Epilepsy Research (2015) 109, 155—162
journal homepage: www.elsevier.com/locate/epilepsyres
Neurodevelopmental outcome of infantile spasms: A systematic review and meta-analysis Elysa Widjaja a,b,∗, Cristina Go b, Blathnaid McCoy b, O. Carter Snead b a b
Diagnostic Imaging, The Hospital for Sick Children, Toronto, Ontario, Canada Division of Neurology, The Hospital for Sick Children, Toronto, Ontario, Canada
Received 15 September 2014; received in revised form 3 November 2014; accepted 11 November 2014 Available online 22 November 2014
KEYWORDS Infantile spasms; Neurodevelopment; Lead time to treatment
Summary Background: The aims of this systematic review and meta-analysis were to assess (i) estimates of good neurodevelopmental outcome in infantile spasms (IS), (ii) if neurodevelopmental outcome has changed since the publication of the first guideline on medical treatment of IS in 2004 and (iii) effect of lead time to treatment (LTTT). Methods: The Medline, Embase, Cochrane, PsycINFO, Web of Science and Scopus databases, and reference lists of retrieved articles were searched. Studies inclusion criteria were: (i) >5 patients with IS, (ii) mean/median follow-up of >6 months, (iii) neurodevelopmental outcome, and (iv) randomized and observational studies. The data extracted included proportion of good neurodevelopmental outcome, year of publication, cryptogenic or symptomatic IS and LTTT. Results: Of the 1436 citations screened, 55 articles were included in final analysis, with a total of 2967 patients. The pooled estimate for good neurodevelopmental outcome was 0.236 (95% CI: 0.193—0.286). There was no difference between the proportions of good neurodevelopmental outcome for the 21 studies published after 2004 [0.264 (95% CI: 0.197—0.344)] compared to the 34 studies published before 2004 [0.220 (95% CI: 0.168—0.283)] (Q value = 0.862, p = 0.353). The pooled estimate of good neurodevelopmental outcome for cryptogenic IS [0.543 (95% CI: 0.458—0.625)] was higher than symptomatic IS [0.125 (95% CI: 0.09—0.171)] (Q value = 69.724, p < 0.001). Risk ratio of LTTT 4weeks for good neurodevelopmental outcome of 8 studies was 1.519 (95% CI: 1.064—2.169).
∗
Corresponding author at: The Hospital for Sick Children, Toronto, Ontario M5G 1X8, Canada. Tel.: +1 416 813 7654; fax: +1 416 8135789. E-mail addresses: [email protected] (E. Widjaja), [email protected] (C. Go), [email protected] (B. McCoy), [email protected] (O.C. Snead). http://dx.doi.org/10.1016/j.eplepsyres.2014.11.012 0920-1211/© 2014 Elsevier B.V. All rights reserved.
156
E. Widjaja et al. Conclusion: Neurodevelopmental outcome was overall poor in patients with IS and has not changed since the publication of first guideline on IS. Although cryptogenic IS has better prognosis than symptomatic IS, the outcome for cryptogenic IS remained poor. There was heterogeneity in neurodevelopmental outcome ascertainment methods, highlighting the need for a more standardized and comprehensive assessment of cognitive, behavioural, emotional and functional outcomes. © 2014 Elsevier B.V. All rights reserved.
Introduction Infantile spasms (IS) constitutes a distinct and often catastrophic form of epilepsy of early infancy, characterized by epileptic spasms, often accompanied by neurodevelopmental regression and an EEG finding of hypsarrhythmia. When all three components are present, the term ‘‘West syndrome’’ is commonly used. The incidence of IS ranges from 2 to 3.5 per 10,000 live births (Luthvigsson et al., 1994; Riikonen, 2001). IS associated with an underlying disorder is known as symptomatic IS (Jellinger, 1987; Riikonen, 2001), and IS with no detectable underlying cause is known as cryptogenic IS. Those with cryptogenic IS were considered to have a better prognosis. The diagnosis and management of IS posed significant challenges to the treating physician. The first evidence-based guideline on the medical management of IS was published in 2004 (Mackay et al., 2004), and the two drugs that were determined most likely to be effective in the short-term management were ACTH and Vigabatrin. The authors also concluded that at the time of publication of the guideline, the effect of medical treatment on long-term neurodevelopmental outcome remains unknown. The aims of this systematic review and meta-analysis were to determine (i) the neurodevelopmental outcome of IS, (ii) if neurodevelopmental outcome has changed since the publication of the guideline on medical treatment of IS in 2004 and (iii) the effect of lead time to treatment (LTTT) on neurodevelopmental outcome.
Materials and methods Data sources A comprehensive literature search of Medline, Embase, Cochrane, PsycINFO, Web of Science and Scopus databases was conducted on Oct 15, 2013 (see Supplementary Table 1). We also searched bibliographies of reviews and original articles for additional articles. Searches were restricted to full-length English articles. No restriction was placed on the time of publication. References were exported and managed using EndNoteX7.
Study selection The titles and abstracts were screened to identify those reporting on original research that involved IS or West syndrome and neurodevelopmental outcome. The full-length articles of abstracts identified in the first screen were obtained and screened for inclusion into the systematic review and meta-analysis. The inclusion criteria were: (i)
reports of >5 patients with IS or West syndrome, (ii) neurodevelopmental outcome reported, (iii) outcomes reported after a mean/median follow-up of >6 months, (iv) the number of patients with good or poor neurodevelopmental outcome was reported, and (v) randomized controlled trials and observational studies. The instrument used to measure neurodevelopmental outcome, and the criteria used to define good neurodevelopmental outcome is shown in Supplementary Table 2. Duplicate publications, that is, studies with any overlapping patient populations from the same centre were excluded. Conference abstracts, unpublished studies, and surgical treatment of IS were excluded. Disagreements of eligibility were resolved through discussion.
Data extraction Two reviewers extracted data from the selected articles. The following data were extracted: study information (author, year), total number of participants, condition information (cryptogenic or symptomatic), estimates of good neurodevelopmental outcome, instrument for measuring development or cognition, and follow-up period. Additional data extracted included LTTT and number of patients with good neurodevelopmental outcome treated within or beyond a pre-defined LTTT.
Study quality The quality of the retrieved articles was assessed using a modified version of the Quality Index (Downs and Black, 1998). For the purpose of this review, the original 27-item has been reduced to 15 items (Ferro and Speechley, 2009). Each item was scored 0 (no/unable to determine) or 1 (yes), with a maximum score of 15. The three subscales are reporting (7 items), external validity (3 items), internal validity (4 items) and study power (1 item) (Ferro and Speechley, 2009). Higher scores indicated studies of higher methodological quality.
Data analyses Data analyses were conducted using the Comprehensive Meta-Analysis software (Englewood, USA). To assess for significance between study heterogeneity, the Q statistic was calculated and I2 was used to quantify the magnitude of between-study heterogeneity. If statistically significant heterogeneity was present (Q statistic p < 0.05), the pooled estimate of the proportion of good neurodevelopmental outcome and 95% confidence intervals (CIs) would be calculated
Neurodevelopmental outcome of infantile spasms using a random effect model. If there was no significance between study heterogeneity, the fixed effect model would be used. The pooled estimates and 95% CIs of studies published before and after 2004 were assessed using mixed effect model and Q statistics. The risk ratio and 95% CI of good neurodevelopmental outcome of patients treated within or beyond a specified LTTT was assessed. Subgroup analysis of patients with cryptogenic and symptomatic IS was also evaluated with mixed effect model and Q statistics. Publication bias was assessed visually using funnel plots and also statistically using Egger’s and the trim and fill tests (Duval and Tweedie, 2000a,b; Egger and Smith, 1997). The trim and fill method uses an iterative procedure to adjust for funnel plot asymmetry by imputing the effect estimates of potentially missing studies and evaluating the influence of these missing studies on the pooled estimate (Duval and Tweedie, 2000a,b).
157 The pooled risk ratio for the 8 studies for good neurodevelopmental outcome for LTTT 4 weeks was 1.519 (95% CI: 1.064—2.169). Auvin et al. (2012) reported 20/54 patients with LTTT 30 days had good neurodevelopmental outcome, hence the risk ratio was 10.370 (95% CI: 1.467—73.315), which was disproportionately higher than the other 7 studies (Cohen-Sadan et al., 2009; Holden et al., 1997; Jeavons et al., 1973; Kivity et al., 2004; Koo et al., 1993; Lombroso, 1983; Singer et al., 1982). If the study by Auvin et al. (2012) was excluded, the pooled risk ratio for the remaining 7 studies for good neurodevelopmental outcome for LTTT 4 weeks was 1.456 (95% CI: 1.076—1.969) (Fig. 4). The pooled risk ratio for the two studies for LTTT 8 weeks was 1.821 (95% CI: 0.944—3.513). There was no difference between the risk ratios for the LTTT of 4 weeks and 8 weeks (Q = 0.226; p = 0.634).
Cryptogenic versus symptomatic infantile spasm
Results The search strategy yielded a total of 3775 citations: 806 from MEDLINE, 675 from EMBASE, 1464 from Scopus, 655 from Web of Science, 162 from PsyINFO, and 13 from Cochrane database (a total of 1436 after duplicates were removed) (Fig. 1). After screening the titles and abstracts, 84 articles met the criteria for full-text review, of which 29 were excluded (in 2 studies follow-up was less than 6 months, in 2 studies follow-up period was not included, in 6 studies the number of patients with neurodevelopmental outcome was not included, in 3 studies the neurodevelopmental outcome was not assessed, in 6 studies only abstracts were available and 10 studies were duplicates). 55 articles were included in the final analyses. There was significant heterogeneity between studies (Q value = 366.849, p < 0.001); hence random effect model was used for the pooled estimate. There were a total of 2967 patients included in the analysis. The pooled estimate for the proportion of good neurodevelopmental outcome was 0.236 (95% CI: 0.193—0.286).
Twenty-five studies included cryptogenic and symptomatic IS, 10 studies reported only symptomatic IS and 7 studies reported only cryptogenic IS. The estimated proportion of good neurodevelopmental outcome in cryptogenic IS patients was 0.543 (95% CI: 0.458—0.625) and in symptomatic IS patients was 0.125 (95% CI: 0.091—0.171) (Q value = 69.724, p < 0.001).
Quality of study 47 of 55 articles were retrospective studies. Of the 8 prospective studies, only one was a randomized controlled study. The mean quality index score (QIS) was 8.6 (standard deviation: 1.5; range: 6—12) (Supplementary Table 2). There was high variability in the instruments used to assess development and/or intelligence quotient (IQ) within and across studies. The instrument used to assess neurodevelopmental outcome was not defined in 27 of the studies. The mean/median follow-up period was highly variable, ranging from >0.75 years to 25 years.
Pre- versus post-2004 studies
Publication bias
There were 21 studies published after 2004 (Fig. 2), and 34 studies were published before or during 2004 (Fig. 3). The estimated proportion of patients with good neurodevelopmental outcome pre 2004 was 0.220 (95% CI: 0.168—0.283) and post-2004 was 0.264 (95% CI: 0.197—0.344) (Table 1). There was no significant difference between the effect size pre- and post-2004 (Q value = 0.862, p = 0.353).
There was no publication bias detected by Egger’s test (t = 1.378, p = 0.174). On visual inspection, the funnel plot appeared slightly asymmetric. The trim and fill method identified 11 missing studies (Fig. 5), and the imputed pooled estimate for good neurodevelopmental outcome using the random effect model was 0.281 (95% CI: 0.233—0.335) (Fig. 5).
Lead time to treatment (LTTT)
Discussion
There were 12 studies that reported LTTT of 3—8 weeks. Eight studies assessed the outcome for LTTT of 4 weeks or 1 month (which would be labelled as 4 weeks for the purpose of this meta-analysis). Two studies assessed the outcome for LTTT of 8 weeks (Eisermann et al., 2003; Matsumoto et al., 1981), one for LTTT of 3 weeks (Glaze et al., 1988) and another for LTTT of 5 weeks (Cusmai et al., 2011).
This systematic review and meta-analysis showed that children with prior history of IS have poor neurodevelopmental outcome and the publication of the practice parameter on medical treatment of IS in 2004 (Mackay et al., 2004) has not changed neurodevelopmental outcome. Shorter LTTT of 4 weeks LTTT. Although
158
E. Widjaja et al.
Figure 1
Flow chart of search strategy.
cryptogenic IS was associated with better neurodevelopmental outcome compared to symptomatic IS, outcome remained poor in cryptogenic IS. The first evidence-based publication on guideline for the medical treatment of IS indicated that ACTH is probably an effective agent in the short-term treatment of IS, and Vigabatrin is possibly effective (Mackay et al., 2004). The US Consensus report on IS showed that there was considerable
Figure 2
variation in the management of IS (Pellock et al., 2010) despite the guideline. We concur that there were persistent variations in medical treatment of IS in the literature even after the publication of the guideline in 2004. At the time of the publication of the first guideline, the authors concluded that there was insufficient evidence to recommend the use of ACTH, corticosteroids, vigabatrin, valproic acid and vitamin B6 to improve long-term outcomes of IS (Mackay et al.,
Forrest plot of proportion of good neurodevelopmental outcome of studies published after 2004.
Neurodevelopmental outcome of infantile spasms
Figure 3
Forrest plot of proportion of good neurodevelopmental outcome of studies published before 2004.
2004). This meta-analysis showed that neurodevelopmental outcome has not changed since the publication of the guideline in 2004. Since then, the updated guideline recommended that hormonal therapy (ACTH or prednisolone) may be considered for use in preference to vigabatrin in cryptogenic IS to possibly improve neurodevelopmental outcome (Go et al., 2012), based largely on the findings of the randomized controlled trial from the United Kingdom Infantile Spasms Study (UKISS) (Darke et al., 2010; Lux et al., 2005). Further study is needed to evaluate the impact of medical treatment and successful control of
Table 1
159
IS, as defined by complete cessation of spasm and EEG normalization, on long-term neurodevelopmental outcome. Although the meta-analysis showed that cryptogenic IS was more likely to be associated with better neurodevelopmental outcome relative to symptomatic IS, the outcome for cryptogenic IS remained poor. Approximately 45% of those with cryptogenic IS continue to have poor neurodevelopmental outcome. It is possible that cryptogenic IS encompasses a range of genetic disorders with similar phenotypic manifestations, as well as those with subtle structural abnormality that are occult on neuroimaging. With
Subgroups of studies published before and after 2004, and those with cryptogenic and symptomatic infantile spasm.
All studies included in analysis Studies published before and up to 2004 Studies published after 2004 Studies with cryptogenic infantile spasm Studies with symptomatic infantile spasm
Number of studies
Total number of patients in studies
Proportion of patients with good neurodevelopmental outcome (95% confidence intervals)
55 34 21 32 35
2967 2004 963 670 1265
0.236 0.220 0.264 0.543 0.125
(0.193—0.286) (0.168—0.283) (0.197—0.344) (0.458—0.625) (0.091—0.171)
160
E. Widjaja et al.
Figure 4 Forrest plot of risk ratio for good neurodevelopmental outcome of lead time to treatment less than 4 weeks relative to greater than 4 weeks.
improvement in diagnostic tests including better imaging techniques, some patients who were previously considered as cryptogenic IS may now be considered as symptomatic IS. Hence, older literature reporting the neurodevelopmental outcome of cryptogenic IS may have underestimated the prognosis of cryptogenic IS. Further studies are needed to identify predictors of poor neurodevelopmental outcome in cryptogenic IS and whether subgroups of symptomatic IS have better prognosis than others. There was lack of consensus in the literature on the effect of LTTT on neurodevelopmental outcome (Askalan et al., 2003; Kivity et al., 2004; O’Callaghan et al., 2011; Singer et al., 1982). In this systematic review and meta-analysis, we have found that LTTT 4 weeks. There was no
Figure 5 Funnel plot with fill and trim showing slight asymmetry in the funnel plot with 11 missing studies (red circles). The pooled estimate of good neurodevelopmental outcome for the reported studies (0.236 (95% CI: 0.193—0.286)) is similar to the imputed estimate of good neurodevelopmental outcome with the missing studies (0.281 (95% CI: 0.233—0.335)). This indicates that the risk of publication bias is low. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
significant difference in the risk ratio for the two studies that reported outcomes for LTTT of 8 weeks (Ito et al., 2002; Matsumoto et al., 1981) (risk ratio = 1.821) compared to the risk ratio for the 8 studies that reported LTTT of 4 weeks (Auvin et al., 2012; Cohen-Sadan et al., 2009; Holden et al., 1997; Jeavons et al., 1973; Kivity et al., 2004; Koo et al., 1993; Lombroso, 1983; Singer et al., 1982) (risk ratio = 1.519), which may be related to the fewer number of studies reporting LTTT of 8 weeks. O’Callaghan et al. (2011) have shown that each category of LTTT (2 months) was associated with a 3.9 (95% CI: 7.3—0.4) point decrease in Vineland Adaptive Behaviour Scales, after controlling for the effects of treatment and aetiology. The updated guideline on the medical treatment of IS recommended that a shorter LTTT of IS with either hormonal therapy or vigabatrin may improve long-term cognitive outcomes (Go et al., 2012). The resource implication and impact of this recommendation on clinical practice needs to be assessed. Substantial heterogeneity in neurodevelopmental outcome ascertainment methods was identified. Some studies used developmental scales or clinical assessment to measure developmental outcome, while others used intelligence quotient (IQ) or educational level as the surrogate measure of cognitive outcome, or a combination of these measures. The length of follow-up may affect whether developmental or cognitive outcome was assessed, as the instruments for developmental outcome was designed for use in the younger children, and other instruments such as the Wechsler Abbreviated Scale of Intelligence for IQ assessment was feasible only above the age of 6 years. Heterogeneity was compounded by variability in scores of developmental quotient and IQ used to define good or poor outcome, with values ranging from 70 to 85 used as the cut-off for defining good or poor neurodevelopmental outcome. The variability used to measure and define neurodevelopmental outcomes may affect the estimates of outcome. Studies that assessed cognitive outcome have used IQ as a surrogate measure of cognition. Most studies did not conduct a comprehensive panel of cognitive evaluation, including memory, language, motor and executive function. Assessment of multiple domains including cognition, behaviour, emotion and function using a variety of validated instruments to capture multi-dimensional constructs is needed to provide a more holistic perspective of the patient’s neurodevelopment.
Neurodevelopmental outcome of infantile spasms The other limitations of this systematic review and metaanalysis included heterogeneity in treatment protocols, with a variety of drugs and dosages used. The follow-up periods were also variable, not only across studies, but also within studies. It was not clear if patients with poor neurodevelopmental outcome have persistent spasms or seizures and whether those with good neurodevelopmental outcome did not have spasms or seizures at the time of neurodevelopmental assessment. Most studies (47/55 studies) were retrospective and only one study was a randomized controlled trial. Given the limitations of retrospective studies, the quality of the vast majority of these studies were at best modest. Despite the limitations, this study has the largest pooled estimate of neurodevelopmental outcome in patients with IS, consisting of 2967 patients, which is not achievable in any prospective study.
Conclusions This systematic review and meta-analysis showed that neurodevelopmental outcome was overall poor in patients with IS, even in those with cryptogenic IS. Further research using a standardized neurodevelopmental outcome measures and a more comprehensive assessment of cognitive, behavioural, emotional and functional outcomes are needed to evaluate the long-term impact of IS on these patients. Further study is also needed to clarify if early versus late cessation of IS, and whether the development of other types of seizures predicted long-term neurodevelopmental outcome. Since IS is a rare disorder, multicentre controlled comparative studies or multicentre patient registries are needed to help inform optimal treatment strategies and predictors of long-term neurodevelopmental outcome.
Acknowledgement We thank Tamsin Adams-Webber for her assistance in literature search.
Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10. 1016/j.eplepsyres.2014.11.012.
References Askalan, R., Mackay, M., Brian, J., Otsubo, H., McDermott, C., Bryson, S., Boyd, J., Snead, C., Roberts, W., Weiss, S., 2003. Prospective preliminary analysis of the development of autism and epilepsy in children with infantile spasms. J. Child Neurol. 18, 165—170. Auvin, S., Hartman, A.L., Desnous, B., Moreau, A.C., Alberti, C., Delanoe, C., Romano, A., Terrone, G., Kossoff, E.H., Del Giudice, E., Titomanlio, L., 2012. Diagnosis delay in West syndrome: misdiagnosis and consequences. Eur. J. Pediatr. 171, 1695—1701. Cohen-Sadan, S., Kramer, U., Ben-Zeev, B., Lahat, E., Sahar, E., Nevo, Y., Eidlitz, T., Zeharia, A., Kivity, S., Goldberg-Stern, H., 2009. Multicenter long-term follow-up of children with idiopathic West syndrome: ACTH versus vigabatrin. Eur. J. Neurol. 16, 482—487.
161 Cusmai, R., Moavero, R., Bombardieri, R., Vigevano, F., Curatolo, P., 2011. Long-term neurological outcome in children with early-onset epilepsy associated with tuberous sclerosis. Epilepsy Behav. 22, 735—739. Darke, K., Edwards, S.W., Hancock, E., Johnson, A.L., Kennedy, C.R., Lux, A.L., Newton, R.W., O’Callaghan, F.J.K., Verity, C.M., Osborne, J.P., 2010. Developmental and epilepsy outcomes at age 4 years in the UKISS trial comparing hormonal treatments to vigabatrin for infantile spasms: a multi-centre randomised trial. Arch. Dis. Child. 95, 382—386. Downs, S.H., Black, N., 1998. The feasibility of creating a checklist for the assessment of the methodological quality both of randomised and non-randomised studies of health care interventions. J. Epidemiol. Community Health 52, 377—384. Duval, S., Tweedie, R., 2000a. A nonparametric trim and fill method of accounting for publication bias in meta-analysis. J. Am. Stat. Assoc. 95, 89—98. Duval, S., Tweedie, R., 2000b. Trim and fill: a simple funnel-plotbased method of testing and adjusting for publication bias in meta-analysis. Biometrics 56, 455—463. Egger, M., Smith, G., 1997. Bias in meta-analysis detected by simple, graphical test. Br. Med. J. 315, 629—634. Eisermann, M.M., DeLaRaillere, A., Dellatolas, G., Tozzi, E., Nabbout, R., Dulac, O., Chiron, C., 2003. Infantile spasms in Down syndrome — effects of delayed anticonvulsive treatment. Epilepsy Res. 55, 21—27. Ferro, M.A., Speechley, K.N., 2009. Depressive symptoms among mothers of children with epilepsy: a review of prevalence, associated factors, and impact on children. Epilepsia 50, 2344—2354. Glaze, D.G., Hrachovy, R.A., Frost Jr., J.D., Kellaway, P., Zion, T.E., 1988. Prospective study of outcome of infants with infantile spasms treated during controlled studies of ACTH and prednisone. J. Pediatr. 112, 389—396. Go, C.Y., Mackay, M.T., Weiss, S.K., Stephens, D., Adams-Webber, T., Ashwal, S., Snead III, O.C., 2012. Evidence-based guideline update-Medical treatment of infantile spasms: report of the guideline development subcommittee of the American Academy of Neurology and the practice committee of the Child Neurology Society. Neurology 78, 1974—1980. Holden, K.R., Clarke, S.L., Griesemer, D.A., 1997. Long-term outcomes of conventional therapy for infantile spasms. Seizure — Eur. J. Epilepsy 6, 201—205. Ito, M., Aiba, H., Hashimoto, K., Kuroki, S., Tomiwa, K., Okuno, T., Hattori, H., Go, T., Sejima, H., Dejima, S., Ikeda, H., Yoshioka, M., Kanazawa, O., Kawamitsu, T., Ochi, J., Miki, N., Noma, H., Oguro, K., Ozaki, N., Tamamoto, A., Matsubara, T., Miyajima, T., Fujii, T., Konishi, Y., Hojo, H., 2002. Low-dose ACTH therapy for West syndrome — initial effects and long-term outcome. Neurology 58, 110—114. Jeavons, P.M., Bower, B.D., Dimitrakoudi, M., 1973. Long-term prognosis of 150 cases of West syndrome. Epilepsia 14, 153—164. Jellinger, K., 1987. Neuropathological aspects of infantile spasms. Brain Dev. 9, 349—357. Kivity, S., Lerman, P., Ariel, R., Danziger, Y., Mimouni, M., Shinnar, S., 2004. Long-term cognitive outcomes of a cohort of children with cryptogenic infantile spasms treated with high-dose adrenocorticotropic hormone. Epilepsia 45, 255—262. Koo, B., Hwang, P.A., Logan, W.J., 1993. Infantile spasms: outcome and prognostic factors of cryptogenic and symptomatic groups. Neurology 43, 2322—2327. Lombroso, C.T., 1983. A prospective study of infantile spasms: clinical and therapeutic correlations. Epilepsia 24, 135—158. Luthvigsson, P., Olafsson, E., Sigurdardottir, S., Hauser, W.A., 1994. Epidemiologic features of infantile spasms in Iceland. Epilepsia 35, 802—805.
162 Lux, A.L., Edwards, S.W., Hancock, E., Johnson, A.L., Kennedy, C.R., Newton, R.W., O’Callaghan, F.J., Verity, C.M., Osborne, J.P., United Kingdom Infantile Spasms Study, 2005. The United Kingdom Infantile Spasms Study (UKISS) comparing hormone treatment with vigabatrin on developmental and epilepsy outcomes to age 14 months: a multicentre randomised trial. Lancet Neurol. 4, 712—717. Mackay, M.T., Weiss, S.K., Adams-Webber, T., Ashwal, S., Stephens, D., Ballaban-Gill, K., Baram, T.Z., Duchowny, M., Hirtz, D., Pellock, J.M., Shields, W.D., Shinnar, S., Wyllie, E., Snead, I.O.C., 2004. Practice parameter: Medical treatment of infantile spasms: report of the American Academy of Neurology and the Child Neurology Society. Neurology 62, 1668—1681. Matsumoto, A., Watanabe, K., Negoro, T., 1981. Long-term prognosis after infantile spasms: a statistical study of prognostic factors in 200 cases. Dev. Med. Child Neurol. 23, 51—65.
E. Widjaja et al. O’Callaghan, F.J.K., Lux, A.L., Darke, K., Edwards, S.W., Hancock, E., Johnson, A.L., Kennedy, C.R., Newton, R.W., Verity, C.M., Osborne, J.P., 2011. The effect of lead time to treatment and of age of onset on developmental outcome at 4 years in infantile spasms: evidence from the United Kingdom Infantile Spasms Study. Epilepsia 52, 1359—1364. Pellock, J.M., Hrachovy, R., Shinnar, S., Baram, T.Z., Bettis, D., Dlugos, D.J., Gaillard, W.D., Gibson, P.A., Holmes, G.L., Nordl, D.R., O’Dell, C., Shields, W.D., Trevathan, E., Wheless, J.W., 2010. Infantile spasms: a U.S. consensus report. Epilepsia 51, 2175—2189. Riikonen, R., 2001. Epidemiological data of West syndrome in Finland. Brain Dev. 23, 539—541. Singer, W.D., Haller, J.S., Sullivan, L.R., Wolpert, S., Mills, C., Rabe, E.F., 1982. The value of neuroradiology in infantile spasms. J. Pediatr. 100, 47—50.
journal homepage: www.elsevier.com/locate/epilepsyres
Neurodevelopmental outcome of infantile spasms: A systematic review and meta-analysis Elysa Widjaja a,b,∗, Cristina Go b, Blathnaid McCoy b, O. Carter Snead b a b
Diagnostic Imaging, The Hospital for Sick Children, Toronto, Ontario, Canada Division of Neurology, The Hospital for Sick Children, Toronto, Ontario, Canada
Received 15 September 2014; received in revised form 3 November 2014; accepted 11 November 2014 Available online 22 November 2014
KEYWORDS Infantile spasms; Neurodevelopment; Lead time to treatment
Summary Background: The aims of this systematic review and meta-analysis were to assess (i) estimates of good neurodevelopmental outcome in infantile spasms (IS), (ii) if neurodevelopmental outcome has changed since the publication of the first guideline on medical treatment of IS in 2004 and (iii) effect of lead time to treatment (LTTT). Methods: The Medline, Embase, Cochrane, PsycINFO, Web of Science and Scopus databases, and reference lists of retrieved articles were searched. Studies inclusion criteria were: (i) >5 patients with IS, (ii) mean/median follow-up of >6 months, (iii) neurodevelopmental outcome, and (iv) randomized and observational studies. The data extracted included proportion of good neurodevelopmental outcome, year of publication, cryptogenic or symptomatic IS and LTTT. Results: Of the 1436 citations screened, 55 articles were included in final analysis, with a total of 2967 patients. The pooled estimate for good neurodevelopmental outcome was 0.236 (95% CI: 0.193—0.286). There was no difference between the proportions of good neurodevelopmental outcome for the 21 studies published after 2004 [0.264 (95% CI: 0.197—0.344)] compared to the 34 studies published before 2004 [0.220 (95% CI: 0.168—0.283)] (Q value = 0.862, p = 0.353). The pooled estimate of good neurodevelopmental outcome for cryptogenic IS [0.543 (95% CI: 0.458—0.625)] was higher than symptomatic IS [0.125 (95% CI: 0.09—0.171)] (Q value = 69.724, p < 0.001). Risk ratio of LTTT 4weeks for good neurodevelopmental outcome of 8 studies was 1.519 (95% CI: 1.064—2.169).
∗
Corresponding author at: The Hospital for Sick Children, Toronto, Ontario M5G 1X8, Canada. Tel.: +1 416 813 7654; fax: +1 416 8135789. E-mail addresses: [email protected] (E. Widjaja), [email protected] (C. Go), [email protected] (B. McCoy), [email protected] (O.C. Snead). http://dx.doi.org/10.1016/j.eplepsyres.2014.11.012 0920-1211/© 2014 Elsevier B.V. All rights reserved.
156
E. Widjaja et al. Conclusion: Neurodevelopmental outcome was overall poor in patients with IS and has not changed since the publication of first guideline on IS. Although cryptogenic IS has better prognosis than symptomatic IS, the outcome for cryptogenic IS remained poor. There was heterogeneity in neurodevelopmental outcome ascertainment methods, highlighting the need for a more standardized and comprehensive assessment of cognitive, behavioural, emotional and functional outcomes. © 2014 Elsevier B.V. All rights reserved.
Introduction Infantile spasms (IS) constitutes a distinct and often catastrophic form of epilepsy of early infancy, characterized by epileptic spasms, often accompanied by neurodevelopmental regression and an EEG finding of hypsarrhythmia. When all three components are present, the term ‘‘West syndrome’’ is commonly used. The incidence of IS ranges from 2 to 3.5 per 10,000 live births (Luthvigsson et al., 1994; Riikonen, 2001). IS associated with an underlying disorder is known as symptomatic IS (Jellinger, 1987; Riikonen, 2001), and IS with no detectable underlying cause is known as cryptogenic IS. Those with cryptogenic IS were considered to have a better prognosis. The diagnosis and management of IS posed significant challenges to the treating physician. The first evidence-based guideline on the medical management of IS was published in 2004 (Mackay et al., 2004), and the two drugs that were determined most likely to be effective in the short-term management were ACTH and Vigabatrin. The authors also concluded that at the time of publication of the guideline, the effect of medical treatment on long-term neurodevelopmental outcome remains unknown. The aims of this systematic review and meta-analysis were to determine (i) the neurodevelopmental outcome of IS, (ii) if neurodevelopmental outcome has changed since the publication of the guideline on medical treatment of IS in 2004 and (iii) the effect of lead time to treatment (LTTT) on neurodevelopmental outcome.
Materials and methods Data sources A comprehensive literature search of Medline, Embase, Cochrane, PsycINFO, Web of Science and Scopus databases was conducted on Oct 15, 2013 (see Supplementary Table 1). We also searched bibliographies of reviews and original articles for additional articles. Searches were restricted to full-length English articles. No restriction was placed on the time of publication. References were exported and managed using EndNoteX7.
Study selection The titles and abstracts were screened to identify those reporting on original research that involved IS or West syndrome and neurodevelopmental outcome. The full-length articles of abstracts identified in the first screen were obtained and screened for inclusion into the systematic review and meta-analysis. The inclusion criteria were: (i)
reports of >5 patients with IS or West syndrome, (ii) neurodevelopmental outcome reported, (iii) outcomes reported after a mean/median follow-up of >6 months, (iv) the number of patients with good or poor neurodevelopmental outcome was reported, and (v) randomized controlled trials and observational studies. The instrument used to measure neurodevelopmental outcome, and the criteria used to define good neurodevelopmental outcome is shown in Supplementary Table 2. Duplicate publications, that is, studies with any overlapping patient populations from the same centre were excluded. Conference abstracts, unpublished studies, and surgical treatment of IS were excluded. Disagreements of eligibility were resolved through discussion.
Data extraction Two reviewers extracted data from the selected articles. The following data were extracted: study information (author, year), total number of participants, condition information (cryptogenic or symptomatic), estimates of good neurodevelopmental outcome, instrument for measuring development or cognition, and follow-up period. Additional data extracted included LTTT and number of patients with good neurodevelopmental outcome treated within or beyond a pre-defined LTTT.
Study quality The quality of the retrieved articles was assessed using a modified version of the Quality Index (Downs and Black, 1998). For the purpose of this review, the original 27-item has been reduced to 15 items (Ferro and Speechley, 2009). Each item was scored 0 (no/unable to determine) or 1 (yes), with a maximum score of 15. The three subscales are reporting (7 items), external validity (3 items), internal validity (4 items) and study power (1 item) (Ferro and Speechley, 2009). Higher scores indicated studies of higher methodological quality.
Data analyses Data analyses were conducted using the Comprehensive Meta-Analysis software (Englewood, USA). To assess for significance between study heterogeneity, the Q statistic was calculated and I2 was used to quantify the magnitude of between-study heterogeneity. If statistically significant heterogeneity was present (Q statistic p < 0.05), the pooled estimate of the proportion of good neurodevelopmental outcome and 95% confidence intervals (CIs) would be calculated
Neurodevelopmental outcome of infantile spasms using a random effect model. If there was no significance between study heterogeneity, the fixed effect model would be used. The pooled estimates and 95% CIs of studies published before and after 2004 were assessed using mixed effect model and Q statistics. The risk ratio and 95% CI of good neurodevelopmental outcome of patients treated within or beyond a specified LTTT was assessed. Subgroup analysis of patients with cryptogenic and symptomatic IS was also evaluated with mixed effect model and Q statistics. Publication bias was assessed visually using funnel plots and also statistically using Egger’s and the trim and fill tests (Duval and Tweedie, 2000a,b; Egger and Smith, 1997). The trim and fill method uses an iterative procedure to adjust for funnel plot asymmetry by imputing the effect estimates of potentially missing studies and evaluating the influence of these missing studies on the pooled estimate (Duval and Tweedie, 2000a,b).
157 The pooled risk ratio for the 8 studies for good neurodevelopmental outcome for LTTT 4 weeks was 1.519 (95% CI: 1.064—2.169). Auvin et al. (2012) reported 20/54 patients with LTTT 30 days had good neurodevelopmental outcome, hence the risk ratio was 10.370 (95% CI: 1.467—73.315), which was disproportionately higher than the other 7 studies (Cohen-Sadan et al., 2009; Holden et al., 1997; Jeavons et al., 1973; Kivity et al., 2004; Koo et al., 1993; Lombroso, 1983; Singer et al., 1982). If the study by Auvin et al. (2012) was excluded, the pooled risk ratio for the remaining 7 studies for good neurodevelopmental outcome for LTTT 4 weeks was 1.456 (95% CI: 1.076—1.969) (Fig. 4). The pooled risk ratio for the two studies for LTTT 8 weeks was 1.821 (95% CI: 0.944—3.513). There was no difference between the risk ratios for the LTTT of 4 weeks and 8 weeks (Q = 0.226; p = 0.634).
Cryptogenic versus symptomatic infantile spasm
Results The search strategy yielded a total of 3775 citations: 806 from MEDLINE, 675 from EMBASE, 1464 from Scopus, 655 from Web of Science, 162 from PsyINFO, and 13 from Cochrane database (a total of 1436 after duplicates were removed) (Fig. 1). After screening the titles and abstracts, 84 articles met the criteria for full-text review, of which 29 were excluded (in 2 studies follow-up was less than 6 months, in 2 studies follow-up period was not included, in 6 studies the number of patients with neurodevelopmental outcome was not included, in 3 studies the neurodevelopmental outcome was not assessed, in 6 studies only abstracts were available and 10 studies were duplicates). 55 articles were included in the final analyses. There was significant heterogeneity between studies (Q value = 366.849, p < 0.001); hence random effect model was used for the pooled estimate. There were a total of 2967 patients included in the analysis. The pooled estimate for the proportion of good neurodevelopmental outcome was 0.236 (95% CI: 0.193—0.286).
Twenty-five studies included cryptogenic and symptomatic IS, 10 studies reported only symptomatic IS and 7 studies reported only cryptogenic IS. The estimated proportion of good neurodevelopmental outcome in cryptogenic IS patients was 0.543 (95% CI: 0.458—0.625) and in symptomatic IS patients was 0.125 (95% CI: 0.091—0.171) (Q value = 69.724, p < 0.001).
Quality of study 47 of 55 articles were retrospective studies. Of the 8 prospective studies, only one was a randomized controlled study. The mean quality index score (QIS) was 8.6 (standard deviation: 1.5; range: 6—12) (Supplementary Table 2). There was high variability in the instruments used to assess development and/or intelligence quotient (IQ) within and across studies. The instrument used to assess neurodevelopmental outcome was not defined in 27 of the studies. The mean/median follow-up period was highly variable, ranging from >0.75 years to 25 years.
Pre- versus post-2004 studies
Publication bias
There were 21 studies published after 2004 (Fig. 2), and 34 studies were published before or during 2004 (Fig. 3). The estimated proportion of patients with good neurodevelopmental outcome pre 2004 was 0.220 (95% CI: 0.168—0.283) and post-2004 was 0.264 (95% CI: 0.197—0.344) (Table 1). There was no significant difference between the effect size pre- and post-2004 (Q value = 0.862, p = 0.353).
There was no publication bias detected by Egger’s test (t = 1.378, p = 0.174). On visual inspection, the funnel plot appeared slightly asymmetric. The trim and fill method identified 11 missing studies (Fig. 5), and the imputed pooled estimate for good neurodevelopmental outcome using the random effect model was 0.281 (95% CI: 0.233—0.335) (Fig. 5).
Lead time to treatment (LTTT)
Discussion
There were 12 studies that reported LTTT of 3—8 weeks. Eight studies assessed the outcome for LTTT of 4 weeks or 1 month (which would be labelled as 4 weeks for the purpose of this meta-analysis). Two studies assessed the outcome for LTTT of 8 weeks (Eisermann et al., 2003; Matsumoto et al., 1981), one for LTTT of 3 weeks (Glaze et al., 1988) and another for LTTT of 5 weeks (Cusmai et al., 2011).
This systematic review and meta-analysis showed that children with prior history of IS have poor neurodevelopmental outcome and the publication of the practice parameter on medical treatment of IS in 2004 (Mackay et al., 2004) has not changed neurodevelopmental outcome. Shorter LTTT of 4 weeks LTTT. Although
158
E. Widjaja et al.
Figure 1
Flow chart of search strategy.
cryptogenic IS was associated with better neurodevelopmental outcome compared to symptomatic IS, outcome remained poor in cryptogenic IS. The first evidence-based publication on guideline for the medical treatment of IS indicated that ACTH is probably an effective agent in the short-term treatment of IS, and Vigabatrin is possibly effective (Mackay et al., 2004). The US Consensus report on IS showed that there was considerable
Figure 2
variation in the management of IS (Pellock et al., 2010) despite the guideline. We concur that there were persistent variations in medical treatment of IS in the literature even after the publication of the guideline in 2004. At the time of the publication of the first guideline, the authors concluded that there was insufficient evidence to recommend the use of ACTH, corticosteroids, vigabatrin, valproic acid and vitamin B6 to improve long-term outcomes of IS (Mackay et al.,
Forrest plot of proportion of good neurodevelopmental outcome of studies published after 2004.
Neurodevelopmental outcome of infantile spasms
Figure 3
Forrest plot of proportion of good neurodevelopmental outcome of studies published before 2004.
2004). This meta-analysis showed that neurodevelopmental outcome has not changed since the publication of the guideline in 2004. Since then, the updated guideline recommended that hormonal therapy (ACTH or prednisolone) may be considered for use in preference to vigabatrin in cryptogenic IS to possibly improve neurodevelopmental outcome (Go et al., 2012), based largely on the findings of the randomized controlled trial from the United Kingdom Infantile Spasms Study (UKISS) (Darke et al., 2010; Lux et al., 2005). Further study is needed to evaluate the impact of medical treatment and successful control of
Table 1
159
IS, as defined by complete cessation of spasm and EEG normalization, on long-term neurodevelopmental outcome. Although the meta-analysis showed that cryptogenic IS was more likely to be associated with better neurodevelopmental outcome relative to symptomatic IS, the outcome for cryptogenic IS remained poor. Approximately 45% of those with cryptogenic IS continue to have poor neurodevelopmental outcome. It is possible that cryptogenic IS encompasses a range of genetic disorders with similar phenotypic manifestations, as well as those with subtle structural abnormality that are occult on neuroimaging. With
Subgroups of studies published before and after 2004, and those with cryptogenic and symptomatic infantile spasm.
All studies included in analysis Studies published before and up to 2004 Studies published after 2004 Studies with cryptogenic infantile spasm Studies with symptomatic infantile spasm
Number of studies
Total number of patients in studies
Proportion of patients with good neurodevelopmental outcome (95% confidence intervals)
55 34 21 32 35
2967 2004 963 670 1265
0.236 0.220 0.264 0.543 0.125
(0.193—0.286) (0.168—0.283) (0.197—0.344) (0.458—0.625) (0.091—0.171)
160
E. Widjaja et al.
Figure 4 Forrest plot of risk ratio for good neurodevelopmental outcome of lead time to treatment less than 4 weeks relative to greater than 4 weeks.
improvement in diagnostic tests including better imaging techniques, some patients who were previously considered as cryptogenic IS may now be considered as symptomatic IS. Hence, older literature reporting the neurodevelopmental outcome of cryptogenic IS may have underestimated the prognosis of cryptogenic IS. Further studies are needed to identify predictors of poor neurodevelopmental outcome in cryptogenic IS and whether subgroups of symptomatic IS have better prognosis than others. There was lack of consensus in the literature on the effect of LTTT on neurodevelopmental outcome (Askalan et al., 2003; Kivity et al., 2004; O’Callaghan et al., 2011; Singer et al., 1982). In this systematic review and meta-analysis, we have found that LTTT 4 weeks. There was no
Figure 5 Funnel plot with fill and trim showing slight asymmetry in the funnel plot with 11 missing studies (red circles). The pooled estimate of good neurodevelopmental outcome for the reported studies (0.236 (95% CI: 0.193—0.286)) is similar to the imputed estimate of good neurodevelopmental outcome with the missing studies (0.281 (95% CI: 0.233—0.335)). This indicates that the risk of publication bias is low. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
significant difference in the risk ratio for the two studies that reported outcomes for LTTT of 8 weeks (Ito et al., 2002; Matsumoto et al., 1981) (risk ratio = 1.821) compared to the risk ratio for the 8 studies that reported LTTT of 4 weeks (Auvin et al., 2012; Cohen-Sadan et al., 2009; Holden et al., 1997; Jeavons et al., 1973; Kivity et al., 2004; Koo et al., 1993; Lombroso, 1983; Singer et al., 1982) (risk ratio = 1.519), which may be related to the fewer number of studies reporting LTTT of 8 weeks. O’Callaghan et al. (2011) have shown that each category of LTTT (2 months) was associated with a 3.9 (95% CI: 7.3—0.4) point decrease in Vineland Adaptive Behaviour Scales, after controlling for the effects of treatment and aetiology. The updated guideline on the medical treatment of IS recommended that a shorter LTTT of IS with either hormonal therapy or vigabatrin may improve long-term cognitive outcomes (Go et al., 2012). The resource implication and impact of this recommendation on clinical practice needs to be assessed. Substantial heterogeneity in neurodevelopmental outcome ascertainment methods was identified. Some studies used developmental scales or clinical assessment to measure developmental outcome, while others used intelligence quotient (IQ) or educational level as the surrogate measure of cognitive outcome, or a combination of these measures. The length of follow-up may affect whether developmental or cognitive outcome was assessed, as the instruments for developmental outcome was designed for use in the younger children, and other instruments such as the Wechsler Abbreviated Scale of Intelligence for IQ assessment was feasible only above the age of 6 years. Heterogeneity was compounded by variability in scores of developmental quotient and IQ used to define good or poor outcome, with values ranging from 70 to 85 used as the cut-off for defining good or poor neurodevelopmental outcome. The variability used to measure and define neurodevelopmental outcomes may affect the estimates of outcome. Studies that assessed cognitive outcome have used IQ as a surrogate measure of cognition. Most studies did not conduct a comprehensive panel of cognitive evaluation, including memory, language, motor and executive function. Assessment of multiple domains including cognition, behaviour, emotion and function using a variety of validated instruments to capture multi-dimensional constructs is needed to provide a more holistic perspective of the patient’s neurodevelopment.
Neurodevelopmental outcome of infantile spasms The other limitations of this systematic review and metaanalysis included heterogeneity in treatment protocols, with a variety of drugs and dosages used. The follow-up periods were also variable, not only across studies, but also within studies. It was not clear if patients with poor neurodevelopmental outcome have persistent spasms or seizures and whether those with good neurodevelopmental outcome did not have spasms or seizures at the time of neurodevelopmental assessment. Most studies (47/55 studies) were retrospective and only one study was a randomized controlled trial. Given the limitations of retrospective studies, the quality of the vast majority of these studies were at best modest. Despite the limitations, this study has the largest pooled estimate of neurodevelopmental outcome in patients with IS, consisting of 2967 patients, which is not achievable in any prospective study.
Conclusions This systematic review and meta-analysis showed that neurodevelopmental outcome was overall poor in patients with IS, even in those with cryptogenic IS. Further research using a standardized neurodevelopmental outcome measures and a more comprehensive assessment of cognitive, behavioural, emotional and functional outcomes are needed to evaluate the long-term impact of IS on these patients. Further study is also needed to clarify if early versus late cessation of IS, and whether the development of other types of seizures predicted long-term neurodevelopmental outcome. Since IS is a rare disorder, multicentre controlled comparative studies or multicentre patient registries are needed to help inform optimal treatment strategies and predictors of long-term neurodevelopmental outcome.
Acknowledgement We thank Tamsin Adams-Webber for her assistance in literature search.
Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10. 1016/j.eplepsyres.2014.11.012.
References Askalan, R., Mackay, M., Brian, J., Otsubo, H., McDermott, C., Bryson, S., Boyd, J., Snead, C., Roberts, W., Weiss, S., 2003. Prospective preliminary analysis of the development of autism and epilepsy in children with infantile spasms. J. Child Neurol. 18, 165—170. Auvin, S., Hartman, A.L., Desnous, B., Moreau, A.C., Alberti, C., Delanoe, C., Romano, A., Terrone, G., Kossoff, E.H., Del Giudice, E., Titomanlio, L., 2012. Diagnosis delay in West syndrome: misdiagnosis and consequences. Eur. J. Pediatr. 171, 1695—1701. Cohen-Sadan, S., Kramer, U., Ben-Zeev, B., Lahat, E., Sahar, E., Nevo, Y., Eidlitz, T., Zeharia, A., Kivity, S., Goldberg-Stern, H., 2009. Multicenter long-term follow-up of children with idiopathic West syndrome: ACTH versus vigabatrin. Eur. J. Neurol. 16, 482—487.
161 Cusmai, R., Moavero, R., Bombardieri, R., Vigevano, F., Curatolo, P., 2011. Long-term neurological outcome in children with early-onset epilepsy associated with tuberous sclerosis. Epilepsy Behav. 22, 735—739. Darke, K., Edwards, S.W., Hancock, E., Johnson, A.L., Kennedy, C.R., Lux, A.L., Newton, R.W., O’Callaghan, F.J.K., Verity, C.M., Osborne, J.P., 2010. Developmental and epilepsy outcomes at age 4 years in the UKISS trial comparing hormonal treatments to vigabatrin for infantile spasms: a multi-centre randomised trial. Arch. Dis. Child. 95, 382—386. Downs, S.H., Black, N., 1998. The feasibility of creating a checklist for the assessment of the methodological quality both of randomised and non-randomised studies of health care interventions. J. Epidemiol. Community Health 52, 377—384. Duval, S., Tweedie, R., 2000a. A nonparametric trim and fill method of accounting for publication bias in meta-analysis. J. Am. Stat. Assoc. 95, 89—98. Duval, S., Tweedie, R., 2000b. Trim and fill: a simple funnel-plotbased method of testing and adjusting for publication bias in meta-analysis. Biometrics 56, 455—463. Egger, M., Smith, G., 1997. Bias in meta-analysis detected by simple, graphical test. Br. Med. J. 315, 629—634. Eisermann, M.M., DeLaRaillere, A., Dellatolas, G., Tozzi, E., Nabbout, R., Dulac, O., Chiron, C., 2003. Infantile spasms in Down syndrome — effects of delayed anticonvulsive treatment. Epilepsy Res. 55, 21—27. Ferro, M.A., Speechley, K.N., 2009. Depressive symptoms among mothers of children with epilepsy: a review of prevalence, associated factors, and impact on children. Epilepsia 50, 2344—2354. Glaze, D.G., Hrachovy, R.A., Frost Jr., J.D., Kellaway, P., Zion, T.E., 1988. Prospective study of outcome of infants with infantile spasms treated during controlled studies of ACTH and prednisone. J. Pediatr. 112, 389—396. Go, C.Y., Mackay, M.T., Weiss, S.K., Stephens, D., Adams-Webber, T., Ashwal, S., Snead III, O.C., 2012. Evidence-based guideline update-Medical treatment of infantile spasms: report of the guideline development subcommittee of the American Academy of Neurology and the practice committee of the Child Neurology Society. Neurology 78, 1974—1980. Holden, K.R., Clarke, S.L., Griesemer, D.A., 1997. Long-term outcomes of conventional therapy for infantile spasms. Seizure — Eur. J. Epilepsy 6, 201—205. Ito, M., Aiba, H., Hashimoto, K., Kuroki, S., Tomiwa, K., Okuno, T., Hattori, H., Go, T., Sejima, H., Dejima, S., Ikeda, H., Yoshioka, M., Kanazawa, O., Kawamitsu, T., Ochi, J., Miki, N., Noma, H., Oguro, K., Ozaki, N., Tamamoto, A., Matsubara, T., Miyajima, T., Fujii, T., Konishi, Y., Hojo, H., 2002. Low-dose ACTH therapy for West syndrome — initial effects and long-term outcome. Neurology 58, 110—114. Jeavons, P.M., Bower, B.D., Dimitrakoudi, M., 1973. Long-term prognosis of 150 cases of West syndrome. Epilepsia 14, 153—164. Jellinger, K., 1987. Neuropathological aspects of infantile spasms. Brain Dev. 9, 349—357. Kivity, S., Lerman, P., Ariel, R., Danziger, Y., Mimouni, M., Shinnar, S., 2004. Long-term cognitive outcomes of a cohort of children with cryptogenic infantile spasms treated with high-dose adrenocorticotropic hormone. Epilepsia 45, 255—262. Koo, B., Hwang, P.A., Logan, W.J., 1993. Infantile spasms: outcome and prognostic factors of cryptogenic and symptomatic groups. Neurology 43, 2322—2327. Lombroso, C.T., 1983. A prospective study of infantile spasms: clinical and therapeutic correlations. Epilepsia 24, 135—158. Luthvigsson, P., Olafsson, E., Sigurdardottir, S., Hauser, W.A., 1994. Epidemiologic features of infantile spasms in Iceland. Epilepsia 35, 802—805.
162 Lux, A.L., Edwards, S.W., Hancock, E., Johnson, A.L., Kennedy, C.R., Newton, R.W., O’Callaghan, F.J., Verity, C.M., Osborne, J.P., United Kingdom Infantile Spasms Study, 2005. The United Kingdom Infantile Spasms Study (UKISS) comparing hormone treatment with vigabatrin on developmental and epilepsy outcomes to age 14 months: a multicentre randomised trial. Lancet Neurol. 4, 712—717. Mackay, M.T., Weiss, S.K., Adams-Webber, T., Ashwal, S., Stephens, D., Ballaban-Gill, K., Baram, T.Z., Duchowny, M., Hirtz, D., Pellock, J.M., Shields, W.D., Shinnar, S., Wyllie, E., Snead, I.O.C., 2004. Practice parameter: Medical treatment of infantile spasms: report of the American Academy of Neurology and the Child Neurology Society. Neurology 62, 1668—1681. Matsumoto, A., Watanabe, K., Negoro, T., 1981. Long-term prognosis after infantile spasms: a statistical study of prognostic factors in 200 cases. Dev. Med. Child Neurol. 23, 51—65.
E. Widjaja et al. O’Callaghan, F.J.K., Lux, A.L., Darke, K., Edwards, S.W., Hancock, E., Johnson, A.L., Kennedy, C.R., Newton, R.W., Verity, C.M., Osborne, J.P., 2011. The effect of lead time to treatment and of age of onset on developmental outcome at 4 years in infantile spasms: evidence from the United Kingdom Infantile Spasms Study. Epilepsia 52, 1359—1364. Pellock, J.M., Hrachovy, R., Shinnar, S., Baram, T.Z., Bettis, D., Dlugos, D.J., Gaillard, W.D., Gibson, P.A., Holmes, G.L., Nordl, D.R., O’Dell, C., Shields, W.D., Trevathan, E., Wheless, J.W., 2010. Infantile spasms: a U.S. consensus report. Epilepsia 51, 2175—2189. Riikonen, R., 2001. Epidemiological data of West syndrome in Finland. Brain Dev. 23, 539—541. Singer, W.D., Haller, J.S., Sullivan, L.R., Wolpert, S., Mills, C., Rabe, E.F., 1982. The value of neuroradiology in infantile spasms. J. Pediatr. 100, 47—50.
