Postoperative cognitive functions in patients with benign intracranial lesions

The aim of this study was to assess pre- and postoperative cognitive functions in patients who underwent surgery for benign intracranial lesions. In total, 58 patients (21 men, 37 women, mean age 51.6 years [range 24–76 years]) with benign intracranial lesions (including benign tumors and vascular lesions) and neuralgia of the trigeminal nerve were included in this prospective study. Extensive cognitive testing was used to categorize attention, memory, and executive functions. Mood and pain were assessed preoperatively (t0, mean 3.7 days before surgery), immediately after surgery/during inpatient stay (t1, mean 7.6 days after surgery), and at first outpatient check-up (t2, mean 99.5 days after surgery). All 58 patients were tested at t0 and t1, but at t2 only 24 patients were available at t2. The data were categorized as improvement/stable condition or deterioration and shown as percentages. The pre- and postoperative values of BDI-II and mood were compared by the Wilcoxon test for paired samples. Binary logistic regression analyses were performed to identify parameters influencing cognition in the subgroup of meningioma patients. Immediately after surgery (t1), the percentage of patients with improvement/stable condition was > 50% in all categories in the majority of subtests (attention: 12/14 subtests, memory: 11/13 subtests, executive functions: 6/9 subtests). Similar results were shown at t2. Mood and pain did not change significantly after surgery. Factors like age, Karnofsky performance status, and tumor volume were not shown as significant influencing factors for cognitive functions in meningioma patients. The results of this study suggest that—in contrast to neuroepithelial tumors—cognitive functions do not deteriorate after surgery of benign intracranial lesions. Further studies are necessary to evaluate the results of this study.

Trail-Making-Test A. The Trail-Making-Test A (TMT-A) task was to connect numbers  in the right order as fast as possible. Time is measured and shown in seconds 31 .
Memory. Wechsler Memory Scale (WMS). The Wechsler Memory Scale Revised 32 is a test battery for assessing different aspects of memory and includes 13 subtests. In this study, the block-and digit-span subtests were used to evaluate verbal and nonverbal short-term memory. A row of digits was read out, or a group of wooden blocks was arranged in a specific order. Patients were asked to repeat the tasks immediately (memory span verbal/nonverbal [ms v/ms nv]) and after a delay (working memory verbal/nonverbal [wm v/wm nv]).
Verbal learning and memory test. The Verbal Learning and Memory Test (VLMT) was used to analyze episodic memory. A learning list of 15 words was read out 5 times, and patients were asked to repeat the words after the words were read out once (Dg1) and after being read out 5 times (Dg5). The gained knowledge was measured (Dg1-5). An interference list with different words was read out, and the patients again were asked to repeat the words from the learning list immediately (Dg6) and after a delay of 30 min (Dg7). The loss of knowledge was measured after the interference list was read (Dg5-6) and after a delay (Dg5-7).
Rey Osterrieth complex figure test (ROCF). The well-known neuropsychological Rey Osterrieth Complex Figure Test (ROCF) was used to analyze the patients' visual memory and visual constructive capacity 33 and included 3 subtests. Initially, the subjects were asked to copy a geometrical figure that was shown to them. This first subtest was not assessed in the present study. In the next 2, subtests, which were analyzed in this study, the subjects were asked to draw a geometrical figure shown to them before, both immediately (ROCF copy) and after a 30-min delay (ROCF delay).
Pain, especially headache, in this cohort was assessed using the IBK, the German version of the Headache Disability Inventory (HDI) 38 . This test was available for 43/58 patients at t 0 , 33/58 patients at t 1 , and 18/24 patients at t 2 . Pain was divided into 4 scales: no headache, slight headache, moderate headache, and severe headache.

Volumetric measurement.
A neuroradiologist performed manual segmentation of the contrast-enhancing part of the intracranial lesion pre-and postoperatively (iPlan Net Cranial 3.0, Brainlab AG, Munich, Germany). No volumetric measurement was performed of vascular lesions or trigeminal neuralgia.

Surgery.
Surgery was performed at the Department of Neurosurgery with the aim of maximum tumor resection in patients with benign tumors. Pituitary adenomas were resected using a transnasal-transsphenoidal approach, and other benign tumors were resected using trepanation. Aneurysms were treated by clipping and pterional trepanation, whereas trigeminal nerve neuralgia was treated by microvascular decompression. Statistical analysis. IBM SPSS Statistics versions 24.0, 25.0, and 26.0 (SPSS Inc., IBM Corp., Armonk, NY, USA) was used for the statistical analysis. Normally distributed data were shown as means/standard deviations, and non-normally distributed data were shown as medians/interquartile range (IQR). The delta between the preand postoperative percentile ranks was recorded, and the patients were divided into improvement/stable condition or deterioration groups. The Wilcoxon test for paired samples was used for pre-and postoperative comparisons of mood and pain and of the basis test battery (MMSE). Binary logistic regression analyses were performed to identify risk factors for postoperative changes of cognitive functions. P < 0.05 was defined as significant.
Ethical approval and informed consent. The study was conducted in accordance with the ethical standards of the 1964 Declaration of Helsinki and its later amendments and approved by the local ethics committee (Ethics committee technical university munich). Informed consent was signed by all study participants.

Results
Patient population. Initially, 81 patients were included in the study. Of them, 23 patients were excluded: 22 patients did not perform neuropsychological testing after surgery due to reduced general condition or lack of retrieved informed consent (for further neuropsychological testing after surgery), and 1 patient was excluded due to missing surgery (neuroradiological intervention).
The main tumor locations were the frontal lobe (23/58) and infratentorial region (16/58). The baseline patient characteristics also included information about initial symptoms and adjuvant treatment, as shown in Table 1. www.nature.com/scientificreports/ surgery to postoperative testing during inpatient stay (t 1 ) was 7.6 days (range 2-55 days), and the mean time from surgery to follow-up at the first outpatient control (t 2 ) testing was 99.5 days (range 61-197 days). Pre-and postoperative comparisons. Analyses were performed for all patients and the patients in the meningioma and pituitary adenoma subgroups ( Table 2). Improvement of cognition was defined as stable/improving cognitive functions in more than 50% of the patients. In the attention category, 12/14 subtests showed early postoperative improvement/stable condition at t 1 , as compared to 11/13 subtests in the memory category and 6/9 subtests in the executive functions category. Similar results were observed in the subgroup of meningioma patients ( Table 2). The percentage of patients with improvement during follow-up (t 2 ) was similar, comprising 11/14 subtests in the attention category and, even higher, 13/13 subtests in the memory category and 9/9 subtests in the executive functions category. Among the meningioma patients, this rate at t 2 was lower in the memory category (8/13 subtests). In the subgroup of patients with pituitary adenomas, 12/14 subtests improvement/stable condition immediately postoperatively (t 1 ) in the attention category, as compared to 7/13 subtests in the memory category and 6/9 subtests in the executive functions category. At t 2 , 12/14 subtests in the attention  www.nature.com/scientificreports/ category showed improvement/stable condition, compared to 12/13 subtests in the memory category and 7/9 subtests in the executive functions category (Table 2). Figure 1 shows the distributions of improvement and deterioration in the subtests of the attention, memory, and executive functions categories at t 1 and t 2 . Figure 2 presents the results of the 8 meningioma patients at different time points in each category (attention, memory, and executive functions).

Mood and pain.
Mood did not change significantly after surgery at t 1 (P = 0.484) and at t 2 (P = 0.306). In addition, pain did not change significantly at t 1 (P = 0.060) or at t 2 (P = 0.564). The distributions of depression and pain at the different time points are shown in Fig. 3. Table 2. Patients with improvement of cognition and/or stable cognition after surgery. Data are shown as numbers of patients with improvement/stable condition of cognitive functions in relation to all patients performing this subtest immediately after surgery and during follow-up. Improvement > 50% shown in italics.  Table 3. No significant influencing factors could be identified in this analysis.   The preoperative neurocognitive functions of patients with benign intracranial lesions were analyzed in a previous study, which showed that age and KPS were the main risk factors for impaired neurocognitive functions before operation 28 . In this study, we analyzed a subgroup of patients with available, extensive postoperative neurocognitive testing.
This study showed improvement or stable condition of cognitive functions in the attention, memory, and executive functions categories. These results agree with those of Tucha et al., who studied (elderly) meningioma patients, showing postoperative improvement mainly in the attention and memory domains and no deterioration www.nature.com/scientificreports/ of preoperative cognitive functions 15,17,39 . Another recent study be Meskal et al. showed postoperative improvement in almost all cognitive domains, except for psychomotor speed and reaction time 16 . The mentioned studies by Tucha et al. reported no improvement in executive functions after meningioma surgery 17,39 . A recent previous study on meningioma patients also reported postoperative improvement of cognitive functions but with lower ongoing cognitive scores as compared to healthy controls 40 .
The rate of improvement/stable condition was even higher during follow-up (t 2 ), as compared to immediate postoperative testing (t 1 ), in this cohort. These results might be explainable by postoperative edema or reduced postoperative functional independence (KPS), a known risk factor for cognitive impairment 7 . Previous studies also showed a transient decline of cognition with recovery at follow-up after surgery and after irradiation 15,19,41 . The time point when postoperative evaluation occurs might be important for these findings. In this study, the first postoperative testing was performed quite early (mean time from surgery to t 1 of 7.6 days). Thus, the aftereffects of surgery might be more prominent at this date. In the other mentioned studies on meningioma patients, postoperative testing was performed later (about 3 months after surgery), which agrees with our follow-up examination (t 2 ) 15,17,39 .
However, the results of the follow-up analyses should be taken with caution. According to the study design (t 1 = postoperative testing during inpatient stay; t 2 = first follow-up testing on outpatient controls), there were major differences in the time periods, with some overlapping results. Furthermore, only patients with a MMSE > 18 underwent extensive neuropsychological testing, thus only selecting patients without major deterioration after surgery.
The highest rate of improvement/stable condition was observed in the ROCF delay subtest, at almost 100%. This improvement might have been because the patients remembered that they had to perform the same task in 30 min. Therefore, these results should be considered with caution.
In contrast to other studies focusing mainly on 1 entity (eg, meningioma or pituitary adenoma), this study included different entities and also vascular lesions. This might introduce a bias due to the high majority of the diseases. However, considering a more heterogeneous group could also provide additional information, as compared to focusing on only 1 entity.
Among the meningioma patient subgroup, tumor size was not observed as a significant influencing factor, contrasting a previous study by Liouta et al. 14 These differences might be explained by the different study designs. The present study focused on patients with benign intracranial lesions and with therefore had a lower number of meningioma patients, whereas Liouta et al. focused only on meningioma patients and therefore had a larger number of patients.
The surgical approach and tumor location might affect cognitive functions. Due to the low number of patients in the present study cohort, no further analyses were conducted regarding these possibilities. Surgery was performed according to the neurosurgical standards at our department and did not significantly differ between the patient subgroups.
This study has several limitations. The high dropout rate during follow-up (at t 2 ) was a main limitation and might have introduced unavoidable bias. Patients with less cognitive deficits might be more likely to perform cognitive testing during follow-up than patients with considerable restrictions would. This might have biased the results and resulted in overestimation of the rate of patients without deterioration after surgery. Further studies with a lower dropout rate are necessary to address this.
Additional limitations include the low numbers of patients with some diseases and the variety of diseases included in this cohort. However, subependymoma, clivus chordoma, arterial-venous malformation, and cavernoma are rare intracranial lesions, and to our knowledge, cognitive functions have not been assessed before in patients with these lesions. To address this, a subgroup analysis was performed with meningioma patients only.
In particular, the number of unruptured aneurysms was very low in this cohort (n = 4), and the study does not add any new findings to the already known results of the ISAT trial, which showed cognitive improvement after endovascular treatment of aneurysms 24 .
The study population remains a very heterogeneous population with small numbers of each individual pathology. The significance of this study might be limited due to this heterogeneous cohort and many other possible confounding variables that would affect cognitive outcomes. However, the aim of this prospective study was to include all types of benign intracranial lesions and not to select special subgroups (eg, meningioma patients) as previous studies on such cohorts already exist. This exploratory study might draw attention to this heterogeneous patient cohort and might be of interest for further (prospective) studies.

Conclusions
Cognitive functions improved or remained stable in the attention, memory, and executive functions categories after surgery of benign intracranial lesions in the majority of our cohort of 58 patients. Due to the high dropout rate and the various intracranial lesions included in this study, the results of this study should be taken with caution, and further studies are necessary to confirm the results. www.nature.com/scientificreports/