Sage Journals: Discover world-class research

Abstract

Sumatriptan, the first marketed triptan, acts predominantly as a peripheral 5-HT_1B/1D agonist. Recently launched (naratriptan and zolmitriptan) triptans with enhanced lipophilicity (TEL) may theoretically reduce central 5-HT levels, potentially exacerbating depressive illness. The aim of this study was to compare depressive illness patient consulting rates (PCR) between the newer triptans and sumatriptan treated patients. Migraine-diagnosed patients were identified from the West Midlands General Practice Research Database (GPRD). Univariate and logistic regression analyses were used to determine the relationship between age, sex, migraine duration, prior depression, prior and current triptan medication and the current PCR for depression. Of a total patient database of 642 000 patients 21 331 (3.3%) had a diagnosis of migraine. From 1993 to 1997, sumatriptan-treated migraineurs had significantly higher depression PCRs (22.3%) compared with non-triptan users (19.3%), a difference of 6.4% (95% confidence interval (CI) 4.6–8.4%, P < 0.001). In the year (April 1997 to March 1998) following the launch of the TELs, depression PCRs were significantly higher among patients using these compounds compared with sumatriptan-treated patients (5.1%, CI 1.8–12.0%, P < 0.05). However, after taking account of prior depression (odds ratio (OR) 6.45, 95% CI 3.63–11.43), TELs were not significantly associated with depression (OR 0.27, 95% CI 0.03–2.13). Furthermore, rates of newly diagnosed depression after treatment were similar in the two triptan groups (sumatriptan 4.2%; TELs 3.9%). Although, the TELs are being prescribed to patients with higher pre-existing rates of depression, they are not associated with subsequently increased consulting for depressive illness compared with patients taking sumatriptan. This study highlights the potential to use GPRD to test targeted hypotheses exploring pharmacovigilance issues for patients using new medicines.

Keywords

Triptan migraine depression pharmacovigilance

Introduction

The introduction of sumatriptan in 1992 was a major therapeutic innovation for the acute treatment of migraine (1). Sumatriptan has, however, been implicated as a possible contributory factor in the pathophysiology of central serotoninergic syndrome, i.e. myoclonus and hyperexcitability (2) and recurrent depression (3). However, both preclinical (4) and clinical data have failed to convincingly demonstrate central activity with sumatriptan (5–7). Central 5-HT_1B/1D receptors clearly do have a role to play in the neuromodulation of serotonin (8). The 5-HT_1B/1D-receptor antagonist GR127935 modifies 5-HT release in the frontal cortex and raphe nucleus (9, 10), and is a potential treatment for both motor disorders and depression (11). An abnormal sensitivity of central 5-HT_1D receptors has been demonstrated in patients with major depression (12). This suggests a potential disease–treatment interaction when patients with underlying depressive illness are prescribed a 5-HT_1B/1D agonist triptan with enhanced lipophilicity (i.e. eletriptan, naratriptan, rizatriptan or zolmitriptan) and thus more central nervous system (CNS) penetration (13). Such interactions can be complex and may involve more than brain penetration issues.

There is strong evidence to support an association between migraine and major depression (14). In a US population-based representative sample (1007 young adults aged 21–30 years) from a Detroit Health Maintenance Organization (HMO) with 400 000 members, structured diagnostic interviews were carried out using International Headache Society (IHS) criteria and current definitions of psychiatric disorders (DSM-III-R:6). Migraineurs had increased lifetime rates of major depression, which when analysed using a longitudinal analysis after 3.5 years found at the age of 33 years that the cumulative incidence of migraine was 19.5% and major depression was 22.5% for both sexes. The relative risk of major depression (adjusted for sex and education using a Cox Proportional Hazards Model) for those with a prior onset of migraine is 3.2 (15). Similarly the relative risk of migraine for patients with depression is greater than expected in any one year. These data suggest a comorbidity based on some underlying shared biology.

The General Practice Research Database (GPRD) systematically provides both drug utilization and diagnostic data from over 4000 UK General Practitioners (GPs) with a population coverage in excess of 3 million. The GPRD is widely regarded as ‘the largest and most valid resource available world-wide for drug safety research and pharmacoepidemiology’ (16). The quality of GPRD data is regularly monitored by the Office of National Statistics and several independent studies have found the data to be of good quality. In the area of psychiatric disorders, Nazareth and colleagues (17) found a high degree of concordance between computer-entered data and case notes.

The aim of this study was to investigate whether the newer triptans, naratriptan and zolmitriptan, are associated with increased consulting for depressive illness than sumatriptan and if so, to investigate confounding factors. Two specific issues are addressed. First, does the 5-year patient consulting rate for depression differ among migraine-diagnosed patients receiving triptans compared with those not receiving triptans? Second, does the 1-year patient consulting rate for depression differ among migraine-diagnosed patients receiving sumatriptan compared with those receiving triptans with enhanced lipophilicity (naratriptan and zolmitriptan)?

Patients and methods

Sampling frame

The sampling frame was the West Midlands GPRD which is owned and maintained by the Medicines Control Agency. The data are held on a Microsoft Access relational database with tables on patient details, prescriptions, and medical records (diagnoses, Hospital referrals). The tables are linked by an anonymized and encrypted patient identifier. The database contains 642 986 individual patient records for patients consulting between 1987 and 1997. The period of observation for the analysis in this study was January 1993 to March 1998. Participating GPs are required to record every prescription and all significant morbidity and consultation outcomes (e.g. Hospital referral). The number of practices participating was 53 in 1993/94, 52 in 1995/96 and 41 in 1997/98.

A comparison of the 1996 GPRD population with the mid-1996 population of England and Wales shows that the distributions are similar with regard to age and sex (18). Socio-economic comparisons cannot be made because the database does not contain any information on patients' socio-economic status. The 1996 GPRD population coverage for the West Midlands was 5.3% compared with 3.9% for the whole of England and Wales.

Study populations

Cohort 1. Registered patients as of January 1993 with a previous diagnosis of migraine.

Cohort 2. Registered patients as of April 1997 with a previous diagnosis of migraine.

Analysis time frames

Cohort 1 was followed from 1993 to 1997. 1993 was the first full year following the introduction of sumatriptan in October 1992 and 1997 is the last full year for the current database. Cohort 2 were followed from April 1997 to March 1998. April 1997 saw the launch of naratriptan and zolmitriptan, while March 1998 is the date of the latest available data.

Data and outcome measures

The 11 diagnostic codes relating to migraine were selected from the 18 695 OXMIS codes (Oxford Medical Information Systems) on the database (OXMIS codes were devised for use by GPs and are based on ICD8 codes and OPCS operation codes). These codes are shown in the Appendix. All diagnoses with these codes were selected from 1993 to 1997. Details concerning triptan medication and depression diagnoses (23 OXMIS codes) were extracted for patients with migraine diagnoses (see Appendix). The study outcome measure is the patient consulting rate (PCR), which is defined as ‘consulting at least once for depression during a defined time period’. In order to address the study objectives, two PCRs were assessed: Cohort 1. Five-year PCR for depression (see Appendix) from 1993 to 1997 for (i) diagnosed migraine patients being prescribed an oral triptan, and (ii) diagnosed migraine patients not receiving any triptan; Cohort 2. One-year PCR for depression from April 1997 to March 1998 for (i) diagnosed migraine patients being prescribed sumatriptan, and (ii) diagnosed migraine patients receiving a triptan with enhanced lipophilicity (TEL) (naratriptan, zolmitriptan). The PCR was also calculated separately for first depression diagnosis subsequent to the first triptan prescription. Patients receiving both sumatriptan and TELs during the study year were included in both subgroups for the analysis.

Analysis

Univariate

χ² tests were performed to assess differences in the PCRs between the subgroups in both cohorts 1 and 2. In addition to statistical significance, a clinically significant difference between groups was prospectively defined as being > 5% at any particular point in time. This is at least 10 times the background incidence of spontaneously reported depression in a New Zealand Intensive Medicines Monitoring Programme (1997) in 3.6 million patients (where the rate per 100 patients was reported as being 0.5 per year).

Multivariate

In order to control for potential confounding factors, logistic regression analysis was performed with the outcome as depression diagnosed during period 2 (April 1997 to March 1998), coded 1 (one or more diagnoses) or 0 (no diagnosis). The independent variables were current triptan medication in period 2, triptan medication from January 1993 to March 1997, depression diagnosed during January 1993 to March 1997. These variables were also coded 1 or 0 depending on their presence or absence during the specified period. The other independent variables were duration since migraine diagnosis at cohort entry, age and gender. The analysis was conducted using SPSS/PC. Crude and adjusted odds ratios (OR) are presented.

Results

The denominators for periods one and two were the number of patients with a diagnosis of migraine in their medical records and registered as of January 1993 (n = 20 059) and April 1997 (n = 12 977), respectively. The fall in the number of patients between 1993 and 1997 was mainly due to practices moving from the GPRD to other software platforms. Figure 1 shows that the cumulative diagnosis rates for headache and migraine were 10.2% and 3.3%, respectively (1987–97).

Figure 1

One year and 5-year patient consulting rates (PCRs) for depression among migraine-diagnosed patients. TEL, Triptans with enhanced lipophilicity; naratriptan & zolmitriptan. ∗The 41 patients receiving sumatriptan and a TEL were included in both groups.

Cohort 1

Migraine-diagnosed patients who received oral triptan were more likely to have had a diagnosis of depression during the same period (6.4% difference, 95% confidence interval (CI) 4.6–8.4%, P < 0.001) than patients who had not received any triptan.

Cohort 2

The 1-year PCR for depression was clinically and statistically significantly higher (5.1%, 95% CI 1.8–12.0%, P < 0.05) for patients receiving a TEL (i.e. naratriptan or zolmitriptan) during this time, compared with those receiving sumatriptan. However, when those patients diagnosed with depression after their first prescription with either sumatriptan or a TEL were compared, then depression rates were no different between the two groups.

Logistic regression analysis

Data for 678 patients receiving triptan medication from April 1997 to March 1998 were entered into the analysis. Sixty patients (8.8%) were diagnosed with depression during this period; of these 30 (50%) were diagnosed depressed during this period after receiving their first triptan prescription. Column A of Table 1 shows that the only significant predictor of depression among migraineurs receiving a triptan between April 1997 to March 1998 was a prior depression diagnosis (January 1993 to March 1997) which had a crude OR of 6.40 (95% CI, 3.68–11.12). The crude OR for TEL use was 1.68 (95% CI, 0.94–3.00). When all variables were entered in the logistic regression model (column B of Table 1), the only significant adjusted OR for prior depression was 6.45 (95% CI, 3.63–11.43). Use of TELs had an OR of 0.27 (95% CI, 0.03–2.13) after controlling for previous depression.

TABLE 1

Crude and adjusted odds ratios (OR) for depression diagnosis, April 1997 and March 1998

Predictor variables	Column ACrude OR	Column BAdjusted OR
Age at April 1997 (n = 678)†	0.98 (0.96–1.01)	0.98 (0.95–1.01)
Duration of migraine diagnosis at April 1997 (n = 678)†	0.99 (0.97–1.02)	0.99 (0.97–1.02)
Gender
Female (n = 550)	1.0	1.0
Male (n = 128)	0.45 (0.19–1.07)	0.49 (0.20–1.22)
Prior depression diagnosis (Jan 1993 to March 1997)
No (n = 546)	1.0	1.0
Yes (n = 132)	6.40 (3.68–11.12)∗	6.45 (3.63–11.43)∗
Prior sumatriptan (Jan 1993 to March 1997)
No (n = 305)	1.0	1.0
Yes (n = 373)	0.74 (0.44–1.26)	0.98 (0.53–1.84)
Current sumatriptan (April 1997 to March 1998)
No (n = 111)	1.0	1.0
Yes (n = 567)	0.41 (0.23–0.74)∗	0.15 (0.02–1.29)
Current TEL (April 1997 to March 1998)
No (n = 526)	1.0	1.0
Yes (n = 152)	1.68 (0.94–3.00)	0.27 (0.03–2.13)

†

Odds ratios for variables coded in years are the odds of receiving a depression diagnosis for a change of 1 year (i.e. as age and duration of migraine diagnosis increase by 1 year, the patient consulting rate (PCR) for depression changes by 0.02 and 0.01, respectively).

∗

P < 0.01.

TEL, Triptans with enhanced lipophilicity.

Discussion

The results of this study show that TELs are associated with a higher prevalence of depressive illness than sumatriptan. However, after taking into account prior diagnosis of depression, there is no elevated risk of depression among migraineurs using TELs. It should be noted that these findings are only generalizable to patients with a recorded diagnosed of migraine and/or depression made by their GP. This would bias the study results if patients receiving the TELs were systematically under-diagnosed for depression, for which we have no evidence. Another possible limitation is that factors associated with medical diagnoses of depression and migraine could influence choice of treatment. For example, patients with pre-existing depression could be prescribed antidepressants which may prevent migraine headaches reducing the requirements for acute treatment such as triptans (19).

The TELs have undoubtedly more potential for central activity than sumatriptan. This is borne out by the selective effect of zolmitriptan on the auditory evoked potential (20, 21) and preclinical data with naratriptan (22). However, when prescribed to migraineurs who are prone to depression, they do not appear to lead to an exacerbation of depression.

Whilst existing pharmacovigilance systems receive and analyse spontaneous adverse reactions to the newly introduced drugs, such as naratriptan and zolmitriptan, there is no provision for detecting population level changes in comorbidity (23). The GPRD enables a priori hypotheses about new medicines to be tested at a population level, with sufficient statistical power to detect clinically important changes. The database also has the advantage of monitoring GPs' actual use of drugs, yet does not interfere in their prescribing decisions. However, this means that the results are not derived from a randomized sample and the potential to control for confounding factors is limited by the available data. In conclusion, the present study has demonstrated that the GPRD can provide longitudinal follow up of patients receiving new drugs, assessing their risks and benefits in relation to clearly defined outcome measures.

Footnotes

Acknowledgements

The authors would like to thank the General Practice Research Database team at the Office of National Statistics, London. The study was conducted following review and approval by the GPRD's Scientific Ethical and Advisory Group.

Appendix. OXMIS codes for migraine and depression migraine diagnoses codes

References

Ferrari

. Migraine—a review. Lancet 1998; 351: 1043 5.

Lane

Baldwin

. Selective serotonin reuptake inhibitor induced serotonin syndrome: review. J Clin Psychopharmacol 1997; 17: 208 21.

La Porta

. Recurrent depression after sumatriptan administration for treatment of migraine. J Clin Psychopharmacol 1995; 15: 81 82.

Kaube

Hoskin

Goadsby

. Inhibition by sumatriptan of newer trigeminal neurones only after blood–brain barrier disruption. Br J Pharmacol 1993; 109: 877 92.

Blier

Bergeron

. The safety of concomitant use of sumatriptan and antidepressant treatments. J Clin Psychopharm 1995; 15: 106 9.

Schoenen

. Measuring newer action of acute antimigraine drugs in humans. Cephalagia 1997; 17: 12 16.

Leung

Ong

. Lack of an interaction between sumatriptan and selective serotonin reuptake inhibitors. Headache 1995; 35: 488 9.

Martin

. Serotonin 5-HT_1B/1D receptor agonists. In: Leff

, ed. Receptor-based drug design. New York: Marcell Dekker, 1998.

Skingle

Sleight

Fenuiuk

. Effects of the 5-HT1D receptor antagonist GR127935 on extracellular levels of 5-HT in the guinea-pig frontal cortex as measured by microdialysis. Neuropharmacology 1995; 34: 377 82.

10.

Starkey

Skingle

. 5-HT1D as well as 5-HT1A autoreceptors modulate 5-HT release in the guinea-pig dorsal raphe nucleus. Neuropharmacology 1994; 33: 393 402.

11.

Hamel

. 5-HT1D receptors: pharmacology and therapeutic potential. Current Drugs-Serotonin ID Res Alert 1996; 1: 19 29.

12.

Cleare

Murray

Sherwood

O'keane

. Abnormal 5-HT1D receptor function in major depression. Psychol Med 1998; 28: 295 300.

13.

Goadsby

. A triptan too far. J Neurol Neurosurg Psychiatry 1998; 64: 143 7.

14.

Breslau

. Psychiatric comorbidity in migraine. Cephalalgia 1998; 18 (Suppl. 22):56 61.

15.

Breslau

Davis

Schultz

Peterson

. Migraine and major depression: a longitudinal study. Headache 1994; 34: 387 93.

16.

Rodriguez

LAG

Gutthann

. Use of the UK General Practice Research Database for Pharmacoepidemiology. Br J Clin Pharmacol 1998; 45: 419 25.

17.

Nazareth

King

Haines

Rangel

Myer

. Accuracy of diagnosis of psychosis on general practice computer system. BMJ 1993; 307: 32 34.

18.

Key Health Statistics from General Practice. Analyses of morbidity and treatment data including time trends, England and Wales. Appendix 2. Office for National Statistics Series MB6 No 1. 1996:83.

19.

Silberstein

. Evidence based guidelines for migraine headache. In: Proceedings of the 42nd Annual Scientific Meeting of the American Headache Society, San Diego, 2000.

20.

Proietti-Cecchini

Afra

Schoenen

. Intensity dependence of the cortical auditory evoked potentials as a surrogate marker of CNS serotonin transmission in man: demonstration of a central effect for the 5-HT_1B/1D agonist zolmitriptan. Cephalalgia 1997; 17: 849 55.

21.

Hughes

Dixon

Yates

Putative central mechanism of action for zolmitriptan in humans. AASH meeting June, 1998.

22.

Williams

Clare

Crisp

Snell

Fuseau

Kempsford

. A study to investigate the central effects of naratriptan in man. J Neurological Sci 1997; 150 (Suppl. S306):S21 24.

23.

Walley

Mantgani

. The UK general practice research database. Lancet 1997; 350: 1097 9.

Are Triptans with Enhanced Lipophilicity Used for the Acute Treatment of Migraine Associated with an Increased Consulting Rate for Depressive Illness?