Keywords
west syndrome - infantile spasms - pulsatile therapy - dexamethasone - corticosteroids
Introduction
Infantile spasms (IS) are an age-specific epilepsy syndrome usually occurring in the first 2 years of life. The disorder manifests clinically with epileptic spasms (ES) and is most commonly associated with a hypsarrhythmia (HA) pattern on electroencephalography (EEG).[1]
[2]
[3] In infants presenting with both ES and HA, the IS syndrome is also known as West syndrome (WS). IS and WS are associated with evolving developmental delay and a poor long-term neurodevelopmental prognosis.[2]
[4] While the etiology of IS can be classified as either genetic, structural/metabolic, or unknown,[5]
[6] the pathophysiology and precise mode of action of current therapies still remain poorly understood. In 1958, Low first described the beneficial effect of hormonal treatment on children with IS, with improvement of both clinical and electroencephalographic abnormalities.[7] Currently, the mainstay therapy approach for IS consists of hormonal treatment (corticotropin [ACTH] or corticosteroids), and/or the anticonvulsant vigabatrin.[8]
[9] Notably, except for vigabatrin, there is no sufficient evidence for the use of other antiseizure medication (ASM).[8]
[10] Despite many studies investigating the therapeutic use of corticosteroids in managing IS, there are still uncertainties about the best treatment approach.[11] Severe adverse reactions due to the long-term application of hormonal therapeutics are a significant limitation. Hence, pulsatile treatment regimens have been assessed in the hope of lowering the complication rate, prolonging the duration of treatment, as well as improving clinical outcomes. To date, there is no evidence indicating which corticosteroid is more effective at treating IS, and not enough data to assess the role of pulsatile treatment.[8] Thus, the aim of this retrospective study was to analyze the efficacy and tolerability of oral pulsatile dexamethasone treatment in children with diagnosed IS or WS. In particular, we evaluated short-term electroclinical responses, long-term neurodevelopment outcomes, and adverse reactions to the therapy. The impact of a prolonged initial pulse was also assessed.
Methods
Patients and Treatment Regime
This retrospective study included 26 patients with IS or WS who were treated with orally administered, pulsatile high-dose dexamethasone at the University Children's Hospital Jena between 2002 and 2021. Patients with tuberous sclerosis or prior treatment with ACTH were excluded. Data were extracted from the local database.
After baseline clinical assessment, EEG recording, and exclusion of contraindications, patients were treated orally for 5 to 7 days with dexamethasone at a daily dosage of approximately 1 to 1.5 mg/kg. The electroclinical response was then evaluated, and an individual decision was made on whether to extend the first pulse for another 5 to 7 days. Patients showing no or an unsatisfying clinical (e.g., persistence of ES, insufficient lowering rate of seizure frequency) or electroencephalographic response (e.g., persistence of epileptogenic discharges or HA in the EEG) after the initial application of dexamethasone received an extended pulse. The decision was made purely clinically. Subsequent pulses were administered every 4 weeks for 5 days. If the response was favorable, defined as stabilization of both seizures and EEG findings, the intervals between the pulses were extended to 6 weeks. The therapy was initially administered and monitored in the hospital, however, later patients were treated on an outpatient basis if the therapy was well-tolerated and the patient had no major adverse reactions. In the final phase, the number of treatment days per pulse was tapered by 1 day with the last pulse consisting of only a 1-day application. Ranitidine was given during the therapy for gastric acid suppression. Patients who did not respond to the therapy were switched to an alternative treatment such as ASM or a ketogenic diet.
Clinical Data and Short- and Long-term Outcome Measures
Pretreatment data collected included gestational age, birth weight, age at IS onset, initial frequency of spasms, as well as number and type of previously given ASM before dexamethasone treatment. All patients underwent the usual diagnostic investigations to determine a possible etiology of the IS, including clinical examination, EEG recording, MRI, and if indicated, genetic and metabolic investigations. Patients were grouped as etiology (either structural/metabolic or genetic origin) or unknown.
The parameters monitored during in-patient therapy included blood pressure, blood glucose, body weight, seizure frequency, and any adverse reactions to therapy. EEG before, during, and after the dexamethasone pulses were analyzed by different, not blinded investigators. Caregivers were instructed to monitor the children regarding seizure frequency and adverse reactions after hospital discharge. This information was collected during readmission to the hospital or at outpatient clinic visits for the next pulse. The number of administered pulses and the total duration of treatment were also monitored.
As short-term electroclinical outcome measures, we evaluated the time from dexamethasone therapy to the initial clinical and EEG response, as well as the quality of each initial response. The clinical response was categorized into one of three groups: complete remission of spasms, reduction of seizure number, and no response up to the third pulse. Complete remission of spasms was defined as a cessation of spasms whereas no initial response was defined as no change in seizure frequency up to the third pulse.
The EEG response was graded as either a good response, a response, or no response. Patients were considered to have a good response if complete resolution of the HA and/or other epileptiform patterns were observed. Cessation of HA in patients with WS who showed persisting epileptiform patterns or a reduction of epileptiform patterns in patients without HA was documented only as a response. No response was defined as no change in EEG effect up to the third pulse. Finally, the time from initiation of therapy to first clinical and EEG relapse was documented, measured in the number of pulses with one pulse equivalent to a time interval of 4 to 6 weeks.
As long-term outcome measures, we analyzed the neurocognitive development and epileptic outcome of each patient at the last follow-up in our pediatric neurology department. We classified the neurocognitive outcome as (1) normal or slightly impaired for patients with an IQ of ≥70, (2) mild to moderately impaired for patients with an IQ between 35 and 69, or for patients who were educated in schools and kindergarten for people with disabilities and educationally handicapped, and (3) severely impaired for those with an IQ of ≤34. Patients required follow-up at least until their third year of life for classification. The long-term epileptic outcome was grouped as (1) no epileptic disorder, (2) persisting epileptic disorder but seizure-free under therapy with ASM, or (3) persisting epileptic disorder with drug-refractory seizures. Epileptic disorders were not stratified into seizure semiologies or epilepsy syndromes.
This study was approved by the Ethics Committee of the Medical Faculty at the Friedrich Schiller University of Jena.
Analysis
Statistical analysis was performed with SPSS version 26. Fisher's exact test was used to determine whether initial clinical and EEG response, cognitive development outcome, and epileptic outcome were associated with factors such as etiology, availability of HA, or duration of initial pulses. Kruskal–Wallis test was used to differentiate whether significant differences in the short- and long-term outcome parameters were associated with time to therapy with dexamethasone. We used Kaplan–Meier curves and the log-rank test to visualize and analyze whether the time to a clinical relapse during therapy, using pulses as a measure of time, was influenced by etiology or therapy duration. Statistical significance was defined as p < 0.05 (two-tailed significance). An exception was made in analyzing the neurocognitive outcome where we used a one-tailed significance test (p < 0.05).
Results
Patient Characteristics
A total of 26 patients (13 females) with IS were included in the study; 19 (73.1%) with WS and the remaining 7 patients all with abnormal EEG backgrounds. The median age of ES onset was 5.5 months (interquartile range [IQR] 4–8). Half of the patients (n = 13) had a known etiology; 9 cases (34.6%) were structural/metabolic and 4 were genetic (15.4%). Most of the patients (n = 17/25, 68%) received a prolonged initial pulse; approximately two-thirds of these patients (n = 11, 64.7%) had a known etiology. The median time from spasm onset to diagnosis was 2 weeks (IQR 0–6) and to dexamethasone treatment 9 weeks (IQR 5–13.25). All infants were initially treated with ASM or other drugs explaining the delay to dexamethasone treatment in the study; most patients (18/24, 70.8%) had previously received two or more drugs. Patient demographics and characteristics are shown in [Table 1]. Notably, nearly all patients (n = 22) received an add-on therapy with ASM during the therapy with dexamethasone; only two patients did not, and for two further cases no information was available. The most frequently used add-on medications were vigabatrin (n = 15, 60%) and valproate (n = 12, 48%; [Table 2]).
Table 1
Patient demographics and clinical characteristics
|
Characteristics
|
Value
|
|
Sex (n/N) [%]
|
|
Female
|
13/26 [50%]
|
|
Male
|
13/26 [50%]
|
|
Median age at IS onset (months) [IQR]
|
5.5 [4–8]
|
|
Mean birth weight (grams), [SD]
|
2,952.4 [774.1]
|
|
Median lead time to diagnosis (weeks) [IQR]
|
2 [0–6]
|
|
Median lead time to treatment (weeks) [IQR]
|
9 [5–13.25]
|
|
Median number of ASM before treatment with dexamethasone [IQR]
|
3 [2–4]
|
|
Etiology (n/N) [%]
|
|
Unknown
|
13/26 [50%]
|
|
Known
|
13/26 [50%]
|
|
Structural/metabolic
|
9/26 [34.6%]
|
|
Genetic
|
4/26 [15.4%]
|
|
Hypsarrhytmia (n/N) [%]
|
19/26 [73.1%]
|
|
Number of ASM prior treatment with dexamethasone (n/N) [%][a]
|
|
≤ 2
|
7/24 [29.2%]
|
|
> 2
|
17/24 [70.8%]
|
|
Prolonged initial dexamethasone pulse n/N [%][b]
|
17/25 [68%]
|
|
Known etiology
|
11/17 [64.7%]
|
|
Unknown etiology
|
6/17 [35.3%]
|
|
Mean lead time to initial clinical response (weeks) [SD]
|
0.72 [2.19]
|
|
Relapse rate after initial cessation of ES
|
38.9%
|
|
Mean lead time to initial electroencephalographic response (weeks) [SD]
|
3.3 [3.6]
|
|
Mean number of pulses [SD]
|
10.8 [6.0]
|
|
Mean duration of treatment with dexamethasone (months) [SD]
|
11.5 [7.4]
|
|
Mean follow-up time (years of life) [SD][a]
|
6.2 [4.0]
|
Abbreviations: ASM, antiseizure medication; ES, epileptic spasm; IQR, interquartile range; IS, Infantile spasm; SD, standard deviation.
a Unknown in two patients.
b Unknown in one patient.
Table 2
Administration of antiseizure medication during dexamethasone therapy
|
Antiseizure medication
|
Number of patients
|
|
Vigabatrin
|
15 (60%)
|
|
Valproate
|
12 (48%)
|
|
Benzodiazepines
|
3 (12%)
|
|
Levetiracetam
|
2 (8%)
|
|
Oxcarbazepine
|
2 (8%)
|
|
Topiramate
|
2 (8%)
|
|
Ethosuximide
|
1 (4%)
|
|
Zonisamide
|
1 (4%)
|
Short-term Clinical and Electroencephalography Response
Almost all patients (25/26, 96.2%) showed some initial clinical response to the therapy up to the third pulse, however, most (n = 17/25, 68%) received a prolonged initial pulse. The mean lead time to a clinical response was 0.7 weeks (standard deviation [SD] 2.2). The treatment algorithm is demonstrated in [Fig. 1]. Altogether 18 (69.2%) patients achieved initial seizure freedom, a further 7 patients (26.9%) showed a reduction in seizure frequency, and only 1 patient (3.8%) showed no initial response to therapy until the third pulse. A higher percentage of patients treated with a prolonged initial pulse were initially seizure-free (14/17, 82.4%) than patients treated with a standard initial pulse (4/8; 50%). Of the 18 patients who were initially seizure-free, 7 (38.9%) relapsed after the first pulse. Favorably, a sustained cessation of ES was achieved in 11 (44%) cases after the initial pulse and in a further 5 cases (total n = 16/25, 64%) after the third pulse. In sum, three patients with a known etiology additionally became sustained seizure-free after the third pulse (n = 6/12, 50%) which equals a doubling in this cohort in comparison to seizure freedom after the first pulse (3/12, 25%). Although patient 8 received only two courses of dexamethasone due to loss of follow-up, we counted him as a non-responder. The quality of the clinical response did not differ significantly between etiological subgroups (p = 0.38) and was not associated with the presence of HA (p = 0.13), duration of the initial pulse (p = 0.16), or lead time to dexamethasone therapy (p = 0.80).
Fig. 1 Treatment algorithm. *missing data in one case. DXM, dexamethasone; IS, infantile spasms; WS, West syndrome.
In our study population, altogether 22 patients showed improved EEG findings on an average of 3.3 weeks (SD 3.6) after the first pulse of dexamethasone. A good response was seen in 40% (n = 10/25) of the cases, some response in 48% (n = 12/25), and no response until the third pulse in 3 (12%) infants. Two patients (20%) with an initially good EEG response relapsed during the treatment. The initial EEG response was associated with etiology (p = 0.028) and duration of the first pulse (p = 0.011). Patients with an unknown etiology and without a prolonged initial pulse were more likely to have a good response. Conversely, the initial EEG response was not associated with the presence of HA (p = 0.32) or lead time to therapy (p = 0.19). All important outcome parameters are shown in [Table 3].
Table 3
Short- and long-term outcome parameters
|
Known etiology
(n = 13)
|
Unknown etiology
(n = 13)
|
p- value
|
Prolonged initial pulse
(n = 17)
|
Standard initial pulse
(n = 8)
|
p-value
|
|
Short-term outcome
|
|
Clinical response n/N [%]
|
|
Seizure free
|
10/13
|
8/13
|
0.38
|
14/17
|
4/8
|
0.16
|
|
Reduction of seizure number
|
2/13
|
5/13
|
3/17
|
4/8
|
|
No response
|
1/13
|
0/13
|
–
|
–
|
|
EEG response n/N [%][a]
|
|
Good response
|
2/12
|
8/13
|
0.028
|
5/16
|
5/8
|
0.011
|
|
Response
|
9/12
|
3/13
|
11/16
|
1/5
|
|
No response
|
1/12
|
2/13
|
0/16
|
2/8
|
|
Cessation of spasms n/N [%][a]
|
|
After first pulse
|
3/12
|
8/13
|
0.11
|
8/16
|
3/8
|
0.68
|
|
After third pulse
|
6/12
|
10/13
|
0.23
|
10/16
|
5/8
|
1.0
|
|
Long-term outcome, n/N
|
|
Epileptic outcome n/N [%][b]
|
3/11
|
8/11
|
0.037
|
6/16
|
5/6
|
0.23
|
|
No epileptic disorder persisted
|
4/11
|
0/11
|
4/16
|
0/6
|
|
Epileptic disorder but seizure-free with ASM therapy
|
4/11
|
3/11
|
6/16
|
1/6
|
|
Neurodevelopment outcome n/N [%][c]
|
|
Normal or slightly impaired cognitive development
|
0/9
|
4/9
|
0.049
|
6/13
|
2/5
|
0.65
|
|
Mild to moderate cognitive impairment
|
4/9
|
2/9
|
5/13
|
1/5
|
|
Severe cognitive impairment
|
5/9
|
3/9
|
2/13
|
2/5
|
Abbreviations: ASM, antiseizure medication; EEG, electroencephalography.
a Missing data in one patient.
b Missing data in four patients.
c Missing data in six patients.
Relapses and Adverse Reactions
During the pulsatile therapy, the patient's clinical status was monitored to see if the status worsened over time or remained stable after any kind of response. The time to first relapse after therapy initiation stratified according to etiology and duration of the initial pulse is shown in [Fig. 2]. Time was measured in the number of pulses with the time interval between two pulses of approximately 4 to 6 weeks. Etiology (p = 0.025) but not a prolonged initial pulse (p = 0.84) was found to have a significant influence on remaining relapse-free. Patients with an unknown etiology were significantly more likely to remain clinically stable: the mean number of pulses till a relapse by unknown etiology was 16.4 pulses (SD = 2.7, 95% CI 11.0–21.8) versus 5.0 pulses (SD = 1.8, 95% CI 1.5–8.5) for a known etiology.
Fig. 2 Kaplan–Meier survival curves with a log-rank analysis comparing time to first relapse for cases with (A) known (mean 5.0 pulses, SD = 1.8, 95% CI 1.5–8.5) versus unknown (mean 16.4 pulses, SD = 2.7, 95% CI 11.0–21.8) etiology (p = 0.025). (B) standard (mean 7.9 pulses, SD = 1.9, 95% CI 4.2–11.5) versus prolonged (mean 9.9 pulses, SD = 2.7, 95% CI 4.6–15.1) initial pulse (p = 0.84). Time interval between 2 pulses: 4 to 6 weeks. SD, standard deviation.
Adverse reactions associated with therapy are shown in [Table 4]. Generally, adverse reactions were mild, and the therapy was well-tolerated. The most frequent reactions were mild infections (n = 21, 80.8%), for example, oral candidiasis, diarrhea, or upper airway infections, irritability (n = 10, 38.5%), and tiredness (n = 7, 26.9%). In 6 cases (23.1%) antibiotic therapy due to an infection was needed. Major adverse reactions such as severe systemic infections, adrenal insufficiency, sustained arterial hypertension, or death were not observed.
Table 4
Adverse reactions
|
Side effects
|
Number of patients (%)
|
|
Mild infection
|
21 (80.77%)
|
|
Irritability
|
10 (38.46%)
|
|
Tiredness
|
7 (26.92%)
|
|
Use of antibiotics due to infection
|
6 (23.08%)
|
|
Sleep disturbance
|
4 (15.38%)
|
|
Asymptomatic transient arterial hypertension
|
2 (7.69%)
|
|
Decreased appetite
|
2 (7.69%)
|
|
Cushingoid habitus
|
2 (7.69%)
|
|
Esophageal reflux
|
1 (3.85%)
|
|
Osteopenia
|
1 (3.85%)
|
|
Obstipation
|
1 (3.85%)
|
Long-term Outcome
Number of pulses, duration of treatment, and long-term treatment outcomes are shown in [Tables 1] and [3]. Four cases were lost to long-term follow-up. Additionally, neurocognitive development status was not investigated in two patients, and two patients were excluded due to young age. Among 18 patients with a mean follow-up of 6.2 years (SD 4.0) and 10.8 pulses (SD 6), 4 (22.2%) had a normal or slightly impaired neurocognitive development, 6 (33.3%) had mild to moderate impaired neurocognitive development, and 8 (44.4%) had a severely impaired neurocognitive development. The neurocognitive outcome was not influenced by initial clinical (p = 0.15) and EEG response (p = 0.367), duration of the initial pulse (p = 0.65), or presence of HA (p = 0.38). The proportion of patients with a normal or slightly impaired neurodevelopment outcome, however, was significantly higher in patients with an unknown etiology (p = 0.049).
Half of the patients with available data (n = 11) had no persisting epileptic disorders at the last follow-up, 4 (18.2%) patients remained seizure-free under therapy with ASM, and 7 (31.8%) infants had persisting drug-refractory epileptic disorders. Patients with an unknown etiology were more likely to have no epileptic disorder at the last follow-up than patients with known pathologies (p = 0.037). Similarly, infants with a good initial EEG response were more likely to be seizure-free at the last follow-up (p = 0.005). The epileptic outcome was not influenced by the duration of the initial pulse (p = 0.231), the initial clinical response (p = 0.27), or the presence of HA (p = 0.16). Time from clinical onset to therapy had no significant impact on long-term neurocognitive development (p = 0.35) nor epileptic outcome (p = 0.92). Patients characteristics may be seen in [Supplementary Table S1] (available in the online version only).
Discussion
Hormonal medication is a first-line treatment for patients with IS and many studies have evaluated the optimal treatment regime. After the United Kingdom Infantile Spasms Study (UKISS) trial demonstrated no significant difference in short-term outcomes between patients treated with ACTH or high-dose prednisolone, corticosteroids became an appropriate alternative in managing these epileptic disorders.[12] Since then, the so-called UKISS protocol has been investigated with different designs and slight deviations in the treatment protocol.[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21] The UKISS protocol consists of orally giving 40 to 60 mg prednisolone for 2 weeks followed by a tapering period over 15 days.[12] In addition, O'Callaghan et al discovered a better spasm control by combining hormonal treatment with vigabatrin in their so-called International Collaborative Infantile Spasms Study (ICISS) trial.[16] Due to the study design, the current treatment proposed by ICISS is still the standard treatment and corresponds to the German treatment guidelines. Some authors have also studied pulsatile treatment with intravenous methylprednisolone using various dosages of 20 mg/kg daily[22]
[23] or 30 mg/kg/d[24]
[25] for three consecutive days and tapering periods of 2 weeks to 2 months. Others have investigated pulsatile intravenously administered dexamethasone using dosages of 0.25 mg/kg on 7 days in 3 months[26] or 20 mg/m2 on 3 days every 4 weeks for at least five times.[27]
Over several years, the University Children's Hospital in Jena has administered an oral pulsatile dexamethasone treatment regime to infants with IS. Analyzing short-term outcomes in these patients, we found that clinical response to therapy occurred on average within 1 week of initiation, matching reports in other studies.[14]
[16]
[23]
[28] Furthermore, the clinical and EEG response rates in our study of 69% and 40%, respectively, are quite similar to clinical response rates (56–80%) and EEG response rates (43.8–64%) in studies using standard treatment with oral prednisolone and are also comparable to rates reported in trials of pulsatile medication regimes (33–64% and 50–65%, respectively).[12]
[13]
[14]
[16]
[17]
[21]
[22]
[23]
[24]
[25]
[27]
[29] The lack of EEG data for one patient and the comparatively long mean time to treatment of 11.8 weeks may explain the lower percentage of patients with a good EEG response in our cohort. Although there is evidence that a short lead time to treatment and an unknown etiology is associated with better initial spasm control,[16] our study does not support these findings. In the event of a poor initial response, we observed that continuation until the third pulse led to additional patients achieving sustained seizure remission. In particular, the number of seizure-free cases among patients with a known etiology doubled.
Notably, we found that infants with IS of unknown etiology were more likely to have a good initial EEG response to therapy, consistent with the findings of Gonzalez-Giraldo et al.[21] The more favorable EEG response in patients who did not receive a prolonged initial pulse may have been influenced by selection bias, as unsatisfactory responses to the initial treatment were a reason for receiving a prolonged pulse.
For patients who showed a response to the therapy in our study, the chances of having a further relapse over the course of the treatment were heavily influenced by the etiology. Infants with an unknown etiology were more likely to remain clinically stable than patients with a known underlying etiology (p = 0.025). Although significantly more patients with a known etiology received a prolonged first pulse (p = 0.03) due to unsatisfying seizure control, a prolonged initial pulse was not associated with a greater likelihood of a further relapse. This indicates that a prolonged initial pulse may lead to spasm control in patients with an unsatisfactory initial response.
A favorable neurodevelopment outcome is another important goal in the treatment of IS. The UKISS trial showed that a lower age of onset of spasms and a longer lead time to treatment is associated with poorer neurodevelopment outcomes at the age of 4 years. Otherwise, the authors demonstrated a higher score in the Vineland Adaptive Behavior Scales (VABS) in patients with an unknown etiology at 14 months and 4 years of age who got a hormonal treatment than those treated with vigabatrin alone.[30]
[31] O'Callaghan et al further showed a higher VABS in patients who achieved a primary clinical response at 18 months of age.[32] In a Finnish population-based study of 214 patients with IS and a follow-up of 20 to 35 years, a favorable neurocognitive outcome with a normal or slightly impaired intelligence (IQ ≥68) was achieved in 24% of the cases. Further, 36% had no epileptic seizures at the last follow-up in adulthood. Among other things, Riikonen underlines the high impact of etiology on the long-term prognosis of IS.[33] The proportion of cases with a normal or slightly impaired neurocognitive development of 22.22% in our study is similar. Our data support the fact that etiology influences neurocognitive outcome significantly as in other studies mentioned.[21]
[30]
[31]
[33] Apart from that, the rate of patients without an epileptic disorder at the last follow-up in our trial is higher than in the Riikonen study. This may be explained by a relatively higher proportion of patients with an unknown etiology. Having said that, we found statistically significant influences on epileptic outcomes by the underlying etiology and the initial electroencephalographic response. Patients with an unknown etiology and good initial response in the EEG were more likely to have no epileptic disorder at the last follow-up than those with a known etiology respectively with a worse initial electroencephalographic response. This also confirms the findings of Gonzalez-Giraldo et al.[21]
Despite difficulties in comparing data from different trials due to different study designs and populations, this study shows that our therapy regimen is safe and well-tolerated without severe complications. The most frequent side effects were infections, for example, oral candidiasis, mild diarrhea, and upper airway infections in 80.77% of the cases which is relatively high compared with other studies that show an infection rate of 0 to 41.7%.[12]
[13]
[14]
[15]
[16]
[19]
[21]
[22]
[23]
[24]
[25]
[26]
[34] Merely, the study of Haberlandt et al discovered a similar infection rate of 71.4% if the rate of virus infections and oral candidiasis count together.[27] Moreover, their treatment protocol is the most similar one to ours. This high rate of mild infections may be explained by the long mean time of treatment of approximately 11 months in our study. Furthermore, upper airway infections are common in preschool children and usually appear several times per year.[35]
In sum, this study shows that our pulsatile treatment regime with dexamethasone is safe and may lead to comparable short-term and long-term outcome parameters as in prescribed treatment regimes, despite long lead-time to treatment and prior drug-refractory seizures. The underlying etiology is the most influencing factor affecting several outcome parameters. Nevertheless, an extension of the first pulse in cases of unsatisfying initial electroclinical response has no significant impact on important outcome parameters. However, it may lead to comparable spasm control between those patients with a less favorable initial response and those with a good one. This study should encourage the reduction of the overall corticoid exposition in children with a good response. Further, prospective randomized controlled trials are needed to analyze whether pulsatile treatment regimes are superior to standard hormonal treatment and to enlighten the optimal duration of therapy.
The main limitations of this trial are its strict retrospective character and the single-center study design which may lead to selection bias and cause the small sample size. Moreover, this is not a randomized, controlled, or blinded trial and the lack of a control group limits the evidence. Additional limitations are the co-medication with ASM before and during the treatment with dexamethasone which may affect the outcome parameters.
