The Impact of Tobacco Cigarette Smoking on Spinal Cord Stimulation Effectiveness in Chronic Spine–Related Pain Patients

CHRONIC AND INTERVENTIONAL PAIN ORIGINAL ARTICLE

Nagy Mekhail, MD, PhD,* Gerges Azer, MD,* Youssef Saweris, MD,*† Diana S. Mehanny, MD,* Shrif Costandi, MD,* and Guangmei Mao, PhD‡

According to the US Centers for Disease Control and Pre- Background and Objectives: Despite the observation that select nic- vention, smoking remains the single most preventable cause otine receptor agonists have analgesic effects, smokers report higher pain of death in the United States.12 Although efforts to reduce the scores and more functional impairments than lifelong nonsmokers, attrib- number of smokers in the United States have been successful, utable to exaggerated stress responses, receptor desensitization, and altered accounting for an overall decrease of smokers from 20.9% to pharmacokinetics compounded by accelerated structural damage result- 18.1% from the year 2005 to 2012,13 a rising trend in the rate of ing from impaired bone healing, osteoporosis, and advancement of disk smoking among chronic pain patients has been observed. Preva- disease. We hypothesized that smoking diminishes the analgesic response lence of smokers in adults with chronic pain has increased from – to spinal cord stimulation (SCS) in patients with chronic spine related 24.2% to 25.7% to 28.3% corresponding to the years 2000, pain conditions. 2005, and 2010, respectively.14 Of interest is the observation that Methods: A retrospective cohort study was performed at Cleveland smokers with chronic pain also have a tendency of consuming Clinic by collecting and assessing data of 213 patients who had been more opioids than former smokers and lifelong nonsmokers.14–16 implanted with SCS for spine-pain indications. History of tobacco smoking In this study, we focus primarily on the spinal cord–related was subcategorized into 3 categories: past (former smoker), present (cur- chronic pain, treated using spinal cord stimulator (SCS), and the rent smoker), or those who had never previously smoked (lifelong non- effect of smoking on the success of this treatment modality. smokers), and a multivariable linear regression was run to measure the Limited information is known about how smoking frequency correlation, if any, between smoking status and numerical rating scale pain and duration impact chronic spine–related pain disorders and their score. In addition, opioid consumption at baseline and 12-month follow- therapeutic options. Generally speaking, there is evidence to sup- up, expressed in milligram oral morphine equivalents, was collected port the association of tobacco use with increased risk of surgical and compared. infection17–19 and delayed wound healing.20–23 On a more specific Results: Adjusted for differences, at 1-year follow-up, current smokers level, focusing on the effect of tobacco on SCS, only 2 studies are (n = 62) reported numerical rating scale pain score of 7.0, which is 1.93 availabletodate.24,25 Although both studies are composed of a lim- (P < 0.001) and 1.32 (P = 0.001) points higher than those of lifelong non- ited sample size (<60 patients), both found an association between smokers (n = 77) and former smokers (n = 74), respectively. Opioid intake smoking and an increased chance of lead migration and revision. was2.4timeshigher(P = 0.004) in smokers than in lifelong nonsmokers. Because of the lack of evidence on the effect of tobacco Conclusions: Among our SCS-implanted sample, a positive correlation smoking on the efficacy of the SCS in terms of pain relief and was observed between tobacco use and degree of pain reduction as early opioid use, utilizing a large sample size, we aimed to further in- as 12 months postimplant; this was evident by the reported higher pain vestigate the correlation, if any, between 3 different smoking scores and opioid use in current smokers in comparison with former categories diagnosed with chronic spinal pain in a retrospective smokers and lifelong nonsmokers. cohort study. Our primary outcome was the degree of change in (Reg Anesth Pain Med 2018;43: 768–775) pain, and our secondary outcome was changes in opioid utilization compared among the 3 groups. We hypothesized that smokers, eculiarly, chronic pain is a more common occurrence than and possibly former smokers, will demonstrate worse pain control Pexpected with a once estimated prevalence of 46.5% among and higher opioid utilization than lifelong nonsmokers. the general population.1 In the United States alone, 116 million Americans are believed to suffer from some form of chronic pain.2 METHODS One common association to chronic pain is tobacco smoking. Despite the observation that smoking may provide analge- Study Population sic effects,3–5 however, smokers have previously reported higher pain scores6,7 and more associated functional impairments than Following institutional review board approval, a retrospec- lifelong nonsmokers.8–11 tive data collection was performed for all patients with chronic spine–related pain who were managed with SCS at Cleveland Clinic Pain Management Department between the years 1997 From the *Evidence-Based Pain Management Research, †Anesthesiology Insti- and 2013. Among the 986 patients who had undergone a suc- tute, and ‡Outcomes Research and Department of Quantitative Health Sciences, cessful SCS trial defined as pain relief of 50% or greater and Cleveland Clinic Foundation, Cleveland, OH. Accepted for publication April 25, 2018. had subsequently received SCS implant, 330 patients were identi- Address correspondence to: Nagy Mekhail, MD, PhD, Cleveland Clinic/C25, fied as having chronic spine–related condition per protocol in- 9500 Euclid Ave, Cleveland, OH 44195 (e‐mail: [email protected]). clusion criteria, of which 117 patients were excluded because The authors declare no conflict of interest. of incomplete records. Therefore, the final analysis included Copyright © 2018 by American Society of Regional Anesthesia and Pain Medicine 213 patients, who were further classified into current smokers ISSN: 1098-7339 (62 patients), former smokers (74 patients), and lifelong non- DOI: 10.1097/AAP.0000000000000870 smokers (77 patients) as illustrated in Figure 1.

768 Regional Anesthesia and Pain Medicine • Volume 43, Number 7, October 2018

FIGURE 1. Flowchart of patient selection. Chronic spine–related pain conditions included post- For comparison and analysis purposes, history of tobacco laminectomy syndrome, cervical/lumbar degenerative disk disease, smoking was subcategorized into 3 categories: past (former cervical/lumbar radiculopathy, cervical/lumbar spondylosis, and smokers), present (current smokers), or those who had never cervical/lumbar spinal stenosis. previously smoked (lifelong nonsmokers). For current and former smokers, the duration of smoking in years and the number of packs smoked yearly along with whether they smoked cigarettes Measurements or cigars were documented. In addition, for former smokers, the Patient's demographic data, smoking status, body mass index number of years of smoking abstinence was also recorded. (BMI), duration of pain, and history of diabetes, disability, and depression were manually obtained from the electronic medical Statistical Analysis records (Table 1), in addition to reported numerical rating scale The patients with missing smoking status, pain scores, and (NRS) and opioid utilization, if any, at baseline and 12 months opioid use at 1-year follow-up were excluded from this study. after SCS implantation. All opioid doses collected were converted Baseline characteristics were compared among 3 study groups into their morphine sulfate (MSO4, mg) equivalent using their using standard descriptive statistics. Categorical data are pre- equianalgesic dosing shown in Table 2.26 sented as number and percent of total. All baseline variables listed All patients checking in for an appointment at Cleveland in Table 1 except for smoking history were used for adjustment in Clinic, once seated in the examination room, are required to all subsequent analyses. answer a set of standardized questions including, but not limited As for the primary hypothesis, the multivariable linear to, NRS pain score and smoking status and frequency. For NRS regression model was used to assess the association between pain score, patients are asked to report their average pain intensity smoking status (current smokers vs never smokers) and pain relief for the last 24 hours on a scale of 0 to 10, with 0 being no pain, after 1 year of SCS therapy. The primary outcome was pain score 5 corresponding to moderate pain, and 10 being the worst pain after 1 year, and the predictor was smoking status. The results imaginable.27 Smoking status is also collected as part of routine were adjusted for baseline pain score, baseline opioid consump- documentation with lifetime nonsmokers documented as such. tion, age, sex, BMI, duration of pain, spine-related pain diagnosis, Former smokers are further asked to provide an estimated quit and diabetes, disability, and depression history. The mean differ- date (of which the years of abstinence were calculated) and num- ence in 1-year pain scores between current smokers and lifelong ber of smoking years prior to quitting. Finally, smokers are further nonsmokers along with 95% confidence interval (CI) was reported. asked to disclose the number of cigars/cigarettes, on average, that For the first secondary analysis, we used linear regression to they smoke per day. Both smokers and former smokers are asked assess if smoking cessation (former smokers vs current smokers if they smoke(d) cigars or cigarettes. and former smokers vs lifelong nonsmokers) was associated with

Copyright © 2018 American Society of Regional Anesthesia and Pain Medicine. Unauthorized reproduction of this article is prohibited. Mekhail et al Regional Anesthesia and Pain Medicine • Volume 43, Number 7, October 2018

TABLE 1. Descriptive Statistics of 3 Study Groups

Factor Current Smoker (n = 62) Former Smoker (n = 74) Nonsmoker (n = 77) Overall P* Sex (male), n (%) 30 (48) 34 (46) 30 (39) 0.499 Age, mean ± SD, y 50 ± 11 57 ± 12 54 ± 11 0.001 BMI, mean ± SD, kg/m2 28 ± 6 30 ± 7 31 ± 6 0.039 Baseline pain score, mean ± SD 7.1 ± 1.7 6.8 ± 1.7 7.7 ± 1.6 0.008 Baseline opioids consumption in mg oral morphine 30 (20, 68) 30 (20, 60) 30 (20, 60) 0.365 equivalency, median (Q1, Q3) Duration of pain, median (Q1, Q3), y 5 (3, 10) 6 (2, 9) 3 (2, 6) 0.001 Pain diagnosis, n (%) 0.215 Postlaminectomy syndrome 57 (92) 62 (84) 63 (82) Degenerative spine disease† 5 (8) 12 (16) 14 (18) Diabetes history, n (%) 8 (13) 13 (18) 10 (13) 0.661 Disability history, n (%) 3 (5) 4 (5) 7 (9) 0.036 Depression history, n (%) 42 (68) 43 (57) 43 (56) 0.300 Smoking history Duration of smoking, median (Q1, Q3), y 25 (18, 32) 20 (10, 30) N/A Amount of smoking, median (Q1, Q3), pack-year 24 (13, 36) 15 (5, 34) N/A Smoking type Cigarette, n (%) 58 (94) 72 (97) N/A Cigar, n (%) 4 (6) 1 (1) N/A Smoking abstinence, median (Q1, Q3), y N/A 16 (8, 23) N/A *Differences among 3 groups were examined using χ2 test or Fisher exact test for categorical variables, analysis of variance, or Kruskal-Wallis test for continuous variables. †Degenerative spine diseases includes degenerative disk disease, cervical /lumbar radiculopathy, spondylosis and spinal stenosis as pain diagnosis at SCS implant treatment.

mean pain at 1-year follow-up with adjustment for potential smokers on pain scores and opioid consumption after 1 year, confounders listed in primary analysis. The mean difference in adjusted for potential confounders listed in primary analysis and pain between 2 pairs (former smokers vs current smokers and the length of smoking. Moreover, we assessed the association former smokers vs lifelong nonsmokers) along with CI was between smoking abstinence and the pain score or opioid consump- reported. Further, we looked at the difference in means of daily tions among former smokers, adjusted for potential confounders opioid consumption at 1-year follow-up on the log scale, compar- listed in primary analysis. ing current smokers with lifelong nonsmokers. We used the Two-tailed model-based Wald tests were used for all primary multivariable linear model that allowed for the same potential and secondary hypotheses. The type I error for the primary confounding adjustment as for the primary analysis. Opioid con- hypothesis was set at the 0.05 level. The type I error for the sumptions were log transformed as log (opioid+1) to meet normal- secondary outcomes was preserved at 0.05 overall by applying ity assumptions for modeling purpose. The ratio of geometric Bonferroni correction and using a significance criterion of means daily opioid dose after 1 year of SCS therapy comparing P < 0.05/3 = 0.017 for each secondary test. Similarly, the signifi- current smokers with lifelong nonsmokers was reported along cance criterion for explanatory analysis was P < 0.05/4 = 0.013 for with CI. each test. To provide more information on the pathogenic role of SAS statistical software version 9.4 (SAS Institute, Cary, smoking on pain management, as an explanatory analysis, we North Carolina) for 64-bit Microsoft Windows and R statistical assessed the impact of smoking amount per day among current software version 2.3.2 for 64-bit Unix operating system (The R

TABLE 2. Opioid Conversion Doses

Oral Dose Equivalent Intravenous Dose Equivalent Drug Name (Medication-MSO4 Ratio) (Parenteral-Oral Ratio) Transdermal Patch Morphine sulfate 30 mg 10 mg (1:3) N/A Hydromorphone 7.5 mg (1:4) 1.5 mg (1:5) N/A Meperidine 300 mg (10:1) 100 mg (1:3) N/A Methadone 12 mg (1:2.5) 10 mg (1:1.2) N/A Oxycodone 20 mg (1.5:1) N/A N/A Hydrocodone 30 mg (1:1) N/A N/A Codeine 200 mg (6.7:1) 120 mg (1:1.67) N/A Fentanyl N/A N/A 109 mg (3.6:1)

FIGURE 2. Comparison of pain scores and opioid consumption after 1 year of SCS therapy among 3 study groups. Pain scores are presented with 95% CI in a bar graph (left). Daily opioid consumption is presented with median in a box plot (right). P values for primary analysis (blue line) and secondary analysis (black line) are shown above the comparisons. All comparisons were adjusted for baseline pain scores, baseline opioid consumption, age, sex, BMI, duration of pain, spine-related pain diagnosis, diabetes, disability, and depression history.

Foundation for Statistical Computing, Vienna, Austria) were used smokers and nonsmokers demonstrated similar pain and opioid for all statistical analyses. consumption levels to each other. Results of the primary and secondary statistical analyses are Power Consideration shown in Table 3. Statistically significant results arose from both Power calculation was based on the primary outcome. As- our primary and secondary analyses. Primary analysis compared suming pain score SD of 2.5, a sample size of 60 smokers and between current and nonsmokers NRS pain scores and the – 60 lifelong nonsmokers provides 90% power to detect difference adjusted mean difference was 1.93 (95% CI, 1.16 2.71; of 1.5 and larger in pain scores at the significance level of 0.05. P < 0.001) at 12 months. Secondary analysis, which compared current smokers with former smokers' NRS pain scores, showed a mean difference between the 2 groups of 1.32, also statistically RESULTS significant (95% CI, 0.38–2.26; P = 0.001). On the other hand, A summary of the various factors and their statistical sig- although not statistically significant, NRS comparison between nificance, all of which were used in the primary and secondary former smokers and lifelong nonsmokers yielded a mean pain analyses for adjustment purposes, is shown in Table 1. Age, score difference of 0.62 units (95% CI, −0.27 to 1.51; BMI, duration of pain, baseline pain score, and proportion of dis- P = 0.096). As for the daily opioid consumption, as previously abled were significantly different among 3 groups (P <0.05).The mentioned, smokers showed statistically significant higher oral majority of our sample was being treated for postlaminectomy morphine-equivalent usage than lifelong nonsmokers with a ratio syndrome. Current smokers had been smoking for a median of of geometric means in milligrams of 2.37 (95% CI, 1.15–4.87; 25 years, and the median quit time for former smokers was P = 0.004), which indicates that the estimated opioid consumption 16 years prior to SCS implant. after 1 year among smokers was 2.37 times higher than that of The NRS pain scores and opioid consumption at the nonsmokers. All comparisons previously stated were adjusted 12-month post-SCS implant benchmark are shown in Figure 2. for sex, age, BMI, duration of pain, baseline opioid consumption Among current smokers, former smokers, and lifelong non- and pain scores, type of pain diagnosis, history of diabetes, smokers, the means (SDs) of pain scores were 7.0 (2.0), 5.2 disability, and depression. An illustrative summary of each com- (2.8), and 5.1 (1.9), respectively, whereas the medians (Q1, Q3) parison can be seen in Figure 2, with CIs marked at their corre- of opioid consumption were 40 (20, 70), 23 (0, 45), and 23 (0, sponding significance level. 45) in mg morphine equivalent. At 12-month post-SCS implant, Last but not least, an additional analysis was performed to current smokers illustrated higher NRS pain scores in addition investigate the correlation of NRS and opioid utilization in 2 sce- to higher levels of opioid consumption in comparison to both narios: amount of cigarettes smoked per day for current smokers former smokers and lifelong nonsmokers. Interestingly, former and the years of abstinence for former smokers. Although none

TABLE 3. Results of Primary and Secondary Analyses

Unadjusted Difference (95% CI) Adjusted Difference (95% CI)* P† Primary analysis‡ Pain scores: NRS scale Difference in means Current vs nonsmoker 1.87 (1.11, 2.64) 1.93 (1.16, 2.71) <0.001§ Secondary analysis|| Pain scores: NRS scale Current vs former smoker 1.74 (0.80, 2.68) 1.32 (0.38, 2.26) 0.001§ Former vs nonsmoker 0.13 (−0.76, 1.02) 0.62 (−0.27, 1.51) 0.096 Opioid consumption, mg Ratio of geometric means (oral morphine-equivalent dose) Current vs nonsmoker 2.79 (1.29, 6.03) 2.37 (1.15, 4.87) 0.004§ *All comparisons were adjusted for baseline pain scores, baseline opioid consumption, age, sex, BMI, duration of pain, spine-related pain diagnosis, diabetes, disability and depression history. †P value is from adjusted analysis. Statistically significant difference. ‡For primary analysis, the significance criterion is P < 0.05, and the 95.0% CI is reported. §Statistically significant difference. ||For secondary analysis, the significance criterion is P < 0.017 by applying Bonferroni correction (ie, 0.05/3) to preserve it at the 0.05 level overall. Correspondingly, 98.3% CIs, as 95% CIs adjusted for multiple comparisons, are reported.

of the findings hereby were found to be statistically significant, smokers and lifelong nonsmokers, which goes in concordance nonetheless, lower NRS pain scores were observed with increased with the majority of literature published about tobacco smoking number of packs smoked following SCS treatment. A decrease in and chronic pain. pain by 1.23 NRS units (95% CI, −2.59 to 0.13; P = 0.023) was The effect of smoking in chronic spinal pain was explored in found per each extra 1 pack-day. Moreover, as for the impact of a cross-sectional postal survey of 29,424 people by Leboeuf-Yde smoking abstinence, we found that 1 more year of quitting et al,45 who found a positive association between smoking and smoking was associated with 0.6-unit decrease (95% CI, −1.65 low-back pain, which was strongest for low-back pain of long du- to 0.43; P = 0.090) in NRS pain scores after SCS implant, adjusted ration. Similar results originated from the systematic review by for all potential confounders. Goldberg et al,46 who concluded that evidence exists supporting the theory that smoking is associated with increasing incidence and prevalence of nonspecific back pain, however, urged the DISCUSSION necessity of additional long-term studies to be able to state Just over 50 years ago, in 1967, the first SCS device was unequivocally that smoking precedes back pain. Another study implanted by Dr Norman Shealy, a Harvard-trained neurosurgeon involving 862 subjects found statistically significant evidence that at Case Western Reserve University.28 In 1989, it was later smoking men had more frequent pain problems compared with approved by the US Food and Drug Administration. Since then, women. In addition, in comparison with nonsmokers, smokers SCS has become increasingly popular with an estimated an- were found to have back pains more frequently.47 Richardson nual patient SCS implant count, in the United States alone of et al48 conducted a randomized, double-blind, placebo-controlled 34,000.29 The efficacy of SCS for chronic spine–related pain crossover study to show the effect of nicotine on pain score level has been supported in several studies as successfully providing in patients with spinal cord injury. The results showed that sufficient pain relief in the majority of cases.30–34 Despite this nonsmokers reported less pain scores on NRS in comparison to success, there remains a portion of patients who fail to respond patients who were acutely exposed to nicotine gum. Similarly, or show decreased efficacy over time to spinal cord therapy with- in a meta-analysis done by Shiri et al,49 patients who have out a well-founded explanation.35,36 Theories involving failure of smoked or are actively smoking demonstrated worse or higher therapy have been attributed to several factors including pain visual analog scale scores throughout the treatment course of etiology36 and duration,35 psychological factors,37–40 and the num- spinal disorders. ber of preimplant surgeries,35 which are all common contributors. The exact cause of the correlation between smoking and Our primary goal in this study was to determine if regular spinal pain is obscure; nonetheless, there are several possible tobacco use has an effect on the response to SCS neuromodulation explanations for the causality. One of the postulations is that therapy specifically in spine-related chronic pain etiologies. The chronic exposure to nicotine leads to altered pain perception.50 acute analgesic effects of nicotine and tobacco in cigarettes have Another theory is that cigarette smokers may develop autonomic consistently been demonstrated in experimental animals,41,42 nervous system imbalance that could result in sympathetic albeit in human subjects the results have been controversial with predominance and aggravated pain sensations.51 the existence of paradoxical evidence.3–5 Whereas some studies Furthermore, the precise underlying mechanism of advanced presented smoking as an adequate pain reliever in humans,5,43,44 disk and intervertebral disk degeneration and its relation to others found that they accentuate the sensation of pain3,4,6,7 as smoking are also not fully understood. One possible explana- well as cause functional impairment.8,11 Findings of our study tion can be linked to nicotine's vasoconstrictive effect, which indicate that chronic spine–related pain patients who were man- may lead to diminution of blood flow to the disk and its aged with spinal cord stimulation do indeed report higher pain endplates, and because the intervertebral disks are avascular scores if they continue to smoke compared with both former and receive their nutrition via diffusion through pores in the

Copyright © 2018 American Society of Regional Anesthesia and Pain Medicine. Unauthorized reproduction of this article is prohibited. Regional Anesthesia and Pain Medicine • Volume 43, Number 7, October 2018 Smoking and Spinal Cord Stimulation endplate, this ultimately may impair the disk and intervertebral the same analgesic effect compared with lifelong nonsmokers. disk metabolism.52 Moreover, smoking-induced chronic cough Vaughan et al67 showed that smoking increases metabolic clear- may result in internal disk disruption via intermittent increase ance of pentazocine, and hence their urine levels of the metabo- in intradiskal pressure.52,53 Nicotine also affects intervertebral lites were reduced by 40%. The effect of smoking on codeine disk health through down-regulation of the proliferation rate was studied by Yue et al68 and concluded that smoking induced and glycosaminoglycan biosynthesis.54 the glucuronidation of codeine.68,69 We can extrapolate from the From another perspective, smoking stimulates neutrophils in previous studies that patients with chronic back pain who smoke the pulmonary capillaries to release elastase and protease enzymes will have increased opioid consumption due to either increased into the systemic circulation.55 This is in addition to cigarette metabolism of the medications or as a mean of compensation for smoking inhibiting α1 antiprotease, which is the most potent diminished social reinforces. protease inhibitor. The latter further aggravates serum proteolytic Explanatory analysis was done to realize if the effect of activity.56,57 This effect is intensified in smokers because of the smoking on SCS therapy was dose dependent. Although we did higher levels of white blood cells residing in the smokers' not find any significant associations in explanatory analyses lungs.55,58,59 This theory is supported by study findings showing (Table 4), we observed that higher smoking amount was moder- that the activities of proteolytic enzymes are higher in degenerated ately related with lower NRS pain scores after SCS treatment disks when compared with normal disks, evidently increasing the with decrease in pain by 1.2 NRS units per each extra 1 pack- grade of degeneration.60,61 day. Moreover, we found that 1 more year of quitting smoking Chronic smoking behavior was associated with maladaptive was associated with 0.6-unit decrease in NRS pain scores after chronic behavioral issues including preference of inactivity, reli- SCS implant. ance on medical management, and a high level of emotional dis- While our study demonstrates the negative impact of tress.62 Jamison et al62 surveyed chronic pain patients and found smoking on SCS therapy, it also raises many questions urging that 54% of patients with back pain were smoking cigarettes. the need for further studies. Fifty-seven percent of the patients reported resorting to smoking when their chronic pain is intensified. As chronic smokers lose Limitations their social reinforcers (eg, work, social recreational activities), As with all retrospective studies, we cannot be certain that they compensate by increasing other reinforcing behaviors such the independent association we found was causal in nature. Al- as eating and smoking. though we accounted for 10 potentially confounding baseline fac- In our study, mean opioid intake was 2.4 mg higher (CI, tors, residual bias due to uncontrolled confounding variables may – 1.2 4.9 mg; P < 0.001) in smokers compared with lifelong non- remain. In addition, because approximately 35% of the sample smokers despite pain scores approaching preimplant NRS reports who had met the inclusion criterion was excluded from the analy- (7.1 ± 1.7). Ironically, smokers end up consuming more analgesics sis because of missing data, results could potentially be biased. despite the fact that nicotine has analgesic properties in animal and Moreover, we noticed the heterogeneity of former smokers on human studies. A synergistic effect would be expected between quit-smoking time, which ranges from 2 months to 54 years with nicotine and opioids to effectively control the pain, which is in a median at 16 years. The effects of quit-smoking time on pain contrary to the clinical reality. reduction after SCS could be further assessed in future studies. There is paucity of studies addressing the effect of smoking on opioids. In a study by Ackerman and Ahmad,63 the smokers group had higher pain scores and higher hydrocodone consumption. On measuring the serum levels of hydrocodone, the smokers CONCLUSIONS group had lower level of hydrocodone, which may explain the Our results demonstrated that the pain scores after 1 year increased intake of opioids. of SCS therapy among smokers were higher than those among The polycyclic aromatic hydrocarbons induce 3 isoforms lifelong nonsmokers. The daily opioid consumption was higher of CYP (especially CYP2E1), which are a vital part of drug meta- for smokers than that for lifelong nonsmokers as well. In addition, bolism. Clinically, dextromethorphan was found to be less the former smokers who quit smoking before SCS implant had effective in chronic smokers when compared with lifelong non- lower pain scores compared with active smokers after 1 year of smokers.64,65 Keeri-Szanto and Pomeroy demonstrated66 in his therapy. Our results suggest a negative association between study that smokers required higher doses of pentazocine to attain smoking and SCS efficacy in chronic spine–related pain patients.

TABLE 4. Results of Explanatory Analyses

Opioid Consumption: mg Oral Pain Scores: NRS Scale Morphine-Equivalent Dose Difference in Means (95% CI) P Ratio of Geometric Means (95% CI) P Amount of daily smoking for current smokers −1.23 (−2.59 to 0.13) 0.023 0.72 (0.23–2.22) 0.460 per 1 pack-day*‡ Smoking abstinence for former smokers per 1 y†‡ −0.61 (−1.65 to 0.43) 0.090 1.02 (0.97–1.07) 0.325 *The association between amount of daily smoking and 2 outcomes was adjusted for baseline pain scores, baseline opioid consumption, age, sex, BMI, duration of pain, spine-related pain diagnosis, diabetes, disability, depression history, and length of smoking. †The association between smoking abstinence and 2 outcomes was adjusted for baseline pain scores, baseline opioid consumption, age, sex, BMI, duration of pain, spine-related pain diagnosis, diabetes, disability, and depression history. ‡Significant criterion is P < 0.013 by applying Bonferroni correction (ie, 0.05/4) to preserve it at the 0.05 level overall. Correspondingly, 98.7% CIs, as 95% CIs adjusted for multiple comparisons, are reported.

REFERENCES 22. Siana JE, Rex S, Gottrup F. The effect of cigarette smoking on wound healing. Scand J Plast Reconstr Surg. 1989;23:207–209. 1. Elliott AM, Smith BH, Penny KI, Smith WC, Chambers WA. The epidemiology of chronic pain in the community. Lancet. 1999;354:1248–1252. 23. Ahn C, Mulligan P. Smoking-the bane of wound healing: biomedical interventions and social influences. Adv Skin Wound Care. 2008;21: 2. Institute of Medicine. Relieving Pain in America: A Blueprint for 227–236. Transforming Prevention, Care, Education, and Research. Washington, DC: The National Academies Press; 2011. 24. De La Cruz P, Fama C, Roth S, et al. Predictors of spinal cord stimulation – 3. Ditre JW, Heckman BW, Zale EL, Kosiba JD, Maisto SA. Acute analgesic success. Neuromodulation. 2015;18:599 602. effects of nicotine and tobacco in humans: a meta-analysis. Pain.2016; 25. Knezevic NN, Lissounov A, Knezevic I, Candido KD. Smoking does not 157:1373. increase the risk of infection but increases the risk of lead migration in 4. Ditre JW, Brandon TH, Zale EL, Meagher MM. Pain, nicotine, and patients with spinal cord stimulator [abstract]. Anesth Analg.2016; – smoking: research findings and mechanistic considerations. Psychol Bull. 123(3S_suppl):392 393. 2011;137(6):1065. 26. McPherson ML. Demystifying Opioid Conversion Calculations: A Guide 5. Pomerleau OF,Turk DC, Fertig JB. The effects of cigarette smoking on pain for Effective Dosing. 1st ed. American Society of Health-Systems and anxiety. Addict Behav. 1984;9:265–271. Pharmacists: Bethesda, MD; 2009. 6. John U, Hanke M, Meyer C, Völzke H, Baumeister SE, Alte D. Tobacco 27. Hawker GA, Mian S, Kendzerska T, French M. Measures of adult pain: smoking in relation to pain in a national general population survey. Prev visual analog scale for pain (VASpain), numeric rating scale for pain (NRS Med. 2006;43:477–481. pain), McGill Pain Questionnaire (MPQ), Short-Form McGill Pain Questionnaire (SF-MPQ), Chronic Pain Grade Scale (CPGS), Short 7. Vogt MT, Hanscom B, Lauerman WC, Kang JD. Influence of smoking on Form-36 Bodily Pain Scale (SF-36 BPS), and Measure of Intermittent and the health status of spinal patients: the National Spine Network database. Constant Osteoarthritis Pain (ICOAP). Arthritis Care Res (Hoboken).2011; Spine. 2002;27:313–319. 63(suppl 11):S240–S252. 8. Goesling J, Brummett CM, Hassett AL. Cigarette smoking and pain: 28. Shealy CN, Mortimer JT, Reswick JB. Electrical inhibition of pain by depressive symptoms mediate smoking-related pain symptoms. Pain. stimulation of the dorsal columns. Anesth Analg. 1967;46:489–491. 2012;153:1749–1754. 29. Thomson S. Spinal Cord Stimulation's Role in Managing Chronic Disease. 9. Hooten WM, Townsend CO, Bruce BK, et al. Effects of smoking status Available at: http://www.neuromodulation.com/spinal-cord-stimulation. on immediate treatment outcomes of multidisciplinary pain rehabilitation. Accessed December 18, 2016. Pain Med. 2009;10:347–355. 10. Weingarten TN, Moeschler SM, Ptaszynski AE, Hooten WM, Beebe TJ, 30. North RB, Kidd DH, Farrokhi F, Piantadosi SA. Spinal cord stimulation Warner DO. An assessment of the association between smoking status, pain versus repeated lumbosacral spine surgery for chronic pain: a randomized, – intensity, and functional interference in patients with chronic pain. Pain controlled trial. Neurosurgery.2005;56:98 107. Physician. 2008;11:643–653. 31. Kumar K, Taylor RS, Jacques L, et al. The effects of spinal cord stimulation 11. Weingarten TN, Podduturu VR, Hooten WM, Thompson JM, Luedtke CA, in neuropathic pain are sustained: a 24-month follow-up of the prospective Oh TH. Impact of tobacco use in patients presenting to a multidisciplinary randomized controlled multicenter trial of the effectiveness of spinal cord – outpatient treatment program for fibromyalgia. Clin J Pain. 2009;25:39–43. stimulation. Neurosurgery. 2008;63:762 770. 12. Centers for Disease Control and Prevention (CDC). Vital signs: 32. Kumar K, Taylor RS, Jacques L, et al. Spinal cord stimulation versus current cigarette smoking among adults aged≥ 18 years—United States, conventional medical management for neuropathic pain: a multicentre 2005–2010. MMWR Morb Mortal Wkly Rep. 2011;60:1207–1212. randomised controlled trial in patients with failed back surgery syndrome. Pain. 2007;132:179–188. 13. Agaku IT, King BA, Dube SR. Current cigarette smoking among adults—United States, 2005–2012. MMWR Morb Mortal Wkly Rep.2014; 33. North RB, Wetzel FT. Spinal cord stimulation for chronic pain of spinal 63:29–34. origin: a valuable long-term solution. Spine. 2002;27:2584–2591. 14. Orhurhu VJ, Pittelkow TP, Hooten WM. Prevalence of smoking in adults 34. Taylor RS. Spinal cord stimulation in complex regional pain syndrome and with chronic pain. Tob Induc Dis. 2015;13:17. refractory neuropathic back and leg pain/failed back surgery syndrome: results of a systematic review and meta-analysis. J Pain Symptom Manage. 15. Hooten WM, Townsend CO, Bruce BK, Warner DO. The effects of 2006;31:S13–S19. smoking status on opioid tapering among patients with chronic pain. Anesth Analg. 2009;108:308–315. 35. Kumar K, Toth C, Nath R, Laing P. Epidural spinal cord stimulation for — 16. Hooten WM, Shi Y, Gazelka HM, Warner DO. The effects of depression treatment of chronic pain some predictors of success. A 15-year – and smoking on pain severity and opioid use in patients with chronic pain. experience. Surg Neurol. 1998;50:110 121. Pain. 2011;152:223–229. 36. Kumar K, Hunter G, Demeria D. Spinal cord stimulation in treatment of 17. Deer TR, Mekhail N, Provenzano D, et al. The appropriate use of chronic benign pain: challenges in treatment planning and present status, a – neurostimulation: avoidance and treatment of complications of 22-year experience. Neurosurgery. 2006;58:481 496. neurostimulation therapies for the treatment of chronic pain. 37. Pincus T, Burton AK, Vogel S, Field AP. A systematic review of Neuromodulation. 2014;17:571–597; discussion 597–598. psychological factors as predictors of chronicity/disability in prospective 18. Moucha CS, Clyburn T, Evans RP,Prokuski L. Modifiable risk factors for cohorts of low back pain. Spine (Phila Pa 1976).2002;27:E109–E120. surgical site infection. Instr Course Lect. 2011;60:557–564. 38. Mannion AF, Junge A, Taimela S, Müntener M, Lorenzo K, Dvorak J. 19. Møller AM, Pedersen T, Villebro N, Munksgaard A. Effect of smoking on Active therapy for chronic low back pain: part 3. Factors influencing early complications after elective orthopaedic surgery. J Bone Joint Surg Br. self-rated disability and its change following therapy. Spine (Phila Pa 2003;85:178–181. 1976). 2001;26:920–929. 20. Guo SA, DiPietro LA. Factors affecting wound healing. JDentRes.2010; 39. Crombez G, Vlaeyen JW, Heuts PH, Lysens R. Pain-related fear is more 89:219–229. disabling than pain itself: evidence on the role of pain-related fear in chronic back pain disability. Pain. 1999;80:329–339. 21. Jonsson KE, Jensen JA, Goodson WH 3rd, et al. Tissue oxygenation, anemia, and perfusion in relation to wound healing in surgical patients. 40. Vlaeyen JW, Linton SJ. Fear-avoidance and its consequences in chronic Ann Surg. 1991;214:605. musculoskeletal pain: a state of the art. Pain. 2000;85:317–332.

41. Sahley TL, Berntson GG. Antinociceptive effects of central and systemic 55. Donaldson K, Brown GM, Drost E, Selby C, MacNee W. Does cigarette administrations of nicotine in the rat. Psychopharmacology (Berl).1979; smoke enhance the proteolytic activity of neutrophils? Ann N Y Acad Sci. 65:279–283. 1991;624:325–327. 42. Craft RM, Milholland RB. Sex differences in cocaine-and nicotine-induced 56. Janoff A. Elastase in tissue injury. Ann Rev Med. 1985;36:207–216. antinociception in the rat. Brain Res. 1998;809:137–140. 57. Laurent P, Bieth JG. Cigarette smoke decreases the rate constant for 43. Fertig JB, Pomerleau OF, Sanders B. Nicotine-produced antinociception the association of elastase with α1-proteinase inhibitor by a in minimally deprived smokers and ex-smokers. Addict Behav.1986;11: non-oxidative mechanism. Biochem Biophys Res Commun.1985; 239–248. 126:275–281. 44. Pauli P, Rau H, Zhuang P, Brody S, Birbaumer N. Effects of smoking on 58. Abboud RT, Fera T, Johal S, Richter A, Gibson N. Effect of smoking on thermal pain threshold in deprived and minimally-deprived habitual plasma neutrophil elastase levels. J Lab Clin Med. 1986;108:294–300. smokers. Psychopharmacology (Berl). 1993;111:472–476. 59. MacNee W, Wiggs B, Belzberg AS, Hogg JC. The effect of cigarette 45. Leboeuf-Yde C, Kyvik KO, Bruun NH. Low back pain and lifestyle: part I: smoking on neutrophil kinetics in human lungs. NEnglJMed. 1989;321: smoking. information from a population-based sample of 29,424 twins. 924–928. Spine (Phila Pa 1976). 1998;23:2207–2213. 60. Kang JD, Georgescu HI, McIntyre-Larkin L, et al. Herniated lumbar 46. Goldberg MS, Scott SC, Mayo NE. A review of the association between intervertebral discs spontaneously produce matrix metalloproteinases, cigarette smoking and the development of nonspecific back pain and nitric oxide, interleukin-6, and prostaglandin E2. Spine (Phila Pa 1976). – related outcomes. Spine (Phila Pa 1976). 2000;25:995 1014. 1996;21:271–277. 47. Lindal E, Stefansson JG. Connection between smoking and back 61. Crean JK, Roberts S, Jaffray DC, Eisenstein SM, Duance VC. — pain findings from an Icelandic general population study. Scand J Matrix metalloproteinases in the human intervertebral disc: role in disc – Rehabil Med.1996;28:33 38. degeneration and scoliosis. Spine (Phila Pa 1976).1997;22: 48. Richardson EJ, Richards JS, Stewart CC, Ness TJ. Effects of nicotine on 2877–2884. spinal cord injury pain: a randomized, double-blind, placebo controlled 62. Jamison RN, Stetson BA, Parris WC. The relationship between cigarette – crossover trial. Top Spinal Cord Inj Rehabil. 2012;18:101 105. smoking and chronic low back pain. Addict Behav.1991;16:103–110. 49. Shiri R, Frilander H, Sainio M, et al. Cervical and lumbar pain and 63. Ackerman WE 3rd, Ahmad M. Effect of cigarette smoking on serum radiological degeneration among fighter pilots: a systematic review and hydrocodone levels in chronic pain patients. Ark Med Soc. 2007;104: – meta-analysis. Occup Environ Med. 2015;72:145 150. 19–21. 50. Girdler SS, Maixner W, Naftel HA, Stewart PW, Moretz RL, Light KC. 64. Jick H. Smoking and clinical drug effects. Med Clin North Am.1974;58: Cigarette smoking, stress-induced analgesia and pain perception in men 1143–1149. and women. Pain. 2005;114:372–385. 65. Boston Collaborative Drug Surveillance Program. Decreased clinical 51. Middlekauff HR, Park J, Moheimani RS. Adverse effects of cigarette and efficacy of propoxyphene in cigarette smokers. Clin Pharmacol Ther. noncigarette smoke exposure on the autonomic nervous system: 1973;14:259–263. mechanisms and implications for cardiovascular risk. J Am Coll Cardiol. 2014;64:1740–1750. 66. Keeri-Szanto M, Pomeroy JR. Atmospheric pollution and pentazocine metabolism. Lancet. 1971;297:947–949. 52. Frymoyer JW, Pope MH, Clements JH, Wilder DG, MacPherson B, Ashikaga T. Risk factors in low-back pain. An epidemiological survey. 67. Vaughan DP, Beckett AH, Robbie DS. The influence of smoking on the J Bone Joint Surg Am. 1983;65:213–218. intersubject variation in pentazocine elimination. Br J Clin Pharmacol. 1976;3:279–283. 53. Frymoyer JW,Pope MH, Costanza MC, Rosen JC, Goggin JE, Wilder DG. Epidemiologic studies of low-back pain. Spine (Phila Pa 1976). 1980; 68. Yue QY, Tomson T, Säwe J. Carbamazepine and cigarette smoking induce 5:419–423. differentially the metabolism of codeine in man. Pharmacogenetics.1994; – 54. Elmasry S, Asfour S, de Rivero Vaccari JP,Travascio F. Effects of tobacco 4:193 198. smoking on the degeneration of the intervertebral disc: a finite element 69. Rogers JF, Findlay JW, Hull JH, et al. Codeine disposition in smokers and study. PloS One. 2015;10:e0136137. nonsmokers. Clin Pharmacol Ther. 1982;32:218–227.