Abstract
The aim of this study was to compare the effectiveness of strong advice to withdraw the overused medication with the effectiveness of two structured pharmacological detoxification strategies in a cohort of patients diagnosed with probable migraine overuse headache (MOH) plus migraine and presenting low medical needs. One hundred and twenty patients participated in the study. Exclusion criteria included: previous detoxification treatments, coexistent medical or psychiatric illnesses and overuse of agents containing opioids, benzodiazepines and barbiturates. Group A received only intensive advice to withdraw the overused medication. Group B underwent a standard out-patient detoxification programme (advice + prednisone + preventive treatment). Group C underwent a standard inpatient withdrawal programme (as in group B + fluid replacement and antiemetics). Withdrawal therapy was considered successful if, after 2 months, the patient had reverted to an episodic pattern of headache and to an intake of symptomatic medication on fewer than 10 days/month. We were able to detoxify 75.4± of the whole cohort, 77.5± of patients in group A, 71.7± of patients in group B and 76.9± of those in group C (P > 0.05). In patients with migraine plus MOH and low medical needs, effective drug withdrawal may be obtained through the imparting of advice alone.
Introduction
In the past 15 years, clinical experience and scientific studies on medication overuse headache (MOH) have accumulated, providing the basis for a revision of the International Classification of Headache Disorders (ICHD-II) (1) diagnostic criteria and for the proposal of other, novel classification systems (2). Although the best system for classifying chronic migraine is still much debated by headache specialists (3), MOH (formerly called ‘drug-induced headache’) is now an International Headache Society-classified clinical entity (1). The introduction of universally accepted operational diagnostic criteria for MOH, which fix at 10–15 days per month the critical intake of symptomatic drugs required for a primary headache to become a chronic form, is expected to help physicians to recognize MOH and thus to remove the diagnostic barriers to the proper care of affected patients. In this scenario there is an increasing need for evidence-based and cost-effective withdrawal strategies for MOH.
Drug withdrawal is the treatment of choice for MOH (4–9). Currently, there are no universally accepted standardized therapeutic protocols and no specific guidelines for controlled trials in MOH (4, 6–8, 10). Almost no prospective placebo-controlled trials have been conducted in this field, in which treatment effectiveness is inferred from the results of open-label trials, retrospective case reviews, anecdotal observations, clinical experience and generalization from the literature on the primary headaches. As a result, therapeutic recommendations for the acute phase of detoxification vary considerably among studies (10–14). Basically, the aims of detoxification therapy are (14): (i) to withdraw the overused drug, and (ii) to treat the withdrawal symptoms by means of a bridging programme of pharmacological and non-pharmacological support, designed to help the patient tolerate the withdrawal process (4, 10, 12–14). Medication withdrawal may be either an abrupt discontinuation or a gradual reduction of the overused drug. Many different, structured strategies have been suggested to treat the symptoms of withdrawal. These include the use of antiemetics, fluid replacement, analgesics or triptans, sedatives, amitriptyline, intravenous dihydroergotamine, valproate and other preventive medications before or after the withdrawal period, neuroleptics, cortisone, relaxation therapies, lidocaine and many others (4, 12–18).
Drug detoxification may be performed as an in-patient or out-patient programme (4, 10, 12–14). A consensus paper by the German Migraine Society recommends out-patient withdrawal for highly motivated patients not overusing barbiturates, opioids or tranquillizers with their analgesics (19). In spite of this, many physicians prefer in-patient programmes (2, 20–22), although they are not a viable option for many patients and their cost-effectiveness has not been evaluated.
There is a general consensus that effective education of MOH sufferers is crucial to proper management of MOH and that most patients respond responsibly when the situation is explained to them (9–12, 23). Some authors feel that, in order to reverse the previously learned illness behaviour and reinforce the patient's active role in the management of their headache, no transitional therapy at all should be given during the acute phase of withdrawal (9, 24).
Years of experience in the management of smoking and other substance abuse disorders have shown that the simple imparting of advice to discontinue the overused or abused substance is one of the most cost-effective interventions in medicine (25, 26). In a retrospective analysis of the effects of drug withdrawal therapy in patients with chronic daily headache and drug misuse, Linton-Dahlöf et al. concluded that out-patient abrupt drug withdrawal constitutes an efficacious and pragmatic first-line treatment of MOH (9).
The effectiveness, in MOH, of doctor's advice alone (i.e. without adjunctive pharmacotherapy) has not yet been established. The aim of this study was to compare the effectiveness of strong advice to withdraw the overused medication with the effectiveness of two structured pharmacological detoxification strategies in a cohort of patients with probable MOH plus migraine, diagnosed according to the ICHD-II criteria and presenting low medical needs, i.e. not requiring specific additional medical interventions.
Methods
Consecutive new patients, aged 16–65 years, suffering from probable MOH plus migraine, diagnosed in accordance with the ICHD-II criteria (1) and attending a subspecialty headache centre (INI, Grottaferrata, Headache Clinic), were evaluated over a 12-month period (September 2004 to August 2005). The diagnosis of MOH plus migraine was made by experienced headache specialists (P.R. and G.N.). During this research period, a revision of the ICHD-II diagnostic criteria for MOH was published (27). The revised criteria did not alter the diagnosis of probable MOH, but for the addition of a subgroup of patients with probable MOH attributed to a combination of acute medications. It has been established that abrupt discontinuation of an overused medication may lead to serious withdrawal symptoms, particularly in patients overusing opioids, barbiturates and tranquillizers, and that these symptoms may be dangerous in patients with significant complicating medical conditions (2, 4, 7, 10, 16). We thus recruited only subjects with low medical needs, unlikely to experience problems as withdrawal treatment out-patients. For ethical reasons and in accordance with the recommendations of the German Migraine Society (19), patients with comorbid major depression, anxiety disorder and substance addiction disorders were not included. Thus, the final set of exclusion criteria were: (i) a current diagnosis or history of coexistent, significant and complicating medical illnesses (which could complicate withdrawal undertaken as an out-patient); (ii) a current diagnosis (fulfilment of diagnostic criteria in the past month) of mood disorder, anxiety disorder or addiction disorder (for substances other than the overused medication), assessed using the Structured Clinical Interview for DSM-IV Axis I Disorders (SCID-I), Clinician Version (28), administered at the end of the initial visit by a trained physician (C.d.L.); (iii) overuse of agents containing opioids, barbiturates (27) and benzodiazepines; (iv) treatment with migraine prophylactic drugs within the past 3 months; (v) previous detoxification treatments; (vi) inability to furnish reliable information about medical history and psychiatric symptoms; and (vii) contraindications to the use of corticosteroids or indomethacin.
The patients were evaluated prospectively and the study lasted 12 weeks for each subject (Fig. 1). After the first visit, eligible subjects were asked to keep a headache diary for 4 weeks simply to record their headache pattern and drug use (baseline period). This period was purely observational and at this time no diagnosis or therapeutic indications were given. At the second visit, 4 weeks later, the patients still fulfilling the inclusion criteria were randomly assigned (using a computer-generated random number sequence), in equal numbers, to three different treatment groups. Group A received only advice to withdraw the overused medication; Group B underwent a standard out-patient detoxification programme consisting of: (i) advice to withdraw abruptly the overused medication, (ii) prednisone p.o. for the first 8 days (60 mg/day, 2 days; 40 mg/day, 2 days; 20 mg/day, 4 days) (15), (iii) personalized preventive treatment, starting on day 1 (the preventive agent was chosen on the basis of the drug's side-effect profile, comorbid and coexistent conditions, and the patient's needs, preferences and previous therapeutic experiences); Group C underwent a standard in-patient detoxification programme. After receiving advice about the need to withdraw symptomatic medications, the patients were admitted to hospital (INI Headache Inpatient Clinic), where they received the following treatment: (i) abrupt discontinuation of the overused medication; (ii) close observation and support for 8 days; (iii) prednisone p.o. (60 mg/day, 2 days; 40 mg/day, 2 days; 20 mg/day, 4 days); (iv) personalized preventive treatment, starting on day 1; and (v) parenteral fluid replacement and administration of antiemetics (metoclorpramide 10 mg i.v., b.i.d.). In order to standardize the educational part of the therapeutic management, all advice was issued by the same physician (P.R.). The advice to discontinue the overused medication was given verbally. The imparting of this advice took about 15 min of the second consultation and was structured as follows:
Outline of the trial.
The role of medication overuse in making headache chronic and in reducing the effectiveness of preventive and behavioural treatments was explained.
The phenomenon and symptoms of withdrawal headache were explained in detail.
The beneficial long-term effects, on migraine natural history, of reducing symptomatic medication intake were emphasized (including the reduction of the reinforcing properties of short-term pain relief).
Anticipatory use of medication was discouraged.
The need for a detoxification programme prior to other therapeutic options was emphasized.
In Group A, the physician stressed the importance of patients playing an active role in the management of their headache, without relying solely on medication.
All the patients were prescribed indomethacin 100 mg (p.r. or i.m.), naproxen 500 mg p.o. or eletriptan 40 mg p.o. for the treatment of withdrawal headache, to be taken at a maximum frequency of two doses per week (the rationale was that patients should not use the drug they had previously abused). In Group A, 22 patients received indomethacin, 10 received naproxen and eight received eletriptan. In Group B, 18 patients received indomethacin, 12 received naproxen and nine received eletriptan. In Group C, 20 patients received indomethacin, 13 received naproxen and six received eletriptan. At the beginning of the study, eletriptan had been randomly selected (by lot) from among the five different triptans currently available in Italy.
The patients were also asked to fill in a detailed headache diary. The patients assigned to the out-patient treatment groups were also told that they could contact the Headache Clinic for medical help and support, should this be needed.
Follow-up visits were scheduled for 1 month and 2 months after the start of the detoxification programme. Patients whose chronic headache showed no improvement 2 months after withdrawal were excluded from the analysis. All patients not attending the follow-up visits were telephoned 3 months after their assignment to the treatment group in order to ascertain their reasons for not adhering to the treatment programme.
The patients were not informed that they were participating in a study assessing the effectiveness of withdrawal therapies for MOH. Our approach was the same as that of Warner (29) and Krymchantowski and Moreira (30) and was authorized by the Ethics Committee of our institute. Therefore, no signed informed consent was obtained. Patients assigned to Groups B and C were told that the prescribed drugs could decrease the expected withdrawal headache and symptoms.
Outcome measures
Our primary outcome measures were: (i) the number of responders, defined as those subjects who, 2 months after the onset of withdrawal treatment, experienced no headache or had reverted to an episodic pattern of headache (<15 headache days/month) and whose intake of symptomatic medication was <10 days/month (we decided to allow for intake of symptomatic medication in the definition of responders because it is a more rigorous measure of successful outcome and because, in some cases, anticipatory use of symptomatic drugs may lead to patients becoming headache free). Patients not experiencing any improvement of their chronic headache after withdrawal were considered not to be affected by MOH and excluded from the statistical analysis; (ii) adherence to the treatment, expressed as the number of patients who completed the study.
Our secondary outcome measures were the percentage reduction, 2 months after the start of the detoxification programme, in the number of headache days/month and in the number of symptomatic medication doses/month, compared with the baseline period.
Power calculation and statistical analysis
The main target parameters were the number of responders and the number of patients completing the study (see Outcome measures). On the basis of the results of previous studies, we assumed a responder rate of 40% in the no-intervention group and of 75% in the structured intervention groups, i.e. the study had to have the power to detect a difference of 35%. In applying the above-mentioned percentages, using a power of 0.8% and a significance level of 5%, a minimum of 35 patients per treatment group was required (Fisher's exact test, two-tailed). We thus needed a total of 105 patients to complete the study. Given the impossibility of performing a satisfactory a priori analysis of the other outcome measures, the adherence to treatment criterion and the secondary target criteria were subjected to exploratory statistical analysis.
Bivariate comparisons among groups were performed using the t-test for continuous variables, non-parametric Wilcoxon and Mann–Whitney tests for ordinal data and the χ2 test and Fisher's exact test (when applicable) for categorical variables. Given that we were performing multiple testing, the level of significance was set at P < 0.025. Analysis was carried out using SPSS 11 for Windows (SPSS Inc., Chicago, IL, USA).
Results
Two-hundred and thirty patients were evaluated during the study period; 103 were excluded because they did not meet the inclusion criteria. Of the remaining 127 patients enrolled in the study, seven withdrew during the baseline period, whereas two patients (one in Group B, one in Group C) did not experience any improvement of their headache 2 months after drug withdrawal and were excluded from the analysis. In these two patients a diagnosis of chronic migraine was made on the basis of the diary data and a diagnosis of MOH was excluded. Table 1 presents the demographic and clinical characteristics of the patients randomized to the three detoxification groups. Of the 118 patients constituting the final sample, 99 were females (83.8%) and 19 (16.2%) males and their mean age was 43.97 ± 12.9 years. Sixty-three (53.5%) patients overused analgesics, 21 (17.9%) overused combinations of analgesics (in all cases, indomethacin plus caffeine plus proclorperazine), three (2.5%) overused ergots, 24 (20.3%) overused triptans and seven (5.6%) overused a combination of acute medications (in all cases, triptans plus analgesics). No significant differences emerged between the three groups with regard to socio-demographic variables, migraine subtype, migraine duration, MOH duration, number of headache days per month or number of doses and type of overused medication (Table 1, P > 0.025).
Demographics and headache characteristics of the study population
Group A, Simple advice; Group B, standard out-patient detoxification programme (advice + preventive treatment + prednisone); Group C, standard in-patient detoxification programme (advice + preventive treatment + prednisone + fluid replacement and antiemetics). NS, Not significant.
One patient in Group B and one patient in Group C were diagnosed at follow-up as affected by chronic migraine and excluded from the statistical analysis.
In Group B, 14 patients received valproic acid (500–1000 mg/day) as preventive medication, 12 received β-blockers (atenolol 75–100 mg/day), 10 received amitriptyline (25–75 mg/day) and three received topiramate (75 mg/day). In Group C, 15 patients received valproic acid (500–1000 mg/day) as preventive medication, 10 received β-blockers (atenolol 75–100 mg/day), nine patients received amitriptyline (30–60 mg/day) and five received topiramate (50–75 mg/day).
Thirteen patients (11%) dropped out of the study (three in Group A, five in Group B and five in Group C, P > 0.025, Table 2). The reasons given for dropping out were: lack of time (four in Group A, one in Group B and two in Group C, all at first follow-up), lack of motivation (two in Group B and one in Group C, all at first follow-up), health problems (two in Group B and one in Group C, all at second follow-up), decision to seek other medical help (two in Group A and one in Group C, all at first follow-up), worsening of headache (two in Group A and one in Group B, all at first follow-up) and fear of side-effects (one in group B, first follow-up). No differences in headache characteristics or demographics were found between the patients from the three groups completing and not completing the follow-up visits (all P > 0.025).
Primary and secondary outcome measures
Group A, Simple advice; Group B, standard out-patient detoxification programme (advice + preventive treatment + prednisone); Group C, standard in-patient detoxification programme (advice + preventive treatment + prednisone + fluid replacement and antiemetics). NS, Not significant.
After two months, 105 patients had completed the follow-up visits (37 in Group A, 34 in Group B and 34 in Group C). Of these, 89 (84.7%) were considered responders (31 of the Group A patients, 28 of the Group B patients and 30 of the Group C patients, Table 2). No statistically significant difference was found between the groups (P > 0.025). Of all the patients enrolled in the study we were able to detoxify 78.8% (85% of the Group A patients, 74.3% of the Group B patients and 76.9% of the Group C patients, with no statistically significant difference emerging between the groups, P > 0.025, Table 2).
The percentage reduction in the number of headache days/month in the 105 patients completing the study was 66.8 ± 26 (67.6 ± 25 in Group A, 61.2 ± 34 in Group B and 73 ± 19 in Group C), with no statistically significant difference emerging between the groups (P > 0.025, Table 2).
The total percentage reduction in the number of doses of symptomatic medication/month was 77 ± 24 (76.6 ± 22 in Group A, 71.7 ± 32 in Group B and 81 ± 13 in Group C, again with no statistically significant difference emerging between the groups, P > 0.025, Table 2).
Discussion
The main finding of this study was that, in patients with migraine plus MOH, low medical needs and no previous detoxification therapy, representing 55% of the MOH patients attending our clinic in the study period, simple, strong advice was as effective as structured in-patient and out-patient detoxification programmes in achieving the withdrawal of the overused medication. In addition, the level of adherence to treatment was comparable in all three strategies (almost 90% of the initial sample completed the study).
These results support the theoretical assumption that adequate patient education and support may play a major role in withdrawal therapy of MOH. Indeed, the use of transitional pharmacological therapies, preventive drugs and hospitalization did not improve the outcome of withdrawal therapy (measured as responder rate and percentage reduction of number of headache days per month and number of doses of symptomatic medication per month). Irrespective of the individual reasons underlying the overuse of symptomatic medications (i.e. anxiety about missing work or socio-relational commitments, fear of an imminent attack, fear of withdrawal headache, the reinforcing properties of pain relief by drug consumption, and others) (10), most headache patients are not aware of the negative role of medication overuse in making their headache chronic and in reducing the effectiveness of preventive and behavioural treatments, or that detoxification should take priority other therapeutic options.
This was certainly true of the MOH patients in this study, from which we excluded subjects with previous experience of detoxification therapy. Our findings support the assumption that uncomplicated MOH patients, at least, will respond responsibly when the situation is explained to them. Further controlled studies are necessary to confirm these findings and to evaluate whether advice alone can be as cost-effective as other detoxification strategies in more complicated MOH patients. These data are needed in order to establish whether advice should be regarded as the first step in a step-care approach to MOH management, in which patients progress through a sequence of treatments (determined by a combination of perceived effectiveness, safety and cost) until they find an effective one, or as just one of the therapeutic options in a stratified-care approach, in which patients, on the basis of the extent of their medical needs, are assigned to different treatments.
The literature on out-patient strategies for detoxifying MOH patients is scant (30). Our study is the first to compare the effectiveness of simple advice with the effectiveness of other detoxification strategies in ICHD-II-diagnosed MOH patients with low medical needs. The use of different diagnostic and inclusion criteria and the adoption of different study designs and outcome measures make it difficult to compare our findings with those of other authors. Nevertheless, there is evidence that the detoxification of MOH patients may be successfully obtained in an out-patient setting with no pharmacological help (9, 30–32).
Linton-Dahlöf et al. (9) conducted a retrospective analysis of the effect of drug withdrawal therapy in 229 patients presenting chronic daily headache with long-term misuse of headache medication, defined as daily or almost daily headache (>15 days/month) accompanied by misuse of symptomatic drugs over the past 12 months. As in our study, these authors recommended withdrawal of the offending drug and provided information, education, support and rescue medication (up to twice per week), but no transitional or preventive therapy (with the exception of amitriptyline, given to nine patients). At follow-up visits, performed after 8–12 weeks, 56% patients experienced a 50% reduction of headache days, with only 48% of patients eligible for evaluation. The better results we obtained in our primary outcome measures may be due to the fact that Linton-Dahlöf et al. used different inclusion criteria (i.e. they included patients with chronic tension-type headache as well as those overusing opioids, and did not filter out those presenting psychiatric or physical comorbidity) and a different follow-up strategy (many patients were referred to local physicians and <50% of the initial sample completed the follow-up visits).
More recently, Krymchantowski and Moreira (30) conducted a prospective randomized open-label trial comparing three different approaches to the out-patient management of 150 patients diagnosed with chronic migraine with medication overuse according to the Silberstein–Lipton criteria (2). Although the authors state that most of the MOH patients enrolled in their study were moderate medication overusers and did not use opioids, barbiturates or tranquillizers, their study was not specifically designed to include patients with low medical needs. After 5 weeks, the treatment strategies did not differ with regard to treatment adherence and effectiveness, but the patients who had not undergone transitional therapies during the first 6 days experienced more withdrawal symptoms and used more rescue drugs. The members of the ‘no intervention’ group analysed in this study cannot be compared to our ‘simple advice’ group because they received multiple preventive drugs; nevertheless, since the beneficial effects of preventive medication are not immediate, their contribution to reducing medication intake remains uncertain. Thus, our findings are in agreement with those obtained by Krymchantowski and Moreira in suggesting that migraine patients with moderate MOH and not using opioids, barbiturates or tranquillizers may be detoxified on an out-patient basis whether or not bridge therapy is used during the initial days of withdrawal. The better results we obtained in all the outcome measures may be explained by our different inclusion criteria, drug consumption profile and study duration.
This study set out to compare the short-term effectiveness of three different strategies of MOH detoxification. The long-term effectiveness of these interventions is currently under investigation. Data from the literature suggest that the strategy chosen (out-patient vs. in-patient) is not a significant predictor of the long-term success of the withdrawal therapy (20, 33, 34), while the long-term effectiveness of the ‘simple advice’ strategy has never been studied.
In theory, patients withdrawing overused medication without pharmacological help should develop a greater sense of being in control of their headache (given that this control is not founded exclusively on drug use) and thus be able to reverse their previously learned illness behaviour. Further studies are needed to evaluate the long-term effectiveness of simple advice in comparison with other withdrawal strategies.
Several possible limitations of our study preclude the drawing of definite conclusions regarding the use of simple advice as a withdrawal therapy for MOH. The authors decided to undertake this study after years of experience of detoxifying patients simply by urging them to withdraw the overused symptomatic medications. In addition, the headache clinic setting may have favoured the productive doctor–patient alliance that is essential for successful educational interventions. Thus, the reproducibility of our findings in different settings, such as primary care, remains to be verified.
In summary, this study suggests that in patients with migraine plus MOH, presenting no overuse of opioids, barbiturates or benzodiazepines, no significant physical or psychiatric comorbidity and no previous experience of detoxification therapy, effective drug withdrawal may be obtained through the simple imparting of advice. Future controlled studies are needed to confirm these observations and to establish the role of simple advice in the management of more complicated forms of MOH.