In the acute phase of anti-NMDAR AE, carbamazepine [CBZ] (n = 11) and oxcarbazepine [OXC] (n = 16) were the most commonly selected ASMs, while valproic acid [VPA] was one of the most commonly used continuous therapy agents during follow-up (10). Melissa Chavez-Castillo et al. examined 31 pediatric patients with anti-NMDAR AE and seizures. A total of 15 (48%) patients were treated with VPA; the most common ASM combination was VPA and levetiracetam [LEV]; 18 (58%) patients were treated with LEV; 13 (42%) patients were treated with phenytoin [PHT]; and 7 (23%) patients were treated with OXC (11). For seizures that remain uncontrolled after the use of one ASM in the acute phase, a combination of ASMs is needed. Haberlandt et al. reported that 15/17 (88%) patients with anti-NMDAR encephalitis took ASMs, with 3/15 (20%) patients taking 1 ASM and 12/15 (80%) patients taking multiple ASMs; 6/12 (50%) of patients taking multiple ASMs (50%) developed resistance, 1/3 (17%) of patients taking 1 ASM developed resistance, and 3 patients on multiple ASMs developed epilepsy (12). Zhong R et al. reported that among 30 patients with anti-NMDAR encephalitis, 15 (50%) were taking 1 ASM, 7 (23.3%) were taking 2 ASMs, and 8 (26.7%) were taking >2 ASMs; they found that for autoimmune encephalitis, including anti-NMDAR encephalitis, the use of more ASMs was independently associated with an increased risk of epilepsy after the acute phase of AE (13). Chavez-Castillo M et al. reported that an average of 2 ASMs were required for seizure control in 30 patients given ASMs during the acute phase of anti-NMDAR encephalitis, and more ASMs were required for patients with drug resistance, with a mean of 2.5 ASMs (11). Liu X et al. reported that in 88 patients with acute seizures in anti-NMDAR encephalitis, for 14 patients treated with a single drug (LEV, valproate, or topiramate), the median time to seizure cessation was 2 weeks; 50 patients (56.8%) received more than one ASM treatment, and the median time from ASM treatment to seizure cessation was 5 weeks (14). The data from 198 patients are shown in Table 2.

When SE occurs in the acute phase of anti-NMDAR AE, treatment with combinations of multiple anticonvulsants, including phenytoin (PHT), LEV, sodium valproate (VPA), topiramate (TPM), ethosuximide, lacosamide, and clobazam, failed to achieve seizure control without propofol and midazolam (MDZ) infusions (15). For refractory SE, high doses of VPA, MDZ, isoproterenol, thiopental sodium, lidocaine, and phenobarbital (PB) may be used. High-dose TPM combined with intravenous high-dose PB or high-dose lidocaine minimizes side effects (15). Liu X et al. reported that 22 patients with SE required ASM and anesthetics, of whom 13 patients (59.1%) had controlled SE, while another 9 patients (40.9%) had unremitting or remitting SE after reduction or discontinuation of anesthetics, relapsed and eventually died in the acute phase (14). Finné Lenoir X et al. reported a case of anti-NMDAR encephalitis in a 17-year-old Asian male in whom a large number of ASMs were ineffective in his prolonged persistent epilepsy (16). In addition to conventional ASMs, Santoro et al. reported ketamine as an effective adjunctive therapy for hyper refractory persistent epilepsy in patients with anti-NMDAR receptor encephalitis, suggesting that administration of a loading dose followed by maintenance infusion resulted in clinical and/or electrogram seizure cessation in less than 48 h (17). Yen HK et al. reported anti-NMDAR encephalitis SE in 6 patients (85.7%) treated with sodium channel blockers and LEV, of which 3 patients (42.9%) received a combination of sodium channel blockers, LEV and sodium valproate, and the duration of SE in patients with combined sodium valproate and PHT at the end of SE was 2.5 ± 0.7 days (n = 2). Both patients had SE within significant control within 1 day of PHT use; in contrast, the duration of SE in patients who did not receive the combination of PHT, LEV and sodium valproate was 4.0 ± 4.7 days (n = 5) (18). Data from these studies are shown in Table 3.

After the acute phase of anti-NMDAR AE, it is not typically necessary to take ASMs for long periods. Early studies have found that ASMs can be gradually reduced during follow-up and that most patients who discontinue the drug within 2 years achieved seizure freedom (19). Furthermore, Zeng W et al. found that most anti-NMDAR AE patients adhered to ASMs for 3–12 months (median 0.5 years), with a mean duration of less than 1 year on ASMs and a low relapse rate (20). Qi Huang et al. reported that only 5.8% of 34 patients with anti-NMDAR AE who discontinued ASMs experienced recurrence at the 1-year follow-up, 14 patients with anti-NMDAR AE were free of seizures for at least 1 year after the last use of ASMs, and only 2 of 23 patients who discontinued ASMs within 3 months experienced recurrence (10). Yao L et al. reported that most of the 60 patients with anti-NMDAR AE discontinued ASM within 1 year, and 49 patients achieved seizure freedom at the final follow-up; however, the study included multiple antibody encephalitis at the same time, and the number of patients with NMDAR encephalitis discontinuation within 1 year was not clearly marked (21). Zhang M et al. found that at the last follow-up (range: 6–39 months), 12 of 14 patients (85.7%) with seizures achieved seizure freedom, and 10 of these 12 patients (83.3%) discontinued ASM within 1 year (22). Another study with a median follow-up of 2 years found that 33 patients (38.3%) on ASM discontinued their medication after 3 months of follow-up, 53 (61.7%) continued on ASM medication after 3 months of follow-up; additionally, no seizures were observed in patients with >2 years of follow-up, short-term (<3 months) ASM treatment was associated with long-term (>3 months) ASM treatment, and there was no difference in seizure risk factors between these groups (14). In another similar study, analysis of NMDAR brain patients with seizures in the acute phase and those who continued to have seizures after the acute phase found that the mean duration of treatment was 3.58 ± 1.08 months (range 1–6 months) for the short-term seizure group (seizures occurred only in the acute phase) and 8.40 ± 1.14 months for the persistent seizure group (seizures continued to occur after the acute phase). 1.14 months (range 7–12 months), and both groups were seizure-free after the discontinuation of ASM (23). Zhang JZ et al. reported that among 34 (68.0%) patients with anti-NMDAR encephalitis who developed seizures, 3 patients had herpes simplex virus encephalitis secondary to anti-NMDAR encephalitis and had persistent seizures; among the other patients, the seizure duration ranged from 1 ~ 47 days, and 16 patients (47.1%) did not continue ASM after discharge (24). The data from 273 patients are shown in Table 4.

According to studies investigating a total of 198 patients with seizures in the acute phase of anti-NMDAR AE (Table 2), valproic acid (VPA), levetiracetam (LEV), carbamazepine (CBZ), and oxcarbazepine (OXC) are the most commonly used ASMs to control these seizures. These studies in Table 2 show that the use of 1–3 ASMs is sufficient, as combining ASMs predicts an increased likelihood of developing epilepsy. VPA, LEV, and PHT can be considered in SE that occurs in the acute phase of anti-NMDAR AE; furthermore, a combination of sodium channel blockers and VPA may be an option worth revisiting for faster seizure control, and adjuvant anesthetics such as MDZ, propofol, and ketamine may be used for patients with refractory SE. These statistics in Table 4 suggest that the optimal duration of ASM use is 3 months for those with short-term seizures or a low frequency of seizures after the acute phase, while patients who still have seizures after the acute phase should slowly stop the drug after 6–12 months, and patients with persistent seizures can use ASM for a longer time.

ASMs are an adjuvant immunotherapy in the treatment of anti-LGI1 AE seizures. In one study, patients with anti-LGI1 AE presented with three main types of seizure: faciobrachial dystonic seizure (FBDS) (44%), medial temporal lobe epilepsy (MTLE)-like seizures (66%), and focal to bilateral tonic–clonic seizures (FBTCS) (77%); they showed only slight improvement after administration of ASMs, whereas immunotherapy was more effective (25). In that study, 4 (44%) patients received corticosteroids combined with intravenous immunoglobulin (IVIG), 4 (44%) patients received corticosteroids, 1 (11%) patient did not receive immunotherapy, and 8 (88%) patients received ASMs for seizures; those who received immunotherapy were significantly healthier and more likely to be free of seizures (25). LGI1 mutation-associated autosomal dominant temporal lobe epilepsy is relatively benign, and standard ASM treatment usually controls seizures (26). Van Sonderen A et al. reported 21 patients with anti-LGI1 AE who received immunotherapy in combination with ASMs over a follow-up period of more than 2 years, with a median follow-up period of 42 months. In the last year of follow-up, 28% of patients were still taking ASMs, and only 14% had seizures in the last year of follow-up (27). Huang Q et al. reported that a 72-year-old man with anti-LGI1 AE who presented with frequent seizures at presentation discontinued ASMs after 3 months, and no recurrence was reported at the 1-year follow-up (10). de Bruijn M et al. reported that for patients with anti-LGI1 AE, the odds of achieving freedom from seizures were higher with immunotherapy than with ASMs alone (19). The study by Feyissa AM et al. also concluded that immunotherapy is the treatment that best achieves freedom from seizures in anti-LGI1 AE (28). Data for these patients are shown in Table 5.

FBDS is the most specific type of attack in anti-LGI1 AE, occurring in approximately half of patients (25). Most studies have concluded that corticosteroids can control FBDS more effectively than ASMs (29). FBDS did not respond well to ASMs in a study by Gao L et al. (30). de Bruijn M et al. reported that focal seizures in patients with anti-LGI1 activity responded relatively well to CBZ, whereas FBDS showed little response to any ASM (19). In patients with FBDS, immunotherapy was associated with higher odds of achieving seizure freedom than ASM treatment (20/30 vs. 3/34, p = 0.001) (28). Some studies have shown positive outcomes for patients with FBDS receiving both immunotherapy and ASMs (31–33). The data from 63 patients are shown in Table 6.

Furthermore, de Bruijn M et al. reported 53 patients with anti-LGI1 AE and recommended the use of ASMs with sodium channel blocking properties (e.g., CBZ or OXC) first in the symptomatic treatment of patients with seizures. CBZ was more effective than LEV in reducing seizure frequency (p = 0.031) and appeared to have the best effect in reducing the frequency of focal seizures, but it tends to cause frequent rashes (19). Irani SR et al. reported that a high percentage (50%) of patients among 9 people with anti-LGI1 activity demonstrated skin reactions caused by the use of ASMs (2 PHT, 2 CBZ, and 1 VPA) (34). Shin YW et al. also reported that 10 of 20 anti-LGI1 AE patients (50%) had their ASMs changed due to adverse cutaneous drug reactions. Eight of them presented with maculopapular eruption, one presented with drug rash with eosinophilia and systemic symptoms syndrome, and one presented with eczema. Causative agents mostly consisted of aromatic ASMs. OXC was discontinued in 2 additional patients due to hyponatremia. Six patients (30%) discontinued their dose of levetiracetam because of psychiatric manifestations, including irritability/aggressive behavior (4 patients), insomnia (1 patient), and depressive mood (1 patient) (35). While the side effects of VPA typically include memory loss and tremors, the side effects of LEV consist of rash and severe behavioral changes, including severe psychotic behavior and suicidal thoughts (19).

According to the reports above, ASMs have a role in the treatment of anti-LGI1 AE seizures, but FBDS with anti-LGI1 AE responds poorly to ASMs. Overall, immunotherapy should take precedence, and ASMs are typically used for adjuvant therapy. Moreover, anti-LGI1 AE patients are prone to skin reactions and hyponatremia with CBZ and OXC or drug side effects such as adverse psychiatric reactions, memory loss, and tremor with VPA and LEV.

Huang Q et al. reported three patients with anti-GABABR autoimmune encephalitis (anti-GABABR AE), one of whom was a 35-year-old female who presented with frequent seizures at presentation and discontinued ASM after 6 months with no reported recurrence during a 1-year follow-up, while two other elderly patients (50 and 64 years old, respectively) presented with intractable seizures (10). Zhong R et al. reported on 14 patients with anti-GABABR AE, of whom 10 (71.4%) developed SE and 6 (42.9%) developed epilepsy after the acute phase of encephalitis, a much higher proportion than in anti-NMDAR AE or anti-LGI1 AE (13). It was also found that patients with anti-GABABR AE were more likely to have acute symptomatic seizures than patients with anti-NMDAR AE and that the occurrence of SE was more frequent in anti-GABABR AE (19). Yao et al. reported that patients with anti-NMDAR AE, anti-LGI1 AE, and anti-Caspr2 encephalitis all achieved effective seizure control within 2 years after concurrent immunotherapy and ASMs, whereas only 55% of patients with anti-GABABR AE had effective seizure control at the last follow-up visit after 2 years (21). Zhu F et al. reported 14 patients with anti-GABABR encephalitis: 5 patients received hormonal shock therapy, 3 patients received gamma globulin therapy, and 6 patients received hormonal shock therapy and gamma globulin. All 14 patients were on ASM, which resulted in a reduction in the number of seizures and varying degrees of improvement in both psychobehavioural disturbances and cognitive decline, but seizure freedom could not be achieved with ASM alone (36). Hainsworth JB et al. reported a case of anti-GABABR encephalitis with refractory SE that occurred with multiple ASMs and antiepileptic narcotics; after treatment with methylprednisolone, intravenous immunoglobulin, plasma exchange and rituximab, i.e., after the end of the rituximab course, the patient stopped the narcotics, had no seizure recurrence, showed significant progressive neurological improvement and was able to perform all activities of daily living independently with a normal EEG 6 months after the visit; subsequently, mycophenolate mofetil as a maintenance immunosuppressant and the antiepileptic regimen was slowly reduced to LEV monotherapy (37). Seizures in patients with anti-GABABR AE can be partially alleviated by immunotherapy combined with ASMs (10, 13, 21, 36, 37). Data for the other 40 patients are shown in Table 7.

Pondrelli F et al. reported 2 cases of anti-VGKC autoimmune encephalitis (anti-VGKC AE), 1 with steroid immunotherapy supplemented with ASM (LEV) with improved outcome, and the other with steroid and imidazolothioprine immunotherapy supplemented with ASM, achieving no seizures (38). Dubey D et al. reported 8 cases of seizure-resistant anti-VGKC AE, most of which used immunotherapy supplemented with a median of 2 ASM, resulting in a 50% reduction in seizures after treatment in 5 (62.5%) cases (39). Anti-Caspr2 AE is a commonly used immunotherapy and ASM (38, 40). The common ASM used is LEV, and other ASMs include sodium channel blockers, VPA and LTG. The treatment results were good, and most patients achieved seizure-free status. Other cell surface antibodies that mediate encephalitis include antibodies against metabotropic glutamate receptor 5 [mGluR5] and voltage-gated calcium channel [VGCC]. They were all treated with immunotherapy and ASMs, and anti-mGluR5 encephalitis was also treated with antitumour therapy. ASMs included MDZ, Fos-PHT and TPM, and the response was good (41, 42). Data for these patients are shown in Table 7.

Based on follow-up observations of seizures in anti-GABABR AE, immunotherapy combined with ASMs was found to be effective in some patients. Patients with anti-GABABR AE are more likely to have acute symptomatic seizures and are more likely to have SE than patients with other common forms of AE, leading to an increased likelihood of long-term ASM use. The duration of ASM administration should be determined by considering all factors, and multimodal immunosuppressive therapy may be considered to improve seizures in patients who develop refractory SE. Seizures in patients with anti-VGKC AE or anti-Caspr2 AE can be mostly alleviated by immunotherapy combined with ASMs. Anti-VGKC AE is not considered definite AE but rather probable or possible antibody-negative AE. The search for VGKC-complex Ab is no longer advised, and patients should be tested for Ab against specific targets (CASPR2 and LGI1) instead (43–45).

The number of studies on other cell surface antibody-mediated AEs is currently low.