The impact of stressful life events on risk of relapse in women with multiple sclerosis: A prospective study

Methods Twenty six ambulating women with relapsing-remitting MS were followed-up for a mean of 56.3 weeks. Patients documented SLEs weekly in self report diaries which were then collected at regular pre-scheduled clinic visits every 4 weeks. SLEs were classified as short-term if they had subjectively no lasting effect and long-term if they had a subjectively felt psychological impact that lasted at least 10–14 days after the event. The severity of SLEs was determined using the Recent Life Change Questionnaire. Results Experiencing three or more SLEs, during a 4-week period, was associated with a 5-fold increase of MS relapse rate (95% CI 1.7–16.4, p = 0.003). The presence of at least one long-term SLE was associated with three times (95% CI 1.01–9.13, p < 0.05) the rate of MS exacerbation during the following 4 weeks. There was no significant association between the severity (95% CI 0.99–1.01, p > 0.05) or the type ( χ 2 = 7.29, df = 5, p > 0.05) of stressor and the risk for relapse. Conclusion Ambulatory women with relapsing-remitting MS who experience cumulative SLEs may be at a greater risk for relapse. Duration is the only stress attribute that seems to increase the risk for relapsing in contrast to stress type and stress severity that were not found to interact with MS exacerbation. Keywords Multiple sclerosis Relapse Risk Stress attributes Stressful life events 1 Introduction Multiple sclerosis (MS) is the most frequent inflammatory, demyelinating disease of the central nervous system affecting approximately 2 million people worldwide with a higher prevalence in northern countries [44] . It affects women at about twice the rate of men, and prevalence appears to be increasing. A course marked by exacerbations and remissions characterizes 85–90% of the MS patients. Clinical relapses reflect an increase of inflammatory activity in the CNS. Every factor that can promote inflammatory activity in the organism may be potentially linked with a relapse [6] . Factors that have been identified as possibly promoting MS relapses are infections, vaccinations, temperature increases, injuries, and the early post partum period [42] . Stressful life events (SLEs) could lead to an increase of inflammatory activity in MS by inducing changes to immune mechanisms that are related to the hypothalamus-pituitary-adrenal axis and to the autonomic nervous system [1,16,29] . Exposure to chronic stress has been shown to reduce the number and function of glucocorticoid receptors on immune cells, reducing their response to cortisol's regulatory control. According to the glucocorticoid resistance hypothesis, an increased concentration of cortisol levels as a response to long-term stress, can induce a down-regulation of glucocorticoid receptors (e.g. reduction in number, binding capacity, and affinity) in immune system cells, thus increasing the risk of inflammation [30] . Additionally, the glucocorticoid resistance produced by a long-term stress may be enhanced by the mild hypercortisolemia generated by a low-grade inflammation process encountered in patients with a relapsing-remitting form of MS [39,41] . This combination effect of long-term SLEs with a small increase in auto-reactive inflammation on the immune system reduces the control of the hypothalamus-pituitary-adrenal axis on inflammation and thus immune cells would be less responsive to the regulatory effects of cortisol [21] . Glucocorticoid resistance has been observed in animals exposed to social stressors and may be a plausible underlying mechanism of the influence of stress in MS [5] . Small increases of cortisol concentrations seen in non-traumatic stress, have been shown to enhance the sensitivity of T cells to a number of cytokines and peptides that promote a proinflammatory response [28,31] . Observations of the relationship of psychological stress to MS can be traced back to the 19th century [33] . However, the validity and the nature of this relationship remain unclear. Some studies indicate a close association between stress and an increased risk for MS exacerbation [2,3,12,27] , while others negate such an association [18,36] or even support, counter-intuitively, a protective role for SLEs [32] . Discrepant results could be attributed to methodological issues, (infrequent monitoring, absence of appropriate control groups, subjective reporting bias, retrospective design, etc.), or to differences in the definition and assessment of “stress”. Franklin et al. evaluated stressful life events in a 20-month follow-up study of 55 patients with MS and found that important events precede clinical relapses [17] . This study was the first to employ a prospective design, however, stressor evaluations were made shortly after the onset of MS symptoms and selective recall of SLEs could have been influenced by the state of relapse. However, other prospective investigations of MS patients, that followed, demonstrated as well a positive relationship between stress and clinical deterioration. Limitations of these studies included not examining specific stressor attributes such as duration, type, or severity of the stressor on MS exacerbations [12,38] , infrequent evaluations of SLEs [9,10] , or study samples which were not homogeneous (various types of MS) [3] . Type of SLEs may be an important issue. Particular types of stressful events such as social conflicts, financial difficulties, physical illness and work-related stress have been reported to be predictive of relapse [26] . However, one study failed to find relations of event type and MS exacerbations [2] . Method of measuring SLEs is another important issue. Interestingly, in one of the studies it was the subjective, rather than the objective, stressor measures that were related to MS disease activity [43] . This may well indicate that the perceived psychological importance of a particular SLE is individualized and may be the crucial factor in determining impact. A recent systematic meta-analysis supports the hypothesis that there is a small but consistent association between SLEs and increased risk for MS exacerbations. In this study reservations are expressed as to whether this relation is causal or indicative of other factors (i.e. underlying inflammatory processes), which may affect both the perception of stress and exacerbation of the disease [28] . In the literature there is a lack of studies clarifying the role of particular features of the stressors involved in the process of MS exacerbation. Stressor type, severity, duration, and timing of stress (in relation to relapse occurrence), may be important predictors of MS relapse and therefore, modification of these factors could influence the disease progress and prognosis [2,3,12] . We hypothesized that the patient`s subjective perception of the duration of the psychological impact of SLEs may be associated with MS relapses. On the basis of the glucocorticoid resistance hypothesis, which indicates that long term stress increases the risk of inflammation, we further hypothesized that long term rather than short term stress will be associated with relapses. Despite strong indications that particular types of SLEs play a role in MS exacerbations, the few studies have examined their impact in the same population have failed to show an association. The categorization of SLEs may have significance for planning therapeutic interventions in clinical practice. Based on the literature [26,40] , we hypothesized that certain categories of SLEs such as those related to health, family, work, and living circumstances, may be associated with MS exacerbation. Data on stressor severity, however, are conflicted. Some have suggested that severity bears a positive relationship to MS relapses, others have demonstrated a relation for moderate stressors only, and others yet, have surprisingly reported an inverse relationship. We therefore included the examination of severity of SLEs in this present study. The findings are far from conclusive. The aims, therefore, of this prospective study were: first, to examine the general relation between SLEs and clinical relapses in ambulatory women with relapsing-remitting MS; second, to investigate the relations of these specific stressor attributes, namely duration, type, and severity, with MS exacerbations. 2 Subjects and methods 2.1 Study participants The recruitment and follow-up details are summarized in Fig. 1 . Participants were female outpatients who met McDonald criteria for definite multiple sclerosis with a relapsing-remitting course [24] . According to McDonald criteria, the diagnosis of MS requires objective evidence of attacks and lesions disseminated in time and space [magnetic resonance imaging (MRI) findings may contribute to the determination of dissemination in time or space; other supportive investigations include cerebrospinal fluid (CSF) and visual evoked potentials (VEPs)]. Subjects agreed to participate in the study for at least 1 year. Patients were randomly recruited from the Outpatient Clinic for Demyelinating Diseases, Eginition Hospital, Athens University Medical School, Greece. All patients were at least 18 years of age and Greek speaking. To enter the study all participants must have suffered at least one relapse in the year preceding the recruitment phase of the study and they had to be currently in remission. To minimize disability bias on perception of stressful life events [35] all participants had to be able to walk independently allowing only for moderate disability in one area (Expanded Disability Status Scale, EDSS ≤3) [23] . Patients were excluded if they had other co-morbid significant medical illness (e.g., thyroid, heart, lung, or kidney disease). Patients were also excluded if they met DSM-IV-TR criteria for substance abuse or dependence, and mental retardation [4] . Exclusion and inclusion criteria were used to ascertain the inclusion in the study of a homogeneous group of stabilized MS patients. Participants who were receiving disease-modifying therapy (e.g. interferon-beta, glatiramer acetate) and/or psychotropic treatment (e.g. antidepressants), were considered stable and were permitted to remain on these medications. Subjects whose immunomodulatory or psychopharmacological regimen was modified or altered during the study were withdrawn at the time of medication change. 2.2 Procedure The study was conducted at Eginition Hospital, of the Athens University Medical School, in Athens, Greece, between December 2005 and July 2007, in accordance with the protocol and ethical principles stated in the Declaration of Helsinki, and the protocol was approved by the institutional review board. Written informed consent was obtained from all participants, following a thorough description of the study procedures and requirements. After a baseline visit, where basic demographic data and disease history information was recorded, regular visits were scheduled every 4 weeks for all patients. Additional visits were to take place within 3 days if patients suspected an exacerbation and the symptoms persisted for more than 24 h. At each regular visit patients had a complete neurological examination while assessment of the neurological impairment was done by using the EDSS. The diaries of self reported SLEs were collected at each regular visit and the examining psychiatrist (always the same person) discussed these events extensively with the patients. Life events that were directly connected to existing signs or symptoms of multiple sclerosis (e.g. missing vacations due to an MS relapse) were excluded from the analysis in order to reduce bias of a possible association between SLEs and MS relapses during a period of high disease activity. 2.3 Assessment 2.3.1 Stressful life events 2.3.1.1 Subjective measures (duration) Each week (every Sunday) patients assessed SLEs in self-report weekly diaries. Patients were asked to describe in their own words any situations or isolated events which occurred within the preceding week and were felt to have a stressful impact on their lives. Patients also reported in the diaries the duration of the subjectively felt psychological impact of the particular SLE after its termination. Specifically, each participant was instructed to characterize a SLE as short-term if it had no lasting effect (impact mainly experienced at the time of the event) and long-term if it implied subjective psychological distress lasting at least 10–14 days after the event (for example in the event of a separation). In this sense duration essentially reflected the chronicity of the psychological impact of the SLE. This classification was based on Brown's definition [7] . The meaning of the review of the diary with the patient was to ascertain, with further probing and questioning, the approximate duration of the psychological impact of the event. The diaries were collected at each regular visit every 4 weeks. If an event lasted more than 4 weeks or was in progress during the interview, it was recorded in the next session. 2.3.1.2 Objective measures (severity, type) The severity of SLEs was determined by the researchers using the Recent Life Change Questionnaire (RLCQ) [25] . The RLCQ, in use since 1975, is a 75-item inventory of potentially stressful life events that has been used in various studies of psychological stress and illness [13,34] . Each event in the RLCQ is assigned a Life Change Unit (LCU). The LCUs are used for the quantification of SLEs and correspond to the psychological energy necessary to adapt to it. A higher LCU score denotes a higher level of stress. The score in RLCQ range from 18 (“correspondence course to help you in your work”) to 123 (“death of an own child”). In order to include all SLEs in the severity measurement, even those that are not described in the RLCQ (e.g. death of a pet), and to avoid underreporting related with checklist methods used by the patients, a method of consensus conferences was employed. The self-report diaries were reviewed and clarified by the interviewing psychiatrist along with the patient, at every regular visit. Consensus conferences were held with the interviewing psychiatrist presenting a formulation of the patient's history to two other psychiatrists (blind to the patients’ physical condition). Each of the two psychiatrists was required to independently rate the severity of any stressful event by assigning a LCU score using the RLCQ. For the SLEs that were not included in the RLCQ, the psychiatrists assigned a score based on the RLCQ. For example, for the death of a pet the consensus conference assigned an LCU score of 47, which was similar to the LCU score of “falling out of a close personal relationship”. This determination was made on the basis of the raters training and judgement. The two psychiatrists that conducted the ratings had become familiar with the performance criteria over a series of competitive training sessions and pre-season friendly matches prior to the testing period, until inter-rater reliability was consistently greater than r = 0.8. To study the possible relation between the type of stress and MS relapses, SLEs were grouped together in six categories: parents/family, social, work/financial, marital/love relationships, health and related problems, living circumstances ( Table 1 ). The formulation of the above categories was based on the RLCQ and on the methodology of previous studies investigating the role of type of stress in MS relapses [2] . 2.3.2 MS disease activity Neurological examination was performed, always by the same neurologist, at baseline and at every regular and additional visit after a suspect exacerbation for the examining neurologist was blind to the SLEs. Exacerbation was defined as a worsening of existing symptoms or appearance of new symptoms lasting more than 24 h, after at least 30 days of improvement or stability, not associated with fever. Following the neurological examination, disability was rated by using the EDSS. The EDDS is the oldest and the most widely used rating system of clinical assessment in MS [37] . The EDSS is rated by half point increments, from 0.0 (normal neurological examination) to 10.0 (death from MS complications). Following the neurological examination, the investigator is required to summarize the results in several “Functional System Scores” which are graded from normal (0) to maximal impairment (5 or 6). The Functional Systems are the following: pyramidal, cerebellar, brain stem, sensory, bowel and bladder, visual, cerebral, and “others”. An overall score for the patient's disability is then obtained by combining the different functional systems grades and the ability to walk, which has to be assessed separately, to provide a score on the full 20-point scale. 2.3.3 Statistical analysis To study the possible association between stressful life events and relapses, the “high risk” interval was set at 4 weeks. After each stressful event patients were considered to enter a period of 4 weeks with high risk of an exacerbation. This choice was based on the “at risk” periods suggested in the literature for various factors affecting the disease process, such as infections (4–5 weeks) [11] , early postpartum period (4 weeks after delivery) [14] , or the possible appearance of gadolinium enhancing lesions on MRI after stressful events (4–8 weeks) [27] . The life event variables included in analyses were the total number of stressors reported by the participant, the duration of each stressor (short-term vs long-term as per patient rating), the severity of each stressor (LCU score as per psychiatrist rating) and the category of each stressor (types) experienced in the 4 weeks preceding MS relapse. For each week of the study patients were grouped according to whether or not they had experienced stress in the 4 weeks immediately preceding it. A Cox univariate regression analysis was used, to calculate if a higher frequency of reported SLEs increases the risk of a relapse. A multivariate Cox regression analysis was used to evaluate simultaneously the association of the high risk period due to the duration (short-long-term) and the severity (LCU score) of the SLEs. Cox Regression analysis is useful for modeling the time to a specified event, based upon the values of given covariates. Even though the number of the study participants was relatively small, the number of ratings, that represent our sample, was high (1250 observations). The longer the study duration, the higher the power is, because a longer period of observation allows us to observe more events. Thus, our sample provided sufficient statistical power for multivariate analysis and protects internal validity. A chi-square test was performed to evaluate whether the type of the event is associated with MS relapses. For the present study it was immaterial if a participant manifested more than one relapse during the observation time. All relapses were evaluated as independent from each other and were not considered cumulative for each patient. In analyzing a new exacerbation, the time axis to originate was considered at the time of the previous exacerbation. Only those exacerbations that occurred after at least 30 days of improvement or stability were eligible for inclusion. Pearson correlations were computed as measures of association for quantitative variables, in order to further explore the relationships between the MS disease activity, SLEs and demographic and clinical characteristics. Paired t -test and t -test were used for determine differences between groups (for normally distributed data). Statistical significance was set as p < 0.05. All statistical analyses were performed using the Statistical Package for Social Sciences for Windows (SPSS 14.0 for Windows, SPSS Inc., Chicago, IL). 3 Results 3.1 Patients’ sociodemographic and clinical characteristics The mean age of participants was 33.5 (SD = 6.7) years and, their mean educational level was 14.1 (SD = 2.2) years. Eighteen of the 26 patients (69.2%) were single, five were married (19.2%), and three were divorced (11.6%). 25 out of 26 (96.1%) women were premenopausal. Their mean duration of disease was 6.5 (SD = 4.4) years, and first manifestation occurred at a mean age of 27 years (SD = 6.1). Patients manifested a mean of 1.2 (SD = 0.5) exacerbations during the previous year. Their mean EDSS score at the beginning of the study was 0.4 (SD = 0.8) and was exactly the same at the end of the study (0.4, SD = 0.7). The fact that no significant change was observed in EDSS between study initiation and completion (paired t -test, t = − 0.49, NS) is consistent with the relapsing-remitting diagnosis. Fourteen of the participants (53.8%) were receiving disease-modifying treatment: nine were receiving interferon beta-1a, four patients were taking interferon beta-1b and, one patient glatiramer acetate. Twelve patients (46.1%) were receiving antidepressants at a mean dose of 133.3 (SD = 44.3) mg/day imipramine equivalents: nine patients were being administered selective serotonin reuptake inhibitors (SSRIs), two were taking noradrenergic and specific serotonergic antidepressants (NaSSAs) and, one patient norepinephrine and serotonin reuptake inhibitors (NSRI). 3.2 Patients, exacerbations and, reported SLEs Twenty one of the 26 (80.7%) patients participated in the study for at least 1 year. Mean follow-up time was 56.3 (SD = 20.8) weeks. Five patients failed to complete at least 1 year of the study follow-up (three were lost to follow-up, one switched disease-modifying therapy and, one increased the antidepressant dosage due to a major depressive episode), but these patients, nevertheless, did participate for a mean of 19.2 weeks (SD = 9.1) per individual. The high variability in mean follow-up times (SD 20.8 weeks) occurred, because the performed statistical analysis (Cox regression) included the drop-outs and those who continued the study for more than 1 year. The demographic and clinical characteristics did not differ significantly between completers ( N = 21) and no completers ( N = 5) of the study. Due to the small sample size of non-completers group, low power comparative analysis was not undertaken. During the study period 44 exacerbations occurred in 21/26 (80.7%) of the patients, and these were confirmed by the study neurologist – a mean of 1.6 (SD = 1.1) exacerbation per year. There was not significant difference between the mean number of MS relapses during the 1 year of the study and the previous year (paired t -test, t = − 1.733, NS). The mean duration of the relapse was 4.1 (SD = 1.6) weeks. The duration of the relapse was calculated from the appearance of a new symptom or the worsening of an existing one, to stabilization of all residual symptoms. The clinical relapses involved the following systems: motor (34.1%, N = 15), sensory (27.2%, N = 12), visual (15.9%, N = 7), brain stem (11.3%, N = 5), cerebellar (9%, N = 4) and, bowel and bladder (2.5%, N = 1). The completed diaries covered 93% (1250/1343) of all weeks of the total follow-up time and were all included in the analysis. In total 507 SLEs were recorded. Sixteen of them were directly related to signs or symptoms of MS and were excluded, resulting in 491 events for analysis. There was a mean 18.7 (SD = 5.8) of reported SLEs per year. A total of 306 of the SLEs (62.3%) had short-term duration and, 185 (37.7%) had a long-term duration. The mean LCU score of SLEs was 38.6 (SD = 12.5). 3.3 Stressor attributes and the risk of exacerbation In total, 90.1% of relapses were associated with one or more SLEs in the previous 4 weeks. Moreover 20.2% of all the reported SLEs were associated with subsequent relapses in the following 4 weeks. Table 2 shows the risks for relapse according to the presence of SLEs in the 4 preceding weeks. Cox univariate regression showed that 1 or 2 SLEs had no significant effect on risk for relapses. Contrarily 3 and ≥4 SLEs were associated with 5.3 and 16.7 times respectively the rate of relapsing during the following 4 weeks. This indicates that stress has a cumulative effect on relapses. A further multivariate Cox regression analysis ( Table 3 ) of the duration and severity of SLEs, showed that the risk for relapsing is influenced only by long-term SLEs and this effect is cumulative. The LCU score of the SLEs and the frequency of short-term SLEs in the precedent 4 weeks had no effect on risk for relapsing. The type of stressor ( Table 1 ) seemed to have no impact on the association between SLEs and subsequent relapses (chi-square test, χ 2 = 7.29, df = 5, NS). There was not a significant correlation between the demographic variables of age [ r (26) = − 0.42, NS], education [ r (26) = −0.14, NS], duration of the disease [ r (26) = − 0.02, NS], and initial EDSS score [ r (26) = − 0.06, NS] with the frequency of the reported SLEs. The number of the SLEs reported in the high risk period did not correlate with the EDSS score [ r (26) = − 0.10, NS], the system involved [ r (26) = 0.01, NS], or the duration of the relapse [ r (26) = 0.02, NS]. The use of antidepressants and disease-modifying treatment, were not associated with a higher frequency of reported stress ( t -test, t 1 = 1.0, NS and t 2 = − 0.95, NS respectively). 4 Discussion We found an association between stressful life events and increased risk for MS exacerbation, in line with a growing body of supporting evidence from other investigators [28] . In our study the effect of stress on relapses was cumulative. Two or fewer SLEs had no effect on risk for relapse. A higher frequency of reported SLEs (three or more) increased the risk of an exacerbation by 5.3 times. These results are consistent with findings from several previous lines of research [2,3,9,10] . There are, however, other studies that have not found that “dose” of stress is relevant to relapses, reporting the relationship as an all-or none phenomenon. Buljevac et al. reported that stress doubled the exacerbation rate during the subsequent 4 weeks and this effect was not “dose dependent” [12] . By examining closely the interactions between stress and MS exacerbations we found a differential effect of stressors. A robust interaction was observed between risk of relapse and perceived duration of stress, but not with type or severity of stressor. The presence of at least one long-term SLE was associated with a 3-fold increase of the risk of exacerbations in the 4 weeks following the stressor. An increase in the rate of long-term SLEs was associated with an increased likelihood of developing a relapse ( Table 3 ). On the other hand, short-term SLEs had no effect in MS disease activity even if present in a high density (≥3 events). According to the previous literature, a possible explanation of this observation could be offered by the glucocorticoid resistance hypothesis [30] . However, the lack of physiological markers of stress in the present study prevents us from further validating this hypothesis. Interestingly there was no significant association between the severity of stressor and the risk for relapse. This is in contrast with the study of Nisipeanu and Korczyn (1993) suggesting that severe stressors may protect MS patients from a relapse [32] . It is known that stress increases the release of cortisol, and cortisol is the main endogenous mechanism of controlling inflammation. Sudden increases of cortisol after traumatic stress as seen in the study of Nisipeanu and Korczyn (during 1 month of missile attacks against Tel Aviv in the first Gulf war) may explain the decreased relapse frequency. This was not the case in our study: war-related stress is not a common condition in our country and such a type of stressor was not experienced by any of the participants of this study, even by patients who had experienced severe stressors (e.g. death of a family member). Furthermore, other studies suggest that it is moderate stressors, rather than severe stressors, that are linked to clinical relapse [8] and new MRI lesions [27] . This is further supported by animal studies of experimental allergic encephalomyelitis, that suggest an association between moderate stress and disease exacerbations while severe negative stressors (e.g. prolonged physical restraint) suppress the clinical and histopathological manifestations of the disease [15] . The results of this study are in accordance with the findings of Ackerman et al., who used a similar categorization of SLEs and found that the type of stress was not significantly related to MS exacerbation [2] . The findings of the present study may enhance the power of those in Ackerman's study, given that the reported life-events were significantly less (215 SLEs) than those in our study (491 SLEs). However, this finding raises some question. It is incongruent with previous studies maintaining that particular types of SLEs (e.g. social conflicts, disruptions in daily routine, death of a close relative) are predictive of disease exacerbations [26,40] . The categorization of SLEs applied in the present study may have played a role in the negative results. Perhaps, other categorizations, the study of particular types of stress separately or a larger number of SLEs reported in every category might have revealed positive associations. On the other hand, subjective experience may play a part on the psychological impact of a particular type of SLE. Indeed in our study it was the subjective measures that yielded positive results. One would need to design a study in which evaluation of the impact of SLEs would score type of event and personal significance simultaneously. Other studies suggest that MS attributes (disability, illness episodes, duration of the disease) [35] and the use of antidepressants and disease-modifying treatment may interfere with the perception of stress [19,22] ; in our study such MS attributes were not associated with a higher density of SLEs. A possible explanation for this observation is that all participants in our study were highly functional (mean EDSS = 0.4) and were not left with a permanent disability status after disease exacerbations. In this study a higher frequency of SLEs 4 weeks before the exacerbation was not associated with major disease activity (increase in EDSS score, system involved, and duration of the relapse). Thus, stress may not be the underlying or the distal cause of MS exacerbation but rather a proximal cause or trigger. Indeed Goodkin et al. demonstrated that changes in white matter are detectable months before the first Gd + MRI lesions [20] . This is one of the first studies to examine prospectively detailed SLEs characteristics and their relationship with relapses in a homogeneous group of patients with relapsing-remitting MS. Every effort was made to maintain a high (weekly) monitoring frequency, in order to minimize subjective recall bias and, to reduce interference of the disease activity on the perception of stress. For this reason SLEs that were directly connected to existing signs or symptoms of the disease were excluded from the analysis. The present study includes both subjective and objective measures of assessment of the SLEs. The use of a checklist method for assessing SLEs, often associated with “underreporting”, was avoided by including in the analysis all the potentially triggering SLEs. Furthermore, the severity rating was made independently (consensus conferences) of the subject's reporting of the events so as to reduce subjective bias. Nevertheless, differences in individual characteristics and contextual details of the SLEs were not taken into consideration in the severity scoring process. This may have introduced some confounders; however, the homogeneity of the sample may have reduced such differences. It should be emphasized that only the subjectively rated measures (number of reported SLEs and duration of stress) rather than the observer-rated ones (severity of stress) were significantly associated with the relapses of MS. Thus, patients’ subjective experience of stress may be of significant importance in the process of relapses. Nevertheless, this approach might be susceptible for biases from other factors that are themselves related to increased risk for exacerbation such as emotional dysregulation. Other study limitations merit acknowledgment: • The number of the participants was relatively small. • A number of SLEs may be the result of subclinical cognitive and behavior changes (e.g. altered attention and concentration) that could precede clinical relapses. • Self report information regarding the timing and duration of SLEs may be influenced by reporter bias. • The sample was heterogeneous with respect to immunomodulatory and antidepressant treatment. • Antidepressants might decrease the relationship between stress and disease exacerbation. • Potential psychological and social mediators (e.g. traumatic history, coping skills, depression) were not directly taken into consideration. • The results may not be fully generalizable to other MS samples (e.g. men, other MS forms, patients with EDSS ≥3). 5 Conclusion In conclusion, we believe that our data provide prospective evidence that SLEs may be a potential trigger of MS exacerbation. Ambulatory women with relapsing-remitting MS who experience cumulative SLEs may be at a greater risk for relapse. To the best of our knowledge, this is the first longitudinal study with a high sampling frequency clearly demonstrating that duration is the only attribute of stress that seems to increase the risk for relapsing, contrary to stress type and severity that appear not to interact with MS exacerbation. A major density of SLEs does not appear to be related to a higher EDSS score and/or duration of the relapse. Our results underscore the importance of early detection and intervention in SLEs in order to achieve a rapid remission of stress and reduce the possibility of relapse. Further long-term clinical trials of preventive psychological interventions (e.g. coping skills training, cognitive behavioral stress management) and/or pharmacological treatments in patients with MS are needed, in order to reduce the likelihood of stress related disease exacerbations. Acknowledgements The authors wish to thank Ms. Eleni Voussoura, for her numerous helpful comments and suggestions. Note: This study was not supported by external funding, nor do the authors have financial or other relationships to disclose relevant to the subject matter of the article. References [1] K.D. Ackerman M. Martino R. Heyman N.M. Moyna B.S. Rabin Immunologic response to acute psychological stress in MS patients and controls J Neuroimmunol 68 1996 85 94 [2] K.D. Ackerman R. Heyman B.S. Rabin B.P. Anderson P.R. Houck E. Frank Stressful life events precede exacerbations of multiple sclerosis Psychosom Med 64 2002 916 920 [3] K.D. Ackerman A. Stover R. Heyman B.P. Anderson P.R. Houck E. Frank 2002 Robert Ader New Investigator award. Relationship of cardiovascular reactivity, stressful life events, and multiple sclerosis disease activity Brain Behav Immun 17 2003 141 151 [4] American Psychiatric Association: Diagnostic and Statistical Manual of Mental disorders 4th ed. 2000 American Psychiatric Association Washington, DC Text revision (DSM-IV-TR) [5] R. Avitsur J.L. Stark J.F. Sheridan Social stress induces glucocorticoid resistance in subordinate animals Horm Behav 39 2001 247 257 [6] A. Bitsch J. Schuchardt S. Bunkowski T. Kuhlmann W. Brueck Acute axonal injury in multiple sclerosis. Correlation with demyelination and inflammation Brain 123 2000 1174 1183 [7] G.W. Brown Life events and measurement G.W. Brown T.O. Harris Life events and illness 1989 Guildford New York 3 45 [8] R.F. Brown C.C. Tennant S.M. Dunn J.D. Pollard A review of stress-relapse interactions in multiple sclerosis: important features and stress-mediating and -moderating variables Mult Scler 11 2005 477 484 [9] R.F. Brown C.C. Tennant M. Sharrock S. Hodgkinson S.M. Dunn J.D. Pollard Relationship between stress and relapse in multiple sclerosis: Part I. Important features Mult Scler 12 2006 453 464 [10] R.F. Brown C.C. Tennant M. Sharrock S. Hodgkinson S.M. Dunn J.D. Pollard Relationship between stress and relapse in multiple sclerosis: Part II. Direct and indirect relationships Mult Scler 12 2006 465 475 [11] D. Buljevac H.Z. Flach W.C. Hop D. Hijdra J.D. Laman H.F. Savelkoul Prospective study on the relationship between infections and multiple sclerosis exacerbations Brain 125 2002 952 960 [12] D. Buljevac W.C. Hop W. Reedeker A.C. Janssens F.G. van der Meche P.A. van Doorn Self reported stressful life events and exacerbation in multiple sclerosis: prospective study BMJ 9 2003 307 310 [13] B. Caplan C.J. Gibson R. Weiss Stressful sequelae of disabling illness Int Rehabil Med 6 1984 58 62 [14] C. Confavreux M. Hutchinson M.M. Hours P. Cortinovis-Tourniaire T. Moreau Rate of pregnancy-related relapse in multiple sclerosis. Pregnancy in Multiple Sclerosis Group N Engl J Med 339 1998 285 291 [15] S.G. Correa M.C. Rodriguez-Galan V.E. Rivero C.M. Riera Chronic varied stress modulates experimental autoimmune encephalomyelitis in Wistar rats Brain Behav Immun 12 1998 134 148 [16] F.W. Foley A.H. Miller U. Traugott N.G. LaRocca L.C. Scheinberg J.R. Bedell Psychoimmunological dysregulation in multiple sclerosis Psychosomatics 29 1988 398 403 [17] G.M. Franklin L.M. Nelson R.K. Heaton J.S. Burkes D.S. Thompson Stress and its relationship to acute exacerbations in multiple sclerosis J Neurol Rehabil 2 1988 7 11 [18] C. Gasperini M.G. Grasso M. Fiorelli E. Millefiorini S. Morino A. Anzini A controlled study of potential risk factors preceding exacerbation in multiple sclerosis J Neurol Neurosurg Psychiatry 59 1995 303 305 [19] J.L. Goeb C. Even G. Nicolas B. Gohier F. Dubas J.B. Garré Psychiatric side effects of interferon-beta in multiple sclerosis Eur Psychiatry 21 2006 186 193 [20] D.E. Goodkin W.D. Rooney R. Sloan P. Bacchetti L. Gee M. Vermathen A serial study of new MS lesions and the white matter from which they arise Neurology 51 1998 1689 1697 [21] C. Heesen D.C. Mohr I. Huitinga F.T. Bergh J. Gaab C. Otte Stress regulation in multiple sclerosis: current issues and concepts Mult Scler 13 2007 143 148 [22] A.C. Janssens D. Buljevac P.A. van Doorn F.G. van der Meché C.H. Polman J. Passchier Prediction of anxiety and distress following diagnosis of multiple sclerosis: two-year longitudinal study Mult Scler 12 2006 794 801 [23] J.F. Kurtzke Rating neurologic impairment in multiple sclerosis: an expanded disability status scale (EDSS) Neurology 33 1983 1444 1452 [24] W.I. McDonald A. Compston G. Edan D. Goodkin H.P. Hartung F.D. Lublin Recommended diagnostic criteria for multiple sclerosis: guidelines from the International Panel on the diagnosis of multiple sclerosis Ann Neurol 50 2001 121 127 [25] M.A. Miller R.H. Rahe Life changes scaling for the 1990s J Psychosom Res 43 1997 279 292 [26] D.C. Mohr L.P. Dick D. Russo J. Pinn A.C. Boudewyn W. Likosky The psychosocial impact of multiple sclerosis: exploring the patient's perspective Health Psychol 18 1999 376 382 [27] D.C. Mohr D.E. Goodkin P. Bacchetti A.C. Boudewyn L. Huang P. Marrietta Psychological stress and the subsequent appearance of new brain MRI lesions in MS Neurology 55 2000 55 61 [28] D.C. Mohr S.L. Hart L. Julian D. Cox D. Pelletier Association between stressful life events and exacerbation in multiple sclerosis: a meta-analysis BMJ 328 2004 731 [29] D.C. Mohr D. Pelletier A temporal framework for understanding the effects of stressful life events on inflammation in patients with multiple sclerosis Brain Behav Immun 20 2006 27 36 [30] D.C. Mohr Stress and multiple sclerosis J Neurol 254 Suppl 2 2007 65 68 [31] A. Munck A. Náray-Fejes-Tóth Glucocorticoids and stress: permissive and suppressive actions Ann N Y Acad Sci 746 1994 115 130 [32] P. Nisipeanu A.D. Korczyn Psychological stress as risk factor for exacerbations in multiple sclerosis Neurology 43 1993 1311 1312 [33] J.M. Pearce Historical descriptions of multiple sclerosis Eur Neurol 54 2005 49 53 [34] V.R. Pendleton E. Willems P. Swank W.S. Poston G.K. Goodrick R.S. Reeves Negative stress and the outcome of treatment for binge eating Eat Disord 9 2001 351 360 [35] C.E. Schwartz F.W. Foley S.M. Rao L.J. Bernardin H. Lee M.W. Genderson Stress and course of disease in multiple sclerosis Behav Med 25 1999 110 116 [36] G. Serlupi Crescenzi Risk factors in multiple sclerosis: implications for pathogenesis G. Serlupi Crescenzi A multidisciplinary approach to myelin disease 1987 Plenum Press New York 227 232 [37] B. Sharrack R.A. Hughes Clinical scales for multiple sclerosis J Neurol Sci 135 1996 1 9 [38] W.A. Sibley Risk factors in multiple sclerosis C.S. Raine H.F. McFarland W.W. Tourtelotte Multiple sclerosis: clinical and pathogenetic basis 1997 Chapman & Hall London 141 148 [39] N.C. Silver C.D. Good G.J. Barker D.G. MacManus A.J. Thompson I.F. Moseley Sensitivity of contrast enhanced MRI in multiple sclerosis. Effects of gadolinium dose, magnetization transfer contrast and delayed imaging Brain 120 1997 1149 1161 [40] E. Stip J.L. Truelle Organic personality syndrome in multiple sclerosis and effect of stress on recurrent attacks Can J Psychiatry 39 1994 27 33 [41] L.M. van Winsen D.F. Muris C.H. Polman C.D. Dijkstra T.K. van den Berg B.M. Uitdehaag Sensitivity to glucocorticoids is decreased in relapsing remitting multiple sclerosis J Clin Endocrinol Metab 90 2005 734 740 [42] T. Vollmer The natural history of relapses in multiple sclerosis J Neurol Sci 256 Suppl 1 2007 5 13 [43] S. Warren K.G. Warren R. Cockerill Emotional stress and coping in multiple sclerosis (MS) exacerbations J Psychosom Res 35 1991 37 47 [44] B.G. Weinshenker Epidemiology of multiple sclerosis Neurol Clin 14 1996 291 308