Predictors of hemodynamic instability during ERCP sedation in older adults: impact of sedation strategy in a retrospective cohort

Abstract

Background

Older adults undergoing endoscopic retrograde cholangiopancreatography (ERCP) are particularly vulnerable to sedation-related hemodynamic instability. This study aimed to evaluate predictors of hemodynamic instability and the association between sedation strategy and hemodynamic outcomes in elderly patients.

Methods

This retrospective cohort study included patients aged ≥ 65 years who underwent ERCP under deep sedation between 2015 and 2025. Hemodynamic instability was defined as either a ≥ 30% reduction in mean arterial pressure or vasopressor requirement. Multivariable logistic regression and inverse probability of treatment weighting (IPTW) were performed.

Results

A total of 150 patients were included, and hemodynamic instability occurred in 46 (30.7%). Advanced age, ASA class III–IV, lower baseline mean arterial pressure, and prolonged procedure duration were associated with increased instability risk. Ketofol use was associated with lower odds of hemodynamic instability than propofol alone (aOR: 0.50, 95% CI: 0.25–0.99, p = 0.047), with consistent findings after IPTW (OR: 0.55, 95% CI: 0.31–0.98, p = 0.043).

Conclusions

Hemodynamic instability is common during ERCP in older adults. Ketofol use was associated with lower observed hemodynamic instability; however, residual confounding cannot be excluded because of the retrospective non-randomized design.

Clinical trial register

Not applicable.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12877-026-08405-7.

Introduction

Endoscopic retrograde cholangiopancreatography (ERCP) is an advanced therapeutic endoscopic procedure widely used in the diagnosis and management of pancreatobiliary diseases. Despite substantial technological advances and improvements in procedural success, ERCP remains a relatively invasive intervention that frequently requires deep sedation to facilitate procedural tolerance, optimize technical performance, and minimize patient discomfort. Appropriate sedation management during ERCP is therefore essential not only for procedural success, but also for maintaining hemodynamic stability and reducing procedure-related adverse events, particularly in vulnerable older adults [1–4]. Older patients represent a uniquely high-risk population in the setting of procedural sedation. Aging is associated with progressive reductions in cardiovascular reserve, impaired autonomic responsiveness, decreased vascular compliance, and diminished physiological compensatory capacity. In addition, elderly individuals commonly present with multimorbidity, polypharmacy, and chronic cardiovascular medication use, all of which may further increase susceptibility to sedation-related hemodynamic disturbances. Consequently, maintaining hemodynamic stability during ERCP in older adults remains a clinically important challenge in contemporary endoscopic practice [5–7].

Propofol is among the most commonly used sedative agents for gastrointestinal endoscopic procedures because of its rapid onset, short recovery profile, and predictable sedative effects. However, propofol is also associated with dose-dependent vasodilation, myocardial depression, and hypotension, which may become more clinically significant in elderly patients with limited physiological reserve. During prolonged or technically complex ERCP procedures, cumulative sedative exposure may further increase the risk of clinically relevant hemodynamic deterioration [4, 8, 9]. Ketamine, in contrast, possesses sympathomimetic properties that may help preserve cardiovascular stability during procedural sedation. The combination of ketamine and propofol, hereafter referred to as ketofol, has therefore gained increasing interest as a potentially balanced sedation strategy that may reduce propofol requirements while maintaining adequate procedural sedation. Previous studies have suggested favorable hemodynamic profiles with ketofol; however, the available evidence remains heterogeneous because of differences in patient populations, sedation depth, drug ratios, and procedural settings. Studies specifically evaluating older adults undergoing ERCP remain limited [10–12].

Most previous investigations of sedation during gastrointestinal endoscopic procedures have focused predominantly on respiratory complications, hypoxemia, or airway-related outcomes [13, 14]. Comparatively less attention has been directed toward hemodynamic instability in older ERCP patients. Furthermore, the contribution of baseline hemodynamic status, procedure duration, systemic disease burden, cardiovascular medication use, and sedation strategy to periprocedural instability has not been fully clarified [13, 15]. Accordingly, this retrospective cohort study aimed to evaluate clinical predictors of hemodynamic instability in patients aged ≥ 65 years undergoing ERCP under deep sedation and to investigate the association between sedation strategy and periprocedural hemodynamic outcomes.

Methods

Study design and ethical approval

This study was designed as a retrospective, single-center observational cohort study conducted in the endoscopy unit of the Department of Gastroenterology at Düzce University Faculty of Medicine. The study included older adult patients who underwent endoscopic retrograde cholangiopancreatography (ERCP) under deep sedation between January 1, 2015 and December 31, 2025. The primary objective was to evaluate predictors of hemodynamic instability during ERCP and to investigate the association between sedation strategy and periprocedural hemodynamic outcomes in elderly patients. The study protocol was reviewed and approved by the Non-Interventional Clinical Research Ethics Committee of Düzce University (approval number: 2026/160; date: May 11, 2026). All study procedures were conducted in accordance with the ethical principles of the Declaration of Helsinki and current institutional regulations. Due to the retrospective nature of the study and the use of anonymized clinical data obtained from existing medical records, the requirement for written informed consent was waived by the ethics committee. All patient data were anonymized prior to analysis.

Study setting and patient population

The study population consisted of patients aged ≥ 65 years who underwent ERCP under deep sedation during the study period. Eligible patients were identified retrospectively through the hospital information management system, anesthesiology records, and endoscopy unit databases. Only the first eligible ERCP procedure for each patient was included to maintain data independence. Patients were categorized according to the sedation strategy administered: propofol-based sedation or ketofol, defined as the combined intravenous administration of ketamine and propofol. Deep sedation was administered by anesthesiology specialists with continuous electrocardiography, non-invasive blood pressure monitoring, pulse oximetry, and heart rate monitoring. A total of 150 patients meeting the eligibility criteria were included in the final analysis. The patient selection process is summarized in Fig. 1.

Fig. 1.

Flow diagram of patient screening, exclusion, and final cohort selection

Inclusion and exclusion criteria

Patients were eligible if they were aged ≥ 65 years, underwent ERCP under deep sedation, received either propofol or ketofol, and had complete clinical, anesthetic, and hemodynamic monitoring data. Patients were excluded for general anesthesia, sedation regimens other than propofol or ketofol, repeated ERCP procedures, incomplete procedural or anesthetic records, missing or unreliable hemodynamic data, or major protocol inconsistencies that could interfere with outcome assessment.

Sedation management and periprocedural monitoring

All ERCP procedures were performed under deep sedation administered by anesthesiology specialists. Sedation strategy was not randomized and was selected according to the attending anesthesiologist’s clinical judgment, taking into consideration patient age, comorbidity burden, baseline hemodynamic status, anticipated procedural complexity, and expected procedure duration. No mandatory institutional allocation algorithm determining propofol versus ketofol use was applied. In the propofol group, propofol was administered intravenously using incremental boluses titrated according to clinical response. In the ketofol group, ketamine and propofol were administered as combined or sequential titrated intravenous boluses according to procedural requirements and the attending anesthesiologist’s assessment. Total administered doses were extracted from anesthesiology records and standardized according to body weight (mg/kg). In patients receiving ketofol, the cumulative ketamine-to-propofol dose ratio was additionally calculated. Sedation was clinically targeted to achieve deep sedation sufficient for procedural tolerance while maintaining spontaneous ventilation. Deep sedation was defined clinically as a level of sedation associated with markedly reduced responsiveness to external stimuli while spontaneous respiration was maintained. Objective sedation-depth measures such as the Ramsay Sedation Scale, MOAA/S, RASS, or bispectral index monitoring were not consistently documented throughout the 10-year study period and were therefore unavailable for standardized retrospective analysis. All patients underwent continuous electrocardiography, non-invasive blood pressure monitoring, pulse oximetry, and heart rate monitoring. Hemodynamic parameters were recorded before sedation initiation and at regular intervals throughout the procedure. Supplemental oxygen was administered when clinically indicated.

Intravenous fluid and vasopressor therapy were administered according to the magnitude and persistence of blood pressure reduction and the patient’s clinical condition. Ephedrine was the predominantly recorded vasopressor used for clinically relevant intraprocedural hypotension. However, because a uniform protocol-defined MAP threshold for vasopressor administration was not consistently applied throughout the entire study period, vasopressor use represented a clinically triggered intervention determined by the attending anesthesiologist.

Data collection and clinical variables

Data were retrospectively collected from the hospital information management system, anesthesiology monitoring forms, endoscopy unit procedure records, and electronic patient files using a standardized extraction form. Demographic variables included age, sex, and body mass index. Clinical variables included ASA physical status, hypertension, diabetes mellitus, coronary artery disease, chronic kidney disease, chronic pulmonary disease, beta-blocker use, and ACEI/ARB use. Overall comorbidity burden was additionally assessed using the Charlson Comorbidity Index (CCI). Procedure-related variables included procedure duration, sedation strategy, administered propofol and ketamine doses, intravenous fluid administration, and vasopressor requirement. Available ERCP-related characteristics included indication, emergency status, difficult cannulation, precut sphincterotomy, stone extraction, hilar intervention, pancreatic intervention, and metallic stent placement. These procedural-complexity variables were incorporated into sensitivity analyses where appropriate. Calendar year of the procedure was also recorded to evaluate potential temporal changes in clinical practice during the extended study period. Hemodynamic variables included baseline systolic blood pressure, baseline mean arterial pressure (MAP), baseline heart rate, lowest recorded systolic blood pressure, lowest MAP, and absolute and percentage changes in MAP. Recovery time was defined as the interval from completion of the procedure to documented readiness for discharge from the post-anesthesia recovery area. Discharge readiness was based on routine institutional clinical criteria, including adequate consciousness, stable respiratory status, and acceptable hemodynamic recovery. A single validated recovery scoring system was not consistently documented throughout the entire study period.

Missing data management

Patients with missing key variables related to sedation strategy, baseline hemodynamic measurements, procedural monitoring, or primary outcome assessment were excluded before the final analysis. Complete-case analysis was performed, and no multiple imputation was applied.

Definition of hemodynamic instability and outcome measures

The primary outcome was hemodynamic instability during ERCP, predefined as a composite endpoint comprising either a ≥ 30% reduction in MAP from the pre-sedation baseline value or administration of a vasopressor for clinically significant intraprocedural hypotension. Baseline MAP was defined as the final recorded measurement obtained before sedation initiation. Secondary outcomes included lowest systolic blood pressure, lowest MAP, absolute and percentage change in MAP, hypotension requiring intervention, bradycardia, intravenous fluid administration, vasopressor requirement, total vasopressor dose, and recovery time. Hypotension requiring intervention was defined as a decrease in blood pressure prompting therapeutic intervention with intravenous fluids and/or vasopressor therapy. Bradycardia was defined as a heart rate < 50 beats/min requiring clinical evaluation or intervention.

Statistical analysis

Statistical analyses were performed using IBM SPSS Statistics version 26.0 (IBM Corp., Armonk, NY, USA). Continuous variables were assessed for normality using the Shapiro–Wilk test and visual inspection and were presented as mean ± standard deviation or median (interquartile range), as appropriate. Categorical variables were expressed as numbers and percentages. Between-group comparisons were performed using the independent-samples t-test or Mann–Whitney U test and the chi-square or Fisher’s exact test, as appropriate. Sedative doses were additionally analyzed as weight-adjusted doses (mg/kg). Because sedation strategy was not randomly assigned, inverse probability of treatment weighting (IPTW) based on propensity scores was used as a sensitivity analysis to reduce measured baseline imbalance between the propofol and ketofol groups. Propensity scores incorporated relevant preprocedural demographic, clinical, hemodynamic, procedural, and calendar-year variables. Covariate balance was evaluated using standardized mean differences (SMDs), with an absolute SMD < 0.10 indicating adequate balance. Weighted logistic regression with robust standard errors was subsequently performed. Multivariable logistic regression was used to evaluate factors associated with hemodynamic instability, including sedation strategy, age, ASA class, CCI, baseline MAP, procedure duration, and calendar year. Adjusted odds ratios (aORs) with 95% confidence intervals (CIs) were reported. Multicollinearity was assessed using variance inflation factors, and model calibration was evaluated using the Hosmer–Lemeshow test. Potential temporal changes during the 2015–2025 study period were evaluated across predefined calendar periods and by including calendar year in adjusted models. Sensitivity analyses additionally incorporated available ERCP procedural-complexity variables. ROC analyses were performed for selected predictors; because cutoff values were derived and evaluated within the same cohort, these analyses were considered exploratory and require external validation. All tests were two-sided, and p < 0.05 was considered statistically significant.

Results

A total of 150 patients aged ≥ 65 years undergoing ERCP under deep sedation were included in the final cohort. Propofol-based sedation was administered to 77 patients (51.3%), whereas 73 patients (48.7%) received ketofol. Hemodynamic instability occurred in 46 patients (30.7%) overall.

Baseline characteristics and propensity score weighting

Baseline characteristics according to sedation strategy are presented in Table 1. Before weighting, patients receiving ketofol tended to be older and had a greater overall comorbidity burden. The prevalence of ASA class III–IV was 53.4% in the ketofol group compared with 41.6% in the propofol group. Chronic kidney disease and beta-blocker use were also numerically more frequent among ketofol-treated patients. Baseline MAP was lower in the ketofol group (92.3 ± 10.9 vs. 96.1 ± 11.4 mmHg). A substantial temporal difference in sedation practice was observed, with ketofol being used more frequently during the later years of the study. Consequently, calendar year demonstrated one of the largest baseline imbalances between treatment groups. After IPTW, balance improved substantially across all prespecified baseline covariates, with absolute SMDs < 0.10. No extreme propensity weights were observed, and the weighted population showed adequate overlap between sedation groups.

Table 1.

Baseline characteristics according to sedation strategy before and after IPTW

Variable	Propofol (n = 77)	Ketofol (n = 73)	Unweighted SMD	IPTW SMD
Age, years	71.8 ± 5.7	73.1 ± 5.9	0.22	0.04
Male sex	43 (55.8)	40 (54.8)	0.02	0.01
BMI, kg/m²	28.0 ± 4.3	28.4 ± 4.2	0.09	0.03
CCI	3 (2–4)	4 (2–5)	0.24	0.05
ASA class III–IV	32 (41.6)	39 (53.4)	0.24	0.04
Hypertension	40 (51.9)	41 (56.2)	0.09	0.03
Diabetes mellitus	23 (29.9)	25 (34.2)	0.09	0.04
Coronary artery disease	19 (24.7)	21 (28.8)	0.09	0.03
Chronic kidney disease	10 (13.0)	15 (20.5)	0.20	0.03
Chronic pulmonary disease	12 (15.6)	13 (17.8)	0.06	0.02
Beta-blocker use	20 (26.0)	25 (34.2)	0.18	0.06
ACEI/ARB use	23 (29.9)	24 (32.9)	0.06	0.03
Baseline MAP, mmHg	96.1 ± 11.4	92.3 ± 10.9	0.34	0.05
Emergency ERCP	13 (16.9)	11 (15.1)	0.05	0.02
Malignant biliary obstruction	19 (24.7)	20 (27.4)	0.06	0.04
Calendar year, median (IQR)	2019 (2017–2022)	2023 (2021–2024)	0.68	0.07

Values are presented as mean ± standard deviation, median (interquartile range), or n (%). SMD: standardized mean difference; IPTW: inverse probability of treatment weighting; BMI: body mass index; CCI: Charlson Comorbidity Index; ASA: American Society of Anesthesiologists; MAP: mean arterial pressure; ACEI: angiotensin-converting enzyme inhibitor; ARB: angiotensin receptor blocker

Sedation exposure, procedural characteristics, and hemodynamic outcomes

Median procedure duration was 38 (30–49) minutes in the propofol group and 42 (32–54) minutes in the ketofol group (p = 0.081). The frequencies of difficult cannulation, precut sphincterotomy, stone extraction, hilar intervention, pancreatic intervention, and metallic stent placement did not differ significantly between sedation groups. When doses were standardized according to body weight, total propofol exposure was significantly greater in patients receiving propofol alone than in those receiving ketofol (2.42 ± 0.58 vs. 1.70 ± 0.42 mg/kg, p < 0.001). Among patients receiving ketofol, the mean ketamine dose was 0.46 ± 0.11 mg/kg. The median ketamine-to-propofol dose ratio was 0.27 (IQR, 0.23–0.31), corresponding to an approximate ketamine ratio of 1:3.7. Patients receiving ketofol had a higher lowest recorded MAP (72.8 ± 9.4 vs. 66.1 ± 10.7 mmHg, p < 0.001) and a smaller percentage decrease in MAP from baseline (20.1 ± 8.4% vs. 30.6 ± 10.3%, p < 0.001).

Hemodynamic instability occurred in 31 of 77 patients (40.3%) receiving propofol and 15 of 73 patients (20.5%) receiving ketofol (p = 0.009). Hypotension requiring intervention occurred in 35.1% and 16.4%, respectively (p = 0.009), while vasopressor administration was required in 29.9% and 12.3%, respectively (p = 0.009). Bradycardia rates were comparable between groups. Recovery time was modestly longer in the ketofol group (Table 2).

Table 2.

Procedural, sedation, and hemodynamic outcomes according to sedation strategy

Variable	Propofol (n = 77)	Ketofol (n = 73)	p value
Procedure duration, min	38 (30–49)	42 (32–54)	0.081
Difficult cannulation	18 (23.4)	19 (26.0)	0.711
Precut sphincterotomy	12 (15.6)	11 (15.1)	0.932
Stone extraction	42 (54.5)	37 (50.7)	0.644
Hilar intervention	8 (10.4)	8 (11.0)	0.909
Pancreatic intervention	5 (6.5)	5 (6.8)	0.938
Metallic stent placement	13 (16.9)	14 (19.2)	0.717
Total propofol dose, mg/kg	2.42 ± 0.58	1.70 ± 0.42	< 0.001
Ketamine dose, mg/kg	—	0.46 ± 0.11	—
Ketamine dose ratio	—	0.27 (0.23–0.31)	—
Lowest systolic BP, mmHg	91.4 ± 12.8	99.6 ± 11.7	< 0.001
Lowest MAP, mmHg	66.1 ± 10.7	72.8 ± 9.4	< 0.001
MAP decrease, %	30.6 ± 10.3	20.1 ± 8.4	< 0.001
Hemodynamic instability	31 (40.3)	15 (20.5)	0.009
Hypotension requiring intervention	27 (35.1)	12 (16.4)	0.009
Vasopressor requirement	23 (29.9)	9 (12.3)	0.009
Bradycardia	9 (11.7)	6 (8.2)	0.476
Recovery time, min	31 (25–39)	36 (30–45)	0.021

Values are presented as mean ± standard deviation, median (interquartile range), or n (%). BP: blood pressure; MAP: mean arterial pressure

IPTW-adjusted analysis

After IPTW, the estimated incidence of hemodynamic instability was 36.8% in the propofol group and 24.1% in the ketofol group, corresponding to an absolute weighted risk difference of − 12.7% points (95% CI, − 24.8 to − 0.6). In weighted logistic regression with robust standard errors, ketofol use remained associated with lower odds of hemodynamic instability compared with propofol alone (weighted OR: 0.55, 95% CI: 0.31–0.98, p = 0.043). The magnitude of the association was attenuated compared with the unadjusted analysis.

Multivariable predictors of hemodynamic instability

Results of the multivariable logistic regression analysis are shown in Table 3. After adjustment for age, ASA physical status, CCI, baseline MAP, procedure duration, and calendar year, ketofol sedation was associated with lower odds of hemodynamic instability compared with propofol sedation (aOR: 0.50, 95% CI: 0.25–0.99, p = 0.047). Increasing age, ASA class III–IV, lower baseline MAP, and longer procedure duration were associated with increased odds of hemodynamic instability. Calendar year was not independently associated with the primary outcome after adjustment (aOR: 0.96 per year, 95% CI: 0.88–1.05, p = 0.389). Addition of ERCP procedural-complexity variables to a sensitivity model did not materially alter the estimated association between sedation strategy and hemodynamic instability (aOR for ketofol: 0.52, 95% CI: 0.27–1.02, p = 0.057).

Table 3.

Multivariable logistic regression analysis of hemodynamic instability

Variable	Adjusted OR	95% CI	p value
Ketofol vs. propofol	0.50	0.25–0.99	0.047
Age, per year	1.06	1.01–1.12	0.021
ASA class III–IV	2.31	1.07–4.98	0.033
CCI, per point	1.14	0.94–1.39	0.188
Baseline MAP, per mmHg	0.96	0.93–0.99	0.008
Procedure duration, per 10 min	1.32	1.08–1.61	0.006
Calendar year, per year	0.96	0.88–1.05	0.389

The model demonstrated acceptable calibration (Hosmer–Lemeshow p = 0.721) and discrimination (AUC: 0.79, 95% CI: 0.71–0.86). All variance inflation factors were < 2.5. OR: odds ratio; CI: confidence interval; ASA: American Society of Anesthesiologists; CCI: Charlson Comorbidity Index; MAP: mean arterial pressure

Temporal trends and sensitivity analyses

Ketofol use increased substantially over the study period. It was used in 10 of 45 procedures (22.2%) performed during 2015–2019, 23 of 50 procedures (46.0%) during 2020–2022, and 40 of 55 procedures (72.7%) during 2023–2025 (p for trend < 0.001). The overall incidence of hemodynamic instability was 35.6% during 2015–2019, 32.0% during 2020–2022, and 25.5% during 2023–2025; this temporal reduction was not statistically significant (p = 0.535). Period-stratified analyses showed a directionally lower incidence of hemodynamic instability with ketofol across all three periods. Hemodynamic instability occurred in 37.1% versus 30.0% of propofol- and ketofol-treated patients during 2015–2019, 40.7% versus 21.7% during 2020–2022, and 46.7% versus 17.5% during 2023–2025. There was no statistically significant interaction between sedation strategy and calendar period (p for interaction = 0.614). The inclusion of calendar year in the multivariable model and propensity score weighting did not materially alter the direction of the association between sedation strategy and hemodynamic instability.

Exploratory ROC and subgroup analyses

Exploratory ROC analyses demonstrated that procedure duration and baseline MAP provided moderate discrimination for hemodynamic instability. The previously identified thresholds of procedure duration > 42 min and baseline MAP ≤ 92 mmHg were retained as sample-derived exploratory thresholds. Because these cutoffs were derived and evaluated in the same cohort, they require external validation before clinical application. Detailed ROC analyses and exploratory subgroup analyses according to age ≥ 75 years, ASA class III–IV, chronic kidney disease, low baseline MAP, and prolonged procedure duration are provided in the Supplementary Material.

Discussion

In this retrospective cohort study evaluating elderly patients undergoing ERCP under deep sedation, we identified several clinically relevant predictors of hemodynamic instability and observed an association between sedation strategy and periprocedural hemodynamic outcomes. Hemodynamic instability occurred in nearly one-third of the study population. Advanced age, higher ASA physical status, lower baseline mean arterial pressure, and prolonged procedure duration were associated with increased instability risk, whereas ketofol-based sedation was associated with comparatively lower rates of hypotension and vasopressor requirement [16, 17]. Importantly, the association between ketofol use and lower hemodynamic instability persisted after multivariable adjustment and remained directionally consistent after IPTW, although its magnitude was attenuated. These findings should therefore be interpreted as observational associations rather than evidence of a causal protective effect.

Aging is associated with reduced cardiovascular reserve, impaired autonomic responsiveness, arterial stiffness, and diminished baroreceptor activity, which may limit compensation for sedative-induced vasodilation and myocardial depression [18, 19]. Previous studies of gastrointestinal endoscopic sedation have similarly reported greater susceptibility to hypotension and cardiovascular adverse events in older adults [20, 21]. Our findings extend these observations by specifically emphasizing hemodynamic vulnerability in an elderly ERCP population.

Higher ASA classification was also associated with hemodynamic instability. ASA class III–IV may reflect reduced physiological reserve and a greater burden of systemic disease, thereby increasing susceptibility to blood pressure reductions during deep sedation [22, 23]. Although overall comorbidity burden assessed using the CCI was incorporated into the revised analyses, ASA physical status remained a clinically relevant marker of hemodynamic risk.

Baseline MAP emerged as another important predictor. Patients with lower preprocedural MAP may have limited hemodynamic reserve because of antihypertensive treatment, autonomic dysfunction, dehydration, frailty, or underlying cardiovascular disease, making additional sedation-related reductions in vascular tone more clinically relevant [17, 24]. Procedure duration was similarly associated with instability. Prolonged ERCP procedures may require greater cumulative sedative exposure and are often associated with increased procedural complexity, both of which may contribute to hemodynamic variability [24, 25]. Sensitivity analyses incorporating available procedural-complexity characteristics did not materially alter the direction of the observed association between sedation strategy and instability. However, the exploratory ROC-derived thresholds for procedure duration and baseline MAP were derived from the same cohort and require external validation before clinical application.

Previous ERCP-specific studies have likewise highlighted the relevance of advanced age, greater systemic disease burden, prolonged procedures, and cumulative sedative exposure to cardiovascular adverse events during sedation. However, direct comparison across studies remains challenging because of differences in patient age, sedation regimens, definitions of hypotension, and procedural complexity [20, 21, 23]. The present study adds to this literature by focusing specifically on hemodynamic instability in an elderly ERCP cohort and by incorporating both conventional multivariable adjustment and propensity-based weighting.

Ketofol use was associated with a more favorable observed hemodynamic profile. Compared with propofol alone, ketofol was associated with lower rates of hemodynamic instability, hypotension requiring intervention, and vasopressor use. The complementary pharmacodynamic properties of ketamine and propofol provide a plausible explanation: propofol reduces systemic vascular resistance and myocardial contractility, whereas ketamine may partially preserve sympathetic tone and cardiovascular function [26–30]. Weight-adjusted dose analyses also demonstrated lower propofol exposure in the ketofol group, which may have contributed to the observed differences.

Previous studies have suggested that ketofol may improve hemodynamic stability during procedural sedation; however, available evidence remains heterogeneous because of differences in ketamine-to-propofol ratios, sedation depth, patient characteristics, and procedural settings [31–34]. The present findings provide additional real-world data from an elderly ERCP population. Nevertheless, sedation strategy in our cohort was selected according to clinical judgment rather than random allocation, creating the potential for confounding by indication despite multivariable adjustment and IPTW.

Recovery time was modestly longer among patients receiving ketofol, suggesting a potential trade-off between hemodynamic stability and postprocedural recovery characteristics. Ketamine-containing regimens may delay complete psychomotor recovery in some older individuals because of age-related changes in pharmacokinetics and central nervous system sensitivity [26, 33]. Sedation strategy should therefore remain individualized rather than being selected solely on the basis of hemodynamic considerations.

Another important consideration was the extended 2015–2025 study period. Ketofol utilization increased substantially over time, reflecting evolving institutional sedation practice. Nevertheless, calendar year was not independently associated with hemodynamic instability after adjustment, and inclusion of calendar year in both propensity score and multivariable models did not materially alter the direction of the ketofol association. Although these analyses reduce concern regarding temporal bias, they cannot completely account for unmeasured changes in ERCP technique, monitoring technology, operator experience, or peri-procedural care during the decade-long study period. Exploratory subgroup analyses suggested directionally lower instability rates with ketofol in several higher-risk categories, including older patients and those undergoing prolonged procedures. However, interaction testing did not demonstrate convincing evidence that the association differed across these subgroups. These findings should therefore be considered hypothesis-generating rather than evidence supporting preferential ketofol use in a specific subgroup [35].

Limitations

Several limitations should be considered when interpreting the present findings. First, the retrospective single-center design limits causal inference and may introduce selection bias. Sedation strategy was determined according to anesthesiologist preference and patient-specific clinical considerations rather than random allocation, resulting in the possibility of confounding by indication. Although multivariable adjustment and IPTW were used to improve comparability between sedation groups, residual confounding from unmeasured factors cannot be excluded.

Second, several clinically relevant geriatric characteristics, including frailty, functional dependency, cognitive impairment, nutritional status, detailed cardiac function, and polypharmacy burden, could not be comprehensively evaluated. Certain sedation-related factors could also not be fully standardized retrospectively. Objective sedation-depth scores or BIS monitoring were not consistently available throughout the study period, and the use and cumulative doses of adjunctive sedative or analgesic agents, including benzodiazepines and opioids, could not be reliably reconstructed for all patients.

Although recovery readiness was documented clinically, a single validated recovery scoring system was not consistently recorded throughout the entire 10-year period. Similarly, vasopressor treatment was guided by the attending anesthesiologist’s clinical judgment rather than by a uniform protocol-defined MAP threshold, which may have introduced variability in the intervention component of the primary outcome.

Although several ERCP complexity characteristics were incorporated into sensitivity analyses, operator experience, procedural stimulation intensity, and other unmeasured technical factors could not be fully standardized. Furthermore, the extended 2015–2025 study period may have included changes in ERCP techniques, monitoring technologies, sedation practices, and operator experience. Calendar year was incorporated into propensity score and multivariable analyses; nevertheless, residual temporal bias remains possible.

Complete-case analysis may have introduced information bias related to missing retrospective data. Subgroup analyses involved relatively small numbers of patients and should therefore be considered exploratory. Similarly, ROC-derived cutoff values were derived and evaluated within the same cohort and require external validation before clinical application. Finally, the findings may not be generalizable to younger patients, other endoscopic procedures, or centers using different sedation practices.

Despite these limitations, the study provides clinically relevant real-world data regarding hemodynamic instability and sedation-related outcomes in a high-risk elderly ERCP population.

Conclusion

In conclusion, hemodynamic instability is a frequent event among older adults undergoing ERCP under deep sedation. Advanced age, higher ASA physical status, lower baseline mean arterial pressure, and prolonged procedure duration were associated with increased instability risk. Ketofol use was associated with lower observed rates of hemodynamic instability, hypotension, and vasopressor requirement compared with propofol alone, and this association remained directionally consistent after multivariable adjustment and IPTW. However, the non-randomized retrospective design and potential residual confounding preclude causal interpretation. These findings support individualized sedation planning, careful preprocedural risk assessment, and close hemodynamic monitoring in older adults undergoing ERCP. Prospective multicenter studies are needed to confirm these associations and determine whether ketofol provides a true hemodynamic advantage in this population.

Supplementary information

Below is the link to the electronic supplementary material.

Author contributions

Conceptualization, A.Ü.E. and S.T.; Methodology, A.Ü.E. and S.T.; Formal Analysis, A.Ü.E.; Investigation, A.Ü.E. and S.T.; Data Curation, A.Ü.E.; Writing – Original Draft, A.Ü.E.; Writing – Review & Editing, S.T.; Visualization, A.Ü.E.; Supervision, S.T.; Project Administration, A.Ü.E. All authors have read and approved the final version of the manuscript.

Funding

This research received no external funding.

Data availability

The data supporting the findings of this study are available from the corresponding author upon reasonable request.

Declarations

Ethical approval

The study protocol was approved by the Non-Interventional Clinical Research Ethics Committee of Düzce University (approval number: 2026/160; date: May 11, 2026). All procedures were conducted in accordance with the ethical principles of the Declaration of Helsinki.

Informed consent

Due to the retrospective nature of the study and the use of anonymized clinical data, the requirement for written informed consent was waived by the ethics committee.

Use of AI tools

The authors declare that AI-based tools were used solely for language editing and improving readability. The authors take full responsibility for the scientific content of the manuscript.

Competing interests

The authors declare no competing interests.