Physical Activity among Lung Cancer Survivors: Changes across the Cancer Trajectory and Associations with Quality of Life

Lung cancer is the top cause of cancer mortality in the United States (1). Individuals diagnosed with regional or distant disease have 5-year survival rates of 20.6% and 2.8%, respectively (2). However, among the 16% of lung cancer patients who are diagnosed with localized disease, the 5-year survival rate is 49.5% (2). Most long-term survivors of lung cancer are diagnosed with early-stage non-small cell lung cancer, for which the primary treatment is surgical resection.

During the acute postoperative period, lung cancer survivors commonly experience symptoms such as pain, dyspnea, and fatigue, and impairments in multiple quality-of-life (QOL) domains are evident (3). Although for many individuals, particularly those who remain free of disease, these acute symptoms and impairments gradually improve, deficits in psychological well-being, physical functioning, and role functioning may persist for several years (4). Lung cancer is largely a disease of older adults, with a median age at diagnosis of 71 years (2). Smoking is the leading cause of lung cancer (5), and many lung cancer survivors have comorbid medical conditions (6).

In tandem with the increasing number of cancer survivors, there has been growing interest in documenting and promoting health-enhancing lifestyle behaviors among survivors. There are multiple potential beneficial effects of physical activity for cancer survivors, including improved cardiorespiratory fitness, body composition, physical functioning, fatigue, vigor, and mood (7-9). There is a small amount of research examining physical activity among individuals with lung cancer. Low preoperative exercise capacity among lung cancer patients is associated with a longer postoperative hospital stay (10), and preoperative exercise training or pulmonary rehabilitation has been found to improve patients' cardiorespiratory fitness, functional ability, and pulmonary function (11-13). Small studies of inpatient pulmonary rehabilitation programs for postoperative lung cancer patients have found positive effects on functional ability, peak exercise capacity, and dyspnea (14, 15). In a study of lung cancer survivors at least 5 years postdiagnosis, greater motivational readiness for physical activity was associated with better symptom control and QOL (assessed using single-item questions; ref. 16).

The results of prior research suggest that lung cancer survivors may benefit from engaging in regular physical activity. However, there is a lack of knowledge regarding potential changes in lung cancer survivors' physical activity across the cancer trajectory, and the degree to which physical activity is associated with QOL among short-term survivors of lung cancer is unknown. We addressed these issues in the current study. Specifically, in a cross-sectional survey of non-small cell lung cancer survivors, we examined reported levels of physical activity for three time points (during the 6 months before diagnosis, during the 6 months after treatment, and at the current time). We anticipated that reported levels of activity would be lower during the post-treatment period compared with before diagnosis and at the current time. We also examined associations between current physical activity and QOL, with the expectation that more active individuals would report better QOL.

After obtaining institutional review board approval, a sample of 514 potentially eligible individuals was identified from clinical and research databases at Memorial Sloan-Kettering Cancer Center. Eligibility criteria included diagnosis of primary pathologic stage IA or IB non-small cell lung cancer, receipt of surgical resection for non-small cell lung cancer, from 1 to 6 years post-treatment, no current evidence of cancer, and oncology treating physician permission to contact the individual. An initial review of medical records revealed that 191 of these 514 individuals were ineligible (see below for full ineligibility data). We mailed a consent form and letter inviting study participation to the remaining 323 potential participants, which was followed up 2 weeks later with a telephone call. Individuals who could not be reached by telephone were mailed another invitation letter. Ultimately, a total of 239 of the 514 potentially eligible individuals were found to be ineligible for study participation. The primary ineligibility reasons were current evidence of cancer (n = 73), being >6 years post-treatment (n = 49), deceased (n = 31), or diagnosed with pathologic stage II to IV disease (n = 30). Of the remaining 275 individuals, 175 provided informed consent and participated in the study (response rate = 63.6%). Reasons for study nonparticipation were as follows: passive refusal (n = 32), unable to reach by telephone (n = 28), no reason provided (n = 12), not interested (n = 11), not wishing to talk about lung cancer (n = 11), current medical issues (n = 5), and unable to recall previous activity level (n = 1).

Most participants (n = 148) completed the survey as a telephone interview, with the remainder (n = 27) completing a mailed paper and pencil survey. Results of a series of independent samples t tests and χ² tests indicated that individuals completing the questionnaire by mail were less likely to be currently employed (χ² = 12.65; P = 0.013), were older (t = 2.39; P = 0.018), and reported a lower level of education (χ² = 9.04; P = 0.029) than those completing a telephone interview, but they did not differ on the other demographic factors or any of the medical characteristics listed in Measures.

Demographics. Participants reported their gender, age, race/ethnicity, education, marital status, and employment status.

Medical Characteristics. Electronic databases and medical charts were examined to obtain data on pathologic stage, preoperative pulmonary function [assessed using forced expiratory volume in the first second (FEV1) % predicted], time since treatment completion, type of treatment received, type of surgical resection, length of hospital stay after surgical resection, and presence of postresection complications. Participants completed questions about their current and prior smoking, their current height and weight (from which body mass index was calculated), and their current comorbid medical conditions (17).

QOL. We used multiple self-report measures to fully assess lung cancer survivors' QOL. Participants completed the SF-36v2 (18), which includes subscales assessing eight dimensions of QOL (α values = 0.74-0.95 in the current study): physical functioning, role limitations due to physical problems, bodily pain, general health perceptions, vitality, social functioning, mental health, and role limitations due to emotional problems. Consistent with the SF-36v2 scoring instructions, subscale scores were calculated as norm-based T scores (18). Dyspnea and fatigue were measured using the 3-item Baseline Dyspnea Index (α = 0.91; ref. 19) and 9-item Brief Fatigue Inventory (α = 0.94; ref. 20), respectively. Anxiety and depression symptoms were assessed using the 14-item Hospital Anxiety and Depression Scale (anxiety subscale α = 0.82 and depression subscale α = 0.77; ref. 21).

Physical Activity. We used a modified version of the Godin Leisure-Time Exercise Questionnaire (22) to measure participants' reported physical activity for three time points: during the 6 months before diagnosis of non-small cell lung cancer, during the 6 months after treatment completion, and at the current time. For each time point, participants reported their weekly frequency and average number of minutes spent engaging in each of light (e.g., easy walking), moderate (e.g., brisk walking), and strenuous leisure-time activities (e.g., running). For each time point, we calculated the number of minutes per week participants engaged in each type of activity. We classified individuals as being sedentary if they reported engaging in no light, moderate, or strenuous physical activities. For each time point, we denoted individuals as meeting national physical activity guidelines if they reported engaging in at least 60 min of strenuous activity or ≥150 min of at least moderate intensity activities per week (23). The Godin Leisure-Time Exercise Questionnaire has good test-retest reliability, has shown convergent validity with objective and other self-report physical activity measures (24), and has been used in prior studies of physical activity in cancer survivors (25-27).

We conducted a series of independent samples t tests and χ² tests to examine potential differences in demographic and medical characteristics between individuals who participated in the study and those who declined participation. For each of the continuous physical activity variables (weekly minutes of light, moderate, and strenuous activity), we used linear regression models estimated by generalized estimating equations assuming unstructured correlation matrices to examine whether activity varied across the three time points. We used robust SEs to account for the nonnormality of the physical activity variables. We confirmed that the time main effects were similar when fitting generalized linear models assuming gamma distributions and log links. The gamma models provide more appropriate distributional assumptions for our nonnegative and skewed physical activity data, but the model parameters are not as readily interpretable as those in the linear models we report. For the dichotomous physical activity variables (whether individuals were sedentary and whether they met physical activity guidelines), we examined changes across the three time points using logistic regressions estimated by generalized estimating equations with unstructured correlation structures and robust SEs.

To examine whether changes in activity across the three time points varied according to select demographic or medical factors, we repeated each of the linear and logistic regression analyses including interactions terms of each demographic/medical factor with time (pre-diagnosis, post-treatment, and current time). Given the lack of prior research on factors associated with physical activity changes among lung cancer survivors, we examined a comprehensive set of demographic and medical factors in the aforementioned analyses: gender, age, pathologic stage, preoperative pulmonary function, time since treatment, type of surgical resection, length of hospital stay postresection, number of surgical complications, current smoking status, current weight status, and number of current comorbid medical conditions. We also report the patterns of activity in terms of meeting/not meeting activity guidelines across the three time points.

To further examine how individuals' current level of physical activity varied as a function of their level of activity before diagnosis and during the post-treatment period, we examined cross-tabulations of individuals' current activity status (whether sedentary, engage in some activity but below guideline levels, or meet physical activity guidelines) with their pre-diagnosis and post-treatment activity status. We used a series of linear regression analyses with robust SEs to examine potential differences in QOL according to individuals' current physical activity status. We also conducted these linear regression analyses while controlling for all of the demographic and medical variables. All statistical analyses were two-sided and a cutoff of P < 0.05 was used to determine statistical significance.

Individuals who declined study participation did not differ from study participants with regard to age, gender, pathologic stage, preoperative pulmonary function, time since treatment completion, type of surgical resection, length of hospital stay, or number of post-resection complications (t values ≤ 1.42, χ² values ≤ 5.58, P values ≥ 0.114).

Participants' demographic characteristics are shown in Table 1. Almost two-thirds of participants were female, half were ages ≥70 years, almost all were White, and most had at least some college education. The medical characteristics of the sample are shown in Table 2. Almost all participants were treated with surgery alone and lobectomy was the most common surgical procedure. Few participants (5.8%) reported currently smoking, although 79.8% were former smokers.

As shown in Table 3, reported levels of light activity did not differ significantly across the three time points and averaged ∼80 min/wk. However, reported levels of moderate activity significantly decreased from pre-diagnosis to post-treatment and increased significantly from post-treatment to the current time. Current levels of moderate activity did not differ significantly from those reported for the pre-diagnosis period. Levels of strenuous intensity activity followed the same pattern as for moderate activity, although the mean number of weekly minutes of strenuous activity was low for all three time points (from a high of 24.5 min for the pre-diagnosis period to 8.1 min for the post-treatment period). The percentage of participants who were sedentary (no reported engagement in light, moderate, or strenuous intensity activities) did not differ significantly across the three time points and varied from 1 in 5 participants at the current time to 1 in 4 during the post-treatment and pre-diagnosis periods. Just over 1 in 4 participants met physical activity guidelines during the pre-diagnosis period. This rate significantly decreased to ∼1 in 8 participants for the post-treatment period and significantly increased back to just over 1 in 4 participants for the current time.

There were no statistically significant differences in the prevalence of physical activity across the three time points for any of the activity measures for the following variables: gender, age, pathologic stage, time since treatment, type of surgical resection, or current smoking status. Table 4 summarizes the variables for which differences were identified in one or more physical activity measures across the three time points. Individuals with fewer comorbid medical conditions increased their engagement in light intensity activities during the post-treatment period, whereas those with more comorbidities decreased such activities after treatment. Individuals with better preoperative pulmonary function (higher FEV1 % predicted) decreased their level of strenuous activities during the post-treatment period. Examination of absolute levels of strenuous activity indicated that those with worse pulmonary function were not engaging in such activities before diagnosis and thus a lack of post-treatment change in their strenuous activities is not surprising.

From the post-treatment period to the current time, individuals with a shorter post-resection hospital stay or fewer surgical complications increased their engagement in strenuous activities, whereas those with a longer hospital stay or more complications did not. For the post-treatment period, individuals having worse preoperative pulmonary function or reporting more comorbid medical conditions were more likely to have become sedentary than individuals with better pulmonary function and those with fewer comorbidities. Comparing the current time with the post-treatment period, individuals who were normal weight were less likely to have become sedentary, but there were no differences for overweight or obese individuals.

The pattern of participants' physical activity in terms of meeting or not meeting physical activity guidelines across the three time points is shown in Table 5. Classification labels for the physical activity patterns were drawn from prior research (28, 29). Almost two-thirds of participants did not meet physical activity guidelines for any of the time points and were denoted as nonexercisers. Almost all of the remaining participants fell equally into one of four activity patterns: met guidelines at all three time points (maintainers), did not meet guidelines during the pre-diagnosis or post-treatment periods but currently met guidelines (adopters), met guidelines during the pre-diagnosis and current time periods but did not meet guidelines during the post-treatment period (temporary relapsers), and met guidelines during the pre-diagnosis period but not during the post-treatment period or currently (permanent relapsers).

Among individuals who were currently sedentary, 62.9% were sedentary and 8.6% met activity guidelines before diagnosis and 57.1% were sedentary and 2.9% met activity guidelines during the post-treatment period. Among individuals who currently met activity guidelines, 6.4% were sedentary and 66.0% met activity guidelines before diagnosis and 16.7% were sedentary and 35.4% met activity guidelines during the post-treatment period.

Table 6 shows the results of analyses examining differences in QOL according to participants' current physical activity status. We first describe the results from the unadjusted analyses. With regard to the SF-36 subscales, compared with less active individuals, those who met physical activity guidelines reported better QOL in terms of physical functioning, role limitations due to physical problems, general health perceptions, and vitality. There were no differences in QOL according to participants' current physical activity status for the SF-36 subscales for bodily pain, social functioning, mental health, and role limitations due to emotional problems. Individuals who met physical activity guidelines reported significantly less dyspnea than sedentary individuals. Additionally, those who met activity guidelines reported a significantly lower level of fatigue than individuals who engaged in low activity. There was no difference in anxiety symptoms according to physical activity status, but individuals meeting activity guidelines reported significantly fewer depression symptoms than sedentary individuals and those who engaged in low activity. The results of the adjusted analyses were largely consistent with those from the unadjusted analyses, indicating that differences in QOL according to physical activity status were not fully explained by potential confounding demographic or medical variables. Of note, all of the statistically significant differences in QOL between sedentary individuals and those who met activity guidelines in the unadjusted analyses were also significant in the adjusted analyses.

Consistent with our expectations, lung cancer survivors reported engaging in fewer minutes of moderate and strenuous intensity activity during the early post-treatment period than before diagnosis and at the current time. This is likely attributable to the typically high prevalence of acute post-treatment symptoms such as pain, dyspnea, and fatigue (3). It was encouraging, however, that survivors' current level of moderate and strenuous intensity activities did not differ from levels reported for the pre-diagnosis period. Individuals with a shorter postoperative hospital stay or fewer surgical complications were more likely to increase their engagement in strenuous activities from the post-treatment period to the current time. This may be due to these individuals having better physical functioning or higher exercise capacity than individuals experiencing a longer hospital stay or more surgical complications, although future empirical research is needed to examine this issue.

Across the full sample, participants' reported engagement in light intensity activities and the percentage of individuals who were sedentary did not differ for the three cancer trajectory time points. However, individuals with more comorbid medical conditions reported a lower prevalence of light intensity activities and were more likely to become sedentary during the post-treatment period. Additionally, individuals with poorer preoperative pulmonary function were more likely to be sedentary during the post-treatment period. This suggests that individuals with more comorbidities or poorer preoperative pulmonary function may be especially in need of post-treatment rehabilitative support and assistance with regard to regaining activity levels.

As anticipated, the percentage of lung cancer survivors who engaged in sufficient activity to meet activity guidelines was lower during the post-treatment period compared with before diagnosis and at the current time. Just under two-thirds of participants did not meet guidelines at any of the three time points. Fewer than 1 in 10 participants met activity guidelines at all three time points. Further assessment of participants' activity status across the three time points indicated a large degree of consistency between individuals' current activity status and their activity status before diagnosis. These results suggest that after the acute post-treatment period, in which engagement in moderate and strenuous activities is uniformly low, most lung cancer survivors resume the level of physical activity (or inactivity) that they engaged in before their lung cancer diagnosis. Approximately 1 in 4 (27.4%) participants currently met physical activity guidelines, which is consistent with evidence that activity guidelines are met by 30.4% of U.S. adults ages 45 to 64 years, 27.0% of those ages 65 to 74 years, and 16.8% of those ages ≥75 years (30).

Individuals who currently met physical activity guidelines reported a higher QOL in several domains compared with less active individuals. In line with the results of several studies in other cancer survivor populations (26, 31, 32) as well as studies in the general adult population (33), there was greater evidence that activity was associated with physical as opposed to mental or emotional QOL domains. However, individuals meeting activity guidelines reported fewer depressive symptoms, suggesting the importance of examining mood as a potential barrier for, or benefit of, physical activity among lung cancer survivors. Across the QOL measures for which statistically significant differences were observed, the magnitude of QOL differences (e.g., in the unadjusted analyses, effect sizes from d = 0.40 to d = 0.80 comparing sedentary individuals and those who met activity guidelines) are suggestive of clinically meaningful differences (34). Overall, combined with the known QOL benefits of physical activity for other cancer survivor populations (7-9), the current study results suggest that interventions that successfully promote lung cancer survivors' engagement in physical activity may increase their QOL in several domains.

Strengths of the study include no indication of differences in available demographic and medical factors between study participants and decliners, a response rate (63.6%) that compares favorably with response rates from similar prior studies (e.g., refs. 25-27, 32, 35, 36), availability of data regarding key pre-resection and post-resection medical factors, and examination of the associations between physical activity and multiple QOL domains. Several limitations to the study should be noted. The cross-sectional research design prohibits examination of the causal direction of associations between physical activity and QOL. Physical activity for the pre-diagnosis and post-treatment periods was measured retrospectively, raising issues of potential recall inaccuracy. Prior studies have documented acceptable reliability and validity of self-reported assessments of physical activity over diverse historical periods, including up to 10 years and beyond (e.g., refs. 37-40), but no prior research has examined this issue in survivors of lung cancer. Lynch et al. (41) examined the test-retest reliability of reported pre-diagnosis physical activity among colorectal cancer survivors. The intraclass correlation coefficients for moderate intensity physical activity and activity categories (sedentary, low activity, and meet activity guidelines) were 0.77 (95% confidence interval, 0.68-0.84) and 0.62 (95% confidence interval, 0.48-0.76), respectively, and the overall reliability of the measure was comparable with that in a population-based sample of adults. However, future research is needed to determine the extent to which the reliability and validity of retrospective physical activity assessments may vary across survivors of varying cancers and according to survivors' demographic and medical characteristics (41). Although physical activity was assessed with a widely used, validated self-report measure, future research would benefit from the addition of objective activity measures. The study participants were mostly well educated and there was limited variability with regard to their race/ethnicity. Thus, the study results may not be generalizable to lung cancer survivor populations with other demographic characteristics.

There is a lack of research regarding the prevalence, patterns, and QOL correlates of physical activity in lung cancer survivors. The current study examined these key issues and fills an important knowledge gap in lung cancer survivorship research, which as a whole is underrepresented within the burgeoning survivorship literature (3). Results suggest that physical activity among lung cancer survivors is particularly low during the early post-treatment period. Current engagement in physical activity is associated with better QOL in several domains. However, the vast majority of lung cancer survivors do not currently meet physical activity guidelines and may benefit from interventions to promote engagement in regular physical activities. With the developing formulation of long-term follow-up care for survivors of lung and other cancers, the current findings suggest that oncologists have an opportunity to recommend increased physical activity as a means of promoting post-treatment QOL. There is a need to develop and test physical activity interventions designed specifically for survivors of early-stage lung cancer. Such interventions should take into account these individuals' tendency to engage in activities of light or moderate intensity as well as their usual older age, frequent comorbid medical conditions, and common physical and QOL impairments. These interventions might also represent a core component of broader pulmonary rehabilitation programs for individuals with lung cancer. In light of documented benefits of such programs with regard to QOL, symptom control, and exercise capacity for individuals with lung diseases such as chronic obstructive pulmonary disease (42), there is reason for optimism for their utility for lung cancer survivors.

No potential conflicts of interest were disclosed.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank Melissa Ozim and Syncia Sabain for assistance with data collection and management, Paul Krebs for assistance with data processing, and the study participants for their valued contribution to this research.