Between July 2018 and March 2019, n = 320 patients undergoing cardiac surgery or percutaneous valve replacement were screened for POD at a university hospital as described in a previous study (10). Of these patients 179 were prospectively included into this study for a one-year follow-up visit (1yFU) between July 2019 and March 2020. 141 patients could not be included in the 1yFU due to refusal to participate (n = 63), death (n = 16), incomplete data (n = 5), inability to attend follow-up due to poor general health (n = 10) or current hospitalization (n = 7) during the study period, unavailability for contact (n = 7), long distance from home (n = 7), and early termination of the study due to the COVID-19 pandemic (n = 26).

The study was approved by the Ethics Committee at Hannover Medical School (local ethics committee approval No. 7876) and was conducted in accordance with the Declaration of Helsinki. Written informed consent was obtained from all study participants.

At baseline and at 1-year follow-up medical history and clinical characteristics such as age, sex, NYHA classification and cardiovascular risk factors were assessed. In addition, perioperative clinical data (type, duration of surgery or number of erythrocyte concentrates received) were collected. Screening for POD was performed twice daily on the first, second and third day after cardiac surgery or percutaneous valve replacement (10). For the detection of POD either the Confusion Assessment Method (3D-CAM) or the Confusion Assessment Method Intensive Care Unit (CAM-ICU) was used, depending on the patient's level of agitation and alertness. If the Richmond Agitation-Sedation Scale (RASS)-Score was between −1 and +1, the 3D-CAM was assessed; otherwise, CAM-ICU was used. The CAM algorithm includes four features: (1) acute onset and fluctuating course, (2) inattention, (3) disorganized thinking, and (4) altered level of consciousness. The presence of both the first and the second feature and either the third or the fourth feature defines a positive screening test for POD (11).

Patient groups were classified according to the presence or absence of diagnosed POD. Group comparisons were performed using the Mann–Whitney U-test for continuous variables and the Pearson chi-squared test for group differences in categorical variables.

To assess whether POD is an independent risk factor for poor neuropsychiatric outcome and health-related quality of life (physical score), a multivariable linear regression analysis was conducted. The outcome parameters at 1yFU were used as dependent variates (MoCA at 1yFU, Score A + B and the Physical Component Summary of the SF-12). Parameters that differed significantly between patients with and without POD [age, difference between MoCA score at 1yFU and preoperative MoCA score (ΔMoCA)] and clinically relevant outcome factors (rehospitalization, NYHA classification) were included as independent variables. Due to a strong association with MoCA at 1yFU and Score A + B, ΔMoCA was excluded as an independent variable from the respective analyses.

Values of p < 0.05 were considered statistically significant. Statistical analyses were performed with SPSS for Windows (IBM Corp. Released 2021. IBM SPSS Statistics for Windows, Version 28.0. Armonk, NY: IBM Corp).

A total of 179 patients received 1yFU after cardiac surgery or intervention. The median age of the cohort was 72.5 years. 35.8% of the patients were female (Table 1). 146 patients underwent elective cardiac surgery [cardiac valve surgery, coronary artery bypass graft (CABG), off-pump CABG, aortic surgery, combined CABG and cardiac valve surgery or transapical valve implantation], while 33 patients underwent transcatheter aortic valve replacement (TAVR) or catheter-based mitral valve reconstruction (MitraClip®, Abbott Vascular, Chicago, USA) (Table 2). Of these 179 patients, 43% (n = 77) had POD at the index event. During the year following the procedure, 61 patients required at least one hospital readmission (34.1%) and 19 patients required an ICU readmission (10.6%). Of the patients who were readmitted, 32.8% (n = 20) had more than one readmission to the hospital. The most common indications for rehospitalization were cardiac or other medical reasons (Table 3).

Patients were compared according to their POD status. Patients with POD were older (p = 0.005) and had worse preoperative cognitive performance [preoperative MoCA: POD 23 (IQR 21–25.5) vs. without POD 25 (IQR 23–26), p < 0.001]. Readmission rates were not significantly different between patients with or without POD (p = 0.379). Clinical characteristics grouped by POD status are shown in Table 1.

134 (74.9%) of 179 patients had at least one neurological or neuropsychiatric symptom one year after surgery or intervention. The most common neurological complaints were gait disturbance (n = 71), abnormal epicritic sensibility (n = 58), paresis (n = 58) and headache (n = 25). 41.9% of patients (n = 75) had chronic pain and 34.1% had sleep disturbances (n = 61) at 1yFU. Of these, 31 patients had difficulty falling asleep, 18 had difficulty staying asleep, and 12 had both problems. Patients experiencing chronic pain had a median score of 3 on the Numeric Pain Rating Scale (NPRS, 0–10). Comparing the median NPRS scores of patients with chronic pain with POD (n = 33, median score: 5) and without POD (n = 42, median score: 3), there was no significant difference (p = 0.088).

We compared the composite neurological and neuropsychiatric scores (score A + B) at 1yFU in patients with and without POD. In univariate analysis, patients with POD had higher score A + B, indicating worse neurologic and neuropsychiatric outcome (p < 0.001). Differences in neurological score (score A) and neuropsychiatric score (score B) remained for each category, with worse results in patients with POD (Table 3).

Multivariable linear regression analysis including the independent variables POD status, age, rehospitalization and NYHA classification revealed that POD, age, NYHA classification and rehospitalization were associated with a higher score A + B at 1 year, indicating a worse neurological and neuropsychiatric outcome (POD status: b:1.172; 95%-CI, 0.070–2.273, p = 0.037; age: b:0.134; 95%-CI, 0.086–0.182, p < 0.001; NYHA classification at 1 year: b:1.998; 95%-CI,1.169–2.828, p < 0.001; rehospitalization b:1.786; 95%-CI, 0.640–2.932, p = 0.002).

The median preoperative MoCA score was 24 (IQR 22–26), while the median MoCA score at 1yFU was 25 (IQR 23–28, p < 0.001). When comparing 1yFU levels with baseline values the median MoCA score improved in both, patients with and without POD although the improvement was less apparent in the delirium group (p = 0.027). Among the individual MoCA items, executive function/visuospatial ability (p = 0.001), verbal ability (p = 0.001), short-term memory (p = 0.003) and orientation (p < 0.001) were the items with lower scores in patients with POD (Table 3). When POD status, age, rehospitalization and NYHA classification at 1yFU were included in a multivariable regression analysis, POD and NYHA classification were independently associated with worse MoCA scores at 1 year [POD status: b:-1.955, 95%-CI: −2.839-(−1.071), p < 0.001, age: b: −0.047, 95%-CI: −0.086-(−0.008), p = 0.018, NYHA classification: b: −1.059, 95%-CI: −1.727-(−0.391), p = 0.002, rehospitalization b: −0.625, 95%-CI: −1.545–0.295, p = 0.181].

At 1yFU, 36 patients (20.1%) had abnormal BDI scores >10, indicating depressive symptoms. 72 patients (40.2%) had abnormal FIS scores >40, indicating the presence of fatigue. When compared bilaterally, the prevalence of abnormal BDI and FIS scores was not significantly different between patients with and without POD (Table 3).

Patients with POD had more frailty (p = 0.028) and a lower Barthel Index (p = 0.025) (Table 3). Both the SF-12 Physical Component Summary (p = 0.022) and SF-12 Mental Component Summary (p = 0.048) were significantly lower in patients with POD. Multivariable regression analysis demonstrated an association of POD, NYHA classification at 1yFU and rehospitalization with the SF-12 physical component summary scores one year after cardiac surgery or intervention [POD status: b:−2.847, 95%-CI: −5,670-(−0,024), p = 0.048; NYHA classification: b:−6.176, 95%-CI: −8.309-(−4.043), p < 0.001; rehospitalisation b:−3.826, 95%-CI: −6.778-(0.875), p = 0.011; age: b:−0.038, 95%CI: −0.166–0.090, p = 0.56; ΔMoCA: b:−0.061, 95%CI: −0.642–0.52, p = 0.836).

A subgroup analysis of delirium subtypes (hyperactive, hypoactive and mixed subtype) was performed (Supplementary Table S1). No significant differences were observed between these subtypes regarding neurologic and neuropsychiatric outcomes.

Consistent with our findings, in a recent retrospective cohort study, including 4,033 patients, Brown et al. discovered an association between POD and new-onset neurocognitive disorders (such as dementia or memory impairment) one year after their ICU stay (16). However, no association was found between POD and depressive episodes, anxiety, and trauma-and-stressor-related disorders. In contrast to our study, the information on the patients’ neuropsychiatric disorders was obtained retrospectively from the medical records, which does not allow any conclusions to be drawn about the patients’ clinical characteristics or even individual symptoms.

In the present study, we found that both, patients with and without POD, complained of various neurological symptoms such as gait disturbance, bradykinesia and paresis one year after surgery. Many patients also experienced symptoms of depression (20.1%) and fatigue (40.2%) at 1yFU. These findings underscore the need for outpatient follow-up for these patients after discharge from the hospital. Various concepts of post-ICU follow-up services are currently being investigated to address various long-term complaints of these patients (17, 18). They vary in frequency of consultation (e.g., monthly, once after six months), specific interventions, and patient population (all patients after ICU stay vs. at-risk patients), but they only partially address neuropsychiatric complaints such as depression or cognitive function (17). POD is one of several factors influencing long-term neurologic and neuropsychiatric outcomes. These outcomes are determined by a complex interplay of multiple elements, including postoperative neurologic events, overall clinical course, and rehospitalizations. Moreover, the majority of patients participated in cardiac rehabilitation, which may be a significant contributor and could have exerted a beneficial effect on neurologic outcomes. Notwithstanding these considerations, our findings indicate that POD serve as a predictor for long-term outcome parameters and could facilitate the identification of patients who may benefit from structured follow-up and intensified postoperative care.

In addition, no significant differences regarding neurologic and neuropsychiatric outcomes are observed between the subtypes of POD (hyperactive, hypoactive and mixed subtype). However, due to smaller subgroup sizes, these results are not yet conclusive. Larger sample sizes will be required for a more robust evaluation of these subgroups.

There are several interpretations and explanations for the association of POD with long-term neurological and neuropsychiatric outcomes. First, POD may reflect pre-existing impaired cerebral function or cognitive reserve prior to surgery or intervention. This pre-existing impairment may be revealed by the effects of non-specific insults, such as surgery-induced neuroinflammation and anesthesia, on the vulnerable brain (19). Affected patients are more likely to suffer from POD and might carry an increased risk of developing neurological and neuropsychiatric symptoms in the long term. The close association of cognitive decline with age and the higher degree of heart failure may reinforce these observations (20, 21).

A second hypothesis is that cerebral microembolization and hypoperfusion during cardiac surgery might significantly damage the brain and thus contribute to both, POD and worse long-term neuropsychiatric outcome (22). Of note, the impact of postoperative hyperperfusion has also been described as a detrimental condition with delayed cognitive recovery, respectively (23). In addition, research indicates that after surgical trauma, peripheral inflammatory factors cross the blood-brain barrier and consecutively trigger post-surgical neuroinflammation, which can cause synaptic dysfunction and neuronal death, ultimately leading to both POD and worse cerebral performance (24, 25).

A third approach is that psychiatric stressors are a strong trigger for prolonged neurological and neuropsychiatric symptoms. In 559 patients after elective noncardiac surgery, Drews et al. investigated the association between the occurrence of POD and posttraumatic stress disorder (PTSD) 3 months after surgery using the PTSS-14 questionnaire. They demonstrated that POD was a risk factor for PTSD at 3 months after surgery (26). A systematic review illustrated the relationship between POD and symptoms of depression (7). The authors showed that there are controversial data on the relationship of POD and symptoms of depression and concluded that they might be risk factors for each other. In our study, we found no significant differences in the occurrence of symptoms of depression as assessed by the BDI in relation to POD status. However, previous studies have shown higher rates of psychiatric comorbidities such as depression and anxiety in patients with cardiac disease (27, 28). This is consistent with the high prevalence of depression symptoms in our patient cohort (BDI >10, n = 36, 20.1%).

Regarding the high prevalence of depression in patients with cardiac diseases, one of the proposed mechanisms focuses on dysregulation of the hypothalamic-pituitary-adrenal axis. Emotional stress is associated with increased cortisol levels and plays an important role in the pathophysiology of depression (29, 30). Elevated cortisol levels can lead to decreased sensitivity of glucocorticoid receptors in the hypothalamic-pituitary-adrenal axis. Therefore, a compromised negative feedback loop may result in cortisol hypersecretion (31). These high cortisol levels are associated with cardiovascular risk factors, such as hypertension and dyslipidemia, and thus contribute to cardiovascular disease (32, 33).

Behavioral factors may also play a role. In a prospective study of 1,017 outpatients with stable coronary artery disease, the authors found an association between physical inactivity in patients with symptoms of depression and adverse cardiac events in their long-term follow-up after a mean of 4.8 years (34).

An association between POD and a decreased quality of life, as measured by the Medical Outcomes Study Short-Form General Health Survey, has also been suggested (35). In the present analysis we found lower scores on the Physical Component Summary of the SF-12, indicating a worse quality of life one year after surgery or intervention in patients with POD compared to patients without POD. This observation has been reported in several studies investigating quality of life after cardiac surgery. In patients with POD, Koster et al. observed a reduced quality of life at 6 months after cardiac surgery using the SF-36 (4), Loponen et al. found lower health-related quality of life in patients with POD up to 3 years after cardiac surgery (36). To address these issues, Ma et al. suggested that comprehensive rehabilitation combined with an educational program could help to reduce the rates of depression and anxiety and might improve quality of life after surgery in patients who underwent CABG (37).

Regarding hospital readmission, there was no significant difference between patients with and without POD in our study. Because the reasons for the readmission of patients in our study were diverse (Table 3), a close pathophysiologic link between POD and readmission is questionable.

Regarding cognitive performance, particularly after cardiac surgery, postoperative cognitive dysfunction (POCD) is a common complication with a prevalence ranging from 10 to 54% within the first few weeks (38). The term is used to describe an impairment of cognitive functions, including for example the executive function or attention, following surgery (39, 40). In the present study we found that the median MoCA score was higher in all patients one year after elective cardiac surgery compared to baseline. However, when comparing the delirious and non-delirious groups, the recovery was less pronounced in the group with POD. Several studies have already suggested a recovery of cognitive performance over time compared to preoperative baseline cognitive levels. Newman et al. found a decrease in the rate of cognitive dysfunction in patients undergoing CABG from 3 months to 6 months after surgery, and then again a higher rate of cognitive dysfunction at 5 years after surgery (41). Sauër et al. also demonstrated recovery of cognitive function one year after cardiac surgery (42).

One explanation for the improved overall MoCA score with 1yFU in our study is that it may be an effect of the testing environment. Preoperatively patients were tested in the hospital three to one day prior to their intervention or surgery, whereas during their follow-up visit, they were tested at home or in the hospital in a calm environment. Hence, anxiety and stress prior to the upcoming surgery or intervention might have affected the preoperative test results. Contradicting this view, data from another group of patients with recent brain injury showed that MoCA scores were not dependent on the testing environment, either in the ICU or in the office (43).

Also, Sacynzky et al. assessed cognitive function in patients undergoing CABG or valve replacement using MMSE (3). They reported that fewer patients with POD had returned to baseline cognitive function at 6 months compared with patients without POD. At 12 months, the proportion of patients returning to pre-surgical cognitive function did not differ between the two groups.

Another explanatory approach might be that improved cardiac output leading to better cerebral perfusion after cardiac surgery or percutaneous valve replacement might improve cognitive performance in the 1yFU (44). However, cardiac parameters were not assessed at the 1yFU in our cohrt.

In the cognitive domain, advanced age and depression must also be considered as independent risk factors for cognitive decline, as these factors may additionally influence the long-term cognitive outcome (45, 46).

Given the high prevalence of POD after cardiac surgery and its short and long-term adverse outcomes, preventive and therapeutic strategies should be considered. Establishing routine screening for delirium in postoperative care has been recommended in various guidelines, but adherence to this recommendation is still insufficient (47).

Follow-up care (e.g., mental health assessments) might be useful to identify patients experiencing neuropsychiatric symptoms, such as cognitive impairment, depression or fatigue after cardiac surgery or percutaneous valve replacement. As part of an enhanced after-care process, the provision of psychological support may be helpful.