Meta-analysis of the incidence and risk factors of postoperative delirium in organ transplant patients

Abstract

BACKGROUND

Postoperative delirium (POD) is a concerning complication of organ transplantation. With organ transplantation offering hope to patients with end-stage organ disease, understanding the incidence and risk factors of POD is crucial, as it can significantly affect patients’ prognosis and healthcare costs.

AIM

To systematically evaluate the incidence and risk factors of POD following organ transplantation to facilitate clinical prevention and optimize patient management and prognosis.

METHODS

Multiple databases such as PubMed and their reference lists were comprehensively searched using a combination of keywords related to organ transplantation and POD. Relevant observational studies on patients who had undergone solid organ transplantation and randomized controlled trials containing relevant analyses were included. Duplicated, data-deficient, non-English, and non-original data studies were excluded. Data were extracted independently by two researchers and then cross-checked. The Newcastle-Ottawa scale was used to evaluate the quality of the included studies. RevMan 5.3 was employed for data analysis. The pooled incidence of POD was calculated according to the data type, and the fixed or random effect model was employed to analyze risk factors based on heterogeneity. Subsequently, sensitivity analysis and publication bias assessments were performed.

RESULTS

A total of 39 relevant literatures were included. The overall incidence of POD in the organ transplant group was 20% [95% confidence interval (CI): 18%-22%]; liver transplant group, 22% (95%CI: 17%-26%); lung transplant group, 34% (95%CI: 23%-45%); and kidney transplant group, 6% (95%CI: 2%-10%). Primary graft dysfunction increased the POD risk, with a pooled odds ratio (OR) (95%CI) of 1.78 (1.09-2.91). A history of hepatic encephalopathy increased the POD risk, with a pooled OR (95%CI) of 3.19 (2.30-4.43). The higher the Acute Physiology and Chronic Health Evaluation II score, the greater the POD risk, with a pooled OR (95%CI) of 1.52 (1.09-2.12). A history of alcohol abuse increased the POD risk, with a pooled OR (95%CI) of 2.84 (1.74-4.65). Thus, the higher the model for end-stage liver disease score, the greater the POD risk, with a pooled OR (95%CI) of 2.49 (1.14-5.43). POD was more likely to develop in patients with preoperative infections, with a pooled OR (95%CI) of 2.78 (1.56-4.97). The use of diuretics increased the POD risk, with a pooled OR (95%CI) of 2.36 (1.38-4.04).

CONCLUSION

In this study, the overall incidence of POD in patients who underwent organ transplantation is 20%. The incidence varies among different types of organ transplantation, and multiple factors can increase the POD risk.

Key Words: Organ transplantation; Postoperative delirium; Incidence; Risk factors; Neurological complications; Meta-analysis

Core Tip: This meta-analysis aimed to precisely determine the incidence and risk factors of postoperative delirium (POD) in patients who underwent organ transplantation. By comprehensively searching multiple databases, 39 relevant studies were included. The overall incidence of POD was 20%, varying among different organ transplants. Factors such as primary graft dysfunction, hepatic encephalopathy history, high Acute Physiology and Chronic Health Evaluation II and model for end-stage liver disease scores, alcohol abuse history, preoperative infections, and diuretic use increased the risk, providing key insights for clinical practice.

INTRODUCTION

Organ transplantation, as one of the significant achievements in modern medicine, has brought hope in extending the life of numerous patients with end-stage organ diseases. However, postoperative complications have always been a crucial issue that affects patients’ prognosis and quality of life. Delirium is a relatively common central nervous system complication after surgery and is frequently observed in patients who undergo organ transplantation[1]. The manifestations of patients range from sluggishness and withdrawal to agitation and aggression.

Postoperative delirium (POD) is characterized by an acute brain dysfunction with disordered consciousness and reduced concentration ability[2]. POD can lead to prolonged hospital stays, increased medical costs, cognitive impairment, long-term functional disabilities, and even an increased mortality rate[3]. For patients who underwent organ transplantation, due to the complexity of surgery, special preoperative physical conditions of the patients, and need for long-term postoperative use of immunosuppressants and other multiple factors, the occurrence of POD may have more severe effects. The reported incidences of POD in patients who underwent organ transplantation vary considerably among medical institutions and research teams, ranging from approximately 10% to 50%[4]. Such differences may stem from various factors such as research designs, patient populations, assessment methods, and types of organ transplantation[5].

Previous meta-analyses on delirium after organ transplantation have made some contributions. For example, some earlier meta-analyses have mainly focused on a single type of organ transplantation, such as liver transplantation, and merely scratched the surface of risk factors, mainly highlighting patient age and preoperative comorbidities[6]. Others have combined data from multiple organ transplantations but failed to comprehensively consider the unique pathophysiological characteristics of different organ transplant scenarios[7]. In contrast, the present study aimed to bridge these gaps. Through a more comprehensive and in-depth meta-analysis, this study integrates data from a wider range of organ transplant types, including liver, lung, and kidney transplants. Moreover, this study will meticulously analyze a more extensive set of risk factors, such as primary graft dysfunction (PGD), history of hepatic encephalopathy, Acute Physiology and Chronic Health Evaluation (APACHE) II score, history of alcohol abuse, model for end-stage liver disease (MELD) score, preoperative infections, and use of diuretics, which were either overlooked or insufficiently explored in previous studies.

In addition, although many studies have attempted to explore the risk factors of POD in patients who underwent organ transplantation, the results remain controversial and were not analyzed comprehensively. Some studies have pointed out that factors such as patient age, graft type, operation duration, and preoperative cognitive function may be related[8]; however, a unified conclusion has not yet been reached. Therefore, through a systematic meta-analysis to integrate and analyze data from existing studies, accurate assessment of the incidence of POD in patients who underwent organ transplantation and elucidation of the related risk factors are significant. This will provide stronger evidence for clinicians in formulating prevention and management strategies and help improve postoperative recovery and long-term prognosis of patients who underwent organ transplantation.

MATERIALS AND METHODS

Search strategy

This review was registered in the International Prospective Register of Systematic Reviews (https://www.crd.york.ac.uk/prospero/, with registration number CRD42024627914). An exhaustive and systematic search was carried out across a multitude of globally recognized and authoritative databases, namely, PubMed, EMBASE, Web of Science, and Cochrane Library. The temporal scope of the search extended from database inception until August 31, 2024, to incorporate the most recent and cutting-edge research outcomes. Keywords pertinent to organ transplantation and POD were strategically selected for the retrieval process. For organ transplantation, specific keywords encompassed “organ transplantation”, “heart transplantation”, “liver transplantation”, “lung transplantation”, and “kidney transplantation” among others. For POD, associated keywords comprised “delirium”, “POD”, “post-surgical confusion”, and “neurologic events”. Concurrently, terms related to “incidence” and “risk factors” were integrated. Boolean logical operators, namely, “AND” and “OR” were meticulously applied to formulate intricate search expressions. For illustrative purposes, a representative search formula was as follows: (“organ transplantation” OR “heart transplantation” OR “liver transplantation” OR “lung transplantation” OR “kidney transplantation”) AND (“postoperative delirium” OR “POD” OR “post-surgical confusion”) OR “neurologic events” AND (“incidence” OR “risk factors”). In addition, a painstaking manual search was conducted on the bibliographies of the retrieved literature references, to uncover any potentially overlooked yet relevant research studies.

Inclusion and exclusion criteria

Inclusion criteria: (1) Study participants: Specifically, patients who have undergone organ transplantation were included, covering but not limited to solid organ transplants such as heart, liver, kidney, and lung transplants, without any limitations on transplantation techniques and modalities; (2) Research content: Studies must contain data on the incidence of POD or an analysis of the risk factors for POD. Whether such content is presented as the primary or secondary research objective will be taken into account; and (3) Study types: These comprise observational studies such as prospective cohort studies, retrospective cohort studies, and case-control studies. Randomized controlled trials that included an analysis related to POD will also be incorporated into the scope of this study.

Exclusion criteria: (1) Duplicate publications: Through a meticulous comparison of literature content, author information, and research data, completely identical, or highly similar duplicate studies are excluded. If a more updated or comprehensive version exists, the latest or most comprehensive one will be selected; (2) Data deficiency or inadequacy: A study that failed to provide sufficient data for calculating the incidence of POD or analyzing risk factors (such as the absence of key variables, unreported sample size, or incomplete statistical results) will be excluded; (3) Non-English literatures: Considering language consistency and resource limitations, only English literatures were included in this study; and (4) Review literatures, comment articles, conference abstracts, etc.: Because these literatures lack original research data or have not conducted independent data analyses, they were not incorporated into this meta-analysis.

Data extraction

Data extraction was independently carried out by two investigators who had undergone systematic training and were well-versed in the research content. Initially, the included studies were imported into literature management software (e.g., EndNote) for unified management. The extracted data encompassed the following aspects: (1) Basic study information: This included the name of the first author, year of publication, country or region where the study was conducted, and study type (such as prospective and retrospective cohort studies); (2) Patient characteristics: This includes the number of patients (N), age, proportion of male patients, and so forth; (3) Organ transplantation-related information: This encompassed the type of transplanted organ; (4) POD-related information: This covered POD assessment methods (such as the Diagnostic and Statistical Manual of Mental Disorders, confusion assessment method, and Richmond agitation-sedation scale), and the number of POD cases (n); and (5) Risk factor information: The risk factors involved in the study, name of each risk factor, measurement method, classification standard, and corresponding effect estimate [odds ratio (OR)] and its 95% confidence interval (95%CI) were meticulously recorded.

After data extraction, the two investigators cross-checked the extracted data. In the case of discrepancies, they were resolved through discussion and reference to the original literature. If an agreement still could not be reached, a third investigator was consulted for adjudication.

Quality assessment

For the observational studies (including cohort studies and case-control studies) included in this research, the Newcastle-Ottawa scale (NOS) was employed for quality evaluation. The NOS scale assesses three aspects: Selection of study participants, comparability between groups, and measurement of exposure or outcome, with a full score of 9 points. The specific evaluation criteria are as follows.

Selection of study participants (4 points): (1) Representativeness (1 point): This measures whether the study population can represent the target population, such as whether it is composed of consecutive cases or obtained through random sampling, etc.; (2) Sample size (1 point): This assesses whether the sample size is sufficient. The determination of whether the sample size meets statistical requirements is based on the study type and research content; (3) Exactness of exposure (1 point): This evaluates whether the diagnostic method for POD is accurate and reliable, and whether there are clear definitions and standards; and (4) Selection of the non-exposed group (1 point): For case-control studies, this examines whether the selection of the non-exposed group is appropriate and whether it is comparable to the exposed group.

Comparability between groups (2 points): (1) Consideration of the most important confounding factors in design and analysis (1 point): This assesses whether the study has adjusted or matched for the important confounding factors (such as age, sex, and underlying diseases) that may affect the results during the design or analysis stage; and (2) Degree of control of confounding factors (1 point): This assesses the adequacy of the control of confounding factors, such as through the implementation of stratified analysis, multivariate analysis, and other control methods.

Measurement of exposure or outcome (3 points): (1) Assessment of exposure (1 point): This item evaluates whether the assessment method for risk factors is accurate and objective, specifically regarding the reliability of obtaining information through medical records, questionnaires, and other means; (2) Assessment of outcome (1 point): This examines whether the assessment method for POD is independent and blinded (e.g., whether the assessor is unaware of the patient’s grouping situation) and whether the assessment tool has been validated; and (3) Follow-up time (1 point): For cohort studies, this evaluates whether the follow-up time is long enough to observe the occurrence of POD, whether there is any loss to follow-up, and whether the handling of the loss to follow-up is reasonable.

Based on the scores obtained from the NOS, the study quality was classified into three categories: High (7-9 points), medium (4-6 points), and low (0-3 points). In the result analysis and discussion, the potential effect of study quality on the results will be considered.

Statistical analysis

RevMan 5.3 was utilized for data analysis. In the analysis of the incidence of POD, corresponding methods were adopted to calculate the pooled incidence according to the data types reported in different studies (such as directly reported incidence, number of cases, and total sample size). If data were presented as a binary variable (such as the occurrence and non-occurrence of POD), the Mantel-Haenszel method was employed to calculate the pooled incidence.

For the analysis of risk factors, a meta-analysis was conducted based on the effect estimates (OR) and their 95%CIs reported in different studies. Initially, the heterogeneity among studies was evaluated by the Q test and I² statistic. The P value of the Q test was used to determine whether heterogeneity existed. If P < 0.1, heterogeneity was present among studies. The I² statistic was used to measure the magnitude of heterogeneity. I² < 25% indicated low heterogeneity; 25% ≤ I² < 50%, moderate heterogeneity; and I² ≥ 50%, high heterogeneity. If the heterogeneity among studies was relatively small (I² < 50%), a fixed effect model was adopted for the analysis. This model assumes that the effect sizes of all studies are identical, and its results are mainly influenced by studies with larger sample sizes. If the heterogeneity was large (I² ≥ 50%), a random effect model was utilized. This model considers the variation of effect sizes among studies, and the results are more conservative.

After the meta-analysis, a sensitivity analysis was performed to evaluate the stability of the results. In this study, the statistical analysis method was adjusted, that is, switching between the fixed effect model and random effect model, to observe whether significant changes occurred in the pooled effect size and other results. Simultaneously, publication bias was assessed. Begg’s and Egger’s tests were conducted, and a P value < 0.05 indicated potential publication bias.

RESULTS

Literature retrieval results

Figure 1 illustrates the literature retrieval process. Thirty-nine studies that met the criteria were screened.

Figure 1 Literature retrieval process.

Literature information extraction

Among the 39 studies, some studies reported the incidence of POD in patients who underwent organ transplantation between 2020 and 2024[9-37] (Table 1). A total of 67925 patients were involved, among whom 2950 experienced POD. Ten studies reported the risk factors of POD[25,38-46] (Table 2). Among them, 1 was about heart transplantation; 22 were liver transplantation; 10 were lung transplantation; 6 were kidney transplantation; 1 was intestinal transplantation; and 1 was heart and lung transplantation. The NOS scores of all the studies were > 6.

Table 1 Literature information reporting the incidence of postoperative delirium in organ transplant patients during 2020-2024.

Ref.	Country/region	Research type	Age (year)	Male	Transplanted organ	POD evaluation	N1	n2	NOS
Ahmed et al[9], 2021	Ontario, Canada	Prospective	57.1 ± 13.6	55%	Heart/lung	NA	78	28	8
Chang et al[10], 2022	Taiwan, China	Retrospective	50.57 ± 7.59	94.05%	Liver	CAM-ICU	84	46	8
Chen et al[11], 2020	Hunan, China	Retrospective	48.29 ± 9.14	84.28%	Liver	CAM-ICU	159	42	8
Cheng et al[12], 2020	Chengdu, China	Retrospective	NA	NA	Liver	CAM-ICU	242	36	7
Chu et al[13], 2022	MD, United States	Retrospective	NA	NA	Kidney	NA	894	43	7
Dalton et al[14], 2022	NC, United States	Retrospective	63.0 ± 6.4	65%	Lung	NA	100	38	7
DeBolt et al[15], 2021	San Francisco, United States	Retrospective	57 ± 11.5	141/143	Lung	CAM	236	34	9
Dragnich et al[16], 2023	TN, United States	Retrospective	49.8 ± 15.7	59%	Lung	CAM-ICU	60	28	9
Ebner et al[17], 2023	FL, United States	Retrospective	NA	NA	Heart	NA	25371	1482	7
Jin et al[18], 2023	Beijing, China	Retrospective	22-67	60.00%	Liver	NA	250	29	8
Kaderi et al[19], 2024	CA, United States	Retrospective	NA	NA	Lung	CAM	27	9	7
Khalil et al[20], 2020	Giza, Egypt	Retrospective	50.16 ± 8.38	90.60%	Liver	NA	149	20	8
Lee et al[21], 2021	Seoul, Korea	Prospective	NA	NA	Liver	NA	201	15	7
Lee et al[22], 2024	Seoul, Korea	Retrospective	NA	NA	Kidney	NA	296	14	7
Long et al[23], 2024	NY, United States	Retrospective	NA	NA	Kidney	NA	735	80	7
Lu et al[24], 2022	Hangzhou, China	Retrospective	52 ± 10	84.30%	Liver	DSM-IV	402	78	9
Ma et al[25], 2023	Changsha, China	Retrospective	47.1 ± 10.0	82.20%	Liver	CAM-ICU	321	62	9
Ma et al[26], 2023	Changsha, China	Retrospective	46.2 ± 9.8	82.94%	Liver	CAM-ICU	211	43	9
Mottaghi et al[27], 2022	Shiraz, Iran	Prospective	NA	NA	Liver	CAM-ICU	227	52	7
Mubashir et al[28], 2022	TN, United States	Retrospective	NA	NA	Lung	CAM-ICU	136	32	7
Ozay et al[29], 2022	Ankara, Turkey	Prospective	mean: 47.56	83.10%	Lung	NA	77	40	9
Park et al[30], 2020	Ansan, Korea	Retrospective	53.66 ± 7.93	27.69%	Liver	DSM-IV	325	69	9
Patel et al[31], 2024	FL, United States	Retrospective	2.0 (0.9, 8.5)	51%	Liver	CAM-ICU	63	47	9
			mean: 14.0	68%	Kidney	CAM-ICU	63	9	9
			mean: 4.0	63.33%	Intestinal	CAM-ICU	30	24	9
Patrono et al[32], 2020	Turin, Italy	Retrospective	NA	74.51%	Liver	CAM	306	41	8
Ri et al[33], 2020	Yangsan, Korea	Retrospective	53.53 ± 8.03	23.53%	Liver	DSM-IV	136	15	9
Ruck et al[34], 2024	United States	Retrospective	≥ 55	NA	Kidney	NA	35800	322	7
Tavabie et al[35], 2020	London, United Kingdom	Retrospective	≥ 18	NA	Liver	NA	793	113	7
Vandiver et al[36], 2022	CA, United States	Retrospective	58.4 ± 11.4	58%	Lung	CAM-ICU	114	50	9
Zhan et al[37], 2024	Hubei, China	Retrospective	53.34 ± 6.97	82.05%	Lung	NA	39	9	8

¹The total number of patients included in the study.

²The number of cases in which postoperative delirium occurred.

POD: Postoperative delirium; NOS: Newcastle-Ottawa Scale; NA: Not available; CAM: Confusion assessment method; ICU: Intensive care unit; DSM: Diagnostic and Statistical Manual of Mental Disorders.

Table 2 presents the literature information regarding the risk factors of postoperative delirium in organ transplant patients.

Ref.	Country/region	Research type	Age (year)	Male	Transplanted organ	POD evaluation	Risk factor	NOS
Anderson et al[38], 2018	PA, United States	Retrospective	mean: 61	63.23%	Lung	NA	a/b	8
Binda et al[39], 2017	Italian	Retrospective	NA	NA	Liver	CAM-ICU	c/d	7
Haugen et al[40], 2018	United States	Retrospective	50.3 ± 13.7	61.40%	Kidney	NA	a	8
Lee et al[41], 2013	Korea	Retrospective	51.00 ± 9.570	74.40%	Liver	DSM-IV	a/f	9
Lee et al[42], 2018	Korea	Retrospective	mean: 54	70.35%	Liver	CAM	f/d/g	9
Lescot et al[43], 2013	France	Retrospective	mean: 60	64.30%	Liver	DSM-IV	d	9
Ma et al[25], 2023	Changsha, China	Retrospective	47.1 ± 10.0	82.2%	Liver	CAM-ICU	i/c/f	9
Smith et al[44], 2018	NC, United States	Retrospective	mean: 54.5	59%	Lung	NA	b/h	8
Wang et al[45], 2014	Taiwan, China	Retrospective	53.40 ± 8.40	74.40%	Liver	RASS	e/c/d/g	9
Yoon et al[46], 2009	Seoul, Korea	Retrospective	50.05 ± 8.05	80.16%	Liver	DSM-IV	e/c/g	9

“e” was reported by Chang et al[10] in 2022; “c/f” was reported by Chen et al[11] in 2020; “a/h” was reported by Lee et al[22] in 2024; “f/c/I” was reported by Ma et al[25] in 2023; “a” was reported by Ruck et al[34] in 2024; “j” was reported by Park et al[30] in 2020 and Ri et al[33] in 2020. POD: Postoperative delirium; NOS: Newcastle-Ottawa Scale; NA: Not available; CAM: Confusion assessment method; ICU: Intensive care unit; DSM: Diagnostic and Statistical Manual of Mental Disorders; RASS: Richmond Agitation-Sedation Scale; a: Age; b: Primary graft dysfunction; c: History of hepatic encephalopathy; d: Acute Physiologic and Health Evaluation II score; e: Alcohol abuse history; f: Preoperative model for end stage liver disease score; g: Endotracheal intubation (day); h: Psychotropic medications use; i: Preoperative infections; j: With diuretics.

Meta-analysis of the incidence of POD in patients who underwent organ transplantation

Among the 29 studies that reported the incidence of POD patients who underwent organ transplantation between 2020 and 2024, 14 were about liver transplantation; 8 were lung transplantation; 4 were kidney transplantation; 1 was heart transplantation; and 1 was heart/lung transplantation, and 1 was simultaneously reported liver/kidney/intestine transplantation.

The results of the subsequent pooled analysis showed a 20% overall incidence of POD in patients who underwent organ transplantation (95%CI: 18%-22%) (Figure 2). The incidence of POD varied among patients with different types of organ transplantation. For example, the incidence of POD in the liver transplant group was 22% (95%CI: 17%-26%); lung transplant group, 34% (95%CI: 23%-45%); and kidney transplant group, 6% (95%CI: 2%-10%) (Figure 3).

Figure 2 Meta-analysis of the incidence of postoperative delirium in organ transplant patients from 2020 to 2024. CI: Confidence interval.

Figure 3 Subgroup analysis of the incidence of postoperative delirium in patients undergoing liver, lung, and kidney transplantations. CI: Confidence interval.

Meta-analysis of patient-related factors of POD in patients who underwent organ transplantation

The variation in age was not associated with the POD risk in patients who underwent organ transplantation, with a pooled OR (95%CI) of 1.11 (0.99-1.26) (Figure 4A). PGD increased the POD risk, with a pooled OR (95%CI) of 1.78 (1.09-2.91) (Figure 4B). A history of hepatic encephalopathy increased the POD risk, with a pooled OR (95%CI) of 3.19 (2.30-4.43) (Figure 4C). The higher the APACHE II score, the greater the POD risk, with a pooled OR (95%CI) of 1.52 (1.09-2.12) (Figure 4D). A history of alcohol abuse also increased the POD risk, with a pooled OR (95%CI) of 2.84 (1.74-4.65) (Figure 4E). Moreover, the higher the MELD score, the greater the POD risk, with a pooled OR (95%CI) of 2.49 (1.14-5.43) (Figure 4F).

Figure 4 Forest plot of patient-related factors. A: Age; B: Primary graft dysfunction; C: History of hepatic encephalopathy; D: Acute Physiologic and Health Evaluation II score; E: Alcohol abuse history; F: Model for end stage liver disease score. CI: Confidence interval.

Meta-analysis of operative-related factors of POD in patients who underwent organ transplantation

Endotracheal intubation was not a risk factor for POD development, with a pooled OR (95%CI) of 3.02 (0.89-10.24) (Figure 5A). The use of psychotropic medications was not associated with POD occurrence, with a pooled OR (95%CI) of 1.67 (0.97-2.88) (Figure 5B). Patients with preoperative infections were more likely to experience POD, with a pooled OR (95%CI) of 2.78 (1.56-4.97) (Figure 5C). The use of diuretics also increased the POD risk, with a pooled OR (95%CI) of 2.36 (1.38-4.04) (Figure 5D).

Figure 5 Forest plot of operative related factors. A: Endotracheal intubation (day); B: Psychotropic medications use; C: Preoperative infections; D: With diuretics. CI: Confidence interval.

Sensitivity analysis and publication bias assessment

The pooled analysis models for the aforementioned risk factors of POD were switched to observe the sensitivity of the meta-analysis results. The results demonstrated that, except for age, endotracheal intubation, and psychotropic medication use, no significant differences were found between the analysis results of the fixed (or random) effect model and the random (or fixed) effect model for the remaining POD risk factors. This finding suggests that the stability of these research results was relatively good (Figures 6 and 7). The publication bias of the included studies was evaluated by Begg’s rank correlation and Egger’s regression methods. The results showed no significant publication bias (Begg’s test, Z = 0.042, P = 0.763; Egger’s test, t = 3.168, P = 0.612).

Figure 6 Sensitivity analysis of patient-related factors. A: Age; B: Primary graft dysfunction; C: History of hepatic encephalopathy; D: Acute Physiologic and Health Evaluation II score; E: Alcohol abuse history; F: Model for end stage liver disease score. CI: Confidence interval.

Figure 7 Sensitivity analysis of operative related factors. A: Endotracheal intubation (day); B: Psychotropic medications use; C: Preoperative infections; D: With diuretics. CI: Confidence interval.

DISCUSSION

In this meta-analysis, the overall incidence of POD in patients who underwent organ transplantation was 20%, and differences were noted in the incidence among different types of organ transplantation. This result holds significant implications for clinical practice. The incidence of POD in the liver transplant group was 22%, which might be related to the crucial roles of the liver in various aspects such as metabolism and detoxification. Impaired liver function could affect the metabolism of neurotransmitters and the function of the blood-brain barrier, thereby increasing the POD risk. Patients with liver diseases often have high blood ammonia levels[47], which can interfere with the neural activities of the brain and may still have effects even after liver transplantation. The incidence of POD in the lung transplant group was as high as 34%, which might be closely associated with the gas exchange function of the lungs. Postoperative pulmonary complications and hypoxemia are relatively common, which can lead to cerebral hypoxia and trigger neurological dysfunction[48]. According to Shen et al[49], hypoxia can disrupt the brain’s energy metabolism and affect the function of nerve cells, thus promoting the occurrence of delirium. The 6% incidence of POD in the kidney transplant group also cannot be ignored. Kidney diseases are often accompanied by impaired water-electrolyte and acid-base balance. The fluctuations in the internal environment during recovery after transplantation may have adverse effects on the brain. The accumulation of toxins and imbalance of the internal environment in patients with chronic kidney diseases can affect the function of the central nervous system. Despite an improvement following kidney transplantation, a residual risk remains.

Organ transplantation is an effective treatment for end-stage organ diseases. However, POD has become a significant complication that affects patient prognosis. This meta-analysis has identified several risk factors associated with POD in patients who underwent organ transplantation. PGD is related to an increased risk of POD, which may be attributed to its disruption of liver metabolism, detoxification, and synthetic functions. For instance, if the liver fails to metabolize neurotoxic substances such as ammonia, ammonia will accumulate in the body and subsequently affect central nervous system function[50], inducing delirium. In addition, PGD-induced internal environment imbalance can alter neurotransmitter levels[51], interfering with the normal signal transduction pathways in the brain and increasing the POD risk. A history of hepatic encephalopathy is an important risk factor. In patients with such a history, neural tissues in the brain have already suffered varying degrees of damage and functional alterations. Even after liver transplantation, the previously impaired neural regulatory mechanisms cannot recover completely in a short period. Under the influence of multiple factors such as postoperative stress, patients are more prone to developing neuropsychiatric symptoms and experiencing delirium. According to Kawaguchi et al[52], the abnormal function of astrocytes in the brains of such patients persists, affecting glial-neuronal interactions and increasing the susceptibility to delirium.

Moreover, the higher the APACHE II score, the greater the POD risk. The APACHE II score comprehensively reflects the severity of the patient’s condition and the degree of physiological disturbance. A high score indicates the presence of multiple organ dysfunctions, severe inflammatory responses, and homeostatic imbalance. Inflammatory factors can cross the blood-brain barrier, activate the microglia, trigger a neuroinflammatory response, and damage neuronal structures and functions, thereby promoting POD occurrence. A study found significantly high levels of inflammatory mediators such as tumor necrosis factor-α and interleukin-6 in patients with a high APACHE II score[53], which are closely related to the occurrence of delirium.

A history of alcohol abuse increases the POD risk. This finding may be related to the fact that long-term alcohol abuse can lead to brain nerve cell damage, neurotransmitter system disorders, and nutrient deficiencies such as vitamin B1. Even after liver transplantation, the damaged nervous system repairs slowly, and the excitability of the central nervous system may change during alcohol withdrawal, inducing delirium. Long-term alcohol abuse causes adaptive changes in γ-aminobutyric acid (GABA)-ergic neurons in the brain. GABA, an inhibitory neurotransmitter, is related to the pathophysiology of neurodegenerative diseases[54].

Moreover, a high MELD score indicates poor liver reserve function and a poor general condition of the patient. This condition may lead to the accumulation of multiple metabolites, insufficient nutrient supply, and impaired immune regulation after surgery. For example, impaired liver function affects tryptophan metabolism, and its metabolites such as kynurenine can affect nerve function and promote delirium development. A study showed that the MELD score is associated with the metabolism of neurotransmitters in the brain (such as serotonin), indirectly demonstrating its relationship with delirium[55]. Patients with preoperative infections are more likely to experience delirium. Preoperative infections trigger a systemic inflammatory response, activate immune cells, and release numerous inflammatory factors. These inflammatory factors can affect the central nervous system through multiple pathways, such as directly acting on nerve cells and changing the permeability of the blood-brain barrier. A recent study indicated that preoperative infection-related inflammation can lead to dysfunction of the neurovascular unit in the brain, affecting cerebral blood perfusion and neural metabolism, which may be one of the molecular mechanisms inducing delirium[56].

The use of diuretics increases the risk of POD. Diuretics may lead to electrolyte disorders, particularly abnormal concentrations of ions such as sodium and potassium. Hyponatremia can cause brain cell edema, change the osmotic pressure balance in the brain, and affect the conduction of nerve impulses. Potassium ion disorders can affect the electrophysiological activity of cardiac myocytes, leading to changes in cardiac pumping function and indirectly affecting cerebral perfusion. In addition, diuretics may affect the stability of the membrane potential of nerve cells by changing electrolyte balance (such as potassium, sodium, chlorine, etc.), which may be a potential mechanism for POD occurrence; however, further exploration is needed.

Notably, several studies have indicated a strong correlation between POD and anesthesia-related factors[57,58]. Although none of the studies included in this study analyzed anesthesia, this aspect must be discussed. Anesthesia techniques can influence cerebral blood flow, oxygenation, and neurotransmitter release. For example, general anesthesia may disrupt the normal function of the central nervous system through its effects on ion channels and synaptic transmission. Different anesthetic agents have varying effect on the brain. Some volatile anesthetics increase the risk of delirium by interfering with the function of the GABA receptor, which is crucial in regulating neuronal excitability. Intravenously administered anesthetics may also affect neurotransmitter metabolism and lead to changes in the brain’s electrical activity. In addition, the duration of anesthesia may be a factor, as longer exposure to anesthetic agents may increase the brain’s vulnerability to delirium. In patients who underwent organ transplantation, the complex physiological state before and after surgery may further interact with anesthesia effects, increasing the risk for delirium. Future research should consider including anesthesia-related factors in the analysis to comprehensively understand the mechanisms of POD in patients who underwent organ transplantation.

The occurrence of POD in patients who underwent organ transplantation is the result of the combined action of multiple factors and involves complex mechanisms such as nerves, inflammation, and metabolism. A deeper understanding of these risk factors and their mechanisms of action helps in formulating targeted prevention strategies in clinical practice and improving the prognosis of patients who underwent organ transplantation. In this meta-analysis, the results of both Begg’s and Egger’s tests indicated the lack of significant publication bias (both P > 0.05). This finding implies that the included studies can, to a certain extent, objectively reflect the real incidence and risk factors of POD in patients who underwent organ transplantation. Although no obvious bias was detected in the current assessment, potential sources of bias that have not been detected cannot be completely excluded. Future related research should continuously focus on publication bias, and the channels and strategies for literature retrieval should be further broadened to ensure the high accuracy and reliability of the research results and provide a solid basis for clinical practice.

This study has some limitations: (1) Literature inclusion: Although 39 studies were included, some studies that met the criteria but were not included, leading to a certain deviation in the results; (2) Data heterogeneity: Studies may have varied patient characteristics, diagnostic criteria, intervention measures, etc., which affect the accuracy of data pooling and, in turn, the reliability of the results; and (3) Potential confounding factors: Not all confounding factors that affect the occurrence of POD have been fully considered, such as the psychological state of patients, making the analysis of the relationship between various risk factors and POD occurrence comprehensive.

CONCLUSION

This meta-analysis has determined the incidence of POD in patients who underwent organ transplantation and identified several risk factors, including PGD, history of hepatic encephalopathy, APACHE II score, history of alcohol abuse, MELD score, preoperative infections, and the use of diuretics. In clinical practice, corresponding preventive and intervention measures should be implemented by targeting these risk factors to reduce the incidence of POD and improve the prognosis of patients who underwent organ transplantation.