Background and purpose
To improve diagnoses of primary brain tumours, knowledge about early indicators is needed. Nationwide Danish health registries were used to conduct a population‐based case–control study including all persons diagnosed with a primary brain tumour between 2005 and 2014 in Denmark.

Methods
All 5135 adults diagnosed with a primary brain tumour in the Danish Cancer Registry were matched to 19 572 general population comparisons from the Danish Civil Registration System. Conditional logistic regression analyses were applied to estimate age‐ and multivariable‐adjusted odds ratios (ORs) for the occurrence of a primary brain tumour up to 10 years after hospital diagnoses or prescription of medications related to nervous system diseases and mental and behavioural disorders.

Results
Increased odds for primary brain tumour after nervous system diseases and mental and behavioural disorders manifested up to 10 years before tumour diagnosis were found. Increased odds were seen especially for hospital contacts for inflammatory nervous system diseases [OR 11.3; 95% confidence interval (CI) 6.5–19.7], epilepsy (OR 9.0; 95% CI 7.6–10.7) and antiepileptic medications (OR 3.6; 95% CI 3.2–4.0), whilst antidementia medications provided a strong, protective association for primary brain tumours (OR 0.5; 95% CI 0.3–0.8).

Conclusions
Sub‐groups of patients diagnosed with or being prescribed certain medications targeting nervous system diseases and mental and behavioural disorders may be at increased risk of being diagnosed with a primary brain tumour. Further studies should disentangle the potential underlying common pathogenetic pathways. The results are important for the development of systematic clinical approaches to ensure early diagnosis of primary brain tumours.