PDF
Citation
Print
Introduction
Moyamoya disease (MMD) is a rare cerebrovascular disease characterized by progressive stenosis or occlusion of the internal carotid artery (ICA) and its main branches as well as the formation of collateral vessels at the skull base. Pediatric patients with MMD usually experience transient ischemic attacks (TIAs) or cerebral infarctions. Revascularization surgery has been established as an effective treatment to prevent ischemic events in MMD [1]. Careful management of patients with MMD is crucial for preventing perioperative ischemic complications. This includes maintaining adequate cerebral perfusion, ensuring normocapnia, and providing sufficient analgesia, all of which are essential for optimizing patient outcomes [2].
Delirium after general anesthesia is a common postoperative problem in children. It is defined as “a disturbance in a child’s awareness or attention to the environment with disorientation and perceptual alterations including hypersensitivity to stimuli and hyperactive motor behavior in the immediate post anesthesia period” [3]. Although various anesthetic-, patient-, surgical-, and medication-related factors have been implicated as risk factors for postoperative delirium [4,5], our understanding of these relationships remains incomplete, particularly in patients with MMD as information is notably limited.
Delirium during emergence from anesthesia, including crying, irritability, or acting out, may increase the risk of hyperventilation, increased intracranial pressure, and subsequent cerebral ischemia in children with MMD [6–8]. However, pharmacological interventions with opioids or sedatives can easily result in excessive sedation and hypercapnia, thereby delaying the evaluation of neurological function. The primary concern is whether this dissociated state of consciousness progresses into more serious neurological events or even cerebral infarction [9,10], as postoperative delirium may suggest an imbalance between cerebral oxygen supply and demand during revascularization surgery. Considering the potential adverse consequences of delirium in patients with MMD, further studies are needed to explore the incidence and risk factors of postoperative delirium in this population.
In the present study, we retrospectively enrolled pediatric patients with MMD who had undergone revascularization. We aimed to investigate the incidence of delirium in the post-anesthesia care unit (PACU) and identify the risk factors associated with PACU delirium. The secondary aim of this study was to evaluate the effects of PACU delirium on postoperative complications.
Materials and Methods
Study population
This manuscript was prepared in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement. Following approval from the Research Ethics Board of the Children’s Hospital of Fudan University (Approval number: 2023-45), patients with MMD aged < 15 years who underwent elective indirect revascularization between January 2014 and October 2023 at our center were identified. The requirement for written informed consent was waived due to the retrospective study design. For inclusion, patients had to have undergone a preoperative diagnostic cerebral magnetic resonance angiography (MRA) or digital subtraction angiography (DSA) to confirm moyamoya pathology. We excluded the following patients: (1) those with no MRA or DSA data and (2) those diagnosed with quasi-moyamoya disease (atherosclerosis, autoimmune disease, meningitis, brain tumors, Down syndrome, head injury, tuberous sclerosis, sickle cell disease, polycystic kidney, and others) according to the moyamoya diagnostic criteria [11].
Anesthesia management
Considering the risk factors for postoperative ischemic complications in patients with MMD, we established a special anesthesia protocol for our patients. Anesthesia management was performed by two experienced anesthesiologists who were mainly engaged in neuroanesthesia. Anesthesia was induced with propofol (3 mg/kg), rocuronium (0.6 mg/kg), and fentanyl (1 μg/kg). After tracheal intubation, pressure-controlled ventilation was provided to each patient. The parameters were adjusted to achieve an end-tidal carbon dioxide partial pressure (EtCO2) of 38–40 mmHg. Anesthesia was subsequently maintained using either total intravenous anesthesia (TIVA) (propofol and remifentanil) or inhalation anesthesia (sevoflurane or desflurane), at the discretion of the anesthesiologist. During inhalation anesthesia, fentanyl was added based on the hemodynamic responses. For TIVA, remifentanil was continuously infused throughout the procedure, without additional fentanyl. A scalp block was performed using ropivacaine before the skin incision.
The baseline blood pressure was measured before anesthesia induction. A noninvasive blood pressure (NIBP) cuff that matched the limb circumference was placed on the upper arm and the mean of three consecutive NIBP readings was recorded as baseline. Invasive arterial blood pressure monitoring was performed after radial artery cannulation. Our goal was to maintain the patient’s mean arterial pressure (MAP) at the baseline value. The intravascular volume was maintained using a balanced salt crystalloid solution. Pulse pressure variation was used to predict fluid responsiveness and was generally maintained below 10%. Hypotension was treated as needed with norepinephrine infusions. Intraoperative blood loss was estimated by the surgical team in conjunction with the anesthesiologist. Packed red blood cells were transfused as necessary to maintain a hematocrit greater than 30%. Body temperature was monitored with an esophageal probe and maintained at 36.5°C to 37.2°C using a Bair Hugger.
PACU management
Patients were extubated in the operating room after surgery and transferred to the PACU. All patients were kept in the PACU until they attained an Aldrete score ≥ 9. Patients were assessed for signs of delirium using the Pediatric Anesthesia Emergence Delirium (PAED) scale [3] by a trained nurse every 10 min until discharged from the PACU (Table 1). At our institution, a PAED score > 12 points [12] meets the criteria for PACU delirium. In the case that delirium persisted despite consoling measures or physical restraint, pharmacological rescue treatments were administered. At the discretion of the PACU anesthesiologist, intravenous dexmedetomidine (0.5 μg/kg) or propofol (1 mg/kg) was administered as the pharmacological rescue treatment of choice. Postoperative pain was evaluated using the Children’s Hospital of Eastern Ontario Pain Scale (CHEOPS) every 10 min from the time the patient entered the PACU. Patients with a CHEOPS score > 6 were treated with a rescue bolus of hydromorphone.
Postoperative management
All patients were admitted to the intensive care unit (ICU) overnight. Blood pressure, urine output, hematocrit, and oxygen saturation were closely monitored. If patients developed any neurological changes, they were immediately evaluated by an attending neurologist and underwent a CT scan.
Data review
Data were obtained from a retrospective review of the electronic medical records at our institution. Patients’ records were reviewed from neurosurgical consultation until hospital discharge. Patient demographics, coexisting diseases, preoperative imaging findings, laboratory data, intraoperative factors, and postoperative recovery were recorded. We focused on information pertaining to PACU delirium. Delirium was defined as a PAED score > 12 at any time point. All images were independently assessed by one experienced neuroradiologist and one experienced neurosurgeon who were blinded to the basic characteristics of the patients with MMD and discrepancies between the two readers were resolved by consensus.
Houkin’s MRA score, which significantly correlates with conventional angiographic staging [13], was used to stage vasculopathy in our patients with MMD. MRA scores were assigned based on the severity of occlusive changes in the ICA, middle cerebral artery, anterior cerebral artery, and posterior cerebral artery and the signals of the distal branches of these arteries (Supplementary Table 1).
The pediatric moyamoya MRI score (PMMS) was used to score the parenchymal findings on brain MRI [14]. Radiological variables chosen for the PMMS were based on the ability to easily and rapidly identify parenchymal findings on brain MRI and the correlation between these variables and outcomes in cerebrovascular diseases. Images were scored from 0 to 5 and included ischemic changes (0–2; 0 = none, 1 = focal, 2 = diffuse), encephalomalacia (0–2; 0 = none, 1 = focal, 2 = diffuse), and hemorrhage (0–1; 0 = not present, 1 = present). Diffuse findings were defined as either bilateral or spanning multiple vascular territories.
Intraoperative invasive arterial pressure was continuously monitored. The data were recorded on an electronic anesthesia sheet at an interval time of 5 min and stored in the electronic medical record system. The time-weighted average MAP (TWA-MAP) was calculated as the area under the curve of the MAP measurements divided by the total measurement time. As blood pressure variability during surgery has been associated with postoperative adverse events and mortality [15,16], the generalized average real variability of the MAP (ARV-MAP), which represents intraoperative MAP variability, was included in the analysis. This metric was calculated using the following algorithm, which was introduced in a previous study [17]:
where T is the total time from the first to the last MAP reading, N is the number of MAP readings, and MAPk represents each set of readings.
Statistical analysis
Categorical variables are expressed as numbers and percentages, and differences between groups were assessed using Pearson’s chi-square test or Fisher’s exact test. The Kolmogorov-Smirnova test was used to assess the distribution of continuous variables. Continuous data showing a normal distribution are presented as the mean and standard deviation, and differences between groups were assessed using the Student’s t-test. Continuous data with a non-normal distribution are expressed as the median and interquartile ranges, and differences between groups were assessed using the Mann-Whitney U test.
Logistic regression analyses were performed to evaluate the association between the variables and PACU delirium. Variables with P values < 0.1 in the univariate analysis were included in the multivariate model with a stepwise backward logistic regression to identify independent risk factors for PACU delirium. Odds ratios (ORs) and 95% CIs were calculated to assess the independent contributions of significant factors. The Hosmer-Lemeshow test was used to evaluate the goodness of fit of the model.
MMD in pediatric patients is typically a bilateral hemispheric disease that requires staged revascularization surgery in both hemispheres. Thus, in this study, each hemisphere was considered a separate operative procedure and PACU delirium was calculated accordingly. Due to the formation of collateral vessels after the reconstruction procedure, patients’ cerebral hemodynamics and subsequent symptoms can vary after each surgery. Therefore, the analysis was conducted in two steps: the first revascularization surgery and the second revascularization surgery. SPSS 26.0 software (SPSS Institute) was used for all statistical analyses. Statistical significance was set at P < 0.05.
Results
In total, 836 medical records of 598 patients were reviewed, 86 of which were excluded because of missing data or meeting exclusion criteria (Fig. 1). Consequently, 750 hemispheric procedures performed on 522 patients were included. PACU delirium occurred in 245 procedures (33% of 750), with a significantly higher incidence after the first revascularization than after the second (37% of 492 vs. 25% of 258; P = 0.002). Whether the surgery was performed in the left or right hemisphere had no significant effect on the occurrence of PACU delirium (34% vs. 31%; P = 0.455). Delirium was present during the PACU stay in 226 (30%), 113 (15%), and 16 (2%) cases at 30 min, 1 h, and 2 h, respectively.
First revascularization surgery
A total of 492 patients underwent their first revascularization, 180 (37%) of whom experienced PACU delirium. The patients’ baseline characteristics and perioperative data are summarized in Table 2. Variables with a P value < 0.1 in the initial univariate analysis that were thus included in the multivariate model were as follows: age, initial presentation, Houkin’s MRA score, PMMS, use of TIVA, intraoperative remifentanil dose, intraoperative fentanyl dose, baseline MBP, TWA-MAP, lowest MAP, and ARV-MAP (Table 3). Multivariate analysis identified that cerebral infarction as initial presentation (OR: 4.64, 95% CI [1.99–10.86]; P < 0.001), PMMS (OR: 2.75, 95% CI [2.10–3.61]; P < 0.001) and ARV-MAP (mmHg/min) (OR: 9.17, 95% CI [3.85–21.82]; P < 0.001) were risk factors for PACU delirium. Conversely, use of TIVA (OR: 0.46, 95% CI [0.28–0.76]; P = 0.003) was associated with a lower incidence of PACU delirium. The results of the Hosmer–Lemeshow test showed that the model fit the data adequately (P = 0.368).
Second revascularization surgery
The incidence of PACU delirium among the 258 cases of second revascularization was 25% (n = 65). Baseline characteristics and perioperative data are presented in Supplementary Table 2. The results of the univariate and multivariate analyses of potential risk factors and PACU delirium are shown in Supplementary Table 3. Independent risk factors included the PMMS and ARV-MAP. TIVA was associated with a lower incidence of PACU delirium. The results of the Hosmer-Lemeshow test showed that the model fit the data adequately (P = 0.499).
Effects of PACU delirium on outcomes
Compared with patients without PACU delirium, patients with PACU delirium had a longer PACU stay, a higher incidence of postoperative TIAs and stroke, and a higher incidence of unplanned ICU readmission (Table 4).
Discussion
In the present study, we found that PACU delirium occurred in 33% of the pediatric patients with MMD after indirect revascularization surgery. Compared to children undergoing the second revascularization surgery, patients who underwent the first surgery had a higher incidence of delirium during their PACU stay. A few risk factors associated with the occurrence of postoperative delirium were identified in our patients with MMD, including cerebral infarction as the initial presentation and a high PMMS, which suggested more preoperative cerebral ischemia and encephalomalacia as detected on MRI. We also found that intraoperative blood pressure variability was independently associated with PACU delirium. We suggest that the use of TIVA may exert protective effects against PACU delirium. In addition, it appears that patients who experience PACU delirium are likely to have more adverse postoperative outcomes.
The lower incidence of PACU delirium in children with MMD undergoing the second revascularization surgery may be due to the formation of collateral vessels and improved cerebral blood supply after the first surgery, allowing the patients to better tolerate potential hemodynamic instability during the second surgery. The first revascularization surgery is also more likely to be performed in the hemisphere with symptoms or hypoperfusion. Moreover, cerebrovascular reactivity and TIA frequency can be improved even in the contralateral hemisphere after unilateral revascularization [18].
The present data indicate that the initial presentation of cerebral infarction and a high PMMS were independent risk factors for PACU delirium. Delirium can result from a wide range of pathophysiological processes in the brain. For example, attention is related to the bilateral frontal regions, the thalamus, and the pons, while cognition is related to the frontal and temporal regions. In previous studies, MRI findings of silent cerebral ischemia [19] and new ischemic lesions [20] were significantly correlated with postoperative delirium in patients who underwent cardiac surgery. We speculate that the patients in our study with severe preoperative cerebral ischemia, cerebral infarction, and encephalomalacia had insufficient compensation for collateral vessels, abnormal regional hemodynamics, cerebral oxygen supply deficits, and impairment of parenchymal functions in certain areas.
Our data showed that intraoperative blood pressure variability was another significant risk factor for PACU delirium. This is consistent with findings in critically ill patients [21–23]. Dysfunctional cerebral autoregulation has been used to explain this association [24]. Impaired cerebral autoregulation could predispose patients with high blood pressure variability to develop microvascular and blood–brain barrier damage from hypertension and cerebral ischemia from hypotension, as enlarged fluctuating loads are inadequately buffered. MMD is a special type of cerebrovascular disease accompanied by dysfunctional cerebral autoregulation. Notable differences in the severity of the deficits of cerebral blood flow (CBF) autoregulation have been seen in adult and juvenile MMD, with more severe deficits in the cerebrovascular response among juvenile cases [25]. Autoregulation relies on the ability of cerebral vasculature to constrict and dilate. We speculate that the arteries of patients with MMD are in a state of maximal vasodilation to maintain CBF. Since blood pressure variability is partially modifiable, the possible causal relationship between blood pressure variability and PACU delirium deserves further exploration.
According to existing research, either no significant difference [26] or a lower incidence [27,28] of emergence delirium (ED) is associated with TIVA compared to volatile anesthesia, which is explained by some inherent properties of volatile anesthetics. In addition, propofol could produce a positive mood or euphoric state postoperatively [29,30]. In our study population, the use of TIVA was associated with a lower incidence of PACU delirium, which may partly be due to the inherent pathophysiological characteristics of MMD. One possible explanation is that volatile anesthetics have a cerebral vasodilatory action and can provoke intracerebral steal from the ischemic area, similar to the effect of hypercapnia [31]. Both cortical blood flow and frontal regional oxygen saturation declined under inhaled anesthesia in patients with MMD [32]. In contrast, intravenous anesthesia with propofol appeared to have the potential to protect the brain and preserve regional CBF in the frontal lobes [33]. However, confounding factors may be present owing to differences in intraoperative opioid use. Remifentanil was used exclusively for TIVA, whereas intermittent fentanyl was administered via inhalation. The results indicate a significant difference in the intraoperative remifentanil and fentanyl doses between patients with and without PACU delirium. Therefore, the observed association between TIVA and a reduced risk of PACU delirium may be due to differences in intraoperative opioid administration, specifically between remifentanil and fentanyl. To address this issue, we developed a multivariate regression model that included TIVA, the remifentanil dose, and the fentanyl dose as covariates. TIVA remained a significant independent factor for PACU delirium, whereas intraoperative opioids (fentanyl or remifentanil) were not included in the multivariate regression equation. However, because remifentanil was exclusively used for TIVA, its impact on the outcomes might be masked by collinearity. Further analysis showed that the variance inflation factor (VIF) of TIVA was 4.882 and the VIF of the intraoperative remifentanil dose was 4.337 for the first revascularization surgery, while for the second revascularization surgery, the VIF of TIVA was 2.095 and the VIF of the remifentanil dose was 1.729. For all the other covariates, the VIF was < 1.5. Consequently, this issue was difficult to evaluate thoroughly, which is one limitation of this study. To properly assess the impact of the type of anesthesia on PACU delirium, a comparison between propofol plus remifentanil and inhalation anesthesia plus remifentanil would be valuable.
Another notable feature is the hemodynamic difference between TIVA and inhalation anesthesia. The ARV-MAP in patients using TIVA (1.48 [1.28, 1.68] mmHg/min) was significantly lower than that in patients using volatile anesthetics (1.61 [1.46, 1.81] mmHg/min; P < 0.001), although no significant difference was found in the baseline MBP, TWA-MAP, or lowest MAP (data not shown). As fentanyl was titrated in response to hemodynamic changes, it may have been administered when hemodynamic fluctuations were observed, resulting in a higher ARV-MAP. In contrast, the remifentanil titration was likely more consistent in the TIVA group. This suggests that differences in intraoperative opioid use may be a significant confounding factor Despite these concerns, both the ARV-MAP and TIVA remained significant independent risk factors for PACU delirium. This indicates that even with opioid use as a potential confounding factor, the ARV-MAP independently affects the occurrence of postoperative delirium. Further well-designed prospective, randomized, controlled trials are required to clarify this association.
Postoperative pain is often considered a significant contributing factor to postoperative delirium [34–36]. Delirium can also occur in children without pain [37]. In our cohort, the pain scores at both PACU admission and discharge were clinically comparable between the patients with and without PACU delirium. Furthermore, no differences in the cumulative hydromorphone dose were found between patients with or without delirium at PACU discharge. All of the above findings suggest that pain was an unlikely cause of delirium in our patients.
Our study has several limitations. First, this was a retrospective analysis of patient data from a single center, which may introduce potential bias, information loss, and the inability to establish causality. The c nature of this study may have limited the generalizability of our findings. The second limitation was selection bias. In the current study, the decision between TIVA and inhalation anesthesia was made by the anesthesiologist, which may have been a source of selection bias. Although no statistically significant differences in age, sex, Houkin’s MRA score, or PMMS were found between the TIVA and inhalation anesthesia groups, the potential for imbalances in other clinical characteristics cannot be ruled out. The impact of any potential imbalance may be a critical factor in the occurrence of PACU delirium. A randomized controlled trial is required to clarify this issue. The third limitation concerns the scale used to assess delirium. The PAED scale has been widely used as a valid rating tool to quantify ED in children. However, a common challenge in using this scale is distinguishing delirium from agitation caused by pain. In a previous study, Somaini et al. [38] emphasized that delirium and pain are two distinct entities that cannot be readily discriminated using the PAED score. Regardless, isolating delirium from pain is clinically important and could improve future treatments of PACU delirium. The items ‘inconsolability’ and restlessness’ have both been associated with delirium but may also reflect pain. ‘No eye contact,’ ‘no purposeful action,’ and ‘no awareness of surroundings’ may be used to identify ED in young children. Fourth, although patients with delirium experienced more adverse postoperative outcomes in our study, this may have been associated with the risk factors for PACU delirium rather than delirium itself. Adequate control of confounding factors is required to determine the true contribution of PACU delirium to adverse postoperative outcomes.
In conclusion, we found that a significant proportion of patients with MMD developed delirium in the PACU after revascularization surgery. This retrospective study suggests that both high intraoperative blood pressure variability and preoperative MRI lesions significantly affect PACU delirium in pediatric patients with MMD. One significant finding was that TIVA was associated with a lower incidence of postoperative delirium. However, this observed association does not directly imply causation and well-conducted, large-scale, randomized, controlled prospective studies are required to confirm these associations.
XML Download