Sage Journals: Discover world-class research

Abstract

Background

Results from studies on diabetes and migraine risk are conflicting, which may be due to methodological limitations. Prospective studies with long follow-up could increase our understanding of the relationship between the two diseases.

Method

We performed a cohort study including the whole Norwegian population alive on 01.01.2004, using prescriptions registered in the Norwegian prescription database to identify individuals developing type 1 diabetes, type 2 diabetes and migraine during follow-up (10 years). We used Cox proportional hazards regression to estimate rate ratios with corresponding 95% confidence intervals for the effect of diabetes on migraine risk, adjusting for age, sex, and educational level.

Result

We identified 7,883 type 1 diabetes patients and 93,600 type 2 patients during the study period. Type 1 diabetes was significantly associated with a subsequent decreased migraine risk during follow-up in the age- and sex-adjusted analyses (0.74; 0.61–0.89). Type 2 diabetes was also associated with a significantly lower migraine risk (0.89; 0.83–0.95). Further adjustment for educational level yielded similar results for both diabetes.

Conclusion

Both type 1 and type 2 diabetes were significantly associated with a decreased risk of migraine. This suggests that diabetes or diabetes treatment may have a protective effect on the development of migraine.

Keywords

Type 1 diabetes Type 2 diabetes migraine risk epidemiology

Introduction

Migraine is a brain disorder characterized by recurrent headache, photophobia and phonophobia (1). It has been suggested that there may be a link between diabetes and migraine, although results from previous studies are conflicting. While higher glucose levels and insulin resistance have been reported in migraine patients compared to healthy individuals in two cross-sectional studies (2,3), other studies have reported lower prevalence of migraine in diabetes patients compared with the general population (4 –8), suggesting a protective role of diabetes in migraine development. Still, a prospective cohort and a cross-sectional study have reported no association between the two diseases (9,10).

Diabetes involves pathophysiological mechanisms that may be relevant for migraine. Some comorbidities associated with diabetes, such as cardiovascular (11,12) and neurologic complications (13), could influence migraine development (14). For example, neuropathy, a well-known diabetes complication, could affect migraine by reducing pain sensation. Despite the possible biologic overlap, little is currently known about a possible link between the two diseases. Data from large longitudinal studies describing the temporal relationship could provide clues on possible mechanisms contributing to the association.

To address this, we conducted a 10 year nationwide registry-based cohort study to examine whether diabetes is associated with the risk of migraine.

Methods

Data sources

For the current study, we included information from the Norwegian Prescription Database (NorPD) (15), the Norwegian National Registry (16) and the Norwegian National Education Database (NNED) (17). Information on drug use during follow-up was retrieved from NorPD. This registry includes complete information on drug prescriptions dispensed at Norwegian pharmacies to individuals since 2004. Drugs are classified according to the Anatomical Therapeutic Chemical classification (ATC), and doses are registered as defined daily dose (DDD). In addition, since 2008, it includes diagnosis codes on prescriptions where the medication was covered by the national insurance scheme. This includes chronic diseases such as diabetes. Information on sex, date of birth, emigration out of Norway, and date of death was retrieved from the Norwegian National Registry. Lastly, we retrieved the highest level of education from NNED as a proxy for socioeconomic status. The information in the registries was linked using the unique national identification number.

Study population

For this study, we included the whole Norwegian population (<80 years old) alive and resident in Norway on 1 January 2004. The oldest age group was excluded, as information on prescriptions in nursing homes is not included in NorPD, and the registry may thus be less complete for the oldest age groups.

Exposure

Information on the use of antidiabetic and antimigraine drugs was retrieved from NorPD and was used as a proxy of disease. Using the national identification number, we generated an exact chronological prescription record for each individual over time.

We defined type 1 diabetes (T1D) patients as individuals treated with at least 180 DDDs of insulin (ATC: A10A) without any oral antidiabetic drugs for the entire follow-up period and type 2 diabetes (T2D) patients as those treated with at least 180 DDDs of oral antidiabetic drugs (ATC: A10B) with or without concomitant insulin prescription. We used 2004 as a washout period to only include incident cases, as disease onset could not be determined for individuals who used antidiabetic drugs in 2004. Individuals were considered exposed from the date of 180 DDDs of any antidiabetic treatment after 2004.

Outcome

New prescriptions of ergotamine and/or triptans (ATC: N02CA and N02CC, respectively) during the follow-up were used as a proxy for incident migraine. Since migraine drugs are specifically used to treat an acute migraine attack, the date at first migraine prescription was considered date of disease onset. To only include incident cases, the year 2004 was used as a washout period.

Statistical analysis

We compared the risk of developing migraine among individuals that developed T1D or T2D during the study period with the rest of the population. Individuals contributed with person time from 1 January 2005 until the date of migraine onset, date of death, date of emigration out of Norway or end of follow-up (31 December 2014), whichever occurred first. We used Cox proportional hazards regression to calculate hazard ratios (HRs), which were interpreted as rate ratios (RR), with 95% confidence intervals (CI) for the effect of T1D and T2D (disease itself or diabetes treatment) on migraine risk, and included the exposure as a time-dependent variable. We categorized age in 5-year intervals and educational level in primary, secondary, undergraduate, and graduate level. These variables were included in two separate multivariable models, the first adjusted for sex and age, and the second additionally adjusted for educational level. We tested for effect modification by age and sex by comparing a model with an interaction term with a model without an interaction term using the likelihood ratio test in separate models. The analyses were repeated among those with specific diagnosis codes (since 2008) on their prescription in order to validate the results from the analyses using prescriptions as proxies for disease. Lastly, we evaluated whether the association between diabetes and migraine is bidirectional by examining whether migraine was associated with an altered subsequent T1D and T2D risk.

Statistical analyses were performed using Stata statistical software version 14.1.

Ethical approval

The Regional Committee for Medical and Health Research Ethics (REK Vest) approved the study.

Results

In this study, we identified 7,883 incident T1D patients, 93,600 incident T2D patients and 4,184,718 non-diabetic individuals during the study period. Individuals developing T1D were younger and had a lower education compared to those developing T2D and controls (Table 1). The proportion of men and women was similar in all groups. The median age at diagnosis was 29 for those developing T1D and 60 for those developing T2D.

Table 1.

Baseline demographic characteristics of study population*.

	Type 1 diabetes	Type 2 diabetes	Non-diabetics
Number	7,883	93,600	4,184,718
Gender
Male (%)	59.7	57.5	50.7
Female (%)	40.3	42.5	49.3
Age, n (SD)	27.7 (21.9)	52.7 (13.9)	34.4 (21.0)
Educational level⁽⁺⁾
Elementary (%)	21.2	35.9	22.9
Secondary (%)	26.5	46.3	34.3
Undergraduate (%)	9.5	12.8	15.6
Graduate (%)	3.5	3.5	4.8
Missing (%)	39.4	1.4	22.5

The complete Norwegian population in 2004 aged 0 to 79.

Type 1 diabetes: Defined as patients treated with at least 180DDD (Defined Daily Dose) of insulin only (A10A).

Type 2 diabetes: Defined as patients treated with at least 180 DDD of oral antidiabetic agents (A10B) with or without insulin prescription.

⁽⁺⁾Only individuals aged 15 years or older are included in the educational registry. Values of polytomous variables may not sum up to 100% due to rounding.

T1D was associated with a significantly lower risk of migraine (RR 0.74; 95% Cl: 0.61–0.89; Table 2). T2D was also associated with a lower migraine risk (RR 0.86; 95% Cl: 0.80–0.92). The risk estimates remained similar after adjustment for educational level.

Table 2.

Hazard Risk (HR) for migraine in diabetes patients compared to the general population.

				HR (95% Cl)
	Persons, n	Migraine, n	Person-years	Age- and sex adjusted^(a)	Multivariable adjusted^(b)
Diabetes diagnosis during follow-up
No	4.184,718	179,429	39,617,964	(Reference)	(Reference)
Yes
Type 1	7,883	109	37,165.2	0.74 (0.61–0.89)	0.74 (0.61–0.89)
Type 2	93,600	930	403,970.1	0.89 (0.83–0.95)	0.86 (0.80-0.92)

Type 1 diabetes: Defined as patients treated with at least 180DDD (Defined Daily Dose) of insulin only (A10A).

Type 2 diabetes: Defined as patients treated with at least 180 DDD of oral antidiabetic agents (A10B) with or without insulin prescription.

^(a)Adjusted for age in 5 year intervals and sex.

^(b)Adjusted for age in 5 year intervals, sex, and educational level.

We observed that the risk estimate in the association between T2D and migraine was lower in men compared to women (Table 3), but the difference was not significant (p for effect modification = 0.36). There was no difference between men and women in the risk estimates for T1D and migraine.

Table 3.

Hazard risk (HR) for migraine in diabetes patients compared to the general population stratified by sex.

	No diabetes		Type 1 diabetes			Type 2 diabetes
Sex	Individuals, n	Migraine, n	Patients, n	Migraine, n	HR (95% Cl)^(a)	Patients, n	Migraine, n	HR (95%Cl)^(a)
Male	2,122,413	47,443	4,703	39	0.75 (0.55–1.03)	53,853	272	0.82 (0.73–0.92)
Female	2,062,305	131,986	3,180	70	0.74 (0.58–0.93)	39,747	658	0.91 (0.84–0.98)

Type 1 diabetes: Defined as patients treated with at least 180DDD (Defined Daily Dose) of insulin only (A10A).

Type 2 diabetes: Defined as patients treated with at least 180 DDD of oral antidiabetic agents (A10B) with or without insulin prescription.

^(a)Adjusted for age in 5 year intervals.

We found a significant effect modification by age in the association between T2D and migraine (p < 0.001), where the inverse association was only found in the oldest age groups (Table 4). There was no significant effect modification by age in the association between T1D and migraine (p = 0.10).

Table 4.

Hazard Risk (HR) for migraine in diabetes patients compared to the general population stratified by age.

	No diabetes		Type 1 diabetes			Type 2 diabetes
Age	Individuals, n	Migraine, n	Patients, n	Migraine, n	HR (95% Cl)^(a)	Patients, n	Migraine, n	HR (95% Cl)^(a)
0–19	1,211,365	57,059	3,576	69	0.84 (0.66–1.06)	1,464	62	1.17 (0.91–1.50)
20–29	590,439	38,259	1,055	16	0.51 (0.31–0.84)	4,167	169	1.37 (1.18–1.59)
30–39	684,388	40,931	889	14	0.61 (0.36–1.02)	10,769	243	1.16 (1.02–1.32)
40–49	595,011	25,682	724	4	0.40 (0.15–1.06)	19,413	210	0.95 (0.83–1.09)
50–59	528,475	12,106	759	4	0.86 (0.32–2.29)	27,273	165	1.03 (0.88–1.20)
≥60	575,040	5,392	880	2	0.94 (0.24–3.78)	30,514	81	0.88 (0.71–1.10)

Type 1 diabetes: Defined as patients treated with at least 180DDD (Defined Daily Dose) of insulin only (A10A).

Type 2 diabetes: Defined as patients treated with at least 180 DDD of oral antidiabetic agents (A10B) with or without insulin prescription.

^(a)Adjusted for sex and age (continuous variable).

The risk estimates in the analyses restricted to those with the diagnosis codes for T1D or T2D on their prescription (from 2008) were consistent with those from the main analyses (Table 5). In the fully adjusted multivariable analyses, the RR for migraine was 0.77 (95% CI: 0.57–1.05) and 0.91 (95% CI: 0.82–1.01) for T1D and T2D, respectively.

Table 5.

Hazard Risk (HR) for migraine patients prescribed antidiabetic drugs specifically for diabetes type 1 or diabetes type 2.

				HR (95%Cl)
	Individuals, n	Migraine, n	Person-years	Age- and sex adjusted^(a)	Multivariable adjusted^(b)
Diabetes diagnosis during follow-up
No	4,042,989	110,231	26,567,479	(Reference)	(Reference)
Yes
Type 1	3,595	42	12,392.7	0.76 (0.57–1.05)	0.77 (0.57–1.05)
Type 2	53,486	377	195,677.7	0.96 (0.86–1.06)	0.91 (0.82–1.01)

Type 1 diabetes: Defined as patients treated with at least 180DDD (Defined Daily Dose) of insulin only (A10A).

Type 2 diabetes: Defined as patients treated with at least 180 DDD of oral antidiabetic agents (A10B) with or without insulin prescription.

^(a)Adjusted for age in 5 year intervals and sex.

^(b)Adjusted for age in 5 year intervals, sex, and educational level.

We observed no association between treatment for migraine and subsequent T1D risk (RR 0.90; 95% Cl: 0.75–1.07). However, migraine was associated with a slightly increased risk of T2D (RR 1.12; 95% Cl: 1.06–1.17).

Discussion

In this nationwide cohort study, we found that prior use of antidiabetic drugs was associated with a significantly decreased risk of medically treated migraine. Both use of insulin alone, as a marker of T1D, and use of oral antidiabetic drugs, a marker of T2D, were significantly associated with a lower migraine risk. Overall, this suggests that diabetes itself, complications due to diabetes or diabetes treatment could have protective effects on the risk of migraine.

Our findings are consistent with two previous cross-sectional studies examining the association between diabetes and migraine (5,7). In a previous study, using prescriptions of antidiabetic and migraine drugs in 2006, we found a lower prevalence of medically treated migraine among persons with diabetes compared to the nondiabetic population (5). Further, a recent cross-sectional study reported an inverse association between T1D and migraine, which remained similar in a multivariable model adjusted for smoking and body mass index (7). However, one study found a lower prevalence of migraine in T2D patients compared to healthy individuals (10). The conflicting results may be due to the different study designs, sample sizes, and inclusion criteria of diabetic patients. Our study adds weight to previous studies suggesting a link between diabetes and migraine, as our findings are less prone to temporality bias, which is a major limitation in cross-sectional studies. To our knowledge, this is the first cohort study on the link between diabetes and migraine risk.

In our study, we also observed that migraine was associated with an increased T2D risk. This is not in line with an earlier study that found no significant association between migraine and T2D risk in women with different types of migraine after adjusting for potential confounders, including body mass index (9). Several studies have examined whether migraine is associated with insulin resistance, which is a predictor of development of T2D (18), but the results are inconsistent. While two studies found higher blood glucose levels in migraine patients compared to controls (2,3), other studies did not find any significant associations (8,19,20). Our study benefits from a large sample size and long follow-up, but we did not have access to blood samples. Thus, the conflicting results may be due to different study outcomes and designs.

Diabetes may involve biological mechanisms relevant to the pathogenesis of migraine. Neuropathy is a well-known diabetes complication characterized by nerve damage with consequent reduction of pain sensitivity (21), which may reduce the sensation or affect the threshold of pain involved in migraine. Additionally, a study suggested a role of the insulin receptors gene (INSR) in migraine development (22). This gene encodes the insulin receptor, which is expressed in different cell types including neurons and endothelial cells. Alterations of its function may influence neuronal excitability and cerebral blood flow related to migraine attacks. Therefore, diabetes drugs affecting glucose blood level, and thus insulin levels, may prevent receptor dysfunction and nerve dysregulation with consequent prevention of the early phase of migraine (1).

Our study has some limitations. As prescriptions were used to classify the disease status, misclassification could have occurred. Nevertheless, the results obtained using this strategy were consistent with the results in the analyses restricted to those with a specific diagnosis code on their prescription. This makes it unlikely that misclassification of disease status can fully explain our findings. Further, we included only patients with at least 180 DDDs, to avoid the inclusion of less certain diagnoses and those on therapy for a short time (e.g. treatment for gestational diabetes). T2D patients managing their disease through lifestyle interventions only or individuals with undiagnosed or subclinical T2D have been misclassified as non-diabetics. Similarly, migraine patients managing their disease with over-the-counter drugs only (e.g. ibuprofen) have been misclassified as free of migraine. However, due to the size of the reference group in our study, it is unlikely that this represents a major source of bias. Additionally, in our study, at least 70% of T2D patients were included, since most of T2D patients in Norway use antidiabetic drugs (23). Although we believe that no major bias has been introduced by the misclassification of persons with migraine or diabetes treating their disease without over-the-counter drugs or lifestyle changes as non-cases and non-exposed persons, our results cannot be generalised to these groups of persons with diabetes and migraine. Lastly, we did not have information on environmental and genetic risk factors relevant for diabetes and migraine, and could therefore not adjust for these in the analyses. However, our results were in line with a previous study where the authors adjusted for body mass index and smoking (7).

In conclusion, we found that both T1D and T2D were associated with a significantly lower risk of migraine. This suggests that diabetes or treatment of diabetes could have a protective role on migraine.

Public health relevance

Type 1 and type 2 diabetes mellitus were associated with a decreased migraine risk in a longitudinal study with 10 years of follow up, suggesting that diabetes or treatment of diabetes may have a protective effect on migraine development.

Type 2 diabetes at young age was associated with increased migraine risk during the follow-up.

Migraine was associated with an increased risk of type 2 diabetes mellitus, suggesting possible common risk factors.

Footnotes

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

References

Kokavec

. Migraine: A disorder of metabolism? Med Hypotheses 2016; 97: 117–130.

Rainero

Limone

Ferrero

et al.

Insulin sensitivity is impaired in patients with migraine. Cephalalgia 2005; 25: 593–597.

Fava

Pirritano

Consoli

et al.

Chronic migraine in women is associated with insulin resistance: A cross-sectional study. Eur J Neurol 2014; 21: 267–272.

Blau

Pyke

. Effect of diabetes on migraine. Lancet 1970; 2: 241–243.

Berge

Riise

Fasmer

et al.

Does diabetes have a protective effect on migraine?

Epidemiology 2013; 24: 129–134.

Burn

Machin

Waters

. Prevalence of migraine in patients with diabetes. Br Med J (Clin Res Ed) 1984; 289: 1579–1580.

Hagen K, Asvold BO, Midthjell K, et al. Inverse relationship between type 1 diabetes mellitus and migraine. Data from the Nord-Trondelag Health Surveys 1995–1997 and 2006–2008. Cephalalgia. 2018; 38: 417–426.

Aamodt

Stovner

Midthjell

et al.

Headache prevalence related to diabetes mellitus. The Head-HUNT study. Eur J Neurol 2007; 14: 738–744.

Burch

Rist

Winter

et al.

Migraine and risk of incident diabetes in women: A prospective study. Cephalalgia 2012; 32: 991–997.

10.

Haghighi

Rahmanian

Namiranian

et al.

Migraine and type 2 diabetes; is there any association?

J Diabetes Metab Disord 2015; 15: 37–37.

11.

Hashemi Madani

Ismail-Beigi

Khamseh

et al.

Predictive and explanatory factors of cardiovascular disease in people with adequately controlled type 2 diabetes. Eur J Prev Cardiol 2017; 24: 1181–1189.

12.

Sattar

. Revisiting the links between glycaemia, diabetes and cardiovascular disease. Diabetologia 2013; 56: 686–695.

13.

Colloca

Ludman

Bouhassira

et al.

Neuropathic pain. Nat Rev Dis Primers 2017; 3: 17002–17002.

14.

Jacobs

Dussor

. Neurovascular contributions to migraine: Moving beyond vasodilation. Neuroscience 2016; 338: 130–144.

15.

Norwegian National Register of Prescription (NorPD), http://norpd.no/ (2010, accessed 3 January 2017).

16.

National Registry, http://www.skatteetaten.no/en/person/National-Registry/ (2014, accessed 3 January 2017).

17.

Norwegian National Education Database (NNED), https://www.ssb.no/en/omssb/tjenester-og-verktoy/data-til-forskning/utdanning (2014, accessed 3 January 2017).

18.

Watson

Peters Harmel

Matson

. Atherosclerosis in type 2 diabetes mellitus: The role of insulin resistance. J Cardiovasc Pharmacol Ther 2003; 8: 253–260.

19.

Sacco

Altobelli

Ornello

et al.

Insulin resistance in migraineurs: Results from a case-control study. Cephalalgia 2014; 34: 349–356.

20.

Cavestro

Rosatello

Micc

et al.

Insulin metabolism is altered in migraineurs: A new pathogenic mechanism for migraine?

Headache 2007; 47: 1436–1442.

21.

Feldman

Nave

Jensen

et al.

New horizons in diabetic neuropathy: Mechanisms, bioenergetics, and pain. Neuron 2017; 93: 1296–1313.

22.

McCarthy

Hosford

Riley

et al.

Single-nucleotide polymorphism alleles in the insulin receptor gene are associated with typical migraine. Genomics 2001; 78: 135–149.

23.

Diabetesforbundet, https://www.diabetes.no/english/type-2-diabetes/ (2016, accessed 1 June 2017).

Diabetes is associated with decreased migraine risk: A nationwide cohort study

Abstract

Background

Method

Result

Conclusion

Keywords

Introduction

Methods

Data sources

Study population

Exposure

Outcome

Statistical analysis

Ethical approval

Results

Discussion

Public health relevance

Footnotes

Declaration of conflicting interests

Funding

References