumour necrosis factor
- IL
Interleukin
Authors' contributions
Jing-xian Zhang conceived this review, collected the data and wrote the main manuscript text. Han-shu Jin, Shuang-shuang Lu and Yu-hui Sun performed the literature search, data collection and data interpretation. Kuang-jing Wang and Jin Huang contributed to the study conception and reviewed and revised the manuscript for important intellectual content. All the authors reviewed and approved the manuscript.
Funding
The author acknowledges financial support from the Science and Technology Project of Changzhou Municipal Health Commission (No. QN202447).
Data availability
The data supporting this study are available from the corresponding author upon reasonable request.
Declarations
Ethics approval and consent to participate
All the experiments were carried out in accordance with the Helsinki Declaration. This study protocol was approved by the Ethics Committee of Changzhou Second People’s Hospital, No. [2025] KY310-01. We registered this study at the Chinese Clinical Trial Registry on March 13, 2026 (No. ChiCTR2600120445).
Not applicable. The need for patient consent was waived because of the retrospective nature of the study.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Contributor Information
Kuang-jing Wang, Email: kuangjing_wang@163.com.
Jin Huang, Email: hj042153@hotmail.com.
References
- 1.Lankisch PG, Apte M, Banks PA. Acute pancreatitis. Lancet. 2015;386(9988):85–96. [DOI] [PubMed] [Google Scholar]
- 2.Calomino N, Kelmendi E, Poto GE, Carbone L, Zanchetta M, Marrelli D. The role of calcium salts in pigment gallstones and their spiculated morphology. Gastrointest Disorders. 2026;8(1):2. [Google Scholar]
- 3.Li CL, Jiang M, Pan CQ, Li J, Xu LG. The global, regional, and national burden of acute pancreatitis in 204 countries and territories, 1990–2019. BMC Gastroenterol. 2021;21(1):332. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Paragomi P, Tuft M, Pothoulakis I, Singh VK, Stevens T, Nawaz H, et al. Dynamic changes in the pancreatitis activity scoring system during hospital course in a multicenter, prospective cohort. J Gastroenterol Hepatol. 2021;36(9):2416–23. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Samanta J, Rana A, Dhaka N, Agarwala R, Gupta P, Sinha SK, et al. Ascites in acute pancreatitis: not a silent bystander. Pancreatology. 2019;19(5):646–52. [DOI] [PubMed] [Google Scholar]
- 6.Huang J, Li L, Chen Y, Mao E, Qu H. Early short-term abdominal paracentesis drainage in moderately severe and severe acute pancreatitis with pelvic ascites. BMC Surg. 2023;23(1):363. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Tao L, Fu J, Wang F, Song Y, Li Y, Zhang J, et al. The application of mirabilite in traditional Chinese medicine and its chemical constituents, processing methods, pharmacology, toxicology and clinical research. Front Pharmacol. 2023;14:1293097. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Wang Y, Zhang X, Li C. Applying hot compresses with rhubarb and mirabilite to reduce pancreatic leakage occurrence in the treatment of severe acute pancreatitis. Iran J Public Health. 2017;46(1):136–8. [PMC free article] [PubMed] [Google Scholar]
- 9.Zhang MJ, Zhang GL, Yuan WB, Ni J, Huang LF. Treatment of abdominal compartment syndrome in severe acute pancreatitis patients with traditional Chinese medicine. World J Gastroenterol. 2008;14(22):3574–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Li J, Hua J, Ruan H, Xu H, Liang C, Meng Q, et al. Topical application of Glauber’s salt accelerates the absorption of abdominal fluid after pancreatectomy. BMC Surg. 2024;24(1):398. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Cai H, Du J, Luo C, Li S. External application of mirabilite before surgery can reduce the inflammatory response and accelerate recovery in mild acute biliary pancreatitis. BMC Gastroenterol. 2023;23(1):264. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Zeng JQ, Zhang TA, Yang KH, Wang WY, Zhang JY, Hu YB, et al. External use of mirabilite to prevent post-endoscopic retrograde cholangiopancreatography pancreatitis in children: A multicenter randomized controlled trial. World J Gastrointest Surg. 2023;15(4):712–22. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Li H, Wu Y, Xu C, An H, Guo C, Cui H. Early ultrasound-guided percutaneous catheter drainage in the treatment of severe acute pancreatitis with acute fluid accumulation. Exp Ther Med. 2018;16(3):1753–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Lee PJ, Papachristou GI. New insights into acute pancreatitis. Nat Rev Gastroenterol Hepatol. 2019;16(8):479–96. [DOI] [PubMed] [Google Scholar]
- 15.Garg PK, Singh VP. Organ failure due to systemic injury in acute pancreatitis. Gastroenterology. 2019;156(7):2008–23. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Wu J, Jin Y, Chen Y, Lu J. A clinical research of external application of mirabilite on intestinal mucosal barrier and prognosis in patients with acute pancreatitis. Chin J Integr Traditional Western Med Intensive Crit Care. 2017;24(6):613–6. [Google Scholar]
- 17.Luo C, Huang Q, Yuan X, Yang Y, Wang B, Huang Z, et al. Abdominal paracentesis drainage attenuates severe acute pancreatitis by enhancing cell apoptosis via PI3K/AKT signaling pathway. Apoptosis. 2020;25(3–4):290–303. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Rogers WK, Garcia L, Intraabdominal hypertension. Abdominal compartment syndrome, and the open abdomen. Chest. 2018;153(1):238–50. [DOI] [PubMed] [Google Scholar]
- 19.Zerem E, Zerem D, Vila Š, Bajgorić S. Ascites in acute pancreatitis: a window into disease severity. World J Gastroenterol. 2025;31(43):112797. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Rao JW, Li JR, Wu Y, Lai TM, Zhou ZG, Gong Y, et al. Ascites characteristics in acute pancreatitis: a prognostic indicator of organ failure and mortality. World J Gastroenterol. 2025;31(28):108926. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Lu Z, Zhu X, Hua T, Zhang J, Xiao W, Jia D, et al. Efficacy and safety of abdominal paracentesis drainage on patients with acute pancreatitis: a systematic review and meta-analysis. BMJ Open. 2021;11(8):e045031. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Sugimoto M, Sonntag DP, Flint GS, Boyce CJ, Kirkham JC, Harris TJ, et al. Better outcomes if percutaneous drainage is used early and proactively in the course of necrotizing pancreatitis. J Vasc Interv Radiol. 2016;27(3):418–25. [DOI] [PubMed] [Google Scholar]
- 23.Wen Y, Zhuo WQ, Liang HY, Huang Z, Cheng L, Tian FZ, et al. Abdominal paracentesis drainage improves outcome of acute pancreatitis complicated with intra-abdominal hypertension in early phase. Am J Med Sci. 2023;365(1):48–55. [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 supporting this study are available from the corresponding author upon reasonable request.