The Effect of Tobacco Smoking on Musculoskeletal Health: A Systematic Review

This systematic review explored associations between smoking and health outcomes involving the musculoskeletal system. AMSTAR criteria were followed. A comprehensive search of PubMed, Web of Science, and Science Direct returned 243 articles meeting inclusion criteria. A majority of studies found smoking has negative effects on the musculoskeletal system. In research on bones, smoking was associated with lower BMD, increased fracture risk, periodontitis, alveolar bone loss, and dental implant failure. In research on joints, smoking was associated with increased joint disease activity, poor functional outcomes, and poor therapeutic response. There was also evidence of adverse effects on muscles, tendons, cartilage, and ligaments. There were few studies on the musculoskeletal health outcomes of secondhand smoke, smoking cessation, or other modes of smoking, such as waterpipes or electronic cigarettes. This review found evidence that suggests tobacco smoking has negative effects on the health outcomes of the musculoskeletal system. There is a need for further research to understand mechanisms of action for the effects of smoking on the musculoskeletal system and to increase awareness of healthcare providers and community members of the adverse effects of smoking on the musculoskeletal system.


Introduction
Tobacco smoke has more than 7,000 harmful chemical compounds that enter a human body either directly through smoking, indirectly through secondhand exposure to smoke exhaled by a smoker, or through downstream smoke released from a cigarette or pipe [1]. Both smokers and nonsmokers are at risk of exposure to the compounds of smoked tobacco that accumulate on the surfaces in a poorly ventilated environment; this method of exposure is known as thirdhand smoke exposure [2]. In the United States, there are approximately 500,000 annual deaths causally related to smoking and secondhand exposure to smoke [3].
Tobacco smoking has known adverse consequences on most human body systems. Researchers have focused more attention on the deleterious effects of smoking for high mortality diseases, such as cancer and diseases of the cardiovascular and respiratory systems, with less research attention on other body systems, such as the musculoskeletal system [3]. The musculoskeletal system is one of the largest human body systems, comprised of bones, joints, muscles, cartilage, tendons, ligaments, and other connective tissues [4]. An intact and functioning musculoskeletal (locomotor) system is necessary to perform activities of daily living and maintain quality of life [5,6]. Several studies have investigated the association between smoking and musculoskeletal disorders. According to the recent Surgeon General report, the causal relationship between tobacco smoking and rheumatoid arthritis, periodontitis, and hip fractures has been confirmed [3]; however, there is inconclusive evidence to support causality between smoking and many other musculoskeletal disorders.

Journal of Environmental and Public Health
Searching online databases revealed significant growth in the body of literature investigating relationships between tobacco smoking and the musculoskeletal system. During our comprehensive online search, we did not encounter any systematic reviews examining those relationships; however, we did find 10 systematic reviews of the effects of tobacco smoking on components of the musculoskeletal system. Five systematic reviews focused on smoking and the effects on dental implants and found smoking increases the risk of peri-implant bone loss and implant failure [7][8][9][10][11]. Another systematic review revealed an association between smoking and lumbar disc herniation [12]. Three other reviews found smoking was related to negative postoperative outcomes on knee ligaments [13], higher complication rates after anterior cruciate ligament (ACL) reconstruction [14], and slowed healing of rotator cuff repair [15]. Also, one review found smoking was associated with rotator cuff tears and other shoulder symptoms [16]. Our review will be the first to collect and assess all the recent literature on the effects of smoking on the musculoskeletal system. This systematic review will orient scientists interested in the health effects of smoking about the state of the science over the last decade as they conduct more advanced research. Also, the amalgamation of these data in one document will be helpful to the research community as there is a high degree of similarity and shared characteristics between musculoskeletal system components.
This systematic review evaluated literature published in the last decade to summarize the evidence regarding the effect of smoking on the musculoskeletal system. This systematic review will answer two main questions: Is there an association between tobacco smoking and musculoskeletal health? What are the effects of tobacco smoking on the musculoskeletal health?

Methodology
This systematic review followed the criteria of A Measurement Tool to Assess Systematic Review (AMSTAR). Before the onset of the systematic review, a specific protocol was developed to minimize bias. This protocol included a priori research questions, a comprehensive literature search, inclusion criteria for studies, screening methods and reasons for exclusion, data abstraction, scientific study quality, data analysis, and synthesis.
A comprehensive literature search using PubMed, Web of Science, and Science Direct was conducted. This search covered 10 years from January 1, 2007, to March 18, 2017, and included only articles written in English. The search strategies included a combination of the following key words: smoking, musculoskeletal system, bone, bones, joints, muscles, tendons, ligaments, and cartilage. Medical Subject Headings (MeSH) were used during the search of PubMed. This step was helpful to expand the search; for example, the entry terms for MeSH of smoking were as follows: smoking, cigar smoking, cigar, tobacco smoking, tobacco, hookah smoking, smoking, hookah, waterpipe smoking, waterpipe, pipe smoking, pipe, cigarette smoking, and cigarette. All retrieved records were pulled from databases using EndNote X7. Duplicated records were removed via EndNote or manually when EndNote failed to recognize duplicates discovered by the authors during title/abstract reviews. After that, abstracts of the retained records were screened for inclusion criteria: English language, human subjects, published January 1, 2007-March 18, 2017, and investigating effects of smoking on the musculoskeletal system. Retained records then underwent full-text screening and records that did not meet the inclusion criteria or were editorials, commentaries, dissertations, case studies, or reviews (e.g., overview, systematic review, and metaanalysis) were excluded. A total of 243 final full-text articles were included in the review and used for data abstraction.
Based on an a priori protocol, data abstraction from selected full-text articles included citation (authors, year), study design, sample characteristics (size, age, sex, race/ethnicity, and type of sampling), study purpose, findings, comments, and/or limitations on data quality and validity. Two independent authors (first and second) extracted data using a standard form. The data abstraction process was piloted for the first 10 articles; it was successful and was used for the remaining articles. Any disagreements between authors were resolved through discussion.
The findings in this review were synthesized qualitatively as there was heterogeneity in study designs and populations. Our narrative analyses considered study design and quality.
This review included studies using various designs: cohort studies (n = 106; 67 were prospective and 39 were retrospective), cross-sectional studies (n = 90), case-control studies (n = 16), randomized control trials (RCTs) (n = 14), and quasi-experimental studies (n = 10). Other study designs included secondary data analysis (n =5) and cross-sequential design (n = 2). Table 1 presents the classification of study designs and related information based on the categories and subcategories. Table 2 summarizes the effect of smoking on major outcomes of musculoskeletal health. Table 3 provides comprehensive information on each study in the review.
Journal of Environmental and Public Health 3 Arthritis (29) B. Osteoarthritis (14) C. Spondylarthrosis (7) D.TMD (4) A. Knee Joint Cartilage (7) B. Spinal Cartilage (12) Study Design   Reasons for exclusion: 35 were reviews (overview, systematic review, meta-analysis), 48 were commentaries, case study/report, or thesis, or symposia, and 20 full texts did not meet inclusion criteria (n = 103) Overall characteristics of these studies were as follows: 15 studies were conducted in males, 12 studies were conducted in females, and 13 included both sexes; 22 studies used data or samples from large-scale longitudinal studies; all studies used self-report to assess smoking habits, with the exception of 6 studies that used objective measures in addition to selfreport: 3 assessed level of cotinine [24,48,49] and 3 assessed level of exhaled carbon monoxide (EXCO) [22,36,41]. Table 3 provides comprehensive details on the findings from those studies for effects of smoking on selected bone-related outcomes. According to a majority of studies, smoking had adverse effects on BMD across age categories and sex. In males, regardless of age, method, and site of measurement for bone density, the cross-sectional studies found smokers had significantly lower BMD than nonsmokers [27, 42-44, 46, 52, 54, 55]. The cohort studies found male smokers exhibited a significant decline in BMD [32,33,50,54]. There was only one cross-sectional study that reported no significant difference 6 Journal of Environmental and Public Health

1.A: Tobacco Smoking and Bone Mineral Density (BMD), Bone Mineral Content (BMC), and Bone Turnover (BT) (40 Articles)
Beyth et al., 2015 [17] Cross-sectional study (i) N = 26 iliac bone marrow samples collected during pelvic surgery: 13  observe the impact of smoking on bone in untreated women (i) During treatment with nasal estradiol, the BMD of the lumbar spine for smokers had increased at 2 years but with less than that in nonsmokers (2.6% versus 3.9%, P = 0.03). Similar trend exhibited among controls (-3.6% versus -2.4%, P = 0.08) (ii) When smokers compared to nonsmokers, there was no difference in the response to estradiol in hip BMD (P = 0.89), whereas the change in the hip on the placebo was similar to that seen in the spine (P = 0.08). Supportive changes were seen within urinary CTX and serum OC Breitling, 2015 [19] Cross-sectional study (i) To examine the relationships between BMD and calcium intake (dietary) and to examine the interaction (smoking and calcium intake) on BMD (i) There was an overall positive trend between calcium intake and bone mineral density among three smoking behavior categories (ii) The interaction of smoking with calcium intake on BMD did not reach statistical significance and the dose-response curves became more similar across smoking behavior strata after adjustment for several factors (BMI and physical activity)  [20] Cross-sectional study [sample from PEAK-25 cohort] (i) N = 1,054: 591 never-smokers, 187 former smokers, 276 current smokers (ii) All participants were 25-year-old women (i) To evaluate the association between smoking and bone mass (BMD and fracture risk) in 25-year-old women (i) BMD and relative fracture risk did not differ between never-smokers and former and current smokers (ii) Among current smokers, BMD (femoral neck) decreased as cigarette consumption increased in dose-response effect (P = 0.037) (iii) BMD was not significantly lower in young women who had smoked for long duration (P = 0.07) or started smoking early (P = 0.64) (iv) After smoking cessation, lower BMD persisted up to 24 months, becoming comparable to never-smokers after 24 months Cross-sectional study (i) N = 96: 32 never-smokers, 32 smokers (>10 pack-year), and 32 mild/moderate COPD (current or former smokers) (ii) The average of age (Mean, range): (49.5, 46-58) for never-smokers, (53, for smokers, and (64.5, 58-74.5) for COPD patients (iii) 28 were males, and 68 were females (i) To investigate the joint effect of smoking and COPD on the health status, body composition, and exercise capacity (i) Compared to never-smokers, patients with COPD had lower fat-free mass (FFM) [P = 0.02] and fat-free mass index (FFMI) [P = 0.008](ii) Six-minute walk distance (6MWD) was lower in COPD and smokers than never-smokers (P = 0.01) (iii) Compared to never-smokers, smokers had worse SF-36 score for functional capacity (P < 0.001) (iv) Compared to smokers and never-smokers, COPD patient had lower SF-36 score for physical functioning (P < 0.001) and role-emotional (P < 0.001) (v) Both COPD diagnosis and smoking inversely associated with FFMI, 6MWD and health status 8 Journal of Environmental and Public Health in postmenopausal women and to assess the relationship between BMD and oxidant/antioxidant parameters (i) The rates of osteopenia and osteoporosis in smokers and nonsmokers were 75% and 52.5%, respectively (ii) The T-scores were significantly lower in smokers than in nonsmokers (median: -2.7 versus -1.4, P < 0.001) (iii) Activities of antioxidant enzymes (superoxide dismutase, glutathione peroxidase, and paraoxonase) were lower and the levels of oxidative stress products (malondialdehyde, nitric oxide) were higher in smokers than in nonsmokers (P < 0.001) (iv) In the smoking group, there was a significant correlation between decreased T-score and oxidative stress parameters Chassanidis et al., 2012 [24] Cross-sectional study (i) N = 105: 45 fractured bone (  (ii) COPD patients had significantly higher proportion of CD(+) T cells expressed RANKL and IL-17 than that of nonsmokers (P = 0.010).
(iii) All groups had similar frequency for RANKL expression in Th17 (P = 0.508).
Journal of Environmental and Public Health 9 years of age (iii) Alcohol intake and anxiety had no effect on bone outcome, and depressive symptoms had no effect on total body BMC 10 Journal of Environmental and Public Health (ii) In white women, the lower levels of total body BMC and hip BMC and BMD were associated with high score of anxiety.
(iii) There were no differences in age-adjusted BMC or BMD between ever-smokers and never-smokers. However, ever-smokers had higher depressive and anxiety symptoms than never-smokers.
(iv) There was no significant interaction between smoking status and depression or anxiety status.
(v) High level of anxiety was associated with lower BMC in group of ever-smoking Drage et al., 2007 [31] Cross-sectional study (i) N = 18: 5 nonsmokers, 13 smokers (ii) The average of age (Mean ± SD): 67.1 ± 12.6 (iii) 9 were males, and 9 were females (i) To investigate the association between jaws BMD and other skeletal site BMD and to investigate the influence of smoking on the BMD (i) The BMD of the ramus was similar to the femur; however it was significantly lower than the BMD of lumbar spine.
(ii) BMD of anterior maxilla was lower than BMD values of femur and ramus. (iii) There was a strong relationship between ramus BMD and spine and hip BMD, but there were no relationships between BMD of other jaw's areas and BMD skeletal sites. (iv) There was an inverse association between increasing age and BMD of both hip and the ramus. However, there was no significant association between BMD of hip, spine, and jaw and years edentulous or cigarette years

12
Journal of Environmental and Public Health (ii) There was a strong correlation between 25 (OH) D and smoking (P < 0.001), and 25(OH)D level was significantly lower in smokers than in nonsmokers.
(iii) In total population, regardless of the age group, there were no correlations between 25(OH)D and BMD in the femoral neck and the lumbar spine. Also, 25(OH)D was not correlated with bone turnover markers: serum osteocalcin, P1NP, and b-CTXs levels. (i) To investigate effects of alcohol and tobacco smoking on BMD and bone metabolism (i) There were no significant differences in BMD of the calcaneus among the 3 groups. However, blood total alkaline phosphatase activity (ALP) was significantly lower in the combined drinking and smoking group than in the control group (P < 0.05).
(ii) There were negative relationships between duration of alcohol consumption and ALP N-mid osteocalcin levels (all P < 0.001).
(iii) Daily cigarette use and smoking duration showed a significantly negative correlation with ALP (P < (iii) With later visits after 3-5 years of GH therapy, total BMC had increased by 1.8% (P < 0.05) in nonsmokers, while the changes in smokers were not statistically significant (P = 0.09).
(iv) Increase in lumbar BMD was seen in both groups, but it was significantly greater among nonsmokers (6.5% versus 2.6%, P < 0.05).
14 Journal of Environmental and Public Health (ii) Both former smokers and never-smokers had similar BMD at any measured sites, indicating that quitting smoking had a positive effect on BMD.
(iii) There was a negative association between smoking and lumbar BMD (r = -0.166, P = 0.004), but there no association was seen between smoking and femoral neck BMD. (iii) Smokers had higher levels of testosterone (total and free) and lower 25-OH-D than nonsmokers and the adjustment for such differences did not alter the associations between smoking and bone parameters.  between current smoking and trabecular microarchitecture of the radius and tibia. Also, to monitor such association after 5 years of follow-up (i) Compared to nonsmoker men at baseline and during follow-up, men started smoking since baseline had smaller increase in the mean of areal bone mineral density (aBMD) at the total body (P < 0.01) and lumbar spine (P = 0.04) and substantially greater decreases in aBMD at the total hip (P < 0.01) and femoral neck (P < 0.01).
(ii) Men who had started to smoke had a minor increment of the tibial cortical cross-sectional area (CSA) (P = 0.03), and a larger decrement of trabecular volumetric BMD (vBMD) (P < 0.001) than men who were nonsmokers.
(iii) The follow-up finding shows that the tibial trabecular vBMD was significantly lower in smokers than in the nonsmokers (7.0%, P < 0.01).

18
Journal of Environmental and Public Health

years
(iii) All participants were males (i) To examine whether smoking and its cessation influence bone status and metabolism in Japanese men ≥ 65 years (i) There were significant differences in the lumbar BMD between the three groups of never-smokers and former and current smokers (P = 0.005). However, total hip BMD was found to be similar between the three groups (P = 0.078).
(ii) Among never-and ever-smokers, lumbar spine and total hip BMD decreased with the number of pack-years or the number of smoking years, respectively. (iii) Smoking did not reveal significant effect for serum osteocalcin (OC) or tartrate resistant acid phosphatase isoenzyme 5b (TRACP 5b). (ii) There was a significant association between smoking and nonunion (P = 0.02).

22
Journal of Environmental and Public Health (ii) Compared to the subjects not exposed to nicotine, subjects who were exposed to nicotine had stronger association between BMD and DMG (OR = 2.31, 95%CI: 1.73-3.07 for exposed, and OR = 1.43, 95%CI: 1.16-1.75 for unexposed, P interaction = 0.008).
(iv) The 10-year risk of major (r = -0.550; P = 0.012) and hip fracture (r = -0.513; P = 0.021) was strongly correlated with femoral neck BMD but not with mechanical performance and to study the influence of several factors (e.g., smoking, hinge fracture, and weight bearing) on the process of gap filling (i) During all time of follow-up (6 and 12 weeks and 6, 12 and 18 months), there was a delay in the osteotomy gap filling rate between smokers but differences were not significant. However, early full weight bearing had no effect on the gap filling rate.
(ii) A fracture of the lateral hinge was found in 39% (27) of the patients. The highest rate was type I fracture (14%), followed by type II fracture (13%), and type III fracture (6%).
(iii) In patient with intact hinge (lateral), the highest rate in gap filling was observed in period between 12 weeks and 6 months after the surgery Taes  (ii) The findings at tibia found that current smoking was associated with a decreased cortical thickness (beta = -0.034 ± 0.01, P = 0.020) and greater endosteal circumference (beta = 0.027 ± 0.009, P = 0.016).
(iii) Smoking at early age (≤ 16 years) associated with higher fracture risk and lesser areal BMD.
(iv) The interaction between current smoking and free estradiol was observed in model used to predict cortical thickness (beta = 0.29 ± 0.11, P = 0.01).

24
Journal of Environmental and Public Health (iv) There was an inverse association between weight gain and risk of fracture in females who were nonsmokers.

1.C: Tobacco Smoking and Alveolar Bone (4 Articles)
Campos et al., 2015 [73] Case-control cross-sectional study (i) 19 heavy smokers, and 19 nonsmokers (never smoked)  with alveolar crest height (ACH) loss and mandibular bone mass (i) ACH loss was greater in older patients with P = 0.012 and patients had fewer mandibular teeth (P < 0.001).
The relationship between ACH loss and tobacco consumption was close to significant (P = 0.079).
(ii) There were significant associations between ACH and the number of mandibular teeth (P < 0.001) and tobacco consumption (P = 0.048).
(iii) There were a significant association between alveolar and basal bone densities and number of mandibular teeth (P = 0.012) and cortical width (P =

26
Journal of Environmental and Public Health (ii) In AgP group only, the mean plaque scores were significantly higher for smoker than nonsmoker (P = 0.04) (iii) The CAL was significantly higher in smokers than in nonsmokers of CP and AgP groups (P = 0.04, 0.01 respectively)  (ii) Calculus visible on radiograph was significantly increased from 2008 to 2013 (22% to 32%, P < 0.05).
(iii) Smoking was the strongest factor associated with ABL; however, socioeconomic factors had limited influence on the severity of ABL. CP in response to scaling and root planing (SRP) (i) In both groups of CP and healthy subjects, baseline salivary sCD44 profiles were significantly higher in smokers than that of nonsmokers (P < 0.001) with the highest levels recorded in smokers within the CP group.
(ii) After treatment, there was a significant decline in salivary sCD44 levels for smokers and nonsmokers in the CP group (P < 0.01); however, the difference was not significant.

30
Journal of Environmental and Public Health (ii) Compared to ex-smokers, smokers had significantly lower bleeding on probing (P < 0.05).
(iii) Salivary ICTP levels were similar among the study groups (P > 0.05). Salivary ICTP levels were correlated negatively with number of teeth (P <
(iii) Compared with smokers and never-smokers in control group, smokers with T2DM had poorer periodontal status (P < 0.05).

32
Journal of Environmental and Public Health (iv) The proportion of six identified pathogens did not differ between current and noncurrent smokers.  and OC compared to the nonsmoker counterparts (i) Periodontal measurements were significantly different between the two groups of healthy control and group of CP (P < 0.001). However, these measurements were not differed between smoker and nonsmokers of CP group (P > 0.05).
(ii) Compared to health control, patients with chronic periodontitis exhibited significantly higher salivary OC (P < 0.05).
(iii) Within the group of CP, smoker had lower salivary OC levels than nonsmoker (P < 0.001).
(iv) Within the healthy control group, smokers revealed higher log ICTP levels than nonsmoker (P < 0.001).
(v) Within the subgroups of nonsmokers, nonsmoker chronic periodontitis exhibited higher log ICTP levels than nonsmoker (P < 0.05).

36
Journal of Environmental and Public Health  (i) To compare the 10-year marginal bone loss rates around implants supporting single-unit crowns in tobacco smokers with and without a history of treated periodontitis (i) There were no significant differences in implant survival between the four groups and the rate ranged between 70% and 100% (P > 0.05).
(ii) Compared with the implants placed in PH patients, implants placed in PC patients had significantly higher marginal bone loss independent of the system used (P < 0.05).

38
Journal of Environmental and Public Health (ii) Anatomic location had significant influence on the implant survival in smokers with smooth-surface implants (P = 0.0047), but it had no influence on implant survival in smokers with rough-surface implants (P = 0.45) and in nonsmokers with smooth-surface implants (P = 0.17). (ii) In analysis based on the stratification, no significant differences were found between the mean GZ scores of implants with TB ≤ 1 mm (thin buccal wall) and TB > 1 mm (thick buccal wall) at baseline and after 2 and 3 years of follow-up.
(iii) The smoking did not seem to influence GZ changes over the follow-up period.

40
Journal of Environmental and Public Health     (ii) All clinical indices were in normal ranges and the Periotest values were decreased with time which indicated that implant becomes more secured in bone.
(iii) The long-term success was 98.6% for immediately loaded implants placed in occlusal function in smokers restored with fixed cross-arch implant-supported restorations. (ii) The marginal bone loss was significantly higher in the 4ITB group than that in the two implant groups.
(iii) The maximal probing depth was correlated with peri-implant bone loss (P = 0.011).
(iv) Smoking found to double the marginal bone loss regardless of the treatment strategy.

46
Journal of Environmental and Public Health (ii) Cytokines were significantly increased in smoker group than in nonsmoker group. The correlations between the cytokine levels and clinical parameters were more marked in smokers. (i) To compare peri-implant microbiota associated with implant transmucosal designs or smoking habits (i) Bone level implant had at least 1 more pathogen than that of tissue level; however, differences in each bacterium were not significant.
(ii) Treponema denticola had significantly more prevalence and abundance in smokers than in nonsmokers (P < 0.05).
(iii) Smokers and nonsmokers exhibited similar peri-implant microbiota based on the PhyloChip Array analysis for 5 smokers and 5 nonsmokers.

48
Journal of Environmental and Public Health

52
Journal of Environmental and Public Health  (iv) In never-smokers, there was no association between the erosive disease and any shared epitopes or antibodies.
(v) There was an association between all measured levels antibody and smoking and shared epitopes.

54
Journal of Environmental and Public Health (ii) There was an inverse association between the change in DAS28 over the first 3 months and number of py (r = -0.28, P = 0.002).

56
Journal of Environmental and Public Health  (ii) In patients with RA, simple erosion narrowing score was higher in current and former smokers than in never-smokers (P = 0.006), and rheumatoid factor titer was higher in current smokers compared with former and never-smokers (P = 0.004).
(iv) In patients with UA, there were no associations between smoking status and parameters of activity or radiographic damage.
( (ii) The follow-ups for EULAR response at 3, 6, and 12 months and at 2 years for the patients who had never smoked and who had been exposed or had not been exposed secondhand to tobacco smoke were found to be not significantly different (P = 0.91, P = 0. 88 (ii) There was a significant association between improvement in DAS28 and HAQ and disease duration up to 12 months, but such trend was not seen among the smokers.

58
Journal of Environmental and Public Health  (iii) Active smoking was associated with the adjusted risk for structural disease progression (OR = 0.50, 95%CI: 0.27-0.93, P = 0.028).
(iv) At 12 months, 16 patients quit smoking; the outcome for those patients was not significantly differed from other patients.

62
Journal of Environmental and Public Health (ii) The outcome scores were improved in all groups of smokers at 1-and 2-year follow-up when compared with the perioperative scores. However, the least improvement was seen among the active smokers. (iii) BMI was correlated in a dose-response pattern with the TJR (P < 0.001).
(iv) There was no association between socioeconomic status and risk for TJR. (i) To estimate the cross-sectional association between cigarette smoking and radiographic OA in Chinese population (i) The estimated prevalence of radiographic knee OA was 28.4%.
(ii) There was an inverse association between smoking and radiographic knee OA (P = 0.019). This association remained significant after adjustments for demographics (age, gender), lifestyle (physical activity, alcohol drinking, and betel quilt chewing), BMI, and educational level (P = 0.016).  Chung et al., 2012 [191] Cross-sectional study

66
Journal of Environmental and Public Health     Melis et al., 2010 [198] Prospective study (i) N = 352: 314 nonsmokers, 38 smokers (38 light smokers, 12 moderate smokers, and 11 heavy smokers) (i) To examine the effect of cigarette smoking on pain intensity in TMD patients. (i) Compared to nonsmokers, smokers had significantly higher overall TMD pain intensity (P = 0.007). Also, such differences were found to be particularly significant when nonsmokers compared with heavy smokers (P = 0.004).

2.D: Tobacco Smoking and Temporomandibular Joint Disorders (TMD) (4 Articles)
(ii) There was positive correlation between cigarettes smoked/day and pain intensity for whole group (P < 0.0001) and the group of females (P = 0.001), but such association was not observed in group of males.
Miettinen et al., 2017 [199] Cross-sectional study (i) N = 8,678: 5, 238  (ii) The effect was attenuated by 45% to nonsignificant level when study adjusted for psychological profile, cytokines, and history of allergy-like conditions. (iii) This study found smoking status was the strongest significant predictor of the fatigue index.

70
Journal of Environmental and Public Health (ii) During incremental exercise, 5 out of 6 smokers demonstrate a decrease in O2Hb throughout the incremental exercise, but 8 out of the 10 nonsmokers demonstrated a gradual increase of O2Hb.

72
Journal of Environmental and Public Health with proinflammatory cytokines, oxidative stress, and changes in muscular and physical performances (i) Smokers showed an increase in the oxidative stress represented by an increase in the levels of thiobarbituric acid reactive substances (TBARS) and a decrease in the total antioxidant capacity of plasma and catalase level (P < 0.05). Also, one of the inflammatory markers (sTNFR1) was increase in the P < 0.05.
(ii) Smokers showed a decrease in the total work which reflects skeletal muscle dysfunction (P < 0.05) (iii) IL-6, IL-10, sTNFR2, superoxide dismutase (SOD), peak torque, VO2 peak, HRmax, and walking distance were similar between groups Orozco-Levi et al., 2012 [212] Case-control study (i) N = 21: 14 male COPD patients who were smokers, 7 healthy nonsmoker controls (ii) The average of age (Mean ± SD) was 67 ± 7 (iii) All participants were males (i) To investigate whether the presence of chronic obstructive pulmonary disease (COPD) is associated with peripheral muscle injury (i) Both groups of control and COPD patients were found to have signs of injury in their skeletal muscles, and such injury was not only shown in cases showing severe airflow obstruction but also in the mild or moderate stages of the disease.
(ii) There was a significant association between current smoking and presence of COPD with the increased injury of the muscle as evidenced by the finding of electron microscopy techniques.

74
Journal of Environmental and Public Health  (iii) There was a negative correlation between pack-years and muscle strength.
(iv) Pack-years were found to be only correlated with CRP levels after the adjustment for age, sex, and BMI.
(ii) There was no differences in the loss of weight, fat mass, lean mass and strength between patients with OLD and never-smoking controls.    (ii) Increasing smoked pack-years were associated with increased medial cartilage volume loss (P = 0.04).
(iii) Compared to never-smokers, the BMLs among ever-smokers were 11.4 more likely to persist over higher thickness for the femoral medial, intercondylar, and lateral cartilage (P = 0.002, P = 0.017, and P = 0.004, respectively).
(ii) Compared with the nonsmokers, smokers had lower strain ratio for the medial distal femoral cartilage (P = 0.003).
(iii) There was significant positive association between amounts of smoking and cartilage thicknesses (P < 0.05).
(iv) There was significant negative association between amounts of smoking and medial cartilage strain ratios (P < 0.05).

78
Journal of Environmental and Public Health (ii) After 2 years of follow-up, smokers had significant improvement in the knee score compared to that of nonsmokers (P < well as pack-years smoked (P = 0.03). However, the study found no association between smoking and presence of tibiofemoral cartilage defects.

4.B: Tobacco Smoking and Spinal Cartilage (12 Articles)
Appaduray et al., 2013 [229] Retrospective cohort study (i) N = 902: 40 patients were diabetic and had positive smoking history, 75 with diabetes, 343 with smoking history, and nonsmokers, and 444 control group (ii) Average of age was 55 ± 17.7 years (i) To determine the effects of diabetes and smoking on the outcomes of lumbar spinal surgery (i) Diabetes was significantly associated with an increased risk of developing single (P = 0.007) or multiple complication (P = 0.008), infectious complication (P = 0.015), and cardiovascular complications (P = 0.015). However, the history of positive smoking was not found to increase the rate of poor outcome. Retrospective cohort study (i) N = 500 who underwent laminectomy at 1, 2, or 3 levels (ii) 284 were males (238 males were nonreoperation group and 46 were in reoperation group), and 216 were females (181 females were nonreoperation group and 35 were in reoperation group) (i) To identify the predictors of an increased risk for reoperation in patients who had underwent lumbar laminectomy (i) Out of 500, 81 patients (16.2%) developed subsequent spinal disorders that required reoperation.
(ii) Smoking was an independent predictor for all reoperations (OR 2.15, P = 0.01) including reoperation after single level (OR 11.3, P = 0.02) and multilevel and after a multilevel (2 or 3 levels) laminectomy (OR 1.98, P = 0.05), but not for nine patients who underwent nondegenerative conditions.
(iii) Also, smoking was a significant predictor for reoperation in 72 with spinal degeneration (OR 2.06, p = 0.04), but it was not a significant predictor for 9 patients that underwent reoperation for nondegenerative conditions Eubanks et al., 2011 [231] Retrospective cohort study (i) N = 158: 117 were nonsmokers, 41 were smokers (ii) Mean of age was 61 with a range of 35-87 years (iii) 93 were males, and 65 were females (i) To assess whether smoking reduces the fusion rate in patients that underwent posterior cervical fusion with lateral mass instrumentation and iliac crest bone grafting, and if it is associated with increased pain, decreased activity level, and decreased rate of return of work (i) The fusion rate (100%) was similar in both groups of smokers and nonsmokers. However, 80% of patients had Grade I or II in Odom Criteria used to assess the clinical outcome.
(ii) Smokers were 5 times more likely to develop advance grades (Grade III or IV) and usually had huge limitation in their physical activities.
(iii) The fusion rate was not influenced by diagnosis, age, and gender

80
Journal of Environmental and Public Health Prospective cohort study (i) N = 50: 34 were nonsmokers, 16 were smokers (ii) Mean of age was 58 years with a range of 29-81 in nonsmokers and it was 47 years with a range of 29-75 in smokers.
(iii) 23 were males, and 27 were females (i) To assess whether smokers yield worse results concerning lumbar interbody fusion than nonsmokers (i) Compared to the smokers, nonsmokers had significantly higher rate of fusion (P = 0.01).
(ii) There were no differences in the in ODI improvement (P = 0.93) and pain reduction (P = 0.54) between smokers and nonsmokers.
(iii) One year after the surgery, the intake of opioids was only reduced in nonsmokers from 32.35% to 29.41%. However, it remained unchanged among the smokers (43.75%) Lau et al., 2014 [235] Retrospective cohort study (i) N = 160: 79 were nonsmokers (never smoke), 41 were quitters (stopped for at least 1 year), and 40 were current smokers (ii) Mean of age (mean ± SD): 52.2 ± 11.3 for nonsmokers, 55.0 ± 11.7 for quitters, and 52.0 ± 12.0 for never-smokers (iii) 93 were males, and 67 were females (i) To examine the effects of smoking on perioperative outcomes and pseudarthrosis rates following anterior cervical corpectomy (i) Current smoking was significantly associated with higher complication rates (P < 0.001) and longer lengths of stay (P < 0.001).
(ii) The complications that were experienced in current smokers were mostly infections (76.5%), and this proportion was greater than that of quitters and nonsmokers (p = 0.013).
(iii) The follow-up at 1 year found that current smoking acts as an independent risk factor for pseudarthrosis (P = 0.012).  Achilles tendons in the proximal, middle, and distal thirds region of the tendons. Also, smoking groups had significant lower strain ratio measurements in the same regions (P < 0.05).
(ii) The thickness of patellar was found to be negatively correlated with smoking amount (all P< 0.05).

84
Journal of Environmental and Public Health  (i) To compare the clinical outcomes of ACL reconstruction between smokers and nonsmokers.
(ii) To find an optimal graft in ACL reconstruction with regard to clinical outcomes for smoking patients. (i) Compared to nonsmokers, the smokers were found to have significant unsatisfactory outcomes regarding the stability (P < 0.001 in side-to-side difference in anterior translation) and functional scores (P < 0.001 in Lysholm Knee Score, IKDC subjective and objective score).
(ii) The Achilles tendon-bone allograft showed the worst outcomes, with statistically significant mean differences for smoking patients in the side-to-side difference in anterior translation, Lysholm knee score, (i) To investigate the association between the amount of smoking and outcomes following ACL reconstruction (i) The three groups differed significantly in terms of stability represented by the postoperative knee translation (P = 0.003) and function represented by Lysholm score (P < 0.001) and IKDC subjective score (P < 0.001) (ii) There was a dose-dependent relationship between pack-years and postoperative anterior translation (P = 0.015) and IKDC objective grade (P = 0.002).  (ii) Breastfeeding was associated with 2-3% increase in BMD in all site except radius and was associated with a 1/3 reduction in risk of fracture (iii) Birth weight was associated with hip, radius, and total body BMD, but such association did not stand after the multivariate analysis. Furthermore, birthweight was not associated with the fracture Kim  (i) Study aims to investigate an intrauterine influence of maternal and paternal smoking during pregnancy on offspring bone mass at mean age 9.9 years (i) Both paternal and maternal smoking were associated with increased total body less head (TBLH) and spine BMC, bone area (BA), and BMD in girls but not boys in calcaneus BMD between 3 groups: an alcohol drinkingonly group, combined alcohol drinking and smoking group, and control nondrinking/nonsmoking [39]. In adolescent females, 2 cross-sectional studies found a high frequency of smoking was associated with lower rate of total body BMC [29] and hip BMD [28,29]; these findings were supported by cohort studies that found initiation of smoking at age 13 affected bone accrual and was associated with low mean BMD at age 17 [33,47]. Another cross-sectional study of adolescent females reported significant linear relationships between urinary cotinine and BMD of the femoral neck, total femur, and lumbar spine [48]. However, only one cross-sectional study in adolescent females found no significant difference in BMC and BMD between smokers and nonsmokers [30]. In premenopausal women, one cross-sectional study reported the BMD of smokers was not significantly different than the BMD of nonsmokers [20].
In postmenopausal women, cross-sectional study findings demonstrated postmenopausal women who smoked had significantly lower BMD than postmenopausal women who did not smoke [23,56] and an increased risk of falls regardless of the BMD T-score [21]. Two randomized control studies were conducted in postmenopausal women. One study found consumption of blackberries was effective in reducing bone loss of the total body BMD in a smokers' group (P = 0.0284) [38]. Another study found quitting smoking significantly associated with increased body weight, fat, muscles, and functional mass that affected BMD [41]. A study using data from one RCT found administration of nasal estradiol for 2 years increased the lumbar spine BMD of smokers (P = 0.03) but did not increase total hip BMD (P = 0.89) [18]. Finally, two cross-sectional studies enrolled both males and females, and one reported BMD and BMC were significantly lower in smokers than those of nonsmokers [26]; the second study used a small sample and found no association between packyears and BMD [31].
The correlational analysis did not find a significant effect for serum osteocalcin (OC) or tartrate resistant acid phosphatase isoenzyme 5b (TRACP 5b) [55]. The interaction between smoking and other compounds found smoking with calcium intake did not reach statistical significance for BMD [19]; however, elderly women with the lowest tertile of choline who were exposed to nicotine had the highest risk for low BMD (OR=4.56, 95%CI: 1.87-11.11) [49]. One prospective study found growth hormone (GH) therapy after 3-5 years did not significantly improve total BMC of smokers (P = 0.09) [40].
The interaction between smoking and genetic factors was also investigated. A potential interaction was reported between smoking and receptor-related Protein 5 (LRP5) C135242T (rs545382) on osteoporosis in postmenopausal women [35] and between smoking and polymorphism of glutathione S-transferases (GSTT1) on bone quality index in young adult men [45]. One cross-sectional study reported a relationship between PTH and hip BMD only for nonsmokers [53]. Finally, in a cross-sectional methodological study, a strong correlation between bioelectrical impedance analysis (sBIA) and dual energy X-ray absorptiometry (DEXA)] regarding whole-body fat mass (FM) and lean mass (LM) (r > 0.9, p < 0.001) was found [51]. (n = 16). The overall characteristics of these studies were as follows: 12 studies were conducted in both sexes: 2 studies in females only [61,71] and 2 studies in males only [62,70]; 12 studies were cohort studies; 3 obtained data or samples from large-scale longitudinal studies; all studies used self-report to assess smoking habits with the exception of one study that assessed levels of cotinine [67]. Table 3 provides comprehensive details on studies which examined the prevalence of fracture in smokers, the association between smoking and fracture risk, fracture healing, the biological mechanism of fracture in smokers, and the interaction of smoking and other fracture risks.
Two prospective cohort studies of both sexes investigating the postsurgery level of serum transforming growth factor-beta 1 (TGF-beta 1) found TGF-beta 1 was lower in smokers than in nonsmokers at 4 weeks [65] and 8 weeks [63]. The trend of lower level of TGF-beta 1 in smokers than that of nonsmokers was observed in both groups of patients with normally healed fractures and delayed healed fractures [65]. Finally, two studies examined interactional effect of smoking and other factors. In the first study, changes in BMI had an effect on fracture risk in nonsmokers, but not in smokers [72]. In the second study, plasma dimethylglycine (DMG) increased the risk of hip fracture in cohort of elderly males and females (HR = 1.70, 95 % CI: 1.28-2.26), and such risks were noticeably increased in women exposed to nicotine (HR = 3.41, 95%CI: 1.40-8.28) [67]. (n = 4). In two cross-sectional studies, one study found smokers had significantly lower alveolar bone density values (P ≤ 0.002) and a greater distance from cemento-enamel-junction to the alveolar bone crest (P < 0.0001) [74]. The second study did not find smoking significantly correlated with alveolar crest height loss [75]. The third cross-sectional study found smoking negatively affected the expression of bone sialoprotein (BSP) and osteocalcin (OC) mRNA (P< .05) and positively altered the expression of Type I collagen (COL-I) (P< .05); however, smoking was not statistically correlated with the expression of mRNA for tumor necrosis factor-alpha (TNF-), transforming growth factorbeta (TGF-), or osteoprotegerin (OPG) [73]. The fourth study, an RCT, found that, despite involvement of smokers in dental hygiene program, smokers had significantly lower density of alveolar bone by Day 365 (P < 0.05) and Day 545 of follow-up during the dental hygiene program (P < 0.01) [76]. Table 3 provides comprehensive details on the 4 studies that examined the effect of smoking on the alveolar bone.

Periodontitis (n=34).
Thirty-four studies examined the prevalence and clinical parameters of periodontitis in smokers, the biological mechanism of smoking, the interaction between smoking and periodontitis risk factors, and interventions to minimize periodontitis in smokers. The characteristics of these studies were as follows: 27 studies were conducted in both sexes; 7 studied males only. Twenty-four studies were cross-sectional, 5 studies were RCTs, 4 studies were cohort, and 1 study was a case-control. Two studies out of 34 obtained data or samples from large-scale longitudinal studies. All the studies used self-report to assess smoking, with the exception of three studies that assessed levels of cotinine [88,99,101]. Table 3 provides comprehensive details on the studies that examined the relationship between smoking and periodontitis.
Eleven studies investigated the prevalence and association of smoking on periodontitis and periodontal parameters. Ten studies were cross-sectional. The comparative analysis found smokers compared to nonsmokers had significantly deeper periodontal pockets [77,89,90,92,108], higher mean clinical attachment loss (CAL) [78,91,92,108], higher mean plaque scores [78,91,92], greater fraction of teeth with apical periodontitis [79], higher marginal bone loss, and a greater number of missing teeth [92]. In a correlational analysis, smoking was strongly associated with alveolar bone loss [84] and the percentage of palatal periodontal pockets ≥ 6 mm [77]. Heavy smoking was also associated with higher prevalence [79,94] and severity of periodontitis [77,94]. A comparative study between different modalities of smoking found cigarette smokers had higher frequency of probing pocket depth ≥ 4 mm and a higher incidence of severe periodontitis compared to nontobacco users [90]. Two studies compared cigarettes smokers to waterpipe and narghile users and found a similarity between groups on most periodontal parameters [92,95].
Fourteen studies examined the potential biological mechanism for smoking in periodontitis and what potential biomarkers may be affected. Thirteen studies were crosssectional designs and enrolled both sexes. Three studies examined smoking and nonsmoking subjects without periodontitis and found smokers had significantly higher synthesis of lipoxygenases and isoprostanes in the extracted periapical granuloma [83], higher whole salivary IL-1 beta and IL-6 (P < 0.05) [93] but a lower total amount of plateletderived growth factor (PDGF-AB) (P = 0.014) in gingival crevicular fluid [85]. Four studies were conducted in smokers and nonsmokers with periodontitis and found smokers had significantly lower levels of salivary osteocalcin (OC) (P < 0.001) [88]; a lower median serum level of OPG (P = 0.0006) [88]; higher levels of prostaglandin E-2, lactoferrin, albumin, aspartate aminotransferase, lactate dehydrogenase, and alkaline phosphatase [96]. The groups had similar levels of salivary C-telopeptide pyridinoline cross-links of Type I collagen (P > 0.05) [88]; median serum receptor activator of nuclear factor kappa-B ligand (RANKL) (P = 0.0942) [99]; and gingival crevicular levels of RANKL and osteoprotegerin (OPG) [107] and similar proportion of identified pathogens [96]. Seven studies enrolled both smokers and nonsmokers in groups with or without periodontitis. Based on periodontal status, the group with periodontitis (smokers and nonsmokers) had significantly higher plasma sRANKL, TNF, a proliferation-inducing ligand (APRIL and BAFF) and lower OPG (P < 0.01) [101], and higher salivary OC (P < 0.05) [102] than the healthy control group (smokers and nonsmokers without periodontitis). Interestingly, these two studies found levels for some of these markers were altered by smoking; sRANKL and TNF concentrations were significantly greater (P = 0.011, P = 0.001; respectively), and OPG concentration was significantly lower (P = 0.001) in smokers with periodontitis; however, such trend was not seen for salivary OC [101,102]. The results from these studies indicated that smokers had more lymphocyte and higher levels of both IFN-and IL-13, regardless of periodontal status [82], had higher salivary sCD44 profiles (P < 0.001) with the highest levels recorded in smokers in the periodontitis group [87], and had significantly higher levels of salivary calcium level (P< 0.05) [97]. A subgroup analysis for smoking and periodontal status found that smokers with chronic periodontitis exhibited significantly higher levels of sIgA [98] and lower plasma OPG concentrations (P = 0.007) but higher sRANKL/OPG ratio (P = 0.01) than smokers without periodontitis [103]; however, smokers and nonsmokers with periodontitis exhibited similar values for plasma sIgA, sRANKL, and OPG concentrations. Smoking is one of the greatest risks for periodontitis and may increase host susceptibility to tissue destruction especially in presence of other factors such as the functional defect of leukocyte and monocyte [98]. These findings indicate periodontal inflammation in smoker with chronic periodontitis patients, as evidenced by high levels of sIgA, seems to lower plasma OPG levels and thereby increase the RANKL/OPG ratio and possibly play a role in the increased susceptibility for alveolar bone destruction in smoker subjects.
One cross-sectional study suggested the interaction between smoking and vitamin D receptor gene polymorphism (CC+CT genotypes of FokI) increased the risk of periodontitis (OR = 9.6, 95%CI: 4.5-20.4). The combined effect was 3.7 times greater than expected from the sum of individual effects [80]. Eight interventional studies examined therapies to manage periodontitis. One observational cohort study that monitored the effect of smoking cessation found quitters had a higher reduction of mean probing depth and CAL relative to nonquitters (P ≤ 0.05) [105]. Another prospective observational study found periodontal maintenance therapy every 3-4 months inhibited the progression of CAL, probing depth, and tooth loss in smokers [86]. Wan et al. [109] found in a prospective cohort study that anterior teeth, sites without plaque, and nonsmoking were significantly associated with a greater reduction in probing pocket depth [109]. Three RCTs found adjunct treatments of low-dose doxycycline for 6 months [100], systemic azithromycin [81], or a daily dose of 325 mg of aspirin [106] did not significantly improve periodontal parameters in smokers with chronic periodontitis. In contrast, 2 RCTs in smokers with chronic periodontitis successfully improved some periodontal parameters. Compared with treatment using only scaling and root planing (SRP), the treatment using Simvastatin (1.2% biodegradable controlled-release gel) as an adjunct to scaling and root planing (SRP) significantly reduced probing depth and significantly increased bone filling (all P < 0.001) [104]. Smoking is one of the greatest risks for periodontitis and is associated with poor periodontal parameters; such finding provides evidence that the treatment used Simvastatin besides SRP in smokers suffering from chronic periodontitis was more effective in reducing the negative effect of smoking on the periodontal parameter than the treatment using only SRP. The second RCT found the treatment using modified YJ (mYJ) Chinese medicinal herbs in a nonsurgical treatment for smokers suffering from periodontitis was associated with higher computer-assisted densitometry values than the treatment using original YJ Chinese medicinal herbs with nonsurgical treatments (P = 0.025) [110]. Also, this finding provides evidence that the use of mYJ Chinese medicinal herbs in a nonsurgical treatment was effective in reducing the negative effect of smoking on the periodontal parameter as evidenced by the increases in radiographic alveolar bone density.
3.1.5. Bone Implants (n = 33). There were 33 studies which investigated implant survival/failure rates, clinical parameters of success/failure, risk factors of implant survival, interaction between smoking and risk factors on the implant survival, effects of implants on surrounding tissue, complications associated with implants, biological mechanisms of smoking effect on implants, and interventions to reduce the effects of smoking and enhance implant survival rate in smokers. The characteristics of these studies were as follows: all 33 studies examined dental implants and enrolled both sexes; 19 were cohort studies; 7 were cross-sectional studies; 4 were RCTs; and 3 were case-control studies. Eleven studies had small samples. All studies used self-report to assess smoking. Table 3 provides comprehensive detail on  these studies exploring the effects of smoking on bone  implants. Thirteen studies examined the effect of smoking on dental implant survival with special consideration of implant type and time of follow-up. Eleven studies were cohort studies. Two studies investigated smoking and early implant failure and the first study found early implant failure was threefold higher in smokers than nonsmokers [141] while the second study found frequency of tobacco smoking was not associated with early implant failure [136]. Ten studies examined longterm survival/failure of dental implants. Two studies reported smoking did not influence implant survival rates [115,123], although 8 studies provided contradictory findings. A correlation analysis found smoking status [112,139], and packyears [139] were inversely associated with dental implant survival. A comparative analysis between smokers and nonsmokers found smokers had lower implant survival rates [114,119,124,130,132,142]. A subgroup analysis based on implant type found smokers had higher failure rates for turned [132] and smooth-surface implants [114]. One study of only tobacco smokers found implant survival with turned or screw surfaces was similar in tobacco smokers regardless of periodontal status [111]. Five of 13 studies measured marginal bone loss; 4 reported smokers demonstrated significantly greater marginal bone loss than nonsmokers [119,123,141,142], and two studies did not report a significant difference [130,131].
Six studies examined the clinical effects of implants on surrounding tissue in smokers. One retrospective cohort study found smoking was associated with overall complications (e.g., implant loss, infection, peri-implantitis, and mucositis) (P = 0.008) [129]. Five studies examined histometric parameters for dental mini-implants, and one case-control study carried out on smokers found mean bone-to-implant contact (BIC%) better in sandblasted acid-etched surfaces than machined surfaces (22.19 ± 14.68% versus 10.40 ± 14.16%, P < 0.001) [117]. Smoking is associated with increased risk of bone implant failure due to its negative effect on tissues surrounding the implant; such finding indicates that the negative effect of smoking on histometric measurements after dental mini-implants was significantly minimized through using of implants with sandblasted acid-etched but it was not improved with the use of implants with machined surfaces. The remaining 4 studies were prospective cohort studies and found smokers had significantly lower BIC% [120,133], lower bone density in thread areas (BA%) [133], less stability at 3, 4, 6, and 8 weeks after surgery [135], and less regrowth of papillae and midfacial soft tissue [128].
Eight studies examined the biological mechanism of smoking on tissue surrounding dental implants and explored potential biomarkers that could be affected by this mechanism. Five of eight were cross-sectional studies; one study analyzed peri-implants fluid of smokers' prior implant placements and found smoking negatively altered the mRNA expression of bone sialoprotein (BSP) and osteocalcin (OC) and positively affected the expression of Type I collagen (COL-I) (P < 0.05). However, smoking was not correlated with the expression of TNF-, transforming growth factorbeat (TGF-), or OPG (P > 0.05) [73]. Four studies analyzed peri-implant fluid and found smokers and nonsmokers had similar levels of pathogens [113,140], OPG, and RANKL/OPG [126]. However, there were contradictory findings regarding the level of cytokines (IL-4, IL-8, or TNF-) in smokers and nonsmokers, as they were reported to be similar in one study [113], significantly lower in one study [126], and significantly higher in yet another study [137]. One casecontrol study found heavy smokers with an IL-1 polymorphism did not increase their risk for peri-implantitis [121]. One short-term prospective study found a 7-day followup for the whole genome array of implant adherent cells was not different between smokers and nonsmokers [138]. The long-term prospective cohort study found smokers with previous periodontal disease had significant clinical signs of inflammation and significantly higher counts of pathogenic bacteria [127].
Four studies were randomized control trials; 2 measured peri-implant parameters for implants with different configurations. One trial found smoking did not influence peri-implant soft tissue response (recession and the papilla index) [116]; a second study found smoking doubled marginal bone loss regardless of treatment [134]. One methodological RCT using stereolithographic surgical guides found smoking was associated with inaccurate implant placement [118]. A therapeutic RCT found mechanical debridement with adjunct antimicrobial did not significantly improve parameters of bleeding on probing, probing depth, or crestal bone loss in smokers [122]. There were positive outcomes (high implant survival, bone level, and low rate of biological complications) reported by one retrospective cohort study where the authors monitored dental implant rehabilitation in patients with systemic disorders and smoking habits [125].
3.1.6. Bone Graft (n = 5). Two of the 5 studies were randomized clinical trials. The first trial found smokers who received an acellular dermal matrix graft (ADMG) with enamel matrix derivative (EMD) had a higher mean gain in recession height and root coverage than smokers who received ADMG alone [143]. This finding indicates that the treatment combining EMD with ADMG was found to be more effective in reducing the negative effect of smoking on root coverage than the treatment using only ADMG. A second trial found regenerative treatment of platelet-rich plasma combined with a bovinederived xenograft did not improve periodontal parameters in smokers [147]. Three long-term prospective cohorts compared smokers to nonsmokers and found smokers had significantly higher marginal bone loss up to 4 years after onlay bone grafting in the atrophic maxilla [145]. These patients also had higher tissue inflammation around augmentation sites once they received bone graft titanium-reinforced ePTFE membranes [144] and had similar survival rates for dental implants after A Le Fort I osteotomy and interpositional bone graft in combination with implants in the atrophic maxilla [146]. Table 3 provides comprehensive details on the 5 studies that examined the effects of smoking on bone graft.

Tobacco Smoking and Joints (n = 54)
3.2.1. Rheumatoid Arthritis (n = 29). The overall characteristics for these 29 studies were as follows: all studies enrolled males and females, 17studies were cohort studies, 7 were cross-sectional studies, 3 were case-control studies, 1 was an RCT, and 1 was a secondary analysis. Twelve studies obtained data or samples from large-scale longitudinal studies and all used self-report to assess smoking habits with an exception of one study that assessed level of cotinine [159]. Table 3 provides comprehensive detail about studies that examined the effect of smoking on several outcomes in patients with RA.
Five studies of varying design (2 cross-sectional studies, 2 prospective cohort studies, and 1 case-control study) enrolled patients from both sexes of similar age. These studies examined the effect of smoking on RA clinical outcomes, such as disease activity, functional capacity, radiographic damage, serology, and existence of extraarticular manifestations. Overall, the collective results were that smokers had significantly higher scores on the Disease Activity Score of 28 joints (DAS 28) [162,167], the functional disability score (Health Assessment Questionnaire) [162], the simple erosion narrowing score [167], CRP [162], and a rheumatoid factor titer [167]. These patients demonstrated severe extraarticular RA [162] and took significantly more disease-modifying antirheumatic drugs (DMARD) [176]. One study reported no difference in DAS28 and radiographic scores between smokers and nonsmokers [176]. Smoking was found to be independently associated with DAS28-CRP3 in human leukocyte antigen-shared epitope (HLA-SE-positive) patients, but not in HLA-SE-negative patients (P for interaction = 0.02) [158], higher Modified Health Assessment Questionnaire [158] scores, and greater number of rheumatoid nodules [161].
Smoking and RA remission were investigated in two studies. A cross-sectional study reported current smokers had higher remission rates than persons who had never smoked or former smokers [154]; however, a prospective cohort study reported lower remission rates in current smokers compared to persons who had never smoked or former smokers at 12-month follow-up [169]. Two prospective studies examined smoking on RA progression, and one study with a large sample size reported radiographic progression for joint damage was not significantly different between smokers and nonsmokers (P = 0.26), but further analysis by authors found smoking intensity (pack-day) to be inversely associated with radiographic progression [152]. Meanwhile, a second study with a small sample size found current smoking associated with radiographic progression [165].
There were 10 studies that examined the effect of smoking on response to RA therapies. A first group of 5 cohort studies was conducted in patients with early stage RA. Four studies found smoking associated with poor response after 3 months of methotrexate or anti-TNF-therapy [166], after 6 months of combined therapy of methotrexate and sulfasalazine [163], and after 12 months of glucocorticoids and DMARDs [169,170]. One study reported no difference in response to therapy after 12 months' follow-up in patients who continued or quit smoking [175]. Similar findings were reported in 5 studies that enrolled RA patients regardless of disease stage.
Four studies reported current smoking was associated with poor response after 3 months [148,160,171] and 12 months [160] of anti-TNF-therapy and after 48 and 102 weeks of therapy that included methotrexate [159]. One of the 5 studies reported exposure to secondhand smoking did not influence the response to RA therapy after 3, 6, and 12 months and 2 years [168].
Seven studies (2 case-control studies, 3 cohort studies, 1 cross-sectional study, and 1 secondary analysis) investigated the interaction between smoking and other factors on RA. The results of these studies found a significant increase in disease activity when there was an interaction between heavy smoking and HLA-DR beta 1 4-amino acid haplotype primarily Positions 11 and 13 [155], between smoking and all positive anti-citrullinated peptide antibodies (ACPA) [153,157,164], and between ever smoking and mannosebinding lectin (MBL2) genotype YA/YA [156]. Further results were increased signs of joint inflammation in first-degree relatives who were younger than 50 and had smoked more than 10 pack-years [172]. There was no interaction found between smoking and endothelial growth factor A haplotype [VEGFA-2578 AA genotype and (A 2578-C 460-G+405)], but endothelial growth factor A haplotype was found to be associated with reduced disease activity in patients of RA who had never smoked [151]. Four studies investigated smoking effects on certain mechanisms and biological markers in patients with RA. One prospective cohort study found smoking and ACPA predicted persistence of high levels of survivin (OR = 4.36, 95% CI: 2.64-7.20, P < 0.001, positive predictive value 0.66, and specificity 0.83) [174]. Two cross-sectional studies compared level of essential and trace elements of smoker and nonsmokers with RA and matched healthy controls of smokers and nonsmokers to determine if there were any associations between toxic elements, cigarette smoking, deficiency of essential trace elements, and risk of arthritis. One study found smokers and nonsmokers RA patients had significantly higher hair levels of toxic elements (Cd and Pb) and lower hair levels of trace elements (Zn, Cu, and Mn) than those of smokers and nonsmokers healthy individuals [149]. The second study found smokers with RA had significantly higher hair and blood levels of toxic element (Cd, Pb, Hg, and AS) and lower hair and blood levels of trace elements (Zn, Cu, Mn, and Se) [150]. Finally, another cross-sectional study reported smoking pack-years was inversely correlated to body fat composition in patients with RA [173]. (n = 14). Three cross-sectional studies provided disparate findings on the effects of smoking and OA. One study reported smoking was not significantly associated with hand OA in a Chuvashian community [182]; two other studies reported an inverse relationship between smoking and radiographic knee OA (P = 0.019) [190] and between indirect smoking and knee and hip OA (OR = 0.271; 95% CI: 0.088-0.828) [183]. One prospective study reported smoking was not significantly associated with the prevalence or incidence of radiographic knee OA [188]. Three prospective studies found smoking associated with higher pain scores [177,178], increased risk for cartilage loss at the medial tibiofemoral joint (OR = 2.3, 95% CI:1.0 -5.4), and increased risk for cartilage loss at the patellofemoral joint (OR = 2.5, 95% CI: 1.1 -5.7) [177]; however, smoking reduced the risk of total joint replacement (TJR) in presence of [rs1051730 T] alleles (HR = 0.84, 95% CI: 0.76 -0.98, per T allele) [181]. Two prospective studies found smoking significantly associated with higher complication rates [189], but not with functional outcomes after a tibial osteotomy in patient with RA [180]. There were five studies that examined the association of smoking with the risk for joint replacement and the risk for complications after joint surgery. Two cohort studies investigated the risks for joint surgery: one prospective cohort study reported smoking increased the risk for total joint replacement (TJR) in males [186]; however, conflicting evidence was reported by another retrospective cohort study that found an inverse association between smoking and TJR (adjusted-HRs: 0.60; 95% CI: 0.48-0.75, and 0.70; 95% CI: 0.56-0.86 in men and women, respectively), but this study investigated the risk for only primary TJR and included both sexes [187]. Another prospective study of patients who underwent total hip or knee arthroplasty found no difference in perioperative mortality rates between smokers and nonsmokers; however, smokers had a higher complication rate [179]. In two retrospective studies, smoking significantly increased the risk for early failure of total hip arthroplasty [185] and wound breakdown after total ankle replacements [184]. Table 3 provides comprehensive details on the 14 studies in this subsection.

Spondyloarthritis (n = 7).
Three of these seven studies investigated the effects of smoking on biological markers in patients with SA; 5 studies investigated effect of smoking on clinical, functional, and imaging outcomes of SA. Three cross-sectional studies found smoking was associated with lower matrix metalloproteinase-generated Type II collagen fragment in patients with SA (P = 0.02) [193] and higher level of vascular endothelial growth factor in patients with ankylosing spondylitis (VEGF) (P < 0.05) [195,196]. Three crosssectional studies, two in patients with ankylosing spondylitis and one in patients with early axial spondyloarthritis, and one prospective cohort study in patients with early axial spondyloarthritis reported smoking was associated with higher pain scores [191,192], disease activity and functional status [191,192,195], poor quality of life [191,192], and spinal radiographic/MRI progression [191,194]. Also, pack-years were positively correlated with duration of inflammatory back pain (r = 0.628, P < 0.001), Bath AS Functional Index (BASFI) (r = 0.443, P < 0.001), and the severity of radiographic damage assessed by the modified Stroke AS Spine Score (mSASSS) (r = 0.683, P < 0.001) [195]. Finally, one case-control study found collagen IX tryptophan (Trp+2) alleles and smoking status did not influence the risk for cervical spondylotic myelopathy (OR = 1.34, 95% CI = 0.85-2.18, P > 0.05); however, smoking intensity with collagen IX tryptophan (Trp+2) exhibited a dose-response relationship with cervical spondylotic myelopathy [197]. Table 3 provides comprehensive details about the 7 studies that examined the effects of smoking on SA.

Temporomandibular Joint Disorders (n = 4).
Of the four studies in this subsection, 4 studies with varying designs compared smokers to nonsmokers and found smokers had higher temporomandibular joint disorder (TMD) pain intensity [198][199][200][201]. Further analysis of these studies found no differences in pain intensity between smokers and nonsmokers after adjustment for demographic variables [201]. The number of cigarettes was associated with pain intensity only in females [198] and females younger than 30 were more likely to develop TMD symptoms than females over the age of 30 [200]. Table 3 provides comprehensive details on the four studies in this subsection. Muscles (n = 20). Compared to bones and joints, few studies investigated the effect of tobacco smoking on skeletal muscles. The overall characteristics of these 20 studies were as follows: 11 studies enrolled both sexes, 8 enrolled only males, and 1 enrolled only females, 10 studies were quasi-experimental, 5 were cross-sectional studies, 4 were cohort studies, and 1 was a case-control study, and all studies used self-report to assess smoking habits. Table 3 provides comprehensive details on these studies which examined the effects of smoking on the anatomical, biological, metabolic, physical, and functional outcomes of skeletal muscles.

Tobacco Smoking and Skeletal
Two studies investigated the association between smoking and the anatomy of skeletal muscles. A cross-sectional study reported smokers had lower Types I and IIa muscle fibers than nonsmokers indicating smokers' skeletal muscles had oxidative fiber atrophy [204]. The prospective cohort study of only males reported rectus femoris volume (RFVOL) at baseline (prior training) was lesser in smokers than in nonsmokers, although RFVOL was significantly increased with training, and due to those authors suggested that training reversed the effects of smoking [214]. Three quasiexperimental studies investigated the biological effects of smoking. The results of those studies reported smokers had decreased local muscle O2Hb [206], thiobarbituric acid [211], and catalase [211] levels, an increase in inflammatory markers (sTNFR1) [211] and similar VO2 [206,211], lactate [206], superoxide dismutase (SOD) [211], and succinate dehydrogenase (SDH) activity [220], inflammatory cytokines (IL-6, IL-10, and sTNFR2), myoglobin concentration [220], and capillarization [220] during leg muscle exercises.
Twelve studies examined the effects of smoking on physical and functional properties of skeletal muscle: 7 studies examined muscle strength, 2 studies examined muscle thickness, and 3 focused on maximal voluntary contraction. The findings of the 3 pretest/posttest, 2 prospective cohort, and 2 cross-sectional studies on muscle strength were that smoking was significantly associated with a reduction in back extensor muscle strength [202,208], grip strength [203,215], and knee muscle strength [207]. One prospective cohort study reported that parameters of body composition and muscle strength were increased in subjects who quit smoking compared to subjects who continued smoking [218]. One study reported an inverse correlation between pack-years and muscle strength [217]. Two pretest/posttest studies were not congruent in terms of findings regarding percentage of change in muscle thickness (PCMT) and relative contribution ratio (RCR) of both internal oblique (IO) and transversus abdominis (TrA) muscles [205,209]. The first study reported PCMT and RCR were not significantly different between smokers and nonsmokers [209]; however, the second study reported significant differences between smokers and nonsmokers in regard to PCMT of the TrA and in RCR of both TrA and IO [205]. Three pretest/posttest studies of male smokers and nonsmokers had similar findings regarding the maximal voluntary contraction for quadriceps muscles [210,221], rectus abdominis, and external oblique [216]; however, maximal voluntary contraction was significantly higher in smokers than in nonsmokers [216].
Three studies (case-control, cross-sectional, and prospective cohort) investigated the interaction of smoking and obstructive lung disease on skeletal muscles. Those studies found smoking in the presence of obstructive lung disease to be significantly associated with increased muscle injury [212] and lower weight and lean mass [219]; however, smoking, regardless of patient spirometry status, was the only independent variable associated with lower quadriceps Klotho levels [213].

Tobacco Smoking and Cartilage (n = 19)
3.4.1. Knee Joint Cartilage (n = 7). Four studies (2 prospective cohorts, 1 case-control, and 1 cross-sectional) described the effect of smoking on knee joint cartilage. A cross-sectional analysis of these studies found both smoking and pack-years were positively associated with the volume of tibia cartilage [228] and femoral medial, intercondylar, and lateral cartilage [226]. There was an inverse association between smoking and cartilage strain ratio [226].There was no consensus regarding the risk of tibiofemoral cartilage defects; one study reported smokers experienced a higher risk for medial and lateral tibiofemoral cartilage defect (OR: 4.91, P < 0.05), and such risk was increased with pack-years (OR 9.90 and OR 12.98, respectively, for heavy smoking versus never smoked, P < 0.05) [225]; another study reported smoking was not associated with tibiofemoral cartilage defect [228]. Interestingly, the prospective cohort study found both smoking and packyears associated with an increased annual loss of medial but not lateral tibia or patellar cartilage [224]. There were 3 studies (2 cohorts and 1 case-control) that reported on the postoperative effects of smoking. Smokers experienced significant early meniscus repair failure (P = 0.0076) [223], less improvement in Modified Cincinnati Knee score after 2 years of autologous chondrocyte implantation surgery for full-thickness chondral defects of the knee (P < 0.05) [227], and a lower satisfaction rate after knee microfracture intervention [222]. Table 3 provides comprehensive details on the 7 studies in this subsection. (n = 12). Twelve studies examined the effects of smoking on spinal cartilage. One secondary data analysis reported smoking was not associated with disc degeneration and low back pain; however, the combination of smoking and hard physical work increased risk of vertebral inflammatory processes (OR = 4.9, 95%CI: 1.6-13.0) [236]. There were 11 cohort (10 retrospective, 1 prospective) studies of patients who underwent spinal surgery. Interestingly, these studies found smoking was significantly associated with an increased risk of reoperation [230,237,238], higher infection rates [235,239], higher risk of 30-day morbidity (P = 0.04) [239], and use of analgesic medication [234,240], but there was no consensus on spinal fusion rate, length of stay, or complication rate. Three studies reported smoking was significantly associated with a lower spinal fusion rate [232,234,235]; however, one study reported spinal fusion rate was not affected by smoking status [231]. One study reported smoking was associated with longer length of hospital stay (P < 0.001) [235], whereas the other study reported no association with length of hospital stay (P = 0.99) [233,237]. Two studies reported smoking was not associated with overall complications [229,233]. Table 3 provides comprehensive details for the 12 studies that enrolled both sexes and investigated the association of smoking with vertebral disc degeneration, pain, and the effect of smoking on spinal surgery outcomes.

Tobacco Smoking and Tendons (n = 6).
Three studies investigated rotator cuff tendons; two cross-sectional studies reported smokers presented with more advanced degenerative changes in their supraspinatus tendons (P < 0.001) [245] and reported that a higher total of smoked cigarettes was associated with the severity of rotator cuff tears (Type II versus Type I, P = 0.032) [242]. The retrospective cohort study in patients with calcified calcific tendinitis of the rotator cuff found smoking was significantly associated with a failure of needle aspiration of calcific deposits (nACD) (adjusted OR = 1.7, 95% CI: 1.0-2.7, P = 0.04) [246]. One case-control study reported smokers had significantly thinner patellar and Achilles tendons in the proximal, middle, and distal thirds region of the tendons and significant lower strain ratio measurements in the same regions (P < 0.05); packyears were inversely related to patellar tendon thickness (P < 0.05) [241]. One cross-sectional study reported smokers had significant improvement in finger range of motion over nonsmokers after tendon grafting [243]. Finally, one prospective study reported that the Constant score was significantly lower in smokers than nonsmokers 1 year postoperatively after rotator cuff reconstruction (71 versus 75, P = 0.017) [244]. Table 3 provides comprehensive details on the 6 studies that investigated the effects of smoking on the anatomical or functional characteristics of tendons. (n = 4). Compared to nonsmokers, smokers were found to have significantly poorer outcomes regarding stability [249], Lysholm Knee Score, International Knee Documentation Committee (IKDC) subjective score, and IKDC objective grade [249,250] after ACL reconstruction. A dose-dependent association was noted between pack-years and postoperative anterior translation (P = 0.015) and IKDC objective grade (P = 0.002) [250]. Tobacco use was associated with a significantly increased risk of postoperative venous thromboembolism (OR = 1.9; P = 0.035) [248] and subsequent ACL reconstruction (OR = 1.7; P < 0.0001) [248]; but it was not found to be significant for postoperative stiffness (OR = 0.9; P = 0.656) [248]. There was no consensus regarding risks of postoperative infection. One study reported tobacco use increased risk infection (OR = 2.3; P < 0.0001) [248], and another study reported smoking was not a significant risk factor (OR=2.5; P = 0.167) [247]. Table 3 provides more details on the 4 cohort studies that enrolled both sexes and investigated effect of smoking on the clinical outcomes and complications after ACL reconstruction. Musculoskeletal System (n = 8). These studies were classified into two groups: first group investigated associations between secondhand smoke and musculoskeletal system disorders, and the second investigated the effects of intrauterine exposure by mothers who smoked or mothers exposed to secondhand smoke and the long-term outcomes on the musculoskeletal system of offspring. The first group had two cross-sectional studies that reported subjects exposed to passive secondhand smoking had significantly lower phalangeal BMD (P < 0.01) [253] and higher risk for femoral neck osteoporosis (OR, 3.68; 95%CI: 1.23-10.92) than unexposed subjects [255]. The second group consisted of 6 studies; 3 studies were cohort studies and reported maternal smoking was significantly associated with lower aerobic fitness of male adolescents [252] and lower total body BMC in male adolescents, but not female adolescents [257]. Maternal smoking was not found to be associated with BMD [254,257] or fractures in adolescents [254]. One prospective cohort and one cross-sectional study reported smoking by both parents during pregnancy had a significant effect on relative leg length (shorter) of offspring at ages 7-10 [258], increased spine BMC, and BMD in girls, but not boys at a mean age of 9.9 years [256]. Finally, one article, a secondary data analysis, reported exposure of nonsmoking pregnant mothers to secondhand smoke from paternal grandmothers was associated with taller girls, and greater bone and lean mass of both sexes at age 17, while exposure of nonsmokers pregnant mothers to secondhand smoking from maternal grandmothers was associated with increased weight of boys at age 17 [251]. Table 3 provides comprehensive details about the 8 studies.

Discussion
This systematic review provides evidence of the substantive negative effects of tobacco smoking on the musculoskeletal system. A majority of studies reviewed (132 of 243) focused on the deleterious effect of tobacco smoking on bones, followed by joints (54 of 243), with less emphasis on muscles, cartilage, tendons, and ligaments. At the bone level, there is sufficient evidence demonstrating tobacco smoking is associated with low BMD, an increased likelihood of fracture, delayed fracture healing, increased alveolar bone loss, increased risk of periodontitis, increased peri-implant bone loss, and implant failure. The inverse association for tobacco smoking with BMD was evident in males across ages, in adolescents of both sexes, and in postmenopausal females; however, these associations were not fully explained by biomarkers monitored to understand the mechanisms of smoking effects on bone metabolism. Studies investigating biological mechanisms were few and were limited by a lack of power or a failure to adjust for confounding variables.

Journal of Environmental and Public Health
The research on periodontitis provided general agreement on clinical outcomes and monitored biomarkers proposed to be affected by tobacco smoking. This may indicate different mechanisms for effects on the alveolar bone than the rest of the body or there may be factors other than smoking that have an isolated effect or interact with the effects of smoking to synergize or diminish the negative effect on human BMD. Our findings regarding the negative outcomes of smoking on bone implants and surrounding tissues were consistent with 5 systematic reviews conducted previously on smoking effects on dental implants [7][8][9][10][11].
The research on joints was more segregated and investigated the effects of smoking on specific joint disorders rather than whole joints. Most studies were focused on RA (29 of 54). There was consensus that smoking is associated with increased disease activity, functional disability, and poor response to therapy. There was also evidence of an interaction between smoking and HLA-DR beta 1 4-amino acid haplotype and ACPA. This interaction was significantly associated with increased RA disease activity. There was evidence of a negative effect on OA outcomes; however, findings were inconsistent. There was evidence that supported the association between smoking and increased pain in patients with TMD, increased disease activity, pain, and poor response to therapy in patients with spondylarthrosis in a pattern similar to that exhibited by patients with RA.
Studies of smoking effects on skeletal muscles provided clear evidence that smoking was associated with poor outcomes, particularly decreased muscle strength. However, these findings were inconsistent on whether smoking is associated with changes in muscle thickness or maximal voluntary contraction. For the effects of smoking on cartilage, this review provided evidence of a harmful association of smoking and pack-years with knee cartilage (increased in cartilage volume, decreased strain ratio, and poor postoperative outcome), low spinal fusion rate, and increased risk of spinal reoperation. A limited number of studies investigated the effect of smoking on tendons and ligaments. For tendons, the results found smoking and pack-years associated with thinner patellar and Achilles tendons, severe rotator cuff tears, and poor postoperative functional outcomes. The results of this review regarding the effect of smoking on tendons are consistent with the findings reported in a previous systematic review by Bishop et al. [16]. For ligaments, smoking was associated with poor functional and stability scores after ACL reconstruction, consistent with the findings in systematic reviews by Kanneganti et al. [13] and Novikov et al. [14] that reported negative effects of smoking on postoperative outcomes.
This systematic review identified few articles on secondhand (4 studies) or intrauterine exposure to smoke (6 studies). In terms of secondhand smoke, reviewed studies reported varying effects. There was an inverse association between secondhand smoke exposure and phalangeal BMD and a positive association with risk of femoral neck osteoporosis. A positive effect reported was the reduction in OA risk, and no effects were found in relation to response to RA therapy. Studies on intrauterine exposure were focused on long-term effects of exposure on the musculoskeletal outcomes of offspring. There was no consensus in the evidence on the effects of smoking on BMD, BMC, relative leg length, or other body composition parameters in male and female offspring.
This systematic review provided evidence of the negative effects of smoking on the musculoskeletal system. Table 3 provides essential information to be considered in future studies. Definitions of smoking status and intensity of smoking based on self-report were inconsistent across studies. For example, when smoking status was treated as binomial category (smoking versus nonsmoking), one study may have added former smokers to the smoking category, while another may have placed former smokers in the nonsmoking category. Similar observations were noted regarding measurement of smoking intensity; one study may define a heavy smoker as an individual who smoked 10 cigarettes or more a day over the last 10 years, while another study may define a heavy smoker as an individual who smoked 10 or more cigarettes a day over the last 5 years. Such variations in classification may lead to a misinterpretation of overall smoking effects and introduce inconsistencies among reported findings. Objective measurements are considered more reliable assessments of smoking exposure; however, objective measurements were reported in only 12 of 243 studies, 9 of these studies measured level of cotinine, and 3 assessed levels of EXCO. We did not encounter any study in this review that measured nicotine dependence (e.g., the Fagerstrom Test for Nicotine Dependence), so we cannot conclude if there were differences in patterns of effects for use or dependence of tobacco smoking. There were a limited number of studies investigating the effects of secondhand smoke and other smoking modalities, such as hookahs, narghiles, or electronic cigarettes. This review encountered only 4 studies on secondhand smoking with varying effects reported. We also only encountered only 2 studies regarding waterpipe smoking and no studies investigated the effects of smoking electronic cigarettes on the musculoskeletal system. This systematic review demonstrates the need for further research to understand the effects of smoking on the musculoskeletal system. Due to the limited evidence on muscles, cartilage, tendons, and ligaments, more studies using different research designs are needed. The need for studies using various research designs naturally extends to study the effects of waterpipe, electronic cigarettes, and secondhand smoke on the musculoskeletal system. Longitudinal observational and experimental studies are needed to conclusively understand the effects of smoking on bone and joint-related outcomes.
There are several factors to be considered in the design of future studies. There needs to be a consistent approach to evaluate self-reported exposure for smoking and a more objective assessment for smoking exposure. There is also a need to assess for other smoking products (polycyclic aromatic hydrocarbons, nitrosamines, etc.) rather than to assess only levels of cotinine or EXCO. There need to be more information and analysis of confounders and genetic factors that may interact with smoking effects on the musculoskeletal system. There is also a need for more and frequent monitoring for changes in smoking status in longitudinal studies. Future research should endeavor to examine more than one musculoskeletal component to shed light on the development and differentiation of cell types of the musculoskeletal system. Finally, there is a need for further research to provide insight into how to minimize the effects of smoking in patients who undergo musculoskeletal surgery. Also, we recommend more ancillary studies as part of large longitudinal studies such as Population Assessment of Tobacco and Health (PATH).
Our research had several limitations. First, this review included only English language articles. Second, in vitro studies were not included. Third, we did not search for specific diseases or disorders under the term musculoskeletal system; however, we believe our search was able to comprehensively capture these disorders that are subcategorized under each section in Table 3. Fourth, this review was focused primarily on the effect of tobacco smoking on musculoskeletal system, and due to that the search method in this review may have missed other studies that have smoking as variable. Finally, we did not find any research on electronic cigarettes with only 2 studies on waterpipes. We believe these areas warrant research and will likely attract attention in the future.

Conclusion
This systematic review provided clear evidence of the negative effects of smoking on the musculoskeletal system. Evidence found smoking associated with lower BMD, and increased risk for fracture, periodontitis, alveolar bone loss and implant failure, increased joint disease, poor functional outcomes, and poor therapeutic response. We also found evidence of adverse effects on muscles, tendons, cartilage, and ligaments, despite the scarcity of studies. As smoking continues to be an important public health concern, there is a need for further research to understand mechanisms of action for the effects of smoking on the musculoskeletal system and to increase awareness of healthcare providers and community members about the deleterious effects of smoking on the musculoskeletal system.