Introduction
Infantile spasm (IS) syndrome is a severe epileptic encephalopathy caused by multiple etiologies. Early and appropriate intervention is critical for better seizure and neurodevelopmental outcomes. Reference Pavone, Polizzi and Marino1 Hormonal therapies (adrenocorticotropic hormone [ACTH] or prednisolone) and vigabatrin are the commonly used first-line therapies. Reference Go, Mackay and Weiss2 The published experience with high-dose (8 mg/kg/day) prednisolone is limited. Reference Eliyan, Heesch and Alayari3,Reference Hussain, Shinnar and Kwong4 This study aimed to study the short-term seizure outcomes following treatment with 8 mg/kg/day prednisolone in children with IS refractory to vigabatrin. We hypothesized that high-dose prednisolone may result in comparable electroclinical remission rates in children with IS when compared to published ACTH rates.
Methods
This retrospective study recruited children seen in the IS clinic at a tertiary-level teaching pediatric hospital between July 2019 and January 2021. All consecutive children with hypsarrhythmia or hypsarrhythmia variant on EEG with/without parentally reported IS, who had been treated with vigabatrin as first-line anti-seizure medication (ASM) followed by high-dose oral prednisolone in cases who did not respond to vigabatrin, were included. The study was approved by the Institutional Ethics Board (ref number 1000072786).
Institutional Infantile Spasm protocol
All children with suspected IS got an urgent 1-hour EEG (with bilateral deltoid EMG) including one sleep–wake cycle. If hypsarrhythmia or hypsarrhythmia variant was detected on EEG, patients were started on vigabatrin and then were referred to the IS clinic. Detailed physical examination coupled with urgent neuroimaging (MRI brain) was done (if already not done). A targeted genetic testing was done in a suggestive phenotype. The children with unexplained etiology underwent extensive metabolic testing and microarray. If these were negative, an infantile epileptic encephalopathy gene panel was sent. The need for whole-exome sequencing was then individualized.
Vigabatrin was started at 50 mg/kg/day and rapidly titrated up (Day 1: 50 mg/kg/day BID; Day 2: 100 mg/kg/day BID; Day 3: 125 mg/kg/day BID; Day 4 and afterward: 150 mg/kg/day BID). If the patient had a partial or complete electroclinical response (ECR) to vigabatrin at 2 weeks, it was continued for 6 months. If no response was noted or the child had intolerable side effects, it was rapidly weaned off in 1 week.
If patients failed vigabatrin therapy, oral prednisolone was started at 8 mg/kg/day (maximum 60 mg/day). The ECR was assessed after 2 weeks. If the patient responded well, it was weaned off over the next 2 weeks. If no response, it was weaned off in 5 days. Blood pressure and urine glucose were monitored weekly and blood work was done every 2 weeks while on prednisolone. A complete ECR was defined as complete cessation of IS as per parental report and resolution of hypsarrhythmia on EEG.
The therapy in children who failed prednisolone was often individualized by the treating neurologist and frequently included topiramate, clobazam, and ketogenic diet. Cases that were amenable to surgery were urgently referred to the epilepsy surgery program.
Objectives
The primary objective was to study the proportion of children with IS, who had failed vigabatrin, and then who achieved ECR (cessation of clinical spasms and resolution of hypsarrhythmia) at 2 weeks and whom this was sustained at 6 weeks (sustained ECR) after the start of oral prednisolone. Other objectives included proportion of children with IS who relapsed after initial response to oral prednisolone, tolerability of high-dose prednisolone, proportion of children with clinical and ECR (cessation of clinical spasms and resolution of hypsarrhythmia) at 2 weeks after the start of vigabatrin, and proportion of children who achieved sustained ECR (cessation of clinical spasms for at least 28 d and resolution of hypsarrhythmia) at 6 weeks after the start of vigabatrin.
Data Extraction
The electronic records of all the consecutive cases seen in the IS clinic who fulfilled the inclusion criteria were studied. Study data were collected, stored, and managed using the Research Electronic Data Capture (REDCap) secure electronic data capture tool. Reference Harris, Taylor and Minor5,Reference Harris, Taylor and Thielke6 Variable collected included age, gestation/birth weight, details of the antenatal period, presence/absence of perinatal HIE, age at onset of spasms, lead time from apparent onset of spasms to treatment, developmental status at the onset of spasms (normal, delayed, regression), family history, details of EEG at time of diagnosis, at 2 weeks, at 4 weeks, at 6 weeks, and at later time points (hypsarrhythmia, hypsarrhythmia variant, normal, focal interictal epileptiform discharges [IEDs], others), treatment response to vigabatrin and oral prednisolone, other therapies used afterward (topiramate, ketogenic diet, epilepsy surgery), details of MRI brain, and genetic testing.
Statistics
Descriptive statistics were used to describe various study variables. A logistic regression model was used to examine the effect of variables (age at spasm onset, delay from spasm onset to initiation of treatment, sex, normal, or abnormal development at spasm onset, structural or nonstructural etiology) on ECR at 2 weeks after vigabatrin and prednisolone.
Results
Seventy-three cases were identified in the IS database during the study period. Eight children were excluded – 2 children with trisomy 21 received prednisolone as first line (both had incomplete ECR on prednisolone and responded to vigabatrin); 3 children (1 – Krabbe’s disease, 1 – ECHS1-related, 1 – unknown etiology) – prednisolone could not be given (after vigabatrin failure) due to active infection; two patients proceeded to surgery after vigabatrin failure (1 – left parieto-occipital polymicrogyria, 1 – perinatal stroke/COL4A1); parent of 1 child with bilateral periventricular leukomalacia (prematurity) refused prednisolone due to ongoing COVID-19 pandemic. Thus, 65 cases were included. The baseline characteristics are summarized in Table 1. Twenty-three patients (35.4%) were on at least one ASM before starting vigabatrin.
Table 1: Salient baseline characteristics of the study participants
IQR = Inter-quartile range; SD = Standard deviation
* The dose of prednisolone for a child with a weight of 7.5 kg (at 8 mg/kg/day) would be 60 mg (highest daily dose used in this study). Children > 7.5 kg would have received < 8 mg/kg/day daily prednisolone
Two children (1 – neonatal hypoglycemia, 1 – methylmalonic acidemia) had no IS; vigabatrin was initiated after the detection of hypsarrhythmia on EEG.
Etiology
Most common etiology was genetic in 25 cases (38.5%): Chromosomal [6 trisomy 21, one each of dup (2) (q11.2q22.1) and 16p13.11 del], single-gene disorders [3 TSC2, 3 KCNQ2, 1 each of others (TSC1, TUBA1A, TUBB2B, SLC35A2, PRUNE1, GRIN2A, TBCK, GMPPB, STXBP1, CASK, DNM1]. Structural causes were seen in 16 cases (24.6%): 5 perinatal strokes, 3 perinatal hypoxic-ischemic encephalopathy (HIE), 2 polymicrogyria (one had ARX gene mutations), 1 lissencephaly [del (17) (p13.2p13.3)], 1 neonatal hypoglycemia sequalae, 1 periventricular leukomalacia (prematurity), 1 hemimegalencephaly, 1 temporo-occipital cortical dysplasia, and 1 Aicardi syndrome. Two cases (3.1%) had metabolic (1 – Zellweger syndrome, 1 – methylmalonic acidemia) etiology. Rest (22, 33.8%) had unknown etiology.
Vigabatrin
Vigabatrin was started in 65 children: 63 with clinical spasms and EEG showing hypsarrhythmia, or hypsarrhythmia variant; 2 children (1 – neonatal hypoglycemia sequalae, 1 – methylmalonic acidemia) had hypsarrhythmia on EEG and no clinical spasms. At 2 weeks after starting vigabatrin, 23/63 (36.5%) had complete clinical response; however, ECR was seen in 20/65 (30.8%; 95% confidence interval [CI]: 20.9%, 42.8%) patients. This response was persistent at 6 weeks with no subsequent relapses. Both children with hypsarrhythmia but no clinical spasms did not show resolution of hypsarrhythmia on vigabatrin.
The logistic regression model showed that only normal development at spasm onset was associated with an ECR with vigabatrin at 2 weeks (β = 1.29, p = 0.04; adjusted OR = 3.66) (Table 2).
Table 2: Multivariable logistic regression model evaluating the relationships between salient patient variables and 2-week electroclinical response to vigabatrin (A) or prednisolone (B)
Coding: Development (0 – Abnormal, 1 – normal), Etiology (0 – structural, 1 – nonstructural), Sex (0 – female, 1 – male)
High-dose Prednisolone
High-dose prednisolone was used in 45 children with spasms who failed (absent or incomplete ECR) vigabatrin. The mean daily dose was 56.4 mg/day (range: 45–60 mg); 32 children received 60 mg/day of prednisolone. At 2 weeks, 37/45 (82.2%) had a complete clinical response. ECR was seen in 35 children (77.8%; 95%CI: 63.7%, 87.5%). This was sustained at 6 weeks in 30 patients (66.7%). Five children (5/35, 14.3%) had a relapse of IS during/after weaning prednisolone. Two children were again started on high-dose prednisolone; one responded well and the other remained refractory. Other three children were switched to topiramate. One of the two children with hypsarrhythmia but no clinical spasms showed resolution of hypsarrhythmia on prednisolone.
It was generally well tolerated. The common adverse effects reported were irritability (15/45, 33.3%), weight gain (10/45, 22.2%), sleep disturbances (6/45, 13.3%), and hypertension requiring transient amlodipine treatment (2/45, 4.4%). Two patients were hospitalized during the weaning period: one child with GMPPB mutation presented with apneas and electrographic seizures (died after 1 month due to multiple organ dysfunction) and one child with STXBP1-related disorder with frequent dystonic episodes. One child with trisomy 21 was hospitalized within 1 week of stopping steroids due to dehydration and hypoglycemia.
The logistic regression model showed that only lesser delay from spasm onset to treatment initiation was associated with an ECR with prednisolone at 2 weeks (β = −0.03, p = 0.02; adjusted OR = 0.97) (Table 2).
Other Therapies
The mean follow-up duration was 13.6 months (standard deviation [SD] 6.6 months). The median number of ASMs at the last follow-up was 1 (inter-quartile range [IQR]:1,2).
Topiramate was used in 25 children. Three children were already on topiramate at the time of starting vigabatrin. In the rest, it was initiated at a mean age of 9.3 months (SD 4.1). Fifteen (15/22) had previous incomplete ECR to or relapse on prednisolone; it was started for new seizure types in the rest. Six (6/22, 27.3%) had complete seizure cessation with topiramate.
Ketogenic Diet
In this cohort, eight patients were started on a classical ketogenic diet at a mean age of 13.3 months (SD 4.7 months). All of them (8/8) had failed vigabatrin. Five out of eight patients had also failed prednisolone in the past. Out of the three prednisolone responders, one relapsed, one evolved into other seizure types, and one presented with electrographic status epilepticus. Six patients had ongoing epileptic spasms at the time of diet initiation; 5/6 had other ongoing seizure types as well.
Six patients were still on diet at the last follow-up. One patient stopped diet after 17 months due to feeding intolerance and irritability. One child died after 1 month of starting the diet. The median follow-up available was 5 months (IQR: 3.8, 9). Six (6/8, 75%) patients had < 50% seizure reduction. Only one patient had > 90% seizure reduction (trisomy 21). Two out of six patients with ongoing spasms had >50% seizure reduction (patients 1 and 8 in Table 3). Two patients are awaiting initiation of the classical diet (1 – TUBA1A, 1 – HIE). Details are summarized in Table 3.
Table 3: Characteristics of study participants who received classical ketogenic diet
ASMs = Anti-seizure medications; B6 = Pyridoxine; CBD = Cannibidiol;CBZ = Carbamazepine; cloba = clobazam; FIAS = Focal impaired awareness seizures; KD = Ketogenic diet; LEV = levetiracetam; MISF = Multiple independent spike foci; MODS = Multiorgan dysfunction syndrome; PHB = Phenobrabital; RUF = Rufinamide; SE = Status epilepticus; TPA = Topiramate; VGB = vigabatrin; VPA = Valproate
Epilepsy Surgery
Three patients had subsequent epilepsy surgery; all seizure free at the last follow-up (Table 4). Four patients are awaiting surgery (three perinatal strokes, one extensive left temporo-occipital polymicrogyria).
Table 4: Characteristics of study participants who underwent epilepsy surgery
FH = functional hemispeherectomy; MCA = Middle cerebral artery
Discussion
Summary of Results
Short-term seizure outcomes following treatment were studied in our cohort of IS. Nearly one-third of patients (38.5%) had genetic etiology. Complete ECR was seen in 30.8% of patients 2 weeks after vigabatrin. This was 77.8% 2 weeks after prednisolone initiation in children who failed vigabatrin. ECR was sustained at 6 weeks after prednisolone in 66.7% cases. Prednisolone was generally well tolerated.
High-dose Prednisolone
Few studies have reported seizure outcomes after the use of 8 mg/kg/day oral prednisolone. Eliyan et al. Reference Eliyan, Heesch and Alayari3 treated 102 children with 8 mg/kg/day of prednisolone as first-line treatment; complete ECR was seen in 59% cases. Among the prednisolone nonresponders, 33% of children responded to ACTH. The significant adverse effects noted with prednisolone were hypertension requiring anti-hypertensives (3), hospitalization due to known or suspected infections, (4) and hyperglycemia (1). In a smaller retrospective cohort Reference Hussain, Shinnar and Kwong4 , complete ECR was seen in 63% (17/27) following prednisolone as first-line therapy.
One prospective study from South Korea Reference Ko, Youn and Chung7 reported ECR in 59.1% cases following 40–60 mg/day of prednisolone in cases who had failed vigabatrin. The adverse effects were common but did not warrant discontinuation of therapy. ECR was seen in one out of five cases (20%) following 4–6 mg/kg/day of prednisolone (after vigabatrin failure) in another smaller study. Reference Jones, Snead and Boyd8 In our study, the ECR occurred in 77.8% cases after prednisolone in cases who had failed vigabatrin. Lesser delay from spasm onset to treatment initiation predicted better response to prednisolone. The rate of ECR is much higher than the reported rates in the sparse literature. We also found adverse effects commonly, but none warranted cessation of therapy. Other studies reporting on outcomes with high-dose prednisolone (> 2mg/kg/day) are summarized in Table 5.
Table 5: Summary of studies using high-dose oral prednisolone/prednisone in children with infantile spasms
KD = Ketogenic diet; LEV = levetiracetam; NR = Not reported; RCT = Randomized controlled Trial; TPA = Topiramate
* Repeat EEG are not done in five children
** Prior exposure to hormonal therapy and vigabatrin was observed among 12% and 35% of patients, respectively
@ Subsequently, 40% (4/10) of prednisolone nonresponders exhibited a complete response after an additional 2-week course with ACTH
# Eleven had a previous ketogenic diet
## 22/66 (33.3%) had ECR with VGB
$ 37/57 (64.9%) had ECR with VGB
Further, Chellamuthu et al Reference Chellamuthu, Sharma and Jain13 showed that high-dose prednisolone (4 mg/kg/day) achieved higher clinical remission in significantly more patients when compared to low-dose (2 mg/kg/day) prednisolone (51.6% vs. 25%, p = 0.03) though electroclinical remission was similar (38.7% vs. 21.9%, p = 0.15).
Prednisolone versus ACTH
Few randomized controlled trials have compared higher doses of prednisolone with ACTH. One trial Reference Wanigasinghe, Arambepola and Sri Ranganathan11 showed that prednisolone (40–60 mg/day) was superior to ACTH; other two trials showed similar efficacy. Reference Lux, Edwards and Hancock10,Reference Gowda, Narayanaswamy and Shivappa12 ECR to prednisolone was seen in 26.7% Reference Gowda, Narayanaswamy and Shivappa12 –71.4% Reference Lux, Edwards and Hancock10 patients in these trials. Besides these, the International Collaborative Infantile Spasm study (ICISS), Reference O’Callaghan, Edwards and Alber18 which explored the effectiveness of hormonal therapy and hormonal therapy/vigabatrin combination for IS, did randomize patients into high-dose prednisolone and ACTH. However, the published data did not report effectiveness outcomes separately for ACTH and prednisolone. Further, a recent systematic review and meta-analysis (six trials) Reference Li, Zhong and Hong19 showed that prednisolone/prednisone (all doses) elicits similar ECR as ACTH. Compared with prednisolone/prednisone, ACTH/tetracosactide was not superior in terms of cessation of spasms at day 14 (relative risk 1.19, 95% CI 0.74–1.92), day 42 (relative risk 1.02, 95% CI 0.63–1.65), and resolution of hypsarrhythmia on electroencephalogram (relative risk 1.14, 95% CI 0.71–1.81). This was further supported by another systematic review and meta-analysis which included both randomized and non-randomized studies. Reference Sánchez Fernández, Amengual-Gual and Gaínza-Lein20 Thus, oral prednisolone (especially high dose) may be a reasonable alternative to ACTH, given the ease of administration and lower cost coupled with similar effectiveness/adverse effect profile. In a recent cost-effectiveness study, prednisolone at 4–8 mg/kg/day was found to be more cost-effective than ACTH. Reference Sánchez Fernández, Amengual-Gual and Gaínza-Lein20
Other Therapies
Vigabatrin resulted in clinical spasm remission in 36.5% cases and ECR in 30.8% cases in this study. Data from past randomized controlled trials is variable; ECR rates have varied from 15.9% to 55.6%. Reference Lux, Edwards and Hancock10,Reference Appleton, Peters and Mumford21–Reference Vigevano and Cilio24
Ketogenic diet was tried in eight patients in our cohort; six had ongoing spasms at the time of diet initiation. Overall, only 2/8 had >50% seizure reduction at the last follow-up. Two recent controlled trials show that 54.1% Reference Zhang, Chen and Wang25 and 65.2% Reference Sharma, Goel and Kapoor26 children showed >50% spasm reduction with ketogenic diet and modified Atkins diet, respectively, in children with epileptic spasms refractory to hormonal therapy.
Epilepsy surgery was done in three cases (two malformations of cortical development, one perinatal stroke) in our cohort; all had excellent short-term seizure outcomes. Similar favorable seizure and neurodevelopmental outcomes have been reported in select cases of drug-resistant IS. Reference Iwatani, Kagitani-Shimono and Tominaga27,Reference Loddenkemper, Holland and Stanford28
Strengths and Limitations
Our study reported the short-term seizure outcomes in a cohort of IS treated at our institution in a protocolized fashion. Although the data collection was retrospective, detailed electronic records were maintained prospectively in the IS clinic. The sample size was relatively small. This study was also limited by the short follow-up period and absence of serial structured neurodevelopmental assessments.
Conclusions
High-dose (8 mg/kg/day) oral prednisolone resulted in complete ECR (at 2 weeks) in more than three-fourth of children with hypsarrhythmia or hypsarrhythmia variant on EEG with/without parentally reported IS who had failed vigabatrin. This was sustained at 6 weeks in two-thirds of the children. It was generally well tolerated and found to be safe. It may be a reasonable and more feasible alternative to ACTH for the treatment of IS. Further trials exploring the comparative effectiveness of 8 mg/kg/day prednisolone with ACTH, dose–response relationship with prednisolone, and exploring various drug combination therapies to improve efficacy are warranted.
Conflicts of interest
None.
Ethics
The study was approved by the Institutional Ethics Board (ref number 1000072786).
Statement of Authorship
All authors provided clinical care to the study participants. WAS collected the data. PJ analyzed the results. WAS and PJ wrote the first draft of manuscript which was then reviewed and approved by all the authors.
