Habitual tea consumption and postoperative delirium after total hip/knee arthroplasty in elderly patients: The PNDABLE study

Abstract Purpose To clarify the effects of habitual tea consumption on postoperative delirium (POD) in elderly patients undergoing total hip/knee arthroplasty. Patients and Methods A prospective cohort study was carried out at Qingdao Municipal Hospital Affiliated to Qingdao University between June 2020 and June 2021. A total of 332 patients aged 65–85 years undergoing total hip/knee arthroplasty under combined spinal and epidural anesthesia were enrolled from the Perioperative Neurocognitive Disorder and Biomarker Lifestyle (PNDABLE) study in the final analysis, consisting of 168 patients with habitual tea consumption and 164 patients with infrequent tea consumption. The primary endpoint was the effects of habitual tea consumption on POD and the incidence of POD, which was assessed by the Confusion Assessment Method (CAM) twice daily during the first 7 postoperative days, and POD severity was measured by the Memorial Delirium Assessment Scale (MDAS). The secondary endpoints were the concentrations of caffeine and tea polyphenols in plasma and cerebrospinal fluid (CSF), which were detected by the enzyme‐linked immunosorbent assay. Results POD occurred in 61 of 332 patients (18.37%), among whom 19 had habitual tea consumption (5.72%) and 42 had infrequent tea consumption (12.65%). Habitual tea consumption (odds ratio [OR] = 0.370, 95% confidence interval [CI]: 0.205–0.670, P = .001) was significantly associated with POD in the logistic analysis, and then after adjusting for age and American Society of Anesthesiologists (ASA) physical status (OR = 0.353, 95% CI: 0.190–0.655, P = .001). Furthermore, caffeine in T 0 plasma (OR = 0.834, 95% CI: 0.752–0.924, P = .001), T 1 plasma (OR = 0.818, 95% CI: 0.738–0.908, P < .001), and CSF (OR = 0.899, 95% CI: 0.820–0.984, P = .022) and tea polyphenols in T 0 plasma (OR = 0.541, 95% CI: 0.416–0.704, P < .001), T 1 plasma (OR = 0.477, 95% CI: 0.359–0.633, P < .001), and CSF (OR = 0.526, 95% CI: 0.397–0.696, P < .001) were associated with POD after adjusting for age and ASA physical status. Conclusion Habitual tea consumption may be associated with a lower incidence of POD in elderly patients.


INTRODUCTION
Postoperative delirium (POD) is acute deterioration of brain function characterized by changes in mental state and attention, as well as disturbance of consciousness (Duque et al., 2018), and the main clinical manifestations of POD are fluctuating levels of consciousness and disorientation after surgery (Inouye et al., 2014). The cognitive function for self-care ability of patients decreased (Hshieh et al., 2017).
Clinical subtyping of delirium according to motor activity is widely used. Hypoactive type is featured by a decreased amount of activity, unawareness, and a decreased amount of speech, while patients with hyperactive type show agitation, irritability, and hallucination. POD is accompanied by the deterioration of prognosis: increased risk of short-term complications, decreased long-term cognitive function, the prolongation of hospital stay, and increased mortality (Oh & Park, 2019;Rudolph et al., 2008;Schenning & Deiner, 2015). However, at the current time, the pathophysiology of POD remains uncertain, and POD is associated with diverse mechanisms, including cholinergic deficiency, inflammatory response, abnormal energy metabolism, and the stress response. Studies have shown that cholinergic deficiency accounts for the majority of POD (Hshieh et al., 2008;Young et al., 1997).
Tea is among the most consumed beverages in the world. It is rich in both caffeine and tea polyphenols. Tea drinking originated in China around 4000-5000 years ago, and spread to the rest of the world.
Caffeine is a central nervous system stimulant, which is naturally found in coffee, cocoa beans, and tea (Heckman et al., 2010). It is a unique compound that can be dissolved in water and nonpolar organic solvents, and distributed rapidly and efficiently throughout the whole body (including the brain) (Kolahdouzan & Hamadeh, 2017). It is generally believed that tea polyphenols, the main substance in tea, are good for health. The tea polyphenols can protect neurons via inhibition of dopamine (DA) oxidation, conjugation with dopamine quinones (DAQ), scavenge of reactive oxygen species (ROS), inhibition of monoamine oxidase B (MAOB), and modulations of nuclear factor erythroid 2-related factor 2 (Nrf2)-Keap1 and proliferator-activated receptor gamma coactivator 1-alpha (PGC-1α) antioxidative signaling pathways, which have strong neuroprotective functions . Some studies have found that caffeine and tea polyphenols exert neuroprotective effects, which can ameliorate cognitive impairment (Abreu et al., 2011;Çakır et al., 2017;Schimidt et al., 2017). Neuroprotection may play a role in the treatment of neuroinflammatory diseases (Venigalla et al., 2016). Namely, caffeine and tea polyphenols exert influence on the inflammatory response of POD via neuroprotective effects.
Then, can caffeine and tea polyphenols influence neurocognitive functions through the cholinergic nervous system?
The vagus nerve stimulates the release of acetylcholine (ACh) via activating the cholinergic pathway, which plays a crucial role in cognitive function. Surgical trauma stimulates the innate immune system to release proinflammatory cytokines and, in particular, interleukin-6 (IL-6) and tumor necrosis factor-α (TNF-α). Increased proinflammatory cytokines can be transmitted to the brain and result in neuroinflammation, leading to blood-brain barrier (BBB) disruption. Central ACh can inhibit the inflammatory response by combining with α7-nicotinic ACh receptors, which reduces BBB destruction and promotes brain protection (Naicker et al., 2016). The study has shown that α7-nicotinic ACh receptors antagonism is involved in the formation of POD in aged rats with tibia fractures . With the increase of age, especially in the elderly, the cholinergic nervous system degenerates. The level of central ACh involved in learning and memory is reduced, which affects cognitive function, and the degradation of cholinergic nervous system is involved in the development of POD Xu et al., 2019).
In other words, it is important to increase the content of Ach, because enhancement of cholinergic neurotransmission may potentially reduce adverse outcomes of POD. There is a dynamic balance between ACh content in plasma and cerebrospinal fluid (CSF) under the interaction of choline acetyltransferase and cholinesterase predominantly including acetylcholinesterase (AChE) and butyrylcholinesterase (BuChE) that can efficiently and rapidly degrade ACh (Vijayaraghavan et al., 2013). Cholinesterase inhibitors increase central ACh levels by prolonging the residence time of ACh in the synaptic cleft, thereby enhancing cholinergic neurotransmission (Okello et al., 2012). An experimental animal study proved that caffeine could inhibit cholinesterase in rat brain (Akomolafe et al., 2017). Another study demonstrated that tea polyphenols could inhibit AChE and BuChE, as well as enhance the cholinergic neurotransmission (Ali et al., 2016). Furthermore, tea polyphenols can play a part in inhibiting cholinesterase activity in vitro and in vivo (Adsersen et al., 2006;Jessen et al., 2014;Kulišić-Bilušić et al., 2008). Therefore, we speculate that caffeine and tea polyphenols, as cholinesterase inhibitors, can reduce cholinesterase activity and ameliorate cognitive impairment. Our study explored neurodegenerative disease from the perspective of diet and lifestyle. Tea consumption is an effective, safe, and appropriate non-pharmacological therapy for older people as well as a promising nonpharmacological intervention for patients with neurodegenerative diseases such as POD. Furthermore, we simultaneously analyzed plasma and CSF samples to evaluate their relationship.
Up to today, empirical evidence is still sparse and the value of tea drinking in the prevention of cognitive decline and POD is far from being firmly established. We aimed to prospectively examine the effects of habitual tea consumption on POD in Chinese elderly.  (Nasreddine et al., 2005), mild cognitive impairment (MCI), which was diagnosed with Nasreddine's criteria (McKhann et al., 2011), and Alzheimer's Disease (AD), which was diagnosed with Guy's criteria (Okello et al., 2004), according to the different criteria.

Participants
The study was conducted at Qingdao Municipal Hospital affiliated to On the basis of inclusion and exclusion criteria, a total of 360 participants were screened for our study, of whom 28 were excluded for various reasons and 332 participants were analyzed in our study ( Figure 1).

Study protocol
This is a prospective cohort study. The MMSE was administered the day before surgery. According to frequency of tea consumption, participants were categorized into habitual tea consumption group (T group) and infrequent tea consumption group (NT group). At the same time,

Habitual tea consumption
Data on tea consumption were obtained using a validated self-reported questionnaire including items like "whether you drink tea or not" and the frequency of tea consumption. Data on tea consumption were collected based on local tea drinking customs and terms such as "green tea," "black tea," and "oolong tea." We divided the frequencies of tea consumption into the following categories: (1) never or <1 time/week; (2) 1-3 times/week; (3) >3 times/week; and (4) almost every day (Ma et al., 2020;Yasutake et al., 2016). A participant was classified as an infrequent tea consumer (NT group) if he/she fell into the Category (1) or (2); a participant was classified as a frequent tea consumer (T group) if he/she fell into the Category (3) or (4).

Anesthesia and surgery
All the participants underwent total hip/knee arthroplasty under combined spinal and epidural anesthesia by the same surgery team. Fivelead electrocardiography, pulse oxygen saturation, and noninvasive blood pressure were monitored continuously during anesthesia and were recorded at fixed intervals of 5 min. The patients inhaled oxygen through the mask at 5 L/min, and the fluctuation of noninvasive blood pressure was kept within ±20% of the baseline value. However, glucocorticoid drugs, nonsteroidal analgesics, and midazolam were avoided during surgeries.

F I G U R E 1
Patients screening and enrollment in the study. Depicted is the study flow diagram, including data on patients who were screened, enrolled, and excluded. CV, coefficient of variation; T 0 , before anesthesia; T 1 , 24 h after surgery.
Postoperatively, the visual analogue scale score of 0-10 was used to assess pain in the patients at the same time. PCIA was used postoperatively for 48 h by all patients. The PCIA opioid consisted of 2.5 μg/kg sufentanil and 5 mg tropisetron (total volume of 100 ml, including 0.9% normal saline, bolus 2 ml, basal rate 2 ml/h, and lockout time 15 min). If patients need it, they were given nonopioid drugs for analgesia, which was recorded.

Sample collection and laboratory testing
Elbow vein blood samples (5 ml

Neuropsychological testing
The MMSE was administered by a neurologist 1 day before the operation, which is widely used as a screening tool for dementia and as a measure of global cognitive function. POD was defined by the Confusion Assessment Method (CAM) (Leung et al., 2008), and POD severity was measured using the Memorial Delirium Assessment Scale (MDAS) (Fabbri et al., 2001) at 10:00 a.m. and 2:00 p.m. twice a day on 1-7 days (or before discharge) by an anesthesiologist postoperatively. The above assessments performed by a neurologist and an anesthesiologist were not involved in intraoperative management of the patients. The  (1) and (2), accompanied by (3)

Participants' demographic characteristics
We screened 360 participants, of whom 28 participants were excluded from the study. And the reasons for dropouts are shown in Figure 1. was significantly associated with habitual tea consumption, and tea polyphenols in T 0 plasma (OR = 1.496, 95% CI: 1.231-1.753, P < .001), T 1 plasma (OR = 2.523, 95% CI: 1.988-3.201, P < .001), and CSF (OR = 1.577, 95% CI: 1.303-1.909, P < .001) were significantly associated with habitual tea consumption (Table 2). Of note, the concentrations of caffeine and tea polyphenols at T 0 and T 1 in P group were significantly lower than those in NP group (P < .05) (Figure 2).

3.3.4
Associations between preoperative plasma biomarkers of tea and other biomarkers  Abbreviations: CI, confidence interval; CSF, cerebrospinal fluid; OR, odds ratio; T 0 , before anesthesia; T 1 , 24 h after surgery.

F I G U R E 2
The concentrations of tea biomarkers in plasma and cerebrospinal fluid between P group and NP group. Comparison of the concentrations of caffeine (a) and tea polyphenols (b) at T 0 and T 1 in plasma and in CSF between P group and NP group. Standard deviations were calculated based on Mann-Whitney U test (**P < .01). CSF, cerebrospinal fluid; NP, non-postoperative delirium; P, postoperative delirium; T 0 , before anesthesia; T 1 , 24 h after surgery.

TA B L E 3 The associations between biomarkers in plasma and cerebrospinal fluid (CSF) (independent variable X) and POD (dependent variable Y)
Biomarkers ( Note: The demographic and perioperative characteristics were included in univariate logistic regression model. The multivariable logistic regression model was adjusted for age and ASA physical status. Abbreviations: CI, confidence interval; CSF, cerebrospinal fluid; OR, odds ratio; POD, postoperative delirium; T 0 , before anesthesia; T 1 , 24 h after surgery. negatively associated with caffeine concentration at T 0 (Figure 3a (Figure 3e).

F I G U R E 3
Correlations between preoperative plasma biomarkers of tea and other biomarkers. The scatter plots depict the associations of the concentrations of AChE (a), BuChE (a), IL-6 (e), and TNF-α (e) and the activities of AChE (b) and BuChE (b) at T 0 with the concentrations of caffeine at T 0 , and the associations of the concentrations of AChE (c), BuChE (c), IL-6 (f), and TNF-α (f) and the activities of AChE (d) and BuChE (d) at T 0 with the concentrations of tea polyphenols at T 0 . AChE, acetylcholinesterase; BuChE, butyrylcholinesterase; ChEs, cholinesterases; IL-6, interleukin-6; T 0 , before anesthesia; T 1 , 24 h after surgery; TNF-α, tumor necrosis factor-α.

DISCUSSION
In this prospective cohort study of Chinese elderly, we evaluated the association between habitual tea consumption and POD in 332 elderly adults who underwent hip/knee arthroplasty under combined spinal and epidural anesthesia. An increase of studies focused on the prediction and treatment of POD, while there was still a lack of evidence focused on correlations between POD and biomarkers of tea. Our study is to explore the impact of habitual tea consumption on POD from the perspective of dietary intake.
Aging and neurodegenerative diseases are well-known risk factors for delirium (Inouye, 1996), and previous cognitive impairment is a major risk factor for POD (Davis et al., 2015). Preoperatively, we selected subjects aged 65-85 years, excluded subjects with cognitive impairment, and compared the MMSE scores of all subjects at baseline.
Our data suggested that there was no significant difference in MMSE score at baseline between patients with and without POD. In this cohort study, we observed that the incidence of POD following total joint arthroplasty was 18.37% (n = 61 of the 332 patients, 5.72% in habitual tea consumers and 12.65% in infrequent tea consumers), which was largely consistent with two previous studies reporting a POD incidence of 16.5% and 17% (Chen et al., 2017;Scott et al., 2015).
As expected, for those who had infrequent tea consumption, we found that they appeared to be more likely to develop POD (12.65%, n = 41 of the 332 patients), which indicated that habitual tea consumption might have an impact on the occurrence of POD. In regression analyses, we found that habitual tea consumption was significantly associated with POD after adjustment for age and ASA physical status. The adjusted OR value was 0.353 (P = .001), and the 95% CI was 0.190-0.655, which indicated that habitual tea consumption was negatively correlated with POD. That is to say, habitual tea consumption after adjustment is still associated with POD and it is a protective factor for reducing the occurrence of POD. Likewise, the same went for biomarkers of caffeine and tea polyphenols in plasma at T 0 and T 1 as well as in CSF.
CSF is in direct contact with the extracellular space of the brain, and it can reflect the biochemical changes that occur in the brain. Compared with biomarkers in plasma, CSF biomarkers have higher clinical significance and diagnostic value (Olsson et al., 2016). Therefore, CSF is the optimal source of POD biomarkers (Xie et al., 2014). Plasma is less invasive and more readily available, and spinal epidural block is a convenient method to collect CSF samples from subjects in our study.
Because CSF could not be collected after operation, we compared and analyzed the preoperative biomarkers in plasma and CSF to verify their correlations so as to estimate their postoperative consistency.
Spearman's rank correlation indicated that the concentrations of preoperative caffeine, tea polyphenols, AChE, and BuChE in plasma were significantly correlated with those in CSF. Additionally, the activities of preoperative AChE and BuChE in plasma were correlated with those in CSF. Consequently, there is a strong consistency of preoperative biomarkers in plasma and CSF, from which it can be inferred that biomarkers in CSF after operation may be also highly consistent with those in plasma.
For Chinese, especially the elderly, tea drinking is a particular dietary habit. Studies have found that drinking tea can improve mood, concentration, and mental performance (Gardner et al., 2007;Gilbert, 2019). Previous research indicated that tea consumption was significantly associated with cognitive impairment and mortality (Qiu et al., 2012;Ruan et al., 2013). All types of tea contain a large amount of caffeine and tea polyphenols. In our study, after excluding confounding variables such as age, gender, height, body weight, and BMI, we revealed that caffeine and tea polyphenols in plasma at T 0 and T 1 and in CSF were significantly associated with habitual tea consumption, which was consistent with our common knowledge about patients who had frequent tea consumption. Previous studies have confirmed that caffeine and tea polyphenols play important roles in prevention and intervention for cognitive decline (Ide et al., 2014;Park et al., 2011). Then, how does habitual tea consumption exert impact on neurodegenerative diseases such as POD?
To our knowledge, ACh is a classic neurotransmitter that is widely distributed in the nervous system and considered to play a key role in the regulation of learning and memory, especially in the regulation of cognitive function (Naicker et al., 2016). As mentioned above, cholinesterase inhibitors increase central ACh levels by prolonging the retention of ACh in the synaptic cleft, thereby enhancing cholinergic neurotransmission (Okello et al., 2012). It has been confirmed that the main treatment of cholinergic impairment is the use of cholinesterase inhibitors (Akomolafe et al., 2017). In our study, we mainly focused on the effects of habitual tea drinking on POD by analyzing the associations of cholinesterase with biomarkers of caffeine and tea polyphenols in the plasma at T 0 among 332 patients.
The concentrations of AChE and BuChE as well as the activities of AChE and BuChE at T 0 were negatively associated with the concentrations of caffeine and tea polyphenols at T 0 , whereas these associations were relatively weak, which may be explained by the inhibition of cholinesterase caused by caffeine and tea polyphenols.
Namely, higher concentrations of caffeine and tea polyphenols were associated with lower concentrations and activities of cholinesterase.
Peripheral inflammation can induce further cytokine release in cerebral tissue through pro-inflammatory cytokines, leading to overactivated microglia and disrupted BBB. Furthermore, overactivated microglia and disrupted BBB create a neurotoxic response, cause neuronal injury, and affect neuronal function (Dantzer et al., 2008;Hughes et al., 2012;MacLullich et al., 2011). Neuroinflammatory factors, such as IL-6 and TNF-α, induce inflammation of the nervous system, which results in dysfunction of neurons and synapses and leads to neurobehavioral and cognitive symptoms of POD (Maldonado, 2013). Cholinergic anti-inflammatory pathway that interrupts a vicious cycle of neuroinflammation plays an important part in inflammatory response on POD. In addition to improving disrupted BBB caused by TNF-α and IL-6, it is theoretically plausible that cholinergic anti-inflammatory pathway prevents neuroinflammatory factors from affecting cognitive function due to entry via disrupted BBB (Evered et al., 2016). In this study, the concentrations of IL-6 and TNF-α at T 0 were also negatively associated with the concentrations of caffeine and tea polyphenols at T 0 , which may be explained by the enhancement of the cholinergic anti-inflammatory pathway induced by caffeine and tea polyphenols.
However, there were weak correlations of the inflammatory factors (IL-6 and TNF-α) with caffeine and tea polyphenols. It is possible that inflammatory response has some special underlying changes in the vulnerable brain, which exerts an influence on correlation analysis in our study. A previous study found, when evaluating plasma and CSF samples for up to 30 h postoperatively, a consistent upregulation of CSF IL-6 and undetectable TNF-α was found. And IL-6 level might correlate with sleep disorders (Buvanendran et al., 2006). Another study has also pointed out that the extent of inflammatory reaction is related to the degree of postoperative cognitive changes (MacLullich et al., 2011).
Probably, in POD patients, there are increased levels of neuroinflammatory factors and suppression of cholinergic nervous system. In other words, inhibiting neuroinflammation or activating the cholinergic nervous system may affect the development of POD, which may be due to changes in AChE, BuChE, and inflammatory factors as indicated by our research. Changes in contents and activities of cholinesterase, caused by habitual tea drinking, probably produce an effect on the cholinergic nervous system.
As for the differences in the associations of POD with the biomarkers of caffeine and tea polyphenols, we speculated that the differences were associated with the different intricate effects of caffeine and tea polyphenols on the central nervous system. ACh levels are regulated not only by the changed cholinergic nerve system but also by adenosine receptors that are combined with caffeine (Acquas et al., 2002;Van Dort et al., 2009).
Inevitably, several potential limitations of this study should be taken into account. First, the current pilot findings do not provide strong evidence that habitual tea consumption can affect the occurrence of POD due to the small sample size. Our results will need to be validated by a larger study. Second, the data on habitual tea consumption in our study were self-reported. Although subjective evaluation of the frequency of tea consumption is likely to be important and valid, this type of measure is not standardized, which may limit comparability with other studies. The measurement for this self-reported tea consumption may be a concern. However, we have little concern about the consistency of such a measurement in the findings because our results are based on the cohort study. Third, although all types of tea are rich in caffeine and tea polyphenols, the associations between various measures of tea consumption and POD, like the amount of tea leaves added, types of tea, and duration of tea consumption, have yet to be examined.
Further research is needed to elucidate potential contributing factors to the pathogenesis of POD. Additional considerations include CSF and plasma sampling over stationary period to consider other than the effect of circadian variations.
In spite of these shortcomings, we have extended the existing research by providing a new estimate of the association between habitual tea consumption and POD among Chinese elderly. The results of this study suggest that habitual tea consumption is probably associated with lower incidence of POD compared with patients who had no tea drinking habit. Because tea is easily available and cheap and has no side effects at normal consumption level, drinking tea is potentially a good preventive measure for neurodegenerative diseases such as POD and may help to reduce healthcare cost during hospitalization. This is especially relevant in the context of a growing and fast-aging world population.

CONCLUSION
Habitual tea consumption may be associated with a lower incidence of POD following knee/hip arthroplasty surgery in elderly patients, which may correlate with inhibition of cholinesterase by caffeine and tea polyphenols, as well as enhancement of cholinergic neurotransmission. To date, it remains unknown whether there are other potential contributing factors to the pathogenesis of POD, pending further investigations.