Six databases were searched electronically with no language or time restrictions up to February 2021. These were Medline via PubMed, Web of Science, Scopus, the Cochrane Central Register of Controlled Trials (The Cochrane Library), and Pro- Quest databases for gray literature. References of the included articles and from the previously published systematic reviews (SRs) were also hand searched to identify randomized clinical trials, cohort, case control studies and case series. Articles were excluded when the number of cigarettes smoked per day was not mentioned and when the participants were not clearly divided into at least 2 subgroups; heavy and light smokers. In vitro, animal, finite element analysis studies, case reports and reviews were all considered ineligible. Based on the eligibility criteria study selection was performed by 2 authors, who critically appraised the included articles and extracted data from them. New castle Ottawa was the used critical appraisal tool, where an article with a score of 6 out 9 means high quality. Disagreements between the authors regarding study selection were resolved by the involvement of a third reviewer. Emails were sent to the authors who had articles with missing or limited data regarding the number of smoked cigarettes or marginal bone loss.

In this SRM, participants were divided into 4 categories based on the number of cigarette consumption: non-smoker, light smoker (< 10 cigarettes/day), moderate smoker (10-20 cigarettes/day), and heavy smoker (> 20 cigarettes/day). Meta-analyses were performed for MBL, and heterogeneity was assessed using the I2. Meta-regressions were also done to explore the possible sources of heterogeneity. Besides, subgroup analyses were conducted to assess the impact of 2 radiographic techniques on the results of the MBL. Attempting to discover the effect of a particular study or number of publications on the results, sensitivity analyses were also conducted. Publication bias was assessed using funnel plot and Egger's and Begg's tests.

The primary outcome was peri-implant MBL. The outcome was not clearly defined and was measured based on the reference points and radiographic techniques reported in the articles. Reference points were implant shoulders, implant-abutment connection, implant threads and a point within implant head. The employed radiographic techniques included periapical radiographs, ortho-pantographs and cone beam computed tomography (CBCT). MBL was reported as a mean difference (MD) measured in mm. The mean follow-up periods ranged from 1–5 years in some studies, however other studies had follow-up periods >5years.

Of the 2512 articles retrieved, only 8 articles, published in 2004 to 2020, were finally included in the qualitative and 5 in the quantitative synthesis of this SRM. One was a randomized clinical trial and the remaining 7 were retrospective cohort studies. The total number of placed implants in the included studies ranged from 80-1126 with a follow-up period that ranged from 1-8 years. The age range of the included participants was 44.2-60.1 years. There was a statistically significant difference between MBL in light smokers (<10 cigarettes) and non-smokers, favoring the latter (MD −0.33, 95% CI −0.69-0.03; P =.072; I ²= 91.2%). Statistically significant higher levels of MBL were found in moderate-heavy smokers (> 10 cigarettes/day) when compared to non-smokers (MD −0.58, 95% CI −0.96–−0.19; P=.003; I ²= 91.2%). No comparison was performed between heavy and light smokers.These results were based on studies of high methodological quality, with a score ranging from 6-9. Subgroup analysis for the used radiographic technique revealed a statistically significant risk of MBL in studies that used ortho-pantographs. These significant findings were totally absent in studies that used periapical radiographs. The results of the sensitivity analysis indicated that none of the studies significantly affected the estimated pooled MD. No evidence for publication bias could be detected by the used tests.

Within the limitations of this study, it might be concluded that the risk of MBL is steadily increasing as daily smoking increases. Patients who smoked > 10 cigarettes/day have a higher MBL than those who smoked < 10 cigarettes/day.