Neurology -- Correspondence for Thijs et al., 66 (7) 1034-1037

Correspondence published:

Syncope in migraine: The population-based CAMERA study: Svetlana Blitshteyn, William P. Cheshire, Jr. (25 September 2006)

Syncope in migraine: The population-based CAMERA study: Imad T Jarjour, MD, E O'brian Smith, PhD, Houston, TX (25 September 2006)

Reply from the Authors: Roland D. Thijs, Mark C. Kruit, Mark A. van Buchem, Michel D. Ferrari, Lenore J. Launer, and J. Gert van Dijk (25 September 2006)

Syncope in migraine: The population-based CAMERA study 25 September 2006

Svetlana Blitshteyn,
Department of Neurology, Mayo Clinic
4500 San Pablo Road, Jacksonville, FL 32224,
William P. Cheshire, Jr.
Send Correspondence to journal:
Re: Syncope in migraine: The population-based CAMERA study

blitshteyn.svetlana{at}mayo.edu Svetlana Blitshteyn, et al.

We read with interest the article by Thijs et al on syncope in migraine. [1] These authors found a higher lifetime prevalence of syncope, frequent syncope and symptoms of orthostatic intolerance among migraineurs compared to control subjects as reported by the patients, without finding objective evidence of autonomic system dysfunction when measured by the autonomic reactivity tests. Our clinical experience suggests a clinical correlation at least as frequent as that reported.
We believe that several factors in the study design may have contributed to an underestimation of the presence of ANS dysfunction signs in migraineurs. First, the cases and controls were somewhat older (mean age 48 +/- 8 years) than the typical patients with POTS and orthostatic intolerance, which commonly occur in women 15-50 years of age. [2] By selecting an older population of patients with migraine, the investigators may have failed to capture potential cases of migraineurs with POTS and other forms of ANS dysfunction.
Second, the orthostatic blood pressure and heart rate measurements in this study were performed while patients remained on anti-hypertensive medications. The use of some anti-hypertensive medications, such as beta- blockers, may blunt the orthostatic blood pressure and heart rate response and therefore may produce unreliable autonomic reactivity test results. [3]
Finally, cardiovascular measurements of supine, standing and post- venipuncture BP and HR measurements may not have been sufficiently sensitive in detecting subtle ANS disturbance that may occur in migraineurs, as determined by previous investigators. [4,5] Other tests of autonomic nervous system function that were not performed in this study, such as Valsalva maneuver, resting rate variation during deep breathing, sustained hand grip and head-up tilt test, may have been more informative in identifying subtle signs of ANS dysfunction in migraineurs.
For example, one study found a statistically significant elevated diastolic blood pressure and lowered resting rate variation during deep breathing, as well as a lower Valsalva ratio, which was not statistically significant, in disabled migraine patients compared to controls. [4] Another study of migraineurs also found signs of ANS dysfunction when using Valsalva maneuver, sustained hand grip and head-up tilt test. [5] We believe that further research in this area is needed.
References
1. Thijs RD, Kruit MC, Van Buchem MA, et al. Syncope in migraine: The population based-CAMERA study. Neurology 2006; 66:1034-1037.
2. Low PA, Opfer-Gehrking TL, Textor SC, et al. Postural tachycardia syndrome (POTS). Neurology. 1995;45(suppl 5):S19-25.
3. Low PA. Testing of the autonomic nervous system. Semin Neurol 2003; 23:407-421.
4. Shechter A, Stewart WF, Silberstein SD, Lipton RB. Migraine and autonomic nervous system function: a population-based, case-control study. Neurology 2002;58:422-427.
5. Mosek A, Novak V, Opfer-Gehrking TL, Swanson JW, Low PA. Autonomic dysfunction in migraineurs. Headache 1999;39:108-117.
Disclosure: The authors report no conflicts of interest.

Syncope in migraine: The population-based CAMERA study 25 September 2006

Imad T Jarjour, MD,
Baylor College of Medicine
6621 Fannin St, CC 1250,Houston, TX 77030-2399,
E O'brian Smith, PhD, Houston, TX
Send Correspondence to journal:
Re: Syncope in migraine: The population-based CAMERA study

jarjour{at}bcm.tmc.edu Imad T Jarjour, MD, et al.

The recent report by Thijs et al [1] demonstrated in a population-based study a significantly higher prevalence of syncope in patients with history of migraine, and found no difference in the prevalence of postural orthostatic tachycardia syndrome (POTS) and orthostatic hypotension (OH) between migraineurs and controls. Certain aspects of this study deserve attention.
The authors described controls as "those who had indicated that they had no severe headaches interfering with daily activities and who had rated any headaches they had as 0 on the pain scale. This effectively excluded people with chronic daily headaches and cluster headache." [1] However, the authors did not emphasize that, "people with minor migraine or an occasional tension headache might have been included" among controls, as the authors of an earlier study on the same cohort of patients stated in their report. [6] This is an important omission.
Moreover, how many controls had migraine? And how many controls with OH or POTS had migraine? If migraine and autonomic dysfunction, manifesting as OH or POTS, share a common pathophysiologic mechanism, then the inclusion of migraineurs among controls would decrease the likelihood of detecting differences in the prevalence of OH and POTS between patients and controls.
The authors state that the diagnosis of migraine in 863 cases was based on the 2004 International Headache Society (IHS) criteria. [7] However, a previous report by the same study group on brain lesions in migraineurs indicated using the 1988 IHS criteria to diagnose migraine. [6,8] The two cohorts of patients appear to be the same in the two reports, [1,6] but the number of migraineurs and the prevalence of migraine with aura are the same in the two reports, despite using a different criteria. Even minor changes in the IHS criteria can result in the exclusion of a small percentage of cases of migraine. [9]
Finally, the authors did not report 95% confidence intervals (CI) and odds ratios (OR). The reporting of CI and OR data is valuable for a better understanding of the range and level of observed differences. To illustrate this point, we calculated the 95% CI for the difference (15%) in prevalence of syncope between patients and controls (46% vs. 31%, respectively) to be 6% to 25%. The OR was 1.93, with 95% CI = 1.26-2.97. Hence, a migraineur has a nearly twofold increased odds of having syncope than controls, but this risk could be close to 1 or up to nearly threefold. One wonders what the OR and CI for the prevalence of syncope would be if migraineurs were excluded from the control group?
References
6.Kruit MC, van Buchem MA, Hofman PA, et al. Migraine as a risk factor for subclinical brain lesions. JAMA 2004;291:427-434.
7.Headache Classification Subcommittee of the International Headache Society. The International Classification of Headache Disorders. 2nd ed. Cephalalgia 2004;24(suppl1):9-160.
8.Classification and diagnostic criteria for headache disorders, cranial neuralgias and facial pain. Headache Classification Committee of the International Headache Society. Cephalalgia. 1988;8(suppl7):1-96.
9. Kelman L. Validation of the classification of migraine without aura (IHS A1.1) proposed in ICHD-2. Headache 2005;45:1339-1344.
Disclosure: The authors report no conflicts of interest.

Reply from the Authors 25 September 2006

Roland D. Thijs,
Leiden University Medical Centre
PO Box 9600, 2300RC Leiden, the Netherlands,
Mark C. Kruit, Mark A. van Buchem, Michel D. Ferrari, Lenore J. Launer, and J. Gert van Dijk
Send Correspondence to journal:
Re: Reply from the Authors

r.d.thijs{at}lumc.nl Roland D. Thijs, et al.

Drs. Blitshteyn and Cheshire raise several points suggesting our study may have underestimated the prevalence of autonomic dysfunction in migraine. Overall, it should be noted that our study was population-based and not a patient population recruited from a headache clinic.
Firstly, whether the relatively high age in our study (mean age 48+/- 8 years, range 30 � 60 years) may have missed POTS cases is debatable. POTS indeed typically presents between the ages of 15 to 50 years. Only limited information is available on the prognosis of POTS, which suggests that it tends to persist for many years [10]; if so, cases would still have been present.
Secondly, antihypertensive medication may affect cardiovascular reactivity measures, but a lowered responsivity seems more likely to result in falsely abnormal results rather than falsely normal results. Such medication may cause clinically relevant orthostatic hypotension. Regardless, there were reasons not to exclude subjects with antihypertensive medication: migraineurs with frequent headaches may use antihypertensive medication for migraine prevention, so excluding them would cause exclusion of disabling migraine. Besides this, the use of antihypertensive medication did not differ significantly between patients and controls, and did not affect the prevalence of syncope.
Finally, the use of additional tests such as a Ewing battery may have resulted in the identification of subtle and subclinical abnormalities. However, this was not the purpose of our study. We focused on syncope as a clinically relevant expression of autonomic dysfunction, and not on autonomic alterations in any form. We accordingly chose to perform tests that would have a more direct relation to the onset of syncope and that could be applied in the context of a large population-based study.
We do agree that more extensive cardiovascular measurements are useful to study migraine, and would suggest that these be directed at the reasons for the increased prevalence of syncope and orthostatic intolerance.
We also thank Drs. Jarjour and Smith for their comments. They state that we omitted a statement concerning the control group from an earlier paper on the CAMERA study. In our publication in JAMA 2004, we stated that the control-definition "effectively excluded people with chronic daily headaches and cluster headache, but people with minor migraine or an occasional tension headache might have been included." [6]
Firstly, let us stress that �minor migraine� does not feature in the classification system of the International Headache Society (IHS); the sentence was added on a reviewer's request, presumably to indicate putative uncertainty regarding the groups. Whereas this uncertainty might indeed hold true for the partially �paper� three- step diagnostic procedure [11], it has no consequences for the current study.
The random inclusion of participants for the CAMERA-study from the larger sample involved a detailed, structured telephone interview. We again asked all 481 subjects about previous and current headaches and when necessary corrected the previous classification. Only 3% of subjects were reclassified, most due to developed migraine headache for the first time in the three to five years in between, or due to changes in migraine subtype. This double-checking makes us confident that the control subjects did not have current or previous migraine fulfilling IHS criteria.
Jarjour and Smith are correct that in 2004 the IHS classification was updated. However, criteria for diagnoses 1.1 and 1.2.1, pertinent to our study, were not changed between the 1988 and 2004 edition. [7,8] Finally, we agree that confidence intervals or odds ratios add information. As for how the odds ratios would behave if migraineurs were excluded from the control group, we can lay Drs. Jarjour and Smith's fears to rest: there were none.
References
10. Kanjwal Y, Kosinski D, Grubb BP. The postural orthostatic tachycardia syndrome: definitions, diagnosis, and management. Pacing Clin Electrophysiol 2003;26:1747-1757.
11.Launer LJ, Terwindt GM, Ferrari MD. The prevalence and characteristics of migraine in a population-based cohort: the GEM study. Neurology 1999;53:537-542.
Disclosure: The authors report no conflicts of interest.