Negative impact of migraine on quality of life after 4 weeks of treatment in patients with major depressive disorder
Shuu-Jiun Wang, MD, Department of Neurology, Neurological Institute, Taipei Veterans General Hospital, No. 201 Shi-Pai Road, Section 2, Taipei 112, Taiwan. Email: email@example.com
Aim: The impact of migraine on health-related quality of life (HRQoL) among patients with major depressive disorder (MDD) after acute antidepressant treatment has not been addressed. The aim of the present study was to investigate whether or not the negative impact of migraine on HRQoL among outpatients with MDD continued to have an effect after 4 weeks of venlafaxine treatment.
Methods: A total of 135 outpatients with MDD were enrolled, who were then treated with venlafaxine 75 mg per day for 4 weeks in the present open-label study. Migraine was diagnosed based on the International Classification of Headache Disorders (2nd edn). Changes in Short-Form 36 (SF-36) and Hamilton Depression Rating Scale (HAMD) scores were the outcome measures. Multiple linear regression was used to assess whether migraine was an independent factor predicting SF-36 score after treatment.
Results: Seventy-two participants (18M/54F) completed the 4-week treatment. Subjects with migraine had a poorer HRQoL in terms of bodily pain and mental health at baseline. Subjects with and without migraine showed significant improvement in all SF-36 subscales and depression after treatment, but subjects with migraine still had a poorer HRQoL regarding bodily pain and physical functioning after treatment as compared with those without migraine. Migraine could predict a negative outcome after treatment in the subscales of physical functioning, role limitations–physical, and role limitations–emotional.
Conclusions: Migraine may have a negative impact on the improvement of partial SF-36 subscales, especially on functional recovery, after acute treatment among outpatients with MDD. Whether additional intervention besides antidepressant treatment for migraine is indicated may need further study.
MOOD DISORDERS AND migraine are correlated and interact.1–3 The association between migraine and major depressive disorder (MDD) is bidirectional, each disorder increasing the risk of first onset of the other at follow up.4 Depressive symptoms are also related to precipitating or aggravating factors of migraine among patients with MDD,5 and patients with migraine who also suffer from depression or anxiety have more somatic symptoms.6 Depression further decreases the quality of life of patients with chronic migraine,7,8 and also has a negative impact on the long-term prognosis of headache disorders among patients with migraine.9,10
Migraine has a negative impact on patients with MDD. Compared with MDD patients without migraine, those with migraine have: (i) more depressive episodes and clinical features of bipolar spectrum traits;11 (ii) more severe depression, anxiety, and somatic symptoms;12–15 and (iii) a poorer health-related quality of life (HRQoL), especially in the physical dimensions.16 Moreover, migraine also has a negative impact on bipolar disorder.17 These studies, however, were all cross-sectional. The impact of migraine on HRQoL or on the outcome of depression after antidepressant treatment has not previously been addressed. In our previous study, it was reported that migraine is one of the most important comorbidities predicting worse bodily pain in HRQoL assessment, even after controlling anxiety comorbidities;18 it is not known whether MDD patients with migraine still have poorer bodily pain after antidepressant treatment as compared with those without migraine. There are many studies related to HRQoL among patients with MDD.19,20 Previous studies have found that migraine has a negative impact on quality of life,21,22 but the impact of migraine on HRQoL after acute antidepressant treatment among patients with MDD has been neglected to date. Therefore, the purpose of the present study was to investigate whether or not the negative impact of migraine on the HRQoL of patients with MDD continued after a 4-week course of venlafaxine treatment. We hypothesized that the negative impact of migraine on the HRQoL of MDD patients may continue to have an effect after a 4-week course of treatment. Venlafaxine was used in the present study because previous studies have reported this drug to be effective in preventing migraine attacks,23 and the use of venlafaxine only and no other antidepressants avoided potential differential treatment efficacy among different antidepressants.
This study, which was approved by the Institutional Review Board of the Chang Gung Memorial Hospital, was conducted in the general psychiatric clinic of the same hospital from January 2004 to January 2005. The inclusion criteria were: (i) consecutive outpatients aged 18−65 years who had not taken antidepressants or any other psychotropic drugs within the previous 2 weeks; and (ii) patients who met the DSM-IV-TR criteria for MDD24 and were experiencing a major depressive episode, as diagnosed using the Structured Clinical Interview for DSM-IV-text revision (TR) Axis I Disorders.25 To prevent HRQoL from being confounded by the effects of other medical conditions, the following exclusion criteria were established: (i) history of substance dependence or abuse of opioids or their derivatives, amphetamine or its derivatives, cocaine, cannabis, hallucinogens, inhalants, or alcohol without full remission in the previous month; (ii) psychotic symptoms, catatonic features, or severe psychomotor retardation with obvious difficulty in being interviewed; and (iii) chronic medical diseases (such as hypertension, diabetes mellitus etc.), with the exception of headache. Written informed consent based on the guidelines regulated in the Declaration of Helsinki was obtained from all subjects prior to study enrollment. Educational level was measured in total years of formal schooling. All patients were assessed using the 17-item Hamilton Depression Rating Scale (HAMD)26 by two psychiatrists who were blind to the headache diagnosis and other data. The two psychiatrists had been trained together to use the HAMD before the study, and the correlation coefficient of the total HAMD scores as rated by the two psychiatrists was 0.91 (P < 0.01).
Based on the aforementioned inclusion and exclusion criteria, 135 subjects (34 men and 101 women; mean age, 30.2 ± 8.4 years; mean HAMD score, 23.4 ± 4.5; mean onset age of the first major depressive episode, 26.9 ± 8.6 years; mean educational years, 13.0 ± 2.4) agreed to participate in the study during the study period.
Assessment of headache
All patients completed a structured headache intake form, which was designed to meet the operational criteria of the International Classification of Headache Disorders (2nd edn; ICHD-2).27 The form emphasized the collection of the information needed to classify migraine and other types of headache. An experienced headache specialist, who was blind to the results of psychiatric evaluation, interviewed all patients after they had completed the headache intake form, and made the headache diagnoses based on the ICHD-2. Subjects evaluated their average headache intensity in the past week using an 11-point (0–10) scale, with 0 representing ‘no pain’ and 10 representing ‘pain as severe as I can imagine’. Headache frequency (measured in days) in the past week was also recorded. The evaluation period of the two headache indices was the previous week in order to make it compatible with the evaluations of the severity of depression and HRQoL. It should be noted that a headache diary was not used in this study, because withholding pharmacological treatment at baseline in order to prospectively observe headache parameters might cause ethical problems owing to the potential suicide risk. Therefore, the headache intensity and frequency measures in the present study might represent a mixture of different headache attacks.
Assessment of HRQoL
The acute version of the Short-Form 36 (SF-36), which is used to evaluate HRQoL in the previous week, was used in order to be compatible with the evaluation period of the HAMD. The SF-36 includes eight domains: physical functioning (PF); role limitations–physical (RP; the impact on capacity to work or other daily activities as a result of physical health); bodily pain (BP); general health perception (GH); vitality (VT); social functioning (SF); role limitations–emotional (RE; the impact on capacity to work or other daily activities as a result of emotional difficulties); and mental health (MH).28 These eight domains can be grouped into two subscales: the physical subscale (PF, RP, BP, and GH) and the mental subscale (VT, SF, RE, and MH). The Taiwanese version of the SF-36 has been found to have good validity and reliability.29
Acute antidepressant treatment
Each patient was treated with venlafaxine (extended release) 75 mg (one capsule) per day, a serotonin and norepinephrine re-uptake inhibitor effective in treating depression and somatic symptoms.30 Zolpidem (in 10-mg tablet form) was prescribed as needed at night,31 because insomnia is a common complaint in patients with MDD. Four weeks later, the SF-36 was administered again and the HAMD re-evaluated by the same psychiatrist. Drug compliance was monitored by pill count, and subjects with a venlafaxine treatment compliance of ≥80% (calculated as total capsules/total days) were categorized as the treatment group for further analyses. The treatment duration was chosen to be 4 weeks, because this time point has been reported to be one of the important milestones after which treatment response can be predicted or observed, or the antidepressant can be changed.32,33
Other medications for the acute treatment of migraine, such as analgesics or triptans, were not prescribed for two reasons: (i) venlafaxine has a migraine-prevention effect;23 and (ii) these medications are not routinely prescribed in most trials of MDD treatment, although pain symptoms are a common problem in patients with MDD.
All statistical analyses were performed using SPSS for Windows 12.0 (SPSS, Chicago, IL, USA). Subjects in the treatment group were divided into the migraine group (MDD subjects with migraine) and the non-migraine group. In each group, paired t-test was used to assess the differences in the SF-36 and HAMD scores between baseline and after the 4-week treatment. The independent t-test was used to assess differences between groups, and Student's t-test with Bonferroni correction (statistical significance: two-tailed P < 0.006 in the present study) was used to identify differences in the scores of the SF-36 subscales between subjects with or without migraine. Pearson correlation was used to examine the relationships between headache indices, SF-36 subscale scores, HAMD score, age, and education. In order to identify whether migraine was an important factor predicting the scores of the SF-36 subscales after the 1-month pharmacotherapy, multiple linear regression with forward selection was used for controlling the scores of these subscales at baseline and the demographic variables. The dependent variable was the SF-36 subscale score after 1-month pharmacotherapy; the independent variables were migraine, age, gender, educational years, and SF-36 subscale score at baseline. In this study, all P were two-tailed results.
Among the 135 subjects, 72 (18 men and 54 women; mean age, 31.2 ± 8.0 years; mean HAMD score, 23.0 ± 4.4; mean onset age of the first major depressive episode, 27.8 ± 8.7 years; mean educational years, 13.2 ± 2.5) with a treatment compliance of ≥80% were categorized as the treatment group; the other 63 subjects, including subjects who dropped out of the treatment course (40 subjects), those who were shifted to other antidepressants (16 subjects), or those who exhibited poor compliance (seven subjects), were categorized as the dropout group and were excluded in further analyses.
Of the treatment group (72 subjects), 38 subjects (52.8%) had migraine (30 without aura, eight both with and without aura; four with chronic migraine, 34 with episodic migraine). Those who did not fulfill the criteria for migraine included 10 (13.9%) with probable migraine without aura; two (2.8%) with chronic tension-type headache (TTH); eight (11.1%) with frequent episodic TTH; two (2.8%) with infrequent episodic TTH; four (5.6%) with probable episodic TTH; and three (4.2%) with headache not elsewhere classified. Five subjects (6.9%) reported no headaches. Notably, in the treatment group, 15 subjects with migraine and three without migraine had taken non-steroidal anti-inflammatory drugs (NSAIDs) or other analgesics in the week previous to enrollment and none of the 72 subjects had been treated with medications for migraine prevention. During the 4-week treatment, three subjects with migraine took NSAIDs, but no subjects without migraine took analgesics. This factor was not taken into consideration because only a small number of subjects with migraine had taken analgesics during the treatment period. Moreover, during the treatment process, there was no statistical difference in venlafaxine compliance or average zolpidem dosage per day between the migraine group and the non-migraine group (4.0 ± 3.6 mg vs 3.0 ± 3.5 mg).
The mean scores of the SF-36 subscales and the HAMD are listed in Tables 1,2. There were no statistical differences in the SF-36 subscale and HAMD scores, two headache indices, mean age of the first major depressive episode, educational years, or age between the treatment and dropout groups; there were also no significant differences in the percentage of migraine sufferers (52.8% vs 39.7% for the treatment vs dropout group) or female patients (75.0% vs 74.6%) between the two groups. In the following sections, only the results of the treatment group are reported.
Table 1. Short Form 36, HAMD, and headache profile scores vs treatment group
|PF||72.1 ± 21.4||74.4 ± 21.1||84.7 ± 15.6|
|RP||27.0 ± 34.9||25.0 ± 33.3||58.3 ± 40.0|
|BP||51.0 ± 26.3||47.3 ± 20.4||63.8 ± 21.2|
|GH||33.8 ± 18.5||31.5 ± 17.7||48.8 ± 20.6|
|VT||19.3 ± 13.2||20.4 ± 15.8||38.3 ± 19.9|
|SF||30.6 ± 22.0||37.3 ± 21.1||57.3 ± 21.2|
|RE||9.5 ± 23.5||7.4 ± 17.0||43.5 ± 38.2|
|MH||21.1 ± 12.7||23.1 ± 13.9||44.7 ± 18.6|
|HAMD||23.9 ± 4.7||23.0 ± 4.4||12.0 ± 6.3|
|Headache intensity||4.8 ± 3.3||5.5 ± 3.4||3.0 ± 2.9|
|Headache frequency||2.9 ± 2.6||2.8 ± 2.4||1.9 ± 2.1|
Table 2. Short Form 36, HAMD, and headache profile scores vs presence of migraine
|PF||70.5 ± 20.2||78.7 ± 21.6||0.10||80.0 ± 18.0**||90.0 ± 10.2||<0.006|
|RP||19.1 ± 29.3||31.6 ± 36.6||0.11||48.7 ± 42.3*||69.1 ± 34.8||0.03|
|BP||39.8 ± 18.6**||55.6 ± 19.3||<0.006||56.4 ± 22.3**||72.0 ± 16.7||<0.006|
|GH||30.8 ± 16.6||32.4 ± 19.1||0.71||46.0 ± 21.0||52.0 ± 19.9||0.22|
|VT||16.2 ± 14.8*||25.1 ± 15.7||0.02||33.4 ± 19.5*||43.7 ± 19.2||0.03|
|SF||31.3 ± 19.4*||44.1 ± 21.1||0.009||52.0 ± 22.0*||63.2 ± 18.7||0.02|
|RE||7.9 ± 18.0||6.9 ± 15.9||0.80||34.2 ± 35.1*||53.9 ± 39.4||0.03|
|MH||17.6 ± 12.3**||29.2 ± 13.2||<0.006||39.4 ± 19.3*||50.6 ± 16.1||0.01|
|HAMD||24.3 ± 4.4*||21.6 ± 4.0||0.009||13.8 ± 6.8*||10.0 ± 5.2||0.01|
|Headache intensity||6.7 ± 3.2**||4.1 ± 3.1||<0.006||4.0 ± 3.3**||1.9 ± 1.9||<0.006|
|Headache frequency||3.6 ± 2.5**||1.9 ± 2.0||<0.006||2.4 ± 2.5*||1.4 ± 1.4||0.047|
Differences in the baseline SF-36 and HAMD scores between groups
In the treatment group, 38 subjects (52.8%) had migraine (categorized as the migraine group) and 34 subjects did not (non-migraine group). At baseline, there were no differences in age or educational years between the migraine and non-migraine groups. The migraine group tended to have lower scores in all SF-36 subscales except the RE subscale in comparison with the non-migraine group (Table 2). The differences at baseline in the SF-36 subscale scores between the two groups reached P < 0.05 for the BP, VT, SF, and MH subscales; after Bonferroni correction, only the differences in the BP and MH scales remained significant (P < 0.006). The migraine group had a higher headache intensity and headache frequency than the non-migraine group.
Differences in SF-36 and HAMD scores between groups after treatment
After the 4-week treatment course, the scores in all of the SF-36 subscales significantly increased in the whole treatment group (Tables 1,3) as well as in the migraine and non-migraine subgroups (all P < 0.001; Tables 2,3). In the whole treatment group, the increased scores ranged from 36.1 ± 39.1 (RE) to 17.8 ± 19.3 (VT) in the mental subscales and 33.3 ± 46.2 (RP) to 10.3 ± 16.3 (PF) in the physical subscales (Table 3). Therefore, a trend of greater improvement was observed in the mental subscales as compared with the physical subscales.
Table 3. Changes in Short Form 36, HAMD, and headache profile scores after 4 weeks of venlafaxine
|PF||10.3 ± 16.3||9.5 ± 15.9||11.3 ± 16.8|
|RP||33.3 ± 46.2||29.6 ± 47.5||37.5 ± 44.9|
|BP||16.5 ± 20.6||16.7 ± 21.7||16.4 ± 19.6|
|GH||17.3 ± 21.1||15.3 ± 18.8||19.6 ± 23.4|
|VT||17.8 ± 19.3||17.2 ± 20.3||18.5 ± 18.3|
|SF||20.0 ± 23.6||20.7 ± 24.9||19.1 ± 22.5|
|RE||36.1 ± 39.1||26.3 ± 36.5*||47.1 ± 39.5|
|MH||21.6 ± 17.8||21.8 ± 19.5||21.4 ± 16.0|
|HAMD||11.0 ± 6.4||10.5 ± 7.1||11.6 ± 5.5|
|Headache intensity||2.5 ± 3.5||2.8 ± 3.8||2.2 ± 3.2|
|Headache frequency||0.9 ± 2.7||1.2 ± 3.2||0.4 ± 1.9|
Compared with the non-migraine group, the migraine group had lower scores in all of the SF-36 subscales after treatment (Table 2), with clinically significant differences (difference >5 points). The statistical differences between the two groups in all of the subscales except the GH subscale reached significance (P < 0.05) before Bonferroni correction; after correction, only the PF and BP subscale scores remained significant (P < 0.006). Table 3 lists the increases in scores of the SF-36 subscales in the whole treatment group and the migraine and non-migraine groups.
The HAMD score and headache intensity decreased significantly (all P < 0.001) after treatment in the whole treatment group and in both the migraine and non-migraine subgroups (Table 3). Headache frequency decreased significantly after treatment in the whole treatment group (P < 0.001) and in the migraine subgroup (P = 0.02), but not in the non-migraine subgroup (P = 0.19). A trend of a lower improvement percentage (difference in post-treatment/baseline scores) in HAMD score in the migraine group than in the non-migraine group was observed (42.5 ± 29.3% vs 53.6 ± 22.7%, P = 0.08).
Correlation of the two headache indices with SF-36 and HAMD scores
In the migraine group, headache intensity and frequency at baseline were significantly correlated with BP subscale score (Table 4) and HAMD score, with correlation coefficients (r) of 0.37 and 0.34, respectively (P < 0.05). In the non-migraine group, the two headache indices were not significantly correlated with any of the SF-36 subscales or the HAMD.
Table 4. Pearson correlation coefficients for Short Form 36 subscales
After the 4-week course of treatment, headache intensity and frequency in the migraine group were significantly correlated with SF-36 subscale scores (Table 4) and HAMD score (r = 0.75 and 0.78, respectively, P < 0.01), with the exception of the two headache indices with the PF subscale and headache frequency with the GH subscale. Conversely, in the non-migraine group, the two headache indices were not correlated with SF 36 subscale scores or HAMD score, with the exception of headache intensity with the RP subscale and headache frequency with the PF subscale and the HAMD (r = 0.38, P < 0.05).
In the whole group, HAMD score was significantly correlated with all SF-36 subscales at baseline (r ranging from −0.28 to −0.42; P < 0.05), with the exception of the VT score, and after treatment (r = −0.39 to −0.78; P < 0.05). Age and educational years were not correlated with SF-36 subscale scores either at baseline or after treatment.
Migraine as an independent factor predicting SF-36 subscale scores after treatment
In the regression model, migraine was identified as an independent factor predicting the post-treatment score in the PF, RP, and RE subscales after controlling for demographic variables and baseline scores (Table 5). In the other five subscales, only the baseline scores were entered into the regression model, and no demographic variables were included in the regression model.
Table 5. Independent factors related to subscales of the SF-36 at end of 4-week treatment
|W4 PF||W0 PF||0.61||0.42||50.0||<0.001|
The MDD subjects with migraine had poorer SF-36 BP and PF subscale scores after the 4-week treatment as compared with those without migraine; moreover, headache intensity in the migraine group was still higher than that in the non-migraine group. Previous studies have found some antidepressants, such as venlafaxine and duloxetine, to be effective in treating depression, somatic symptoms, and headaches;23,30,34 a 4-week venlafaxine treatment course, however, did not erase the differences in some subscale scores or headache intensity between the two groups. The present study has demonstrated that the negative impact of migraine on some SF-36 subscales and headache intensity among patients with MDD persisted after 4 weeks of venlafaxine treatment.
After controlling for baseline subscale scores and demographic variables, migraine independently predicted lower scores in the PF, RP, and RE subscales after the 4-week treatment. PF, RP and RE are related to functional impairment as a result of physical and emotional problems, which raises the possibility that migraine may hinder recovery after functional impairment. A previous study also found that patients with migraine have a poorer HRQoL, especially in terms of RP and RE.35 In fact, functional impairment resulting from migraine encompasses multiple dimensions, such as worrying about further attacks and limitation in social and family interaction, as well as capacity to work.36 These results also were compatible with the Lipton et al. study, which indicated that migraine and depression independently predict impaired HRQoL in the general population.22
There are four clinical implications of the present study. First, previous studies have reported that depression has a negative impact on the prognosis of migraine.9,10 The present results have demonstrated the interaction of depression and migraine from the other direction; that is, the negative impact of migraine on depression outcome after acute treatment. Second, pain symptoms, especially headache and muscle soreness, are associated with a poor response to treatment for depression.37 MDD patients with migraine had more severe headaches and more pain symptoms at baseline, which remained the case after treatment. This raises some interesting questions: whether more active treatment of migraine may improve treatment response, and whether or not additional intervention besides antidepressant treatment is indicated. And third, the two headache indices were significantly correlated with most SF-36 subscales and the HAMD in the migraine group after the 4-week treatment but not in the non-migraine group, which implied that the impact of headache intensity and frequency on HRQoL was more significant in patients with migraine as compared to those without migraine. And fourth, subjects in both the migraine and non-migraine subgroups exhibited significant improvement in all subscales of the SF-36, the HAMD, and headache intensity after the 1-month treatment. This finding was compatible with previous studies showing that venlafaxine was effective in the treatment of depression and somatic symptoms, including headache.23,30 There was a trend, however, towards the improvement in the physical subscales being smaller than the improvement in the mental subscales. Greco et al. reported that physical or pain symptoms were more difficult to treat than emotional symptoms among patients with depression.38 This might explain in part the fact that patients with migraine had a persistently poorer HRQoL in the PF and BP subscales as well as a higher headache intensity as compared with patients without migraine.
The present study had several methodological issues or limitations: (i) this open-label study with the exclusion criteria described was performed in a medical center, which might introduce bias, and the small sample size precluded statistical significance; (ii) although the present study was an open-label study, all subjects accepted the same treatment, and the raters of the HAMD were blind to headache diagnosis; (iii) although early treatment response is one predictive factor of prognosis,39 it is unknown whether the present results would be consistent if the treatment period was extended to 2–3 months; (iv) the dosage of venlafaxine used in the present study was 75 mg; whether or not a higher dosage of venlafaxine could compensate for the pre-treatment discrepancy between the two groups needs further study; (v) regarding the evaluation of headache intensity and frequency, a more reliable method would be the use of a prospective headache diary; (vi) the duration of headache disorders was not recorded because the headaches of most of the present subjects were episodic and fluctuating, and therefore a reliable headache duration could not be provided; and (vii) although zolpidem is not an analgesic or a medication for migraine prevention, improvement of insomnia may indirectly improve headache; the migraine group appeared more likely to use zolpidem, and this may cause bias. In future studies, a larger sample size, a longer treatment period, and a double-blind placebo-controlled trial design are indicated in order to investigate the impact of migraine on the treatment prognosis of MDD.
In conclusion, MDD patients with migraine still exhibited a poorer HRQoL in terms of the BP and PF subscales than those without migraine after 4 weeks of treatment. Migraine independently predicted a poor HRQoL as measured by the PF, RP and RE after 4 weeks of treatment after controlling for demographic variables and scores at baseline. Whether comorbid migraine is a negative factor of treatment outcome and of functional recovery in patients with MDD and whether additional intervention for MDD patients with migraine is indicated require further study.
This study was supported in part by grants from the National Science Council of Taiwan (NSC 93–2314-B-182A-200 and NSC 94–2314-B-182A-207) and from Ministry of Education, Aim for the Top University Plan.