3.1. Postoperative Cognitive Decline (Deficit)
With the increasing number of non-cardiac surgeries performed worldwide, with over 300 million procedures annually, also in the high-risk and aging population, POCD could pose a significant threat to global health issues [26
]. Surprisingly, there is a limited number of studies focused on assessing the role of intraoperative BP drop and the development of POCD, despite its clear pathophysiological association [27
Between 1993 and 1995, a randomized controlled trial including 235 patients older than 50 years (mean age 72 years) undergoing total hip replacement with neuraxial anesthesia was performed [28
]. Subjects were randomly assigned into two arms: ‘controlled IOH’ group, with a mean arterial pressure (MAP) of 45–55 mmHg or a group with 55–75 mmHg. The researchers found no significant differences between those two groups in terms of either early (i.e., 1 week after surgery) or late (i.e., 4 months after surgery) POCD. POCD was defined as a within-subject change in score for each of 10 neuropsychological tests over a 4-month period. It is vital to note that it is difficult to draw reliable conclusions based on the above-mentioned data. Although all subjects were hypotensive during surgery, no other adverse events were assessed, including MINS and AKI. Also, no control group (i.e., without hypotension) was applied to verify if MAP of 45–55 mmHg was safe to maintain adequate organ perfusion.
Moller et al. published the results of a multicenter study, in which 1218 surgical patients aged ≥ 60 years undergoing major abdominal, thoracic and orthopedic procedures under general anesthesia between 1994 and 1996, were tested for POCD after 1 week and 3 months after procedure using a battery of neuropsychological tests [12
]. In their study, IOH was defined as an MAP drop of at least 60% from baseline value for at least 30 min. No statistical significance was observed in POCD occurrence between ‘hypotensive’ and ‘non-hypotensive’ subjects, at either 1 week (i.e., ‘early POCD’) or 3 months (‘late POCD’) postoperatively. Lack of association between hypotension and cognitive decline might have resulted from a unique definition of IOH, which was very strict and incomparable with other research cited in this review, and different from current recommendations of the POQI [4
Brazilian researchers, in a study of 55 patients undergoing orthopedic, urologic, general or vascular surgeries in general (n
= 28) or regional (n
= 27) anesthesia, assessed the incidence and risk factors of POCD defined as Mini-Mental State Examination (MMSE) scores below 24/30 or by a decrease of 4 or more points during follow-up evaluations [29
]. The incidence of IOH, defined as systolic BP decrease of more than 30% as compared to baseline values, lasting more than 5 min or needing vasopressor or inotropic support, did not differ between subjects who developed or did not develop POCD (p
= 0.7). The study was underpowered (n
= 55) and the results were not adjusted for a plethora of possible confounding factors.
In 2009, Yocum et al. published a post-hoc cohort analysis in which 24 hypertensive and 21 normotensive patients aged ≥40 years were screened for POCD after lumbar laminectomy or microdiscectomy [30
]. A battery of 5 neuropsychometric tests was used for assessing cognitive performance preoperatively (baseline) at 1 day, and 1 month postoperatively. The blood pressure was recorded as a fraction of preoperative MAP (i.e., % of baseline MAP). The lowest fractional MAPs, together with several other factors (age, hypertension, years of education, duration of surgery, etc.) were put into multivariate analysis in which the sole presence of either hypertension or lowest fractional MAP was not associated with POCD. However, hypertension crossed with minimum MAP was significantly associated with cognitive decline, both 1 day (p
= 0.001) and 1 month (p
= 0.02) after surgery. With these findings, the authors emphasized the role of impaired cerebrovascular autoregulation in developing POCD, as hypertensive patients are often less adaptive to lower blood pressures. These results, however, should be interpreted with caution, since the study population was small (n
= 45) and the history of antihypertensive drug treatment was not taken into consideration.
During 2009–2010, Panda et al. in a sample of 30 hypertensive and 30 non-hypertensive patients, aged 40–59 years, undergoing elective non-cardiac procedures under general anesthesia lasting ≥2 h observed a postoperative significant decline in cognitive function in both groups [31
]. However, the POCD pattern differed between groups, based on Z-scores in a battery of 7 neuropsychological tests. Importantly, there was a noticeable difference in minimal intraoperative MAP, i.e., 93.8 mmHg vs. 68.6 mmHg (p
< 0.001). The reader should bear in mind two major limitations of the study, including lack of adjustment for hypotensive events during surgery and a short observation period, lasting only 7 days.
Heyer et al. assessed the influence of intraoperative BP on early cognitive deficit in 183 patients undergoing cross-clamp open carotid endarterectomy under general anesthesia between 1995 and 2012 [32
]. They found that patients whose intraoperative MAP was maintained ≥ 20% of baseline values, exhibited an 82% lower risk of POCD (p
< 0.001). In their study, POCD was assessed by a local neuropsychologist using self-developed tests. Nevertheless, as the procedure itself is usually burdened with sudden changes in BP supply to the brain, causing possible disturbances in cerebral blood flow, higher MAP should be ensured to guarantee adequate perfusion.
Another interesting study was set up to verify the relationship between perioperative hemodynamic management and the risk of new-onset dementia [33
]. The study was performed in 696 individuals averagely aged 84 years undergoing hip fracture surgery in either regional (8%) or general (92%) anesthesia. Within 6 or 12 months after surgery, dementia was diagnosed in 12% of patients using ICD-10 research criteria for dementia. Surprisingly, patients with MAP ≥ 130 mmHg at admission had over 5-times higher risk of dementia compared to subjects with MAP < 110 mmHg (p
= 0.035). In addition, those whose MAP at admission was >100 mmHg had over 10-times higher risk of dementia compared to individuals with MAP < 80 mmHg (p
= 0.009). This phenomenon can be explained by the fact that previous long-lasting arterial hypertension, which was diagnosed in 37% subjects preoperatively, might alter vascular autoregulation and made the cardiovascular system more susceptible to the effects of intraoperative BP drops. The results should be interpreted with caution because the assessment was performed in two different cohorts of patients hospitalized in 2005/2006 (n
= 364) and 2009/2012 (n
Wang H et al. published the results of their research on the influence of controlled hypotension on postoperative cognitive function in 20 elderly spinal surgery patients, performed from 2013 to 2014 [34
]. Controlled hypotension was defined as a MAP not less than 55 mmHg but no upper BP limit was set up. Importantly, intraoperative MAP was over 25 mmHg lower than the value recorded directly after induction of anesthesia (61.4 mmHg vs. 87.2 mmHg). The authors revealed that the crude score in MMSE was significantly lower on a postoperative day 1 and POCD (i.e., MMSE score at least 2 points lower than the preoperative one) occurred in 20% of subjects but that effect disappeared on day 7. The sample size was small and no effect of demographic of clinical data was examined.
Ni C et al. enrolled 78 orthopedic patients (mean age 70 years) undergoing total knee arthroplasty between 2014 and 2015, and screened them for POCD on days 1 and 6 after surgery [35
]. The patients received regional anesthesia and no sedative agents were applied. POCD was diagnosed based on the battery of ISPOCD neuropsychometric tests. In subjects who developed POCD, a statistically significant decrease in MAP was observed after induction of anesthesia, with a mean drop of 22.24 mmHg compared to 13.22 mmHg in patients without POCD (p
A group of Turkish investigators performed a prospective, randomized clinical trial to compare cognitive function in 46 patients undergoing moderate hypotensive (i.e., MAP 60–70 mmHg) or normotensive anesthesia for endoscopic sinus surgery [36
]. The results were published in 2005. There were no significant differences between the two groups in MMSE and the Visual Aural Digit Span Test (VADST) 2 h and 24 h postoperatively. Unfortunately, the study group was too small, BP drop was rather too low and the observation period too short to draw reasonable conclusions.
In 2015, Deiner et al. published their results on a study recruiting patients older than 65 years scheduled for major general, spine, urologic or thoracic surgery under general anesthesia [37
]. POCD was defined based on a complex neuropsychological assessment with a battery of tests, MMSE and Confusion Assessment Method. Subjects with POCD were exposed to IOH on a similar extent as patients who did not develop cognitive decline (5 incidents in both groups, p
= 0.78). Hypotension was defined as a number of incidents of MAP < 55 mmHg lasting for 5 min. Two limitations are crucial: significant impact of a type of anesthesia (total intravenous anesthesia, TIVA vs. volatile) and a small study group in a 3-month follow-up (n
In the next study, Chuan et al. explored a role of cerebrovascular autoregulation impairment in developing POCD in 140 subjects (mean age 68 years) undergoing abdominal, thoracic and vascular procedures using general anesthesia between 2014 and 2016 [38
]. They found that defective cerebrovascular autoregulation (using tissue oxygenation index of dynamic autoregulation, TOx) was linked to the failure of cognitive recovery, assessed 3 days after surgery using the cognitive domain of the Postoperative Quality of Recovery Scale. The TOx in patients with cognitive recovery compared with those with impaired cognition was statistically significantly lower (0.06 vs. 0.18, p
= 0.02). Those findings shed light on the role of the management of intraoperative BP to ensure adequate cerebral perfusion and cognitive performance in the postoperative period.
Finally, between 2014 and 2016, Langer et al. performed a randomized controlled trial in which 101 patients aged 80 ± 4 years were allocated into two groups [39
]. The first group—the ‘target group’—had a personalized strategy of keeping MAP ≥ 90% of preoperative values, whereas the second group (the control one) had a liberal approach of intraoperative BP management. Despite having a higher intraoperative MAP (93 mmHg vs. 85 mmHg), no significant improvement was found in the prevalence of POCD after 3 months (11% vs. 7%, p
= 0.5) in the study group. POCD was diagnosed by an experienced neuropsychologist using a battery of 7 tests. Some drawbacks should be taken into account in data interpretation, including relatively small sample size and a 24% drop-out rate for the neuropsychological testing after 3 months. Additionally, no standardized protocol for hemodynamic management was implemented.
Available studies on the association between intraoperative hypotension (IOH) or intraoperative blood pressure drop and postoperative cognitive decline (deficit) (POCD) are summarized in Table 5
3.2. Postoperative Delirium
One of the first attempts to assess the impact of intraoperative hemodynamic parameters on the development of POD was made by Marcantonio et al. [40
] who analyzed this issue in a group of 1341 patients over 50 years of age undergoing major elective non-cardiac surgeries. POD was diagnosed in only 9% of subjects, using a structured interview. There was no association between IOH and the development of delirium (POD vs. non-POD: 8% vs. 9%, p
= 0.4). The duration of IOH also did not affect the number of POD episodes.
Over 10 years later, Tognoni et al. [41
] identified IOH as one of the predictors of POD in 90 patients aged 74 ± 0.4 years undergoing urological procedures. In their study, hypotension was defined as a drop in systolic BP <90 mmHg requiring implementation of fluid therapy or infusion of catecholamines. POD, diagnosed by Confusion Assessment Method (CAM) algorithm, developed in 8.8% of patients, usually on postoperative day 1 and lasted for 3 ± 0.8 days. IOH occurred more frequently in the POD group (66% vs. 24%; p
= 0.02); however, BP values were recorded in 15-min time intervals, what is a serious limitation in reliable concluding.
In a study published in 2011, Patti et al. [42
] assessed 100 patients aged over 65 years undergoing colorectal oncological surgeries under general anesthesia. IOH was considered as a decrease in MAP <60 mmHg or the need for infusion of catecholamines. The screening for POD was performed twice a day with the CAM algorithm, and allowed to diagnose delirium in 18% of patients. It was found that IOH was a strong independent factor of POD (OR = 9.74; 95%CI 2.5–37.9, p
In 2015, Hirsch et al. [43
] diagnosed POD with CAM in more than a half (i.e., 54%) of 540 patients aged 73 ± 6 years, who underwent major non-cardiac procedures. The authors revealed that IOH, defined as an absolute MAP drop <50 mmHg or relative decrease in MAP or systolic BP by more than 20, 30 or 40% from the baseline value, had had a significant impact on POD development (p
> 0.05 for all). In addition, higher intraoperative variability of BP, calculated as a variance from all intraoperative BP measurements, was associated with a slightly higher risk of delirium (OR = 1.04; 95%CI 1.01–1.07, p
Wang J. et al. [44
] investigated the impact of 12 perioperative variables on POD, that was assessed in 200 geriatric patients aged over 65 years who underwent elective orthopedic procedures. Seven factors have been identified as risk factors for POD, namely: age, type of anesthesia (general vs. regional), duration of surgery, intraoperative hypercapnia, IOH, altered preoperative mental state and postoperative sleep disorders. However, the authors did not provide the IOH diagnostic criteria, which is a serious limitation for drawing reliable conclusions.
In addition, in the elderly orthopedic population (i.e., n
= 103, average age 82 ± 7 years), Wang NY. et al. [45
] evaluated intraoperative BP as a function of mean intraoperative MAP (msMAP) and the percentage change in msMAP in relation to the baseline BP value. For each patient, the average msMAP was calculated as the area under the MAP curve formed by the 5-min recordings during surgery, divided by the length of surgery. Baseline MAP value was calculated as the mean of BP measurements from the moment of patient’s admission to the hospital until induction of anesthesia. The assessment for POD was performed on day 2 after surgery with CAM, what resulted in diagnosing delirium in 23% of patients. It was shown that in the group of subjects with msMAP <80 mmHg the increase of msMAP by 10 mmHg decreased the risk of POD (OR = 0.21, 95%CI 0.05–0.86, p
= 0.03), whereas in the group of patients with msMAP ≥80 mmHg the increase of msMAP by 10 mmHg increased the risk of POD (OR = 2.34, 95%CI 1.11–4.94, p
Contrarily, in a study published in 2016, recruiting 480 patients aged 81 ± 6 years undergoing major non-cardiac surgery, it was found that IOH, defined as a decrease in baseline MAP >30%, had no impact on the frequency of POD episodes (5.8 vs. 6.1 episodes; p
= 0.9) [46
]. POD diagnosed with CAM occurred in 28.5% of individuals. The serious drawback of the study is the fact that baseline MAP was determined based on the measurements taken only 10-min before anesthesia induction.
Interestingly, secondary analysis of data from a study of van Grootven and colleagues [47
] revealed that among 86 patients aged 80 ± 7 years undergoing orthopedic surgery, the lowest intraoperative diastolic BP was the protective factor for POD (OR = 0.92; 95%CI 0.85–0.99, p
= 0.03) and the protective impact of the highest systolic BP was of a borderline significance (OR = 0.97; 95%CI 0.94–1.0; p
= 0.06). The assessment of POD with CAM was performed in a complex manner on days 1, 3, 5 and 8 post-surgery. POD was diagnosed in 12.2%, 15.3%, 6.1% and 4.5% of subjects, respectively. This phenomenon is difficult to explain and was not discussed properly by the authors.
Soh et al. [48
] investigated the issue of POD in 109 subjects aged over 60 years. IOH was defined as a decrease in MAP <80% of preoperative value or systolic BP < 90 mmHg, or absolute MAP <60 mmHg. IOH episodes were rare, however, were more frequent in those who developed POD (4 vs. 2 episodes; p
= 0.01). The reader should bear in mind two major limitations of the study, including monitoring of BP in 15-min time intervals and the assessment of POD lasting only 48 h after surgery.
In the next study performed in the spinal surgery setting, Jiang et al. [49
] retrospectively evaluated clinical charts of 451 patients aged 65 ± 18 years who underwent spinal surgery. POD, diagnosed with CAM was found in 9.3% of patients. Systolic BP <80 mmHg was considered as IOH. IOH was described by the authors as one of the main factors influencing the development of POD but the interpretation of statistics is rather not so convincing (i.e., OR = 7.5; 95%CI 0.18–17.93, p
In the study published in 2017 by Neerland et al. [33
], no association between the lowest intraoperative MAP and POD was found among 696 elderly orthopedic patients, i.e., aged 85 years (IQR 78–89) with femoral neck fracture (POD vs. non-POD: 63 mmHg vs. 67 mmHg, p
= 0.24). Screening for POD using CAM was performed once a day, except for the weekends, which should be taken into account in data interpretation.
In 2017, Guo et al. [50
], basing on data from 385 patients with a mean age of 48 ± 13 years undergoing escharotomy, designed a predictive model to assess the probability of POD. One of the variables included in the model was IOH, defined as MAP drop <55 mmHg (OR = 27.07; 95%CI 7.13–102.83, p
< 0.05). Patients with POD, representing 14.6% of the studied population had significantly more incidents of IOH (78% vs. 9%; p
< 0.001). One ought to notice that the evaluation of POD using CAM was performed only once a day.
More recently, Langer et al. [39
] performed a study covering 101 patients aged 80 ± 4 years scheduled to various non-cardiac surgeries. The authors investigated two strategies of the intraoperative BP control. In the first group, personalized management consisting of maintaining MAP ≥90% of preoperative value was applied, while in the second group, no target values were defined, leaving MAP values for the anesthesiologist’s decision. Interestingly, both strategies had similar and non-significant effect on the risk of POD (p
= 0.2), which was observed in 9.9% of all patients using CAM-ICU for evaluation.
Additionally in 2019, in a group of 323 patients at a mean age of 60 years undergoing laryngectomy, POD by CAM occurred in 8.7% of them [51
]. POD was more frequent in subjects who had IOH (p
= 0.03). IOH was defined as BP lower than 70% of the baseline value lasting for more than 30 min.
For the study of Maheshwari et al. [52
], aiming to assess the effect of IOH on POD, 1083 patients aged 62 ± 14 years undergoing organ transplantation (20%), gastrointestinal surgery (28%), orthopedic surgery (22%) or other non-cardiac surgery (30%) were recruited. Hypotension was considered as a MAP drop <65 mmHg. The assessment for delirium was performed twice a day with CAM-ICU. The delirium was found in 35% of individuals. It was shown that IOH was a significant risk factor for POD (HR = 1.11; 95%CI 1.03–1.20, p
= 0.009). Moreover, the study demonstrated that POD was more frequent in patients with postoperative hypotension: the risk increased significantly with the decrease of the lowest postoperative MAP (HR = 1.12; 95%CI 1.04–1.20, p
In a small study covering 39 patients (median age 52 years, IQR 47–57) undergoing elective endovascular intracranial aneurysm treatment [53
], the maximal intraoperative decrease of systolic BP was significantly higher in those who developed POD (33 mmHg vs. 25 mmHg, p
= 0.002), but this effect was not confirmed in multivariate analysis (p
= 0.5). POD was diagnosed in 35% of patients by using the CAM algorithm.
Finally, Radinovic et al. [54
] in 277 orthopedic patients aged over 60 years aimed to assess the effects of a single episode of IOH, BP variations and heart rate changes on POD development. POD was evaluated using CAM with identification of its type (i.e., hypo- or hyperactive). POD was diagnosed in 53% of patients, 75% of which had the hypoactive type. Among patients who developed POD, lower BP values were observed compared to patients without POD, respectively: 128 mmHg vs. 122 mmHg (p
= 0.001) for systolic, 78 mmHg vs. 76 mmHg (p
= 0.01) for diastolic and 95 mmHg vs. 91 mmHg (p
= 0.001) for mean BP. Although the authors failed to demonstrate a cut-off point for the MAP value below which the risk of POD increased, they showed that the likelihood of POD increased in patients with MAP interval of 75–80 mmHg. Interestingly, the lowest MAP had a slightly protective effect on the occurrence of any type of POD (OR = 0.94, 95%CI 0.89–0.99, p
= 0.01), which is difficult to interpret. It must be underlined that BP values in the POD group did not meet the IOH criteria that are usually adopted in all previously mentioned studies. However, we may hypothesize that for selected patients (e.g., those with preoperative hypertension) even a decrease in MAP to 80 mmHg may cause clinically significant organ hypoperfusion. Unfortunately, this study lacked the evaluation of hemodynamic parameters in the postoperative period, which might help to explain those interesting but vague associations.
The above-mentioned studies on the association between intraoperative hypotension (IOH) or intraoperative blood pressure drop and postoperative delirium (POD) are summarized in Table 6