Purpose: To evaluate the influence of cigarette smoking on the survival of dental implants with a retrospective observational study of 5 years. Materials and Methods: A total of 1727 consecutively treated patients at four private practices were divided into non-smokers (NS group, 1178 patients) and smokers (S group; 549 patients) according to what they declared prior to implant placement. Non-smokers received 4460 implants and 2583 implant-supported prostheses, whereas smokers received 2260 implants and 1292 implantsupported prostheses. Various implant systems and procedures were used. Outcome measures were prosthesis and implant survival. Results: Over the 5 years after loading, 159 (17%) non-smokers and 91 (13%) smokers were lost to follow-up; 20 (0.9%) prostheses could not be placed or failed in 15 non-smokers and 12 prostheses (1.2%) could not be placed or failed in 12 smokers. One hundred and twelve (2.9%) implants failed in 105 non-smokers and 107 (5.5%) implants failed in 75 smokers. Most of the implant failures (90%) occurred before implant loading. Fitting a logistic regression for early implant failures and total implant failures, taking into account the clustering of implants in patients, there were no statistically significant differences for prosthesis failures (P value not calculated as too few failures) and early implant failures between the two groups (P = 0.13). However, when considering all implant failures up to 5 years after loading, significantly more failures (5.5%) occurred in smokers compared with non-smokers (2.9%) (OR 1.72; 95% CI 1.20 to 2.50; P = 0.003). Conclusions: Due to the retrospective nature of this study, conclusions have to be interpreted with caution. Five years after loading, smokers experienced almost twice as many implant failures compared with non-smokers. Non-statistically significant trends in favour of non-smokers were observed for early implant failures and prosthesis failures.