Journal of Cancer Survivorship

, Volume 2, Issue 4, pp 262–268

Predicting recreational difficulties and decreased leisure activities in women 6–12 months post breast cancer surgery

Authors

    • Dalhousie University Family Medicine Teaching UnitDr. Everett Chalmers Regional Hospital
  • Ryan Hamilton
    • Dalhousie University Family Medicine Teaching UnitDr. Everett Chalmers Regional Hospital
  • Sue Tatemichi
    • Dalhousie University Family Medicine Teaching UnitDr. Everett Chalmers Regional Hospital
  • Roanne Thomas-MacLean
    • University of Saskatchewan
  • Anna Towers
    • Faculty of Medicine, McGill University
  • Thomas F. Hack
    • Faculty of Nursing, University of Manitoba
  • Andrea Tilley
    • Atlantic Health Sciences Corporation
  • Winkle Kwan
    • British Columbia Cancer Agency
Article

DOI: 10.1007/s11764-008-0068-8

Cite this article as:
Miedema, B., Hamilton, R., Tatemichi, S. et al. J Cancer Surviv (2008) 2: 262. doi:10.1007/s11764-008-0068-8

Abstract

Introduction

A Canadian research team is conducting a multi-centered, non-interventional national study with the objective of charting the course of arm morbidity after breast cancer surgery. This paper examined the relationship between arm morbidity and leisure and recreational activities of affected women.

Methods

Five hundred and forty seven women with stage I-III breast cancer were recruited in four centers across Canada: Surrey (BC); Winnipeg (MB), Montreal (QB) and Fredericton (NB). Participants were enrolled in the study 6–12 months post surgery. Physical examination was used to assess arm and shoulder functioning and questionnaires were used to assess disability, pain, and participation in recreational and leisure activities.

Results

At the first clinical assessment (T1), the mean number of months post breast cancer surgery was 8.4. At T1 49% of women reported difficulty with recreational activities that involved “some force or impact” and 29% experienced negative changes to their involvement in leisure activities. A hierarchical multiple regression analysis found that several arm morbidity variables were significant predictors of difficulty with participation in recreational activities. A second hierarchical regression found also that arm morbidity factors were significant predictors of negative changes in leisure activities. Follow-up analyses found that arm morbidity, was most closely related to difficulty with recreational activities requiring free movement of the arm and using force.

Conclusion

Many women treated for breast cancer experience arm morbidity. Arm morbidity is related to difficulties with recreational activities and negative changes in leisure activity participation.

Implications

Breast cancer survivors should engage in recreational and leisure activities that are compatible with reduced range of motion and pain, and avoid those that exacerbate their arm morbidity.

Keywords

Breast cancerArm morbidityRecreational and leisure activities

Introduction

With advances in early detection and treatment strategies for breast cancer, the population of survivors is growing rapidly as incidence and mortality rates decline. Although most treatment options available today are beneficial and lifesaving, they are associated with physical, psychological and social sequelae. Surgery, lymph node dissection and radiation can all have long-term effects on arm and shoulder function and can cause pain, lymphedema, heaviness and a limitation in range of motion creating serious arm morbidity [8, 21]. Since there is no cure for arm morbidity, it is important to emphasize prevention; however, prevention is difficult when we lack a full understanding of its multifactorial aetiology [17].

Breast cancer survivors’, just like all cancer survivors’ return to day-to-day activities represents a return to ‘normal’ and may be a crucial component for a satisfactory quality of life [16]. A component of these day-to-day routines are recreational and leisure activities. However, it is challenging for breast cancer survivors to know which recreational or leisure activities are safe to engage in and which may trigger or exacerbate arm morbidity. Breast cancer survivors may receive conflicting messages; some agencies advocate for and others advocate against strenuous physical activities [2, 6]. In the late 1990’s with the introduction of “Dragon Boat Racing” for breast cancer survivors, it was argued that strenuous physical activity was not related to arm morbidity such as lymphedema and women were encouraged to engage in strenuous activities [10, 15] however, other studies caution against strenuous activities [23].

Current estimates of the prevalence of arm morbidity (range of motion problems, pain or lymphedema) resulting from breast cancer treatment vary greatly depending on the definitions and measurements used (self-reports or clinical measurements). Based on several studies as discussed in a review article, the prevalence of arm morbidity, including pain, range of motion problems and swelling (lymphedema) ranges from 5%–80% [5, 9, 18, 19, 22]. Most of the studies, however, were based on self reported surveys regarding swelling, ROM problems and pain. Few studies are prospective in nature and include clinical assessments of range of motion measurements [18].

To the best of our knowledge, no prospective study has been completed that examines the incidence rates of clinically determined arm morbidities such as lymphedema, pain, and changes in arm ROM (abduction and rotation) as well as self reported social issues such as disability, and participation in leisure and recreational activities after breast cancer surgery. We feel that this data will help provide a clearer picture of the relationship between arm morbidity in breast cancer patients and participation in recreational and leisure activities.

Methods

Data collection

Collection of data consisted of three components: review of medical chart information, clinical assessment and the administration of surveys. Medical chart reviews were carried out to supplement and corroborate medical information regarding diagnosis and treatment reported by the participants. Assessments are carried out twice a year, one clinical and telephone interview assessment, for up to 5 years. The first assessment (T1) was clinical and all measures were administered including: range of motion, swelling, height and weight, Profile of Moods (POMS), Quick-Disabilities of Arm, Shoulder & Hand (Quick DASH), Short Form McGill Pain Questionnaire (MPQ-SF) and a survey that was developed for this study that looks at medical, demographic and social information. For the analysis in this paper, we are using the range of motion, swelling and body mass index variables as well as components from the Quick-DASH and the MPQ-SF and the social survey. All the data was entered in the statistical computer program SPSS 15.0. This paper is based on T1 data (cross sectional) and a select number of variables and measures.

Measures

In this study, arm morbidity was measured in three ways: (1) presence of lymphedema, (2) range of motion (ROM), and (3) a validated measure of the presence of pain (MPQ-SF) [7]. Lymphedema was operationalized as a percentage of arm volume in excess of that of the unaffected side. This involved obtaining sequential circumferential arm measurements which were entered into a software package that uses a formula to calculate arm volume in millilitres. The unaffected arm volume was compared to the affected arm volume to provide a percentage volume increase of the affected arm [11]. We followed the International Society of Lymphology definition of lymphedema [1], but took into account the natural variability in size between the two arms: < 10% excess was considered no lymphedema; between 10% and 19% excess volume was considered to be minimal lymphedema; 20–40% was considered moderate and greater than 40% was considered severe lymphedema. Range of motion (ROM) was assessed through measures of shoulder abduction and external rotation. This approach was congruent with a prior study completed by one of our co-investigators [14]. Research associates were trained to complete a goniometric assessment of ROM. The impairment in the affected arm was measured through comparison with the unaffected arm (degrees of motion lost). Two measures were taken, one to assess available ROM and the other to assess the sensation point for discomfort. In order to assess the relationship between clinically measured arm morbidity and difficulties with recreational activities, we administered the Quick-DASH [4]. In addition, the study team developed a social questionnaire that included items measuring both past and current recreational and leisure activity participation.

Recruitment

The women were recruited from four sites across Canada: Surrey (British Columbia), Winnipeg (Manitoba), Montreal (Québec) and Fredericton (New Brunswick). In three sites women were recruited through cancer clinics and at the last site, they were recruited in surgeons’ offices because of the lack of a provincial cancer centre. The inclusion criteria for recruitment were: 1) diagnosis of stage I to III breast cancer; 2) breast surgery a maximum of 12 months prior to first assessment; 3) ability to communicate in either English or French; 4) 18 years of age; and 5) ability to give informed consent. Exclusion criteria were: women with in-situ, Stage IV or bilateral disease. The research protocols were approved by research ethics boards in all four sites.

Analysis

In addition to frequency calculations, two hierarchical multiple regression analyses were conducted to assess the strength of the relationship between the independent and dependent variables. Both regressions had the same set of independent variables/predictors with 1) range of motion 2) working for pay, 3) presence of pre-existing arm problems, 4) axillary dissection, 5) surgery type (radical or modified radical mastectomy), 6) private health insurance, 7) age, 8) BMI, 9) present pain, and 10) lymphedema entered on the first step. The interaction between present pain and range of motion was entered on the second step. The dependent or criterion variables were “difficulties with recreational activities” and “negative changes to leisure activities.”

Three follow-up simultaneous regressions were conducted using only the arm morbidity variables (range of motion, pre-existing arm problems, present pain, lymphedema) as predictors and DASH 17 (difficulty with low impact recreational activities), DASH 18 (difficulty with recreational activities requiring force or impact), and DASH 19 (difficulty with recreational activities requiring free movement of the arm) as the criterion variables.

Results

The average time elapsed between surgery and assessment (T1) for the 5741 women in the study was 8.4 months. More than 88% of the women had either Stage I or II breast cancer. The majority of the participants lived in urban areas. Surgical procedures included modified radical mastectomy (22%), lumpectomy (76%) or radical mastectomy (2%). Of the women, 98% underwent lymph node dissection, 94% were treated with radiation and 75% of the women were taking hormonal therapy, usually Tamoxifen or Exemestane (see Table 1). At T1 almost a quarter of the women did not engage in exercise that involved the affected arm and less than half exercised between 4 and 7 times weekly involving their affected arm.
Table 1

Demographic, diagnostic and treatment participant profile

Demographics

n

%a

Surrey (British Columbia)

164

28

Winnipeg (Manitoba)

84

15

Montreal (Quebec)

295

52

Fredericton (New Brunswick)

31

5

Mean Age = 54.29 (SD = 11.95)

  

Living in a urban area

447

84

Income < $40,000

115

33

Income between $ 40,001–$80,000

178

38

Income > $ 80,001

137

29

No exercise using arms

128

24

Exercise 1–3 times weekly using arms

149

28

Exercise 4–7 times weekly using arms

257

48

Diagnostic stage

  Stage I

243

43

  Stage II

256

45

  Stage III

64

11

Treatments

  Sentinel lymph node dissection only

150

27

  Sentinel & axillary lymph node dissection

132

24

  Axillary lymph node dissection only

273

49

aPercentages do not always add up to 100 due to rounding and missing data

Independent variables

Eleven independent variables (one of which was an interaction) were used as predictors in the regression analyses. These variables are described in Table 2. More than one third (34.5%) of the women reported that they experienced current pain on the McGill Present Pain Index. More than half (55%) of the women had restrictions in lateral rotation of the affected arm and 47% had reduced abduction of the affected arm. Mild lymphedema was found in 14 women (2.4%) and moderate to severe lymphedema was discovered in 3 women (0.5%) for a total of 17 cases or 3%. Early evidence in our study shows that the four types of arm morbidity are discrete conditions at T1. A composite arm morbidity variable has been created using the dichotomizations values of “present pain” (yes/no); “lymphedema” (<10% = no, > 10% = yes) “abduction restriction” (yes/no) and “rotation restriction” (yes/no). The composite variable was created to reflect the fact that women were largely reporting pain, ROM restrictions and lymphedema at the first point of data collection [24]. Overall, 73% of the women reported some form of arm morbidity.
Table 2

Independent variables

Continuous variables

Potential range

Actual range

Mean / SD

1

Range of motion restriction level

1–3

1–3

1.97 (0.82)

2

Present pain

1–5

1–5

1.53 (0.84)

3

Age

NA

28–85

54.67 (11.09)

4

BMI

NA

17–49

26.78 (5.02)

Dichotomous variables

# Yes (%)

# No (%)

 

5

Private Health Insurance

428 (75)

143 (25)

 

6

Working for Pay

175 (30.5)

398 (69.5)

 

7

Pre-existing arm problems

202 (35.6)

365 (64.4)

 

8

Axillary Dissection

405 (73)

150 (27)

 

9

Radical or modified radical mastectomy

139 (24.3)

432 (75.7)

 

10

Lymphedema

17 (3)

557 (97)

 

Interaction

Potential Range

Actual Range

Mean / SD

11

Range of motion restriction & present pain Interaction

1–15

1–15

3.17 (2.42)

Recreational activities

Three of the DASH questions (17, 18 and 19) asked about the women’s involvement in recreational activities. Question 17 inquired about participation in activities that require “little effort” (e.g., playing cards); question 18 asked about activities that require “some force or impact” (e.g., playing golf or tennis) and question 19 asked about participation in activities where the affected arm is “used freely” (e.g., playing Frisbee). Few women (8.2%) reported difficulty participating in recreational activities that involved “little effort”. Almost half of the women, (48.5%) reported that they had difficulties carrying out activities that required “some force or impact” and 44% of the women reported that they experienced difficulty partaking in activities that involved “moving their arm freely.” We have used the total score of the three aforementioned DASH variables in the first regression. The potential range on this composite variable was 3–15 (actual range observed in study 3–14) and a mean score of 4.76 (SD = 2.23) was obtained reflecting mild disability.

Negative change in leisure activities

Twenty-nine percent (n = 169) of the women had experienced a negative change in their degree of participation in leisure activities since breast cancer treatment. When asked what kind of change of activity that involved, 60% of these women (n = 101) said “they slowed down a bit, but have not stopped”. The leisure activities in which the women were involved were varied and included a number of sports (e.g., swimming and cycling), sculpting and painting.

Regression analyses

The first analysis used the “difficulties with recreational activities” variable as the criterion. The set of independent variables entered on the first step were found to have a predictive relationship with the “difficulties with recreational activities” variable and explained a significant amount of variance (R2 = .241 (24%), F = 16.84, p < .001). The F change from step 1 to step 2 was examined to see if the interaction between pain and range of motion had predictive value beyond the set of 10 independent variables entered on the first step. The F square change was not significant (F Change = 0.055, p = .849) and as a result only the first step was analyzed in greater detail. Examination of the semi-partial correlations revealed that five of the ten predictors had significant unique contributions in explaining the variance in difficulties with recreational activities. The presence of pre-existing arm problems (semi-partial correlation = .130, p = .001) and the arm morbidity variables “present pain” (semi-partial correlation = .259, p < .001), “presence of lymphedema” (semi-partial correlation = .096, p = .011), and “restricted range of motion” (semi-partial correlation = .270, p < .001) all had significant unique relationships with recreational difficulties. Thus, increased pain and restricted ROM were positively related to increased difficulty with recreational activities. Age, on the other hand had a significant negative relationship with the criterion (semi-partial correlation = −.085, p = .026); indicating older women reported fewer problems with recreational difficulties.

The second hierarchical multiple regression analysis used “negative changes in leisure activities” as the criterion variable. The set of independent variables entered on the first step were found to have a predictive relationship with the “negative changes in leisure activities” variable and explained a significant amount of variance (R2 = .090, 9.0%), F = 5.21, p < .001). The F change from step 1 to step 2 was examined to see if the interaction between pain and range of motion had predictive value beyond the set of 10 independent variables entered on the first step. The F square change was not significant (F Change = 0.004, p = .948) and as a result only the first step was analyzed in depth. Examination of the semi-partial correlations revealed that three of the ten predictors had significant unique contributions in explaining the variance in negative changes in leisure activities. Whether a participant was working for pay had a significant relationship with negative changes to leisure activities (semi-partial correlation = −.113, p = .007) with those still working for pay experiencing fewer negative changes. Two arm morbidity variables were found to have significant positive relationships with “negative changes to leisure activities” with “present pain” (semi-partial correlation = .178, p < .001) and “restricted range of motion” (semi-partial correlation = .092, p < .027) each making unique contributions.

Follow-up analyses

The arm morbidity variables emerged as significant and interesting predictors in the first two regression analyses. As a result we decided to conduct three follow-up regression analyses using the arm morbidity variables (range of motion, lymphedema, pain) and pre-existing arm problems as predictors. All four predictors were entered on the same step in each analysis. The criterion variables in these follow-up analyses were the different types of recreational difficulties measured by the DASH.

The first follow-up simultaneous regression used the four predictor variables discussed above and DASH 17 (difficulty with activities requiring little effort) as the criterion. The four independent variables were found to be significant predictors of difficulties with recreational activities requiring little effort (R2 = .093, F = 14.45, p < .001). An analysis of the semi-partial correlations found that all four variables had significant unique predictive ability (pre-existing arm problems semi-partial correlation = .114, p = .005; present pain semi-partial correlation = .113, p = .005; lymphedema semi-partial correlation = .181, p < .001; restriction of range, semi-partial correlation = .134, p = .001), although the semi-partial correlations were relatively weak compared to those observed for the more strenuous activities described below.

The second follow-up simultaneous regression used the same four predictors and DASH 18 (difficulty with activities requiring some force or impact) as the criterion. The four independent variables as a set were found to be significant predictors of difficulties with recreational activities requiring little effort (R2 = .189, F = 32.74, p < .001). An analysis of the semi-partial correlations found that two of the four variables had significant unique predictive ability. Present pain was found to have a significant positive semi-partial correlation (semi-partial correlation = .255, p < .001) with recreational difficulties requiring force and/or impact as did restriction of range (semi-partial correlation = .264, p < .001).

The final follow-up simultaneous regression used the same four independent variables and DASH 19 (difficulty with activities requiring moving the arm freely) as the criterion. The four predictors as a set were found to be significant predictors of difficulties with recreational activities requiring free movement of the arm (R2 = .211, F = 37.49, p < .001). An analysis of the semi-partial correlations found that three of the four variables had significant unique predictive abilities. Present pain was found to have a significant positive semi-partial correlation (semi-partial correlation = .278, p < .001) with recreational difficulties requiring force and/or impact as did restriction of range (semi-partial correlation = .253, p < .001) and pre-existing arm problems (semi-partial correlation = .113, p = .003).

Discussion

In our study, 73% of women 6–12 months post breast cancer surgery either reported pain or were found to have ROM restrictions in the affected arm; however, the incidence of lymphedema was low. Arm morbidity had a statistically significant relationship with recreational and leisure activities. Pain and ROM restrictions had the most significant relationship with recreational activity difficulty, particularly those activities involving “some force or impact” or while using the arm “freely.” There is a range of variables that can explain difficulties with recreational activities; however, having arm pain and/or restricted range of motion appears to be of much importance in trying to explain why some individuals modify their recreational pursuits. Age had a negative relationship with difficulties with recreational activities. This finding may be explained by a possible natural reduction in more strenuous recreational activities as people age. Surprisingly, “type of surgery,” “type of node dissection,” “BMI” and “having supplemental health insurance” did not have a significant unique relationship with the experience of “difficulties while engaged in recreational activities.”

Arm morbidities (present pain and ROM restrictions) were significantly (statistically) related to “negative change in leisure activities.” The good news is that the majority of these women reported that they had merely “slowed down and not stopped” leisure activities. A regression analysis was only able to explain 9% of variance on the negative changes in leisure activities variable. Although statistically significant, in real life it can be argued that arm morbidity does not have a huge impact on day-to-day leisure activities of the breast cancer survivors in this study. Women who work for pay seem to have less negative change in leisure activities. It is difficult to ascertain the directional relationship between working and “negative change in leisure activities;” however, we assume that women who went back to work, be it part- or full-time felt mentally and physically better and therefore reported less “negative change in leisure activities.” “Type of surgery,” “type of node dissection,” “BMI”, and “age” were not associated with negative change to leisure activities.

Using objective ROM restriction measures and self-reported pain measures it was found that women with arm morbidity were likely to experience difficulties with all types of recreational activities. The relationship between arm morbidity and difficulties with recreational activities was more pronounced when increasingly strenuous recreational activities were considered. Hence, we believe that women who do not have arm morbidity (e.g., pain, ROM restrictions) can enjoy strenuous activities such as Dragon Boat Racing; however, those who do have arm morbidity may want to refrain from more strenuous types of recreational activity.

For all breast cancer survivors, recreational and leisure activities offer much in terms of physical exercise, health maintenance, social interaction, support and personal meaning [13, 20] The opportunity to partake in leisure or recreational activities for breast cancer survivors may be an important priority in the management of their follow-up care. We suggest that women engage in activities that will not exacerbate their arm morbidity, but will ameliorate their pain and range of motion restrictions. These women may want to work under the guidance of physiotherapists or other physical health specialists to responsibly engage in various physical activities. Many women in our sample did not engage in any exercise at all involving the affected arm, and less than half indicated that they engaged in exercise involving the affected arm on a regular basis. Engagement in physical activities is an important aspect of overall health and well-being and an important aspect of being ‘normal’ again.

Our study outcomes differ somewhat from other studies of the functional sequelae of breast cancer treatments. In a study by Andrykowski et al. [3], it was reported that breast cancer survivors had returned to their pre-baseline activities two to six months following the completion of their adjuvant breast cancer therapy. Our study findings are more in accordance with the outcomes of the Kärki et al. study [12]. These authors argue that many breast cancer survivors, 6 months post surgery, experience impairment of the affected arm and some had terminated all involvement in recreational and leisure activities. Our findings were also similar to those of Shimozuma et al. [21] who reported a decline in recreational and social activities in women who had undergone modified radical mastectomy by 18.5% and those after breast conserving treatment by 24% one year post surgery. The problems reported by Kärki et al. [12] were somewhat different from the problems documented in our study. They reported that 6 months post-surgery the women in their study complained primarily of scar tightness, axillary edema and neck–shoulder pain and at 12 months post-surgery limb numbness replaced the axillary edema as a more common complaint [12]. There was concurrence between our study and that of Kärki et al. in the finding that some women terminated certain activities, and most lowered the level of participation in recreational and leisure activities.

Limitations

The women in this study were only 6–12 months post surgery and some may still have been receiving adjuvant therapy. The importance of this will be borne out as the study progresses and we will be able to track women’s level of participation in recreational and leisure activities over time, thus at this time it is not possible to determine causation.

Conclusion

Almost three-quarters of breast cancer survivors 6–12 months post breast cancer surgery presented with arm morbidity. Arm morbidity in the form of lymphedema but primarily pain and range of motion restriction was found to be related to increased difficulty with recreational activities and negative changes to leisure participation. The relationship between arm morbidity and recreational difficulties was the strongest for those activities requiring force or impact and free movement of the affected arm. The arm morbidity variables had a stronger relationship with recreational activity difficulties than with negative changes to leisure activities. This may indicate that despite difficulty, participants were still engaging in their regular activities or that they had found alternative activities to replace those causing difficulty. Breast cancer survivors who have pain and ROM restrictions should engage in recreational and leisure activities that ameliorate pain and range of motion problems and not exacerbate them. Finally, this study points to the need for a well designed randomized controlled trial to examine in greater detail the relationships between treatment and arm functioning.

Implications for cancer survivors

A large number of women in this study presented with arm morbidity 6 to 12 months post surgery. Because the arm morbidity created difficulties with recreational and (some) leisure activities, we believe that it is important that breast cancer survivors engage in activities based on their comfort level. Physical therapists may benefit from being cognizant of arm morbidity and thus assist breast cancer survivors in activities that help ameliorate pain and ROM restrictions.

Footnotes
1

Of the sample, 16% were diagnosed outside the 6 and 12 months post surgery range: 60 (10%) women <6 months and 35 (6%) >12 months. Statistics are based on all 574 women.

 

Acknowledgement

Canadian Institutes of Health Research (CIHR) MOP 68883

Copyright information

© Springer Science+Business Media, LLC 2008