Smoking and spinal fusion with rhBMP-2: long-term clinical, functional, and occupational outcomes

Attention: open in a new window. PDFPrintE-mail



Smoking impairs bone healing and is associated with pseudoarthrosis and inferior clinical outcomes. In the era of BMP, the efficacy of spinal fusion in smokers has only been studied in one small series, with no available data on clinical outcomes.



This is the first report of long-term clinical, occupational, and functional outcomes comparing smokers to nonsmokers who were fused with posterior fusion and TLIF using BMP.


Study Design

Retrospective review of prospectively collected outcomes


Patient Sample

545 consecutive adults age 61 years (19-88) who underwent spinal fusion for degenerative, spondylolisthesis, or deformity at one center.


Outcome measures

Clinical: VAS pain scores, pain medication records; Functional: Oswestry Disability index (ODI); Occupational: work status pre-op and long-term, work type (sedentary, medium, heavy; retired, student, unknown), time to return to work.



All patients underwent spinal fusion for spondylolisthesis (180 patients, average 3 levels), degenerative disease (193 patents, average 3 levels), or deformity (172 patients, average 9 levels); All were fused with BMP-2, local autograft, and corticocancellous allograft: posterior fusion only- 98 patients, additional TLIF+BMP- 447 patients at average 1.7 levels. Smokers-55, Nonsmokers-490; smokers were not discouraged from having surgery. Prior surgery: smokers-24, nonsmokers-210. Work status preop: working -195 (sedentary-102, medium-61, heavy-9, student-4, unknown-19), retired-276, disabled-73. Excluded: non-fusion surgery, age < 18 years. Clinical and radiographic data were collected preop, 1 year, 2 years, and latest follow-up.



Follow-up averaged 5 years (24-112months). There was no significant difference in complications between smokers and nonsmokers: nonunion-23 (4%), adjacent fracture-16(3%), infection-13 (2%), whether primary or revision surgery. Significant improvement was noted in VAS, Oswestry (ODI), and pain medication records among smokers and nonsmokers for all diagnostic groups, whether primary or revision surgery. SMOKERS: Primary surgery (VAS preop-7.1, 2 year- 4.1, p=0.0026; ODI preop-51, 2 year-32, p=0.0433), Revision surgery (VAS preop-7.3, 2 year-3.7 p=0.0034; ODI preop-56, 2 year-32, =0.0001). NONSMOKERS: Primary surgery (VAS preop- 5.8, 2 year-2.8,p<0.001; ODI preop-44.9, 2 year-24.8, p<0.001), Revision surgery. (VAS preop-6.6, 2 year-4.0 p<0.001; ODI preop-55, 2 year-36, p<0.001). WORK: Returned and remained working long-term: Smokers- 20/24 (83%), Nonsmokers- 123/167(74%) (p=0.450, Fischer's exact test), no difference in time to return. Filed for disability after surgery: 4/24 smokers (17%), 5/167 nonsmokers(3%) (p=0.016, Fischer's exact test).



Smokers and nonsmokers undergoing primary or revision spinal fusion for scoliosis, spondylolisthesis, and degenerative disease showed significant clinical improvement and a strong ability to return to work and remain working long-term in the largest series in the literature. Smokers had slightly more pain pre-op than nonsmokers, and were more likely to claim disability despite similar VAS and ODI improvement. BMP appears to negate the influence of smoking on clinical, functional, and occupational outcomes after instrumented spinal fusion, regardless of fusion length, diagnosis, or revision surgery. Other factors besides clinical and functional outcomes likely account for smokers being more likely to file for disability.