Our search identified 604 potentially relevant studies. Of these, 14 studies (15 interventions) were RCTs and met our inclusion criteria. The numbers of participants were 352, 138 and 1745 for aspirin, steroid and NSAIDs groups, respectively. One selected study comprised two separate interventions. Interventions assessed in these studies were grouped into four categories: aspirin (three interventions), steroids (one intervention), traditional NSAIDs (six interventions), and selective cyclooxygenase-2 (COX-2) inhibitors (five interventions). All studies were evaluated for internal validity using a risk of bias assessment tool. The risk of bias was low for five studies, high for seven studies, and unclear for two was no significant improvement in cognitive decline for aspirin, steroid, traditional NSAIDs and selective COX-2 inhibitors. Compared to controls, patients receiving aspirin experienced more bleeding while patients receiving steroid experienced more hyperglycaemia, abnormal lab results and face edema. Patients receiving NSAIDs experienced nausea, vomiting, elevated creatinine, elevated LFT and hypertension. A trend towards higher death rates was observed among patients treated with NSAIDS compared with placebo and this was somewhat higher for selective COX-2 inhibitors than for traditional NSAIDs.
Because non-genomic pathways include any mechanism that is not a genomic effect, there are various non-genomic pathways. However, all of these pathways are mediated by some type of steroid hormone receptor found at the plasma membrane.  Ion channels, transporters, G-protein coupled receptors (GPCR), and membrane fluidity have all been shown to be affected by steroid hormones.  Of these, GPCR linked proteins are the most more information on these proteins and pathways, visit the steroid hormone receptor page.