Hormonal contraceptive use and risk of glioma among younger women: a nationwide case–control study

Abstract

AIM

Oral contraceptive use influences the risk for certain cancers. However, few studies have examined any link with risk of central nervous system tumours. We investigated the association between hormonal contraceptive use and glioma risk among premenopausal women in a population-based setting.

METHODS

Using national administrative and health registries in Denmark to conduct a nationwide case–control study, we identified all women ages 15 to 49 years with a first time diagnosis of histologically verified glioma between 2000 and 2009. Each case was age-matched to eight population controls using risk set sampling. Based on prescription data, exposure until 2 years prior to the index date was categorized according to hormonal contraceptive type, i.e. combined oestrogen-progestagen or progestagen only, and duration of use (<1, 1 to <5, ≥5 years). We used conditional logistic regression to compute odds ratios (ORs) with 95% confidence intervals (CIs) for glioma associated with hormonal contraceptive use, adjusting for potential confounders.

RESULTS

We identified 317 cases and 2126 controls. Ever use of hormonal contraceptive was associated with an OR of 1.5 (95% CI 1.2, 2.0) and the OR increased with duration of use (long term, ≥5 years: OR 1.9; 95% CI 1.2, 2.9). The association between long term hormonal contraceptive use and glioma risk was most pronounced for progestagen only therapy (OR 2.4; 95% CI 1.1, 5.1), especially when this regimen constituted the sole hormonal contraceptive therapy (OR 4.1; 95% CI 0.8, 20.8).

CONCLUSION

Long term hormonal contraceptive use may increase the risk of glioma.

WHAT IS ALREADY KNOWN ABOUT THIS SUBJECT

• Studies of glioma risk associated with use of hormonal contraceptives are limited by their focus on post-menopausal women and, typically, self-reported drug information.

• Prospectively collected data on the association between hormonal contraceptive use and glioma risk among younger women are scarce.

WHAT THIS STUDY ADDS

• In this Danish registry-based study, we found that ever use of hormonal contraceptives was associated with a moderately increased risk of glioma among younger women.

• A nearly two-fold increased risk of glioma was observed among long term users of hormonal contraceptives.

Introduction

Gliomas are the most common type of malignant brain tumours in the Nordic countries [1,2]. Glioma incidence is lower among women than among men [1,2], which has led to the suggestion that female sex hormones or reproductive factors may protect against glioma development [3,4]. Epidemiological studies have reported either no association or a weak inverse association between contraceptive use and glioma risk [5–13], but these findings are limited by their focus on post-menopausal women only [10,12] or by relatively small subsamples of premenopausal women [5–9,11,13].

Use of combined oestrogen-progestagen contraceptives slightly increases the risk of in situ and invasive cancer of the uterine cervix, hepatocellular carcinoma and breast cancer [14], but the excess risk of breast cancer pertains only to young women with current or recent contraceptive use [14]. In addition the link between hormone replacement therapy (HRT) and breast cancer diminishes markedly within relatively few years following cessation of the therapy [15]. The potential for this pattern to exist for glioma and female sex hormone influences suggests a need to focus on premenopausal women who are, or have recently been, exposed to oral contraceptives.

Previous studies of oral contraceptive use and glioma risk have provided only sparse and inconclusive information on associations according to contraceptive type and length of contraceptive use [3,4]. The lack of clarity may derive partly from the limitation of self-reporting, raising methodological concerns, which may be further accentuated among patients with glioma due to impaired cognitive skills [16].

To address these gaps, we conducted a large nationwide register-based case–control study in Denmark to investigate the association between contraceptive use and glioma among premenopausal women.

Methods

Setting

Our case–control study was based on the following Danish population-based administrative and health registries: the Danish Cancer Registry (Cancer Registry) [17,18], the Danish Civil Registration System [19], the Danish National Prescription Registry (Prescription Registry) [20], the Danish National Patient Register (Patient Registry) [21] and Statistics Denmark. Complete cancer, hospital and prescription histories could be obtained and linked at the individual level through the unique civil registry number assigned to all Danish citizens since 1968 [19].

Case ascertainment

The Cancer Registry has recorded incident cases of cancer on a nationwide basis since 1943 and offers accurate and almost complete ascertainment [17,18]. Cancer diagnoses are recorded according to the International Classification of Diseases (ICD), version 10 (ICD-10) and the ICD for Oncology (ICD-O-3) for topography and morphology codes. Reporting of glioma is mandatory regardless of tumour evolution.

Eligible cases were women aged 15–49 years with a histologically verified first diagnosis of brain glioma during 2000–2009, ascertained from the ICD-10 and ICD-O-3 morphology codes (see Appendix). We required cases to have no previous history of cancer (including central nervous system tumours and excluding non-melanoma skin cancer). The date of diagnosis was defined as the index date.

Selection of population controls

Using risk set sampling [22] and applying the same selection criteria as for cases, we sampled eight birth year-matched controls from the Danish female population for each case. The index dates for controls were identical to the dates of diagnosis of the corresponding cases.

The Danish Civil Registration System contains continuously updated information on vital status and migration. We used this information to restrict cases and controls to individuals who had resided in Denmark for at least 10 consecutive years before the index date. This restriction was imposed after the sampling of controls. Therefore the final ratio of cases to controls deviated slightly from 1:8.

Contraceptive exposure assessment

The Prescription Registry contains information on all prescriptions for drugs, including hormonal contraceptives, dispensed at Danish pharmacies since 1995 [20]. The data include the date of dispensing and a full product description, including the anatomical therapeutic code [23], size and number of dose units, dosage form (e.g. tablets) and the total number of defined daily doses (DDD) in the dispensed package. The DDD is the assumed average maintenance dose per day for a drug used for its main indication in adults [23]. From this registry, we obtained information on redeemed prescriptions of hormonal contraception from 1995–2009 and assessed exposure during the period from 1995 until 2 years prior to index date. Prescriptions within 2 years prior to the index date were disregarded to minimize the impact of potential differential diagnostic intensity according to use of hormonal contraceptives and to ignore exposure unlikely to have influenced the development of glioma cases. We defined ever use of hormonal contraceptives as ≥two prescriptions and non-use as <two prescriptions with the exception of hormonal intrauterine devices (IUD), for which ≥one prescription was defined as ever use. We further classified ever use into current/recent use (≥two prescriptions overall and ≥one prescription 2 to <5 years before index date) and past use (≥two prescriptions overall but no recorded prescriptions during the period 2 to <5 years before index date), and by type of hormonal contraceptive (combined oestrogen-progestagen, progestagen only or mixed use; >one type of contraceptive recorded). Furthermore, we classified hormonal contraceptives by usage form into tablets and IUD. The remaining types of hormonal contraceptives (vaginal ring n = 8; implant n = 7; injection n = 4; transdermal n = 1) were disregarded in usage form analyses. We also categorized hormonal contraceptive use according to cumulative duration of use. The coverage of each prescription was calculated as the sum of DDDs for the prescription, with one DDD set to be equivalent to 1 day of use, plus a grace period of 120 days (1 year for IUD) to allow some degree of non-adherence. For women with more than one treatment period, the duration of the periods was added. Finally, the cumulative duration of hormonal contraceptives use was categorized as <1 year, 1 to <5 years or ≥5 years.

Potential confounders

We used length of schooling according to annually updated information from Statistics Denmark [24] as a proxy for socioeconomic status. The number of years of schooling was categorized into three strata (7–10, 11–12, 13+ years). Some studies have suggested that parity is associated with risk of glioma [3]. We therefore obtained information on fertility status from the nationwide Danish Fertility Database [25], and women were classified according to number of live births: 0 (nullipara), 1, 2, 3+ or ‘missing information’. From the Patient Registry, which holds data on all non-psychiatric hospitalizations and outpatient visits in Denmark [21], we obtained information on medical history of allergy or asthma. Allergies are inversely associated with glioma risk [26]. Finally, we assessed use of potentially confounding drugs, including HRT, antidepressants, antihistamines and anti-asthma medications, and non-aspirin non-steroidal anti-inflammatory drugs (NSAIDs). Exposure to a potential confounder drug was defined as ≥two prescriptions. An exception was non-aspirin NSAIDs, for which use was defined as ≥10 prescriptions to capture chronic use.

Statistical analysis

We computed the proportions of cases and controls in the categories for each study variable. Conditional logistic regression was used to estimate odds ratios (ORs) with 95% confidence intervals (CIs) for glioma associated with the use of hormonal contraceptives. We calculated both crude ORs, adjusted by design for age and index year, and multivariable ORs, further adjusted for years of schooling, and history of allergy or asthma. We also calculated ORs with additional adjustment for use of antihistamines, anti-asthma medication, HRT, antidepressants and non-aspirin NSAIDs. However, all of these covariates had little influence (<5% individually and combined) on the ORs and were not included in the final model. Stratified analyses were specified by the duration and type and form of hormonal contraceptive use.

In supplemental analyses we (i) restricted cases to glioblastoma multiforme, (ii) increased the exposure lag time to 5 years, (iii) reduced the length of the grace period (30 days for tablets, 0 days for IUD), (iv) restricted the study population to women aged 15–44 years (to decrease the likelihood of menopause) and (v) restricted the study population to women identified in 2002–2009, yielding a minimum of 7 years of prescription history. In addition, we stratified the analyses according to education level (years of schooling; women aged 35–49 years included).

All analyses were performed using Stata SE 12.1 (StataCorp, College Station, TX, USA). The study was approved by the Danish Data Protection Agency and the Danish Medicines Agency.

Results

Our study population consisted of 317 cases and 2126 controls, which were similar for distribution of age, parity and years of schooling (Table 1). Compared with controls, cases had a lower prevalence of history of asthma or allergy and use of anti-asthma and antidepressant drugs, and slightly higher prevalence of antihistamine, non-aspirin NSAID, and HRT drug use.

Table 1.

Characteristics of women aged 15–49 years with glioma and their matched controls, 2000–2009. Numbers (%)

Characteristics	Cases (n = 317)	Controls (n = 2126)
Age (years)
15–34	114 (36.0)	751 (35.3)
35–44	115 (36.3)	782 (36.8)
45–49	88 (27.8)	593 (27.9)
Calendar year
2000–2004	164 (51.7)	1107 (52.7)
2005–2009	153 (48.3)	1019 (47.9)
Parity, number of childrena
0	93 (29.3)	611 (28.7)
1	54 (17.0)	366 (17.2)
2	112 (35.3)	743 (34.9)
3+	48 (15.1)	344 (16.2)
Missing	10 (3.2)	62 (2.9)
Schooling, number of years
7–10	82 (25.9)	587 (27.6)
11–12	139 (43.8)	872 (41.0)
13+	90 (28.4)	616 (29.0)
Missing	6 (1.9)	51 (2.4)
Sex hormone use
Contraceptivesc	186 (58.7)	1065 (50.1)
HRT	9 (2.8)	52 (2.4)
Comorbidity indicators
Allergy and asthma hospital contactsc	9 (2.8)	86 (4.0)
Anti-asthma drugsd	47 (14.8)	358 (16.8)
Antihistaminesd	42 (13.2)	261 (12.3)
Antidepressantsd	28 (8.8)	218 (10.3)
NA-NSAIDd,e	16 (5.0)	86 (4.0)

^aNo data available on births in 2009.

^bAccording to Patient Registry.

^cDenotes variables with P < 0.05 (chi-squared test).

^dAccording to Presciption Registry.

^eNon-aspirin non-steroidal anti-inflammatory drugs.

Overall, 58.7% of cases and 50.1% of controls were ever users of contraceptives (P = 0.004) (Table 1), yielding an OR of 1.5 (95% CI 1.2, 2.0) for glioma associated with ever use of hormonal contraceptives. The OR was higher for current/recent use (OR 1.7; 95% CI 1.3, 2.4) than for past use (OR 1.2; 95% CI 0.8, 2.0) (Table 2). In analyses defining hormonal contraceptive use according to type and form, the OR for glioma was elevated with use of combined oestrogen and progestagen (OR 1.4; 95% CI 1.0, 1.8) and highest for progestagen only (OR 2.8; 95% CI 1.6, 5.1). Mixed use of both types of contraceptives was also associated with an increased OR (OR 1.5; 95% CI 0.8, 3.0).

Table 2.

Risk of glioma by recency, type and form of hormonal contraceptive use by women aged 15–49 years

Contraceptive use	Cases	Controls	Crude odds ratioa (95% CI)	Adjusted odds ratiob (95% CI)
Non-use	131	1061	1 (reference)	1 (reference)
Ever use	186	1065	1.5 (1.2, 2.0)	1.5 (1.2, 2.0)
Current/recentc	147	791	1.7 (1.3, 2.3)	1.7 (1.3, 2.4)
Pastd	39	274	1.2 (0.8, 2.0)	1.2 (0.8, 2.0)
Type and form
Oestrogen and progestagen onlye	145	895	1.4 (1.0, 1.8)	1.4 (1.0, 1.8)
Tablet onlyf	142	890	1.4 (1.0, 1.8)	1.4 (1.0, 1.8)
Progestagen onlye	21	76	2.9 (1.6, 5.2)	2.8 (1.6, 5.1)
Tablet onlyf	14	34	3.3 (1.6, 6.9)	3.2 (1.5, 6.8)
IUDg onlyf	6	38	2.4 (0.8, 6.8)	2.4 (0.8, 6.8)
Mixed useh	20	94	1.6 (0.8, 3.0)	1.5 (0.8, 3.0)

^aAdjusted for age and calendar year by design.

^bAdditionally adjusted for schooling and hospital contacts for allergy and asthma.

^cLast prescription for contraceptive dispensed <5 years prior to index date.

^dLast prescription for contraceptive dispensed ≥5 years prior to index date.

^eOnly use of a single type of contraceptive recorded in Prescription Registry.

^fOnly use of a single form of contraceptive recorded in Prescription Registry.

^gIntrauterine device.

^hMore than one type of contraceptive recorded in Prescription Registry.

ORs for glioma increased with duration of use of contraceptives of any type from 1.4 (95% CI 0.8, 2.3) with use <1 year to 1.9 (95% CI 1.2, 2.9) with long term use (≥5 years) (Table 3). A similar risk pattern was observed for combined oestrogen-progestagen (ever use ≥5 years: OR 1.8; 95% CI 1.2, 2.9; exclusive use ≥5 years: OR 1.7; 95% CI 1.1, 2.8). Increasing ORs with duration of hormonal contraceptive use were also seen with progestagen only use, although only with exclusive use of this agent. Tests for trend were statistically non-significant in all strata.

Table 3.

Risk of glioma by duration of hormonal contraceptive use by women aged 15–49 years

Hormonal contraceptive use (HC)	Cases	Controls	Adjusted odds ratioa (95% CI)	P for trend
Non-use	131	1061	1 (reference)
Ever use
One or more types of HCb, years
Any type of HC
<1	27	143	1.4 (0.8, 2.3)
≥1, <5	96	563	1.5 (1.1, 2.2)
≥5	63	359	1.9 (1.2, 2.9)	0.6
Oestrogen and progestagen
<1	19	118	1.1 (0.6, 2.1)
≥1, <5	86	525	1.4 (1.0, 2.1)
≥5	60	346	1.8 (1.2, 2.9)	0.7
Progestagen only
<1	8	28	2.5 (1.0, 6.2)
≥1, <5	18	87	1.8 (0.9, 3.4)
≥5	15	55	2.4 (1.1, 5.1)	0.2
Single type of HCc, years
Oestrogen and progestagen
<1	19	115	1.2 (0.6, 2.2)
≥1, <5	78	476	1.5 (1.0, 2.1)
≥5	48	304	1.7 (1.1, 2.8)	0.7
Progestagen only
<1	8	25	2.7 (1.1, 6.8)
≥1, <5	10	38	2.9 (1.2, 6.9)
≥5	3	13	4.1 (0.8, 20.8)	1.0

^aAdjusted for age and calendar year by design; additionally adjusted for years of schooling and hospital contacts for allergy and asthma.

^bWomen with dispensed prescriptions for one or more types of contraceptives in 1995 to index date.

^cWomen with dispensed prescriptions for only a single type of contraceptive in 1995 to index date.

In analyses for specific histological types of glioma, the highest ORs were observed for glioblastoma multiforme (ever use: OR 1.5; 95% CI 0.9, 2.3; ≥5 years: OR 2.5; 95% CI 1.3, 4.9; P for trend = 0.3) (Table 4). Elevated ORs with long term (≥5 years) use of hormonal contraceptives were also seen for astrocytoma but not for oligodendroglioma.

Table 4.

Risk of glioma associated with hormonal contraceptive use by women ages 15–49 years stratified according to histological type

Contraceptive use	Cases	Controls	Adjusted odds ratioa (95% CI)	P for trend
Glioblastoma multiforme
Non-use	48	377	1 (reference)
Ever use	55	313	1.5 (0.9, 2.3)
Duration of use (years)
<1	7	51	1.0 (0.4, 2.4)
≥1, <5	26	173	1.3 (0.7, 2.4)
≥5	22	89	2.5 (1.3, 4.9)	0.3
Astrocytoma grade II and III
Non-use	35	279	1 (reference)
Ever use	59	353	1.4 (0.8, 2.4)
Duration of use (years)
<1	10	36	2.0 (0.7, 5.6)
≥1, <5	30	181	1.4 (0.7, 2.6)
≥5	19	136	1.6 (0.7, 3.6)	0.2
Oligodendroglioma
Non-use	25	208	1 (reference)
Ever use	41	226	1.5 (0.8, 2.8)
Duration of use (years)
<1	4	28	1.1 (0.3, 4.6)
≥1, <5	23	128	2.0 (1.0, 4.1)
≥5	14	70	1.1 (0.4, 3.0)	0.4
Variousb
Non-use	23	197	1 (reference)
Ever use	31	173	1.8 (0.9, 3.8)
Duration of use (years)
<1	6	28	1.9 (0.6, 5.8)
≥1, <5	17	81	1.9 (0.8, 4.4)
5+	8	64	2.6 (0.7, 9.9)	0.1

^aAdjusted for age and calendar year; additionally adjusted for years of schooling and hospital contacts for allergy and asthma.

^bPilocytic astrocytoma (n = 17), ependymoma (9), mixed glioma (7), glioma malignant (4), choroid plexus papilloma (4), dysembryoplastic neuroepithelial tumour (4), gliosarcoma (3), myxopapillary ependymoma (2), subependymal glioma (2), fibrillary astrocytoma (1), and gliomatos.

The majority of supplementary analyses yielded estimates similar to those of the main analyses, with two exceptions. First, increasing the lag time to 5 years reduced the risk of glioma associated with ever use of hormonal contraceptives (OR 1.3; 95% CI 1.0, 1.8). Second, whereas only little variation in use of hormonal contraceptives was observed by education level (low education level 40.3%; middle 44.3%; high 39.6%), ORs for glioma associated with hormonal contraceptive use varied according to education level (low: OR 0.8; 95% CI 0.4, 1.6; middle: OR 2.0; 95% CI 1.2, 3.1; high: OR 1.6; 95% CI 0.9, 2.8). However, these analyses were based on small numbers.

Discussion

Long term use of hormonal contraceptives was associated with an increased risk of glioma that increased with duration of use. The risk estimates were particularly elevated for use of progestagen only contraceptives. In analyses according to histological type of glioma, slightly higher risk estimates were observed for glioblastoma multiforme.

Previous case–control [5–7,9,11] and cohort studies [8,10,12,13] have reported either no association or a weak inverse association between ever use of hormonal contraceptives and glioma risk, without any apparent duration–risk pattern. In addition, previous results addressing recency of hormonal contraceptive use have been inconsistent, with two studies indicating a slightly increased risk [7,13] and one reporting a substantially reduced risk of glioma associated with current use [11]. We found that recent use of hormonal contraceptives was associated with an increased risk of glioma, whereas the risk estimate arising from former use was elevated only slightly. The majority of women in previous studies were over 50 years, and thus only former contraception was captured, which might explain the finding of no association. Moreover, most studies that included younger women did not stratify by age, except for Hatch et al., who reported an association of ever use of hormonal contraceptives with a higher risk of glioma among women under age 50 years (OR 1.25; 95% CI 0.67, 2.34) and an increased risk with duration of use (≥10 years: OR 1.70; 95% CI 0.73, 3.94). They found an inverse association, however, among women aged 50 years or over (ever use: OR 0.37, 95% CI 0.20, 0.69; 10+ years: OR 0.37, 95% CI 0.11, 0.22) [7]. Our results are thus in line with the only previous study that specifically addressed the risk of glioma in women younger than 50 years.

We also evaluated the effect of hormonal contraceptive type on glioma risk, which has not been reported previously to our knowledge. Although we observed elevated risk estimates for combined contraceptive or progestagen only, the estimate associated with the latter type was clearly higher for both ever and exclusive use. Our data indicate that particularly progestagen exposure may confer an increased risk of glioma. Progesterone increases proliferation of high grade astrocytoma cell lines [27,28] as well as growth factor levels [29], which is in line with increased progesterone receptor protein or mRNA with glioma grade [30,31]. In contrast, studies on the influence of oestrogen or inhibition of oestrogen on glioma cell growth or apoptosis in vitro have shown differential effects [32–34]. However, limited statistical precision of our results hindered any robust conclusion regarding the influence of progestagen only use on glioma risk. Furthermore, our finding of an increased risk of glioma associated with IUD use may indicate an impact of unmeasured confounding because IUD therapy is associated with very low systemic concentrations of progesterone [35]. A potential unmeasured confounder could be obesity because of preferential prescribing of progestagen only contraceptives to women who are overweight and/or at increased risk of arterial or venous thrombosis, as recommended in the Danish drug manual [36] and by international health authorities [37,38]. Although the observed association may be non-causal, IUDs have been associated with increased breast cancer among post-menopausal women, which might conversely indicate that even a low but constant level of progestagen exposure could have adverse health effects [39].

Our study strengths include reliance on a nationwide prescription registry, which provided detailed long term drug use histories and eliminated recall bias. Our use of other nationwide registries with virtually complete coverage and continuously updated data on demographics, hospital contacts and cancer outcomes minimized selection bias. The detailed information from the Cancer Registry allowed us to evaluate the association of hormonal contraceptive use by histological type of glioma.

Our study also had some potential limitations. First, our results could be biased if contraceptive users were more likely to undergo brain imaging because of socioeconomic factors. However, the Danish National Health Service provides equal tax-supported access to health care to Danish citizens, and our study thus was performed in a setting with free access to health services independent of income. A previous Danish study reported no major association between socioeconomic status and incidence of central nervous system tumours [40]. Second, we cannot exclude that women using hormonal contraceptives may seek medical consultation more often than women in general, leading to greater possibility of brain tumour detection in this population (surveillance bias). Making this less probable is that the excess risk associated with long term hormonal contraceptive use was highest for glioblastoma multiforme, which is highly unlikely to remain undetected because of its rapid and aggressive course. Third, the Prescription Registry was established in 1995 and contraceptive use prior to 1995 was misclassified as non-use. However, such non-differential misclassification would lead to attenuation of risk estimates. Fourth, brain tumours could conceivably influence adherence to medication use prior to diagnosis by affecting cognitive skills, for example, but the effect would have been a reduction in risk estimates. Finally, except for parity, we could not adjust for reproductive factors such as age of menarche, or for anthropometric measures, such as body mass index (BMI), because the registries do not contain this information. Although BMI was not associated with glioma risk in a recent prospective study from Norway [41], others have found an association between obesity and glioma risk [10]. Some women with high BMI may use progestagen only because of the relative contraindication for hormonal contraceptives in this population, but women using combined hormonal contraceptives have on average lower BMIs than non-users [13,42]. We cannot exclude a confounding effect of body composition with regard to progestagen only use. However, we find it less likely that such anthropometric measures influenced our results for combined contraceptive use.

In conclusion, long term use of hormonal contraceptives was associated with an increased risk of glioma among younger women. Considering the extensive use of hormonal contraceptives our finding merits further investigation. If the finding is confirmed, an important next step will be evaluation of the influence of previous hormonal contraceptive use on the prognosis of glioma.

Appendix

Codes used in the analyses

Cancer codes

Glioma

ICD-10

C71.0-C71.9, D33.0-D33.2, D43.0-D43.2

ICD-O-3 morphology

94403—glioblastoma multiforme

94003, 94013, 94103, 94113—astrocytoma grade II and III

94503, 94513, 94603—oligodendroglioma grade II and III

93801,93803, 93813, 93823, 93831,93900–94001, 94121–94401,

94413–94501—‘other’

ATC codes

Contraceptives

G03AA, G03AB, G03HB01, G02BB – Oestrogen and progestagen combined

G03AC, G02BA03, G03DA02 – Progestagen-only

Post-menopausal replacement therapy

G03C, G03DC, G03F

Non-aspirin non-steroidal anti-inflammatory drugs

M01A (including Cox2 inhibitors), excluding M01AX

Antidiabetics

A10A – insulin

A10B – oral antidiabetics

Allergy and asthma

R06A – antihistamines

R03 – anti-asthma drugs

Psycholeptics

N06 – antidepressants

Hospital discharge codes

Allergy and asthma

ICD-8: 493, 507, 708, 69200, 69202, 69210, 69212, 69220, 69222, 69230, 69232, 69260, 69262

ICD-10: J45, J46, J30, L50, L23

*Preparations without official DDD value from WHO's index of January 1 2012 and without fixed DDD-values from WHO's list for combined products were assigned values (national DDK) values available at: http://www.ssi.dk/Sundhedsdataogit/Dataformidling/Laegemiddelstatistikker/Leksikon/∼/media/Indhold/DK%20-%20dansk/Sundhedsdata%20og%20it/NSF/Dataformidling/Leksikon/Nationalt%20tildelte%20DDK-vaerdier.ashx

Competing Interests

All authors have completed the Unified Competing Interest form at http://www.icmje.org/coi_disclosure.pdf (available on request from the corresponding author) and LA, SF, BWK and DG declare no support from any organization for the submitted work, no financial relationships with any organizations that might have an interest in the submitted work in the previous 3 years and no other relationships or activities that could appear to have influenced the submitted work. JH declares no support from any organization for the submitted work. JH has participated in research with funding from MSD, Pfizer and Takeda paid to the institution where he is employed. JH has personally received travel reimbursement from Pfizer. PR declares no support from any organization for the submitted work. PR has been an investigator in a clinical trial supported by MSD (HPV studies). PR has personally received honoraria from Bayer-Schering, Nycomed, MSD, Organon and Pharmakon and is a member of the Advisory Board for Amgen (osteoporosis).

Funding

This study was supported by grants from the Danish Cancer Society (grant nr. R56-A2879), Odense University Hospital and the University of Southern Denmark. The funding sources had no role in the design, analysis and interpretation of the results. Thus, the authors were independent from the funding sources.