Introduction
Endoscopic retrograde cholangiopancreatography (ERCP) is a commonly used procedure
to evaluate and/or treat biliopancreatic disorders. Some adverse events (AEs) that
can occur during or after ERCP include: perforation, bleeding, cholecystitis, and
pancreatitis. In some patients, post-ERCP pancreatitis (PEP) can become severe enough
to necessitate Intensive Care Unit (ICU) monitoring or even result in death. The estimated
incidence of PEP is approximately 8.8 % in average-risk patients to 14.1 % in high-risk
patients [1 ]. Factors which place patients at a high risk of PEP include both patient-related
factors and procedure-related factors. Patient-related factors include: suspected
sphincter of Oddi dysfunction, female gender, previous pancreatitis, previous PEP,
younger age, non-dilated extrahepatic bile ducts, absence of chronic pancreatitis,
and normal serum bilirubin. Procedure-related factors include duration and number
of cannulation attempts, pancreatic duct (PD) guidewire passages > 1, PD contrast
injection, precut sphincterotomy, pancreatic sphincterotomy, biliary balloon dilatation,
intraductal ultrasound, and failure to clear bile duct stones [2 ]
[3 ]
[4 ]
[5 ].
There has been significant headway in reducing the incidence and severity of PEP with
multiple interventions, which includes PD stent placement, nonsteroidal anti-inflammatory
drugs (NSAIDs) per rectum, and aggressive peri-procedure intravenous fluids [5 ]
[6 ]
[7 ]
[8 ]
[9 ]
[10 ]. Both the American College of Gastroenterology and American Society for Gastrointestinal
Endoscopy endorse the use of these items to reduce the risk of severe PEP in high-risk
patients [9 ]
[10 ]. In addition, the European Society of Gastrointestinal Endoscopy have recently updated
their guidelines in 2020 in which they recommend aggressive hydration with Lactated
Ringer’s solution (LR) in patients with contraindication to NSAIDs, provided that
they are not at risk of fluid overload and that a prophylactic PD stent is not placed
[2 ].
A current topic of interest is understanding the optimal type and rate of peri-procedural
intravenous hydration to prevent PEP. There are multiple studies and meta-analyses
indicating that aggressive peri-procedural hydration, specifically LR, may significantly
lower the incidence of PEP [11 ]
[12 ]. In the majority of these studies, aggressive IV hydration with LR is compared to
either standard IV hydration with LR or normal saline (NS). Talukdar et. al found
that a higher mean volume of IV fluid (IVF) (752 ± 783 mL vs 1012 ± 725, P = 0.036) and use of LR (OR, 0.56; 95 % confidence intervals (CI) 0.31–0.99) was associated
with a lower risk of PEP in a secondary analysis of an international multicenter PEP
prevention trial in high-risk patients [13 ].
To date, there have only been two studies which evaluated the efficacy of aggressive
NS when compared to aggressive LR for PEP prevention. In the first study, there were
four patient groups, of which only two of the patient groups received both rectal
indomethacin and 1 L of either NS or LR within 30 minutes of their ERCP. Thirteen
percent of the NS group and 6 % of the LR group experienced PEP, which was not statistically
significant and the number of patients in each arm was low [14 ]. A subsequent prospective randomized controlled trial (RCT) evaluated aggressive
LR hydration vs aggressive NS hydration vs standard LR hydration. While there was
a statistical difference in lowering PEP between the aggressive and standard groups,
there was no difference in the PEP rates between the aggressive LR and aggressive
NS groups. In addition, this study examined both average and high-risk patients who
underwent ERCP [15 ].
To our knowledge, there has not been a study with a head-to-head comparison of PEP
occurrence with aggressive NS infusion vs aggressive LR infusion for specifically
high-risk patients whom all received rectal indomethacin. We performed the first RCT
to compare aggressive NS infusion vs aggressive LR infusion for the prevention of
PEP in high-risk patients all receiving rectal indomethacin and undergoing ERCP. A
part of this manuscript was published as an abstract in the journal of Gastrointestinal
Endoscopy [16 ].
Methods
Study design
A randomized, blinded, controlled trial was performed at a single tertiary care center
(Tampa General Hospital) in the United States from June 2017 to June 2019. This study
was in compliance with the Declaration of Helsinki and approved by our institution’s
investigational review board (approved 7 /2017). This trial was registered online
at clinicaltrials.gov (NCT03215862) before enrollment of any patients. A data and
safety monitoring board provided oversight.
Patients
All consecutive adult (> 18 years of age) patients scheduled to undergo ERCP were
screened for exclusion criteria. These criteria were: age less than 18, age greater
than or equal to 75, inability to provide informed consent, pregnancy, acute pancreatitis,
chronic pancreatitis, any contraindication to aggressive IVF hydration [evidence of
clinical volume overload (peripheral or pulmonary edema)], respiratory compromise
(oxygen saturation < 90 % on room air), chronic kidney disease (creatinine clearance
< 40 mL/min), systolic congestive heart failure (ejection fraction < 45 %), cirrhosis,
severe electrolyte disturbance with sodium < 130 mEq/L or > 150 mEq/L, cholangitis,
and/or a true NSAID allergy.
Patients who did not meet the exclusion criteria were consented for the study. Patients
were eligible to be randomized to either NS or LR group if they met one major or two
minor validated patient or procedural risk factors for developing PEP. Major criteria
include: sphincter of Oddi dysfunction, personal history of PEP, total bilirubin less
than 1.0, more than eight cannulation attempts or more than 10 minutes, precut sphincterotomy,
endoscopic papillary balloon dilation of an intact sphincter, pancreatic sphincterotomy,
and ampullectomy. Similarly, possession of two or more of the following minor criteria:
female sex, age under 50 years, personal history of recurrent acute pancreatitis,
PD injection leading to “acinarization” or over three PD injections, or PD guidewire
insertion twice or more [4 ]
[9 ]. These inclusion factors are associated with high risk for PEP [2 ]
[5 ].
Intervention
Once written informed consent was obtained, patients were randomized in a 1:1 ratio
to receive either NS infusion or LR infusion. Each ERCP procedure was performed by
1 of 3 therapeutic endoscopists who perform > 200 ERCPs per year. Postgraduate year-6
and postgraduate year-7 gastroenterology fellows were involved in all cases and were
allowed up to seven attempts at cannulation. All patients that were eligible for randomization
and that received pre-procedural fluids were given 100 mg rectal indomethacin just
prior to start of ERCP procedure. Prophylactic PD stents were placed if the PD was
cannulated twice or more, and/or at the discretion of the endoscopist.
Administration of intravenous fluid
The IVF infusion was initiated by nurses in the pre-procedure area prior to the procedure.
The infusion was started at a rate of 3 mL/kg/h, and this was continued throughout
the procedure. A 20 mL/kg bolus was then given at the end of the procedure. After
this bolus was complete, the infusion was continued at a rate of 3 mL/kg/h. If the
patient remained hospitalized, the post-procedure infusion was continued for 8 hours
[17 ]
[18 ]. The maximum rate of the intravenous fluids was set at 500 mL/h. If the patient
was discharged home after the procedure, then the infusion was only continued for
the duration that the patient remained in our post-procedure area. In our institution,
these patients are monitored for 45–60 minutes based on alertness, vital signs, and
appropriate verbal response before being discharged home.
Outcomes
The primary study outcome was the occurrence of PEP. This was defined by: (1) new
or worsening abdominal pain that is clinically consistent with acute pancreatitis,
combined with (2) associated pancreatic enzymes elevation ≥ 3 times the upper limit
of normal 24 hours after the procedure, and (3) resultant or prolongation of existing
hospitalization of ≥ 2 nights [2 ]. Secondary study outcomes included: severe acute pancreatitis (persistent organ
failure ≥ 48 hours), localized AEs (pseudocyst formation or walled off pancreatic
necrosis), and death. Other recorded events were AEs related to IVF: peripheral/pulmonary
edema and hypoxia. Also recorded were AEs related to NSAIDs: anaphylaxis, gastrointestinal
bleeding, and acute kidney injury.
Inpatient participants were all seen and examined within 24 hours following the procedure
and asked/evaluated for new abdominal pain, shortness of breath, chest pain, and/or
extremity swelling. Outpatient participants were called by one investigator (RP) on
the phone within 72 hours after the procedure. Patients were asked if they experienced
any new abdominal pain, new shortness of breath, extremity swelling, or chest pain.
Sample size
It was estimated that a total of 242 subjects would have been required to detect a
difference in PEP rates of 0.16 in the NS group versus 0.05 in the LR group assuming
a medium effect size with 80 % power using Fisher’s exact test.
Methods to reduce bias
The randomization sequence was generated using a computer. Randomization was performed
as block randomization in random block sizes in a 1:1 allocation ratio using sealed
envelopes. Patients, nurses, endoscopists, outcome assessors and data analysts were
blinded to the treatment assignment. Both types of intravenous fluids (LR and NS)
were available in the pre-procedural area. The pre-op nurse covered the intravenous
fluid bag with a dark opaque bag which obscured the name/type of fluid from vision.
The endoscopists, investigators, and patients were therefore not aware of the allocation.
Statistical analysis
Patient and demographic characteristics were summarized as mean and standard deviation
for continuous variables and as rates for categorical variables. All analyses were
performed following the intent-to-treat principle for benefits. The difference in
primary outcomes across compared groups was assessed using a Fisher’s exact test and
summarized as relative risk (RR) along with 95 % CI. Additional measures of relative
(RRR), absolute risk reduction (ARR) and number needed to treat (NNT) along with 95 %
CI was also calculated. The statistical significance was set at 5 % for all comparisons.
The adjusted and unadjusted difference in primary outcomes was assessed using the
binary logistic regression. All analyses were performed using the IBM SPSS version
25 statistical analysis package.
Results
Patients
A total of 763 subjects were screened ([Fig. 1 ]). Of these, 596 met exclusion criteria. The remaining 167 patients were enrolled
and provided written informed consent. Six of these patients did not show up after
consent was obtained and were not able to be reached via telephone despite multiple
attempts. One hundred sixty-one patients underwent randomization to either LR or NS
and all received rectal nonsteroidal anti-inflammatory drugs (NSAIDS). After the procedure,
another 25 patients were excluded as they did not meet patient or procedural risk
factors that were high-risk for developing PEP. Thus, a total of 136 patient were
included in the final data analysis. Baseline characteristics of the 136 included
patients were similar across the two randomized groups ([Table 1 ]). The mean age was 51.9 years in the LR group and 51.8 years in the NS group (P = 0.97). There were a similar percentage of females and males enrolled in the LR
(64 % and 36 %, respectively) and NS groups (61 % and 39 %, P = 0.72). Both the major criteria and minor criteria of risk factors for developing
PEP were similar between the two groups. The distribution of PD stent placement between
LR and NS groups was 21 (29 %) and 13 (20 %), respectively. Finally, there were 43
(50.5 %) ambulatory cases in the LR group, and 41 (53.9 %) in the NS group.
Fig. 1 Study flowchart detailing the number of patients initially screened, enrolled, and
included in the prospective analysis.
Table 1
Baseline patient characteristics (high-risk patients).
Lactated Ringer’s
Normal saline
Age (mean)
51.9
51.8
Female (%)
64 %
61 %
Male (%)
36 %
39 %
Outpatient procedures (n)
43
41
Race
Black
6
1
Asian
0
4
Latino
4
2
White
53
48
Other/Not Recorded
9
9
Total
72
64
n = 136
Risk factors
Major inclusion criteria
Suspected SOD
8
6
Personal history of PEP
2
5
Bilirubin < 1
50
41
More than 8 cannulation
26
24
Precut sphincterotomy
12
6
Balloon dilation of intact
1
0
PD sphincterotomy
5
4
Ampullectomy
6
6
Minor inclusion criteria
Female
47
39
Age < 50
35
26
Personal history of
7
7
PD injection
24
20
PD cannulation with wire × 2
23
18
# of patients with prophylactic PD stents placed
21
13
SOD, sphincter of Oddi; PEP, post-ERCP pancreatitis; PD, pancreatic duct.
Outcomes
As highlighted in [Table 2 ], the incidence of PEP was 4 % (3/72 patients) in the LR group versus 11 % (7/64
patients) in the NS group resulting in a RR of 0.38 (95 % CI 0.10 to 1.42; P = 0.19). The relative risk reduction (RRR) was 0.62 (95 % CI –0.41 to 0.90) along
with an ARR of 0.07 (95 % CI –0.025 to 0.17) and an NNT of 15 (95 % CI –41 to 6).
Given the difference in PD stent placement between groups, we performed a multivariate
analysis which showed that PD stent did not have an impact on the rate of pancreatitis
(RR adjusted 0.34, 95 % CI (0.08–1.26); P = 0.11).
Table 2
Primary outcome: occurrence of PEP.
Lactated Ringer’s (n = 72)
Normal saline (n = 64)
P value
PEP cases
3
7
4.2 %
10.9 %
0.131
Outpatient PEP cases (discharged home)
1
4
0.49
Inpatient PEP cases (returned to hospital floor)
2
3
PEP, post-ERCP pancreatitis.
Regarding secondary outcomes ([Table 3 ]), there was one case of pancreatitis leading to pseudocyst formation in the NS group.
Otherwise, there were no cases of severe pancreatitis, walled off pancreatic necrosis,
death, events related to IVF infusion, or events related to NSAID use. There were
reports of procedural related events: two cases of bacteremia in the NS group, one
case of fever/sepsis in each group, one case in the LR group of post-sphincterotomy
bleed, and three cases in the LR group and two cases in the NS group of post-procedural
abdominal discomfort leading to admission to the hospital for observation.
Table 3
Secondary outcomes.
Lactated Ringer’s
Normal saline
Severe acute pancreatitis
0
0
Pseudocyst formation
0
1
Peripancreatic abscess formation
0
0
Walled off pancreatic necrosis
0
0
Death
0
0
Reported events related to intravenous fluid infusion
0
0
Reported events related to NSAID use
0
0
Presented within 2 days with bacteremia
0
2
Presented within 2 days with fever/sepsis
1
1
Sphincterotomy bleed requiring repeat EGD
1
0
Outpatient admitted to hospital after ERCP due to post-procedural abdominal discomfort
3
2
NSAID, nonsteroidal anti-inflammatory drug; EGD, esophagogastroduodenoscopy; ERCP,
endoscopic retrograde cholangiopancreatography.
[Table 4 ] gives a breakdown of the risk factors of each patient with PEP. Each PEP patient
had at least three risk factors (minor or major). The most common risk factors were
bilirubin < 1, more than eight cannulation attempts, or more than 10 minutes spent
cannulating, female gender, PD injection, and PD wire insertion × 2.
Table 4
Risk factors for PEP present in each of 10 patients who experienced PEP, as well as
the type of IVF they received.
Patient
1
2
3
4
5
6
7
8
9
10
Type of fluid received
NS
LR
NS
NS
NS
LR
NS
LR
NS
NS
Prophylactic PD stent placed
No
Yes
No
No
No
Yes
No
Yes
No
Yes
SOD dysfunction suspicion
X
X
Personal history of PEP
X
Bilirubin < 1
X
X
X
X
X
X
X
X
X
More than 8 cannulation
X
X
X
X
X
X
Precut sphincterotomy
X
X
Balloon dilation of intact
PD sphincterotomy
X
Ampullectomy
X
X
X
Female
X
X
X
X
X
X
Age < 50
X
X
X
X
Personal history of
X
X
PD injection
X
X
X
X
X
PD cannulation with wire × 2
X
X
X
X
X
X
PEP, post-ERCP pancreatitis; IVF, intravenous fluid; SOD, sphincter of Oddi dysfunction;
NS, normal saline; LR, Lactated Ringer’s; PD, pancreatic duct.
Discussion
The topic of our research study is the optimal type of intravenous hydration peri-procedure
to prevent PEP in patients who are at high risk of PEP. To date, there has not been
a study on prevention of PEP with a head-to-head comparison of aggressive IV NS compared
to aggressive IV LR for specifically high-risk patients who all received rectal indomethacin.
We think it is important to specifically examine patients at a high risk of PEP who
are all receiving rectal indomethacin, since this group can benefit the most from
establishing the optimal way to administer IVF. With this in mind, we performed a
RCT to assess aggressive LR infusion compared to aggressive NS infusion for the prevention
of PEP in high-risk patients undergoing ERCP. Overall, the incidence of PEP was lower
in the group receiving an aggressive LR infusion (4 %) compared to NS infusion (11 %),
with a RRR of 62 %, an ARR of 7 % and NNT of 15. However, the difference was not statistically
significant potentially due to poor accrual, thereby impacting the power of the study.
Furthermore, the distribution of outpatient procedures (50.5 % in LR vs 53.9 % in
NS group) may have contributed to our findings. More patients in the NS group were
outpatient, received less total volume of fluid and thus, may have been more likely
to develop PEP. However, the difference in the percentage of ambulatory cases between
the LR and NS group is likely not great enough to fully account for the difference
in rates of PEP.
The recently published FLUYT trial (fluid hydration to prevent PEP in average- to
high-risk patients receiving prophylactic rectal NSAIDs) assessed the value of aggressive
hydration in patients receiving prophylactic rectal NSAIDs. They included 826 moderate-
to high-risk patients all receiving prophylactic rectal NSAIDs randomized to either
a control group receiving maintenance NS or aggressive LR [19 ]. Interestingly, they found that aggressive peri-procedure hydration did not reduce
the incidence of PEP in patients who routinely received prophylactic rectal NSAIDs.
A large, multicenter, RCT such as the FLUYT trial may overpower our efforts to determine
the efficacy of aggressive hydration. Nonetheless, our study shows that there is no
statistical difference between the two forms of aggressive hydration, which is congruent
with their findings. In addition, while the aforementioned trial compares aggressive
LR to maintenance NS, it does not evaluate the role of aggressive NS infusion and
its comparison to aggressive LR infusion, which is unique to our study. Aggressive
NS (as opposed to maintenance NS) was used in our study to equalize the effect of
aggressive hydration between both groups, and thus create a more direct comparison
between LR and NS.
There were several strengths of this study. First, the study design of an RCT reduces
the risks of known and unknown biases that can potentially confound results. This
is the first study to compare aggressive LR hydration with aggressive NS hydration
in specifically high-risk patients for PEP. The endoscopists were blinded to which
fluid was given, limiting any bias on their part. However, there are few limitations
as well. A key limitation of the study is unexpected poor accrual of patients. We
planned for 242 patients with expected rates of pancreatitis being 5 % in the lactated
Ringer’s arm versus 16 % in the NS arm. However, the enrollment was stopped at 136
patients within the given timeframe of the study (2 years). The poor accrual can be
attributed to several factors including the limited timeframe of fellowship training,
as this study was led by a trainee in a 3-year fellowship program, and lack of resources
including personnel etc. However, these findings are important to inform power calculations
for future studies to provide reliable answers. For example, based on the estimates
from previous studies we powered the study for an absolute difference of 11 % for
the rate of pancreatitis which was higher than the observed rate of 7 % resulting
in post hoc power of 35 % versus the planned 80 % at a significance level of 5 %.
Nevertheless, this is one of the largest RCTs on this issue. We have also considered
patient allocation after randomization as a limitation of our study. Due to the nature
of the study in capturing only high-risk patients, we had to exclude patient’s that
did not meet procedural criteria for high-risk after randomization. Some high-risk
factors for PEP cannot be determined until after the ERCP is complete, and thus, those
patients that did not meet procedural criteria for PEP had to be excluded after randomization.
However, the number of patients who were excluded was similar between groups. Another
possible limitation is the study being performed at a single institution and all procedures
performed under the supervision of three endoscopists. The single institution participation
and limited number of endoscopists may be associated with reduced practice variation.
This same issue possibly limits the generalizability of these findings. However, these
findings provide real-world evidence of the efficacy of LR versus NS for the prevention
of PEP.
Conclusions
In conclusion, there is insufficient evidence to support the use of aggressive LR
infusion over aggressive NS peri-procedure in patients receiving rectal NSAIDs and
known to be at high risk for PEP.
