Abstract
Background:
Patients with glioblastoma multiforme (GBM) who undergo radiation often require anticonvulsants during treatment. The aim of this study was to determine the effects of anticonvulsants on GBM clinical outcomes.
Methods:
A retrospective analysis was performed using the SEER-Medicare database. All patients with GBM who were treated with radiation and concurrently taking an anticonvulsant were included in final analysis. Each class of medication was further subdivided by mechanism of action. Descriptive statistics were performed for all variables. Kaplan Meier survival curves were generated for each class of medication and Cox regression analysis was performed to assess the effect of each individual variable on survival.
Results:
There were 1561 patients available for final analysis. On multivariate Cox regression analysis, GBM patients taking sodium/calcium (Na/Ca) channel blocker anticonvulsants during radiation therapy demonstrated both improved overall survival (OS) (HR, 0.799; 95% CI [0.716, 0.891]; P<0.001) and cancer specific survival (CSS) (HR, 0.814; 95% CI [0.727, 0.911]; P<0.001).
Conclusion:
OS was significantly better in patients taking NA/Ca channel blockers among patients with GBM who were concurrently undergoing radiation therapy.
Keywords: GBM, radiation, anticonvulsants
Introduction
Radiation is an important part of treatment for many patients with glioblastoma multiforme (GBM). Many of these patients often present with neuro or psychiatric symptoms which require medical management. Seizures are thought to occur among 40–60% of patients with brain tumors whereas psychiatric symptoms may occur in up to 80% of patients 1,2. Whether there is any definitive interaction between radiation and use of various central nervous system medications remains unknown. However, both in vivo and in vitro studies indicate that certain central nervous system (CNS) medications may have neuroprotective effects or even radiation sensitizing effects with cranial irradiation 3–5.
While such data may lead to development of new treatment strategies and ways to prevent radiation induced neurotoxicity in the future, it is important to have clinical data to ensure there are no unexpected interactions between such medications and radiation efficacy. For example, while lithium has been implicated in preventing neurocognitive decline, one concern with the use of lithium concurrently with radiotherapy is the possibility that it may reduce the effects of radiation induced cell killing by reducing free radical damage 6.
The aim of this study was to perform a retrospective analysis of the effects of anticonvulsants on both overall survival (OS) and cancer specific survival (CSS) among patients with brain GBM. The Surveillance, Epidemiology and End Results (SEER)-Medicare national cancer database was used to perform this analysis.
Materials and Methods
The Surveillance, Epidemiology, and End Results (SEER)-Medicare linked database was used to identify all patients with central nervous system tumors diagnosed between 2007 and 2014. The SEER-Medicare database consists of two components with SEER covering approximately 35% of the U.S. cancer population 7. Medicare is a federal insurance program which covers 94% of elderly patients aged 65 and older. Claims data in Medicare is linked to the SEER registry to allow for epidemiological and cancer research 7.
Only patients diagnosed with glioblastoma multiforme who had undergone radiation therapy were included in the final cohort. Moreover, patients with more than one primary, unknown month/year of diagnosis, or who had a diagnosis on autopsy or death certificate were excluded. Patients were required to have continuous Parts A and B coverage one year prior to diagnosis and at least one year after diagnosis or until time of death or censoring. Part A includes hospital, skilled nursing, hospice, and home health services while part B includes physician and outpatient services. Patients were also required to have continuous Part D coverage, which covers prescription drugs, for at least one year after diagnosis or until time of death or censoring. Finally, patients with HMO enrollment during this time frame were excluded.
The Patient Entitlement and Diagnosis Summary File (PEDSF) was used to extract all patient demographics and survival data including the following: age, sex, marital status, race, histology, survival months, overall survival, and cancer specific survival. Patients missing any of these variables were excluded from analysis. Both radiation therapy and radiation therapy were identified using ICD-9, ICD-10, HCPCS/CPT, and revenue center codes (Supplementary Table 2).
Medicare part D data was used to extract information regarding anticonvulsants. Anticonvulsants were further grouped into sodium calcium (Na/Ca) channel blockers, GABA receptor enhancers, and combination drugs defined as those with dual action on both Na/Ca channels and GABA receptors (Supplementary Table 1). Patients were regarded to have been actively taking any one of these medications if there was at least one day of overlap between prescription start and end dates and radiation start and end dates.
The primary outcome measures were times in months from diagnosis to death secondary to any cause for OS and secondary to the cancer diagnosis for CSS. Descriptive statistics were calculated for all variables. Survival curves were generated using the Kaplan Meier method with differences in survival calculated using the log rank test. Multivariable analysis was performed with Cox proportional hazards regression models to determine the predictive performance of covariates with respect to OS and CSS. Comparisons were considered statistically significant at P < 0.05. All statistical analyses were performed in SPSS, version 24 (IBM Corporation) and SAS (SAS Institute, Cary, NC). A waiver was obtained by our institutional review board.
Results
A total of 1561 patients met inclusion criteria with 842 patients (53.9%) 72 years or younger and 806 (51.6%) males (Table 1). There were 1252 patients (80.2%) who underwent surgery to the primary and 402 patients (25.8%) who underwent chemotherapy. Overall, there were 1233 patients (79.0%) on anticonvulsants with 1055 (67.5%) on Na/Ca channel blockers, 140 (8.9%) on GABA receptor enhancers, and 38 (2.4%) on combination drugs. Additional details are summarized in Table 1.
Table 1:
Baseline Demographics
| N | % | |
|---|---|---|
| Age | ||
| ≤72 | 842 | 53.9 |
| >72 | 719 | 46.1 |
| Sex | ||
| Male | 806 | 51.6 |
| Female | 755 | 48.4 |
| Marital Status | ||
| Married | 961 | 61.6 |
| Single/Other | 600 | 38.4 |
| Race | ||
| White | 1442 | 92.4 |
| Black | 65 | 4.2 |
| Other/Unknown | 54 | 3.5 |
| Surgery | ||
| Yes | 1252 | 80.2 |
| No/Unknown | 309 | 19.8 |
| Chemotherapy | ||
| Yes | 402 | 25.8 |
| No | 1159 | 74.2 |
| Anticonvulsants | ||
| Na/Ca Channel Blocker | 1055 | 67.5 |
| GABA Receptor Enhancer | 140 | 8.9 |
| Combination | 38 | 2.4 |
| None | 328 | 21.0 |
OS was significantly better in patients taking Na/Ca Channel blockers among GBM patients taking anticonvulsants as demonstrated in figure 1, with 1- and 2-year OS rates of 34% and 12% among those taking Na/Ca channel blockers vs 23% and 7% among those not on these medications, respectively (Figure 1, P <0.001). Similar associations were observed with CSS among those taking Na/Ca channel blockers vs those not on these meds with 1- and 2-year CSS rates of 36% and 14% and 27% and 9%, respectively (P <0.001). No other significant associations between either OS or CSS were seen among the other classes of anticonvulsants (Supplementary Figures 1 and 2).
Figure 1:
Overall survival (a) and cancer specific survival (b) among GBM patients taking anticonvulsants (Na/Ca channel blockers)
On multivariate Cox regression, the following variables were significant predictors of OS among GBM patients: age, marital status, race, surgery, chemotherapy, and anticonvulsant use (Table 2). Specifically, OS was significantly better in patients taking Na/Ca channel blockers as compared to patients who were not on these medications (HR, 0.799; 95% CI [0.716, 0.891]; P<0.001). Similar to the effects observed with OS, CSS was significantly better in patients taking Na/Ca channel blockers among GBM patients (HR, 0.814; 95% CI [0.727, 0.911]; P<0.001).
Table 2:
Multivariate Cox Regression of baseline variables with overall survival
| Age | Hazard Ratio | 95% CI | P-value |
|---|---|---|---|
| ≤72 | Ref | - | - |
| >72 | 1.331 | (1.200, 1.477) | <0.001 |
| Sex | |||
| Male | Ref | - | - |
| Female | 1.003 | (0.902, 1.116) | 0.949 |
| Marital Status | |||
| Married | Ref | - | - |
| Single/Other | 1.119 | (1.003, 1.249) | 0.044 |
| Race | |||
| White | Ref | - | - |
| Black | 0.929 | (0.718, 1.202) | 0.576 |
| Other/Unknown | 0.620 | (0.466, 0.824) | 0.001 |
| Surgery | |||
| No/Unknown | Ref | - | - |
| Yes | 0.521 | (0.458, 0.592) | <0.001 |
| Chemotherapy | |||
| No | Ref | - | - |
| Yes | 0.736 | (0.655, 0.827) | <0.001 |
| Anticonvulsants | |||
| Na/Ca Channel Blocker | |||
| No | Ref | - | - |
| Yes | 0.799 | (0.716, 0.891) | <0.001 |
| GABA receptor enhancer | |||
| No | Ref | - | - |
| Yes | 0.947 | (0.791, 1.133) | 0.551 |
| Combination | |||
| No | Ref | - | - |
| Yes | 0.985 | (0.706, 1.376) | 0.932 |
Discussion
Primary tumors of the CNS remain associated with a poor prognosis. Thus, new forms of treatment are necessary to help improve survival. Ideally, new therapeutic modalities should have a synergistic, or at least additive, effect with currently accepted treatments. Of course, this also means that new treatments should not have any unexpected antagonistic effects with our current standard of care which could potentially have detrimental effects on survival. Radiotherapy, for example, works through the formation of free radicals which leads to DNA damage and ultimately, cell death 8. However, certain medications may act as free radical scavengers/antioxidants; thus, these would conceivably interfere with radiation’s mechanism of action. For example, antipsychotics, anticonvulsants, and lithium have all been implicated in the scavenging of free radicals and producing antioxidant effects 9–11.
Interestingly, OS and CSS were both improved among patients taking the class of anticonvulsants that block sodium and/or calcium channels on both Kaplan Meier survival analysis and Cox regression. Psychiatric and CNS medications have demonstrated evidence of anticancer activity 12–15. Moreover, certain medications may even potentiate the effect of radiation. Animal studies indicate that the anticonvulsant, valproate, may sensitize glioblastoma tumor cells to radiation and even protect hippocampal neurons without compromising radiation efficacy 4. Valproate’s effects have been attributed to its interaction with various proteins and biochemical pathways including histone deacetylase inhibitors and the Wnt/β catenin pathway, leading to cell cycle arrest, inhibition of angiogenesis, and apoptosis 12–15.
Valproate is also one of the most widely studied anticonvulsants in the clinical setting with various retrospective studies demonstrating improved survival among GBM patients taking valproate 16–19. In their retrospective analysis of GBM patients undergoing radiation, Barker et al. demonstrated improved OS among patients taking concurrent valproate. Among the five most commonly prescribed anticonvulsants which included valproate, phenytoin, levetiracetam, carbamazepine, and phenobarbital; only valproate was significantly associated with OS on Cox regression 18.
Despite these promising observations, there is conflicting data from pooled analyses. In their meta-analysis of 1634 patients from 5 studies, Yu et al. demonstrated improved survival with valproate (HR, 0.56; 95% CI [0.44, 0.71]) 20. However, no similar association was demonstrated by Happold et al, who performed a pooled analysis from four randomized controlled trials consisting of 1869 patients 21. In their final cohort of GBM patients treated with chemoradiation and temozolomide, no significant improvement in either overall or progression free survival was demonstrated among patients taking valproate or levetiracetam.
While some studies have demonstrated improved outcomes associated with anticonvulsant use, these observations may have been due to the treated conditions rather than the medications themselves. Epilepsy, for instance, has often been linked to improved survival 22–25. While this association may be due to earlier diagnosis and treatment and/or the anti-neoplastic effects of anticonvulsants, many studies are retrospective in nature and thus definitive conclusions are difficult to establish 22,25. Thus, further prospective studies are necessary to help validate these results. ChiCTR2100049941 is enrolling patients in a double blind randomized clinical trial to determine the effect of temozolomide and levetiracetam in GBM patients and will help provide further information 26.
Limitations of this study include those that are inherent to retrospective analyses. It is possible that any differences in survival that were observed may have been due to underlying selection biases that could not be accounted for in our analysis. For example, there was a lack of data regarding molecular histological features such as IDH mutation and MGMT methylation status in SEER-Medicare and thus, could not be controlled for in analysis. Similarly, it is possible that certain GBM cases were due to secondary transformation from lower grade gliomas which could also impact results. Moreover, we looked at a specific subgroup of patients, specifically those aged 65 and above with CNS primaries who had undergone radiation as part of their treatment regimen. Thus, our results may not be applicable to those who do not fall into this category. We also studied classes of drugs rather than individual drugs which may further influence results.
Moreover, not all medications that affect the CNS were included in analysis. Finally, we collected data regarding both radiation therapy and simultaneous anticonvulsant to observe any potential interactions between the two. However, while SEER-Medicare provides prescription and end dates, compliance is not recorded. Thus, it is possible, that some patients may not have been actively taking their prescribed anticonvulsants during radiation despite having an active prescription.
Nonetheless, the strength of our study lies in the large patient population available for analysis. To our knowledge, this is the first large national database study analyzing the effects of various CNS medications on both OS and CSS. While the results of this study indicate that there may be a significant association between CNS medications and survival among patients undergoing radiation for CNS primaries, additional prospective studies and randomized controlled trials are necessary to validate our results.
Supplementary Material
Table 3:
Multivariate Cox Regression of baseline variables with cancer specific survival
| Glioblastoma Multiforme | |||
|---|---|---|---|
| Age | Hazard Ratio | 95% CI | P-value |
| ≤72 | Ref | - | - |
| >72 | 1.312 | (1.179, 1.460) | <0.001 |
| Sex | |||
| Male | Ref | - | - |
| Female | 1.005 | (0.901, 1.122) | 0.923 |
| Marital Status | |||
| Married | Ref | - | - |
| Single/Other | 1.104 | (0.986, 1.236) | 0.088 |
| Race | |||
| White | Ref | - | - |
| Black | 0.867 | (0.659, 1.141) | 0.309 |
| Other/Unknown | 0.602 | (0.447, 0.809) | <0.001 |
| Surgery | |||
| No/Unknown | Ref | - | - |
| Yes | 0.516 | (0.451, 0.589) | <0.001 |
| Chemotherapy | |||
| No | Ref | - | - |
| Yes | 0.755 | (0.670, 0.851) | <0.001 |
| Anticonvulsants | |||
| Na/Ca Channel Blocker | |||
| No | Ref | - | - |
| Yes | 0.814 | (0.727, 0.911) | <0.001 |
| GABA receptor enhancer | |||
| No | Ref | - | - |
| Yes | 0.930 | (0.771, 1.121) | 0.444 |
| Combination | |||
| No | Ref | - | - |
| Yes | 0.970 | (0.685, 1.374) | 0.864 |
Funding:
Department of Radiation Oncology, University of Arkansas for Medical Sciences
Footnotes
Conflict of Interest: The authors declare no conflicts of interest.
References
- 1.Madhusoodanan S, Ting MB, Farah T, et al. : Psychiatric aspects of brain tumors: A review. World journal of psychiatry 5:273–285, 2015 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Vecht CJ, Kerkhof M, Duran-Pena A: Seizure prognosis in brain tumors: new insights and evidence-based management. The oncologist 19:751–759, 2014 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Malaterre J, McPherson CS, Denoyer D, et al. : Enhanced lithium-induced brain recovery following cranial irradiation is not impeded by inflammation. Stem Cells Transl Med 1:469–79, 2012 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Thotala D, Karvas RM, Engelbach JA, et al. : Valproic acid enhances the efficacy of radiation therapy by protecting normal hippocampal neurons and sensitizing malignant glioblastoma cells. Oncotarget 6:35004–22, 2015 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Yazlovitskaya EM, Edwards E, Thotala D, et al. : Lithium treatment prevents neurocognitive deficit resulting from cranial irradiation. Cancer Res 66:11179–86, 2006 [DOI] [PubMed] [Google Scholar]
- 6.Khasraw M, Ashley D, Wheeler G, et al. : Using lithium as a neuroprotective agent in patients with cancer. BMC Medicine 10:131, 2012 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Enewold L, Parsons H, Zhao L, et al. : Updated Overview of the SEER-Medicare Data: Enhanced Content and Applications. J Natl Cancer Inst Monogr 2020:3–13, 2020 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Tashiro S, Nishibuchi I, Wondergem J: Why Radiotherapy Works Chapter 6. International Atomic Energy Agency (IAEA), IAEA, 2017 [Google Scholar]
- 9.Zhao Q-L, Ito H, Kondo T, et al. : Antipsychotic drugs scavenge radiation-induced hydroxyl radicals and intracellular ROS formation, and protect apoptosis in human lymphoma U937 cells. Free Radical Research 53:304–312, 2019 [DOI] [PubMed] [Google Scholar]
- 10.Machado-Vieira R, Andreazza AC, Viale CI, et al. : Oxidative stress parameters in unmedicated and treated bipolar subjects during initial manic episode: a possible role for lithium antioxidant effects. Neurosci Lett 421:33–6, 2007 [DOI] [PubMed] [Google Scholar]
- 11.Mori A, Noda Y, Packer L: The anticonvulsant zonisamide scavenges free radicals. Epilepsy Research 30:153–158, 1998 [DOI] [PubMed] [Google Scholar]
- 12.Zhuo C, Xun Z, Hou W, et al. : Surprising Anticancer Activities of Psychiatric Medications: Old Drugs Offer New Hope for Patients With Brain Cancer. Frontiers in pharmacology 10:1262–1262, 2019 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Blaheta RA, Cinatl J Jr.: Anti-tumor mechanisms of valproate: a novel role for an old drug. Med Res Rev 22:492–511, 2002 [DOI] [PubMed] [Google Scholar]
- 14.Chen Y, Tsai YH, Tseng SH: Valproic acid affected the survival and invasiveness of human glioma cells through diverse mechanisms. J Neurooncol 109:23–33, 2012 [DOI] [PubMed] [Google Scholar]
- 15.Osuka S, Takano S, Watanabe S, et al. : Valproic acid inhibits angiogenesis in vitro and glioma angiogenesis in vivo in the brain. Neurol Med Chir (Tokyo) 52:186–93, 2012 [DOI] [PubMed] [Google Scholar]
- 16.Oberndorfer S, Piribauer M, Marosi C, et al. : P450 enzyme inducing and non-enzyme inducing antiepileptics in glioblastoma patients treated with standard chemotherapy. J Neurooncol 72:255–60, 2005 [DOI] [PubMed] [Google Scholar]
- 17.Weller M, Gorlia T, Cairncross JG, et al. : Prolonged survival with valproic acid use in the EORTC/NCIC temozolomide trial for glioblastoma. Neurology 77:1156–1164, 2011 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Barker CA, Bishop AJ, Chang M, et al. : Valproic acid use during radiation therapy for glioblastoma associated with improved survival. Int J Radiat Oncol Biol Phys 86:504–9, 2013 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Kerkhof M, Dielemans JC, van Breemen MS, et al. : Effect of valproic acid on seizure control and on survival in patients with glioblastoma multiforme. Neuro Oncol 15:961–7, 2013 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Yuan Y, Xiang W, Qing M, et al. : Survival analysis for valproic acid use in adult glioblastoma multiforme: a meta-analysis of individual patient data and a systematic review. Seizure 23:830–5, 2014 [DOI] [PubMed] [Google Scholar]
- 21.Happold C, Gorlia T, Chinot O, et al. : Does Valproic Acid or Levetiracetam Improve Survival in Glioblastoma? A Pooled Analysis of Prospective Clinical Trials in Newly Diagnosed Glioblastoma. J Clin Oncol 34:731–9, 2016 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Berendsen S, Varkila M, Kroonen J, et al. : Prognostic relevance of epilepsy at presentation in glioblastoma patients. Neuro Oncol 18:700–6, 2016 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Dobran M, Nasi D, Chiriatti S, et al. : Prognostic Factors in Glioblastoma: Is There a Role for Epilepsy? Neurol Med Chir (Tokyo) 58:110–115, 2018 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Fan X, Li Y, Shan X, et al. : Seizures at presentation are correlated with better survival outcomes in adult diffuse glioma: A systematic review and meta-analysis. Seizure 59:16–23, 2018 [DOI] [PubMed] [Google Scholar]
- 25.Toledo M, Sarria-Estrada S, Quintana M, et al. : Epileptic features and survival in glioblastomas presenting with seizures. Epilepsy Res 130:1–6, 2017 [DOI] [PubMed] [Google Scholar]
- 26.Sun M, Huang N, Tao Y, et al. : The efficacy of temozolomide combined with levetiracetam for glioblastoma (GBM) after surgery: a study protocol for a double-blinded and randomized controlled trial. Trials 23:234, 2022 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.