There is evidence, although not overwhelming and unanimously shown in prospective trials, that withdrawal therapy is the best treatment for MOH. However, all experts and headache centers agree that withdrawal therapy should be offered to patients with MOH. The goal of this treatment is not only to detoxify the patients and to stop the chronic headache but also, probably, to improve responsiveness to acute or prophylactic drugs .
The recommended procedures for withdrawal of patients with MOH vary, and no study has compared abrupt withdrawal treatment with tapered withdrawal in prospective randomized trials. Most headache specialists favor the abrupt discontinuation of pain medication under the impression that abrupt withdrawal is associated with faster resolution of the drug-induced pain-coping behavior . However, tapered withdrawal seems to be recommendable for opioids, for barbiturates, and for benzodiazepines. Main withdrawal symptoms are worsening of the headache, nausea, vomiting, arterial hypotension, tachycardia, sleep disturbances, restlessness, anxiety, and nervousness. These symptoms normally last between 2 and 10 days but can persist for up to 4 weeks. The withdrawal headache was shorter in patients having taken triptans (mean 4.1 days) than ergotamine derivatives (mean 6.7 days) or NSAIDs (mean 9.5 days) .
The outcome of withdrawal therapy in patients with MOH followed up by a neurologist as compared to a primary care physician did not differ significantly for calculated mean headache and improvement of headache days . Therefore, it is suggested that a primary care physician can follow these patients after detoxification, which was made in this study in hospital, as well as a neurologist or a pain specialist.
With regard to non-pharmacological approaches of treating MOH, combined short-term psychodynamic psychotherapy and pharmacological therapy improved headache in MOH, and the combination of both had been superior to pharmacological therapy alone for reducing long-term relapses and reduction in quality of life in a non-randomized study . In another study, 120 uncomplicated patients with MOH were treated with three different modalities: (i) only strong advice to withdraw overused medication; (ii) standard outpatient detoxification programme (rapid withdrawal of overused medication plus oral prednisolone for 8 days plus personalized prophylactive drugs); (iii) inpatient programme (rapid withdrawal of overused medication plus oral prednisolone for 8 days plus personalized prophylactive drugs plus parenteral fluid and antiemetics plus close observation for 8 days). The percentages of patients achieving successful withdrawal and the headache frequency were not different between the groups during the follow-up period of 60 days after withdrawal .
A direct comparison between inpatient withdrawal and outpatient withdrawal treatment showed that both methods revealed a significant decrease in headache days per month after 12 months and a reduction in the scores of migraine disability without superiority of one method . Following this study, the outpatient withdrawal is less expensive and as successful in a motivated patient group than inpatient withdrawal. Advantages of the inpatient withdrawal are the close monitoring of medication intake and the clinical state, professional psychological support, an immediate treatment of withdrawal symptoms, and eventually the administration of intravenous drugs. Overusing opioids, barbiturates, or benzodiazepines, psychological problems, severe medical comorbidities, severe withdrawal symptoms (e.g., vomiting, status migrainosus), or previous medication withdrawal failure are reasons for inpatient treatment according to expert consensus or national guidelines [30–32]. However, this recommendation is not supported by randomized prospective trials.
A recent prospective, multicenter study investigated three relatively small groups: (i) only personalized preventive medication from day 1 (n = 17); (ii) abrupt withdrawal plus rescue medication (n = 20); (iii) no preventive medication plus no advice to stop overused drugs (n = 19) . The primary end-point, change in headache days, did, however, not differ significantly between all three groups. Because of the more pronounced reduction in the headache index of the first group in comparison with the second group, there might be an advantage for a personalized preventive medication without abrupt withdrawal. In another study, advice alone was successful as withdrawal therapy in nearly all patients with simple MOH but significantly less successful in patients with complicated MOH . Further larger, prospective trials are necessary to answer these questions.
Studies on a specific preventive therapy of MOH are missing. Therefore, the choice of the preventive agent in MOH should be based on the primary headache (e.g., migraine vs. tension-type headache), the possible side effects of the drugs, the comorbidities, and the patient’s preference and previous therapeutic experience. Several open-label trials showed positive effects of different substances such as valproic acid and topiramate in the prophylactic treatment of chronic daily headache with excessive medication intake. A double-blind trial in patients with the specific diagnosis of chronic migraine and medication overuse showed a significant reduction in the mean number of migraine days per month by topiramate (range 50–200 mg/day) in comparison with placebo (−3.5 ± 6.3 vs. −0.2 ± 4.7; P < 0.05). However, side effects were reported by 75% of the patients in the topiramate group compared with 37% in the placebo group . The headache reduction was nevertheless not big enough to change the chronic headache into an episodic form. In a similar study on chronic migraine, topiramate achieved a significant reduction in migrainous days per month by 6.4 as compared to placebo which achieved a reduction by 4.7 days/month .
In a large-scale study of 335 patients with MOH from the Danish Headache centre, where abrupt detoxification was initiated, the headache frequency was reduced by 67% in migraine patients and by 37% in those with combined migraine and tension-type headache after a 2-month observation period without prophylactic medication . In a recent project with two large studies on the efficacy of onabotulinum toxin A in the treatment of chronic migraine, also patients with medication overuse were treated . Although no specific data on the efficacy of onabotulinum toxin in this specific group of patients (between 63% and 69% of all patients in the different treatment arms) is given, the studies give evidence that onabotulinum toxin A is efficacious in the reduction of headache days in MOH. In summary, it is suggested that detoxification prior to initiating prophylactic therapy may not be required in all patients with MOH , whereas other studies support the importance of initial detoxification [20,23,37].
Treatment of withdrawal headache
Because most drugs helpful for the treatment of withdrawal headache can cause MOH themselves, corticosteroids were regarded as an option for the treatment of withdrawal headache [40,41]. The only controlled, randomized, double-blind study that investigated oral prednisolone during the first 6 days after medication withdrawal revealed no effect on a combined primary end-point. Of total 97 patients, 49 received prednisolone (60 mg on days 1 and 2, 40 mg on days 3 and 4, and 20 mg on days 5 and 6) and 48 placebo . Conversely, a large open-label trial on patients with chronic daily headache and medication overuse showed that treatment with 60 mg prednisone for 2 days and tapering down by 20 mg every other day effectively reduced rebound headache and withdrawal symptoms . Recently, in a small proof-of-concept study, nine patients each with MOH received either placebo or 100 mg prednisone for 5 days . The duration of withdrawal headache was significantly lower in the prednisone group as compared to the placebo group. Taken these results together, there might be an efficacy of corticosteroids on withdrawal symptoms in patients with MOH but high-quality placebo-controlled trials are needed.
There are no other controlled trials on the specific treatment of withdrawal headache or of other symptoms during withdrawal therapy. One open study suggested the combination of intravenous hydration, dexamethasone, metoclopramide, and benzodiazepines for 7–15 days . Very early studies suggested that also (subcutaneous) sumatriptan, naproxen (500 mg), and amitriptyline (10–50 mg) were effective in ameliorating withdrawal headache [40,45,46]. However, all these studies were not placebo controlled. Therefore, by expert consensus, headache drugs and analgesics are not recommended for the treatment of headache during withdrawal therapy except single intravenous administrations in very severe cases.
Prognosis of withdrawal therapy
The relapse rate of MOH is about 30% (range between 14% and 41%) after 1 year regardless whether inpatient, outpatient, or advice alone treatment were applied . Further, the relapse rates do not differ significantly when a short or a long observation period is used, and most studies indicate that the eventual relapse occurs at an early stage (i.e., within few months) after detoxification. In one study, for example, the relapse rate was 23% both after 2 months and after 1 year in the same sample;  in another example, the relapse rate was 41% after 1 year and 44% after 4 years . Overall, the detoxification is fairly successful in most patients, and all patients with MOH should be informed and encouraged to discontinue their overuse. In the general population, simple advice regarding MOH was sufficient to result in a successful treatment of MOH in 76% of all patients after 1.5 years .
In an Italian study on different ways of withdrawal therapy, a long duration of migraine before medication overuse, a higher frequency of migraine after withdrawal therapy, and a greater number of previous preventive treatments were associated with a higher risk for relapse of MOH . In other studies, predictors of relapse were male sex, intake of combination analgesics after withdrawal therapy, nicotine and alcohol consumption, and taking the former medication again after withdrawal therapy [51,52]. Recently, use of codeine-containing drugs, low self-reported sleep quality, and high self-reported bodily pain as measured by the quality of life tool SF-36 were predictors for a poor outcome . In some studies, the prognosis was better for patients with migraine as the underlying primary headache disorder than for patients with tension-type headache and for ergotamine or triptan withdrawal than for analgesic withdrawal [37,48]. It is likely that the different results in these studies with respect to predictors of relapse are because of different study designs and different background populations.