OnaBotulinumtoxin-A in Chronic Migraine Patients With Short or Long Disease History (the BACH Study)

Onabotulinumtoxin-A (OBT-A) is a worldwide approved preventive treatment for patients suffering from Chronic Migraine (CM, 1.3, ICHD III 2018). Its effectiveness in reducing headache frequency and disability has been demonstrated by two well-designed phase III clinical trials, PREEMPT 1 and 2 and, more recently, by some real-life studies.

However, the use of OBT-A in the real world clinical practice has been still raising some issues: i.e. searching for more adequate patients' selection, when discontinuing the OBT-A treatment and when resuming it, finding associated patients' psychopathological profile possibly influencing the treatment, and so on. Moreover, finding some indicators of good or bad prognosis coming from OBT-A real-life experience could be of help for the clinicians to individuate the best candidates to the treatment.

The aim of the present study is mainly to investigate whether the history of Chronic Migraine and, more precisely, its duration for over or less than 10 years, can predict OBT-A treatment effectiveness. In other words, the study will aim to establish whether starting OBT-A treatment earlier in CM patients' life could make it more effective.

Since psychiatric symptoms may represent prognostic factors influencing treatment effectiveness, the present study will also aim at evaluating if the psychopathological profile of the enrolled patients may influence the outcome.

Consecutive CM patients treated in Italian tertiary level Headache Centers and classified according to CM history (shorter/longer than 10 years) will be enrolled.

Aims

To compare OBT-A treatment effectiveness in patients with shorter (1-9 yrs) and longer (10 or longer) years history of CM, in terms of:

• monthly migraine (MMD) or headache (MHD) days
• monthly acute medications (MAM)
• disability scales (Headache Impact Test, HIT-6; Migraine Disability Assessment MIDAS questionnaire) scores
• pain intensity and quality

Objectives

• to investigate whether the history of Chronic Migraine (over or less than 10 years) can predict OBT-A treatment effectiveness
• to find possible indicators of good or bad prognosis coming among psychiatric comorbidities of the enrolled patients

Consecutive CM patients (according to ICHD-3 1.3) treated in Italian headache centers, undergoing OBT-A, will be enrolled.

Methods All patients will fill a headache diary to report MMDs and acute medications taken every month.

Disability will be assessed by means of HIT-6 and MIDAS. Moreover, the intensity and quality of perceived pain will be evaluated by means of Numerical Rating Scale (NRS), 11-point Box Scale (BS-11), Present Pain Intensity (PPI), Behavioral Rating Scale (BRS-6) e Short-form McGill Pain Questionnaire (SF-MPQ).

The psychopathological profile will be evaluated at different times of the study by means of the following psychological questionnaires: Beck Depression Inventory (BDI-II), State-Trait Anxiety Inventory (STAI-Y), Toronto Alexithymia Scale (TAS-20).

At enrolment (T0), before starting the treatment, patients will have to fill a questionnaire about the previous 3 months (screening phase) including:

• MMD and MHD
• MAM
• NRS, BS-11, PPI, BRS-6 e SF-MPQ scales
• MIDAS e HIT-6 scale
• BDI-II, STAI-Y, TAS-20 questionnaires After completing the diary and scales above, patients will receive OBT-A treatment according to PREEMPT protocol.

Patients will be divided into 2 groups according to their CM history (duration): A) those from 1 to 9 years and B) those with 10 or more years of CM history.

Patients will be re-evaluated every 3 months, at each OBT-A cycle, during the whole year of treatment. At every evaluation time patients will show their diary of headaches with MMD and MAM taken.

At T3 and T9 (3 and 9 months respectively after T0), patients will fill the questionnaire including MMD and MAM taken in the previous 3 months, pain evaluation scales, MIDAS e HIT-6 scales.

At T6 and T12 (6 and 12 months respectively from T0, T12 being the final evaluation) patients will have to fill all questionnaires/scales as at T0.

Outcomes Primary endpoint Change in the mean number of MMD (or MHD) from baseline to months 6 through 12 between groups

Secondary

1. 50% or greater reduction in MMD from baseline to months 6 through 12 between groups
2. change in the number of MAM from baseline to months 6 through 12 between groups
3. change in pain intensity and quality scores from baseline to months 6 through 12 between groups
4. change in terms of psychological profile (STAI-Y, TAS-20 scales) in the stratified subgroups based on Z score of BDI-II scale (< 0 ≥ 1,8) between groups

Statistical analysis The data will be collected in an electronic database in anonymous form that will be protected by a password for each participant center.

On the basis of primary outcome (change during the follow-up vs. baseline of MMD), a general linear model with mixed effects will be applied, which will allow to evaluate, considering the within-subject dependencies and allowing to use all the observations available for each subject (even if for example the measurement of a subject at the time Tx should be missing), the interaction time X group and estimate if and when during follow-up a significant difference between the two groups will occur. According to a previous study (Vernieri et al, Headache 2019), the distribution of MMD is expected to follow a Gaussian distribution, therefore the general linear model will be a mixed-effects ANOVA. For some of the other variables (such as the number of analgesics or MIDAS) a generalized model will be applied to take into account the non-Gaussian nature of their distributions.

The sample size has been predefined according to the primary outcome with the following approach. Thanks to the previous study (Vernieri et al, 2019), it was possible to estimate the standard deviations (SD) of the changes of MMD between baselines and subsequent times. The maximum standard deviation was observed between baseline (T0) and T4 (1 year) and was 8 (with very similar values in two independent centers: 7.8 and 8.2). Since this SD was estimated on 100 patients, the 95% confidence interval for SD is 7.0-9.3. In order to avoid underestimation of the variability of the primary outcome, we have therefore set SD equal to the upper limit of the confidence interval built on the maximum SD found of the variations (vs. baseline) during the follow-up and that is SD=9.3.

In the following table it is reported the sample size required to have a power of 90% to recognize as statistically significant (at two-tails alpha level set at 0.05) the differences between the expected decreases specified in the first two columns. Since the evaluation times at follow-up will be 4 (T3, T6, T9, T12), the probability of error of type I has been divided by 4, in order to keep the overall alpha error below 5%.

In the cited study a decrease after 6 months of treatment (and confirmed at 12 months) of about 7 days of migraine/month was observed. This absolute decrease corresponded to a percentage change of about 30-35%. The hypothesis is that chronic patients from more than 10 years may benefit less from treatment and, in particular, that between the two groups there is a difference in efficacy of 4 (hypothesis 1) and 6 (hypothesis 2) days/month.

Hypothesis 1 Expected pre-post decrease Group1 (>=10y) 5 (SD=9.3) Expected pre-post decrease Group1 (<10y) 9 (SD=9.3) Power (1-b) 90% a (two-tails) 0.0125 N Group1 115 N Group1 115

Hypothesis 2 Expected pre-post decrease Group1 (>=10y) 4 (SD=9.3) Expected pre-post decrease Group1 (<10y) 10 (SD=9.3) Power (1-b) 90% a (two-tails) 0.0125 N Group1 52 N Group1 52

With these numbers it is also possible to calculate that there will be a probability of 90% to recognize as statistically significant (at alpha level 0.05) standardized effect size f=0.08 (hypothesis1) and f=0.12 (hypothesis 2) for the interaction Time X Group. The calculation for the interaction was based on an ANOVA with 4 repeated measurements after T0 (T3, T6, T9, T12), assuming the same correlation between the baseline follow-up assessments observed in the study by Vernieri et al, 2019 and applying a correction for lack of sphericity epsilon=0.9. Considering that f=0.10 is conventionally classifiable as "small", the indicated number of values therefore allow to be sensitive to small variations (even if clinically relevant).

Reference

Vernieri et al. Onabotulinumtoxin-A in Chronic Migraine: Should Timing and Definition of Non-Responder Status Be Revised? Suggestions From a Real-Life Italian Multicenter Experience. Headache. 2019 Sep;59(8):1300-1309.