Academiejaar 2015 – 2016

Postoperative cognitive dysfunction after shoulder surgery in beach chair position: a

pilot study

Jan FIERENS

Promotor 1: Prof. dr. Stefan De Hert

Promotor 2: Dr. Tom Jacobs

Masterproef voorgedragen in de master in de specialistische geneeskunde Anesthesie - Reanimatie

Academiejaar 2015 – 2016

Postoperative cognitive dysfunction after shoulder surgery in beach chair position: a

pilot study

Jan FIERENS

Promotor 1: Prof. dr. Stefan De Hert

Promotor 2: Dr. Tom Jacobs

Masterproef voorgedragen in de master in de specialistische geneeskunde Anesthesie - Reanimatie

Table of contents

Abstract ...................................................................................................................... 1

Introduction ................................................................................................................. 3

Beach chair position ................................................................................................ 3

Near-infrared spectroscopy: technical aspects and clinical use ............................. 5

Postoperative decline in cognitive function ............................................................. 7

Goal of the study ...................................................................................................... 11

Materials and methods ............................................................................................. 12

Results ...................................................................................................................... 15

Discussion ................................................................................................................ 18

References ............................................................................................................... 21

Nederlandse samenvatting ....................................................................................... 27

1

Abstract

Background/Goal

Major shoulder surgery in beach chair position has been associated with severe

adverse neurologic events even in healthy middle-aged patients. This has been

attributed to the beach chair position itself as it reduces cerebral oxygenation. The aim

of the present study is to identify whether intra-operative episodes of cerebral

desaturation are associated with postoperative cognitive decline.

Methods and Materials

After obtaining ethics committee approval and informed consent, 11 adult

patients (4 males and 7 females with mean age=65;0 y, SD=12.4y) without prior

neurological and psychiatric disorder were enrolled prospectively. Neurocognitive

function was assessed preoperatively, at the 3rd postoperative day and at 3 months

postoperatively. Memory and attention performance were evaluated by means of the

Auditory Verbal Learning Test (AVLT), the Stroop Color Word Test (SCWT) and the

Symbol Digit Modalities Test (SDMT). Near-infrared spectroscopy (NIRS) was used to

determine intra-operative episodes of cerebral desaturation, defined as a decrease in

cerebral tissue oxygen saturation (rScO2) of more than 20% compared to the baseline

value. Routine anesthesia was applied. A repeated measures multivariate analysis of

variance with time as within-subject variable (pre-surgery, post-surgery, and follow up

assessment) was performed.

Results

Cerebral desaturation occurred in all patients. No significant changes in overall

cognitive test performance was observed over time, neither was there any significant

interaction between time and age or cerebral desaturation. A significantly increased

performance was observed between the postoperative and the follow-up assessment

on the AVLT (p=.014), which may possibly be attributed to a test-retest effect.

Conclusion

These preliminary results suggest no association between episodes of cerebral

desaturation and postoperative cognitive dysfunction. The current sample is yet too

2

small to come to definite conclusions, and confirmation in an adequate sample-sized

study population (as is planned in this ongoing study) is mandatory.

3

Introduction

Beach chair position

Surgery on the shoulder joint is one of the most common surgical procedures

performed by orthopedic specialists. Since its introduction in the early eighties beach

chair position is the most frequently preferred position for both open and arthroscopic

shoulder surgery. Patients are positioned sitting up more than 70 degrees upright with

the operative arm draped free to be manipulated by either a mechanical arm rest or a

surgical assistant. The head rests on a well-padded head block.

Figure 1: Classical beach chair position (picture copyrighted by Ao Foundation)

The advantages of the classic beach chair position include easier access to the joint,

lack of brachial plexus strain, less risk for neurovascular trauma and if necessary an

easy conversion from arthroscopic to open procedure (Skyhar et al., 1988; Gelber et

al., 2008).

4

The safety of this position has been well established. Most patients who undergo

shoulder surgery in this position do not show adverse sequelae (Friedman et al., 2009).

Nonetheless severe adverse neurologic events have occurred in a relatively healthy

middle-aged population supposedly not at risk for severe cerebrovascular incidents. A

case of visual loss and ophthalmoplegia has been described after a surgically

uneventful rotator cuff repair in 2003 (Bhatti and Enneking, 2003). Pohl and Cullen

have reported four cases of ischemic brain damage and spinal cord injury, resulting in

a permanent vegetative state and death in two cases (Pohl and Cullen, 2008). The

American Shoulder and Elbow Surgeon’s Association has reported eight cases of

cerebrovascular events during shoulder surgery in beach chair position (Friedman et

al., 2009). The etiologic mechanisms are largely unknown. It is postulated that the

upright position induces unfavorable hemodynamic changes. In healthy awake

volunteers, assuming a sitting position induces a sympathic nervous system response.

Heart rate and systemic vascular resistance rise to maintain adequate mean arterial

pressure and cardiac output (Smith et al., 1994; Edgell et al., 2004). This autonomous

effect does not diminish with increasing age (Edgell et al.2004; Murell et al., 2011).

During general anesthesia these sympathic responses are diminished due to the

vasodilation effect and blunting of the baroreceptor reflex of anesthetics, resulting in

an attenuated increase in system vascular resistance, decrease in mean arterial

pressure and a decreased cardiac output. A high incidence of severe hypotension is

not surprising during general anesthesia in a sitting position (Potter et al., 1999). These

profound physiological changes may lead to undetected cerebral hypotension. This

reduction in cerebral perfusion pressure below a critical threshold of duration and

severity may induce neurological damage. This threshold is felt to be the lower limit of

cerebral autoregulation, which according to recent publications is a subject of

physiological heterogeneity. While it attempts to maintain a constant cerebral perfusion

pressure over a wide range of mean arterial pressures, both lower and upper limits are

subject of interpatient variability, taking patient’s age, comorbidities and pharmacologic

therapies in consideration (Drummond, 1999; Sanders et al, 2011). Cerebral

hypotension may go unnoticed using standard intraoperative monitoring. Cerebral

perfusion pressure can be estimated using mean arterial pressure (MAP) (Murphy and

Szokol, 2011). However cerebral perfusion pressure can be severely underestimated

using brachial cuff readings, due to hydrostatic pressure differences between head and

5

arm/heart level. This difference can be as much as 20 to 40 mmHg depending on the

elevation of the head (Koh et al, 2012)

Near-infrared spectroscopy: technical aspects and clinical use

Near-infrared spectroscopy (NIRS) is a non-invasive technique allowing

measurements of tissue oxygenation, including brain oxygenation. It became possible

to measure regional cerebral oxygen saturation (rScO2) after the innovative work of

Jöbsis. He demonstrated that light in the near-infrared spectrum can penetrate the

bone of the cranium, allowing measurements of oxyhemoglobin and deoxyhemoglobin

inside the brain (Jöbsis, 1977). Non-invasive, self-adhesive optical sensors are applied

to the skin of the forehead. These optrodes emit a light in the near-infrared spectrum.

Its reflection is measured by sensors at fixed distances from the light source. The

strength of the reflected light detected by these sensors is inversely related to the

concentration of light-absorbing molecules in the underlying tissue. Oxyhemoglobin

and desoxyhemoglobin absorb light at a wavelength of 800 nm in the near infrared

spectrum, but they both have a different characteristic peak absorption. Using a

modified Beer-Lambert law, NIRS provides a measurement of the concentration of

oxygenated and deoxygenated hemoglobin to total hemoglobin concentration.

Algorithms are used to distract absorption of superficial extracranial tissue from that

obtained from deeper cerebral tissue to obtain a cerebral saturation value of the

superficial frontal cortex (Zheng et al., 2013). Traditional pulse oximetry, which uses

light in the red and infrared spectrum, distinguishes between venous and arterial blood.

NIRS does not make this differentiation. Approximately 70 to 80 % of cerebral blood is

venous blood, consequently rScO2 provides an indication of the regional oxygen supply

and demand (McCormick et al, 1991; Murking and Arango, 2009). Continuous

measurements of cerebral oxygenation allow for an accurate recognition of episodes

of cerebral desaturation. NIRS is used extensively in major surgery at high risk for

adverse neurologic outcome. It has found its way into the arsenal of the cardiac

anesthesiologist. NIRS may predict postoperative neurological dysfunction after

cardiac surgery and allows detection of intraoperative cannula misplacement during

cardiopulmonary bypass (Zheng et al, 2013). During pediatric cardiac surgery, NIRS

6

is the most widely applied and accepted tool in multimodal neuromonitoring. Some

evidence exists in the pediatric population that decreased cerebral oxygen values are

associated with brain abnormalities on postoperative MRI scans, though no conclusive

evidence exists on the neurologic outcome after perioperative intervention solely

based on NIRS values (Clark et al., 2012). NIRS not only has its part in cardiac surgery,

but it is widely used during major vascular procedures. NIRS offers the possibility to

predict neurologic events during carotid endarterectomy. About ten percent of patient

suffer from postoperative neurologic changes triggered by carotid clamping. A

decrease in rScO2 is related to potential postoperative neurologic deterioration and

might imply the need for a different anesthetic approach (Pennekamp et al, 2009),

using deliberate hypertension, or surgical approach, using a carotid shunt (Nielsen,

2014). NIRS also opens the possibility to predict a hyperperfusion syndrome in the

immediate postoperative period after releasing the carotid clamp (Pennekamp et al.

2012). Similar applications of NIRS are performed during abdominal surgery, both

laparoscopic and open. Inducing a head up position for laparoscopic procedures

induces a decrease in rScO2, up to 80% of baseline value in 20% of patients (Gipson

et al., 2006). Similar results were found during laparoscopic gynecological procedures

(Lee et al., 2006). Clamping the inferior caval vein during orthotopic liver

transplantation, induces a decrease in rScO2. A cerebral desaturation of 13% has been

associated with release of neuron specific enolase, a biomarker of neuronal damage

(Plachky et al., 2004). This makes NIRS a viable option for monitoring cerebral

autoregulation and subsequently hemodynamic management during major abdominal

and transplantation procedures.

Few studies exist assessing events of cerebral desaturation and their

neurocognitive sequelae. A systematic review by Zheng and colleagues only cited low-

level evidence that interventions to improve cerebral saturation could prevent stroke or

post-operative cognitive dysfunction during cardiac surgery with cardiopulmonary

bypass (Zheng et al., 2013). Two studies examine the effects of cerebral desaturation

and occurrence of neurocognitive dysfunction in abdominal surgery. They both

associated cerebral desaturation with postoperative cognitive decline, as evidenced by

a lower mini-mental state examination (MMSE) score or the occurrence of

postoperative delirium (Casati et al, 2007; Morimoto et al., 2009). This association was

also confirmed during thoracic surgery (Tang et al., 2012). Several observational

7

studies have investigated the use of NIRS during shoulder surgery in different

positions. During arthroscopic procedures in lateral decubitus rScO2 is maintained. On

the other hand beach chair position may induce a dramatic decrease in rScO2 (Fisher

et al., 2009; Dippmann et al. 2010, Tange et al, 2010; Lee et al. 2011; Yadeau et al.,

2011; Jeong et al., 2012; Ko et al., 2012; Moerman et al. 2012; Salazar et al, 2013;

Pant et al, 2014). Lowered cerebral saturation corresponded with a decreased MAP.

A restoration of MAP to preoperative value restored rScO2 (Lee et al., 2011).

Postoperative decline in cognitive function

Postoperative decline in cognitive function is becoming increasingly recognized

in Postoperative cognitive dysfunction (POCD) is broadly defined as a decline in

cognition temporally associated with surgery. Cognitive function is a rather ‘simple’

name for a broad and elaborate spectrum of capabilities such as learning,

understanding, perception, attention, verbal abilities, executive functions and abstract

thinking. Unlike delirium, no elaborate definition exists to accurately describe cognitive

dysfunction. It is considered a mild neurocognitive disorder and, according to DSM IV

criteria, it can only be identified as an exclusion diagnosis. Other neurocognitive

disorders such as delirium, dementia and other amnestic disorders have to be

excluded first (Alcover et al, 2013). While it is categorized as a trivial brain dysfunction,

it has been associated with an increased one year mortality, impairment of daily

functioning, risk of leaving the labor market or early retirement, increased health care

cost and dependency on government support payments (Deiner and Silverstein, 2009).

Cognitive deterioration after cardiac surgery has been described in about three

to eight patients out of ten, with ongoing cognitive deficits after six months in up to 60%

of these patients. POCD had been described in one out of four elderly patients in the

first seven days after major non-cardiac surgery. The cognitive symptoms persist in

about ten to twenty percent following major surgery after six months. Incidence is lower

in younger patients and minor surgery (Steinmetz and Rasmussen, 2016).

8

The pathophysiologic mechanisms leading to POCD after surgery are poorly

understood. Animal studies attribute an extensive role to the immune system

succeeding surgical ‘trauma’. Peripheral surgery in a mice model activates the

inflammatory TNFα/NF-κB pathways with cytokines release. In turn these cytokines

disrupt the integrity of the blood-brain barrier, which allows macrophages to invade the

hippocampal formation more easily, impairing memory formation. Cognitive functioning

remains preserved when cytokine release is blocked by activation of the anti-

inflammatory cholinergic pathways (Terrando et al., 2011). POCD has been

extensively studied in cardiac surgery. Longer duration of cardiopulmonary bypass,

poor left ventricular function, manipulation of an atherosclerotic aorta, cerebral

hyperthermia and valve surgery have all been associated with postoperative cognitive

deterioration (Tomotko and Maekawa, 2014). Although cognitive dysfunction has been

less comprehensively studied and identified in non- cardiac surgery, numerous risk

factors have been identified in more recent studies. Table 1 gives an overview of the

most common risk factors (Rundshagen, 2014).

Table 1: Risk factors for POCD

Patient Advanced age; pre-existing cerebral, cardiac or vascular disease;

preoperative mild cognitive impairment; low educational level; history

of alcohol abuse; genetic predisposition

Surgery Extensive surgical procedure; intra- or postoperative surgical

complications; secondary surgery; cardiac surgery (long

cardiopulmonary bypass, poor cardiac function, valve surgery and

atherosclerotic aorta handling)

Anesthesia Long-acting anesthetics; marked disturbances of homeostasis

(hypoxemia, hypotension), intra- or postoperative anesthesia

complications

Several observational studies have revealed an association between extensive

surgery, secondary surgery and postoperative complications and an increased

incidence of POCD (Steinmetz and Rasmussen, 2016). Age is a major risk factor for

9

POCD. Increasing age implies a decreased cognitive performance and ability to

compensate for cognitive deficits. This coincides with the ‘brain reserve’ theory, which

shows a greater vulnerability for POCD in brains with a lower reserve. This also

includes patients with a pre-existing mild cognitive impairment (Deiner and Silverstein,

2009). A low educational level is a further risk factor for POCD. Several studies have

presumed the expression of the APOE 4 allele as a risk factor (Alcover et al, 2013).

Hudetz et al showed that patients with a positive history for alcohol abuse had worse

cognitive impairment after surgery. A severe disruption of homeostasis, such as severe

hypotension or hypoxemia, have been blamed for postoperative cognitive deterioration

(Steinmetz and Rasmussen, 2016; Rundshagen, 2014). Even though general

anesthesia has often been blamed for causing neurotoxicity, currently no sufficient

evidence exists to link POCD to general anesthesia. The residual effect of the

anesthetics solely depends on its pharmacokinetic properties. A shorter duration of

action implies a shorter impairment of cognitive function in the postoperative period.

Diagnosis of postoperative cognitive dysfunction requires a preoperative

baseline cognitive function analysis and a determination of what will be defined as a

significant decline in cognitive functioning. Cognitive deterioration is a subtle process

and can only be detected using a combination of neuropsychological tests. Current

test batteries for POCD have several issues. First there is the flooring effect, i.e. what

is the correct technical approach for a patient with an already low baseline score.

Secondly there is currently no real consensus about which tests have to be used to

diagnose POCD. A 1995 recommendation lists four essential tests to diagnose POCD:

Rey Auditory Verbal Learning test (a word learning test), Trail Making test (ability to

perform combined tasks), Groove Pegboard test (manual dexterity) and Digit Span test

(ability to remember a sequence of numbers) (Murkin et al., 1995). This variety of tests

may be due to the absence of theoretically substantiated model that accurately

describes causes and outcomes of cognitive changes due to surgical trauma. This

confusing recommendation resulted in a lot of combinations of different tests have

being in clinical studies, making data interpretation and comparison a real challenge.

A possible explanation is the duration of these combinations of tests. A

comprehensive neuropsychological examination takes about three hours, which is not

always suitable in a perioperative setting. Currently diagnosis of POCD implies testing

10

in the following domains: learning and memory, attention, executive functioning and

language, supplemented by a decline in perception and abstract thinking (Tsai et al,

2010). Finally, there is no clear agreement about the optimal timing when to test

patients at risk for POCD. Patients undergoing their preoperative cognitive assessment

may not score as well as expected due to pre-procedural anxiety or effects of

premedication. Patients who are tested shortly after surgery may have substandard

scoring due to residual effect of the medication or anesthetics, pain and general health

status. Trials requiring long-term follow-up may experience loss to follow up in patients

suffering from POCD, leading to confounding by attrition.

Postoperative cognitive decline had been extensively tested in cardiac surgery,

but less in non-cardiac surgical patients. Several studies have tried to associate the

occurrence of POCD with various adjustable surgical and anesthetic factors. Currently

only two clinical studies have addressed the association between cerebral desaturation

during shoulder surgery in beach chair position and postoperative cognitive

dysfunction. One study did not find any significant change in postoperative MMSE

score (Salazar et al., 2013). The other study combined both executive function tests

and serum biomarkers of brain injury. No differences in cognitive outcome were found

between the beach chair group and supine position control group (Laflam et al, 2015).

11

Goal of the study

This study aims to identify episodes of cerebral desaturation using NIRS during

shoulder surgery in beach chair position and will try to associate these episodes to

potential postoperative neurocognitive dysfunction using a combination of

neurocognitive tests. These will test executive functions mediated by the frontal lobe,

i.e. the location where NIRS is used to measure cerebral oxygenation. Considering the

underlying shoulder pathology, all tests will be motor free. Furthermore patient risk

factors for events of cerebral desaturation and postoperative neurocognitive

dysfunction will be investigated. We hypothesize that events of cerebral desaturation

are connected to neurocognitive dysfunction in the postoperative period.

12

Materials and methods

Approval from the local ethics committee was obtained. Written informed

consent was obtained from study candidates. All consecutive patients planned for total

shoulder arthroplasty, were screened for inclusion after referral to the preoperative

consultation. The average time interval between this consultation and date of surgery

was ten days. Exclusion criteria were age younger than 18, prior neurologic event,

amourosis fugax, apparent cognitive dysfunction, carotid stenosis, previous head and

neck surgery, cervical disc herniation, recent acute myocardial infarction and syncope.

Elderly patient aged 75 or more were screened using a mini-mental state examination

for gross neurocognitive deficit prior to inclusion in the study. A score lower than 24 is

compatible with a preoperative existing neurocognitive deficit, thus these patients were

excluded. Collected demographic data included age, gender, level of education,

smoking status, prior health problems (diabetes mellitus, chronic obstructive

pulmonary disease, metabolic syndrome, pre-existing cardiovascular disease), length,

weight, body mass index (BMI) and current medication use. Education level was

defined as years in any form of education (primary education, secondary education,

college/university, doctorate, night classes). A baseline preoperative neurocognitive

function was assessed using the ‘Auditory Verbal Learning Test’ (AVLT), ‘Stroop Color

Word Test’ (SCWT) and ‘Symbol Digit Modalities Test’ (SDMT). All neurocognitive

tests evaluate frontal lobe mediated executive functions. All tests were motor free,

considering the underlying shoulder pathology. Patients did not receive any

psychoactive medications (e.g. anxiolytics) before performing this neurocognitive

assessment.

No premedication or anxiolytics were administered pre-operatively. After arrival

in the operating room, the standard protocol for shoulder surgery was followed. An

intravenous line was placed and normal saline was used as a standard. A three or five

leads ECG, depending on the cardiovascular status of the patient, was applied to the

patient. A pulse oximeter was secured to the index finger of the non-operated hand. A

blood pressure cuff was positioned on the non-operated arm. Non-invasive blood

pressure was measured every three minutes. All three were connected to the

anesthesia monitor. Before induction of general anesthesia an ultrasound guided

interscalene block was placed using a mixture of plain levobupivacaine 0.5% and

13

mepivacaine 2%. Two INVOS sensors (Covidien, Medtronic, Minneapolis, USA) and a

bispectral index (BIS) sensor (Covidien, Medtronic, Minneapolis, USA) were applied

on the patient’s forehead. Anesthesia was induced with sufentanil (0.1 µg/kg), propofol

(2 – 3 mg/kg) and cisatracurium (0.1 mg/kg). After reaching adequate depth of

anesthesia and neuromuscular blockade an endotracheal tube was placed. Anesthesia

was maintained with a mixture of 50% air and 50% inspired oxygen and sevoflurane

(age corrected MAC value of 1) and adapted to obtain BIS values between 40 and 60.

Additional doses of muscle relaxants and opiates were at the discretion of the treating

anesthesiologist. An oral thermometer was placed and patient’s temperature was kept

at a minimum of 35.5°C using forced air warming. A baseline cerebral oximetry value

was obtained after obtaining adequate depth of anesthesia and before positioning.

Afterwards the patient was positioned in the beach chair position. Patients were raised

to a 60 to 70 degrees sitting position with the head fixed in midline position. Continuous

cerebral oximetry was obtained during the entire anesthetic and surgical procedure.

All personnel in the operating room was blinded for cerebral oximetry measurements.

Likewise all auditory signals of the device were disabled.

Cerebral desaturation was defined as a decline in cerebral oxygen saturation of

20%, compared to baseline value. An area under the curve combining the duration and

severity of cerebral desaturation (AUCrScO2) was calculated. Intraoperative hypotension

was defined as systolic blood pressure lower than 80 mmHg. Bradycardia was defined

as a heart rate below 45 beats per minute. They were treated with ephedrine and

atropine respectively, according to current standard of care guidelines in our operating

theatre. In view of the current debate concerning the use of phenylephrine and its effect

on cerebral oxygenation, this drug was not our first choice of medication. Standard

postoperative nausea and vomiting and pain protocols were followed. The amount of

blood loss and fluids given to the patient as well as the doses of all drugs administered,

were recorded. After the surgical procedure, the patient was awakened and

transported to the postoperative care unit, where they were discharged after reaching

discharge criteria (Aldrete score > 8).

The first postoperative testing took place on the third postoperative day

(concordant with the day of hospital discharge in the majority of standard cases). The

‘Auditory Verbal Learning Test’ (AVLT), ‘Stroop Color Word Test’ (SCWT) and ‘Symbol

14

Digit Modalities Test’ (SDMT) were repeated using alternative versions where

appropriate to minimize test-retest effects.

The second postoperative testing took place three months post-surgery. This

coincided with a follow-up consultation with the treating orthopedic surgeon in the

outpatient clinic. The ‘Auditory Verbal Learning Test’ (AVLT), ‘Stroop Color Word Test’

(SCWT) and ‘Symbol Digit Modalities Test’ (SDMT) were assessed again. After this

follow-up moment, the study was ended. All neurocognitive tests were performed by

the same person.

Statistical analysis was performed using the SPSS Statistics version 23 (2015)

(IBM, New York, USA). Data were tested for normal distribution using the Shapiro –

Wilk test. Normally distributed data are reported as mean and standard deviation (SD).

The incidence of cerebral desaturation in patients older than 65 and patients suffering

from hypertension was analyzed using the Fisher Exact test. Correlation were

calculated using Pearson correlation coefficient. The pre-surgery, post-surgery and

follow-up test results were analyzed using a repeated multivariate analysis of variance

(repeated ANOVA). Age and degree of cerebral oxygen desaturation were included as

covariates. A p - value lower than 0.05 was considered statistically significant.

15

Results

21 consecutive patients, planned for total shoulder arthroplasty, were screened

for inclusion. After medical evaluation 12 (57.1%) were included in the study. One

patient did not complete the second postoperative neurocognitive evaluation and was

considered lost to follow-up. Table 2 shows relevant demographic data of included

patients. Consequently statistical analysis was performed on the data of 11 patients,

who completed all three neurocognitive evaluations after surgery. Two patients

(18.2%) required additional surgery in beach chair position because of surgical

complications. One patient required shoulder reduction under general anesthesia two

months postop after prosthesis luxation. The other patient required evacuation of a

hematoma at the surgical site under general anesthesia four days after the initial

surgery.

Table 2: Patients characteristics

Sex, male/female (n (%)) 4 (36.4%) / 7 (63.6%)

Age, years (mean ± SD) 65 ± 12.39

Education level, years (mean ± SD) 10.6 ± 3.8

Hypertension, yes/no (n (%)) 5 (45.5%) / 6 (54.5%)

Table 3 shows relevant information about anesthesia and surgery. All patients

received crystalloid solutions. Following current controversy about fluid administration,

colloids were only used in severe blood loss, which none of our patients suffered from.

The surgical procedure did not necessitate deliberate hypotension, so current standard

practice in case of low blood pressure was followed. The use of phenylephrine as first-

line treatment of hypotension was discouraged. 7 out of 11 patients (63.63%) received

ephedrine (6 – 33 mg) and 2 patients (18.2%) received phenylephrine (100 – 700 µg)

for systolic blood pressure lower than 80 mmHg. Phenylephrine was only used as a

second line agent in hypotension refractory to ephedrine. Two patients (18.2%)

received a fluid bolus of crystalloids for the treatment of hypotension. No patients

received atropine for a heart rate lower than 45 beats per minute.

16

Table 3: Anesthesia and surgical characteristics

Duration BCP, min (mean ± SD) (min – max) 110 ± 32 (48 – 192)

Blood loss, ml (mean ± SD) (min – max) 325 ± 352 (120 – 1200)

Total fluids, ml (mean ± SD) (min – max) 789 ± 487 (360 – 2250)

Drugs use, n (%) Atropine: 0 (0%)

Ephedrine: 7 (63.63%)

Phenylephrine: 2 (18.2%)

Fluid bolus: 2 (18.2%)

The descriptive statistics of the neurocognitive tests at the three different points

in time are displayed in table 4. The AVLT score represents a learning score. The total

score represents the number of correct answer during the first five AVLT test runs. The

SCWT is scored as the number of correct answers. The SDMT represent the time in

seconds to finish the entire test.

Table 4: Descriptive statistics of the neurocognitive tests

SCWT, n (mean ± SD) T1 55.52 ± 35.67

T2 64.74 ± 43.89

T3 56.35 ± 45.0

AVLT, n (mean ± SD) T1 38.18 ± 10.27

T2 38.18 ± 12.41

T3 48.73 ± 13.07

SDMT, s (mean ± SD) T1 42.36 ± 14.73

T2 42.18 ± 15.79

T3 47.64 ± 10.45

No patient experienced gross neurological complications. Mean baseline rScO2

before induction was 60.9% and 62.4%, respectively on the left and right side. Before

positioning mean rScO2 on the left side was 71.2% and 75.2% on the right side. After

positioning to beach chair positioning, there was a relative decrease of rScO2 of more

17

than 20% in 3 patients (27.3%). During the overall surgical procedure in beach chair

position there was a relative decrease of rScO2 of more than 20% in all eleven patients

(100%). The mean AUCrScO2 during the entire surgical procedure in beach chair

position was 73.42 %*sec on the left side and 180.92 %*sec on the right side. Age

older than 65 did not appear to be a significant risk factor for cerebral desaturation after

positioning in beach chair (χ² = 0.244; P = 0.62). Likewise, adequately treated

hypertension did not appear to be a risk factor for cerebral desaturation after

positioning (χ² = 0.749; P = 0.39). Similarly no correlation was found between age and

AUCrScO2 on the right (r = 0.081, P = 0.81) and left side (r = -0.040, P = 0.91).

A repeated measures multivariate analysis of variance with a Greenhouse-

Geisser correction with time as within-subject variable (pre-surgery, post-surgery, and

follow up assessment) was performed. Age and degree of rScO2 desaturation were

included as covariates and SCWT, SDMT and AVLT as measures. The analysis

revealed no significant changes in overall cognitive test performance over time. In

addition, no significant interaction between time and age or rScO2 desaturation was

unveiled. Post-hoc tests only revealed a significant rise in score between the

postoperative and follow-up assessment on the AVLT performance (P = 0.014), which

can be attributed to a test-retest effect.

18

Discussion

To our knowledge, this is the first study to examine the incidence of cerebral

desaturation and its association with postoperative neurocognitive sequelae in such a

homogenous surgical group. All patient underwent the same surgical procedure, a total

shoulder arthroplasty, with a similar amount of surgical time and duration of beach

chair position. As demonstrated in previous research, a high incidence of cerebral

desaturation occurs when a patient’s position is changed from supine to beach chair.

Our results confirmed a cerebral desaturation during the surgical procedure in beach

chair position in all of our patients. This corresponds with previous research (Fisher et

al., 2009; Dippmann et al. 2010, Tange et al, 2010; Lee et al. 2011; Yadeau et al.,

2011; Jeong et al., 2012; Ko et al., 2012; Moerman et al. 2012; Salazar et al, 2013;

Pant et al, 2014). Episodes of cerebral desaturation were similar in patients with and

without chronic hypertension. Finally our study showed no decline in cognitive outcome

after shoulder surgery in beach chair position.

The exact frequency of neurologic complications in patients undergoing surgery

in beach chair position is unknown. Major stroke occurs in 0.0004% of patients

according to a survey of the American Shoulder and Elbow Surgeons (Friedman et al.,

2009). When the treating anesthetist fails to appreciate the effect of hemodynamic

changes caused by the gravitational effect of the head-up position, cerebral

hypoperfusion may ensue. Blood pressure decreases proportionally to the weight of

the ‘fluid’ column. A subtraction of 1.35 mmHg per cm of head elevation is advised to

estimate the actual blood pressure at the brain. Although we were not able to provide

beat-to-beat blood pressure measurements and link them directly to episodes of

cerebral desaturation, several previous studies have shown the connection between

cerebral desaturation and hypotension. Not all patients will subsequently develop

cerebral ischemia, if compensatory mechanism such as a higher oxygen extraction

reserves are sufficient. At risk are patients with altered cerebral autoregulation (such

as untreated hypertension), cerebral atheromatosis and other problems with the circle

of Willis.

19

Cerebral desaturation was not associated with neurocognitive deficit in our

study population. In the majority of studies trying to find an association between

neurocognitive dysfunction and operations in beach chair position, no systematic

neuropsychological testing was performed. This is the second study that uses a test

battery of validated neuropsychological tests. Although our neuropsychological testing

is compatible with the 1995 consensus guideline and subsequent revisions, we chose

not to include motor testing, thus disabling some executive functions and

frontotemporal interactions. We opted to retest our patient on the third postoperative

day to reduce the possibility of residual effect of medication and anesthesia, which

could affect test scores negatively. Postoperative cognitive dysfunction may not always

be apparent during the hospital stay. We opted to retest at the third month

postoperative. A significant retesting effect was found in the AVLT, a test with memory

forming and recall as prime goal. Substitution to a more difficult test could possibly

eliminate this effect and improve test quality. A possible addition to our neurocognitive

testing could be the use of specific enzymatic markers of cerebral damage. Neuron-

specific enolase (NSE) is a neuronal isoform of the glycolytic enzyme enolase and is

more specific for grey matter lesions. Protein S100B is an intracellular calcium-binding

dimer, implicated in the proliferation and differentiation of neurons. It is more specific

for white matter. Considering the different characteristics, the more prudent approach

would be to analyze both, though current evidence has yet to define a clear consensus

about cut - off values compatible with neuronal damage, especially in the case of the

more subtle neurocognitive dysfunction. Other biomarkers of brain damage are

currently under investigation in a post-resuscitation and traumatic brain injury setting

(Tomaszewski, 2015) Only two studies show an association between POCD and

elevated serum markers of brain damage (Xie et al., 2013; Lagram et al., 2015).

Another possible option is the electroencephalogram (EEG). Patients who develop

cognitive dysfunction show less burst suppression and states of deep sleep during

anesthesia (Deiner et al., 2015). Thus far no study has investigated the use of EEG in

the postoperative diagnosis of POCD.

Our study has several limitations. The small sample size and lack of power could

explain the absence of a statistically significant result. This has several implications.

Our smaller study population did not allow for any proper subgroup analysis. Perhaps

the effect of cerebral desaturation does not have significant cognitive implications in

20

healthy patients, but may prove to be important in patients with cardiovascular risk

factors, such as peripheral artery atheromatosis, diabetes mellitus with associated

autonomous neuropathy, severe untreated hypertension and coronary artery disease.

No analysis was possible in the group of older patients (> 65 years old), who might be

at increased risk for the development of neurocognitive dysfunction. Our inclusion

criteria were strict. We chose not to admit any patients with a pre-existing

neurocognitive deficit or psychiatric disorder to the study protocol. Nonetheless both

patients group have an impaired cerebral reserve and are especially at risk for the

development of POCD. Another limitation is the lack of control group. We did not

compare our beach chair group with a similar non-beach chair group. Prolonged

shoulder surgery in lateral decubitus or other major joint surgery could be considered

as a control group in later study protocols.

In conclusion, a large fraction of our study patients undergoing shoulder surgery

in beach chair position are prone to experience episodes of cerebral desaturation at

positioning and during surgery. In our patient population, this cerebral lack of oxygen

was not correlated with any development of postoperative cognitive dysfunction. In

patients at risk for cerebral desaturation, specific monitoring and appropriate

hemodynamic interventions seems warranted as long as a potential relation between

cerebral desaturation and POCD has not been definitively excluded.

21

References

ALCOVER L., BADANES R., MONTERO M.J., SORO M., BELDA F.J.: Postoperative

delirium and cognitive dysfunction. Trends Anaesth Crit Care, 2013, 199-204.

BHATTI M.T., ENNEKING F.K.: Visual loss and ophthalmoplegia after shoulder

sugery. Anesth Analg, 2003, 899-902.

CASATI A., FANELLI G., PIETROPAOLI P., PROIETTI R., MONTANINI S.: Monitoring

cerebral oxygen saturation in elderly patients undergoing general abdominal sugery: a

prospective cohort study. Eur J Anaesthesiol., 2007, 59-65.

CLARK J.B., BARNES M.L., UNDAR A., MYERS J.L.: Multimodality neuromonitoring

for pediatric cardiac surgery: our approach and critical appraisal of the available

evidence. World J Pediatr Congenit Heart Surg, 2012, 87-95.

DEINER S., LUO X., SILVERSTEIN J.H., SANO M.: Can intraoperative processed

EEG predict postoperative cognitive dysfunction in the elderly?, Clin Ther, 2015, 2700-

2705.

DEINER S., SILVERSTEIN J.H.: Postoperative delirium and cognitive dysfunction. Br.

J. Anaesth, 2009, i41-i46.

DIPPMANN C., WINGE S., NIELSEN H.B.: Severe cerebral desaturation during

shoulder arthroscopy in the beach-chair position. Arthroscopy, 2010, 148-150.

DRUMMOND J.C.: The lower limit of cerebral autoregulation: time to revise our

thinking. Anesthesiology, 197, 1481-1483.

EDGELL H., Robertson A.D., HUGHSON R.L.: Hemodynamics and brain blood flow

during posture change in younger woman and postmenopausal women compared with

age-matched men. J Appl Phsyiol, 2012, 1482-1493.

ELY E.W., SHINTAI A. TRUMAN B., SPEROFF T., GORDON S.M., HARRELL F.E.J.,

INOUYE S.K., BERNARD G.R., DITTUS R.S.: Delirium as a predictor of mortality in

mechanically ventilated patients in the intensive care unit. JAMA, 2004, 1753-1762.

22

FISHER G.W., TORRILLO T.M., WEINER M.M., ROSENBLATT M.A.: The use of

cerebral oximetry as a monitor of the adequacy of cerebral perfusion in a patient

undergoing shoulder surgery in the beach chair position. Pain Pract, 2009, 304-307.

FRIEDMAN D.J., PARNES N.Z. ZIMMER Z., HIGGINS L.D., WARNER J.J.P.:

Prevalence of cerebrovascular events during shoulder surgery and association with

patient position. Orthopedics, 2009.

GELBER P.E., REINA F., CACERES E., MONLLAU J.C. A comparison of risk between

the lateral decubitus and the beach-chair position when establishing an antero-inferior

shoulder portal: a cadaveric study. Arthroscopy, 2007, 522-528.

GIPSON C.L., JOHNSON G.A., FISHER R., STEWART A., GILES G., JOHNSON J.O.,

TOBIAS J.D.: Changes in cerebral oximetry during peritoneal insufflation for

laparoscopic procedures. J Minim Access Surg, 2006, 67-72.

GUENTHER U., RIEDEL L., RADTKE F.M.: Patients prone for postoperative delirium:

preoperative assessment, perioperative prophylaxis, postoperative treatment. Curr

Opin Anaesthesiol, 2016, ePub ahead of print.

HUDETZ J.A., IQBAL Z., GHANDHI S.D.: Postoperative cognitive dysfunction in older

patients with a history of alcohol abuse. Anesthesiology, 2007, 423-430.

JEONG H., JEONG S. LIM H.J, LEE J., YOO K.Y.: Cerebral oxygen saturation

measured by near-infrared spectroscopy and jugular venous bulb oxygen saturation

during arthroscopic shoulder surgery in beach chair position under sevoflurane-nitrous

oxide or propofol-remifentanil anesthesia. Anesthesiology, 2012, 1047-1056.

JOBSIS F.F.: Noninvasive, infrared monitoring of cerebral myocardial oxygen

sufficiency and circulatory parameters. Science, 1977, 1264-1267.

KO S.H., CHO Y.W., PARK S.H., JEONG J.G., SHIN S.M., KANG G.: Cerebral

oxygenation of patients during arthroscopic shoulder surgery in the sitting position.

Korean J Anesthesiol, 2012, 297-301.

23

KOH J.L., LEVIN S.D., CHEHAB E.L., MURPHY G.S.: Cerebral oxygenation in the

beach chair position: a prospective study on the effect of general anesthesia compared

with regional anesthesia and sedation. J Shoulder Elbow Surg, 2013, 1325-1331.

LAFLAM A., JOSHI B., BRADY K., YENOKYAN G., BROWN C., EVERETT A.,

SELNES O., MCFARLAND E., HOGUE C.W.: Shoulder surgery in the beach chair

position is associated with diminished cerebral autoregulation but no differences in

postoperative cognition or brain injury biomarker levels compared with supine

positioning: the Anesthesia Patient Safety Foundation Beach Chair Study. Anesth

Analg, 2016, 176-186.

LEE J.H., MIN K.T., CHUN Y.M., KIM E.J., CHOI S.H.: Effects of beach-chair position

and induced hypotension on cerebral oxygen saturation in patients undergoing

arthroscopic shoulder surgery. Arthroscopy, 2011, 889-894.

LEE J.R., LE P.B., DO S.H., JEON Y.T., LEE J.M., HWANG J.Y., HAN S.H.: The effect

of laparoscopic surgery on cerebral oxygenation. J Int Med Res, 2006, 531-536.

LIN S.M., LIU C.Y., WANG C.H., HUANG C.D., HUANG P.Y., FANG Y.F., SHIEH M.H.,

HUO H.P.: The impact of delirium on the survival of mechanically ventilated patients.

Crit Care Med, 2004, 2254-2259.

MOERMAN A.T., DE HERT S.G., JACOBS T.F., DE WILDE L.F., WOUTERS P.F.:

Cerebral oxygen desaturation during beach chair position. Eur J Anaesthesiol, 2012,

82-87.

MORIMOTO Y., YOSHIMURA M., UTADA K., SETOYAMA K., MATSUMOTO M.,

SAKABE T.: Prediction of postoperative delirium after abdominal surgery. J Anesth,

2009, 54-56.

MCCORMICK P.W., STEWART M., GOETTING M.G., DUJOVNY M., LEWIS G.,

AUSMAN J.L.: Noninvasive cerebral optical spectroscopy for monitoring cerebral

oxygen delivery and hemodynamics. Crit Care Med, 1991, 89-97.

MILLBRANDT E.B., DEPPEN S., HARRISON P.L., SHINTANI A.K., SPEROFF T.,

TRUMAN B., BERNARD G.R., DITTUS R.S., ELY E.W.: Costs associated with delirium

in mechanically ventilated patients. Crit Care Med, 2004, 955-962.

24

MURKIN J.M., ARANGO M.: Near-infrared spectroscopy as an index of brain and

tissue oxygenation. Br J Anaesth, 2011, 3-13.

MURKIN J.M., NEWMAN S.P., STUMP D.A., BLUMENTHAL J.A.: Statement of

consensus on assessment of neurobehavioral outcomes after cardiac surgery. Ann

Thorac Surg, 1995, 1289-1295.

MURPHY G.S., SZOKOL J.W.: Blood pressure management during beach chair

position shoulder sugery: What do we know? Can J Anaesth, 2011, 977-982.

MURRELL C.J., COTTER J.D., GEORGE K., SHAVE R., WILSON L., THOMAS K.,

WILLIAMS M.J., AINSLIE P.: Cardiorespiratory and cerebrovascular responses to

head-up tilt I: influence of age and training status. Exp Gerontol., 2001, 9-17.

NIELSEN H.B.: Systematic review of near-infrared spectroscopy determined cerebral

oxygenation during non-cardiac surgery. Frontiers Physiol, 2014.

PANT S., BOKOR D.J., LOW A.K.: Cerebral oxygenation using near-infrared

spectroscopy in the beach-chair position during shoulder arthroscopy. Arthroscopy,

2014, 1520-1527.

PENNEKAMP C.W.A., BOTS M.L., KAPPELLE L.J., MOLL F.L., DE BORST G.J.: The

value of near-infrared spectroscopy during carotid endarterectomy in perioperative

stroke prevention. Eur J Vasc Endovasc Surg, 2009, 539-545.

PENNEKAMP C.W.A., IMMINK R.V., DEN RUIJTER H.M., KAPPELLE I.J., FERRIER

C.M., BOTS M.L., BUHRE W.F., MOLL F.L., DE BORST G.J.: Near-infrared

spectroscopy can predict the onset of cerebral hyperperfusion syndrome after carotid

endarterectomy. Cerebrovasc Dis, 2012, 314-321.

PLACKY J., HOFER S., VOLKMANN M., MARTIN E., BARDENHEUER H.J.,

WEIGAND M.A.: Regional cerebral oxygen saturation is a sensitive marker of cerebral

hypoperfusion during orthotopic liver transplantation. Anesth Analg, 2004, 344-349.

POHL A, CULLEN D.J.: Cerebral ischemia during shoulder surgery in the upright

position: a case series. J Clin Anesth, 2008, 463-469.

25

POTTER J.M., PIDGEON C., CUNNINGHAM A.J.: The sitting position in

neurosurgery: a critical appraisal. Br J Anesth, 1999, 117-128.

RUNDSHAGEN I.: Postoperative cognitive dysfunction. Dtsch Artzebl, 2014, 119-125.

SALAZAR D., SEARS B.W., AGHDASI B., ONLY A., FRANCOIS A., TONINO P:

Cerebral desaturation events during shoulder arthroscopy in the beach chair position:

patient risk factors and neurocognitive effects. J Shoulder Elbow Surg, 2013, 1228-

1235.

SALAZAR D., SEARS B.W., ANDRE J., TONINO P, MARRA G.: Cerebral desaturation

during shoulder arthroscopy: a prospective observational study. Clin Orthop Relat Res,

2013, 4027-4034.

SANDERS R.D., DEGOS V., YOUNG W.L.: Cerebral perfusion under pressure: is the

autoregulatory ‘plateau’ a level playing field for all? Anaesthesia, 2011, 965-976.

SKYHAR M.J., Altchek D.W., WARREN R.F., WICKIEWICZ T.L., O’BRIEN S.J.

Shoulder arthroscopy with the patient in the beach-chair position. Arthroscopy, 1988,

256-259.

SMITH J.J., PORTH C.M., ERICKSON M.; Hemodynamic response to the upright

position. J Clin Pharmacol, 1998, 375-386.

TANG L., KAZAN R., TADDEI R., ZAOUTER C., CYR S., HEMMERLING T.M.:

Reduced cerebral oxygen saturation during thoracic surgery predicts early

postoperative cognitive dysfunction. Br J Anaesth, 2012, 623-629.

TANGE K., KINOSHITA H., MINONISHI T., HATAKEYAMA N., MATSUDA N.,

YAMAZAKI M.: Cerebral oxygenation in the beach chair position before and after

general anesthesia. Minerva Anesthesiol, 2010, 485-490.

TERRANDO N., ERIKSSONS L.I., RYU J.K., YANG T., MONACO C., FELDMANN M.,

JONSSON FAGERLUND M., CHARO I.F., AKASSOGLOU K., MAZE M.: Resolving

postoperative neuroinflammation and cognitive decline. Ann Neurol, 2011, 986-995.

TOMOTKO G, MAEKAWA K.: Cerebral dysfunction after coronary bypass surgery. J

Anesth, 2014, 242-248.

26

TSAI T.L., SANDS L.P., LEUNG J.M.: An update on postoperative cognitive

dysfunction. Adv Anesth, 2010, 269-284.

WITLOX J., EURELINGS L.S.M., DE JONGHE J.F.M., KALISVAART K.J.,

EIKELENBOOM P., VAN GOOL W.A.: Delirium in elderly patients and the risk of

postdischarge mortality, institutionalization and dementia. JAMA, 2004, 443-451.

XIE Z., MCAULIFFE S., SWAIN C.A., WARD S.A., CROSBY C.A., ZHENG H.,

SHERMAN J., DONG Y., ZAHGN Y., SUNDER N., BURKE D., WASHICOSKY K.J.,

TANZI R.E., MARCANTONIO E.R.: Cerebrospinal fluid aβ to tau ratio and

postoperative cognitive change. Ann Surg, 2013, 364-369.

YADEAU J.T., LIU S. S., BANG H., SHAW P.M., WILFRED S.E., SHETTY T.: Cerebral

oximetry desaturation during shoulder surgery performed in a sitting position under

regional anesthesia. Can J Anaesth, 2011, 9574-9577.

ZHENG F., SHEINBERG R., YEE MS., ONO M., ZHENG Y., HOGUE C.W.: Cerebral

near-infrared spectroscopy monitoring and neurologic outcomes in adult cardiac

surgery patients: a systematic review. Anest Analg, 2013, 663-676.

27

Nederlandse samenvatting

Achtergrond/doel

Majeure schouderchirurgie in strandstoelhouding is geassocieerd met ernstige

neurologische gevolgen, zelfs in een gezonde patiëntenpopulatie van middelbare

leeftijd. De strandstoelhouding zelf wordt vaak aangehaald, aangezien deze de

cerebrale oxygenatie in het gedrang kan brengen. Deze studie heeft als doel de

associatie aan te tonen tussen peroperatieve episodes van cerebrale desaturatie en

postoperatieve cognitieve dysfunctie.

Materialen en methoden

Na toestemming van het Ethisch Comité en het afnemen van geïnformeerde

toestemming, werden 11 volwassen patiënten (4 mannen, 7 vrouwen met gemiddelde

leeftijd van 65 jaar met een SD van12.4 jaar) zonder voorafgaande neurologische of

psychiatrische pathologie op prospectieve wijze geïncludeerd. Zowel preoperatief, als

op de derde dag en drie maanden postoperatief werd de neurocognitieve functie

beoordeeld. Geheugen en aandacht werden geëvalueerd door middel van de ‘Auditory

Verbal Learning Test (AVLT)’, de ‘Stroop Color Word Test (SCWT)’ en de ‘Symbol

Digit Modalities Test (SDMT)’. Nabij-infrarood spectroscopie (NIRS) werd gebruikt om

peroperatieve episodes van cerebrale desaturatie op te sporen, gedefinieerd als een

daling van de basiswaarde met meer dan 20%. Standaard anesthetische zorg werd

verleend. Een herhaalde meting multivariaat analyse van variantie met tijd als ‘within-

subject’ variabele werd uitgevoerd.

Resultaten

Cerebrale desaturatie kwam voor bij alle patiënten tijdens de chirurgische

procedure. Er werden geen significante veranderingen in cognitief vermogen

gevonden tijdens de drie evaluatiemomenten. Eveneens was er geen significante

interactie tussen tijd en leeftijd met cerebrale desaturatie. Een significante verbetering

in test score voor de AVLT test postoperatief werd opgemerkt. Dit is meer dan

waarschijnlijk het gevolg van een belangrijk test-hertest effect.

28

Conclusie

Deze voorlopige resultaten tonen geen associatie tussen episodes van

cerebrale desaturatie en postoperatieve cognitieve dysfunctie. De huidige

patiëntenpopulatie is echter nog te klein om tot definitieve conclusies te komen.

Bevestiging van de bekomen resultaten in een grotere populatie is noodzakelijk.