The effect of smoking on bone healing

Search strategy

Five electronic databases were searched: MEDLINE, Web of Science, The Cochrane Library, SCOPUS and EMBASE. All databases were searched from their earliest records until August 2012. The full search strategy developed for MEDLINE is shown in Table I. This search strategy was customised according to the database being used. The searches were restricted to English language publications.

Study inclusion and exclusion criteria

Prospective and retrospective clinical studies assessing bone healing in smokers and non-smokers were included. The studies involved clinical interventions, ranging from conservative treatment such as cast immobilisation to surgical interventions, for example internal fracture fixation. In addition, arms of clinical trials comparing different interventions that reported results separately for smokers and non-smokers were included. The following inclusion criteria were applied: 1) publications written in the English language; 2) human studies; and 3) studies categorising subjects into at least two groups (non-smokers and smokers).

Studies were excluded if they contained inadequate data to allow a clear comparison of bone healing in smokers and non-smokers following treatment.

Outcome measures

Included studies needed to report one or more of the following primary outcomes, which were: 1) clinical indicators of bone healing (including measures of time to clinical union and diagnoses of non-, delayed or malunion); and 2)radiological indicatorsof bone healing (including measures of time to radiologically defined union).

Secondary outcomes were: 1) complications of bone healing (including diagnoses of infection or osteomyelitis); and 2) patient based-outcomes (when not used as part of a clinical indicator for bone healing), which included pain and functional status.

Study selection, data collection and analysis

In the first phase of study selection, the titles and abstracts of all identified publications were independently screened by two reviewers (RAP and RMP). Disagreement between reviewers was resolved by discussion.

The full texts of all studies of possible relevance were obtained for independent assessment against the inclusion criteria stated. Studies rejected at this stage were recorded in a rejection table with reasons for rejection. Again disagreement between reviewers was resolved by discussion.

Inter-reviewer agreement was calculated with Cohen’s κ score for each screening stage. Where possible, authors of studies were contacted to resolve doubts about study design, patient populations and to request missing information.

A data extraction form was used to collect information from the included studies. The data collected included study characteristics, outcome measures, treatment characteristics, results, quality assessment data and other general information.

Quality assessment

The quality of the included studies was assessed according to: 1) similarity of the smoking and non-smoking groups at baseline; 2) masking of smoking status to the clinician(s) assessing the outcome(s); 3) reproducibility of the outcome measure(s); and 4) completeness of follow-up and explanations for dropouts. Studies with a retrospective design were not assessed with reference to completeness of follow-up.

When all criteria were met the risk of bias was estimated as low. A moderate risk of bias was assigned when one or more criteria were partly fulfilled and a high risk assigned when one or more criteria were not met. The quality assessment was not used to exclude any studies qualifying for the review on the basis of their inclusion criteria.

Data synthesis

A summary table was constructed using information collected on the data extraction forms. The pooled data was analysed in a descriptive format. Studies were analysed for similarities and suitability for meta-analysis.

Characteristics of study designs and settings

Smoking was reported as a primary focus in all the studies. One study was multi-centred with eight participating centres.⁷ Six studies were conducted within single institutions.^8-13 Two studies were unclear as to whether they were multi-centred or conducted in single institutions.^14,15

Two studies^14,15 were published by the same authors and included similar patients treated at the same institution at overlapping times. After contacting the authors it was established that there was an overlap in the patients used between these two studies.

Characteristics of participants

One study¹⁵ only included male patients, whereas all others included both male and female patients. The mean age of participants was provided in six studies.^{7,9,10,12,14,15} This ranged from 33 years⁹ to 53 years.^14,15 Schmitz et al¹¹ reported the mean age of the smoking group as 35.6 years and the non-smoking group as 35.8 years. Adams et al⁸ reported the mean age of smokers was 38.7 years and non-smokers was 39.2 years. One study reported the mean age seperately for men and women (40.3 years and 43.7 years, respectively).¹³

One study⁷ used a subgroup of patients who were participating in a lower extremity assessment project. This project was a cohort study of lower extremity trauma patients who were at risk of amputation.

The number of patients in the studies ranged from 33¹² to 273.⁸ The number of smokers in the studies ranged from 13¹² to 140.⁸

The studies categorised patients based on a self-reported smoking history, but the definition of ‘smoking’ varied. Six studies were unclear on their reported definition of a smoker,^7,10,12-15 but more definite descriptions included > five cigarettes per day,¹¹ ≥ ten cigarettes per day,⁸ or the use of any amount of tobacco in the form of cigarettes, a pipe or cigars.⁹ The two studies by W-Dahl and Toksvig-Larsen^14,15 clearly defined non-smokers as patients who at a pre-operative examination stated they had never smoked or had stopped smoking > six months ago.

Characteristics of clinical state and intervention

Six studies investigated fractures of the tibia^7-12 and three were based on tibial osteotomies.^13-15 Of the fracture studies, three included open and closed fractures,^10-12 whereas the remaining three studies^7-9 included only open fractures.

Schmitz et al¹¹ investigated closed and Gustillo Grade I open fractures, whereas Castillo et al⁷ investigated only Grade III open fractures. Harvey et al⁹ included only Gustillo Grade II, IIIA and IIIB fractures.

Three studies^9,11,12 investigated fractures in the tibial shaft while Ristiniemi et al¹⁰ investigated distal tibia fractures (defined as fractures within 5 cm of the ankle joint).

In two studies.^8,11 fractures were treated by intramedullary rod fixation, external fixation or cast immobilisation. Fractures in Harvey et al⁹ were treated with external fixation or intramedullary nailing. In one study¹⁰ fractures were treated with two-ring Ilizarov hybrid external fixation. One study⁷ did not state the form of fracture stabilisation used.

Characteristics of outcome measures

The majority of studies investigating tibia fractures assessed union based on clinical and radiological data.^7-10 In these four studies fractures were described as united when there was radiological evidence of bridging of cortices and there was no pain at the fracture site upon weight bearing. One study¹¹ defined and measured clinical union and radiological union separately. Alemdaroglu et al¹² assessed consolidation based on radiographic measures only.

All studies investigating tibia fractures described in some format the time to union for smokers and non-smokers. Several studies described an average time to union while one study⁹ grouped the time to union into three different time periods (< six months, six to nine months and > nine months). One study¹⁰ did not provide data on time to union between smoking and non-smoking groups, but included the results from a univariate analysis of the difference.

Harvey et al⁹ also described union rates between smokers and non-smokers. Four other studies investigating tibia fractures described the proportion of delayed unions or nonunions in the smoking and non-smoking groups.^7,8,10,11

Two studies based on tibial osteotomies presented data on the mean time in external fixation.^14,15 The time in external fixation was once again based on both a clinical and radiological assessment of healing. Both studies also described the incidence of delayed healing amongst smokers and non-smokers. Meidinger et al¹³ reported the percentage of nonunions and consolidations.

Secondary outcomes

Three studies reported on infection as a secondary outcome. These studies documented the incidence of infection amongst smokers and non-smokers, defined as deep infection in two studies,^8,9 and osteomyelitis in one study.⁷

Length of follow-up

The exact length of follow-up was not clearly specified in several studies.^9,12-15 Schmitz et al¹¹ described follow-up until complete healing or for at least a year. Adams et al⁸ described follow-up until fracture union or clinical intervention for non-union was required, this study had a mean follow-up of 21.6 months. Ristiniemi et al¹⁰ document a mean follow-up time of 41 months, while Castillo et al⁷ had a mean follow-up of 24 months. Alemdaroglu et al¹² described follow-up for at least six months after union.

Conclusions drawn on the effect of smoking on bone healing

Details of the conclusions drawn on the effect of smoking on bone healing for each tibia study are presented in Table II.

Five of the nine studies concluded that smokers took significantly longer to heal than non-smokers.^{8,10,11,14,15} Two of these studies also noted that smokers were more likely to have delayed healing.^14,15 These two studies defined delayed healing as > 16 weeks in external fixation.

Castillo et al⁷ did not present a direct stastistical analysis of the time to fracture healing between smokers and non-smokers. The study however reported that non-smokers are more likely to heal by 24 months than smokers (p = 0.01).

Meidinger et al¹³ reported the percentage of smokers in the nonunion and consolidation groups. They reported a significant difference in the the percentage of smokers in the nonunion group compared with the percentage in the consolidation group (p < 0.05).

Harvey et al⁹ did not describe the average time to union, instead results were presented as four different healing groups; timely union, delayed union, late union and un-united. The study reported that more smokers had late union or un-united fractures than non-smokers. The union rate was not significantly different between the smokers and non-smokers (p = 0.10).

Adams et al⁸ found a non-statistically significant difference in the percentage of smokers and non-smokers with nonunions. Alemdaroglu et al¹² noted no significant difference in the healing time for smokers and non-smokers treated by circular external fixation.

Overall all tibia studies except one¹² reported a negative effect of smoking on bone healing.

One study,⁷ concluded that smokers were more likely to develop osteomyelitis. Harvey et al⁹ noted no significant difference in the incidence of deep infections in smokers and non-smokers. Adams et al⁸ did not include a statistical analysis of the incidence of deep infections among smokers and non-smokers.

Risk of bias in included studies

Details of the quality assessment for each included tibia study is presented in Table IV. Fracture characteristics among other things were noted not to be significantly different between smokers and non-smokers in three studies.^8,9,11 It was unclear in six studies^7,10,12-15 whether the smoking and non-smoking groups were comparable at baseline. Although two studies^14,15 provided baseline data on the mean age and BMI of the smoking and non-smoking groups, a statistical analysis to assess a significant difference between the two groups was not reported.

Femoral fractures

Giannoudis et al¹⁶ presented a retrospective study aimed to assess the factors that affected union of the diaphysis of the femur by comparing 32 patients with nonunion and 67 patients with united fractures. The fracture union and nonunion groups were comparable regarding gender, Injury Severity Score and soft-tissue injury. The assessment of healing was poorly described, but the definition of nonunion was based on clinical and radiological criteria. Overall the study concluded that smoking was not a statistically significant factor for nonunion of the femoral diaphysis.

Femoral distraction

Kenawey et al¹⁷ studied 35 patients treated with intramedullary femoral lengthening at a single institution. This prospective study assessed regenerate failure. Insufficient bone regenerate developed in eight cases, three of whom were smokers. Normal regenerate developed in 29 cases, of which two were smokers. It was concluded that smoking was associated with a higher risk of insufficient bone regeneration.

Fibular fractures

One study investigated fibular fractures.¹⁸ This study retrospectively examined 52 patients with minimally displaced isolated fibular fractures treated by either internal fixation or cast immobilisation. Healing was based on radiological assessment. The study concluded that smokers displayed an increased time to radiological bone healing compared with non-smokers (p = 0.034).

Ulna osteotomy

One study was conducted on patients who had undergone an elective procedure on the ulna.¹⁹ This retrospective study involved ulna-shortening osteotomies on 39 patients to treat ulna impaction syndrome. Union was defined by radiological assessment and a clinician masked to the patient’s smoking status interpreted all radiographs. One smoker underwent bilateral procedures and bilateral results were included. Smokers had a longer time to union and higher incidence of delayed union or nonunion. The mean time to union was 7.1 months in smokers and 4.1 months in non-smokers (p = 0.016), and 30% of smokers experienced delayed union or nonunion compared with none of the non-smokers (p = 0.02).

Subtalar arthrodesis

Chahal et al²⁰ investigated subtalar arthrodesis. This multicentre study was carried out at two hospitals. Radiological outcomes were independently assessed by musculoskeletal radiologists. The results showed that smokers were 3.8 times more likely to have a nonunion than non-smokers. The study concluded that smokers had a significantly lower rate of union after subtalar arthrodesis.

Ankle arthrodesis

Two retrospective cohort studies investigated arthrodesis of the ankle.^21,22 The majority of patients in the study by Collman et al²¹ were diagnosed with post-traumatic ankle arthritis. As the study was based on a small patient population (11 smokers and 28 non-smokers), analysis of results was limited to observational trends. Smokers attained union in almost all cases and did not show a trend towards nonunion. In the study the term ‘smokers’ was poorly defined, and may have included individuals with a previous smoking history. The authors also acknowledge that the outcome analysis was limited by a small patient population. Non-significant differences between the groups may have been due to the small study size rather than a lack of association.

Perlman and Thordarson²² studied 67 fusions in 61 patients. The rate of nonunion was higher in the smoking group (32.5% of smokers vs 22% of non-smokers), but statistical significance was not reached. However, p-values were not reported in the study, and the results from bilateral fusions in the same patient were included.

Elective foot surgery

Krannitz et al²³ investigated the effect of cigarette smoking on radiological bone healing after elective Austin bunionectomies with internal screw fixation. The study included 17 self-reported smokers and 17 non-smokers. This study was unique among those included in this review in that it confirmed smoking status with a urine cotinine test. The study concluded that smokers displayed an increased time to radiological bone healing compared with non-smokers (p < 0.001). The urine cotinine level in the smoking group was highly correlated with prolonged bone healing.

Meta-analysis of the data presented in the studies

A meta-analysis was not possible for any of the orthopaedic studies, due to heterogeneity in study design, data collection and data presentation.

Analysis of study designs

It is important to note that the effect of smoking was reported as a primary focus for all the included studies. However the studies were less clear in regards to their design. It could not be established whether one study conducted by Risiniemi et al¹⁰ was prospective or retrospective. The study by Adams et al⁸ was partially prospective and partially retrospective. Overall the studies were predominantly retrospective and cohort in design.

Analysis of study participants

The number of patients in the studies ranged from 33 to 273.¹² The only study to report no negative effect of smoking on bone healing was also the study with the smallest sample size,¹² and the authors acknowledged that this may have prevented a statistical difference in healing times from being recorded.

The conclusions drawn in this review are based on studies of middle-aged adults. It is however interesting to note the conclusions of Rajan et al.²⁴ Their study retrospectively reviewed patients who were found to have a prolonged bone healing index (BHI) after limb deformity corrective surgery. The study demonstrated that BHI was increased in both active and passive smokers. The 17 smoking patients in their study comprised 16 adolescents and one nine-year-old, an age distribution substantially different to the studies included in this review.

The definition of a smoker varied significantly between the studies in this review, ranging from any history of tobacco use to thresholds of ten or 20 cigarettes a day. A surprisingly large proportion of studies (eight of the 17) did not define a smoker in their publications.

All of the included studies categorised patients based on self-reported smoking status. Krannitz et al²³ acknowledged in their discussion that smoking status could be misclassified when based on statements by the patient, and that patients could mislead physicians about their tobacco use. Lores Obradors et al²⁵ investigated patients attending a clinic for respiratory medicine, and demonstrated that 21 of 125 patients (17%) smoked while denying doing so.

Scott, Palmer and Stapleton²⁶ identified that most studies rely on patient reported smoking habits, which can be unreliable. They recommended biochemical analysis as the gold standard of assessing smoking status. Biochemical assessment of current smoking status can be achieved by measurement of systemic levels of cotinine,^23,27 which provides the most accurate assessment of the level of exposure. It is also important to record the number of years a patient has smoked to provide an estimate of pack-years, as the duration and magnitude of smoking exposure will have a greater impact on healing.²⁶ Misclassification of the smoking status of patients would tend to favour smokers being included in the non-smoking group, which would reduce the chances of finding an effect of smoking upon outcomes. This may also be affected by inclusion of former smokers in the study.

The study by Krannitz et al²³ was the only one to confirm smoking status with a biochemical analysis. A urine cotinine test was performed on patients pre-operatively and during the post-operative healing period. It involved a urine dipstick, with the strip changing colour depending on the concentration of cotinine in the urine. Results showed a correlation between the concentration of cotinine in the urine and the time to healing.²³

Finally the participants also differed between trials in terms of their initial clinical diagnoses.

Analysis of interventions

The majority of studies were based on fractures, and studies that included revision surgery were excluded. The tibial fractures were treated either by internal fixation, external fixation or cast immobilisation. Internal fixation was achieved by different methods.

Several studies reported on elective procedures, such as osteotomies. These procedures are good models to assess bone healing as the surgical procedure performed is standardised and weight-bearing can be assessed at regular intervals.

Analysis of outcome measures

The assessment of union was similar in the studies dealing with tibial fractures, based on both clinical and radiological assessment. The other studies commonly defined union through radiological assessment only.

Schmitz et al¹¹ when investigating the effect of smoking on tibial shaft healing, only included closed and Grade I open tibial fractures. The reason for excluding Grade II and III open fractures was to avoid confounding variables such as the severity of soft-tissue damage and the timing of wound coverage. A review in the medical literature has noted that smokers presenting with an open tibial fracture will in particular suffer the negative effects of their smoking behaviour, because these fractures also involve significant soft-tissue injury.²⁸ This demonstrates that healing of bone should not be considered as an independent process and undoubtedly events within the soft tissues influence bone healing.

Overview of conclusions

The findings of this review are in agreement with the recent literature review by Al-Hadithy et al.¹ This paper concludes that smoking has a significant effect on fracture union, particularly in tibial shaft fractures, spinal and foot and ankle fusions. The delay in union was reported to be more apparent in those cases requiring bone grafts, as there is an increased chance of devascularising the graft. The paper suggests smokers have a 40% increased time to union and chance of nonunion compare with non-smokers.

Another recent study of note is that of Moghaddam et al.²⁹ This study was excluded as it combined the results of smoker and ex-smokers, however, it demonstrated similar finding to that of this review. Current and previous smokers exhibited a significantly higher proportion of delayed union and nonunion of tibial shaft fractures. This difference was highly significant (p = 0.0007), which indicated that the time that was necessary for bone healing was significantly increased (p = 0.0008).

Overview of quality assessment

Three of the 17 studies reported on the degree of difference at baseline in the smoking and non-smoking groups. This information, required to establish whether the groups were comparable, was provided in less than half of the included studies.

The included prospective studies had a high level of follow-up and generally all dropouts were accounted for. The reproducibility of outcome measures was poorly reported. Also in general the smoking status of the patient was not concealed from the examiner. Due to the less than ideal methodological quality of the included studies, the results presented in this review should be interpreted with caution.

In several studies bilateral defects were treated, and these were managed statistically in different ways. Schmitz et al¹¹ randomly excluded a tibia from bilateral patients in their study in order to avoid statistical dependence between the observations. In comparison, Harvey et al⁹ reported on a cohort of 105 patients and analysed data from 110 tibial fractures. Bilateral fractures were included in this study, and data analysed based on fractures rather than patients.

Overview of the effect on ex-smokers

One study considered the effect on bone healing on previous smokers. Castillo et al⁷ reported that previous smokers were at increased risk of delayed union after tibia fractures, but their risk was not as great as current smokers.

Conclusion

Smoking negatively influences healing of the tibia. It is difficult to draw conclusions on the effect of smoking on bone healing in general. The evaluation of other orthopaedic studies provided less definitive conclusions and did not strengthen the findings seen in the tibia studies. Differences in study designs, methods of measuring bone healing and presentation of data precluded a complete pooling of data for a more robust analysis of all the information.

Clinical implications

The review strengthens our knowledge of the impact of smoking on bone healing. The reduced bone healing potential in smokers, suggests that smoking cessation advice should be offered to smokers before elective procedures. This review however has not investigated the effect of smoking cessation on bone healing potential, and so the advantages can only be speculated.

Implications for future research

One recurring problem in this review was the variability in study design and outcome measures. It would be recommended that future studies record time to healing and present data transparently using mean values with a measure of the spread of data.

Studies evaluating the effect of smoking on treatment response should be based on more reliable methods of assessing smoking exposure rather than sole reliance on patient-reported data. Other methods include the measurement of exhaled carbon monoxide or assessment of salivary/serum levels of cotinine.