Serum Glial Fibrillary Acidic Protein (GFAP) as a Potential Biomarker for Monitoring Postoperative Complications in Deep Brain Stimulation Surgery

doi:10.21203/rs.3.rs-4968505/v1

Download PDF

Article

Serum Glial Fibrillary Acidic Protein (GFAP) as a Potential Biomarker for Monitoring Postoperative Complications in Deep Brain Stimulation Surgery

https://doi.org/10.21203/rs.3.rs-4968505/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

Deep brain stimulation (DBS) is an efficient treatment for movement disorders, most commonly Parkinson’s Disease (PD), dystonia and essential tremor. DBS surgery carries risks, e.g. the risk of delayed peri-lead edema (PLE) and the risk of postoperative cognitive decline. The mechanisms of these complications are not fully understood and there is no established biomarker to screen for these complications after DBS surgery. To assess whether serum glial fibrillary acid protein (sGFAP) might constitute a potentially useful biomarker to detect complications after DBS surgery, we measured sGFAP and serum neurofilament light chain (sNfl) in 58 patients undergoing DBS at our center. Both serum markers increased transiently after surgery. Serum GFAP returned to baseline after weeks, whereas sNfl only returned to baseline after months. Patients with poorer preoperative cognitive performance had higher postoperative sGFAP values, and the relationship of sGFAP with preoperative patient characteristics was closer than for sNfl. These properties suggest that sGFAP can be a valuable biomarker to monitor patients for delayed complications after DBS surgery.

Health sciences/Neurology/Neurological disorders/Movement disorders

Health sciences/Neurology/Neurological disorders/Parkinsons disease

Parkinson’s Disease

dystonia

tremor

DBS

sNfl

sGFAP

Deep brain stimulation (DBS) of target nuclei such as the globus pallidus pars interna (GPi), nucleus subthalamicus, (STN) and thalamus (ventral intermediate, VIM) is an established and efficient therapy for movement disorders. Most commonly, DBS is considered in essential tremor (ET), dystonia and Parkinson’s disease (PD).^1–4 In general, DBS is safe and effective also in older patients and is still considered in patients with cognitive impairment.^5,6

DBS surgery carries risks, and some of them are known to be influenced by preoperative patient characteristics. In people with Parkinson’s disease (PwPD), there is a known risk of postoperative cognitive decline, which increases with age.⁷ Preoperative cognitive impairment^8–10 or genetic risk factors¹¹ also confer an increased risk of subsequent cognitive decline after surgery. The causal mechanism for this is still unknown and thus, it is currently not possible to accurately predict postoperative cognitive decline preoperatively.

Improved imaging and surgical techniques have contributed to a lower risk of hemorrhage associated with DBS surgery, as trajectories can avoid blood vessels and fewer microelectrode tracks are necessary for accurate lead placement.¹² However, delayed-onset peri-lead edema (PLE) is still observed after DBS surgery with one study reporting symptomatic PLE in 7% of patients 6 weeks after DBS surgery.¹³ The pathophysiology of PLE is not fully understood, and it is commonly attributed to a foreign body reaction.¹⁴ Still, other mechanisms are conceivable, e.g. the implantation of DBS electrodes could be akin to processes described in traumatic brain injury (TBI), in which a role of the response of glial cells to the injury in addition to neuronal damage is assumed.¹⁵

In a previous study, we showed that neuronal injury as reported by serum neurofilament (sNfL) is caused by DBS surgery itself and not by chronic neurostimulation.¹⁶ The half-life of sNfL is long and increased levels of sNfl only return to baseline values months after surgery. Serum NfL is therefore not an ideal marker to assess perioperative damage in DBS surgery.

In this context, serum glial fibrillary acidic protein (sGFAP), a commonly used and reliable biomarker for glial injury, might offer a useful alternative. It is FDA-approved for clinical use in traumatic brain injury (TBI) due to its favorable relation to clinical outcomes and its kinetics after injury.¹⁷ Studies in Multiple Sclerosis (MS) suggest a better correlation of sGFAP values with factors of disease progression than of sNfL.^18,19 In addition, sGFAP could potentially be useful to unravel a role of glial injury in the pathogenesis of PLE in comparison to sNfl, which reports axonal damage.

To examine whether sGFAP might constitute a suitable marker of peri- and postoperative damage of DBS surgery, we measured sGFAP along with sNfl pre- and postoperatively in DBS patients receiving either GPi, VIM or STN-DBS. In addition, we assessed whether preoperative patient characteristics or biomarker measurements increase the risk of neuronal or glial injury as reported by sNfl and sGFAP.

Serum neurofilament after DBS surgery

A total of 58 patients undergoing DBS surgery in our center were prospectively enrolled in this study, comprising 47 patients with PD and 11 patients with “non-degenerative” diseases, i.e., dystonia or ET. Their demographic characteristics are displayed in table 1.

Table 1

Demographic and clinical characteristics

	PwPD (n=47)	Non-degenerative (n=11)
Age at DBS implantation, years	63.3 (8.8)	59.3 (15.9)
Sex, n male/female (% female)	31/16 (34%)	8/3 (27%)
Disease duration, years	9.5 (3.9)	27.9 (23.1)
Hoehn and Yahr stage, median (range)	2 (1-4)	n/a
UPDRS III OFF total score	39.6 (16.3)	n/a
MoCA total score	26.8 (2.5)	n/a
preoperative sNfl, pg/ml	18.8 (7.4)	17.9 (7.4)
preoperative sGFAP, pg/ml	139.8 (71.7)	153.5 (97.1)

Note. Data are “mean (SD)” unless otherwise indicated. UPDRS: Unified Parkinson's Disease Rating Scale; MoCA: Montreal Cognitive Assessment; sNfl: serum neurofilament light; sGFAP: serum glial fibrillary acidic protein.

To report neuronal injury during and after DBS surgery, sNfL was used as previously described.¹⁶ In patients with PD, sNfL increased significantly (Friedman’s test and Nemenyi-Friedman post-hoc test, p<0.001) between preoperative baseline and 3 – 5 days after surgery (figure 1). The mean preoperative baseline sNfL was 18.8 pg/ml (95% CI 16.24 – 21.36), sNfL increased to 40.92 pg/ml (33.071 – 48.78) at the first postoperative time point. At the second postoperative time point, mean sNfL was 79.88 pg/ml (68.72 – 91.043) and significantly higher (p<0.001) than preoperative baseline or the first and fourth postoperative time points. There was a tendency of sNfL to decline at the third postoperative time point (mean sNfl 75.64 pg/ml (62.67 – 88.61), p=0.755 for comparison with time point 2). The fourth postoperative measurement was obtained 3 – 6 months after surgery and the mean sNfL was 23.34 pg/ml (19.21 – 27.46), which was not significantly different from the preoperative baseline (p=0.619), thus indicating a return to baseline values. In patients with non-degenerative diseases, the results were equivalent. Due to the small sample size, the preoperative baseline vs. the first postoperative time point was not significantly different, the other results described above also apply to this part of our cohort.

A two-way ANOVA between the PD and the non-degenerative group was significant for the interaction between the time points after surgery and log(sNfL) measurements in both groups (p<0.001), i.e. it showed an effect of the surgery on log(sNfL) values. There was no significant effect of either the group itself or the interaction between the group and the time point on log(sNfL) measurements.

Taken together, the time course of sNfL after DBS surgery in patients with PD was similar as described previously,¹⁶ findings in patients with essential tremor or dystonia were similar to those observed with in PwPD.

Serum GFAP after DBS surgery

The slow dynamics of sNfL after surgery result mainly from its long half-life and not from continual neuronal damage – as discussed previously.¹⁶ To obtain more direct information about surgical trauma resulting from DBS implantation, we measured sGFAP. Since GFAP is expressed in astrocytes, it reports glial activation, which is another difference from sNfL. The time course of sGFAP was markedly different from sNfL in PD patients (figure 1).

In PD patients, only the first postoperative time point differed significantly from preoperative baseline (p=0.001) and subsequent time points (p for all pairwise comparisons ≤ 0.005). Mean sGFAP was 136.65 pg/ml (113.51 – 159.78) at preoperative baseline, 540.06 pg/ml (361.5 – 718.62) at the first postoperative time point, 166.56 pg/ml (134.27 – 198.85) at the second postoperative time point, 150.35 pg/ml (115.26 – 185.44) at the third postoperative time point and 145.43 pg/ml (111.37 – 179.49) at the fourth postoperative time point. Postoperative time points 2-4 were not significantly different from preoperative baseline. The time course for the non-degenerative cohort was similar in that the first postoperative measurement was the only one that differed significantly from baseline.

Again, a two-way ANOVA revealed no significant interaction between the group (i.e. PD and non-degenerative) and postoperative log(sGFAP) measurements and a significant effect of the time point after surgery and log(sGFAP) measurements (p<0.001).

Note that log(sNfL) and log(sGFAP) values for PwPD are displayed in figure 1 in relation to mean baseline values to enable comparability between serum marker kinetics. The results of post hoc tests comparing measurements between time points were identical when using log(sNfL) and log(sGFAP) instead of raw sNfl or raw sGFAP.

Baseline values of sNfL and sGFAP are affected by age and BMI, but not disease-specific factors in PwPD

The correlation between log(sNfl) and log(sGFAP) of the same patient at preoperative baseline was moderate in PwPD (Spearman’s rho, r=0.55, p=.002 in PwPD; r=0.45, ns in non-degenerative patients). This could result from a floor effect or from the fact that sNfL reports neuronal damage and sGFAP reports glial damage. This finding is consistent with the observation that the time course of the two parameters and the factors that influence them are different.^20–22

Baseline values of sNfl and sGFAP are influenced by factors that may increase their release into the blood, including age and the presence of neurodegenerative diseases, and by factors that may affect their distribution or elimination, like body mass index (BMI) or renal function.^19,23,24

Indeed, baseline log(sNfL) correlated negatively with BMI (figure 2a; PwPD: r=-0.36, p=0.012; non-degenerative patients: ns). Baseline log(sGFAP) showed the same trend but did not reach significance in PwPD (r=-0.34, p=0.061), but in non-degenerative patients (r=-0.7, p=0.016). In PwPD, baseline log(sNFfL) and log(sGFAP) values also correlated with age (figure 2b, log(sNfl) in PwPD: r=0.46, p=0.001, log(sNfl) in non-degenerative patients r=0.69, p=0.018; log(sGFAP) in PwPD r=0.4, p=0.022, log(sGFAP) in non-degenerative patients: ns).

There was a borderline significant correlation between MoCA and baseline log(sNfl) values (r=-0.32, p=0.043), as has been described previously.^25,26 There was no significant correlation between MoCA scores and baseline log(sGFAP). Baseline log(sGFAP) and log(sNfl) did not correlate significantly with further PD-specific assessments (UPDRS III total score, Hoehn and Yahr-stage, presence of dyskinesias or other disease-related complications, not shown). Also, baseline log(sGFAP) and log(sNfl) did not correlate significantly with comorbidities, such as diabetes, arterial hypertension, uremia and nicotine abuse (Spearman’s rho), independent of group (PwPD vs. non-degenerative patients, Mann-Whitney-U).

Perioperative damage as reported by sGFAP is influenced by age, motor symptoms, cognitive impairment and baseline sGFAP

We observed a strong correlation between values at baseline and the first postoperative time point in PwPD, both for log(sNfl) (log(sNfl) r=0.71, p<0.001) and for log(sGFAP) (r=0.72, p<0.001). These correlations were also observed in non-degenerative patients (only significant for log(sNfl) r=0.8, p=0.003, log(sGFAP): ns). Patients with higher values at baseline thus showed higher postoperative values of log(sNfl) and log(sGFAP).

To identify risk factors of increased perioperative damage, we aimed to predict the postoperative change in log(sGFAP) and log(sNfl) values in PwPD from baseline data by using a random forest regressor which predicted postoperative values at chance level. Using gradient boosting, however, a model incorporating age, UPDRS III in the OFF condition, the total MoCA score and the baseline log(sGFAP) value explained about 41% of the variance in postoperative log(sGFAP) values (r²=0.41). Of the factors in this model, the combined relative importance of preoperative MoCA and UPDRS III scores was more than 95%, with age and preoperative log(sGFAP) values having a combined importance <0.03. This could not be replicated for log(sNfl) values, indicating that GFAP could be better suited to report differences in perioperative damage between individual patients.

To explore these findings in more detail, we examined individual correlations with log(sNfl) and log(sGFAP) values. In contrast to the findings with gradient boosting, the degree of change of log(sNfl) and log(sGFAP) did not correlate significantly with age. MoCA total values were moderately negatively correlated to the degree of change in log(sGFAP) values with r=-0.61 (figure 2c, p=0.001), i.e., patients with cognitive impairment showed more perioperative damage as reported by log(sGFAP). GPi is often used as a DBS target in patients with cognitive impairment. To ensure that the correlation of log(sGFAP) and MoCA was not biased by this, we computed the correlation without PwPD that received GPi-DBS. The correlation remained moderate with r=-0.54 (p=0.008). There was a weak, but significant correlation of MoCA with the log(sNfl) change at the second postoperative timepoint (r=0.36, p=0.034), which was not significant when controlling for baseline log(Nfl) values. Correlations of baseline log(sNfl) and log(sGFAP) with the maximum change from baseline (second postoperative time point for log(sNfl) and first postoperative time point for log(sGFAP)) were also not significant.

The change in log(sGFAP) or log(sNfl) after surgery did not correlate significantly with further PD-specific assessments (UPDRS III total score, Hoehn and Yahr-stage, presence of dyskinesias or other disease-related complications, not shown), or with comorbidities (diabetes, arterial hypertension, uremia, nicotine abuse).

BMI was moderately positively correlated to the degree of change of log(sNfl) in PwPD (r=0.5, p=0.02), but not in non-degenerative patients. There was no significant correlation with the degree of change of log(sGFAP). This correlation was reversed, but not significant when controlling for preoperative log(sNfl), indicating that this correlation is driven by baseline values (partial correlation r=-0.59, p=0.07). The observation itself might be explained by an increased distribution volume in patients with a higher BMI.

Factors of the DBS surgery itself, i.e., the duration of the operation, the number of microelectrodes used or general anesthesia/awake surgery and the total energy delivered by the stimulation at different time points, did not show any significant association (Spearman’s rho) or significant differences (Mann-Whitney-U) between groups.

In this study, we performed longitudinal measurements of serum biomarkers of neuronal (sNfl) and glial (sGFAP) injury after DBS surgery in a cohort of patients with movement disorders, namely PD, dystonia and ET. To the best of our knowledge, this is the first study reporting prospective postoperative measurements of sGFAP in DBS patients. Both sNfl and sGFAP transiently increased after surgery, but sGFAP showed much faster kinetics in comparison to sNfl. We found no relevant differences in plasma marker measurements between PD as a progressive neurodegenerative disorder and non-degenerative disorders, i.e., dystonia and ET.

Baseline values of sNfl were higher in older participants than in younger participants (figure 2b), consistent with previous findings by others.²⁷ We also observed a negative association between sNfl and BMI (figure 2a), which has been explained by the increased distribution volume of NFL released from the nervous system.²⁸ Except for a non-significant correlation between BMI and sGFAP in PwPD, these observations also apply to sGFAP (figure 2a and b), as has been shown in MS.^19,23

Perioperative change in sGFAP was not associated with the number of microelectrode tracks. This finding is consistent with the hypothesis that the perioperative increase results from a foreign body reaction¹⁴ and not consistent with the hypothesis that it results from intraoperative trauma, which would be more severe with a higher number of microelectrode tracts. It should be noted, however, that the variance of microelectrode tracks was small. The perioperative change in sGFAP was associated with baseline sGFAP, suggesting that it is regulated, in part, by the same factors, which could include elimination mechanisms as shown for sNfl.²⁹

Even within the narrow range of MoCA scores covered in our cohort, a lower preoperative MoCA score was associated with a higher perioperative increase in sGFAP (figure 2c). This finding suggests that patients with even moderate cognitive impairment are at increased risk for perioperative damage, which is in line with the practice of triaging patients with significant cognitive impairment from DBS surgery.³⁰ Our findings nonetheless indicate that there is no threshold for a “safe” DBS surgery. Rather, our findings suggest that safety declines gradually with cognitive performance, thus constituting an argument for “early” DBS surgery in patients that are expected to benefit from the procedure. Our findings also underline the importance of preoperative cognitive testing in PD, which should be conducted in a standardized way.³¹ Apart from cognitive status, we were not able to identify further risk factors robustly associated with increased sGFAP/sNfl values.

sGFAP transiently increased after DBS surgery, returning to baseline values just weeks after surgery (figure 1), comparable to sGFAP measurements in patients after brain tumor operations.³² The kinetics of sNfl were markedly different, as sNfl only returned to baseline 3 – 6 months after surgery (figure 1 and ¹⁶). Elevated levels of sGFAP^33,34 and sNfl³⁵ are associated with clinical milestones of PD (dementia, falls, nursing home placement and hallucinations) and with limited survival.^36,37 The fact that both biomarkers returned to baseline confirms the long-term safety of DBS therapy – albeit in patients carefully selected for this procedure. In particular, the lack of an increase of sGFAP or sNfl after initiation of stimulation and the lack of an association of both biomarkers to the total energy delivered by DBS demonstrates that the stimulation is not harmful and does not by itself confer a risk of accelerated neurodegeneration. Due to the rapid postoperative normalization, this finding is clearest for sGFAP (figure 1); the time course of sNfl confirms our previous findings.¹⁶

We hypothesized that the faster kinetic of sGFAP as opposed to sNfl might facilitate the detection of delayed injury after DBS surgery. PLE constitutes a rather commonly observed adverse event after DBS surgery, with at least 14.7% of DBS patients developing PLE after surgery, only half of which are symptomatic.¹³ One study even reported some degree of PLE in all consecutively investigated DBS patients (n=18) postoperatively³⁸ and thus, PLE may be underdiagnosed in clinical practice. As discussed above, the pathophysiology of PLE is not fully understood. The increase of sGFAP after DBS surgery as shown here suggests a glial involvement in the injury caused by the procedure and possibly in the development of PLE, as has been shown for other forms of brain edema.³⁹ As sGFAP shows a favorable kinetic to detect additional injury after DBS surgery and indicates glial damage, it might be suitable to examine PLE more systematically.

PLE typically occurs with a delay of up to 3 weeks after surgery,³⁸ hence at a time point when sGFAP has already returned to baseline in most cases (figure 1). Increasing or persistently increased sGFAP values during this time could therefore indicate additional injury. Due to its long half-life, sNfl measurements are not to useful detect additional neuronal injury after DBS surgery, like PLE, because – as shown here and elsewhere^40,41 – sNfl remains elevated for months after neuronal injury, thus potentially masking additional increases caused by a second injury. Thus, sGFAP as a blood-based biomarker could be used for safety assessments, in particular if imaging should be avoided, is not available or inconclusive.

There are limitations to this study. First, this was a single center study, and the results should be confirmed in a multicenter study. Second, a substantial number of measurements at time point 3 are missing (35%) due to organizational aspects, as the first activation of DBS in a few PwPD and in most patients with non-degenerative diseases was done in an outpatient setting and this did not always coincide with the prespecified timepoint of the measurements after surgery. Third, the number of non-degenerative patients included in our cohort was low. These patients mainly represent a control group in comparison to PwPD, so our study was not able to identify possible risk factors for surgery unique to either dystonia or ET.

In summary, our study showed a transient increase of markers of neuronal injury following DBS surgery. Serum GFAP declines rapidly after surgery in comparison to sNfl. Furthermore, we observed a negative correlation between preoperative cognitive function as measured by MoCA and sGFAP, which could not be described for sNfl. The best performing model could predict 41% of the variance of postoperative sGFAP measurements, which was not replicable for sNfl. This indicates that sGFAP is more reliably associated with preoperative risk factors and therefore better suited to detect differences in perioperative damage and risk than sNfl. Thus, sGFAP seems to be better suited as a marker of additional postoperative damage as opposed to sNfl. A single pre- and postoperative measurement of sGFAP might indicate a second injury such as PLE, making sGFAP potentially more feasible for clinical use.

Study population and design

To investigate differences in neuronal vulnerability upon a defined central nervous trauma (i.e., implantation of DBS electrodes), patients with an underlying neurodegenerative disease, i.e., PD, and non-neurodegenerative disease, i.e., essential tremor and cervical dystonia, were recruited. Serum NfL and GFAP were used as proxies for neuronal damage and glial activation respectively. Patients were recruited at the Department of Neurology at the University Hospital Dresden. The study was approved by the institutional review board of the Technische Universität Dresden (EK533122019, EK487122016, IRB00001473). Written informed consent was obtained from all participants before inclusion in the study. All study procedures were performed following relevant local guidelines and regulations. Serum samples were collected at defined time points to map the dynamics of these two processes in the postoperative course. These time points were as follows: (i) preoperative baseline (time point 0): within 2 days before surgery (n=58), (ii) postoperative time point 1: 3 – 5 days postoperatively (n=50), (iii) postoperative time point 2: 6 – 8 weeks after DBS surgery, before activation of stimulation (n=49), (iv) postoperative time point 3: up to one week after activation of stimulation (n=38) and (v) postoperative time point 4: 3 – 6 months after activation of stimulation (n=51). This work is based on a preliminary observation and therefore includes a subset of pre-published data (perioperative sNfl values of 21 patients).¹⁶

Demographic data were collected from all participants. For PD patients, this included further clinical information about motor and cognitive status (Hoehn & Yahr stage, Unified Parkinson’s Disease Rating Scale (UPDRS) part III in the OFF and ON condition, Montreal Cognitive Assessment (MoCA)) and the presence of disease-related complications. For all patients, we obtained information about common cardiovascular risk factors such as arterial hypertension, diabetes, uremia, BMI and smoking status.

A total of 64 patients were recruited for this study. The data of 6 patients were excluded from the analysis (all of them with PD), since either the baseline measurement was missing or less than half (i.e. 1 or 2) measurements had been performed. Thus, data of 58 patients were included in the analysis. Of the remaining 47 PwPD in our study, 44 received STN-DBS and 3 received GPi-DBS. All 3 dystonia patients received GPi-DBS and all 8 ET patients received VIM-DBS. Patients received implants of the following manufacturers: 45 Boston Scientific, 6 Medtronic, 4 Abbott and 3 Aleva Neurotherapeutics.

sNfl and GFAP measurements

Serum samples were stored at − 20 °C after preparation. NFL measurement was performed as described previously.⁴² In short, the Advantage NF-Light Singleplex-Kit on a Simoa HD-1 instrument (Quanterix) was used. Calibrators and diluted serum samples were measured in duplicates. The lower threshold of quantification was 0.775 pg/ml. The mean intraassay coefficient of variation of duplicates and the mean interassay coefficient of variation were < 10%. Serum GFAP measurement was performed using the GFAP Singleplex-Kit from Quanterix on the same HD-1 instrument.

Statistical Analyses

Comparisons between groups were carried out using the Mann–Whitney-U-Test, adjusted with Bonferroni correction where applicable. To assess differences between different time points, a Friedman’s test was performed. To assess differences between different time points, a Nemenyi-Friedman post-hoc test was used, which inherently corrects for multiplicity. We used a two-way ANOVA to determine the influence of the group (i.e. PD or non-degenerative) and the time points after surgery on sNfl and sGFAP measurements.

To normalize the right-skewed distribution of NFL in the further analysis, natural log-transformation (log(sNfl)) was used, as described by others. The same transformation was applied to serum GFAP measurements. Bivariate correlations were assessed with Spearman’s rank test. Since there is no global null-hypothesis assuming all correlations in the dataset are r = 0, these do not need to be corrected for multiplicity.⁴³

To assess whether baseline data could predict postoperative log(sNfl) and log(sGFAP) values, a random forest regressor and gradient boosting were used.

Statistical analyses were performed with Python. Sample size was not determined by a separate sample size estimate, as we tried to include all patients receiving DBS at our center from 2018 and 2023 that were available for assessments postoperatively and agreed to be included in this study.

Founding sources and Conflict of Interest

No specific funding was obtained for this study, there are no competing interests to report.

Financial Disclosures for the previous 12 months

BF reports no funding related to the conduct of this study. Outside of the submitted work, he reports grants from the German Research Foundation (DFG) and speaker honoraria from AbbVie, Stadapharm, Desitin, Zambon and Bial. NS has nothing to report. All other authors have nothing to report.

Author Contribution

1. Research project: A. Conception, B. Organization, C. Execution2. Statistical Analysis: A. Design, B. Execution, C. Review and Critique3. Manuscript Preparation: A. Writing of the first draft, B. Review and CritiqueNS: 2A, 2B, 2C, 3A, 3BJB: 2A, 2B, 2C, 3BJA: 1B, 1C, 2B, 2C, 3BMD: 1B, 1C, 2B, 2C, 3BLK: 1A, 1B, 1C, 2C, 3BWHP: 1C, 3BKA: 1B, 1C, 3BTZ: 1B, 1C, 3BBF: 1A, 2B, 2C, 3BAF: 1A, 1B, 1C, 2C, 3B

Acknowledgement

We would like to thank our patients for participating in this study.

Data Availability

The data collected in this study are available from the corresponding author upon reasonable request.

Limousin, P. & Foltynie, T. Long-term outcomes of deep brain stimulation in Parkinson disease. Nat. Rev. Neurol. 15, 234–242 (2019).
Flora, E. D., Perera, C. L., Cameron, A. L. & Maddern, G. J. Deep brain stimulation for essential tremor: a systematic review. Mov. Disord Off J. Mov. Disord Soc. 25, 1550–1559 (2010).
Volkmann, J. et al. Pallidal neurostimulation in patients with medication-refractory cervical dystonia: a randomised, sham-controlled trial. Lancet Neurol. 13, 875–884 (2014).
Tisch, S. & Kumar, K. R. Pallidal Deep Brain Stimulation for Monogenic Dystonia: The Effect of Gene on Outcome. Front. Neurol. 11, 630391 (2021).
Azevedo, P., Aquino, C. C. & Fasano, A. Surgical Management of Parkinson’s Disease in the Elderly. Mov. Disord Clin. Pract. 8, 500–509 (2021).
Al Ali, J. et al. Cognitive outcomes in patients with essential tremor treated with deep brain stimulation: a systematic review. Front. Hum. Neurosci. 18, (2024).
Olson, M. C., Shill, H., Ponce, F. & Aslam, S. Deep brain stimulation in PD: risk of complications, morbidity, and hospitalizations: a systematic review. Front. Aging Neurosci. 15, 1258190 (2023).
Kleiner-Fisman, G. et al. Subthalamic nucleus deep brain stimulation: summary and meta-analysis of outcomes. Mov. Disord Off J. Mov. Disord Soc. 21 (Suppl 14), S290–304 (2006).
Rački, V. et al. Cognitive Impact of Deep Brain Stimulation in Parkinson’s Disease Patients: A Systematic Review. Front. Hum. Neurosci. 16, (2022).
Cavallieri, F. et al. Predictors of Long-Term Outcome of Subthalamic Stimulation in Parkinson Disease. Ann. Neurol. 89, 587–597 (2021).
Pal, G. D. et al. Cognitive Effects of Subthalamic Nucleus Deep Brain Stimulation in Parkinson’s Disease with GBA1 Pathogenic Variants. Mov. Disord. 38, 2155–2162 (2023).
Kimmelman, J. et al. Risk of Surgical Delivery to Deep Nuclei: A Meta-analysis. Mov. Disord Off J. Mov. Disord Soc. 26, 1415–1421 (2011).
Whiting, A. C. et al. Peri-Lead Edema After Deep Brain Stimulation Surgery: A Poorly Understood but Frequent Complication. World Neurosurg. 124, e340–e345 (2019).
Cuba, C. M. K. E. et al. Idiopathic delayed-onset edema surrounding deep brain stimulation leads: Insights from a case series and systematic literature review. Parkinsonism Relat. Disord. 32, 108–115 (2016).
Mira, R. G., Lira, M. & Cerpa, W. Traumatic Brain Injury: Mechanisms of Glial Response. Front. Physiol. 12, 740939 (2021).
Frank, A. et al. Serum neurofilament indicates that DBS surgery can cause neuronal damage whereas stimulation itself does not. Sci. Rep. 12, 1446 (2022).
Abdelhak, A. et al. Blood GFAP as an emerging biomarker in brain and spinal cord disorders. Nat. Rev. Neurol. 18, 158–172 (2022).
Axelsson, M. et al. Glial fibrillary acidic protein: a potential biomarker for progression in multiple sclerosis. J. Neurol. 258, 882–888 (2011).
Meier, S. et al. Serum Glial Fibrillary Acidic Protein Compared With Neurofilament Light Chain as a Biomarker for Disease Progression in Multiple Sclerosis. JAMA Neurol. 80, 287–297 (2023).
You, H. et al. Evaluation of Blood Glial Fibrillary Acidic Protein as a Potential Marker in Huntington’s Disease. Front. Neurol. 12, (2021).
Castaño-Leon, A. M. et al. Serum assessment of traumatic axonal injury: the correlation of GFAP, t-Tau, UCH-L1, and NfL levels with diffusion tensor imaging metrics and its prognosis utility. J. Neurosurg. 138, 454–464 (2022).
Ferrari, F. et al. Quantification and prospective evaluation of serum NfL and GFAP as blood-derived biomarkers of outcome in acute ischemic stroke patients. J. Cereb. Blood Flow. Metab. 43, 1601–1611 (2023).
Yalachkov, Y. et al. Effect of Estimated Blood Volume and Body Mass Index on GFAP and NfL Levels in the Serum and CSF of Patients With Multiple Sclerosis. Neurol. Neuroimmunol. Neuroinflammation. 10, e200045 (2023).
Wang, X., Shi, Z., Qiu, Y., Sun, D. & Zhou, H. Peripheral GFAP and NfL as early biomarkers for dementia: longitudinal insights from the UK Biobank. BMC Med. 22, 192 (2024).
Ma, L. Z. et al. Serum Neurofilament Dynamics Predicts Cognitive Progression in de novo Parkinson’s Disease. J. Park Dis. 11, 1117–1127 (2021).
Ma, W. et al. Elevated Levels of Serum Neurofilament Light Chain Associated with Cognitive Impairment in Vascular Dementia. Dis. Markers 6612871 (2020). (2020).
Khalil, M. et al. Serum neurofilament light levels in normal aging and their association with morphologic brain changes. Nat. Commun. 11, 812 (2020).
Manouchehrinia, A. et al. Confounding effect of blood volume and body mass index on blood neurofilament light chain levels. Ann. Clin. Transl Neurol. 7, 139–143 (2020).
Tang, R. et al. Association of neurofilament light chain with renal function: mechanisms and clinical implications. Alzheimers Res. Ther. 14, 189 (2022).
Defer, G. L., Widner, H., Marié, R. M., Rémy, P. & Levivier, M. Core assessment program for surgical interventional therapies in Parkinson’s disease (CAPSIT-PD). Mov. Disord. 14, 572–584 (1999).
Foley, J. A., Foltynie, T., Limousin, P. & Cipolotti, L. Standardised Neuropsychological Assessment for the Selection of Patients Undergoing DBS for Parkinson’s Disease. Park. Dis. 4328371 (2018). (2018).
Baumgarten, P. et al. Pre- and early postoperative GFAP serum levels in glioma and brain metastases. J. Neurooncol. 139, 541–546 (2018).
Tang, Y. et al. Plasma GFAP in Parkinson’s disease with cognitive impairment and its potential to predict conversion to dementia. Npj Park Dis. 9, 1–5 (2023).
Lin, J. et al. Plasma glial fibrillary acidic protein as a biomarker of disease progression in Parkinson’s disease: a prospective cohort study. BMC Med. 21, 420 (2023).
Frank, A. et al. Serum neurofilament indicates accelerated neurodegeneration and predicts mortality in late-stage Parkinson’s disease. Npj Park Dis. 10, 14 (2024).
Schnalke, N. et al. Morbidity Milestones Demonstrate Long Disability-Free Survival in Parkinson’s Disease Patients with Deep Brain Stimulation of the Subthalamic Nucleus. Mov. Disord Clin. Pract. 10, 569–578 (2023).
Kempster, P. A., O’Sullivan, S. S., Holton, J. L., Revesz, T. & Lees, A. J. Relationships between age and late progression of Parkinson’s disease: a clinico-pathological study. Brain. 133, 1755–1762 (2010).
Borellini, L. et al. Peri-lead edema after deep brain stimulation surgery for Parkinson’s disease: a prospective magnetic resonance imaging study. Eur. J. Neurol. 26, 533–539 (2019).
Li, D. et al. Neurochemical regulation of the expression and function of glial fibrillary acidic protein in astrocytes. Glia. 68, 878–897 (2020).
Graham, N. S. N. et al. Axonal marker neurofilament light predicts long-term outcomes and progressive neurodegeneration after traumatic brain injury. Sci. Transl Med. 13, eabg9922 (2021).
Shahim, P. et al. Neurofilament light as a biomarker in traumatic brain injury. Neurology. 95, e610–e622 (2020).
Akgün, K. et al. Profiling individual clinical responses by high-frequency serum neurofilament assessment in MS. Neurol. Neuroimmunol. Neuroinflammation. 6, e555 (2019).
García-Pérez, M. A. Use and misuse of corrections for multiple testing. Methods Psychol. 8, 100120 (2023).

No competing interests reported.

Download PDF

Reviews received at journal
01 Oct, 2024
Reviewers agreed at journal
12 Sep, 2024
Reviewers agreed at journal
06 Sep, 2024
Reviewers invited by journal
05 Sep, 2024
Editor assigned by journal
05 Sep, 2024
Editor invited by journal
02 Sep, 2024
Submission checks completed at journal
31 Aug, 2024
First submitted to journal
24 Aug, 2024

You are reading this latest preprint version

Serum Glial Fibrillary Acidic Protein (GFAP) as a Potential Biomarker for Monitoring Postoperative Complications in Deep Brain Stimulation Surgery

Status:

Version 1

Abstract

Figures

Introduction

Results

Serum neurofilament after DBS surgery

Serum GFAP after DBS surgery

Baseline values of sNfL and sGFAP are affected by age and BMI, but not disease-specific factors in PwPD

Perioperative damage as reported by sGFAP is influenced by age, motor symptoms, cognitive impairment and baseline sGFAP

Discussion

Conclusion

Methods

Statistical Analyses

Declarations

Founding sources and Conflict of Interest

Financial Disclosures for the previous 12 months

Author Contribution

Acknowledgement

Data Availability

References

Additional Declarations

Status:

Version 1