Medication overuse headache (MOH) is a secondary headache that is classified by the third edition of the International Classification of Headache Disorders (ICHD-III) as a group of headaches attributable to the administration or discontinuation of various substances. MOH occurs ≥15 days per month in patients with preexisting headache. It occurs as a result of regular (at least 3 consecutive months) overuse (10 or 15 days, depending on the type of medication) of drugs used as an acute or symptomatic headache therapy (
1).
This type of headache most commonly occurs between ages 30 and 50, more commonly in women (the ratio of women to men is 3–4:1). The estimated global prevalence of MOH is about 3%. The burden on the patients and their families is a consequence of the poor quality of daily life. MOH represents a significant burden on the health system and society as a whole and affects work ability, with frequent absenteeism. The negative effect of MOH on the patient is far greater than the negative effect of migraine or tension-type headaches (
2–
4).
Although MOH can occur from any type of headache, it most commonly arises from the chronicity and transformation of migraine. Currently, there is no comprehensive account of the pathophysiological process underlying MOH. However, central sensitization processes, such as suppression of endogenous pain modulation and neuronal excitability due to imbalances of neurotransmitters (adrenaline, serotonin), are likely contributors to MOH (
5).
MOH therapy involves educating the patient on the need for limited use of analgesics, with the implementation of detoxification therapy and prophylactic support for coanalgesic therapy. This may be performed in an outpatient or hospital setting, abruptly or gradually discontinuing overdose. Neither immediate nor delayed therapeutic treatments are fully successful (
6).
Previous research has mainly examined the risk factors for the onset and effective treatment of MOH (
7–
9). Insufficient research has been dedicated to examining the parameters that determine the effect of MOH on activities of daily life and on different domains of patient health (
10–
12). An examination of the association between the various parameters and the disability caused by MOH would be useful for identifying patients at risk for developing severe disability. The recognition of potentially variable risk factors would make it possible to develop additional therapy and help patients. Therefore, the aim of this study was to analyze the association between demographic, clinical, and paraclinical parameters and the effect of MOH on daily life, and different domains of health, depression, anxiety, and stress in MOH patients.
Methods
The study was approved by the Internal Review Board of the Faculty of Medicine, University of Niš (Serbia). Patients provided voluntary informed consent to participate in the study in accordance with the Declaration of Helsinki.
This was an observational, 1-year cross-sectional study. The study included all patients examined in the Headache Center of the Neurology Clinic at the Clinical Center in Niš during 2019 (January through December), in whom MOH was first diagnosed during this period. The Clinical Center in Niš is a tertiary health care institution serving about 2 million inhabitants of southeastern Serbia. Headache outpatients include patients referred by a primary care physician or by a specialist in neurology, internal medicine, or related specializations. Patients with previously diagnosed MOH were not included in this study, in order to avoid the influence of applied pharmacoprophylaxis and other therapies on the parameters examined in this study.
The diagnosis of MOH was made according to the diagnostic criteria of the ICHD-III (
1). Secondary etiology of the headaches was excluded after complete diagnostic processing. Computerized tomography or MRI was performed in all study patients. For all patients, the diagnosis of headache was made by a physician specialist in neurology and pain medicine working in this headache center (S.L.).
Data collected from patients included demographic characteristics, educational level, marital, family, and work status, number of family members, residence, personal history, past medical history, headache history (primary and/or secondary headaches), headache type, characteristics, duration, frequency, the type(s) and effectiveness of symptomatic and preventive therapy, and habits and risk factors (physical activity, cigarette smoking, use of alcohol, excessive use of caffeine [i.e., >3 cups of coffee per day]).
Data collected regarding MOH and previous headache were for duration, pain location (frontal, temporal, parietal, occipital), lateralization (unilateral, diffuse), character (dull, pulsating pain), pain intensity (using a numerical pain assessment scale), associated symptoms and signs (nausea/vomiting, photophobia, phonophobia, diplopia, neck and shoulder stiffness, blurred vision, tinnitus, hypoacusis), type of analgesic therapy used, preventive therapies, frequency of use of this therapy (number of days per month), and therapy effectiveness (i.e., on pain intensity reductions and/or on associated symptoms).
The effect of headache on activities of daily living was assessed with the Headache Impact Test–6 (HIT-6; 6TM, version 1.1). HIT-6 is a tool used to measure the effect of headaches on the ability to function in a job, at school, at home, and in social situations. For each of the six questions, patients choose one of the offered answers (never, often, sometimes, very often, and always). HIT-6 ratings are obtained by scoring each answer with 6, 8, 10, 11, or 13 points. The score range is 36–78. Higher scores indicate a greater effect of headache on patient life; a score of 60 or more indicates that headaches severely affect family, work, school, or social activities (
13).
The Depression Anxiety Stress Scale (DASS) was used to assess depression, anxiety, and stress. The patient answers 42 questions by selecting a statement (0=did not apply to me at all, 1=applied to me at some degree or for some of the time, 2=applied to me at a considerable degree or for a good amount of time, 3=most accurately describes my feelings in the past week). The maximum score on this scale is 42 for each item tested (depression, anxiety, and stress). Scores indicating a normal range are 0–9 (for depression), 0–7 (for anxiety), and 0–14 (for stress). Higher scores indicate severe depression (>28), anxiety (>20), and stress (>34); values between the high and normal ranges indicate mild or moderate depression, anxiety, and stress (
14). Patient testing was performed by a psychologist (M.L.).
The Quality of Life Assessment was conducted with the SF-36 Questionnaire (version 1.0) (
15). This questionnaire tests eight domains of health (quality of life): physical functioning, role functioning/physical functioning, role functioning/emotional functioning, energy/fatigue, emotional well-being, pain, general health, and health change. The SF-36 also examines the perception of change in health. The answers received are translated into numerical values by the evaluation key. High values are the result of better health; the lowest and highest possible values are 0 and 100. This survey was administered by a psychologist (M.L.). All tests were performed close to the time of MOH diagnosis.
Data are presented as arithmetic mean and standard deviation, respectively, as medians and 25th and 75th percentiles. Normally distributed data were examined using t test and analysis of variance. Nonnormally distributed data were examined using the Mann-Whitney U test and the Kruskal-Wallis H test. A chi-square or Fisher’s test was used in the analysis of categorical data. An exploratory logistic regression analysis (enter method) was conducted to further evaluate the significant associations between levels of HIT-6, depression, anxiety, and stress scores and demographic and clinical characteristics. From these analyses, variables with a p value <0.10 were retained for the subsequent multivariable model (backward Wald method). The Hosmer-Lemeshow test was used to test the calibration ability of the model. A complete case analysis was performed. The hypothesis was tested with a significance threshold of p<0.05. All statistical analyses were performed using R software (version 3.4.3, R Foundation for Statistical Computing, Vienna, Austria).
Results
During the study period, a total of 982 patients (including both newly referred and established patients) were examined in the headache center, 89 of whom were diagnosed with MOH for the first time. Of these patients, six were excluded because of unwillingness to participate in this study. The excluded patients did not differ in age and sex distribution when compared with the study population (Fisher’s test, p=0.12; chi-square test, p=0.15).
Eighty-three patients met study criteria, 72 of whom were women. The mean age of the study population was 40.54 years (SD=11.58 years; youngest age, 18 years; oldest age, 71 years). Most patients were from a city (N=65), with a high school level education (N=44), married (N=57), with one or more children (N=56), employed (N=55), and without comorbidity (N=47). In the study population, 34 patients were smokers, four patients consumed alcohol regularly, 74 patients consumed caffeinated beverages regularly (≥3 cups of coffee per day), and eight patients practiced regular physical activity.
The average length of MOH until diagnosis was 5 years (mean=5.58 years, SD=3.52). On average, 19 days (19.40 days, SD=6.07) per month were filled with headache. A single headache attack had an average duration of 11 hours (mean=10.71 hours, SD=1.90). The headache had a sharp character in 42 patients. Temporal localization of headache was dominant in the study population (N=54). Bilateral presentation of headache was reported in 47 patients. Neck and shoulder stiffness were identified in 71 patients. Forty-four patients reported analgesic therapy up to 15 days per month. Combined analgesics (N=44) with partial therapeutic efficacy (N=65) were most commonly used. The mean duration of previous chronic headache was 7.71 years (SD=6.79). In 76 patients, the average monthly burden of this headache type was <15 days. Half of the patients used symptomatic therapy, such as common analgesics, for their headache. Tricyclic antidepressants were the most commonly used prophylactic drugs (N=69) (
Table 1). Most of the MOH patients had migraine (N=53) before the development of MOH; a smaller number had tension-type headache (N=21) and other symptomatic headache (N=9).
The average HIT-6 score in the study population was 65.39 (SD=5.45; minimum score, 49; maximum score, 77). The mean values obtained on the DASS were 12 for depression (6–23; mild depression), 15 for anxiety (9–22; moderate anxiety), and 22 for stress (15–31; moderate stress). The SF-36 scores were lower than would be expected in the general population. SF-36 scores were significantly lower among patients with a HIT-6 score ≥60 for physical functioning (p=0.018), role functioning/physical functioning (p=0.001), energy/fatigue (p=0.049), and pain (p=0.043) (
Table 2). Among patients with a HIT-6 score ≥60, migraine (with and without aura) was significantly more frequent compared with patients with a HIT-6 score <60 (p=0.025). Migraine patients had statistically significantly higher HIT-6 scores than patients with tension-type headaches (p=0.018) and patients with secondary headaches (p=0.003). The efficacy of symptomatic therapy in prior headache was statistically significantly lower in patients with HIT-6 scores ≥60 (66.2% versus 100.0%, p=0.011).
Patients with comorbidities had significantly higher values on depression (p=0.007), anxiety (p=0.001), and stress (p=0.015). Patients who consumed alcohol had significantly lower depression (p=0.028) and anxiety scores (p=0.018). Anxiety was significantly higher among women (p=0.022) (
Table 3).
HIT-6 scores were significantly different with respect to headache character (p=0.001), lateralization (p=0.046), and localization (p=0.047); the type of previous chronic headache (p<0.001); and the efficacy of symptomatic therapy on previous chronic headache (complete versus partial versus absent, p=0.027). The scores for depression, anxiety (except for the efficacy of symptomatic therapy on a previous chronic headache), and stress did not differ significantly from the clinical parameters examined (
Table 4).
Univariate logistic regression analysis showed that statistically significant risk factors for HIT-6 scores >60 were alcohol use (odds ratio=0.145, p=0.047) and decrease in SF-36 scores for domains such as physical functioning (odds ratio=0.970, p=0.027), role functioning/physical functioning (odds ratio=0.972, p=0.003), and social functioning (odds ratio=0.954, p=0.007). Risk factors for depression were decreases in the value of the following domains of health: physical functioning (odds ratio=0.977, p=0.013), role functioning/emotional functioning (odds ratio=0.978, p=0.012), energy/fatigue (odds ratio=0.942, p=0.003), emotional well-being (odds ratio=0.905, p<0.001), social functioning (odds ratio=0.935, p<0.001), pain (odds ratio=0.930, p<0.001), general health (odds ratio=0.959, p=0.007), and the presence of comorbidity (odds ratio=3.900, p=0.008). Risk factors for anxiety were age (odds ratio=1.047, p=0.030), number of children (odds ratio=1.808, p=0.020), and decreases in the value of all health domains tested, which were as follows: physical functioning (odds ratio=0.965, p<0.001), role functioning/physical functioning (odds ratio=0.984, p=0.035), role functioning/emotional functioning (odds ratio=0.976, p=0.001), energy/fatigue (odds ratio=0.950, p=0.003), emotional well-being (odds ratio=0.925, p<0.001), social functioning (odds ratio=0.942, p<0.001), pain (odds ratio=0.935, p<0.001), general health (odds ratio=0.953, p=0.001), and the presence of comorbidity (odds ratio=3.832, p=0.005). Risk factors for stress were the duration of the previous headache (odds ratio=1.107, p=0.031) and decreases in the value of the following health domains: role functioning/emotional functioning (odds ratio=0.983, p=0.017), energy/fatigue (odds ratio=0.943, p=0.001), emotional well-being (or 0.902, p<0.001), social functioning (or 0.962, p=0.001), pain (or 0.962, p=0.004), and general health (odds ratio=0.974, p=0.041) (
Table 5).
Multivariate analysis included all risk factors from univariate analysis for which p<0.10. The backward elimination (Wald) method showed statistically significant risk factors for HIT-6 scores >60: role functioning/physical functioning (odds ratio=0.977, p=0.024) and social functioning (odds ratio=0.963, p=0.032). Multivariate analysis revealed that the risk factors for depression were emotional well-being (odds ratio=0.928, p=0.007), social functioning (odds ratio=0.950, p=0.009), and the presence of comorbidity (odds ratio=5.417, p=0.013). Multivariate analysis revealed that risk factors for anxiety were age (odds ratio=1.091, p=0.007), MOH duration (odds ratio=1.422, p=0.047), emotional well-being (odds ratio=0.933, p=0.012), and social functioning (odds ratio=0.943, p=0.001). Emotional well-being (odds ratio=0.902, p<0.001) was a statistically significant risk factor for stress (
Table 5).
Correlation analysis showed that the HIT-6 score correlated significantly with the duration of previous chronic headache before the onset of MOH (r=0.327, p=0.003). The HIT-6 score correlated significantly with depression (r=0.306, p=0.025), physical functioning (r=−0.347, p=0.001), role functioning/physical functioning (r=−0.449, p<0.001), social functioning (r=−0.336, p=0.002), and pain (r=−0.270, p=0.013). Depression, anxiety, and stress correlated statistically significantly with all domains tested (see Table S1 in the online supplement).
Discussion
The present findings demonstrate that MOH has a significant negative impact on the personal, family, and social life of the patient and that it is associated with depression, anxiety, and stress. The most significant changes in examined health-related domains were emotional well-being and social functioning. Emotional well-being was associated with depression, anxiety, and stress, whereas social functioning was associated with depression and anxiety. In this study, the presence of comorbidities was associated with depression, whereas patient age and longer duration of MOH were associated with anxiety.
The results of this study are consistent with the results of previous population-based studies finding a higher prevalence of MOH in women and the occurrence of MOH at the age of 30–50 years (
2,
16). The results presented here are consistent with earlier findings from a population-based door-to-door survey that MOH is more prevalent in urban areas than in rural areas (
17). Our results are also in line with research findings that MOH is far more common in people who are employed (
2). Previous studies provided no clear evidence of an association between lifestyle (smoking, obesity, physical activity) and MOH (
2,
18). Because most of these studies, similar to our study, were cross-sectional studies, it is not possible to examine the causality of this relationship, but it is important to examine these factors in the context of other risk factors for MOH.
The results presented here indicate mild depression, moderate anxiety, and moderate stress in the study patients and indicate that the highest proportion of patients had chronic migraine prior to the development of MOH. Previously, it has been shown that a significant percentage of MOH is caused by migraine transformation. In a longitudinal population-based cohort study, comorbidities and the existence of psychological distress modified the course of migraine and affected its chronicity and transformation into MOH (
18). A cross-sectional survey sampling from the adult population found a higher prevalence of depression in patients with chronic migraine who developed MOH than in patients who did not develop MOH (
19). Results from the series of consecutive patients with chronic headache indicated that chronic migraine is more commonly associated with psychiatric comorbidities than other chronic headaches (
20). Some retrospective analyses have identified depression as an independent predictor of MOH (
21). It has been shown that comorbidities such as psychiatric disorders may strongly influence the effect of MOH and MOH management (
5).
On the other hand, anxiety is often associated with an episodic primary headache that evolves into MOH (
22–
24). Some results from the cross-sectional studies suggest an association between panic disorder and MOH (
20,
22); another prospective study suggests an association between borderline personality disorder and primary headache (
25). Our results find the presence of depression and anxiety in MOH, which is consistent with some similar previous research (
26–
28). In this study, age and duration of MOH were recognized as risk factors for MOH anxiety. Disruption of endogenous pain modulation mechanisms and development of central sensitization have been confirmed in the elderly and in chronic pain conditions. The functional association of these mechanisms with mechanisms of cognitive behavioral functioning (
2) could be an explanation for the development of anxiety in elderly patients with long-term duration of MOH.
The relationship between stress and MOH has been studied among patients in specialized headache centers and longitudinal population-based studies (
29,
30). Headache patients who have less capacity to cope with stress have been shown to use analgesics more often when coping with stress (
31). It has already been shown that the most common trigger for different types of headache is stress (
32–
34). Some of these studies analyzed only clinical-based populations (
32). There is direct evidence of an association between MOH and stress (
2,
29,
35). It has been shown that stress can also potentiate other behavioral factors for headache, such as lack of sleep and diet (
29). It has also been shown that stress can be associated with the chronicity of migraine and its transformation into MOH (
36). The putative mechanism by which stress leads to headache and to chronicity and transformation in MOH is the activation of the hypothalamic-pituitary axis and the sympathetic nervous system and the disruption of activity of various brain regions involved in the pain matrix and cognitively affective phenomena. The influence of cognition (locus of control and self-efficacy) and negative emotions (depression, anxiety, anger) on the onset of headache, pain perception, attachment to prescribed therapy, outcome of therapy, and headache-related disability was supported (
37,
38). It has been documented that strengthening coping mechanisms in patients with chronic headache can significantly affect functioning quality and pain perception (
39). Patients who do not adequately manage stress and negative emotions (anger) and anxiety have been shown to have an increased risk of developing chronic headache and MOH (
40,
41).
In this study, the effect of MOH on the daily life of the patient was assessed as highly negative. Results indicate that patients with MOH have a higher degree of functional disability over time than patients with chronic headache without MOH (
42,
43). The results of this study indicate that emotional and social support could be significant in reducing the negative effect of MOH on all domains of patient health and on the onset of depression, anxiety, and stress. Some authors have pointed to a significant effect of psychological support in patients with chronic headaches and chronic pain conditions in general (
44,
45). It has also been shown that psychotherapy can be effective in the therapeutic control of different types of primary and secondary headaches (
46). Psychological support has been shown to be significant for successful detoxification and effective prevention of MOH recurrence (
6,
47) and for the prevention of headache chronicity and the onset of MOH (
48).
Some limitations of the study must be noted. Selection bias may influence the present findings, as the study cohort was derived from a tertiary headache center. As such, the symptoms and conditions may have been worse than what may be observed in other settings (e.g., primary care). However, consecutive recruitment of participants may have reduced further biases associated with ascertainment of patient subsets in this tertiary headache. This is a single center study and has all the inherent limitations of that kind of study design. Further, the nature of this study does not allow an examination of causality, only association.
By contrast, the relatively large sample size in the present study permitted examination of the relationship between a relatively large number of headache parameters and the effect of MOH on depression, anxiety, and stress as well as the identification of predictors of these relationships. These analyses suggest the existence of subgroup of MOH patients who are at increased risk for greater negative impact of MOH. These patients could have special benefits from additional interventions aimed at improving the physical, emotional and social quality of life, thus improving the treatment of MOH and its comorbidities.
Conclusions
In conclusion, MOH has a significant negative impact on the personal, family, and social life of patients and is associated with depression, anxiety, and stress. Patient’s age, duration of MOH, the presence of comorbidities and the negative effects of physical, emotional, and social functioning are particularly important contributors to the significant impact of these co-occurring conditions. The present findings suggest that there is a need for early diagnosis and treatment of MOH and comorbidities, as well as a need to provide emotional and social support especially in selected MOH patients to prevent and mitigate the effects of MOH, depression, anxiety, and stress. The findings further suggest that such prevention and mitigation efforts are likely to require multidisciplinary, individually oriented treatment.