ASSOCIATION OF SMOKING HISTORY ON HEALTH-RELATED QUALITY OF LIFE IN PATIENTS UNDERGOING RADICAL CYSTECOMY

Abstract

Purpose:

Radical cystectomy (RC) has the potential to impact health-related quality of life (HRQOL). Many patients who undergo RC are current or former smokers. To better inform pre-operative patient counseling, we examined the association between smoking status and HRQOL after RC.

Materials and Methods:

A secondary analysis was performed on a prospective, longitudinal study (2008-2014) examining HRQOL in patients undergoing RC for bladder cancer. We analyzed 12 validated patient-reported outcome measures that focused on functional, symptomatic, psychosocial, and global HRQOL domains. Measures were collected pre-operatively and 3-, 6-, 12-, 18-, and 24-months post-operatively. For each HRQOL domain, we estimated the mean domain scores using a generalized estimation equation linear regression model. Each model included survey time, smoking status, and time-smoking interaction as covariates. Pairwise comparisons of current, former, and never smokers were estimated from the models.

Results:

Of the 411 patients available for analysis, 29% (n=119) never smoked, 59% (n=244) were former smokers, and 12% (n=48) were current smokers. Over the follow-up period, never smokers compared to current smokers had better global QOL scores (mean difference = +8.9; 95% CI 1.3 to 16; p = 0.023) and lower pain levels (mean difference = −10; 95% CI −19 to −0.54; p = 0.036). Compared to current smokers, former smokers had marginal improvements in global QOL (+6.9 points) and pain (−7.5 points) during the follow-up period.

Conclusions:

Current smokers reported worse HRQOL recovery in the 24-months after RC. These findings can be used to counsel patients who smoke on recovery expectations.

1. INTRODUCTION

Smoking is the leading cause of bladder cancer and many patients undergoing radical cystectomy (RC) are current or former smokers. Smoking is associated with higher tumor grade and stage at initial diagnosis,¹ more advanced disease at the time of RC, and an increased risk of cancer recurrence.² Following RC, smokers also have an increased risk of peri-operative complications and need for re-operative surgery.^{3, 4}

The trajectory of health-related quality of life (HRQOL) recovery after RC has been previously evaluated.⁵ However, among patients undergoing RC, there is a complex relationship between patient comorbidities, convalescence after surgery, and smoking status that has not been explored. In other smoking-associated cancers, such as lung cancer, active smokers reported worse recovery in HRQOL up to 12 months following lobectomy and pneumonectomy.⁶ Given the similar comorbidity profile and the surgical complexity of RC, we hypothesized that there may be a similar negative association between smoking status and HRQOL after RC.

The pre-RC period is a time to optimize modifiable comorbidities, address inherent surgical risks, and establish recovery expectations. Patient education is paramount to ensure patients are active participants in the medical decision-making and treatment processes. Understanding the implications of smoking status on HRQOL recovery after surgery can better inform patients about their expected post-operative course. Therefore, our objective was to examine the association smoking status has on HRQOL before and after RC for bladder cancer in order to improve pre-operative counseling. Given this emphasis on patient counseling, we did not attempt to determine the causal relationship between smoking and post-operative functional recovery.

2. MATERIALS AND METHODS

Following institutional review board approval (IRB Protocol No. 08-076), a secondary analysis was performed on a prospective, longitudinal study open between 2008-2014, which examined HRQOL in patients undergoing RC for bladder cancer. The study methods and primary outcome of this study has been previously published.⁵ For this analysis, we analyzed 12 a priori chosen HRQOL domains captured via validated patient-reported outcome (PRO) measures: Cognitive Functioning, Emotional Functioning, Fatigue, Financial Difficulties, Pain, Physical Functioning, Global QOL, Role Functioning, and Social Functioning from the European Organization for Research and Treatment of Cancer Quality of Life Questionnaire (EORTC QLQ-C30)⁷; Fear of Recurrence from the Fear of Recurrence Questionnaire (FRQ)⁸; Mental Health via the Mental Health Inventory (MHI)⁹; and Satisfaction with Life via the Satisfaction with Life Scale (SWLS).¹⁰ All HRQOL domains are measured between 0 and 100, except Satisfaction with Life (5-35) and Fear of Recurrence (0-92). Higher scores on the functional domains indicate superior HRQOL, whereas higher symptom domain scores indicate worse symptoms. Measures were collected pre-operatively and at 3-, 6-, 12-, 18-, and 24-months post-operatively. FRQ, MHI, and SWLS measures were not collected at 3-months post-operatively.

For each HRQOL domain, we built a generalized estimation equation (GEE) linear regression model with an autoregressive correlation structure. The model included HRQOL domain as the outcome. Survey time, smoking status, and time-smoking interaction were covariates. Survey time was included in the model as a categorical variable. From the model, the predicted mean HRQOL value at each time point was estimated and plotted in figures to illustrate HRQOL scores over time among the smoking groups. To test for differences in HRQOL over time, we quantified mean HRQOL values over the follow-up period again utilizing an autoregressive GEE model with smoking status as the covariate. Pairwise comparisons among the three smoking groups (current, former, or never smoker) included Tukey’s adjustment. Smoking group categorization was based on self-reported smoking status at the time of initial consultation. Smoking status remained static throughout the study period.

All analyses were conducted using R version 4.2.1 with the geepack (v1.3.4), emmeans (v1.7.5), gtsummary (v1.6.1) and tidyverse (v1.3.2) packages.^11-15

3. RESULTS

Of the 411 patients available for analysis, 29% (n=119), 59% (n=244), and 12% (n=48) were never, former, and current smokers, respectively. Clinicopathological characteristics of the cohort are listed in Table 1. Compared to never and former smokers, current smokers were younger and had higher rates of clinically and pathologically advanced disease. Responses to baseline PRO measures had an approximately 70% completion rate with completion rates at subsequent study time periods ranging from 47% to 59% (Supplemental Table 1).

Table 1.

Baseline cohort characteristics based on smoking status.

Characteristic	Never, N = 1191	Former, N = 2441	Current, N = 481
Age at Surgery	66 (60, 74)	68 (63, 74)	61 (55, 68)
Male	90 (76%)	199 (81%)	35 (73%)
BMI	27 (24, 31)	29 (26, 32)	28 (25, 32)
Pre-RC eGFR	71 (57, 87)	69 (55, 81)	70 (55, 83)
Urinary Diversion
Continent Diversion	64 (54%)	117 (48%)	25 (52%)
Ileal Conduit	55 (46%)	127 (52%)	23 (48%)
Clinical T Stage
Ta or Tis	17 (15%)	18 (7.4%)	0 (0%)
T1	52 (45%)	88 (36%)	15 (32%)
T2-T4	47 (41%)	136 (56%)	32 (68%)
Unknown	3	2	1
Prior Intravesical Therapy	57 (48%)	90 (37%)	8 (17%)
Neoadjuvant Chemotherapy	45 (38%)	110 (45%)	23 (48%)
pN+	16 (13%)	37 (15%)	11 (23%)
Adjuvant Chemotherapy	6 (5.0%)	16 (6.6%)	1 (2.1%)

¹Median (IQR); n (%)

Baseline differences were observed in certain HRQOL domains among the smoking groups. Never smokers reported better cognitive functioning and less pain, but lower social functioning; former smokers reported better emotional functioning (Figure 1).

Figure 1.

Radar plot constructed to compare baseline mean survey scores based on smoking status. Higher scores on the functional domains indicate superior health-related quality of life (HRQOL). Higher symptom scores indicate worse symptoms.

Over the follow-up period, there were differences in global QOL based on smoking status (p = 0.023). Compared to current smokers, never smokers had higher global QOL scores (mean difference = +8.9; 95% CI 1.3 to 16), indicating better overall quality of life. Former smokers, compared to current smokers, had marginal improvements in global QOL scores (mean difference = + 6.9; 95% CI −0.38 to 14). Compared to baseline levels, never smokers had a steady improvement in global QOL scores throughout the 24-month follow-up period. In contrast, current smokers experienced a 10-point decrease in global QOL domain score by 3-months post-operatively with limited recovery thereafter (Figure 2). No statistically significant differences were detected in mean cognitive, emotional, physical, social, or role functioning domain scores based on smoking status throughout the follow-up period (Table 2).

Figure 2.

Generalized estimating equation model estimates for never smokers (red), former smokers (green), and current smokers (blue). Selected questionnaire results are shown. Associated p-values, which test for any difference in the mean health-related quality of life (HRQOL) domain among the smoking groups over the follow-up period, are listed in Table 2.

Table 2.

Mean health-related quality of life (HRQOL) values over the follow-up period.

			Mean Over Follow-up Period			Pairwise Mean Differences (95% CI)
HRQOL Domain	Scale	Interpretation	Never	Former	Current	Never - Former	Never - Current	Former - Current	p- value5
Global QOL1	0-100	Higher scores = better function	77	75	68	2.0 (−2.4, 6.4)	8.9 (1.3, 16)	6.9 (−0.38, 14)	0.023
Pain1	0-100	Higher scores = worse symptoms	10	13	20	−2.5 (−6.9, 1.8)	−10 (−19, −0.54)	−7.5 (−17, 1.8)	0.036
Mental Health2	0-100	Higher scores = better health	78	77	69	1.4 (−2.9, 5.7)	9.5 (0.42, 19)	8.1 (−0.66, 17)	0.050
Cognitive Functioning1	0-100	Higher scores = better function	89	87	82	2.4 (−1.0, 5.9)	7.1 (−0.77, 15)	4.7 (−3.1, 12)	0.054
Satisfaction with Life3	5-35	Higher scores = better satisfaction	14	15	18	−0.61 (−2.5, 1.3)	−3.4 (−6.9, 0.05)	−2.8 (−6.1, 0.47)	0.069
Fear of Recurrence4	0-92	Higher scores = worse fear	64	64	65	0.06 (−1.4, 1.5)	−1.7 (−3.7, 0.41)	−1.7 (−3.6, 0.17)	0.094
Emotional Functioning1	0-100	Higher scores =better function	83	81	74	1.4 (−3.4, 6.2)	9.2 (−1.6, 20)	7.8 (−2.7, 18)	0.14
Fatigue1	0-100	Higher scores = worse symptoms	23	23	28	0.08 (−4.3, 4.5)	−4.9 (−13, 3.7)	−5.0 (−13, 3.4)	0.4
Financial Difficulties1	0-100	Higher scores = worse symptoms	15	15	22	−0.65 (−7.2, 5.9)	−7.3 (−20, 5.6)	−6.7 (−19, 5.6)	0.4
Physical Functioning1	0-100	Higher scores = better function	88	86	84	1.3 (−2.6, 5.2)	3.2 (−3.4, 9.8)	2.0 (−4.5, 8.4)	0.5
Social Functioning1	0-100	Higher scores = better function	78	79	75	−0.66 (−6.3, 4.9)	3.1 (−7.8, 14)	3.8 (−6.7, 14)	0.7
Role Functioning1	0-100	Higher scores = better function	82	83	83	−0.81 (−6.1, 4.5)	−0.41 (−9.7, 8.8)	0.40 (−8.2, 9.0)	>0.9

¹European Organization for Research and Treatment of Cancer Quality of Life Questionnaire (EORTC QLQ-C30)

²Mental Health Inventory (MHI)

³Satisfaction with Life Scale (SWLS)

⁴Fear of Recurrence Questionnaire (FRQ)

⁵Test of any difference among smoking groups

Following RC, we also observed differences in pain scores based on smoking status (p = 0.036). Compared to current smokers, never smokers had improved pain scores (mean difference = −10; 95% CI −19 to −0.54). Former smokers, when compared to current smokers, also had improved pain, albeit to a lesser degree (mean difference = −7.5; 95% CI −17 to 1.8). Regardless of smoking status, pain scores expectedly worsened 3-months following RC. By 6-months, pain scores returned to baseline for never and former smokers. For current smokers, pain scores returned to baseline by 12-months before worsening at 18- and 24-months (Figure 2). No differences were detected in mean fatigue symptom scores based on smoking status throughout the follow-up period.

Although not reaching traditional statistical significance (p = 0.050), both never and former smokers had higher estimated mental health scores during the follow-up period. Compared to current smokers, never smokers reported improvement in mental health scores (mean difference = +9.5; 95% CI 0.42 to 19). Former smokers, when compared to current smokers, experienced a smaller improvement in mental health scores (mean difference = +8.1; 95% CI −0.66 to 17). We found no clinically or statistically significant differences throughout the follow-up period in mean satisfaction with life, fear of recurrence, or financial difficulties domain scores based on smoking status.

4. DISCUSSION

In this study, we evaluated the association between smoking status and HRQOL following RC for bladder cancer. Over the 24-month follow-up period, we observed that current smokers, compared to never smokers, had limited recovery in global QOL scores and slower return to pre-RC pain levels. Following RC, pain scores for current smokers worsened more sharply, took longer to recover to baseline, and then subsequently deteriorated between 18- and 24-months. A similar trend among current smokers was noted for global QOL.

While our findings may reflect a higher frequency of peri-operative complications, more advanced disease, or an increased comorbidity burden among patients who smoke (compared to those who have never smoked), for this to be true we would have expected appreciable differences in other functional domains, including physical functioning. However, never and current smokers recovered to baseline physical functioning by 12-months. Additionally, the rate of adjuvant chemotherapy utilization was not increased amongst current smokers. It is more likely that these findings reflect the complex relationship between tobacco dependence, chronic pain, recovery after surgery, and general quality of life. Previous studies have suggested that patients with chronic pain engage in tobacco smoking as a coping mechanism.¹⁶ Paradoxically, these patients report worse pain levels and greater pain interference in daily activities.¹⁷

When compared to former smokers, we also observed that current smokers experienced worse recovery in global QOL and pain levels throughout the follow-up period. However, these differences were smaller and had wider confidence intervals that crossed zero, indicating a marginal and less conclusive clinical decline. Our data showed some evidence that mental health scores throughout the follow-up period were higher in never and former smokers, compared to current smokers. Again, differences between the former and current smoker groups were not statistically significant. The clinical implication of smoking status on post-operative mental health therefore, remains unclear.

Our findings of worse HRQOL recovery in current smokers can be used in the pre-operative period to inform patient counseling on post-RC convalescence. While our findings do not establish a causal relationship between smoking and worse functional recovery after surgery, with up to 25% of patients who undergo RC being active smokers,¹⁸ the pre-operative period presents an opportune time to address the known oncologic and overall health benefits of smoking cessation. A discussion regarding smoking cessation also affords the urologist an opportunity to address comorbid factors that may influence pain control. Simultaneous smoking cessation and pain intervention programs have previously been shown to be feasible, efficacious, and a method of eliminating tobacco as a means of managing pain.¹⁹

It is important to consider the effects of patient age when assessing the association between smoking status and post-operative outcomes. Haeuser et al. utilized the National Surgical Quality Improvement Program (NSQIP) database to analyze 10,528 patients who underwent RC between 2011 and 2017, of which approximately 22.8% were active smokers.²⁰ They found that as patients aged, smoking was associated with a higher probability of developing a major complication within 30 days of surgery. For patients in their 50s and 60s, there was no difference in observed complications between active smokers and non-smokers. However, for patients in their 70s, active smokers had an increased probability of developing a major complication (OR 1.28, 95% CI 1.12-1.47; p < 0.001). The probability increased even higher for patients in their 80s (OR 1.45, 95% CI 1.17-1.80; p = 0.001). In our study, the median age of current smokers was 61 years, compared to 66 and 68 years for never and former smokers, respectively. The younger age distribution among active smokers could explain why our findings were limited to only two HRQOL domains.

Providing a clinical interpretation of mean score changes on the EORTC QLQ-C30 can be challenging. Previous studies have examined clinically based benchmarks and subjective significance questionnaires to create a generalizable scoring rubric.^{21, 22} However, these methods vary based on the cancer population, treatment offered, and domain being evaluated. Focusing on the patient experience can be the most beneficial. For example, a nine-point increase in global QOL that never smokers reported at 24-months (compared to baseline) is equivalent to overall health (“during the past week”), improving from below average to average.

Our study has several limitations. First, we did not account for smoking duration and intensity among current and former smokers, nor did we account for the length of time since quitting among former smokers. These parameters are important because cumulative smoke exposure has been shown to be associated with worse oncologic and comorbid outcomes.^{2, 23} Patients with a remote and light smoking history are probably not similar to those with an intense smoking history who quit very recently. Prior data show that some peri-operative complications can be mitigated by quitting as close as four weeks before surgery,²⁴ but it is not clear how the recency of quitting influences longer-term HRQOL. This lack of granularity may have introduced heterogeneity within the former smoker group, potentially limiting conclusive findings between former and current smokers. Systematic collection and standardized reporting of tobacco use information, as outlined in the Cancer Patient Tobacco Use Questionnaire,²⁵ are needed in future studies to capture the full extent of cumulative smoke exposure and to explore the presence of heterogeneity between smoking groups. Second, our smoking status exposure was held static throughout the study period, thus reflecting only a patient’s baseline smoking status and not accounting for the small number of patients who may have quit smoking or relapsed during the study period. Despite these limitations, we present a robust description of post-operative HRQOL covering 24-months from the time of RC. This is the first study to evaluate smoking status with HRQOL following RC and can be used in the pre-operative period to counsel active smokers on the potential for worse HRQOL after surgery, as well as to address the known health benefits of smoking cessation.

5. CONCLUSIONS

The negative impact of smoking extends beyond traditional oncologic and surgical outcomes to include worse HRQOL following RC. However, clinically meaningful differences between never and former smokers remain unclear and require additional study. Further research is needed to explore the causal mechanisms for worse recovery in those with a smoking history and to better understand the relationship of smoking exposure and its impact on HRQOL.