Role of Early Percutaneous Drainage of Symptomatic Acute Fluid Collections in Moderately Severe and Severe Acute Pancreatitis: A Randomised Study

Abstract

Objective:

The step-up approach for severe acute pancreatitis (SAP) has gained considerable traction. However, there is no consensus on percutaneous drainage (PCD) of early (<4 weeks) acute fluid collections (AFCs), some of which may produce mass effects and worsening inflammation. We designed a prospective randomised study to determine the role of PCD in symptomatic AFCs in SAP.

Materials and Methods:

Our prospective randomised study divided patients of moderately SAP (MSAP) and SAP with symptomatic (mass effect, fever, anorexia, and poor organ scores) AFCs into two groups:– Group 1 for conservative management (CM) and Group 2 for PCD. Requisite ethics and scientific clearances were obtained. The primary outcome variable was hospital stay; some other recorded variables included mortality, organ scores, sequential organ failure assessment (SOFA), and intra-abdominal pressure (IAP).

Results:

Of 78 patients with acute pancreatitis in 18 months, 38 (49%) fulfilled the inclusion criteria. Group 1 or CM had 20, and Group 2 or PCD had 18 patients; 16 from each group were eventually compared. No significant differences were observed in any clinical or outcome parameter except better day 7 SOFA scores and IAP in Group 2 (SOFA: 1.69 ± 1.55 vs. 3.33 ± 1.35, P = 0.013; IAP in mm Hg: 9.29 ± 1.82 vs. 13.81 ± 2.71, P < 0.001). Overall, four patients (10.5%) underwent necrosectomy, while mortality and hospital stay did not differ significantly between the groups.

Conclusion:

PCD in appropriately selected patients of MSAP and SAP with AFCs can hasten recovery, slow down organ damage, and decrease IAP significantly. More studies with a larger sample are needed in this area.

Introduction

The worldwide incidence of acute pancreatitis (AP) is rising steadily, with about 3040 cases per 100,000 population per year.[1] There are notable differences over geographical areas with respect to incidence and aetiology; however, the two most common causes remain cholelithiasis and ethanol abuse. Cholelithiasis is rising as the most common aetiology in the West, and in India.[2,3] The underlying mechanism of injury in AP is premature activation of pancreatic enzymes and auto-digestion. Subsequent inflammation results in pancreatic oedema, haemorrhage, and ultimately necrosis. The spectrum of local and systemic injury is vast, prompting a standardised stratification of severity – the modified Atlanta system which is still relevant today.[4] Most series still report a high morbidity of almost 50% and mortality of up to 20% in moderately severe and severe AP (MSAP and SAP).[1,3]

Amongst local complications, acute fluid collections (AFCs) and subsequent pseudocyst formation (>4 weeks) are the most frequent, starting in 30–50% of patients within 48 h of onset. Most of these collections are in the lesser sac and the anterior pararenal space, infrequently tracking into the abdominal cavity, pelvis, or mediastinum. The traditional management of AFCs has been conservative till at least 4 weeks since most are expected to resolve. Even subsequent pseudocysts after 4 weeks mostly resolve slowly in another 6 weeks. Conservative management (CM) is multidisciplinary, involving gastroenterologists, interventional radiologists, and surgeons.[5,6] This traditional approach was intended to minimise further injury in an acute inflammatory setting.

Over the last decade, some modifications have been suggested in this approach. Some authors believe that the 4-week cut-off is a bit arbitrary and abrupt. This is because of demonstrable partial or complete encapsulation of the AFC in the third week.[7] Also, AFCs may be symptomatic, with mass effects like bowel or biliary obstruction, persistent systemic inflammatory response syndrome (SIRS), anorexia, intractable pain, weight loss, or, abdominal compartment syndrome and organ failure. These may be hard to manage by an inflexible and dogmatic approach. With the successful adoption of the step-up approach for drainage of pancreatic necrosis worldwide after the multi-centre pancreatitis, necrosectomy versus step up approach (PANTER) trial, a plethora of options are available to drain even symptomatic AFCs.[7,8,9] In the step-up approach, percutaneous catheter drainage (PCD) under ultrasound/computer tomography (CT) guidance is the first modality. Recently, Zhang et al.[10] and Wang et al.[11] hypothesised that PCD for AFCs may decrease intra-abdominal pressure (IAP) and/or inflammatory mediators in the AFCs. Their studies showed a good success rate; however, both were retrospective in design. One of the few prospective series on the subject from Chandigarh, India, recruited 101 consecutive patients of AP demonstrated an initial PCD success rate of 67% and showed that volume reduction and organ failure resolution predicted PCD success.[12] There have been other recent encouraging results too, albeit with some concerns about infection and fistula formation in the PCD tract.[13,14] Our clinical experiences with many difficult symptomatic AFCs prompted us to design this prospective randomised study to determine the role of early PCD of symptomatic AFCs in improving outcomes in MSAP and SAP.

Materials and Methods

A prospective randomised study was designed in the Department of Surgery of our tertiary teaching institute in north India between August 2022 and January 2024 (18 months). Due approval was obtained from the institutional ethics committee (IEC-TP (MD\MS) 127\2022)/IEC\ABVIMS\RMLH\970 dated July 2022 and the institutional review board, as per the Declaration of Helsinki guidelines. Patients above 18 years of age presenting to our hospital with a diagnosis of MSAP or SAP, as per the modified Atlanta classification,[4] were recruited in the study after written informed consent. Patients with AP of more than 4 weeks duration, or, with a previous intervention were excluded from the study.

After a clinical/biochemical/radiological diagnosis of MSAP or SAP, patients were admitted to a high-dependency unit and administered CM with oxygen therapy, antibiotics, intravenous fluids, blood products if required, and analgesics. Ultrasound abdomen on admission and serially was performed in all patients, while a CT scan was performed at admission only for diagnostic uncertainty and later after a minimum of 4 days after diagnosis. The severity of illness was gauged with modified Marshall, bedside index of severity in AP, sequential organ failure assessment (SOFA), and CT severity index (CTSI) scores. We then identified MSAP or SAP patients with AFCs on imaging with persistent symptoms/signs, viz. mass effects (bowel or biliary obstruction), persistent SIRS, anorexia, intractable pain, fever, weight loss, abdominal compartment syndrome, worsening organ scores, or any parameter suggestive of poor treatment response in the clinician’s opinion. These patients were randomised into two groups using computer-generated random number tables – Group 1 or the control group where patients were managed conservatively (CM group) and Group 2 or the study group which was planned for PCD in the AFCs (PCD group). The sealed envelope technique was used for allocation concealment and to avoid bias in the initial allotment. Subsequent blinding was not possible due to the nature of the interventions planned. The primary outcome variable was total hospital stay, and the secondary variable was mortality.

In the study group, patients underwent PCD with the ultrasound guidance of a 2.5 MHz frequency linear probe under local anaesthesia in the radiology department. CT guidance was also used as per necessity, as was upgradation in management by multiple drainage tubes. Fluid accumulation in the space most adjacent to the abdominal wall was determined. The presence of any part of the intestine, stomach, or liver between the sensor and the fluid accumulation was considered a contraindication for puncture. After achieving a safe window, the tip of the puncture needle was inserted with a “hands-free” technique which allowed constant monitoring of the needle position. Then, with the Seldinger technique and guidewires, a pigtail catheter of 16 Fr or abdominal drain of 28–32 Fr was used for drainage (the largest size deemed accessible was chosen). In this study, PCD was performed for ANY symptomatic intra-abdominal collection and not just lesser sac AFCs. Symptomatic patients with respiratory distress due to pleural effusion underwent chest tube insertion with 28–32 Fr and were included in the study. The drain fluid was sent for culture, fluid amylase, and total counts. Sequential clinical and biochemical monitoring continued after the procedure, and further step-up procedures (videoscopic drainage or necrosectomy) were performed if indicated.

In the control group (as well as the study group), CM included intravenous fluids, monitoring, adequate analgesia, oxygen therapy, broad-spectrum antibiotics, gastrointestinal decompression, initial fasting then gradual enteral or parenteral nutrition, blood gas analysis to guide fluid and other therapies, and life support like haemodialysis or mechanical ventilation if required. If patients in the control group had no improvement or worsening of the symptomatic AFCs after 72 h and had a “drainable” AFC, they were offered drainage by the clinical team. Conversely, if patients in the study group did not have a safe window for drainage, they were continued on CM. Both groups also underwent serial evaluation with IAP measurement using a urinary catheter connected to a water manometer bedside.

Statistical analysis

Descriptive statistics, like mean and standard deviation or median with interquartile range, were used for continuous variables, and they were compared between the two groups using the Student t-test for parametric data or Mann–Whitney U test for non-parametric data. Categorical variables were depicted by numbers and proportions and compared using the chi-square test for parametric data or Fischer’s exact test for non-parametric data. A P value of less than 0.05 was taken as significant.

Assuming a population variance of 49 (standard deviation = 7 units) and an anticipated effect size difference of at least 1 week (mean hospital stay difference of 7 days between the two groups), to achieve a power of 80% and a level of significance of 5% (two-sided), the study was calculated to require a sample size of 32, that is, 16 for each group, assuming equal groups.

Results

During the period of the study, a total of 78 patients of AP were admitted to the units conducting the study. Of these, 49 (63%) patients were MSAP or SAP, and 38 of the 49 (77.5%) fulfilled the inclusion criteria of the study. After randomisation, the patients were divided into two groups; Group 1 or CM group had 20 patients, while Group 2 or PCD group had 18 patients. The mean duration of symptoms that qualified the patients into the study (mass effects due to bowel or biliary obstruction, persistent SIRS, anorexia, intractable pain, fever, weight loss, abdominal compartment syndrome, and worsening organ scores) was 7.2 days. Contrast CT scan was performed in 32/38 (84%) patients on an average of 5.2 days after admission. Six patients did not undergo CT scans due to early intervention, mortality, or logistics reasons. The CT findings were suggestive of a mean CTSI of 8.4 (range 4–10), lesser sac AFCs in 16/32 (50%) patients, other AFCs in 6/32 (19%), necrosis in 10/32 (31%), and generalised ascites in three patients [Figure 1].

Figure 1.

Symptomatic acute fluid collections (AFCs, <4 weeks after acute pancreatitis) in varying circumstances, clockwise “a” to “f”, starting upper left corner. (a) An AFC in the lesser sac tracking into the parietal wall anteriorly (marked with dotted white two-headed arrow) presenting as crepitus and erythema. (b) A copious amount of pus was drained, and necrosectomy was needed later as well. (c) Another AFC (marked “afc”) was drained with a transgastric drain (marked with white arrow). (d) Resolution of symptoms and size of AFC using transhepatic percutaneous drainage (PCD) (marked with white arrow). (e) Symptomatic AFC with no safe window for percutaneous drain (PCD, lack of window depicted with dotted white two-headed arrow). (f) A large infected AFC in the lesser sac without a safe window for PCD needed early necrosectomy

In Group 1, CM had to be abandoned in four patients due to deterioration in organ scores or rapidly progressive symptoms in 48 h. One patient was drained with PCD successfully, while one underwent PCD but succumbed in the next 48 h. No radiological window could be identified in two patients for PCD; one patient died in 24 h, while one underwent an emergency necrosectomy and died after 4 days. In Group 2, two patients could not be drained with PCD due to absence of a safe radiological window; one underwent successful necrosectomy, while one died in 48 h due to multi-organ failure. These crossover patients were not analysed by intention-to-treat analysis due to small overall numbers in the study, and the subsequent comparisons in Tables 1 and 2 are between the 16 remaining patients in each group.

Table 1.

Demographic and clinical comparison of patients of MSAP/SAP with symptomatic AFCs between Group 1 (conservative management) and Group 2 (drain)

Parameter	Group 1 (n = 16)	Group 2 (n = 16)	P value
Mean age (years ± SD)*	38.69 ± 7.59	32.88 ± 7.76	0.028
21–30 age group	2 (12.5%)	7 (43.8%)
31–40 age group	9 (56.2%)	5 (31.2%)
41–50 age group	5 (31.2%)	4 (25.0%)
Gender
Male	10 (62.5%)	12 (75.0%)	0.446
Female	6 (37.5%)	4 (25.0%)
Mean BMI (kg/m2 ± SD)	21.52 ± 1.66	21.62 ± 2.08	0.836
Aetiology
Gallstones	13 (81.2%)	9 (56.2%)	0.212
Alcohol	3 (18.8%)	4 (25.0%)
Trauma	0 (0.0%)	3 (18.8%)
Symptoms
Pain abdomen	16 (100.0%)	16 (100.0%)	1.000
Fever	5 (31.2%)	5 (31.2%)	1.000
Abdominal distension	16 (100.0%)	16 (100.0%)	1.000
Obstipation	8 (50.0%)	6 (37.5%)	0.476
Vomiting	12 (75.0%)	11 (68.8%)	1.000
Respiratory distress	5 (31.2%)	4 (25.0%)	1.000
Inability to accept oral feed	10 (62.5%)	8 (50.0%)	0.476
Signs
Fever	2 (12.5%)	3 (18.8%)	1.000
Tachycardia	14 (87.5%)	12 (75.0%)	0.654
Hypotension	4 (25.0%)	3 (18.8%)	1.000
Tachypnea	5 (31.2%)	5 (31.2%)	1.000
Crepts/reduced air entry in lung fields	5 (31.2%)	4 (25.0%)	1.000
Lump/phlegmon	1 (6.2%)	2 (12.5%)	1.000
Oliguria	3 (18.8%)	4 (25.0%)	1.000

MSAP: moderately severe acute pancreatitis, SAP: severe acute pancreatitis, AFCs: acute fluid collections, SD: standard deviation, BMI: body mass index

Student t-test (or Mann–Whitney U test) was used for comparison of continuous variables with differences in means, while the chi-square test (or Fischer’s exact test) was used for categorical variables

^*Denotes when P value was significant at <0.05

Table 2.

Clinical progression, hospital stay, and other outcome parameters comparison of patients of MSAP/SAP with symptomatic AFCs between Group 1 (conservative management) and Group 2 (drain)

Parameter	Group 1 (n = 16)	Group 2 (n = 16)	P value
Mean CTSI ± SD	6.88 ± 1.93	7.62 ± 2.33	0.301
Mean serum amylase on DOA (IU/L ± SD)	1018.31 ± 1008.24	1342.12 ± 971.47	0.341
Mean serum lipase on DOA (U/L ± SD)	970.94 ± 1396.27	649.00 ± 630.58	0.925
Mean peak IAP (mm Hg ± SD)	14.50 ± 2.19	13.19 ± 2.23	0.191
Mean IAP on day 0 (mm Hg ± SD)	12.56 ± 2.39	12.75 ± 2.46	0.829
Mean IAP on day 2 (mm Hg ± SD)	13.00 ± 2.00	13.06 ± 2.26	0.954
Mean IAP on day 7 (mm Hg ± SD)*	13.81 ± 2.71	9.29 ± 1.82	<0.001
Mean IAP on day 14 (mm Hg ± SD)	11.43 ± 3.46	9.12 ± 2.10	0.066
Mean peak SOFA score ± SD	3.56 ± 1.50	3.25 ± 1.84	0.660
Mean SOFA score on day 0 ± SD	1.25 ± 1.06	1.62 ± 1.45	0.568
Mean SOFA score on day 2 ± SD	2.94 ± 1.12	3.19 ± 1.80	0.575
Mean SOFA score on day 7 ± SD*	3.33 ± 1.35	1.69 ± 1.55	0.013
Mean SOFA score on day 14 ± SD	2.40 ± 1.88	1.89 ± 1.76	0.647
Mean SOFA score on day 21 ± SD	2.00 ± 2.32	1.43 ± 2.15	0.560
>2 organ failure present	10 (62.5%)	7 (43.8%)	0.288
Mean time to mortality (days ± SD)	21.33 ± 7.51	12.40 ± 8.91	0.189
Mean hospital stay (days ± SD)	16.08 ± 5.31	16.27 ± 10.66	0.662
Mortality	3 (18.8%)	5 (31.2%)	0.685

MSAP: moderately severe acute pancreatitis, SAP: severe acute pancreatitis, AFCs: acute fluid collections, SD: standard deviation, CTSI: CT severity index, DOA: day of admission, day 0: day of inclusion into group1/2, SOFA: sequential organ failure assessment, IAP: Intra-abdominal pressure

Student’s t-test (or Mann–Whitney U test) was used for comparison of continuous variables with differences in means, while the chi-square test (or Fischer’s exact test) was used for categorical variables

^*Denotes when P value was significant at <0.05

The tables show that apart from age, all other demographic and clinical variables were comparable between the groups. After the insertion of PCD, the only parameters that significantly decreased were the IAP on day 7 and the SOFA score on day 7. The IAP showed a decrease on other days too, however, these changes were not statistically significant. Figure 2 denotes that the extent of fall in IAP and SOFA scores is more in Group 2 (PCD) than in Group 1. Also, there is a significant correlation between the IAP and SOFA scores on day 7 (when they are significantly low) in Group 2 as compared to Group 1 (Spearman correlation coefficient ρ = 0.3, P value = 0.223 for Group 1 as compared to ρ = 0.7, P = 0.007 for Group 2), thereby suggesting that the changes are not due to random chance.

Figure 2.

Graphs depicting relationships between intra-abdominal pressure (IAP) and organ failure scores (SOFA) in the two groups, clockwise “a” to “d”, starting upper left corner. (a) Better decrease of mean SOFA scores with time in the percutaneous drainage (PCD) group. (b) Better decrease of mean IAP values with time in the PCD group. (c) Significant and strong correlation between IAP and SOFA at day 7 in the PCD group (Spearman’s correlation coefficient “ρ” = 0.7, P = 0.007). (d) Non-significant weaker correlation between IAP and SOFA at day 7 in the other group (“ρ”= 0.3, P = 0.223)

PCD was performed on a total of 17 patients (16 + 1). Of these, four patients underwent only intercostal drainage for respiratory symptoms, while eight patients had insertion of multiple drains. Two patients had insertion of only abdominal drain for generalised collection. The mean time of drainage was 10 days (range 8–17 days) after admission, with worsening organ scores as the commonest indication in 70% of patients. Other frequent indications were abdominal distension (increased IAP) and ileus. More than 60% of patients had multiple indications. The average initial drainage amount was 1200 mL, with mostly cloudy fluid. Cultures were positive in only two patients, with the offending organisms as Klebsiella and Enterococcus species. Drain amylase levels could not be calibrated satisfactorily, and this parameter was not used further in the study. The mean time to drain removal was 12 days (range 6–14 days) after the amount had reduced to less than 10 mL/day. There were no complications like infections or bowel injuries due to PCD; only one patient developed a pancreatic fistula which healed on CM in 3 months. Transgastric or transhepatic passage of the PCD also did not have any adverse consequences [Figure 1]. Overall, four patients underwent exploratory laparotomy and necrosectomy, two as earlier described in crossover patients. The other two patients were both following PCD in Group 2; one had a rapid extension of toxic fluid into the parietal abdominal wall [Figure 1A], while one did not improve after PCD. Both these patients recovered eventually. The overall mortality was 11/38 (29%); there was no significant difference in either time to death or mortality rate between the two groups. Most of the deaths in 7/11 (64%) were due to acute respiratory distress syndrome (ARDS) or respiratory failure, one death was following acute tubular necrosis and hyperkalaemia due to rhabdomyolysis in traumatic pancreatitis, and the remaining three were in the “crossover” patients as described above due to progressive necrosis and organ failure.

Discussion

The traditional management of AP, especially with increasing levels of severity and systemic involvement, has been CM in the intensive care or high-dependency unit. Aggressive intervention was reserved as a last resort for severe sterile necrosis and proven infected necrosis. This intervention was primarily pancreatic debridement and necrosectomy, which was associated with a high rate of procedure-related morbidity and mortality.[2,5] A landmark study by van Santvoort et al.,[8] known as the PANTER trial (minimally invasive step-up approach vs. maximal necrosectomy in patients with acute necrotising pancreatitis), was instrumental in changing the prevalent practices. The authors recruited 88 patients of necrotising pancreatitis and randomly assigned them to an open necrosectomy or step-up approach (PCD the minimally invasive retroperitoneal necrosectomy if required). The primary endpoint was a new-onset multiple-organ failure, which occurred significantly less often in the step-up approach.[4] This trial paved the way for acceptance of the step-up approach. However, this acceptance was for necrotising pancreatitis not responding to CM for at least the initial 3–4 weeks.

Debate over drainage of early fluid collections

As mentioned above, the rationale of delayed interventions in MSAP and SAP till at least 4 weeks is to avoid a “second hit” in an inflammatory and unstable situation. However, this inflexible approach does not account for a subset of patients who do not respond to initial CM. In our clinical experience, many patients with large AFCs in the first 2–3 weeks have significant abdominal distension, fever, and inability to accept oral feeding. Enteral feeding is a proven important parameter in reversing the inflammatory and catabolic state. Collections in the pleural space also hamper adequate oxygenation. We, thus, felt the pressing need to modify the traditional approach for this subset of patients. The use of PCD has been studied as early as 1998 by Freeny et al.[15] as an alternative to primary surgical necrosectomy in infected acute necrotising pancreatitis. The rationale for PCD was to interrupt sepsis, and thereby postpone the need for surgical intervention. They demonstrated success in sepsis management in 25/34 (74%) patients and avoided surgical intervention in 47%, without any catheter-related complications.[15] Till 2014–15, the results of the PANTER trial and similar studies convinced most clinicians to adopt minimally invasive drainage. However, this drainage was still restricted to necrosis around the pancreas and was performed after 4–6 weeks. Subsequently, many authors felt the need to drain AFCs and not just necrosis with PCD.[10,11,12,13,14,16,17,18,19,20]

One of the first attempts was by Mallick et al. from India, who compared the outcomes of PCD for AFC and walled-off necrosis (WON) in retrospective data of 650 cases. Differences between the need for additional surgical necrosectomy [14% with acute necrotic collections (ANC) and 12% of WON; P = 0.364] and mortality (19% with ANC and 13.7% with WON; P = 0.132) were higher in PCD with AFC, though not significant. The external pancreatic fistula was significantly more frequent with WON than in those with ANC (24.4% vs. 34.2%, respectively, P = 0.034). It was concluded that persistent organ failure is more often an indication of PCD in patients with ANC, while suspected infection is more commonly an indication in WON. A large number 44/258 patients needed drainage in AFC for pressure symptoms. Early PCD was found to be as efficacious as delayed PCD. This data, though retrospective, resonates greatly with our clinical experience.[17]

Can abdominal paracentesis be a bridge between conservative management and further step-up?

Liu et al.[21] analysed retrospective data of 102 patients of AP; 53 cases underwent abdominal paracentesis drainage (APD) followed by PCD, and 49 cases were managed only PCD. They found significantly lower mortality, delayed further interventions, and lower inflammatory markers in the APD group.[21] The same authors later identified predictors for PCD after APD as infected AFCs, collections of more than 100 mL, and reduction of AFCs by 50% after APD.[14] Zerem et al.[13] recruited a prospective cohort of 118 cases, where 96 underwent APD in the peritoneal fluid collection; 42 were managed with APD alone, and 54 underwent PCD after APD (14 required necrosectomy after initial PCD). The remaining 22 patients directly underwent PCD for AFCs. It was seen that APD led to a large decrease in levels of the initial severity scores and laboratory variables in both groups and that the application of APD ahead of PCD is safe and beneficial in the management of AP AFCs but there were no relevant predictors that suggest whether APD is indicated or not.[13] A meta-analysis done on the topic suggests that early application of APD in AP is associated with reduced all-cause mortality, expenses, and duration of hospitalisation, however, the included studies are low in number and weak in design.[22] We tried APD in only the four cases where a safe window for initial PCD was not available by ultrasound, however, it did not aid eventual management. In fact, lack of a window for drainage proved to be a significant hindrance in the management of these patients, and three of them died. Due to high recurrence and irregular location of AFCs, we do not consider APD to be a useful tool in the management of MSAP and SAP.

Mechanism of benefit after drainage of acute fluid collections

The onset of infection is recognised as the “tipping point” tilting the balance in AP towards organ dysfunction. Wang et al.[11] hypothesised that PCD in SAP with sterile AFCs may reduce further infection. In 452 patients of SAP with PCD, patients were into three groups – sterile, secondary infection, and primary infection with mortality of 4.1%, 10.9%, and 18.6%, respectively. Logistic regression analysis indicated that multi-organ dysfunction syndrome, catheters located intra-abdominally, and intra-abdominal hypertension were predictors for infection after PCD. The authors felt that due to reduction in IAP, intra-abdominal PCD for sterile AFCs seems to be as effective as peripancreatic PCD. Patients with a dramatic decrease in IAP after PCD had a lower incidence of infection and better alleviation of organ failure.[11] In our study too, IAP and SOFA score reduction correlates well with better outcomes. Other studies have also demonstrated a decrease in inflammatory markers in the AFC following drainage, which is associated with better outcomes.[12,16,23] Bellam et al.[12] documented the role of IAP reduction and C-reactive protein (CRP) levels in the success of PCD. Of note, Mallick et al.[23] demonstrated a significant decrease in CRP, IL-6, and IL-10 levels correlating with the outcome in 59 patients managed with PCD. Though the concept is scientifically plausible, we could not consistently perform tests of inflammatory markers on the drain fluid, and it would be an area for future research. Serum amylase showed a faster-decreasing trend in our study after PCD, though most studies do not attribute amylase levels to severity, except Formanchuk et al.[16] We have included all types of drainage, namely intercostal drainage, percutaneous lesser sac AFC drainage, and generalised abdominal fluid drainage in a single intervention group, as we believe that the mechanisms of benefit, as discussed above, are relevant for all these types. In our data, only 4/38 patients (10.5%) have undergone exploratory laparotomy and necrosectomy, suggesting that PCD seems to have prevented many potentially morbid and fatal operations. The mortality in our patients (11/38; 29%) is high, although cause-specific analysis reveals most mortality due to ARDS and respiratory failure, or inability to perform timely drainage due to lack of a window. Insertion of PCD, when appropriately indicated, can reduce disease progression and mortality; however, this may be more readily appreciable in a larger sample size.

In advocating PCD for sterile yet symptomatic AFCs, a few ethics issues arise because of the possibility of the recommendations being extrapolated to all patients. The complications of unnecessary drainage are infection and pancreatic fistula formation. These are rare (<10%) in all the mentioned studies, but if present, can be difficult to manage. We had one patient with pancreatic fistula secondary to PCD, and it took more than 6 weeks to heal. Hence, patient selection is very important. Early PCD should be adopted only for MSAP or SAP with sterile yet symptomatic AFCs and not in any milder scenario. Zhang et al.[10] studied the efficacy of PCD in 361 cases of AP of varying severity; early PCD contributed to decrease in operation rate, infection rate, and mortality in SAP. In MSAP with sterile AFCs, early PCD actually led to increased operation rate and infection rate.[10] Consistency of the drainable fluid and a safe radiological window are also necessary parameters to be evaluated. Most studies have used CT-guided drainage since CT delineates retroperitoneal structures better. However, ultrasound can also be used effectively, as seen in most of our cases. Homogeneous fluid accumulations are easier to evacuate, while necrotic debris may need many sessions of therapy. Sometimes multiple catheters or upscaling of catheter diameter may be required.[16,24] Irrigation and saline debridement of the area through an early PCD cannot be recommended since adequate sealing off of the greater sac has not occurred by this time (<4 weeks).

Apart from ethical issues as mentioned above, there are other challenges and limitations in our study. The inclusion criteria are strict, and that restricts the sample size. Most studies cited previously are retrospective studies, and there are only a few prospective studies on the role of PCD for AFCs.[16,19] These studies have a small sample size and are also not randomised trials. Randomisation will always be challenging in this scenario because, in the event of the failure of CM, the clinician is ethically bound to offer an intervention. Patients of AP also have respiratory, renal, and other complications that lead to poor outcomes, and these will be confounding factors in the evaluation of outcomes. We also do not have consistent follow-up data after discharge to evaluate long-term outcomes. A multi-centric randomised study would help to resolve many queries in this evolving area.

Conflicts of interest

There are no conflicts of interest.

Funding Statement

Nil.