Early Use of Fentanyl Increases Risk of Delirium in Patients Transfer to the intensive Care Unit After Cardiac Surgery Immediately: A Target Trial Emulation

ABSTRACT

Aim

In the post‐operative period of cardiac surgeries, the majority of patients are admitted to the intensive care unit (ICU) for ongoing management, where delirium frequently occurs as a complication. However, the association between various analgesic regimens and the onset of postoperative delirium in ICU patients following cardiac surgery remains poorly understood. The present study employs a target trial emulation (TTE) framework to examine the effects of early postoperative pain management on the incidence of delirium in patients immediately admitted to the ICU after cardiac surgery.

Patients and Methods

Study participants were selected from the MIMIC‐IV version 3.1 database, comprising 5356 adult patients who were admitted for the first time and underwent cardiac surgery, followed by immediate transfer to the intensive care unit (ICU). The TTE framework was applied, utilizing the clone‐censor‐weighting (CCW, refer to the Statistical Analysis section) method for data analysis.

Results

In this study of 5356 adult patients who underwent cardiac surgery and were immediately transferred (within 24 h post‐surgery) to the intensive care unit postoperatively, in comparison with the morphine group, patients receiving fentanyl exhibited a significantly elevated risk of delirium within seven days postoperatively (hazard ratio [HR] = 1.69; 95% confidence interval [CI]: 1.41–1.95) and demonstrated an earlier onset of delirium.

Conclusion

The findings of this study, conducted within a TTE framework, indicate that the utilization of fentanyl for pain management during the initial 24 h following cardiac surgery is associated with a higher incidence and earlier onset of postoperative delirium compared to morphine.

Abbreviations

CABG

coronary artery bypass grafting

CAM‐ICU

the Confusion Assessment Method for the ICU

CCW

clone‐censor‐weighting

CIs

confidence intervals

CVD

cardiovascular and vascular diseases

GCS

Glasgow Coma Scale

HR

hazard ratio

ICDSC

the ICU Delirium Screening Checklist

ICU

intensive care unit

IQR

interquartile range

MIMIC IV

Medical Information Mart for Intensive Care IV

RCT

randomized controlled trial

RMST

restricted mean survival time

SAVR

surgical aortic valve replacement

SOFA

Sequential Organ Failure Assessment

TTE

target trial emulation

1. Introduction

As the global population continues to age at an increasingly rapid rate, the incidence of cardiovascular and vascular diseases (CVD) remains a major global health concern, with a continuous upward trend being observed on a worldwide basis [1]. Cardiac surgery has become a pivotal intervention for reducing mortality in patients with cardiovascular disease, and it is now widely performed across many countries [2]. It is evident that coronary artery bypass grafting (CABG) and surgical aortic valve replacement (SAVR) are among the most prevalent surgical procedures, and they have been instrumental in enhancing patient survival rates and quality of life [3, 4]. Nevertheless, the long‐term prognosis following cardiac surgery remains uncertain [1].

Delirium is a clinical manifestation of acute encephalopathy, with an incidence of up to 50% among critically ill patients [5, 6]. As a prevalent syndrome within the intensive care units, it is characterized by acute alterations in mental status, inattention, disorganized thinking, and other unexplained disturbances in consciousness [7]. The onset of delirium has been demonstrated to be associated with a multitude of adverse outcomes, including increased mortality, cognitive impairment, organ dysfunction, prolonged mechanical ventilation and the intensive care unit (ICU) stay, as well as elevated healthcare costs [8, 9].

The pathogenesis of delirium remains unclear, although it is widely believed to result from the interplay of multiple contributing factors [6, 9]. A plethora of studies have identified various medications, notably analgesics, sedatives, and benzodiazepines, as potential risk factors [9, 10]. Opioids, which are commonly administered for the control of pain following surgery, for pain management in mechanically ventilated patients, and for the management of other acute pain conditions in the intensive care unit (ICU), have been frequently examined in this context. A number of cohort studies have been conducted in order to ascertain the potential association between opioid use and the development of delirium in critically ill patients [11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21]. While some findings suggest a possible link, current evidence remains insufficient to establish a definitive causal relationship between opioid exposure and delirium in the ICU setting [9, 22, 23].

The majority of patients undergoing cardiac surgery are transferred directly to the ICU for postoperative care. Opioids are extensively utilized in both intraoperative anesthesia management and postoperative analgesia in this population. Nevertheless, conducting prospective studies in this setting poses significant challenges. The critical condition of ICU patients, the concurrent use of multiple medications, as well as issues such as ICU discharge or loss to follow‐up, can all introduce confounding factors that may influence study outcomes [22].

The extant research on the impact of early (first 24 h) analgesic treatment with morphine and fentanyl on the development of postoperative delirium in patients transferred to the ICU after cardiac surgery remains limited. Although randomized controlled trial (RCT) is widely regarded as the gold standard for evaluating the efficacy of interventions, they frequently encounter difficulties in reflecting the diversity of real‐world clinical practice. Elderly patients and those suffering from multiple comorbidities are frequently excluded due to eligibility constraints. Consequently, the study population is not random and the external validity and generalizability of RCT findings are reduced [24].

Cardiovascular diseases are prevalent among elderly patients with multiple comorbidities, posing significant challenges to the implementation of appropriate RCTs. At present, there is a conspicuous absence of RCT evidence concerning the impact of morphine and fentanyl on postoperative delirium in patients transferred to the ICU following cardiac surgery. The utilization of existing observational data to estimate real‐world causal effects presents a valuable alternative to traditional RCTs. In consideration of the particular clinical characteristics exhibited by cardiac surgery patients and the conventional utilization of morphine and fentanyl for the management of pain, this study employs a target trial emulation (TTE) framework to simulate an RCT. The objective of this study is to assess the effects of early postoperative opioid administration, within the first 24 h, on the incidence of delirium in patients immediately admitted to the ICU following cardiac surgery.

2. Methods

2.1. Study Design and Setting

The present study employs an observational data‐driven TTE to examine the effects of morphine/fentanyl on pain management during the early postoperative period (first 24 h) in patients immediately transferred (within 24 h post‐surgery) to the intensive care unit after cardiac surgery. Patients who underwent cardiac surgery during their initial admission were identified from the Medical Information Mart for Intensive Care (MIMIC‐IV) version 3.1 database using ICD‐9 and ICD‐10 codes (Table S1) [25]. The analgesics analyzed include morphine and fentanyl. Morphine administration records were extracted from the “inputevent” table (itemid = 225154), while fentanyl records were obtained from the same table with itemid 221744 or 225942 within the MIMIC‐IV ICU module. The primary outcome of this study is the occurrence of delirium. The identification of delirium was based on methods developed by the research team in collaboration with established diagnostic tools: the ICU Delirium Screening Checklist (ICDSC) and the Confusion Assessment Method for the ICU (CAM‐ICU). Despite the fact that the MIMIC‐IV database exclusively records CAM‐ICU assessments, delirium cases were additionally identified using nursing documentation by searching for key diagnostic terms. These terms were (1) delirium, (2) confusion, (3) agitation, and (4) altered mental status. These were extracted from the “chartevents” table (itemid = 220001). A meticulous review of each patient's medical record was conducted to ascertain that delirium assessments corresponded to an appropriate level of arousal. Patients lacking delirium assessments or without any record of evaluation were excluded from the study. Furthermore, during the process of data extraction, the Richmond Agitation‐Sedation Scale (RASS) scores of patients were verified to be ≥ –3, in order to ensure the presence of valid assessment conditions.

2.2. Data Source

All data for this study were obtained from the MIMIC‐IV, a publicly available database that offers comprehensive support for medical research. MIMIC‐IV (version 3.1) contains hospitalization records for approximately 364,627 patients who received emergency or intensive care at Beth Israel Deaconess Medical Center in Boston between 2008 and 2022. The database contains de‐identified patient information, thus obviating the need for informed consent. Following the completion of the requisite training programmer and the attainment of certification, one of our authors, Weisheng Chen, was granted access.

2.3. Eligibility Criteria

The present study employed a TTE framework, and the participants were patients over 18 years old who were admitted to the ICU for the first time, underwent cardiac surgery, and were immediately transferred back to the ICU postoperatively, with a minimum ICU stay of 24 h. “immediately transferred back to the ICU postoperatively” refers to patient admission to the ICU within 24 h after surgery, encompassing the operative day and the subsequent calendar day. In an effort to minimize the impact of any potential bias resulting from prior opioid exposure, the study's participants were restricted to individuals who received their initial dose of morphine or fentanyl on the day of surgery or within the subsequent 24 h. The exclusion criteria comprised the following: (1) patients who received their first documented administration of morphine or fentanyl before the surgery; (2) patients with other opioid use prior to cardiac surgery or within the admission to the ICU; (3) patients who experienced delirium before surgery or before receiving morphine or fentanyl (Figure 1).

FIGURE 1.

Inclusion and exclusion flowchart of the study. ICU, intensive care unit; MIMIC‐IV, Medical Information Mart for Intensive Care IV.

2.4. Treatment Strategy and Assignment

In this study, patients were categorized into two distinct groups, namely the “morphine analgesia group” and the “fentanyl analgesia group”, based on their utilization of morphine/fentanyl for the management of pain during the initial 24 h after the start of cardiac surgery and subsequent transfer to the ICU.

2.5. Follow‐Up and Outcomes

The follow‐up period extended from the time of ICU admission to 168 h (7 days) post‐transfer, with the primary outcome being the risk of postoperative delirium.

2.6. Statistical Analysis

The present study employs a TTE framework, utilizing the clone‐censor‐weighting (CCW) method to mitigate immortal time bias (Figure 2). A new‐user design was adopted in order to eliminate the biases that may have been associated with current users. In view of the observational nature of the data, it is not possible to ascertain the patients' treatment intentions. Consequently, causal inference in this emulated trial is based on per‐protocol effects [26].

FIGURE 2.

Study Design. RCT, randomized controlled trial.

The initial step in the methodology is to clone each patient in the original dataset, thereby creating two identical copies with the same baseline characteristics. One of these is then assigned to the “morphine analgesia group” and the other to the “fentanyl analgesia group”. The second step involves the application of artificial censoring to the cloned dataset, a process which involves the removal of patients who deviate from their assigned treatment protocol. Specifically, during the follow‐up grace period, if a patient is transferred to the morphine group from the fentanyl group, or vice versa, their data are censored. In order to account for the variability between the duration of opioid administration and the length of stay in the ICU, a 24‐h grace period is incorporated. The third step employs inverse probability of censoring weighting to correct for selection bias introduced by censoring. This approach adjusts for baseline, time‐varying, and post‐baseline covariates, with weights truncated at the first and 99th percentiles. Subsequent analyses employ the weighted dataset to generate cumulative risk curves, thereby comparing the risk of postoperative delirium between the morphine and fentanyl analgesia groups. The 95% confidence intervals (CIs) for the differences in delirium incidence and restricted mean survival time (RMST) were derived via 1000 iterations of the nonparametric bootstrap method.

The statistical analysis was conducted using R software (version 4.4.2, https://www.r‐project.org/). A p‐value less than 0.05 was considered to be statistically significant.

2.7. Sensitivity and Subgroup Analyses

In order to assess the robustness of the findings, this study conducted sensitivity analyses by varying the follow‐up grace periods, selecting 12 and 36 h to examine whether early morphine or fentanyl use alone for pain management in the ICU continued to influence postoperative delirium under different timeframes (Tables S2 and S3, Figures S1 and S2). All sensitivity analyses were performed using the CCW method and the same analytical procedures described above. Subgroup analyses were conducted based on age, gender, race, Sequential Organ Failure Assessment (SOFA) score, Glasgow Coma Scale (GCS) score, sepsis status, dexmedetomidine use, blood lactate and pH levels to evaluate the impact of these covariates on the results. A sensitivity analysis was added to include patients who received other opioid therapy before or within 24 h of ICU admission, acknowledging the clinical reality of multimodal analgesia in this typical cardiac surgery population. The calculation of all 95% CIs was conducted using 1000 nonparametric bootstrap samples.

3. Results

3.1. Baseline Clinical Characteristics

The final number of patients included in the study was 5356 (see Figure 1), with a median age of 69.00 years (interquartile range (IQR): [62.00, 76.00]). Of these, 1554 (29.01%) were female and 3802 (70.99%) were male. The baseline characteristics of the study participants are outlined in Table 1. The median SOFA score upon ICU admission was 4.00 (IQR: [3.00, 6.00]), and the median GCS score was 15.00 (IQR: [14.00, 15.00]). Within the cohort, 1309 patients (24.44%) were diagnosed with congestive heart failure, 1445 (26.98%) had a history of myocardial infarction, and 521 (9.73%) had been diagnosed with cerebrovascular disease.

TABLE 1.

Baseline characteristics of the study population.

Variable	Overall (n = 5356)
Personal characteristics
Age (years old)	69.00 [62.00, 76.00]
Sex (%)
Male	3802 (70.99)
Female	1554 (29.01)
Race (%)
White	3921 (73.21)
Other& a	1435 (26.79)
Scores on the first day in ICU
CCI	4.00 [3.00, 6.00]
SOFA	4.00 [3.00, 6.00]
GCS	15.00 [14.00, 15.00]
Blood gas and laboratory on the first day in ICU
Lactate (mmol/L)	2.60 [2.00, 3.40]
PH	7.45 [7.42, 7.48]
PaO2/FiO2	200.00 [139.00, 268.25]
Baseexcess (mmol/L)	−3.00 [–5.00, –1.00]
Glucose (mg/dL)	176.00 [156.00, 202.00]
White blood cell (109/L)	15.80 [12.60, 19.80]
Bun (mg/dL)	17.00 [14.00, 21.00]
Creatinine (mg/dL)	0.90 [0.80, 1.10]
Comorbidities
Cerebrovascular disease (%)
Yes	521 (9.73)
No	4835 (90.27)
Myocardial infarct (%)
Yes	1445 (26.98)
No	3911 (73.02)
Congestive heart failure (%)
Yes	1309 (24.44)
No	4047 (75.56)
Chronic pulmonary disease (%)
Yes	1084 (20.24)
No	4272 (79.76)
Renal disease (%)
Yes	796 (14.86)
No	4560 (85.14)
Metastatic solid tumor (%)
Yes	15 (0.28)
No	5341 (99.72)
Liver disease (%)
Yes	152 (2.84)
No	5204 (97.16)
Dementia (%)
Yes	26 (0.49)
No	5330 (99.51)
Diabetes (%)
Yes	1925 (35.94)
No	3431 (64.06)
Peptic ulcer disease (%)
Yes	27 (0.50)
No	5329 (99.50)
Paraplegia (%)
Yes	43 (0.80)
No	5313 (99.20)
Sepsis (%)
Yes	2706 (50.52)
No	2650 (49.48)
Drugs records before using morphine or fentanyl
Antipsychotics (%)
Yes	28 (0.52)
No	5328 (99.48)
Benzodiazepines (%)
Yes	326 (6.09)
No	5030 (93.91)
Dexmedetomidine (%)
Yes	581 (10.85)
No	4775 (89.15)
Vasopressors (%)
Yes	1479 (27.61)
No	3877 (72.39)
Outcomes
Incidence of delirium occurrence (%)	605 (11.30)

^a Other& includes individuals identified as: Black/African, Black/African American, Black/Cape Verdean, Black/Caribbean Island (n = 206).

Asian, Asian—Asian Indian, Asian—Chinese, Asian—South East Asian, Asian—Korean (n = 124); Hispanic or Latino, Hispanic/Latino—Central American, Hispanic/Latino—Columbian, Hispanic/Latino—Cuban, Hispanic/Latino—Dominican, Hispanic/Latino—Guatemalan, Hispanic/Latino—Honduran, Hispanic/Latino—Mexican, Hispanic/Latino—Puerto Rican, Hispanic/Latino—Salvadoran, Portuguese, South American (n = 178); American Indian/Alaska Native, Native Hawaiian or other Pacific Islander (n = 10); multiple race/ethnicity, other, patient declined to answer, unable to obtain, unknown (n = 917).

3.2. Primary and Secondary Outcomes

All patients included in this study were immediately transferred (within 24 h post‐surgery) to the ICU following cardiac surgery, with the time of transfer serving as the starting point for follow‐up. Within seven days of being admitted to the ICU, 605 patients (11.30%) developed delirium, with 211 (3.94%) of these patients receiving morphine analgesia and 394 (7.36%) receiving fentanyl analgesia. As demonstrated in Figure 3, the weighted cumulative risk curve indicates a higher risk of delirium in the fentanyl group compared to the morphine group.

FIGURE 3.

Association between morphine or fentanyl analgesia alone in patients transfer to the intensive care unit after cardiac surgery immediately (within 24 h post‐surgery) and the development of delirium within 168 h (7 days). CCW, clone censor weighting; SOFA, Sequential Organ Failure Assessment; GCS, Glasgow Coma Scale.

Following the implementation of CCW, patients in the fentanyl group increased to 1.69 times risk of delirium within seven days following surgery when compared to those in the morphine group (hazard ratio [HR]: 1.69, 95% CI: 1.41 to 1.95) (Table 2). Furthermore, the restricted mean survival time (RMST) to first delirium onset was 6.20 h shorter in the fentanyl group (95% CI: 4.78 to 7.97), indicating an earlier onset of delirium symptoms. While the absolute difference was modest, it remained statistically significant.

TABLE 2.

Outcomes in patients with only use morphine/fentanyl in the ICU after clone‐censor‐weight (follow‐up grace period of 24 h).

Outcomes	Morphine group (n  = 5356)	Fentanyl group (n  = 5356)
Primary outcome
Delirium (%)	7.43%	12.20%
Differences in the HR of delirium (95% CI a )	1.69 (1.41 to 1.95)
Difference in RMST (days, 95% CI a )	6.20 (4.78 to 7.97)

Abbreviations: CI, confidence interval; ICU, intensive care unit; RMST, restricted mean survival time.

^a All 95% CIs were calculated from 1000 bootstrap replicates.

3.3. Sensitivity and Subgroup Analyses

In this study, two sensitivity analyses were performed by adjusting the follow‐up grace periods. The findings remained consistent with those of the primary analysis (see Tables S2 and S3, Figures S1 and S2). When the aforementioned grace period was set to 12 h, patients in the fentanyl analgesia group increased to 1.54 times risk of developing delirium within 7 days following surgery in comparison to those in the morphine group (HR: 1.54, 95% CI: 1.26 to 1.81), accompanied by a RMST disparity of 6.28 h (95% CI: 4.04 to 8.65) (Table S2, Figure S1). Following the extension of the grace period to 36 h, the risk in the fentanyl group increased to 1.77 times that of the morphine group (HR: 1.77, 95% CI: 1.50 to 2.04), with a RMST difference of 6.24 h (95% CI: 4.73 to 7.77) (Table S3, Figure S2). These findings serve to reinforce the robustness of the primary results, thereby demonstrating a consistently higher risk and earlier onset of delirium associated with early fentanyl use alone. Results from the multimodal analgesia sensitivity analysis (Tables S4–S7, Figures S3–S5) showed consistent trends: compared to morphine, fentanyl was associated with an increased incidence and an accelerated onset of delirium.

In the sensitivity analysis, the increased risk of developing delirium within 7 days postoperatively and the shorter RMST observed in the fentanyl analgesia group remained statistically significant when compared with the morphine analgesia group. The findings of this study indicate that patients administered fentanyl alone are more prone to developing postoperative delirium and experience its onset at an earlier stage following surgery.

As illustrated in Figure 4, the association between the early postoperative administration of morphine or fentanyl alone and the subsequent 7‐day delirium risk is investigated across diverse subgroups.

FIGURE 4.

Forest plot for subgroup analysis after using the clone‐censor‐weight method. SOFA, Sequential Organ Failure Assessment; GCS, Glasgow Coma Scale. Other&: See the Table 1 notes for details.

Figure 4 demonstrates the association between early analgesia with morphine or fentanyl monotherapy alone and the risk of delirium onset within 7 days across all subgroups. This observation is consistent with the primary analysis. The delirium risk showed significant differences between analgesic groups when stratified by variables including age, sex, race, sepsis status, dexmedetomidine use, SOFA score, GCS score, and key laboratory parameters (lactate, pH). This finding reinforces the robustness of the observed association across diverse clinical profiles.

4. Discussion

In the context of the expanding global utilization of opioids in clinical practice, there has been an observed escalation in adverse outcomes related to opioids, including hallucinations and delirium [27]. Morphine, a naturally occurring opioid with a long history of clinical use and broad availability, has traditionally been one of the most frequently implicated agents in cases of opioid‐induced delirium [28, 29, 30]. However, with the increasing utilization of potent and fast‐acting synthetic opioids, such as fentanyl, the differential impact of natural versus synthetic opioids on delirium risk has garnered increasing attention [27]. Fentanyl and morphine exhibit distinct pharmacological profiles. Fentanyl is a highly lipophilic semi‐synthetic opioid that rapidly crosses the blood–brain barrier. It has a fast onset of action, distributes quickly into lipid‐rich tissues, and is metabolized in the liver. Its distribution into tissues leads to a prolonged context‐sensitive half‐time, thereby extending its duration of action during prolonged infusions [31]. In contrast, morphine is primarily metabolized in the liver (>90%), where it is rapidly converted into active glucuronide metabolites: mainly morphine‐3‐glucuronide (M3G), with smaller amounts of morphine‐6‐glucuronide (M6G) and normorphine. All three metabolites are pharmacologically active. M6G is thought to contribute to morphine's analgesic effects, whereas M3G exhibits neuroexcitatory (potentially proconvulsant) properties. These metabolites can accumulate, particularly in patients with renal impairment, and may prolong effects even after infusion is stopped or the rate is reduced. In patients with renal failure, the clearance of morphine metabolites is markedly altered, which can lead to toxicity due to metabolite accumulation [32].

In this context, the present study, grounded in a TTE framework and closely aligned with real‐world clinical practice, aimed to investigate the effects of early (within the first 24 h) postoperative analgesia using morphine or fentanyl alone on the risk of delirium in patients immediately transferred(within 24 h post‐surgery) to the ICU following cardiac surgery.

The present study adopts a TTE framework, grounded in the principles of RCT design, to estimate the causal effects of interventions using observational data. The TTE is comprised of two fundamental stages. The initial stage encompasses the specification of the hypothetical randomized trial, herein designated as the “target trial.” This trial is predicated on the objective of elucidating the causal research question. This step necessitates meticulous delineation of all critical trial elements, encompassing explicit inclusion and exclusion criteria, treatment strategies and assignment mechanisms, well‐defined follow‐up initiation and endpoints, outcome measures, causal contrasts of interest, and appropriate statistical methodologies. The second stage involves the analysis of observational data in a manner that closely aligns with the design specifications of the emulated trial, ensuring consistency and methodological rigor throughout the simulation process.

In clinical practice, treatment decisions are often complex, and RCTs may not be feasible or generalizable to certain populations, such as elderly patients, who are frequently underrepresented in trials. Moreover, the evaluation of the sustained effects or long‐term efficacy of pharmacological interventions poses additional challenges, thereby complicating the replication of the rigor of RCTs. In such contexts, the necessity for empirical decision‐making becomes apparent, requiring the utilization of high‐quality observational data to generate credible evidence. Target trial emulation provides a structured approach to achieving this objective, enabling researchers to explicitly define the circumstances in which study findings may inform clinical decision‐making. In this study, a target trial emulation protocol was developed to assess the impact of using morphine or fentanyl alone on the risk of postoperative delirium among patients who were immediately transferred to the ICU following cardiac surgery. The target trial clearly delineated the study population, intervention strategies, follow‐up period, and outcome of interest. The findings of this study indicate that, in comparison with morphine alone, the utilization of fentanyl alone for the purpose of analgesia during the initial 24 h following ICU admission after cardiac surgery is associated with an elevated incidence of delirium and a more precipitous onset of delirium symptoms within 7 days following surgery.

A comparison of the present study with previous research indicates partial consistency, whilst significant disparities remain. While some studies have reported no significant association between opioid use after cardiac surgery and the incidence of postoperative delirium, other research has suggested a more nuanced relationship [33, 34]. For instance, a prospective study encompassing patients from internal medicine, surgical, cardiovascular, and neurology ICUs found that both natural opioids (e.g., morphine) and synthetic opioids (e.g., fentanyl) were associated with an increased risk of delirium in critically ill ICU patients [22]. It has been indicated by prior prospective studies in surgical and trauma ICU settings that commonly used opioid analgesics are effective for pain management. However, these studies have also demonstrated that they pose a potential risk for delirium. It is noteworthy that in the surgical SICU, the administration of fentanyl has been associated with a significantly elevated risk of delirium, while morphine has been observed to exert a protective effect. One postulated explanation for this discrepancy is that the continuous administration of high doses of fentanyl, in comparison to the typically lower doses of morphine, may be a contributing factor [17]. Moreover, earlier prospective studies on surgical patients have also corroborated the finding that morphine use may be associated with a reduced incidence of delirium [35].

A recent retrospective study examining the perioperative use of morphine and fentanyl in ICU patients undergoing cardiac surgery demonstrated that perioperative fentanyl use was associated with a higher risk of postoperative delirium and poorer short‐term outcomes compared to morphine [36]. Furthermore, other studies have suggested that, although preoperative opioid use may not be significantly associated with the incidence of postoperative delirium following cardiac surgery, patients receiving higher intraoperative doses of fentanyl were more likely to develop postoperative delirium than those who received lower doses [33, 37].

At present, there is an absence of robust randomized controlled trial evidence regarding the impact of using morphine or fentanyl alone for pain management during the early postoperative period (first 24 h) following cardiac surgery, when patients are immediately transferred(within 24 h post‐surgery) to the ICU. This study, utilizing a target trial emulation framework, is pioneering in its demonstration that, in contrast to use morphine alone, early postoperative administration of using fentanyl alone is associated with a heightened incidence and accelerated onset of delirium in patients undergoing cardiac surgery. The findings of this study suggest that, in the context of ICU pain management after cardiac surgery, morphine may serve as a safer alternative to fentanyl in reducing the risk of postoperative delirium. Future studies should focus on optimizing opioid use strategies, taking into account such factors as timing, dosage, and patient‐specific issues, with a view to supporting more individualized and safer analgesic regimens.

5. Strengths and Limitations

The present study boasts several notable strengths. Firstly, it employs a target trial emulation design, which assists in mitigating confounding bias that is inherent in observational data and enables more accurate estimation of causal effects. Secondly, the analysis is based on the MIMIC‐IV database, which provides a large and diverse sample of cardiac surgery patients, thereby enhancing the generalizability of the findings. Thirdly, the incorporation of sensitivity analyses with varying follow‐up grace periods enables a robust evaluation of the impact of early (first 24 h) postoperative analgesia using morphine or fentanyl alone on the risk of postoperative delirium. This further reinforces the credibility of the results.

It is evident that this study is not without its limitations. Firstly, as with all observational studies, there remains a risk of residual confounding due to unmeasured variables, such as specific intraoperative management strategies or additional postoperative complications, which could not be fully accounted for. Secondly, although the MIMIC‐IV database provides high‐quality data, it represents a single‐center dataset from a large academic medical center, which may limit the generalizability of the findings to other healthcare settings with more diverse patient populations. The validation of these results in future multicenter studies is indicated. Thirdly, while the study concentrated on the early (first 24 h) administration of morphine or fentanyl following cardiac surgery, it did not consider the potential influence of other opioids or adjunctive factors such as hemodynamic stability, both of which may substantially impact delirium risk. In conclusion, the 7‐day follow‐up period‐imposed limitations on the assessment, confining it to short‐term outcomes. This precluded the evaluation of the potential long‐term effects of natural versus synthetic opioid use on postoperative recovery and prognosis. It is recommended that future research incorporate extended follow‐up periods with the aim of better understanding the enduring impact of early postoperative analgesia strategies in this patient population.

6. Conclusion

The findings of this study, conducted within a TTE framework, indicate that using fentanyl alone for pain management during the initial 24 h following cardiac surgery is associated with a higher risk and earlier onset of postoperative delirium compared to use morphine alone. The findings of this study indicate that, in the management of postoperative pain in cardiac surgery patients admitted to the ICU, natural opioids such as morphine may be preferable to synthetic alternatives like fentanyl, with a view to mitigating the risk of delirium.

Funding

This work was supported by the Science and Technology Projects in Guangzhou (Nos. 2023A03J0592 and 2025A03J4246) and the National Natural Science Foundation of China (Nos. 82572508).

Ethics Statement

The Medical Information Mart for Intensive Care IV (MIMIC‐IV) database was supported by grants from the National Institute of Biomedical Imaging and Bioengineering (NIBIB) of the National Institutes of Health (NIH) under award numbers R01‐EB001659 (2003‐2013) and R01‐EB017205 (2014‐2018) and approved by the Institutional Review Boards of Beth Israel Deaconess Medical Center (Boston, MA) and the Massachusetts Institute of Technology (Cambridge, MA).

Informed Consent

Data extracted from the MIMIC‐IV database do not require individual informed consent because the research data are publicly available, and all patient data are de‐identified.

Conflicts of Interest

The authors declare that they have no competing interests.

Supporting information

Table S1: ICD‐9 and ICD‐10 codes for identifying cardiac surgeries.

Table S2: Outcomes in patients with only use morphine/fentanyl in the ICU after clone‐censor‐weight (follow‐up grace period of 12 h).

Table S3: Outcomes in patients with only use morphine/fentanyl in the ICU after clone‐censor‐weight (follow‐up grace period of 36 h).

Table S4: Baseline characteristics of the study population.

Table S5: Outcomes in patients with only use morphine/fentanyl in the ICU after clone‐censor‐weight (follow‐up grace period of 12 h).

Table S6: Outcomes in patients with only use morphine/fentanyl in the ICU after clone‐censor‐weight (follow‐up grace period of 24 h).

Table S7: Outcomes in patients with use morphine/fentanyl in the ICU after clone‐censor‐weight (follow‐up grace period of 36 h).

Figure S1: Association between morphine or fentanyl analgesia alone in patients transfer to after cardiac surgery immediately (within 12 h post‐surgery) and the development of delirium within 168 h (7 days).

Figure S2: Association between morphine or fentanyl analgesia alone in patients transfer to after cardiac surgery immediately (within 36 h post‐surgery) and the development of delirium within 168 h (7 days).

Figure S3: Association between morphine or fentanyl analgesia alone in patients (not censored at opioid switching or initiation of another opioid therapy) transfer to after cardiac surgery immediately (within 12 h post‐surgery) and the development of delirium within 168 h (7 days).

Figure S4: Association between morphine or fentanyl analgesia alone in patients (not censored at opioid switching or initiation of another opioid therapy) transfer to after cardiac surgery immediately (within 24 h post‐surgery) and the development of delirium within 168 h (7 days).

Figure S5: Association between morphine or fentanyl analgesia alone in patients (not censored at opioid switching or initiation of another opioid therapy) transfer to after cardiac surgery immediately (within 36 h post‐surgery) and the development of delirium within 168 h (7 days).

Data Availability Statement

The data were available on the MIMIC‐IV website at https://mimic.physionet.org/. The data in this article can be reasonably applied to the corresponding author.