The present analysis is devoted to the elucidation of the role of age, age at seizure onset, and epilepsy duration on HRQOL in an Italian multicentre study on epilepsy patients based on the validated and specific Epi-QoL questionnaire. It emerges that, if no other known correlates of the overall Epi-QoL score are considered, age and duration of epilepsy can be expected to have a significant consistent negative association, and age at onset a positive association of limited significance (i.e. significant in one model only), with HRQOL in epilepsy. However, none of the age-related variables showed a significant effect on the overall Epi-QoL score, after adjusting for a selected set of confounders including demographic and clinical factors, and, in particular, the seizure frequency in the preceding 12 months.
The main conclusion of a modest role of age-related factors as determinants of HRQOL is consistent with the vast majority of the literature reports on the determinants of HRQOL in epilepsy patients [11
], including the only available study published on the same topic to our knowledge [10
], where the effects of age at onset and duration were no longer significant, after adding Profile of Mood States Depression/Dejection, Adverse Events Profile, comorbidities, and antiepileptic drugs to the regression models.
Nonsignificance of age-related factors has rarely been reported in simple models [10
], but is more common when results come from multiple models including age-related factors and other clinical variables, such as measures of psychiatric comorbidity, number of antiepileptic medications, and seizure severity, as in Szaflarski and coauthors [10
]. The instrument used to measure HRQOL may also play a role, as specific measures were found to be more sensitive in detecting variation in age at onset and other clinical variables than generic measures [14
Significant but modest effects of age-related factors on HRQOL in epilepsy patients have been reported in other papers, even after adjustment for other known important determinants [14
]. However, some inconsistencies emerged in the sign of the effect of the age-related factors on HRQOL.
In our analysis, age was a relevant determinant of HRQOL: its negative impact on HRQOL emerged from all the three type-1 models, with statistical significance of the coefficients given in two of them. This is in accordance with findings from an analysis based on adolescents [19
], where older patients (14-17 years), independent of epilepsy severity, reported worse overall HRQOL than did their younger counterparts, and with several studies on chronic health problems and HRQOL [45
]. However, in one of the few reported studies that focuses on age in epilepsy adults, Pugh and colleagues [18
] reported that older adults (65 years and older) seemed to have less compromised QOL (as measured using a generic health status measure, the SF-36) than young (18-40 years) and middle-aged (41-64 years) adults. Ageing is associated with a mean decline in learning and memory performances, and, although this decline was apparently similar for epilepsy patients and healthy subjects, the former group reached poor performance levels much earlier than the latter, as epilepsy patients may fail to build up adequate learning and memory performance during childhood and adolescence [9
]. This may have important implications in terms of a limited HRQOL. Moreover, age is generally associated with an increased number of chronic health problems, and multiple chronic health problems are significantly associated with a reduced HRQOL, in particular with reduced levels of physical and emotional functioning [45
]. In patients with a previous diagnosis of epilepsy, the ability of coping with new chronic health problems and their consequences, together with epilepsy severity, may be crucial in the evaluation of the role of age on HRQOL.
From our analysis, it emerged that duration of epilepsy had a consistent significant negative effect on HRQOL in both single and pairwise unadjusted models. This is still in agreement with findings from studies on epilepsy adolescents, where a longer duration of epilepsy had a significant negative impact on the memory and concentration subscale of QOLIE-AD-48 [19
] and on the overall HRQOL [17
], and with various reports on the effects of disorder duration for other chronic conditions, such as Parkinson's disease [46
]. However, in Szaflarski and coauthors [10
], duration had a significant positive effect on HRQOL in the simple regression model. Even in presence of working adjustments of the social and psychological consequences of the disorder and coping mechanisms, cognitive decline and emotional-behavioral distress, together with a potentially increased number of drugs, may lead epilepsy subjects to give a poorer evaluation of HRQOL. Indeed, some studies assessing verbal learning and memory skill of epilepsy patients with ad-hoc inventories suggested a slow cognitive decline with increased duration of epilepsy [32
]. A relevant investigation [28
] demonstrated that increasing years of duration of temporal lobe epilepsy was modestly associated with increased and generalized self-reported emotional-behavioral distress, even after adjustment for potentially confounding clinical variables (including age of onset), and that this comorbid emotional-behavioral distress was significantly associated with a poorer HRQOL.
Finally, in our paper, age at onset emerged as a positive predictor of limited significance (in one model out of three) for the overall Epi-QoL. Accordingly, consistent evidence has been reported from studies on epilepsy children and adolescents, with two studies [14
] suggesting that an earlier age of epilepsy onset was associated with poorer HRQOL. In one study on adults who had operations for refractory extratemporal epilepsy [24
], onset at an older age predicted a minimal increase in the overall health scores, but not in the total HRQOL score, in both single and multiple models. On the contrary, the paper by Szaflarski and coauthors [10
] and two others [21
] suggested that adding years to age at onset decreased HRQOL. In detail, in the same population of young epilepsy adults with mild intellectual disabilities, epilepsy/disability onset in adolescence (after the age of 10 years, compared with disability onset during childhood) was associated with poorer HRQOL, both independently and in interaction with neuroticism [21
]. The first set of studies, including our analysis, provided data supporting the association between poorer seizure control, diminished neuropsychological functioning, and low HRQOL in patients with an earlier onset of epilepsy. However, the role of age at epilepsy onset deserves further attention, especially in its joint effect with age, in consideration of how epilepsy can interfere with brain maturation and then can affect cognitive functioning in the long term [9
A possible explanation of the differences in significance and sign of the age-related beta coefficients across the papers may be found in the different set of covariates included as potential confounders in the regression models, and in the different populations under consideration. For instance, while Szaflarski and coauthors [10
] restricted the analyses to adult patients with medication-resistant epilepsy from a US epilepsy monitoring center, we based our analyses on the overall study population of adult patients recruited in secondary and tertiary Italian centers for the care of epilepsy. When we selected the subgroup of medication-resistant epilepsy patients and carried out the same pool of analyses, the significant associations previously identified in the type-1 models disappeared, and none of the age-related factors were found to be relevant in the explanation of HRQOL. Although nonsignificant, the signs of the linear associations between age-related factors and HRQOL were more similar to those in Szaflarski et al. [10
]: age at onset emerged as negatively associated with HRQOL across all the fitted models, and the estimated beta coefficient for duration changed its sign to positive in the pairwise model with age (data not shown).