Methods: We conducted a nested case-control study using data from electronic medical records of general practitioners, participating in NIVEL Primary Care Database, to evaluate timing and numbers of visits for 192 symptoms and diseases up to nine years before diagnosis of IA. To this end we used the International Classification of Primary Care (ICPC-1) coding system. 2772 patients who received a new diagnosis of IA between 2012 and 2014 were matched (ratio 1:2) with controls on  age, gender, general practice and retrospective duration of follow-up. The frequency of primary care visits between the IA patients and controls were compared using logistic regression in different time periods before diagnosis.

Results: The consultation rate for musculoskeletal symptoms was increased in IA patients in the last 1.5 years before diagnosis with odds ratios (ORs) of 1.8 (confidence interval; CI: 1.6-2.1, p-value<0.05), 1.4 (CI 1.2-1.6, p<0,05) and 1.3 (CI 1.1-1.5, p<0.05), respectively, at 6, 12 and 18 months before diagnosis. For infections, the consultation rate was significantly higher 6 and 18 months prior to diagnosis (OR=1.2; both CI: 1.1-1.4, p-value<0.05). Finally, for IA-related disease and other chronic diseases a significant difference was observed only 3 months before diagnosis with ORs of 1.2 (CI 1.02-1.3, p<0.05) and 1.3 (CI 1.1-1.5, p<0.05) respectively. All ORs are corrected for age and gender. Important contributors to the above mentioned significance levels were presence of shoulder complaints (16.1% in the IA-patients versus 9.6% in the controls; chi2 73.9, p<0.001), hand/finger complaints syndrome (12.2% versus 5.6%; chi2 112.5, p<0.001), carpal tunnel syndrome (5% versus 2.5%; chi2 37.1, p<0.001) and foot/toe complaints (15.2% versus 9.2%; chi2 67.0, p<0.001).

Conclusion: We found significantly increased consultation rates in general practice for musculoskeletal symptoms and infectious diseases prior to the diagnosis of IA. This diverging trend started 4-6 years before diagnosis, but becomes statistically significant around 1.5 years preceding diagnosis. Possibly, these symptoms can be used to develop methods for earlier detection of IA in general practice.