Smoking cessation correlates with a decrease in infection rates following total joint arthroplasty

Abstract

Background

The impact of tobacco use on perioperative complications, hospital costs, and survivorship in total joint arthroplasty (TJA) is well established. The aim of this study is to report the impact of tobacco cessation on outcomes after TJA and to measure the impact of a voluntary smoking cessation program (SCP) on self-reported smoking quit rates in a premier academic medical center.

Methods

A seven-year (2013–2019) SCP database was provided by the Integrative Health Promotion Department and Infection Prevention and Control Department. We evaluated program and smoking status, patient demographics, length of stay (LOS), and 90-day post-operative infection rates and readmission rates. The primary outcome was quit rates based on SCP enrollment status. The secondary outcomes measured infection rates, readmission rates, and LOS based on enrollment status and/or quit rate.

Results

A total of 201 eligible patients were identified: 137 patients in the SCP (intervention) group and 64 in the self-treatment (control) group. SCP patients trended towards higher quit rates (43% vs 33%, p = 0.17), shorter LOS (2.47 vs 2.62 days, p = 0.52), lower infection rates (7.3% vs 12.5%, p = 0.27) and slightly higher readmission rates (5.8% vs 4.7%, p = 0.73). In a sub-analysis, self-reported smokers demonstrated statistically significant decrease in infection (3.7% vs 12.5%, p = 0.03).

Conclusion

There was a statistically significant decrease in infection rates in patients who self-reported quitting tobacco prior to TJA. Additionally, quit rates for patients who participated in a voluntary SCP trended towards increased pre-operative cessation. Further efforts to increase tobacco cessation prior to TJA and examine the impact on patient outcomes are needed.

1. Introduction

In an effort to optimize patient outcomes in total joint arthroplasty (TJA), modifiable risk factors associated with short term perioperative complications, such as smoking, have come to the forefront of our attention. Smokers undergoing TJA are associated with over 50% increased risk of perioperative complications and around $5000 increase in hospital costs relative to nonsmokers.^1,2 Lavernia et al. demonstrated that active tobacco use translated into 16% high hospital costs for elective TJA ($35,628 vs $30,706).¹ As there is more data available regarding specific methods to implement significant improvement of comorbidities such as obesity and diabetes in the TJA population, there is an opportunity to explore how to better control smoking in order to achieve similar results.^3,4

In response, we developed a voluntary perioperative tobacco and SCP at our medical center for smokers undergoing elective total knee (TKA) and total hip arthroplasty (THA) since 2013. The SCP itself aims to assist smokers in quitting tobacco before surgery and to maintain quit status after surgery, as well as to reduce perioperative complications, including readmissions, and hospital length of stay (LOS). The aim of this study is to report the impact of tobacco cessation on outcomes after TJA and to measure the impact of a voluntary smoking cessation program (SCP) on self-reported smoking quit rates in a premier academic medical center. We hypothesize that engagement in the program will increase cessation rates and that patients who quit smoking prior to surgery will have decreased complications, readmission rates, and LOS.

2. Methods

2.1. Study design and patient population

After receiving Institutional Review Board (IRB) approval, an SCP database with de-identified patient information over a 7-year time period (2013–2019) was collected and provided by the Integrative Health Promotion Department and Infection Prevention and Control Department at our institution. The consecutive patients reviewed from this database include smokers greater than or equal to 18-years-old undergoing primary, elective TJA. In regards to smokers information we collected estimated start year of smoking, pack years, method of smoking (cigar, cigarettes), any attempts at quitting and currentsmoking status with packs per day. Patients were excluded from the study if they were non-smokers, were less than 18-years-old, or underwent non-elective TJA. The variables collected in this study were program status, baseline patient characteristics (age, BMI, American Society of Anesthesiologists (ASA) score, sex, and race), smoking status (either quit or did not quit smoking), LOS, infection rates (either superficial or deep surgical site infection within a 90-day post-operative period), and readmission rates (within 90-days after surgery). All causes of readmission were documented, such as debridement and implant retention (DAIR), manipulation under anesthesia (MUA), and fracture.

Patients who either completed the program (n = 82), enrolled in the program (n = 14), or enrolled but withdrew in the program (n = 41) were considered to be in the experimental (SCP) group (total n = 137). We chose these sub-groups as the experimental group, including any patients who participated in the program but withdrew later on, with an intention-to-treat analysis in mind as a way to minimize potential bias in the reporting of our results. Patients who declined to enroll (n = 30), were not eligible for the program at the time due to their scheduled time of surgery (n = 17), or with whom we had unsuccessful contact, (n = 17) made up our self-treatment (control) group (total n = 64). We chose these sub-groups as the control or more accurately named the non-intervention group as they were smokers who underwent TJA, but did not participate in the program. Our primary outcome looked at smoking cessation (quit) rates in patients depending on their SCP enrollment status. Our secondary outcome evaluated the impact on readmission rates and LOS depending on patient SCP enrollment status and respective quit rate.

2.2. Peri-operative smoking cessation program

Current tobacco users seen at our institution's Medicaid ambulatory clinic were identified and referred to the SCP program prior to surgery. When a tobacco user decided to proceed with scheduling an elective TJA, the surgeon informed the patient about the risks associated with perioperative tobacco use and encouraged the patient to quit before surgery. Due to variability in office staffing and resources, attending surgeons were encouraged but not required to refer smokers to the SCP. Smokers who did not participate in the SCP were still informed, counseled, and encouraged to quit smoking prior to surgery via our standard preoperative optimization protocols.

SCP interventions are initiated 4–6 weeks prior to surgery, with a goal of being tobacco free 1–2 weeks prior to surgery. The protocol consists of four pre-operative telephone sessions including assessment, education, counseling, nicotine replacement therapy (NRT), and two follow-up sessions within 30 days after surgery. The SCP team is comprised of nurse practitioners, licensed clinical social workers, physician's assistants, and physicians. The SCP team contacts the orthopedic surgeon one week prior to surgery to update the surgeon on the patient's progress in the program, and a decision is made to proceed with or to postpone surgery. Patients are instructed to stop NRT the night before surgery. Upon surgical admission, the patient's self-reported current tobacco use (cigarettes per day or equivalent) is recorded. NRT is resumed after hospital discharge as needed. The SCP follows up with the patient within 30 days post-operative to reassess tobacco use. After this point, any further intervention is passed to the patient's primary care physician. This 30-day mark is the last standardized follow up for the program.

2.3. Statistical analysis

Descriptive statistics were used to report primary and secondary outcomes between the two groups. Continuous variables were represented as means ± standard deviation while categorical data was represented as counts (%). These variables were analyzed using unpaired T-tests for continuous data, such as age, BMI, ASA, and LOS and chi-squared tests for categorical data, such as sex, race, quit rate, infection rate, and readmission rate. A sub-group analysis using chi-squared test was conducted to identify readmission and reoperation in patients who quit smoking and were either enrolled or not enrolled at our institution's SCP. Another sub-analysis was done to identify infection rates, readmission rates, and LOS in patients who quit smoking regardless of SCP enrollment. All statistical analyses were performed using SPSS v25 (IBM Corporation, Armonk, New York). A p-value of less than 0.05 was used to determine the statistical significance of a finding. Additionally, a post-hoc power analysis (α = 0.05 and power of 80%) was conducted using G*Power version 3 (Erdfelder, Faul, & Buchner, Germany) to determine the power of our study's findings with the current proportions.

3. Results

A total of 201 patients were identified and analyzed, with 137 patients in the SCP group and 64 patients in the self-treatment group. The SCP group was comprised of 52% males and 48% females with an average age of 57.7 years (±9.1), average BMI of 30.2 kg/m² (±6.1), and average ASA of 2.4 (±0.6). The majority of patients in the SCP group were White (46%), followed by African American (34%), and other (20%). The self-treatment group consisted of 52% males and 48% females with an average age of 57.5 years (±8.7), average BMI of 31.5 kg/m² (±6.2), and average ASA of 2.3 (±0.56). Whites made up 45% of the self-treatment cohort, followed by African American (38%), and other (17%). No significant differences were found between the two groups for any demographic data, including age (p = 0.86), BMI (p = 0.21), ASA (p = 0.32), gender (p = 0.97), and race (p = 0.8). Our results showed that patients who enrolled in the SCP had higher quit rates vs those who did not enroll in the SCP, however this difference was not significant (43% vs 33%, p = 0.17). Of note, all patients included in this study were offered the opportunity to enroll in the program and were encouraged to quit smoking, so this may have caused an increase in cessation rates even for those who declined the program itself. Infection rates based on the SCP enrollment status were lower in the SCP group vs those in the self-treatment group, although not statistically significant (7.3% vs 12.5%, p = 0.27) as our study was not powered to detect differences in infection rates. See Table 1 for reference.

Table 1.

Patient demographics, smoking status, and infection rates by SCP enrollment status, represented as means (±SD) or counts (%).

Variables	SCP (n = 137)	Self Treatment (n = 64)	p-value*
Patient Characteristics
Age	57.7 (±9.1)	57.5 (±8.7)	0.86
BMI	30.2 (±6.1)	31.3 (±6.2)	0.21
ASA	2.4 (±0.6)	2.3 (±0.56)	0.32
Gender			0.97
Male	71 (52)	33 (52)
Female	66 (48)	31 (48)
Race			0.8
White	63 (46)	29 (45)
African American (Black)	46 (34)	24 (38)
Other Race (including Asian and Hispanic/Latino)	28 (20)	11 (17)
Smoking Status			0.17
Quit	59 (43)	21 (33)
Smoke	77 (57)	43 (67)
Length of Stay (LOS)	2.47 (±1.6)	2.62 (±1.4)	0.52
Infection Rates	10 (7.3)	8 (12.5)	0.27
Readmission Rates	8 (5.8)	3 (4.7)	0.73

*p-values are derived from two-tailed t-test or chi-test for categorical values.

A sub-group analysis was performed measuring the readmission rates and LOS in patients in the SCP group who successfully quit smoking to determine whether smoking cessation alone affects readmission rates or LOS in our study population rather than SCP enrollment status alone. These findings show that patients who quit smoking in the SCP group had a lower rate of infection (3.4% vs. 4.8%), yet these findings were not statistically significant (p = 0.13 in the SCP group and p = 0.19 in the self-treatment group) as they were underpowered. Moreover, patients who quit smoking in the SCP group were less likely to experience readmissions within 90-days (3.4%) vs those in the self-treatment group (4.8%), however this was not statistically significant (p = 0.29 in the SCP group and p = 0.98 in the self-treatment group). Hospital LOS was shorter in patients who stopped smoking in the SCP group (2.4 days) vs self-treatment group (2.5 days), although not statistically significant (p = 0.82 in the SCP group and p = 0.68 in the self-treatment group), see Table 2.

Table 2.

Infection rates, readmission rates, and LOS by smoking status and SCP enrollment status, represented as means (±SD) or counts (%).

SCP (n = 137)
	Quit	Smoke	p-value*
Infection Rate	2 (3)	8 (10)	0.13
Readmission Rate	2 (3)	6 (8)	0.29
LOS	2.4 (±1.3)	2.5 (±1.8)	0.82
Self Treatment (n = 64)
	Quit	Smoke	p-value*
Infection Rate	1 (5)	7 (16)	0.19
Readmission Rate	1 (5)	2 (5)	0.98
LOS	2.5 (±1.1)	2.7 (±1.6)	0.68

Another analysis was done to assess whether smoking status alone, regardless of SCP enrollment, affects readmission rates and LOS in our study population. The results of this sub-analysis showed that patients who quit smoking were less likely to experience infection (3.7% vs 12.5%, p = 0.03), readmissions (4.9% vs 9.2%, p = 0.26), and had shorter LOS (2.5 vs 2.6 days, p = 0.65) than smokers, see Table 3.

Table 3.

Infection rates, readmission rates, and LOS by smoking status alone, represented as means (±SD) or counts (%).

	Quit	Smoking	p-value*
Infection Rate	3 (4)	15 (12)	0.03
Readmission Rate	3 (4)	8 (7)	0.38
LOS	2.5 (±1.2)	2.6 (±1.7)	0.65

Last, causes of readmissions within a 90-day post-operative period are detailed in Table 4, with DAIR occurring as the most common cause of readmission in the SCP group (75%) at a rate of 4.4% followed by MUA and trauma evenly (12.5%) at a rate of 0.7% each. DAIR, MUA, and trauma were each equally a cause of readmission in our self-treatment group (33%) at a rate of 1.6%. There were no significant differences detected between the SCP group and the self-treatment group for any cause of readmission.

Table 4.

Causes of Readmission by SCP enrollment status, represented as counts (%) and range (months).

Cause of Readmission	SCP (n = 137)	Mean Follow-Up Time (range)	Self Treatment (n = 64)	Mean Follow-Up Time (range)	p-value*
Debridement and Implant Retention (DAIR)	6 (4)	1.4 (0.9–2.4)	1 (2)	1.2	0.31
Manipulation under anesthesia (MUA)	1 (1)	2.6	1 (2)	2.7	0.58
Traumaa	1 (1)	0.6	1 (2)	1.1	0.58

^aTrauma due to peri-prosthetic fracture.

4. Discussion

We encouraged smoking cessation and initiated a voluntary program due to the known risks of smoking within arthroplasty. All patients, whether enrolled in our SCP or not, were encouraged to quit smoking. We report a 23% increase in quit rate in patients enrolled in our program versus those who did not enroll. Although not statistically significant, the findings of our study showed positive trends in each of our study outcomes. Self-reported quitting combined with participation in the program was clinically associated with higher quit rates, shorter LOS, and lower infection rates compared to those who did not enroll. In a sub-analysis looking at smoking status alone regardless of SCP participation, patients who quit smoking were more likely to experience lower infection rates, less readmissions, and shorter LOS than smokers. As our SCP increased success in smoking cessation substantially it has the potential to further improve these outcomes for patients undergoing TJA.

Although our results were not statistically significant, this does not subtract from the clinical significance of smoking cessation and its impact on decreased perioperative complications, increased hospital costs, length of stay, and decreased TJA survivorship that is actively being explored in the Orthopedic field.5, 6, 7, 8, 9 Boylan et al. generated a cost analysis model and found that the presence of a PJI averaged between $105,079 (for DAIR) and $111,717 (for two-stage exchange).⁵ By using these numbers as estimates, one can easily extrapolate the mass amount of savings possible with decreasing the number of complications, even if the actual number decreased does not prove to be statistically significant.

There is a variety of literature detailing the different ways smoking cessation significantly affects outcomes for both patients and the healthcare system. Many authors have reported a reduction in complications such as wound complications, infections, readmissions in patients who have quit smoking prior to elective TJA.^7,10, 11, 12 In a review of patients undergoing elective TJA, Moller et al. found that tobacco use was the single most important risk factor for the development of post-operative complications including increased wound-related and cardiopulmonary complications, intensive care unit (ICU) admission, and LOS.¹³ Tobacco use is a potentially modifiable risk factor that presents an opportunity to reduce complication rates, improve clinical outcomes, and decrease medical costs through initiating programs to identify tobacco users and promote smoking cessation prior to surgery. In a recent model exploring the short term cost effectiveness of mandatory smoking cessation programs (SCP), the average 90-day cost was $32 less for patients enrolled, and participation could reduce the rate of PJI (periprosthetic joint infection) by 25% for former vs current smokers with the cost of intervention averaging less than $219.⁵ A prospective, randomized, controlled trial in 2008 demonstrated that just four weeks of structured intervention prior to surgery resulted in reduction of complication rate of 41%–21% (p = 0.03).¹⁴ A retrospective cohort study of 3300 patients undergoing primary THA found that former and current smokers had a 43% and 56% increased risk of perioperative complications, respectively, compared to nonsmokers.¹⁵ Our data showed similar trends towards decreased LOS, infections, and readmission rates in quitters vs non-quitters. Additionally, Singh et al. have demonstrated that smokers undergoing TJA were more likely than never smokers to develop surgical site infection (odds ratio (OR) 1.41), pneumonia (OR 1.53), stroke (OR 2.61), and 1-year mortality (OR 1.63).⁸ Former smokers experienced these adverse outcomes less frequently than current smokers, but more often than never smokers. Another group demonstrated that current smokers have an increased risk of pulmonary complications relative to never smokers after elective non-cardiac surgery (OR 5.5).¹⁶ In 2018 there was a large retrospective study published by Bedard et al. reviewing the effect of smoking on revision TKAs. They found that out of 8776 revision TKAs, the currently smoking cohort had higher rates of wound complications (3.8% vs 1.8%, P < 0.0001), deep infection (2.5% vs 1.0%, P < 0.0001), pneumonia (1.3% vs 0.4%, P < 0.0001), and reoperation (5.0% vs 3.1%, P = 0.001). This demonstrates how smoking further complicates revision cases compared to the higher risks it poses for primaries.⁶ Tobacco use has also been linked with the survivorship of knee implants; 46-month survivorship of TKAs was significantly lower among smokers than non-smokers.⁹ Our sub-analysis adds onto the previous literature, showing that smoking cessation, even in this 6 week perioperative window, is associated with lower complication rates, thus highlighting the benefits of smoking cessation prior to TJA as well as the importance of documenting former tobacco use even for patients who successfully quit before surgery. Although our study's sample size is small compared to the previously enumerated studies, our results echo their results. Moreover, our study shows that even in this narrow time window, patients may experience better outcomes if they can even pause their tobacco use around the time of surgery.

The current study reports a cessation rate of 43% in patients who enrolled in our institution's SCP vs 33% in patients who did not, although not statistically significant (p = 0.17), this is a positive trend towards increased quit rates. Patients may find success in our program due to the multiple resources available to cease smoking such as additional telephone counseling and education, NRT, the possibility of surgeons postponing cases if patients did not adhere to abstinence as an additional motivating factor, and the option for patients to choose to enroll and participate. Surprisingly, 33% of patients who did not enroll in our SCP found success in smoking cessation prior to surgery, which may be attributed to marital status and education level, which are known to influence smoking cessation success.¹⁷ These increased rates may also be related to our introduction of the idea and encouragement of smoking cessation, even if they did not ultimately enroll. These rates may also just simply be reflective of our study's sample size and the possibility of self-reporting bias.

Although our study has several strengths including the multiple resources available to the patients to cease smoking, there are several limitations in our study to note. First, cessation rates reported in this study may be susceptible to self-reporting bias in which patients may either over- or under-report their smoking behaviors. Thus, we cannot confirm that these patients have in fact quit smoking without the presence of chemical testing. One way to possibly solve this is to incorporate biochemical confirmation testing (urine nicotine tests) with the pre-operative labs. This includes urine anabasine and cotinine. The urine anabasine signifies nicotine use and the cotinine signifies nicotine replacement therapy use.¹⁸ If performed the week prior to surgery, it is known whether or not the patients are actively smoking and subsequently can decide whether or not they would consider postponing surgery. There is also the possibility of using carbon monoxide breath tests, as previously cited studies used to evaluate the recent smoking activity. Carbon monoxide maybe a better option as it is only positive for smoking-not for NRT or e-cigarettes which is a challenge regarding the use of cotinine. Neither of these tests have long enough half-lives to completely exclude the possibility of patients quitting only several days before they are tested, pre- and post-operatively and continuing to smoke in the intervals. These tests could also be used in a future study looking at complete nicotine cessation. It would be interesting to see if nicotine replacement therapy itself introduces risk factors for patients, and by using these tests we could ensure people not only quit smoking, but also were not using any nicotine therapies to see if this further affected outcomes. Another challenge is the inconsistency in how the program is recommended to patients by surgeons. The result is a less efficient enrollment process and increased challenges during the enrollment process which feeds into a small sample size. This process is currently being streamlined so that enrollment better represents the number of arthroplasty surgeries occurring at our institution. Lastly, as not all patients were required to undergo the smoking cessation program, it is possible that characteristics of program participants differed from those of smokers who did not participate in the program. We tried to counteract this by reviewing a large number of characteristics as well as previously established patient classifiers (ASA score etc.) to verify similarities between the groups.

In conclusion, our study reports that smoking cessation, even within a small perioperative window, alone significantly decreases infection rates following total joint arthroplasty, and that there are positive trends towards improved readmission rates and reduced length of stay. Further efforts to encourage and support smoking cessation are warranted, including development of optimal tobacco cessation strategies for the total joint arthroplasty population.

Disclaimers

The authors have not received any grant support or research funding and do not have any proprietary interests in the materials described in the article.