Sixty-three publications were identified, of which 18 met the selection criteria for the review (one of which reported on 2 different studies which were then reviewed separately). Five studies compared NSAIDs to placebo or to other treatment modalities, while 14 compared two types of NSAIDs.
Only 5 of the 19 trials received a methods score of over 50 points (50%). The design or conduct of the trials was often inadequate with respect to loss to follow-up, sample size and analysis according to intention-to-treat.
The 5 trials comparing NSAIDs to placebo or to other treatment modalities demonstrated variations in success rates with NSAIDs from 20% to 86%. However, the studies varied considerably regarding type of NSAID, dosage, diagnosis or indication, duration of complaints and exact definition of success. 4 of the studies scored 50 points or more, of which 3 demonstrated superior short-term efficacy of NSAIDs. 2 out of 4 studies comparing NSAIDs with other treatment modalities found corticosteroid injections to be more effective than NSAIDs.
Success rates in the 14 studies comparing two types of NSAIDs had success rates ranging from 10% to 79% in the intervention groups. This may be due to similar variations among studies to those discussed above. Few differences between the various types of NSAIDs were demonstrated. Only 3 of the studies had sufficient power to demonstrate a clinical difference in success rates between treatments of 25%.
Adverse reactions were reported by 8-76% of patients receiving NSAIDs (mainly gastrointestinal or sometimes skin rash, dizziness or headache). These reactions were usually considered to be moderate or mild in nature and resulted in withdrawal from treatment in less than 10% of patients in most trials.