A Survey of Joint and Muscle Aches, Pain, and Stiffness Comparing Women With and Without Breast Cancer

Transcription

1 Vol. 46 No. 4 October 2013 Journal of Pain and Symptom Management 523 Original Article A Survey of Joint and Muscle Aches, Pain, and Stiffness Comparing Women With and Without Breast Cancer Deborah Fenlon, PhD, RGN, Julia M. Addington-Hall, PhD, BA, Alison C. O Callaghan, MPhil, BSc, BA, Joanne Clough, MBBS, DM, FRCA, MRCP, FRCPCH, Peter Nicholls, PhD, MSc, BSc, and Peter Simmonds, FRACP, MBBS Faculty of Health Sciences (D.F., J.M.A.-H., A.C.O., P.N.), University of Southampton; and Cancer Research UK Clinical Unit (P.S.), Cancer Sciences Division, University of Southampton and Southampton University Hospitals NHS Trust, Southampton, United Kingdom Abstract Context. Joint and muscle aches, pain, and stiffness have been reported to be a problem for some women after adjuvant breast cancer treatment; however, the extent and impact of this problem are unknown. Objectives. The purpose of this study was to determine the prevalence of this problem in comparison with women of a similar age without breast cancer. Methods. Two hundred forty-seven women attending breast cancer follow-up clinics were invited to complete pain and quality-of-life measures. A comparison group of 274 women of similar age was drawn from women attending breast screening and benign breast clinics. Prevalence and severity of pain were compared between the two groups. Results. The mean age of all women in the study was 59 years (range 30e86 years). The median time since diagnosis of cancer was 28 months (range 2e184 months). Adjuvant treatments included radiotherapy (79%), chemotherapy (45%), and hormone therapy (81%). Sixty-two percent of women with breast cancer reported pain today compared with 53% of women without breast cancer (P ¼ 0.023). Significant predictors of pain in both patient groups were cancer, age, and arthritis. For the cancer cases, significant predictors of pain were age, arthritis, taxane chemotherapy, aromatase inhibitors, and tamoxifen. Quality of life (measured by the Short Form-36) was significantly worse for women with breast cancer compared with controls and was significantly worse in the breast cancer cases with pain. Conclusion. Treatment with tamoxifen, taxane chemotherapy, and aromatase inhibitors for breast cancer is predictive of joint pain, which may have an impact on women s lives for some years after breast cancer. J Pain Symptom Manage 2013;46:523e535. Ó 2013 U.S. Cancer Pain Relief Committee. Published by Elsevier Inc. All rights reserved. Address correspondence to: Deborah Fenlon, PhD, RGN, Faculty of Health Sciences, Highfield Campus, University of Southampton, Southampton Ó 2013 U.S. Cancer Pain Relief Committee. Published by Elsevier Inc. All rights reserved. SO17 1BJ, United Kingdom. dfenlon@ soton.ac.uk Accepted for publication: October 23, /$ - see front matter

2 524 Fenlon et al. Vol. 46 No. 4 October 2013 Key Words Arthralgia, breast cancer, joint pain, survey, tamoxifen, aromatase inhibitors Introduction Breast cancer now affects one in eight U.K. women, 1 and almost two-thirds of newly diagnosed women are now likely to survive for at least 20 years. 2 A greater proportion of women with early breast cancer are now receiving adjuvant chemotherapy and/or hormone therapy. There are reports in the literature that, after primary breast cancer treatment, up to threequarters of women may complain of joint aches, pain, and stiffness, 3 and it is becoming apparent that this problem is a more important clinical issue than has been realized to date. These problems are being reported more frequently because the use of aromatase inhibitors (AIs) has been taken up more widely. Adjuvant AI treatment has been found to be more effective in preventing breast cancer recurrence than tamoxifen 4 and is now the treatment of choice for adjuvant therapy in postmenopausal women, after primary treatment for hormone receptor-positive breast cancer. Joint pain may cause or contribute to patients stopping treatment; an online survey conducted by Breast Cancer Action found that 30% of respondents reported that they had discontinued adjuvant AI use because of adverse effects, of whom 47% did so because of joint-related problems. 4 A cohort study in California showed that 61% of patients on AIs developed AI-related arthralgia, and 20% of women stopped AI treatment because of persistent joint pain. 5 Theories about the cause of this pain largely relate to early menopause and low estrogen levels, which indicate that the use of AIs may exacerbate this problem. The ATAC (Arimidex Ò, Tamoxifen, Alone, or in Combination) study reported joint symptoms in women taking Arimidex (anastrozole) of 35.6% compared with 29.4% in women taking tamoxifen. 6 This study reported that there were no differences in quality of life (QoL) between women taking tamoxifen or anastrozole; however, the instruments used for measuring QoL were symptom oriented and did not include a measure for musculoskeletal symptoms. A more recent study of a small cohort of Dutch women found that 74% of women on AIs developed arthralgia, and it significantly impacted their household and family life, recreation, and occupation. 7 A larger cross-sectional survey of women receiving AI therapy found that 47% of patients attributed their current arthralgia symptoms to AI therapy, 8 and another cohort study found that 45.5% of women developed severe arthralgia meeting the criteria for rheumatologic referral. 9 A blinded case-control study of 120 women found that women taking AIs had thicker tendon sheaths than controls and electromyelogram findings consistent with carpel tunnel syndrome. 10 It also is apparent that women treated with AIs have an associated decline in bone mineral density as a result of low estrogen levels. AIs lower plasma estradiol, estrone, and estrone sulfate concentrations by up to 98%, which has a negative impact on bone remodeling. 11,12 There also are some reports in the literature that adjuvant chemotherapy may be associated with joint pain. 13,14 To date, there is little evidence to establish the extent of joint pain in women who have had breast cancer in comparison with joint pain that might be expected as a natural consequence of aging or menopause. This study set out to establish the prevalence of joint pain in women treated for early breast cancer compared with a control group of women of similar age without breast cancer and begin to explore the relationship between this pain and breast cancer treatments. Methods Study Design and Participants The study protocol was reviewed and approved by the local research ethics review committee (LREC no. 06/Q1701/157). A cross-sectional survey design was used to capture information from participating women at a single time point. Two cohorts were recruited to this study. Women with a history of early-stage (Stages IeIII) breast cancer who had completed planned surgery, adjuvant chemotherapy, and

3 Vol. 46 No. 4 October 2013 Joint and Muscle Aches, Pain, and Stiffness in Women With and Without Breast Cancer 525 radiotherapy were approached in two breast cancer follow-up clinics. During the same period, women under the age of 50 years attending a diagnostic clinic with benign breast disease and, during a one week period, all women attending for mammographic screening (aged $50 years) without a subsequent diagnosis of cancer were approached to form a control group. No information was collected from the benign group as to prior breast biopsies. To be eligible, women also had to be $18 years of age and able to read English. Any prior cancer was an exclusion criterion in both groups. Potential participants were given verbal and written information about the study, and after giving informed consent, they were asked to complete a questionnaire about joint and muscle aches, pain, and stiffness and return this to the researchers by post. Measures Measures used were the Nordic Musculoskeletal Questionnaire, 15,16 the Brief Pain Inventory (BPI), the Short Form-36 Health Survey (SF- 36), and a short questionnaire to collect information on coexisting conditions that could cause joint pain, past and current medications, and other factors that may have a bearing on this pain, including lymphedema and weight. Data on preexisting arthritic conditions were collected by self-report. Women were asked if they had been diagnosed with rheumatoid arthritis, osteoarthritis, fibromyalgia, fibrositis, gout, or any other form of arthritis. Other relevant clinical information, including information on surgery and radiotherapy, and details of chemotherapy and hormone therapies used were abstracted from clinical records. The Nordic Musculoskeletal Questionnaire has been used extensively to ascertain musculoskeletal pain by collecting data on the presence or absence of pain in 12 body sites and on the impact of this pain on daily life. 17e19 The BPI 20 measures both pain intensity and its interference with function. There also are questions on pain relief, pain quality, and patients perceptions of the causes of pain. It is widely used among cancer patients, with data published for postbreast cancer surgery. 21 It also has been validated among people suffering from osteoarthritic conditions. 22 The SF is a generic 36-item health status questionnaire. There are normative data available, 24 including data for breast cancer populations. 25 Statistical Analysis Demographic, clinical, and medical characteristics were compared between breast cancer cases and controls using the Chi-squared test or t-test for age, weight, and body mass index. Data from the BPI and Nordic Musculoskeletal Questionnaire were compared using the Chisquared test for the presence of pain, and Mann-Whitney U-test for continuous data (severity scores and number of sites with pain) as distributions of these were skewed. Similarly, as the subscale scores from the SF-36 QoL questionnaire were skewed, the data were summarized using medians and interquartile ranges, and groups were compared using the Mann-Whitney U-test. Patient, clinical, and treatment characteristics were tested for association with the presence or absence of pain using logistic regression; results were summarized using odds ratios (ORs) and 95% CIs, and the significance of effects was tested using the Wald test. Each characteristic was first tested in a univariate analysis and then all those which were statistically significant at a level of P < 0.10 were included in turn in the multiple logistic regression analysis to see which remained significant and, therefore, could be said to have an independent effect on the presence of pain. For the logistic regression analysis, only patients with data available on all the characteristics of interest were included in the univariate and multiple regression. The same procedure was followed to determine which characteristics were significantly associated with severity of pain, using the measure of pain on average from the BPI, for those who reported pain today, but using linear regression methods as the outcome, variable was a scale measure (graded from 0 to 10). Similarly, linear regression was used to compare the subscale scores from the SF-36 between groups after adjusting for factors that were significantly different between the groups, as described previously. Results Of the 317 eligible women with early-stage breast cancer attending the hospital clinic for

4 526 Fenlon et al. Vol. 46 No. 4 October 2013 follow-up, 303 were approached to participate (Fig. 1). Twelve women declined, 291 questionnaires were given out, and 204 (71%) were returned. A further 43 participants were recruited from a second hospital, but response rate figures are not available for this group. Recruiting for the control arm of the study took place from a mobile breast screening unit and benign breast clinics. Four hundred eighty-seven women attended for screening at the screening unit in the week of the study, 31 were missed, and 79 refused (Fig. 2). Questionnaires were given out to 377 women; 239 (63%) of these were returned. A further 35 participants were recruited from benign breast clinics, but response rate figures are not available for this group. Demographics Characteristics of the study population are shown in Table 1. The mean ages of the breast cancer and control populations were 58.7 (SD ¼ 10.9) and 56.6 (SD ¼ 8.8) years, respectively (P ¼ 0.022). The mean weights were 72.2 (SD ¼ 13.9) and 69.0 kg (SD ¼ 12.7) (P ¼ 0.018), respectively. Significantly more women with a history of breast cancer were postmenopausal at the time of completing the questionnaire, had illnesses other than breast cancer, were currently taking any prescribed or over-the-counter medication, and were not currently employed compared with those without cancer. Number of eligible patients in clinic: 317 (1) + 43 (2) Patients missed: 14 Patients refused: 12 Questionnaires given out: 291 (1) + 43 (2) Questionnaires not returned: 87 Questionnaires returned: 204 (1) + 43 (2) Final sample: 247 Fig. 1. Breast cancer patients recruited to the study (1) Center 1. (2) Center 2. Number of eligible patients invited to screening unit: a Patients did not attend: 120 Patients missed: 31 Patients refused: 79 Questionnaires given out: a Questionnaires not returned: 148 Questionnaires returned: 274 Final sample: 274 Fig. 2. Control group (non-breast cancer) patients recruited to the study. a Patients recruited through benign breast clinics. Treatment Details Clinical and treatment details for those women with a history of breast cancer are given in Table 2. Women with breast cancer had a median of 28 months (range 2e184 months) from diagnosis. Only one patient did not have primary surgery (59% had wide local excision and 36% had mastectomy), 79% had radiotherapy, 45% had chemotherapy, and 81% had hormone therapy. In the 112 patients who received chemotherapy, the most common regimen was sequential epirubicin, cyclophosphamide, methotrexate, and 5-fluorouracil (n ¼ 110); patients were treated with taxanes. Of the 200 patients who received hormone therapy, 168 had been treated with tamoxifen and 63 had been treated with AIs. As women were some time since diagnosis, tamoxifen was the first-line treatment of choice for adjuvant therapy for many of the women. It is not recorded how many had switched from AIs to tamoxifen as a result of symptoms experienced. Treatment details are missing for 13 patients in the cancer group. Presence and Severity of Pain On the BPI, 153 (62%) women with a history of breast cancer reported that they were experiencing pain today compared with 145 (53%) of the controls (P ¼ 0.023) (Table 3). Factors that

5 Vol. 46 No. 4 October 2013 Joint and Muscle Aches, Pain, and Stiffness in Women With and Without Breast Cancer 527 Characteristics Table 1 Demographic, Clinical, and Medical Characteristics of Women With and Without Breast Cancer Women With Breast Cancer, N ¼ 247; n (%) Women Without Breast Cancer, N ¼ 274; n (%) P-value a Age (yrs), mean (SD) [range] b 58.7 (10.9) [30e86] 56.6 (8.8) [18e72] c Ethnicity White 201 (81.4) 189 (69.0) Other 5 (2.0) 7 (2.6) Unknown 41 (16.6) 78 (28.5) Marital status Married/partner 180 (72.9) 189 (69.0) Other 62 (25.1) 84 (30.7) Unknown 5 (2.0) 1 (0.4) Job status Employed 97 (39.3) 141 (51.5) Not employed/retired 142 (57.5) 127 (46.3) Unknown 8 (3.2) 6 (2.2) Weight (kg), mean (SD) [range] d 72.2 (13.9) [49e122.5] 69.0 (12.7) [45.4e115] c BMI (kg/m 2 ), mean (SD) [range] e 27.3 (5.0) [17.6e48.5] 26.5 (4.5) [17.6e43.3] c Menopausal status <0.001 Postmenopausal 193 (78.1) 140 (51.1) Premenopausal 16 (6.5) 37 (13.5) Perimenopausal 3 (1.2) 7 (2.6) Unknown 35 (14.2) 90 (32.8) Previous surgery on bones or joints No 187 (75.7) 214 (78.1) Yes 53 (21.5) 58 (21.2) Unknown 7 (2.8) 2 (0.7) Had any serious illness other than cancer No 170 (68.8) 215 (78.5) Yes 68 (27.5) 53 (19.3) Unknown 9 (3.6) 6 (2.2) Arthritis No 186 (75.3) 195 (71.2) Yes 58 (23.5) 75 (27.4) Unknown 3 (1.2) 4 (1.5) Currently taking any prescribed/ OTC medication <0.001 No 38 (15.4) 104 (38.0) Yes 202 (81.8) 167 (60.9) Unknown 7 (2.8) 3 (1.1) Currently taking any complementary or herbal medicines No 149 (60.3) 151 (55.1) Yes 90 (36.4) 121 (44.2) Unknown 8 (3.2) 2 (0.7) BMI ¼ body mass index; OTC ¼ over the counter. a Chi-squared test unless otherwise indicated. Unknown categories excluded from significance tests. b Age unknown for 15 women with cancer and 23 women without cancer. c t-test. d Weight unknown for 91 women with cancer and 33 women without cancer. e BMI unknown for 98 women with cancer and 33 women without cancer. were independently significantly associated with the presence of pain as reported on the BPI were history of breast cancer (OR ¼ 1.71; 95% CI ¼ 1.14e2.57; P ¼ 0.009), preexisting arthritic conditions (OR ¼ 4.56; 95% CI ¼ 2.71e7.69; P < 0.001), and age (OR ¼ 0.98; 95% CI ¼ 0.96e1.00; P ¼ 0.023) (Table 4). Of the women who reported pain today on the BPI, there was no significant difference in severity of pain between breast cancer cases and controls except for the assessment of pain right now, which was worse for the cases (P ¼ 0.012). Factors that were independently significantly associated with the severity of pain on average were history of breast cancer

6 528 Fenlon et al. Vol. 46 No. 4 October 2013 Table 2 Clinical and Treatment Characteristics of Women With Breast Cancer (N ¼ 247) Characteristics n (%) Time since diagnosis (months), median (IQR) [range] 28 (14e48) [2e184] Primary surgery None 1 (0.4) Wide local excision 145 (58.7) Mastectomy 88 (35.6) Unknown 13 (5.3) Axillary surgery No 3 (1.2) Yes 219 (88.7) Unknown 25 (17.4) Radiotherapy No 38 (15.4) Yes 196 (79.4) Unknown 13 (5.3) Chemotherapy a No 120 (48.6) Yes 112 (45.3) E-CMF 110 Other nontaxane regimen 12 Taxane-containing regimen 17 Unknown 15 (6.1) Hormone therapy a No 34 (13.8) Yes 200 (81.0) Aromatase inhibitors 63 Tamoxifen 168 Ovarian suppression 14 Other 12 Unknown 13 (5.3) Chemotherapy and hormone therapy Neither 3 (1.2) Chemotherapy only 31 (12.6) Hormone therapy only 117 (47.4) Both chemotherapy and hormone therapy 81 (32.8) Hormone therapy but unknown chemotherapy 2 (0.8) Unknown for both 13 (5.3) Lymphedema No 180 (72.9) Yes 56 (22.7) Unknown 11 (4.5) IQR ¼ interquartile range; E ¼ epirubicin; C ¼ cyclophosphamide; M ¼ methotrexate; F ¼ 5-fluorouracil. Time since diagnosis unknown for 14 cases. a Some women had been treated with more than one chemotherapy or hormone therapy. (P ¼ 0.045), increased weight (P < 0.001), and preexisting arthritic conditions (P ¼ 0.014) (Table 5). Sites and Impact of Pain From the Nordic Musculoskeletal Questionnaire, pain in at least one site over the last seven days was reported in 83% of women with breast cancer and 83% of controls (P ¼ 0.957); the OR (95% CI) adjusted for patient characteristics significantly associated with the presence of pain (age, weight, and arthritis) was 1.02 (0.58e1.79). Significantly more women who had been treated for breast cancer reported experiencing pain in the hands and upper back over the preceding seven days than those with no history of cancer (Table 6). Women without breast cancer were significantly more likely to report pain in the neck over the last seven days and over the last 12 months (Table 6). Pain in at least one site over the last 12 months was reported in 88% of women with breast cancer and 91% of controls (P ¼ 0.207); the OR (95% CI) adjusted for age, weight, and arthritis was 0.54 (0.28e1.06). Although significantly more women who had been treated for breast cancer than controls reported pain on the BPI, for those women who had pain, no difference was observed in the impact of this pain on daily activities (Table 3). On the Nordic Musculoskeletal Questionnaire (Table 6), 50% of women with breast cancer who reported pain in at last one site over the last 12 months stated that it had prevented normal activities in the past 12 months compared with 42% in the noncancer group (P ¼ 0.120), although breast cancer patients reported significantly more sites with pain affecting activity over the same time period (P ¼ 0.021). Quality of Life QoL as measured by the SF-36 was significantly worse for women with breast cancer whether or not they had joint pain. Compared with the controls, women with breast cancer had significantly worse scores for six of the eight domains including pain, physical functioning, role limitation because of physical problems, social functioning, and energy/vitality resulting in a lower general health perception in this group, which remained in all but energy/vitality and general health perception after adjusting for differences between the groups. In contrast, no significant difference was observed in mental health and role limitation because of emotional problems (Table 7). For women with a history of breast cancer, those with joint pain had worse QoL than those without pain, reporting significantly lower scores for all domains, which remained after adjusting for factors associated with pain (Table 8).

7 Vol. 46 No. 4 October 2013 Joint and Muscle Aches, Pain, and Stiffness in Women With and Without Breast Cancer 529 Table 3 Presence of Pain and Impact on Activity (BPI) Comparing Women With and Without Breast Cancer Presence and Impact of Pain Women With Breast Cancer, N ¼ 247 Women Without Breast Cancer, N ¼ 274 P-value a Do you have pain? n (%) b No 85 (34.4) 124 (45.3) Yes 153 (61.9) 145 (52.9) Unknown 9 (3.6) 5 (1.8) For those with pain, c median (IQR) Pain at worst in past 24 hours 5 (3e6) 4 (3e6) Pain at least in last 24 hours 2 (1e3) 2 (0e3) Pain on average 4 (2e5) 3 (2e5) Pain right now 3 (1e5) 2 (1e4) Pain interfered with activity 2 (0e5) 2 (0e5) Pain interfered with mood 2 (0e4) 1.5 (0e3) Pain interfered with walking 2 (0e5) 1 (0e4) Pain interfered with work 3 (1e5.75) 2 (0e5) Pain interfered with relations 0 (0e2) 0 (0e2) Pain interfered with sleep 3 (1e5) 2 (0.5e5) Pain interfered with enjoyment of life 2 (0.75e5) 2 (0e4) Receiving treatments or medications for pain d 72 (47.1) 88 (60.7) d For those receiving pain treatments or medications Relief with medication; e median (IQR) 50 (30e80) 50 (20e80) BPI ¼ Brief Pain Inventory; IQR ¼ interquartile range. Missing data: Relief with medication unknown for eight women with breast cancer and 15 women without breast cancer; for all the other pain scales, the number of women with missing data ranged from three to nine for women with cancer and one to four for women without cancer. a Mann-Whitney test unless otherwise indicated. b Chi-squared test. c Pain scored from 0 to 10; a higher score indicates poorer health status. All responses relate to past 24 hours. d Patients reported receiving treatments or medication for pain in a free-text field on the BPI questionnairedthere is no tick box for not receiving such treatments, so no statistical comparison is made for these data. e Relief with medication scored from 0 (no relief) to 100% (complete relief). Causes of Pain in Breast Cancer Cases In the breast cancer group, the factors that were independently significantly associated with the presence of pain were younger age (OR ¼ 0.97; 95% CI ¼ 0.94e0.99; P ¼ 0.027), preexisting arthritic conditions (OR ¼ 5.28; Table 4 Results of Logistic Regression Analyses of Associations With Presence of Pain (From the BPI) in Breast Cancer and Non-Breast Cancer Groups Characteristics Crude OR (95% CI), P-Value Adjusted OR a (95% CI), P-Value Breast cancer P ¼ P ¼ Yes 1.50 (1.02e2.21) 1.71 (1.14e2.57) Age P ¼ (0.98e1.01) P ¼ (0.96e1.00) Previous surgery on bones or joints P ¼ P ¼ Yes 1.53 (0.95e2.46) 1.29 (0.78e2.13) Had any serious illness other than cancer? P ¼ P ¼ Yes 1.13 (0.72e1.78) 1.00 (0.61e1.64) Arthritis P < P < Yes 3.79 (2.30e6.24) 4.56 (2.71e7.69) Currently taking any prescribed/otc medication? P ¼ P ¼ Yes 1.57 (1.03e2.40) 1.26 (0.79e2.01) Currently taking any complementary/herbal medicines? P ¼ P ¼ Yes 1.02 (0.69e1.50) 0.99 (0.66e1.48) BPI ¼ Brief Pain Inventory; OR ¼ odds ratio; OTC ¼ over the counter; BMI ¼ body mass index. A total of 440 patients with data available on all the characteristics listed in the table were included in the univariate and multiple regression models. Menopausal status, weight, and BMI were tested in univariate analyses but were not statistically significant and so were omitted from the analyses presented above because of the extent of missing data, which would have greatly reduced the data set available for analysis. a Adjusted for cancer, age, and arthritis.

8 530 Fenlon et al. Vol. 46 No. 4 October 2013 Table 5 Results of Linear Regression Analyses of Associations With Severity of Pain on Average in Breast Cancer and Non-Breast Cancer Groups for Those Who Reported Presence of Pain (From the BPI) Characteristics Regression Coefficient (95% CI), P-Value Breast cancer P ¼ (0.04 to 1.14) Age P ¼ ( 0.02 to 0.03) Weight P < (0.02 to 0.06) Previous surgery on bones or joints P ¼ ( 0.05 to 1.18) Had any serious illness other than cancer? P ¼ ( 0.44 to 0.86) Arthritis P ¼ (0.15 to 1.28) Currently taking any prescribed/otc medication? P ¼ ( 0.09 to 1.18) Currently taking any complementary/herbal P ¼ medicines? 0.51 ( 1.06 to 0.05) Adjusted Regression Coefficient a (95% CI), P-value P ¼ (0.01 to 1.08) P ¼ ( 0.04 to 0.02) P < (0.02 to 0.05) P ¼ ( 0.34 to 0.88) P ¼ ( 0.61 to 0.65) P ¼ (0.14 to 1.23) P ¼ ( 0.31 to 0.91) P ¼ ( 0.91 to 0.16) BPI ¼ Brief Pain Inventory; OTC ¼ over the counter. A total of 207 patients who reported the presence of pain and with data available on all the characteristics listed in the table were included in the univariate and multiple regression models. A positive regression coefficient indicates an increase in severity of pain on average with the presence of the characteristic (or higher values for age and weight). a Adjusted for cancer, weight, and arthritis. 95% CI ¼ 2.18e12.76; P < 0.001), prior treatment with taxane chemotherapy (OR ¼ 6.22; 95% CI ¼ 1.27e30.38; P ¼ 0.024), use of AIs (OR ¼ 3.18; 95% CI ¼ 1.38e7.30; P ¼ 0.006), and treatment with tamoxifen (OR ¼ 2.56; 95% CI ¼ 1.25e5.27; P ¼ 0.011) (Table 9). Discussion We undertook this study to evaluate the incidence and impact of joint and muscle aches, pain, and stiffness in women with early-stage breast cancer. In the absence of validated tools in this setting, we opted to use two different assessment tools that have been validated in other clinical settings, together with a QoL assessment tool that is well validated in the cancer setting. The BPI is a generic pain questionnaire, but participants were aware that the focus of this study was in relation to joint and muscle aches, pain, and stiffness. We compared a group of women who had completed surgery and adjuvant chemotherapy/radiotherapy with a control group of women of similar age without breast cancer and found that the results of these two assessments were largely consistent. Although we found a high incidence of selfreported symptoms in our control group, joint and muscle aches, pain, and stiffness were significantly more common in women who have previously been treated for early-stage breast cancer, and this had a significantly detrimental effect on their QoL. In a logistic regression analysis, previous breast cancer treatment was associated with nearly a doubling in the risk of having joint pain. This concurs with a case-control study in Maryland where 28% of cancer patients had a two point or greater increase in visual analogue scale scores for pain compared with 14% of controls. 27 The patterns of joint pain differed between women with a history of breast cancer and controls. Specific sites affected more commonly in those with previous breast cancer were the hands and upper back, whereas those without breast cancer were more likely to report neck pain. Women treated for early-stage breast cancer had worse overall QoL scores compared with those in the control group. In particular, they reported worse physical functioning, greater role limitation because of physical problems, and a more detrimental effect of pain on QoL and social functioning. Within the group of women treated for breast cancer, those experiencing joint pain had significantly worse overall QoL scores (reflecting worse scores for all domains) than those who were not. Taken together, these results suggest that joint

9 Vol. 46 No. 4 October 2013 Joint and Muscle Aches, Pain, and Stiffness in Women With and Without Breast Cancer 531 Site Affected Table 6 Results From the Nordic Musculoskeletal Questionnaire Women With Breast Cancer, Number With Pain/Total (%) Women Without Breast Cancer, Number With Pain/Total (%) Crude OR (95% CI) P-value a In last seven days Any site 202/242 (83) 223/269 (83) 1.04 (0.65e1.66) Neck 68/243 (28) 99/269 (37) 0.67 (0.46e0.97) Either shoulder 92/243 (38) 89/269 (33) 1.23 (0.86e1.77) Either elbow 36/240 (15) 34/268 (13) 1.21 (0.73e2.01) Either hand 100/242 (41) 84/267 (31) 1.53 (1.07e2.21) Upper back 50/240 (21) 30/268 (11) 2.09 (1.28e3.41) Lower back 104/243 (43) 120/268 (45) 0.92 (0.65e1.31) Hips 77/241 (32) 69/268 (26) 1.35 (0.92e1.99) Knees 87/242 (36) 92/268 (34) 1.07 (0.75e1.54) Feet 77/241 (32) 78/269 (29) 1.15 (0.79e1.68) In last 12 months Any site 213/243 (88) 246/269 (91) 0.66 (0.37e1.18) Neck 93/240 (39) 129/269 (48) 0.69 (0.48e0.98) Either shoulder 109/240 (45) 114/269 (42) 1.13 (0.80e1.61) Either elbow 54/242 (22) 49/269 (18) 1.29 (0.84e1.99) Either hand 116/240 (48) 109/269 (40) 1.37 (0.97e1.95) Upper back 66/238 (28) 57/269 (21) 1.43 (0.95e2.14) Lower back 132/241 (55) 163/269 (61) 0.79 (0.55e1.12) Hips 98/240 (41) 99/269 (37) 1.18 (0.83e1.69) Knees 116/242 (48) 121/267 (45) 1.11 (0.78e1.57) Feet 95/241 (39) 92/267 (34) 1.24 (0.86e1.78) For those with pain at individual sites in last 12 months, did it affect activity? Any site 105/210 (50) 104/246 (42) 1.36 (0.94e1.98) Neck 21/90 (23) 29/129 (22) 1.05 (0.55e1.99) >0.99 Shoulders 37/107 (35) 32/114 (28) 1.35 (0.77e2.40) Elbows 13/51 (25) 13/49 (26) 0.95 (0.39e2.32) >0.99 Hands 43/112 (38) 29/109 (27) 1.72 (0.97e3.04) Upper back 21/63 (33) 11/57 (19) 2.09 (0.90e4.85) Lower back 53/130 (41) 59/162 (36) 1.20 (0.75e1.93) Hips 41/98 (42) 33/99 (33) 1.44 (0.81e2.57) Knees 36/114 (32) 35/121 (29) 1.13 (0.65e1.98) Feet 35/94 (37) 21/92 (23) 2.01 (1.06e3.81) Total number of sites with pain In last seven days b 0 40/242 (16) 46/269 (17) 1 1e2 65/242 (27) 93/269 (35) 0.80 (0.46e1.41) 3e4 71/242 (29) 69/269 (26) 1.18 (0.67e2.10) 5þ 66/242 (27) 61/269 (23) 1.24 (0.69e2.24) Median (IQR) [range] 3 (1e5) [0e11] 2 (1e4) [0e12] c In last 12 months b 0 30/243 (12) 23/269 (9) 1 1e2 48/243 (20) 85/269 (32) 0.43 (0.21e0.87) 3e4 61/243 (25) 59/269 (22) 0.79 (0.39e1.60) 5þ 104/243 (43) 102/269 (38) 0.78 (0.41e1.50) Median (IQR) [range] 4 (2e6) [0e12] 3 (1.5e6) [0e12] c Affected activity in last 12 months b 0 105/210 (50) 142/246 (58) 1 1e2 55/210 (26) 65/246 (26) 1.14 (0.72e1.82) 3e4 29/210 (14) 26/246 (11) 1.51 (0.81e2.82) 5þ 21/210 (10) 13/246 (5) 2.18 (0.99e4.86) Median (IQR) [range] 0.5 (0e2) [0e9] 0 (0e2) [0e9] c OR ¼ odds ratio; IQR ¼ interquartile range. Denominators vary because of missing data. The presence of pain at different sites of the body comparing women with and without breast cancer over the last seven days, in the last 12 months, and whether pain has affected activity in the last 12 months. a Chi-squared test unless otherwise indicated. b Chi-squared test for trend. c Mann-Whitney test. and muscle aches, pain, and stiffness are a significant problem associated with breast cancer treatment. Although QoL scores were clearly poorer for women with pain after breast cancer, this did not appear to be reflected in women s self-

10 532 Fenlon et al. Vol. 46 No. 4 October 2013 SF-36 Domain a Table 7 Impact of Breast Cancer on Quality of Life (SF-36) Women With Breast Cancer Median (IQR) Women Without Breast Cancer Median (IQR) P-value b (crude comparison) P-value c (adjusted) Physical functioning 75 (55e90) 85 (70e95) <0.001 <0.001 Role limitation because of emotional problems 100 (67e100) 100 (67e100) Role limitation because of physical problems 100 (25e100) 100 (75e100) <0.001 <0.001 Social functioning 89 (67e100) 89 (78e100) Mental health 76 (64e84) 76 (64e84) Energy/vitality 60 (41e70) 60 (50e75) Pain 67 (44e89) 78 (55e89) General health perception 72 (57e82) 77 (62e87) SF-36 ¼ Short Form-36 Health Survey; IQR ¼ interquartile range. Missing data: The number of women with missing data per domain ranged from 2 to 24 for women with breast cancer and from 5 to 13 for women without breast cancer. a All quality of life domain scores are on a scale from 0 to 100, and a lower score reflects poorer health status. b Mann-Whitney test. c Wald test from multiple linear regression, adjusting for age, previous surgery, other illness, arthritis, and medication. report of how much their pain had interfered with daily activities compared with the pain experienced by women without breast cancer, when measured by either the Nordic Musculoskeletal Questionnaire or BPI. In fact, the only significant difference measuring severity and impact of pain between women with cancer and without cancer was for pain right now. There were no significant differences in other self-reports of pain, including average pain or in any aspect of impact on activity as measured by the BPI. This could be explained in one of two ways. Either the pain experienced by women after breast cancer does not have a materially different impact on daily life than the aches and pains of normal aging or that the women with breast cancer view the impact of pain differently once they have been exposed to a diagnosis of cancer. After cancer, the impact that joint pain is perceived to have on life may appear relatively insignificant in comparison with the cancer diagnosis. It is suggested that any future research takes into account the experience of cancer diagnosis in study design. A preexisting diagnosis of arthritis was the strongest predictor for joint pain in the whole study cohort and also was associated with a fivefold increased risk of pain in those treated for breast cancer. No independent verification was SF-36 Domain a Table 8 Impact of Pain on Quality of Life (SF-36) in Women With Breast Cancer Women With Breast Cancer and Pain, b Median (IQR) Women With Breast Cancer and No Pain, b Median (IQR) P-value c (Crude Comparison) P-value d (Adjusted) Physical functioning 65 (50e80) 85 (75e100) <0.001 <0.001 Role limitation because of emotional problems 100 (41.7e100) 100 (100e100) Role limitation because of physical problems 75 (25e100) 100 (75e100) <0.001 <0.001 Social functioning 78 (56e100) 100 (78e100) < Mental health 72 (60e84) 80 (72e89) < Energy/vitality 55 (35e65) 70 (55e80) <0.001 <0.001 Pain 56 (44e67) 89 (72e100) <0.001 <0.001 General health perception 67 (47e82) 67 (47e82) < SF-36 ¼ Short Form-36 Health Survey; IQR ¼ interquartile range. Missing data: The number of women with missing data per domain ranged from 0 to 14 for women with breast cancer and from 0 to 7 for women without breast cancer. a All quality of life domain scores are on a scale from 0 to 100, and a lower score reflects poorer health status. b The presence or absence of pain reported on the Brief Pain Inventory. c Mann-Whitney test. d Wald test from multiple linear regression, adjusting for age, previous surgery, arthritis, lymphedema, and ever use of taxanes, aromatase inhibitors, and tamoxifen.

11 Vol. 46 No. 4 October 2013 Joint and Muscle Aches, Pain, and Stiffness in Women With and Without Breast Cancer 533 Table 9 Results of Logistic Regression Analyses of Associations With the Presence of Pain (From the BPI) in the Breast Cancer Group Characteristics Crude OR (95% CI), P-value Adjusted OR a (95% CI), P-value Age P ¼ (0.96e1.01) P ¼ (0.94e0.99) Arthritis P ¼ P < Yes 3.90 (1.72e8.83) 5.28 (2.18e12.76) Time since diagnosis P ¼ (0.99e1.01) P ¼ (0.99e1.01) Primary surgery P ¼ P ¼ Wide local excision 1 1 Mastectomy 0.97 (0.54e1.75) 0.97 (0.50e1.89) Radiotherapy P ¼ P ¼ Yes 0.74 (0.33e1.63) 0.81 (0.34e1.92) Ever use of taxanes P ¼ P ¼ Yes 4.43 (0.98e19.91) 6.22 (1.27e30.38) Ever use of aromatase inhibitors P ¼ P ¼ Yes 2.36 (1.13e4.94) 3.18 (1.38e7.30) Ever use of tamoxifen P ¼ P ¼ Yes 1.61 (0.86e2.99) 2.56 (1.25e5.27) Lymphedema P ¼ P ¼ Yes 1.93 (0.96e3.91) 1.95 (0.91e4.18) BPI ¼ Brief Pain Inventory; OR ¼ odds ratio. A total of 211 patients with data available on all the characteristics listed in the table were included in the univariate and multiple regression models. a Adjusted for age, arthritis, and ever use of taxanes, aromatase inhibitors, and tamoxifen. gained of diagnosis of arthritis; however, it is known that the degree of arthritis pain does not correlate well with radiologic evidence of arthritis, and it is the presence of pain rather than disease that is the focus of this article. The interaction between preexisting arthritis and breast cancer treatment needs to be further explored as the impact of treatment on preexisting arthritis is not well described, and it is unclear whether those with this problem are more likely to experience impaired function or QoL after breast cancer treatment. Although age is a predictive factor for joint pain, it has a relatively small effect, which would have little impact clinically. It also is thus unlikely that the small (although significant) difference in mean age between our two patient cohorts accounts for the observed difference in the incidence of joint pain as indicated by the fact that in a logistic regression analysis that included age, breast cancer treatment remained a significant predictor of pain. Menopausal status also was significantly different between the two groups, although the effect of this is not certain as there were missing data on menopausal status for the noncancer group. Menopausal status also was taken into consideration in the regression analysis and was not predictive of pain. Nevertheless, it is possible that this had an influence and should be further investigated in any future research. There was also a difference in weight between the two groups, which may have an impact on pain reported as the result of arthritis. However, when both weight and arthritis were included in the logistic regression model, only arthritis was predictive of pain. There are missing data on weight, which may have a bearing on the finding, and this would be important to take into consideration in future research. We note that more people without breast cancer were taking pain medication than those with breast cancer, although it cannot be determined whether this was significant as it was recorded as free text. Nor do we have any indication about the amount or strength of medication that was being taken. It is not known whether this indicates that non-breast

12 534 Fenlon et al. Vol. 46 No. 4 October 2013 cancer-related pain is easier to control or that the actual prevalence of pain in the control group is higher than reported. The prevalence of joint pain reported in this study is much higher than that in previous treatment studies but similar to previous cross-sectional surveys of joint symptoms in patients receiving AIs. 7e10,28,29 This probably reflects the fact that it represents patientreported data captured using questionnaires specific to this problem and, therefore, is more likely to reflect the reality of the problem in routine clinical practice. Adjuvant chemotherapy for early-stage breast cancer is usually administered over a period of three to seven months, whereas adjuvant hormonal therapy is usually administered for at least five years, with current recommendations tending toward even long-term treatment. The median time since diagnosis of breast cancer in our patient cohort was 27 months, suggesting that joint pain may be a long-lasting problem for women who have had breast cancer, perhaps particularly when associated with adjuvant hormonal therapy. This may be one of the factors that contribute to a significant proportion of patients discontinuing adjuvant hormonal therapy earlier than intended. 4 Although breast cancer treatment is a strong predictor of joint pain for individual patients, it can sometimes be difficult to discern which aspect of treatment is responsible. Our data agree with previously published studies that suggest that both chemotherapy (specifically taxanes) and AIs may cause joint and muscle aches, pain, and stiffness. Our study also showed tamoxifen use to be associated with an increased risk of joint pain. This finding might indicate why the problem of joint pain has been underestimated in previous studies in which new therapies, such as the AIs, have been compared with tamoxifen when looking at incidence of joint pain. 6 These results correspond with what we have observed clinically, although anecdotally we also have seen some patients with quite severe joint pain and stiffness during and after nontaxane-containing adjuvant chemotherapy, although in this study nontaxane-containing chemotherapy regimens were not predictive of pain. Joint pain is infrequently reported as a side effect of tamoxifen and clinically regarded as a less prominent concern so that tamoxifen may be offered as an alternative to those women taking AIs who experience joint pain. 26 Conclusion This research shows that women who have been treated for breast cancer may continue to experience significant problems because of joint and muscle aches, pain, and stiffness, which have a significant impact on their health for some years after primary treatment has been completed. There is an indication in our findings that hormone therapies, both tamoxifen and AIs, and also adjuvant chemotherapy, particularly taxanes, are independently predictive of pain. Further research is required to understand the experience of this problem from the individual perspective. More detailed information about specific causes of joint pain is required to facilitate decision making around treatment options. Longitudinal work also is required to enhance our knowledge of the onset, pattern and duration of these symptoms, and risk factors for their occurrence to adequately inform women about to commence treatment for early-stage breast cancer. Disclosures and Acknowledgments This work was supported by an investigatorinitiated grant from The National Cancer Research Institute Supportive and Palliative Care capacity building initiative. In the last five years, Dr. Simmonds has received honoraria from Novartis for several presentations on management of gastrointestinal stromal tumors and honoraria from Novartis, Pfizer, and Roche for attending advisory boards on letrozole, sunitinib, and trastuzumab and has been supported to attend a number of conferences funded by Novartis and Pfizer. Dr. Fenlon has received an honorarium from Roche to present on hormone treatment in breast cancer. All the other authors have no conflicts of interest. The authors thank the clinicians who supported recruitment into this study, the women with cancer who participated, and Jo Haviland who assisted with statistical support. References 1. CancerResearchUK. Breast cancer factsheet Available from

13 Vol. 46 No. 4 October 2013 Joint and Muscle Aches, Pain, and Stiffness in Women With and Without Breast Cancer 535 rchuk.org/aboutcancer/statistics/. Accessed March 10, CancerResearchUK. Incidence of cancer Available from aboutcancer/statistics/incidence. Accessed March 5, Carpenter JS, Andrykowski MA. Menopausal symptoms in breast cancer survivors. Oncol Nurs Forum 1999;26:1311e Zivian MT, Salgado B. Side effects revealed: Women s experiences with aromatase inhibitors. San Francisco: Breast Cancer Action, Presant CA, Bosserman L, Young T, et al. Aromatase inhibitor-associated arthralgia and/or bone pain: frequency and characterization in nonclinical trial patients. Clin Breast Cancer 2007;7: 775e Howell A, Cuzick J, Baum M, et al. ATAC Trialists Group. Results of the ATAC (Arimidex, Tamoxifen, Alone or in Combination) trial after completion of 5 years adjuvant treatment for breast cancer. Lancet 2005;365:60e Boonstra A, van Zadelhoff J, Timmer-Bonte A, et al. Arthralgia during aromatase inhibitor treatment in early breast cancer patients: prevalence, impact, and recognition by healthcare providers. Cancer Nurs 2013;36:52e Mao JJ, Stricker C, Bruner D, et al. Patterns and risk factors associated with aromatase inhibitorrelated arthralgia among breast cancer survivors. Cancer 2009;115:3631e Henry NL, Giles JT, Ang D, et al. Prospective characterization of musculoskeletal symptoms in early stage breast cancer patients treated with aromatase inhibitors. Breast Cancer Res Treat 2008; 111:365e Dizdar O, Ozçakar L, Malas FU, et al. Sonographic and electrodiagnostic evaluations in patients with aromatase inhibitor-related arthralgia. J Clin Oncol 2009;27:4955e Dowsett M, Jones A, Johnston SR, et al. In vivo measurement of aromatase inhibition by letrozole (CGS 20267) in postmenopausal patients with breast cancer. Clin Cancer Res 1995;1:1511e Geisler J, Haynes B, Anker G, Dowsett M, Lønning PE. Influence of letrozole and anastrozole on total body aromatization and plasma estrogen levels in postmenopausal breast cancer patients evaluated in a randomized, cross-over study. J Clin Oncol 2002;20:751e Smith D. Additional cases of postchemotherapy rheumatism. J Clin Oncol 1993;11:1625e Siegel J. Postchemotherapy rheumatism: is this a menopausal symptom? J Clin Oncol 1993;11: Kuorinka I, Jonsson B, Kilbom A, et al. Standardised Nordic questionnaires for the analysis of musculoskeletal symptoms. Appl Ergon 1987;18:233e Dickinson CE, Campion K, Foster AF, et al. Questionnaire development: an examination of the Nordic Musculoskeletal Questionnaire. Appl Ergon 1992;23:197e Porter JM, Gyi DE. The prevalence of musculoskeletal troubles among car drivers. Occup Med (Lond) 2002;52:4e Malchaire J, Piette A, Cock N. Associations between hand-wrist musculoskeletal and sensorineural complaints and biomechanical and vibration work constraints. Ann Occup Hyg 2001;45:479e Miranda H, Viikari-Juntura E, Martikainen R, Riihim aki H. A prospective study on knee pain and its risk factors. Osteoarthritis Cartilage 2002; 10:623e Cleeland CS, Gonin R, Hatfield AK, et al. Pain and its treatment in outpatients with metastatic cancer. N Engl J Med 1994;330:592e Carpenter JS, Sloan P, Andrykowski MA, et al. Risk factors for pain after mastectomy/lumpectomy. Cancer Pract 1999;7:66e Williams VSL, Smith MY, Fehnel SE. The validity and utility of the BPI interference measures for evaluating the impact of osteoarthritic pain. J Pain Symptom Manage 2006;31:48e Ware JE, Sherbourne CD. A 36 item Short Form Health Survey (SF-36): conceptual framework and item selection. Med Care 1992;30:473e Walker-Bone K, Reading I, Coggon D, Cooper C, Palmer KT. The anatomical pattern and determinants of pain in the neck and upper limbs: an epidemiologic study. Pain 2004;109:45e Conde DM, Pinto-Neto AM, Cabello C, et al. Menopause symptoms and quality of life in women aged 45 to 65 years with and without breast cancer. Menopause 2005;12:436e Poole CJ, Earl HM, Hiller L, et al. NEAT Investigators and the SCTBG. Epirubicin and cyclophosphamide, methotrexate, and fluorouracil as adjuvant therapy for early breast cancer. N Engl J Med 2006;355:1851e Helzlsouer KJ, Gallicchio L, MacDonald R, Wood B, Rushovich E. A prospective study of aromatase inhibitor therapy, vitamin D, C-reactive protein and musculoskeletal symptoms. Breast Cancer Res Treat 2012;131:277e Crew KD, Greenlee H, Capodice J, et al. Prevalence of joint symptoms in postmenopausal women taking aromatase inhibitors for early-stage breast cancer. J Clin Oncol 2007;25:3877e Kanematsu M, Morimoto M, Honda J, et al. The time since last menstrual period is important as a clinical predictor for non-steroidal aromatase inhibitorrelated arthralgia. BMC Cancer 2011;11:436.