To examine whether prescription patterns complied with recommendations not to use non-steroidal anti-inflammatory drugs (NSAIDs) in patients with cardiovascular contraindications. Moreover, we examined predictors for initiation and prescriber responsibility. We used Danish medical databases to identify all patients with first-time cardiovascular disease during 1996-2017 (n = 628 834). We assessed standardized prevalence proportions, predictors from logistic regression, and prescriber identifiers. One-year prevalence of NSAID initiation increased 3.4% from 1996 (19.4%) to 2001 (22.7%) and declined by 2.7% thereafter until 2017 (13.5%). Trends were independent of age, sex, and disease subtype, although larger annual declines occurred for heart failure (3.9%) and ischaemic heart disease (3.5%) since 2002. One-year prevalence remained highest among patients with venous thromboembolism (16.6%) and angina (13.8%), and lowest for ST-segment elevation myocardial infarction (7.0%) and heart failure (8.8%). Initiators were predominantly prescribed ibuprofen (59%), diclofenac (23%), and etodolac (6%). Diclofenac and coxib use declined, while ibuprofen and naproxen use increased. Median prescribed pill dose of ibuprofen declined after 2008 from moderate/high (600 mg) to low (400 mg). Treatment duration declined for all NSAIDs, except celecoxib. Rheumatic, obesity, and pain-related conditions predicted NSAID initiation. General practitioners issued 86-91% of all NSAID prescriptions, followed by hospital prescribers (7.3-12%). Initiation of NSAIDs in patients with cardiovascular disease declined since 2002. Shorter treatment duration, declining COX-2 inhibition, and increasing use of naproxen and low-dose ibuprofen suggest adherence to guidelines when NSAIDs cannot be avoided. Still, NSAID use remained prevalent despite cardiovascular contraindications, warranting awareness of appropriateness of use among general practitioners in particular.