Background: We determined prospectively the long-term outcomes of nonoperative treatment of chronic patellofemoral pain syndrome.Methods: Of forty-nine patients in a prospective, randomized, double-blind study of unilateral chronic patellofemoral pain syndrome in the knee, forty-five were reexamined seven years after the initial trial of treatment. In the earlier trial, the short-term (six-month) effects of intra-articular injections of glycosaminoglycan polysulfate combined with intensive quadriceps-muscle exercises were compared with those of injections of a placebo combined with exercises and with those of exercises alone. At seven years, the follow-up consisted of standardized subjective, functional, and clinical assessments and muscle-strength measurements as well as magnetic resonance imaging, radiography, and bone-densitometry measurements of the knee.Results: At six months, complete subjective, functional, and clinical recovery had occurred in almost three-fourths of the patients and, with the numbers available for study, neither significant nor clinically important differences among the three initial treatment groups were detected. The subjective and functional parameters showed few changes between six months and seven years; almost three-fourths of the patients still had full subjective and functional recovery at the time of the latest follow-up. However, according to the physician's clinical evaluation, the number of patients who had no symptoms on the patellar compression and apprehension tests decreased over time, from forty-two (93 percent) and forty (89 percent) of forty-five patients at six months to thirty (67 percent) and thirty-one (69 percent) at seven years; these changes were significant (p = 0.002 and p = 0.023, respectively). The number of patients who had crepitation on the patellar compression test increased over time, from twenty-six (58 percent) at six months to thirty-six (80 percent) at seven years (p = 0.021). The physician's overall assessment showed a similar trend, with thirty-four patients (76 percent) having had complete recovery at six months compared with thirty (67 percent) at seven years; however, with the numbers available, this change was not significant (p = 0.420). Magnetic resonance imaging, performed for thirty-seven patients, revealed no abnormalities in twenty-four (65 percent), mild abnormalities in four (11 percent), moderate abnormalities (a 25 to 75 percent decrease in the thickness of the cartilage) in seven (19 percent), and overt patellofemoral osteoarthritis in two (5 percent) at seven years.Conclusions: The seven-year overall outcome was good in approximately two-thirds of the patients. However, the remaining patients still had symptoms or objective signs of a patellofemoral abnormality.