PlasmaLyte, Lactated Ringer’s, or Normal Saline resuscitation in acute pancreatitis: a single-center, pragmatic, cluster-randomized trial

Abstract

Background/aims

The optimal resuscitation fluid for acute pancreatitis (AP) remains uncertain. This study evaluated the effects of PlasmaLyte (PL), Lactated Ringer’s solution (LR), and Normal Saline (NS) in patients with early AP.

Methods

A single-center, pragmatic, unblinded, monthly cluster-randomized trial compared PL, LR, and NS in patients with AP who presented within 6 h of symptom onset and had no evidence of overt severe disease at enrollment. The primary outcome was the change in systemic inflammatory response syndrome (SIRS) prevalence at 24 h. Secondary outcomes included clinical outcomes and biochemical changes after fluid resuscitation.

Results

Sixty-six patients were enrolled (PL 21, LR 24, NS 21). Baseline SIRS prevalence was 23.8% in the PL group, 29.2% in the LR group, and 19.0% in the NS group; at 24 h, SIRS prevalence decreased to 4.8%, 12.5%, and 9.5%, respectively. This corresponded to relative reductions of 80.0% in PL, 57.1% in LR, and 50.0% in NS, without a statistically significant between-group difference (P = 0.668). No significant differences were observed in major clinical outcomes. At 24 h, potassium was higher in PL than in LR, chloride was higher in NS than in PL, and pH, bicarbonate, and base excess were lower in NS than in balanced crystalloid groups. These biochemical differences should be interpreted cautiously because of the small sample size and cluster design.

Conclusions

In this small pragmatic trial, PL, LR, and NS showed no significant difference in the primary clinical endpoint. However, balanced crystalloids were associated with more favorable acid-base profiles than NS. These findings are exploratory and hypothesis-generating and require confirmation in larger multicenter trials with appropriate cluster-adjusted or individually randomized designs.

Trial registration

Clinical Research Information Service (CRIS), KCT0007422. Registered June 20, 2022, https://cris.nih.go.kr.

Introduction

Acute pancreatitis is a common gastrointestinal disorder that can lead to substantial morbidity and mortality, particularly when pancreatic inflammation is accompanied by local complications or organ failure [1–3]. Appropriate early fluid resuscitation remains a central component of supportive management because maintenance of pancreatic microcirculation may reduce progression of systemic and local injury [4–6].

Although crystalloid fluids are widely recommended for initial resuscitation, the optimal type of crystalloid remains uncertain. Guidelines and prior studies have often favored LR over NS because balanced crystalloids may better preserve acid-base homeostasis, but the overall evidence remains limited and not entirely consistent [7–10]. A recent meta-analysis of randomized trials comparing LR with NS found that LR was associated with lower odds of moderately severe or severe pancreatitis, ICU admission, and local complications, whereas SIRS and organ failure did not differ significantly [11–13].

PlasmaLyte is a balanced crystalloid with a composition closer to that of plasma than either LR or NS [14, 15]. Compared with NS, it contains less chloride and has a physiologic pH; unlike LR, it contains acetate and gluconate as buffer anions and also contains magnesium [16]. These features suggest that PL may differ from both LR and NS in effects on electrolyte and acid-base balance, but prospective comparative data in AP are lacking.

Therefore, this study aimed to evaluate whether clinical and biochemical outcomes differed according to initial resuscitation with PL, LR, or NS in patients with early AP.

Materials and methods

Study design and participants

This was a single-center, pragmatic, unblinded, monthly cluster-randomized trial conducted at The Catholic University of Korea, Uijeongbu St. Mary’s Hospital, from October 2022 to July 2023.

The study protocol was approved by the institutional review board of Uijeongbu St. Mary’s Hospital (UC22MISI0038), and written informed consent was obtained from all participants. The trial was registered at the Clinical Research Information Service (KCT0007422) and reported in accordance with CONSORT principles.

Patients were eligible if they presented to the emergency department within 6 h of symptom onset and met the diagnostic criteria for AP. Diagnosis of AP required at least two of the following: characteristic abdominal pain, serum amylase or lipase greater than three times the upper normal limit, or imaging findings consistent with AP. Because final severity classification according to the revised Atlanta classification cannot be determined reliably at the time of very early presentation, patients were enrolled if they had AP without evidence of overt severe disease at baseline and were then classified during follow-up according to the revised Atlanta criteria [17].

Inclusion criteria were as follows: (1) age 19 years or older, (2) diagnosis of AP within 6 h of symptom onset, and (3) willingness to participate and provide written informed consent. Exclusion criteria were as follows: (1) clinically overt severe AP at presentation, including persistent shock, respiratory failure, or other organ failure requiring immediate intensive management; (2) New York Heart Association class greater than II heart failure; (3) pulmonary edema; (4) renal failure with serum creatinine greater than 1.9 mg/dL; (5) abnormal liver function with albumin less than 3 mg/dL or liver cirrhosis; (6) clinical signs of hypervolemia, defined as peripheral edema, pulmonary rales or crackles, clinically evident elevated jugular venous pressure, or clinical or radiographic evidence of pulmonary edema; (7) pregnancy; (8) age younger than 19 years or older than 80 years; and (9) refusal to participate.

Randomization and interventions

Patients were assigned by monthly cluster randomization to receive one of three study fluids: PL, LR, or NS. In this pragmatic design, each calendar month was prospectively assigned to a single study fluid according to a pre-specified random schedule, and all eligible patients presenting during that month were treated with that fluid as the default initial resuscitation fluid. This approach was chosen because AP often presents during night-time or early-morning hours, making individual real-time randomization difficult in the emergency setting.

The study was unblinded. Patients, treating clinicians, investigators, and outcome assessors were aware of group assignment. The cluster design therefore reflected real-world practice but also introduced a risk of temporal confounding and performance bias.

The fluid resuscitation protocol was based on a modified goal-directed hydration strategy informed by previous AP studies [18–20]. Initially, patients received a 10 mL/kg bolus over 1 h, followed by continuous infusion at 3 mL/kg/h for 11 h. At 12 h, laboratory reassessment included blood urea nitrogen (BUN), creatinine, and hematocrit. If at least two of these three parameters increased compared with baseline values, an additional 10 mL/kg bolus was administered over 1 h, followed by infusion at 3 mL/kg/h for another 11 h. If at least two of the three parameters decreased compared with baseline, the infusion rate was reduced to 1.5 mL/kg/h for 12 h without an additional bolus. After 24 h, fluid therapy was managed according to the patient’s clinical condition and physician judgment.

Because of the pragmatic emergency-room setting, some patients may have received an initial small amount of fluid before formal study enrollment. Only patients who could continue treatment under the assigned monthly study protocol were enrolled in the trial.

Study outcomes

The primary outcome was the change in SIRS prevalence 24 h after fluid resuscitation. SIRS was defined as two or more of the following: heart rate greater than 90 beats/min, respiratory rate greater than 20 breaths/min or PaCO2 less than 32 mmHg, temperature greater than 38 °C or less than 36 °C, or white blood cell count less than 4,000/mm3, greater than 12,000/mm3, or greater than 10% bands [17].

Secondary outcomes included C-reactive protein level at 24 h, local complications (acute peripancreatic fluid collection, pancreatic necrosis, pseudocyst, or walled-off necrosis), organ failure at 48 h, progression to moderately severe or severe AP, ICU admission, invasive intervention, length of hospital stay, in-hospital mortality, 30-day mortality, 30-day readmission, metabolic acidosis, fluid overload, and changes in electrolyte and acid-base parameters at 24 h. Final severity classification was determined during follow-up according to the revised Atlanta classification. Local complications were assessed by contrast-enhanced computed tomography or other clinically indicated imaging during hospitalization, according to routine clinical practice rather than a fixed study-mandated schedule.

Sample size estimation

The sample size calculation was based on the primary outcome of change in SIRS prevalence. At the time of study design, no prospective comparative data were available for PL in AP. Therefore, the expected effect size was derived from prior LR-versus-NS literature, particularly the study by Wu et al., which reported marked differences in SIRS reduction at 24 h [20]. Baseline SIRS prevalence was assumed to be 50%, and PL was assumed to have efficacy at least comparable to LR because both are balanced crystalloids. Using G*Power 3.1.9.7 with 80% power and a two-sided alpha of 0.05, 20 patients per group were required; assuming a 10% dropout rate, 22 patients per group were planned, for a total of 66 patients. This assumption was indirect and is acknowledged as a limitation.

Statistical analysis

All analyses were performed on an intention-to-treat basis. Continuous variables were tested for normality with the Shapiro-Wilk test and are presented as mean ± standard deviation or median with interquartile range, as appropriate. Normally distributed variables were compared using one-way analysis of variance followed by Tukey post hoc testing, whereas non-normally distributed variables were analyzed using the Kruskal-Wallis test followed by Bonferroni-adjusted pairwise comparisons. Categorical variables were compared using the chi-square test or Fisher’s exact test. For three pairwise nonparametric post hoc comparisons, P less than 0.017 was considered statistically significant. All analyses were performed using IBM SPSS software version 29 (IBM Corp., Armonk, New York, USA).

Because the study used monthly cluster randomization with a small number of clusters, formal cluster-adjusted analysis and reliable estimation of intracluster correlation coefficients were not feasible. Accordingly, all P-values should be interpreted with caution, as standard patient-level analyses may underestimate uncertainty in small cluster-randomized trials.

Results

Baseline characteristics

A total of 66 patients were enrolled and analyzed: 21 in the PL group, 24 in the LR group, and 21 in the NS group (Fig. 1). Baseline demographic and clinical characteristics were generally similar among groups (Table 1). Mean age ranged from 47.1 to 55.1 years, and most patients were male. Alcohol-related and biliary pancreatitis were the most common etiologies. At enrollment, SIRS prevalence was 23.8% in the PL group, 29.2% in the LR group, and 19.0% in the NS group (P = 0.731). Baseline BISAP scores, comorbidities, and laboratory values were also similar across groups.

Fig. 1.

CONSORT diagram of study participant flow. PL PlasmaLyte, LR lactated Ringer’s solution, NS normal saline

Table 1.

Baseline demographics and clinical characteristics of study participants

	PL (n = 21)	LR (n = 24)	NS (n = 21)	P- value
Age (years) (mean ± SD)	54.1 ± 20.3	47.1 ± 17.7	55.1 ± 20.0	0.315
Male	17 (81.0%)	14 (58.3%)	15 (71.4%)	0.252
Body mass index (kg/㎡) (median) (IQR)	26.8 (23.4, 30.1)	24.0 (21.4, 28.9)	26.2 (23.6, 28.2)	0.368
Etiology				0.569
Alcohol	10 (47.8%)	9 (37.5%)	5 (23.8%)
Biliary	4 (19.0%)	8 (33.3%)	9 (42.9%)
Idiopathic	4 (19.0%)	2 (8.3%)	4 (19.0%)
Drug	0	1 (4.2%)	0
Hypertriglyceridemia	2 (9.5%)	4 (16.7%)	2 (9.5%)
Other	1 (4.6%)	0	1 (4.6%)
SIRS present at enrollment	5 (23.8%)	7 (29.2%)	4 (19.0%)	0.731
BISAP at enrollment (median) (IQR)	1 (0, 1)	0 (0, 0.75)	0 (0, 1)	0.320
Comorbidities	4 (19.0%)	9 (37.5%)	8 (38.1%)	0.314
BUN (mg/dl) (median) (IQR)	14.7 (10.3, 17.1)	13.7 (11.3, 15.8)	14.3 (10.6, 17.2)	0.539
Creatinine (mg/dl) (median) (IQR)	0.85 (0.68, 0.99)	0.78 (0.65, 0.90)	0.90 (0.74, 1.05)	0.123
Hematocrit (%) (mean ± SD)	43.9 ± 3.6	41.7 ± 6.5	43.4 ± 5.8	0.378
CRP (mg/dl) (median) (IQR)	0.56 (0.12, 2.45)	0.40 (0.14, 0.78)	0.30 (0.08, 0.83)	0.576
Na (mEq/L) (median) (IQR)	138.0 (134.75, 140.75)	138.0 (133.5, 140.0)	140.0 (137.5, 141.5)	0.104
K (mEq/L) (mean ± SD)	4.010 ± 0.4323	4.125 ± 0.3166	3.943 ± 0.3529	0.249
Cl (mEq/L) (median) (IQR)	102.0 (98.5, 105.0)	102.0 (96.3, 104.0)	102.0 (100.0, 104.5)	0.746
Mg (mg/dl) (mean ± SD)	1.881 ± 0.2105	1.909 ± 0.3029	1.971 ± 0.2648	0.568
Ca (mg/dL) (median) (IQR)	9.150 (8.800, 9.400)	9.100 (8.900, 9.700)	9.100 (8.950, 9.400)	0.648
pH (median) (IQR)	7.39900 (7.38750, 7.42600)	7.41150 (7.39075, 7.43400)	7.39100 (7.37950, 7.43150)	0.453
Bicarbonate (mmol/L) (median) (IQR)	24.200 (23.500, 25.450)	24.300 (23.125, 25.125)	24.800 (22.750, 25.400)	0.905
BE (mmol/L) (median) (IQR)	-0.200 (-0.950, 1.250)	-0.200 (-1.475, 0.875)	0.400 (-2.000, 1.150)	0.863

PL PlasmaLyte; LR lactated Ringer’s solution, NS normal saline, SD standard deviation, IQR interquartile range, SIRS systemic inflammatory response syndrome, BISAP Bedside Index of Severity in Acute Pancreatitis, BUN blood urea nitrogen, CRP C-reactive protein, Na sodium, K potassium, Cl chloride, Mg magnesium, Ca calcium, BE base excess

Primary outcome

SIRS prevalence decreased in all three groups after 24 h of fluid resuscitation (Fig. 2). At 24 h, SIRS was present in 1 of 21 patients (4.8%) in the PL group, 3 of 24 patients (12.5%) in the LR group, and 2 of 21 patients (9.5%) in the NS group. Relative reductions from baseline were 80.0% for PL, 57.1% for LR, and 50.0% for NS. The between-group difference was not statistically significant (P = 0.668). Given the small number of SIRS events, these findings should be interpreted as exploratory rather than confirmatory.

Fig. 2.

Impact of fluid type on prevalence of SIRS at enrollment and 24 h after fluid resuscitation. PL PlasmaLyte, LR lactated Ringer’s solution, NS normal saline, SIRS systemic inflammatory response syndrome

Secondary clinical outcomes

No significant differences were observed among groups in CRP level at 24 h, total fluid volume administered during the first 24 h, ICU transfer rate, need for invasive treatment, length of stay, or mortality (Table 2). Median hospital stay was 4.0 days (IQR 3.0-5.5) in the PL group, 3.5 days (IQR 3.0-5.75) in the LR group, and 4.0 days (IQR 3.0-7.5) in the NS group (P = 0.252).

Table 2.

Clinical outcomes and adverse events according to the type of fluid therapy

	PL (n = 21)	LR (n = 24)	NS (n = 21)	P- value
CRP (mg/dL) (median) (IQR)	5.07 (3.515, 8.585)	5.9 (2.565, 12.205)	3.86 (1.685, 14.165)	0.923
Total volume of fluid during initial 24 h (mL) (mean ± SD)	4607.0 ± 825.1	4341.2 ± 1254.1	4304.0 ± 985.6	0.455
Local complications
Acute peripancreatic fluid collection	17 (81.0%)	12 (50.0%)	15 (71.4%)	0.076
Pancreatic necrosis	2 (9.5%)	1 (4.2%)	2 (9.5%)	0.728
Pseudocyst	0	0	0	1
Walled-off necrosis	0	0	2 (9.5%)	0.110
Length of stay (day) (median) (IQR)	4.0 (3.0, 5.5)	3.5 (3.0, 5.75)	4.0 (3.0, 7.5)	0.252
ICU transfer	2 (9.5%)	4 (16.7%)	4 (19.0%)	0.877
Need for invasive treatment	0	0	2 (9.5%)	0.110
Organ failure at 48 h	0	0	0	1
In-hospital mortality	0	0	0	1
30-day mortality	0	0	0	1
Fluid overload	0	0	1 (4.8%)	0.636
Metabolic acidosis	0	0	2 (9.5%)	0.110
Readmission within 30 days	0	1 (4.2%)	2 (9.5%)	0.515

PL PlasmaLyte, LR lactated Ringer’s solution, NS normal saline, SD standard deviation, IQR interquartile range, SD standard deviation, ICU intensive care unit

Local complications were observed in all groups. Acute peripancreatic fluid collection was numerically more frequent in the PL group (81.0%) than in the LR group (50.0%) or NS group (71.4%), but this difference did not reach statistical significance (P = 0.076). No patients developed organ failure at 48 h, and there were no in-hospital or 30-day deaths. Two patients in the NS group required invasive treatment, two developed metabolic acidosis, and one had fluid overload; none of these events occurred in the PL or LR groups, although differences were not statistically significant. Thirty-day readmission occurred in two NS patients, one LR patient, and no PL patients.

Electrolyte and acid-base parameters at 24 h

Several biochemical parameters differed among groups at 24 h (Fig. 3). Sodium and chloride were highest in the NS group. Potassium was higher in the PL group than in the LR group. Magnesium was highest in the PL group, and calcium was highest in the LR group. However, many of these differences likely reflect the intrinsic electrolyte composition of the administered fluids and remained within clinically usual ranges.

Fig. 3.

Electrolytes at 24 h after fluid resuscitation. Na sodium, K potassium, Cl chloride, Mg magnesium, Ca calcium

Acid-base parameters were more clinically informative (Fig. 4). pH was lower in the NS group than in the balanced crystalloid groups (P = 0.030). Bicarbonate and base excess were also lower in NS than in LR, with PL values lying between LR and NS. Two cases of metabolic acidosis occurred in the NS group. These findings are consistent with the known tendency of chloride-rich fluids to promote hyperchloremic metabolic acidosis, but the small sample size and lack of cluster-adjusted analysis preclude firm conclusions about clinical benefit.

Fig. 4.

Acid-base parameters at 24 h after fluid resuscitation. BE base excess

Discussion

In this pragmatic monthly cluster-randomized trial comparing PL, LR, and NS for early AP resuscitation, no significant difference was observed in the primary endpoint of SIRS prevalence at 24 h or in major clinical outcomes. Nevertheless, balanced crystalloids were associated with a more favorable acid-base profile than NS, and the PL group showed a numerically greater reduction in SIRS prevalence, although this trend was not statistically significant and should be regarded as hypothesis-generating only.

The finding that PL had the numerically greatest SIRS reduction must be interpreted in the context of the unexpectedly low baseline prevalence of SIRS in the study cohort. The sample size calculation assumed a 50% baseline SIRS prevalence on the basis of prior LR-versus-NS literature [20], but the observed prevalence was approximately 24%. As a result, the number of primary outcome events was much smaller than anticipated, substantially reducing statistical power. Similar underpowering due to lower-than-expected SIRS prevalence has been described in prior AP resuscitation studies [20].

The absence of significant differences in major clinical outcomes is broadly consistent with some prior randomized studies and the recent WATERLAND trial comparing LR and NS [20–23], although recent meta-analyses suggest that LR may reduce moderately severe or severe pancreatitis, ICU admission, and local complications relative to NS [11–13]. In the present study, the higher frequency of local complications and moderately severe AP in the PL group may reflect imbalance related to the small number of monthly clusters rather than a true detrimental effect of PL. Indeed, if the PL group contained a greater proportion of patients predisposed to complications, the numerically more favorable SIRS trend in that group may have been biased toward the null.

The most consistent signal in this study involved acid-base and selected electrolyte parameters. Compared with NS, balanced crystalloids better preserved pH, bicarbonate, and base excess. This pattern is biologically plausible, given the high chloride content of NS and its established association with hyperchloremic metabolic acidosis [14, 16, 24]. In AP, acid-base disturbances may be clinically relevant because acidosis has been linked experimentally to pancreatic injury.

PL and LR may not be physiologically identical despite both being balanced crystalloids. LR uses lactate as its principal buffer anion, whereas PL contains acetate and gluconate [15, 16]. Prior work in burn patients suggested no clear base excess advantage of PL over LR despite the higher nominal pH of PL, possibly because of gluconate accumulation after large-volume administration [25]. In the present study, LR had numerically higher bicarbonate and base excess than PL, although the difference between PL and LR was not statistically significant in post hoc testing. This observation may be related to differences in buffer metabolism, but it should be interpreted cautiously because this trial did not measure gluconate or acetate concentrations.

This study has several strengths. It is, to our knowledge, the first prospective trial directly comparing PL, LR, and NS in AP. In addition, its pragmatic design reflects real-world emergency care, where immediate individual randomization may be difficult.

However, the limitations are substantial. First, this was a single-center study with a small sample size. Second, the monthly cluster-randomized design created only a small number of clusters, increasing the risk of temporal confounding and cluster-related imbalance [26]. Third, formal cluster-adjusted analyses and reliable intra-cluster correlation estimation were not feasible, which weakens confidence in nominal P-values and raises the possibility of type I error for secondary biochemical comparisons. Fourth, the unblinded design may have introduced performance bias, although major outcomes were largely objective [26]. Fifth, the cohort consisted mostly of patients without severe disease at baseline, limiting generalizability to severe AP and reducing the ability to detect differences in hard clinical endpoints. Finally, several electrolyte differences reflected the known composition of the administered fluids and may not themselves represent clinically meaningful benefit.

Taken together, these findings do not support a definitive conclusion that one fluid was clinically superior to the others in this study. Rather, they suggest that balanced crystalloids, including PL and LR, may better preserve acid-base homeostasis than NS in early AP resuscitation. Larger multicenter trials using either individual randomization or adequately powered cluster-adjusted designs are required to determine whether these metabolic differences translate into meaningful clinical benefit.

Conclusions

In this small single-center pragmatic trial, PL, LR, and NS did not differ significantly in SIRS resolution or major clinical outcomes in early AP. Balanced crystalloids were associated with more favorable acid-base profiles than NS, but these findings are exploratory and should not be overinterpreted. Larger well-powered multicenter studies are needed before firm conclusions can be drawn regarding the optimal resuscitation fluid in AP.

Authors’ contributions

H.-K.K. and S.-H.S. conceptualized and designed the study, analyzed the data, and drafted the manuscript. H.-H.C. and S.-S.K. performed the study, collected, and interpreted the data, and revised the manuscript. All the authors agree to be fully accountable for ensuring the integrity and accuracy of this work. All the authors have read and approved the final version of the manuscript.

Funding

The authors received no funding for this work.

Data availability

The data sets used and analyzed are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

This study was performed in accordance with the Declaration of Helsinki and was approved by the Institutional Review Board of The Catholic University of Korea, Uijeongbu St. Mary’s Hospital (UC22MISI0038). All patients provided written informed consent.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.