Treatment of Medication Overuse Headache: Effect and Predictors After 1 Year

A Randomized Controlled Trial

Louise N. Carlsen MD, PhD; Carolien Rouw MD; Maria L. Westergaard MD, PhD; Mia Nielsen MD; Signe B. Munksgaard MD, PhD; Lars Bendtsen MD, PhD, Dr.Med.Sci.; Rigmor H. Jensen MD, Dr.Med.Sci.


Headache. 2021;61(7):1112-1122. 

In This Article


Study Population

Patients with MOH were recruited consecutively from October 2016 to November 2018. The final follow-up visit occurred in November 2019. Of the 120 included patients, 96 completed 1-year follow-up (n = 29 in W+P, n = 33 in P+pW, and n = 34 in W+pP) (Figure 2). The overall dropout rate was 20% (28% in W+P, 18% in P+pW, and 15% in W+pP). Baseline characteristics were similar in terms of age and sex: mean age (SD) of 44.4 (12.0) years and 76 (79.2%) women, respectively (Table 1). The dropout rate between the three groups was not significant (p = 0.335). Baseline characteristics for dropouts are presented in Table S2 without any remarkable differences between the groups. Tables S5 and Table S6 show the missing data at baseline and after 12 months, respectively.

Figure 2.

CONSORT flow diagram. Patient flow from inclusion at baseline to 6-months follow-up including all contacts (visits at baseline, 2- and 6-months follow-up and phone calls at 1 and 4 months). P+pW, preventive treatment with potential withdrawal after 6 months; W+P, withdrawal and preventives from start; W+pP, withdrawal with optional preventive treatment after 2 months. 1Reasons for in-patient care: Significant opioid overuse (n = 41), severe physical or psychiatric co-morbidity with need for in-patient care (n = 11), difficult social situation (n = 10), complicated by pregnancy (n = 2), unclear or severe pre-existing headache diagnosis (n = 13), previously failed out-patient withdrawal (n = 2), unknown (n = 20). 2Not fulfilling inclusion criteria: Based on pre-existing headache diagnoses, e.g., post-traumatic headache or cluster headache. 3Exclusion criteria: Physical or psychiatric co-morbidity (n = 36), pharmacological preventive treatment for headache at inclusion time (n = 17), pregnancy or breast feeding (n = 5), language barriers (n = 57). Reasons for drop-out W+P: 4Lost to follow-up (n = 2), did not want to start preventive treatment (n = 2), continued the medication overuse (n = 3); 5did not want to start preventive treatment (n = 1), cluster headache diagnosis after withdrawal and no migraine and/or tension-type headache (n = 1); 6Lost to follow-up (n = 2). P+pW: 7Lost to follow-up (n = 3); 8Lost to follow-up (n = 1), did not want to start preventive treatment (n = 1); 9Lost to follow-up (n = 1); 10Lost to follow-up (n = 1). W+pP: 11Lost to follow-up (n = 2), diagnosis of malignancy after inclusion (n = 1); 12Pregnancy (n = 1); 13Lost to follow-up (n = 1), underwent a second withdrawal (n = 1)

No harms or adverse events occurred during the trial period.

Course of Treatment

Description of withdrawal, treatment with preventive medication, and multidisciplinary treatment during the 1-year follow-up period are presented in Table 2.

In the P+pW group, nine (28%) patients who received preventive medication still had MOH after 6 months of treatment. They were, therefore, offered withdrawal therapy. Out of these nine patients, five accepted withdrawal therapy.

At 1-year follow-up, the percentage of patients using preventives was 66% in the W+P group, 85% in the P+pW group, and 62% in the W+pP group. Candesartan was by far the most used preventive medication, used by 43% of the total study population.

Approximately 71% of patients (79% in the W+P group, 64% in the P+pW group, and 71% in the W+pP group) were referred to physiotherapy, and 34% of patients were referred to therapy sessions with a psychologist.

Treatment Effect for the Total Study Population

In the total study population, monthly headache days were reduced from 24.6 days (95% CI: 23.4–25.8) to 15.0 days (95% CI: 13.0–17.0) at 1-year follow-up (p < 0.0001) (Table 3). Figure 3 illustrates reduction in headache days over the 1-year study period.

Figure 3.

Change in headache days. Data are shown as mean values with 95% CI (error bars) [Color figure can be viewed at]

Monthly migraine days decreased from 9.9 days (95% CI: 8–11.7) to 5.6 days (95% CI: 4.3–6.9) (p = 0.001). Days with acute medication were reduced from 22.2 days (95% CI: 20.9–23.5) to 10.8 days (95% CI: 9.1–12.5) (p < 0.0001). Furthermore, headache pain intensity, SDS score, and HURT score were reduced by a statistically significant degree, whereas overall QoL and self-rated health showed a statistically significant improvement (Table 3).

No significant differences from baseline to 1-year follow-up were found in HADS depression score, HADS anxiety score, or PSS score (Table 3).

Comparison of the Three Treatment Strategies

After 1 year, reduction in monthly headache days was 10.3 days (95% CI: 6.7–13.9) in the W+P group, 10.8 days (95% CI: 7.6–14) in the P+pW group, and 7.9 days (95% CI: 5.1–10.7) in the W+pP group, without any difference between the groups (p = 0.377) (Figure 3, Table S3).

Noteworthily, the number with resolved MOH patients was high in all three groups with 26 (90%) in the W+P group, 31 (94%) in the P+pW group, and 28 (82%) in the W+pP group (p = 0.322). This corresponds to relapse rates of 10%, 6%, and 18%, respectively. In summary, no differences were observed between the treatment groups (Table S3).

Prognostic Factors for Still Having Chronic Headache After 1 Year

After 1 year, 39 out of 96 patients (41%) had chronic headache. Baseline characteristics for patients with episodic and chronic headache after 1-year follow-up are presented in Table S4. In both the univariate and multiple regression analyses, higher number of headache days (aOR 1.19 [95% CI: 1.09–1.31]; p < 0.001), higher number of days with acute medication (aOR 1.11 [95% CI: 1.03–1.19]; p = 0.004), higher pain intensity (aOR 1.04 [95% CI: 1.01–1.08]; p = 0.007), and having depression (aOR 4.7 [95% CI: 1.38–18.95]; p = 0.018) predicted a chronic headache form after 1 year (Table 4). In contrast, higher self-rated health (aOR 0.61 [95% CI: 0.36–0.97]; p = 0.043) and daily consumption of at least 200 mg caffeine (aOR 0.40 [95% CI: 0.16–0.96]; p = 0.043) predicted a lower risk of still having chronic headache at 1-year follow-up. Age, sex, type of primary headache, migraine days/month, duration of medication overuse, higher SDS score, lower overall QoL, lower HURT score, higher risk of having anxiety, and higher perceived stress score were not associated with remaining chronic headache.