Also taking this into consideration, we believe that LGS should be perceived as an age-dependent electroclinical pattern rather than an epilepsy disease proper. Such an electroclinical pattern (syndrome) can be found in a variety of distinct diseases such as tuberous sclerosis complex or Sturge-Weber syndrome, as well as in association with developmental abnormalities of the brain, both extended or relatively focal. In fact, it is the conjunction of certain clinical and EEG features that allow the diagnosis of LGS in patients with a variety of underlying neurological disorders (Arzimanoglou et al. 2004).

Issues related to prognosis

A comprehensive debate on prognosis should imply that we are dealing with a well-defined entity and that prospective data are available on the natural evolution of the disease for a statistically significant number of patients, rigorously corresponding to predefined criteria. The number of parameters to be considered when evaluating complex epilepsies, such as LGS, makes prospective large-scale studies very difficult to realize, not only because of the rarity of the syndrome but also because one would need to integrate the potential role of different aetiologies, the presence of various types of seizures with an onset at various phases of the disorder and at different ages, and the use of a number of antiepileptic drugs (AEDs) (order of prescription, which combination(s), at what dose(s), for which type(s) of seizures, etc.). The reasons include (1) the long-term evolution and prognosis were poorly investigated and (2) the reports on large numbers of patients remain relatively few (Chevrie and Aicardi 1972; Yagi 1996; Heiskala 1997).

Short-term mortality has been estimated to be 4.2% to 7% of those affected, in part as a result of tonic status. Up to 10% of children with LGS die prior to age 11 years (Beaumanoir et al. 1988; Rai et al. 1988).

Most of the published data on the prognosis of LGS come from retrospective series. The inherent risk with such studies is to include patients not on the basis of the hypothesized diagnosis at onset but after some years of evolution, influenced by what is already considered being the syndrome’s evolution.

PROGNOSIS OF EPILEPTIC SEIZURES IN LGS

Patients with tonic seizures, episodes of non-convulsive or tonic status epilepticus, and rapid discharges during sleep are most likely as adults to manifest the same seizure types. Beaumanoir (1981) studied a group of 103 patients followed up for a mean of 19.7 years and divided it into 2 subgroups:

1. Patients with a typical picture including tonic seizures, non-convulsive or tonic status, and 10-Hz discharges during slow sleep, and who continued to exhibit an unchanged pattern at the end of follow-up in three of four of cases

In the second group, with a better outcome, LGS seemed to represent only an age-related mode of expression of a focal epilepsy with temporary diffusion of the EEG abnormalities.

Similar results have been reported (Yagi 1996) on the persistence of tonic seizures in typical patients. Furthermore, in typical cases, the clinical pattern changes with age. Between 15 and 20 years old, the overall seizure frequency usually diminishes. Atypical absences and drop attacks may become rare, but all other types of seizures, including tonic seizures at sleep, persist. In older patients, tonic seizures during sleep are probably under-reported.

Prognosis may also be different in the group of patients who develop atonic and/or tonic seizures in the course of focal epilepsies (Pazzaglia et al. 1985; Roger et al. 1987). Gastaut and Zifkin (1988) drew similar conclusions in their study of 40 patients. In all studies, the occurrence of tonic and atonic attacks was associated with the appearance in the EEG records of bilateral fast or slow spike–wave complexes and/or bursts of 10-Hz recruiting rhythm, especially during sleep. Such secondary diffusion of initially focal epilepsy may be very difficult to distinguish it from classic LGS. In fact, most of such cases fulfil all or most of the criteria for the syndrome, of which they may constitute a subgroup and should not be regarded as a different entity. Their practical interest is that they may be due to localized lesions that could conceivably be amenable to surgical treatment. It is, indeed, possible that many cases of LGS are due to undetectable focal lesions or regions of dysfunction.

Prognosis is difficult to establish for patients presenting at onset only with myoclonic-astatic seizures. Most of them will have a long-term evolution that differs from the typical evolution of LGS. However, some patients will later also present tonic seizures and/or bursts of rapid rhythms during sleep, thus fulfilling the criteria for LGS diagnosis (Kaminska et al. 1999; Arzimanoglou et al. 2009; Kaminska and Oguni 2013).

PROGNOSIS AND THE ROLE OF AEDS

Regarding overall efficacy, the vast majority of patients with LGS remain intractable to AEDs or to other treatment options such as ketogenic diet or vagal nerve stimulation.

Undoubtedly, a small number of patients experience a reduction in seizure frequency, particularly when appropriately treated from onset with large-spectrum drugs such as valproate or lamotrigine, alone or in combination. For a greater proportion of patients, drop attacks could be controlled, thus influencing the quality of life. However, none of the new or old AEDs has brought a meaningful change in the overall prognosis of the syndrome (Arzimanoglou et al. 2009). For the very few patients that rapidly became seizure free, we do not dispose of any controlled data on the evolution of the other components of the syndrome, such as intellectual disability and behavioural problems.