Comparative study. To compare the radiographic outcome of patients undergoing long spinal deformity surgery to the sacrum/ilium, using either rhBMP-2 without iliac crest bone graft (ICBG) or ICBG without rhBMP-2. rhBMP-2 has been shown to be more effective in promoting successful bone union in patients undergoing single level lumbar spinal fusion than ICBG. However, to the best of our knowledge, there are no studies that compare the efficacy of rhBMP-2 versus ICBG in long spinal deformity surgery. To obtain uniform background, we selected patients with adult spinal deformity who underwent primary spinal fusion from the thoracic spine to the sacrum/ilium and had a minimum 2-year follow-up. Fifty-five consecutive patients, consisting of 32 patients who underwent a fusion using ICBG without rhBMP-2 (ICBG group) and 23 patients who underwent a fusion using rhBMP-2 without ICBG (BMP group) were analyzed. The 2 groups were similar with respect to age, gender, smoking history, comorbidity, and body mass index. The average number of vertebrae fused (11.3 in both groups) and the degree of preoperative deformity (major Cobb angle 58.3 degrees in ICBG group vs. 54.2 degrees in BMP group) were also similar in both groups. All but 2 patients had both anterior and posterior surgery. Both groups were similar in terms of final deformity correction. The average total amount of rhBMP-2 used in the BMP group was 119.2 mg (anterior 11.6 mg/level; posterior 10.0 mg/level). Of the 32 patients in the ICBG group, 9 patients (28.1%) developed a pseudarthrosis, while only 1 of 23 patients (4.3%) in the BMP group developed a pseudarthrosis with the caveat that the follow-up period was shorter in the BMP group (average follow-up of 4.9 vs. 2.7 years). The pseudarthrosis rate in the BMP group compares favorably to pseudarthrosis rate in ICBG group, suggesting that the use of rhBMP-2 without iliac harvesting leads to a competitive fusion rate in long adult spinal deformity surgery, while avoiding ICBG harvest site morbidity.