Characteristics of Affected Limb Pain and Associated Factors in Young Adult Survivors of Lower Limb Osteosarcoma in Japan

Abstract

Purpose:

This study aimed to characterize affected-limb pain and its associated factors among adolescent and young adult (AYA) survivors of lower extremity osteosarcoma in Japan. We hypothesized that both physical and psychosocial factors would be associated with pain, with an interaction between pain intensity and pain interference.

Methods:

A cross-sectional study was conducted among osteosarcoma survivors aged 18–39 years who had completed treatment at least 2 years earlier. Self-administered questionnaires assessed pain intensity and interference, physical function, fatigue, sleep disturbance, anxiety, depression, and perceived social support. Multiple regression analyses were performed to identify factors associated with pain outcomes. Semistructured interviews were also conducted to provide supplementary insights into pain characteristics and contextualize the quantitative findings.

Results:

Among 64 participants, 65.6% reported pain or pain interference during the preceding 7 days. Pain intensity was significantly associated with poorer physical function, older age at diagnosis, and lower perceived social support, whereas pain interference was primarily associated with greater fatigue severity. Pain intensity and pain interference were strongly correlated. Qualitative analysis identified four pain characteristics prosthesis-related pain, phantom limb pain, secondary pain, and weather-related pain. Interviews further illustrated how physical limitations, fatigue, and social support shaped pain experiences.

Conclusion:

Pain among long-term osteosarcoma survivors is influenced by multiple biopsychosocial factors. Although average pain severity was relatively mild, substantial interindividual variability was observed. Comprehensive survivorship care should address physical function, fatigue management, and social support to optimize long-term pain management and self-care among AYA osteosarcoma survivors.

Keywords

cancer survivorship osteosarcoma adolescent and young adult pain fatigue social support

Introduction

Survivors of lower limb osteosarcoma may experience persistent pain in the affected limb following limb-sparing surgery or amputation with tumor resection.^1–4 Persistent pain is influenced by physical, psychological, and social factors and can interfere with activities of daily living and the ability to fulfill social roles.^5,6

Osteosarcoma is a rare malignancy that predominantly affects adolescents and young adults, most commonly in their teens and twenties. In Japan, approximately 2,300 cases of osteosarcoma were reported among individuals aged 0–39 years between 2006 and 2022.⁷ The extremities are the most frequent tumor site (45%).⁸ While amputation was historically the primary treatment, advances in surgical techniques have enabled more than 80% of patients to undergo limb-sparing procedures as first-line treatment.^9,10 Adolescent and young adult (AYA) survivors face psychosocial challenges during critical developmental periods, including reduced autonomy, difficulties with identity formation, academic and employment challenges, interpersonal difficulties, and fertility concerns.^11,12

Existing studies on pain in malignant bone tumors have largely focused on postoperative pain,^1,2 with limited research on chronic affected-limb pain among long-term survivors. Previous studies include retrospective cohort analyses based on medical records,¹³ quality-of-life comparisons between limb-sparing and amputation groups in which pain is assessed as a single item,¹⁴ and studies focusing on phantom limb pain.^15–17 Survivors reporting pain have been shown to have poorer physical function and greater uncertainty,⁴ and pain may further impair daily activities and social participation.⁶ However, no studies have directly measured affected-limb pain in long-term osteosarcoma survivors or explored its characteristics and associated factors.

Pain is defined by the International Association for the Study of Pain as “an unpleasant sensory and emotional experience associated with, or resembling that of, actual or potential tissue damage.”^18,19 Pain is inherently subjective, and individuals’ self-reports should be respected.^18–20 Pain reflects an interaction of sensory, emotional, cognitive, and social factors rather than tissue damage alone.^20,21 Assessing both pain intensity and pain interference is essential for a comprehensive understanding of pain.^22,23 The biopsychosocial model provides a holistic framework for chronic pain assessment, such as physical injury, neural function, psychological responses, and social context, rather than focusing exclusively on nociceptive signal transmission.²⁰ This model has identified anxiety, depression, sleep disturbance, and fatigue as relevant factors in pediatric and AYA cancer survivors.¹³ This study hypothesized that pain would be associated with physical and psychosocial factors, with an interaction between pain intensity and interference. Supporting survivors experiencing affected limb pain requires clarifying not only the presence of pain but also its characteristics and associated factors from both subjective and objective perspectives.

This study aimed to identify affected-limb pain and its associated factors in osteosarcoma survivors and to explore the pain characteristics and subjective pain experience of associated factors through interviews. This approach provides a comprehensive perspective on pain in the affected limb among osteosarcoma survivors, highlighting potential challenges and opportunities for enhanced support.

Materials and Methods

Study design

The study primarily employed a quantitative cross-sectional design, supplemented by interviews to explore the characteristics of affected-limb pain. Guided by the biopsychosocial model, we hypothesized that pain in the affected limb would be associated with physical, psychological, and social factors (Fig. 1). In this exploratory study, hypotheses were tested using quantitative data, and survivor-narrated interview data were subsequently analyzed to deepen understanding of the findings and to interpret the characteristics of affected-limb pain and its associated factors from multiple perspectives.

FIG. 1.

Hypothesis of related factors for pain.

Participants

Participants who met the selection criteria were recruited from four oncology hospitals in the Tokyo metropolitan area. Inclusion criteria were as follows: survivors aged 18–39 years who had been diagnosed with lower extremity osteosarcoma, had been informed of their diagnosis, had completed osteosarcoma treatment, had remained disease-free for at least 2 years, and were under routine follow-up at the time of the study. Exclusion criteria included individuals with osteosarcoma arising outside the lower extremity and those considered unable to respond due to evident psychiatric disorders or cognitive developmental impairments, as determined by their treating physicians.

Procedures

Potential participants meeting the eligibility criteria were identified by coauthors who were orthopedic surgeons at oncology hospitals. A written invitation and detailed study information were mailed to eligible individuals. Those who provided informed consent were enrolled, and self-administered questionnaires were distributed by mail. Upon questionnaire return, participants were asked about their willingness to participate in interviews; those who agreed were contacted by e-mail, and interviews were conducted either online or in person. All questionnaire respondents and interview participants received an honorarium. The study was conducted between May 2022 and August 2023 and was approved by the Ethics Review Committee of Tokyo Medical and Dental University (now the Institute of Science Tokyo) and collaborating institutions.

Self-administered questionnaire

Demographic data (gender, education, employment status), as well as disease-related information, were collected. The following factors: pain intensity and pain interference from Patient-Reported Outcome Measurement Information System (PROMIS),^24,25 physical function using Toronto Extremity Salvage Score (TESS),^26,27 fatigue and sleep disturbances from PROMIS, anxiety and depression (The Japanese version of the Hospital Anxiety and Depression Scale),²⁸ and social support (The Japanese version of Zimet’s perceived social support scale)²⁹ were considered (Table 1).

Table 1.

Measurements for Pain Intensity, Pain Interference, and Dependent Variable Scales

Variable	Instrument	Subscales/Items	Scoring scheme	Reliability
Pain intensity	Patient-Reported Outcome Measurement Information System (PROMIS) Japanese version	3 items	Total scores were converted into T-scores using a conversion table. Cutoff point: ≦55: within normal limits; 55–60: mild; 60–70: moderate; ≧70: severe	Cronbach’s α = 0.83 in this study
Pain interference	Patient-Reported Outcome Measurement Information System (PROMIS) Japanese version	6 items	Total scores were converted to T-scores using a conversion table Cutoff point: ≦55: within normal limits; 55–60: mild; 60–70: moderate; ≧70: severe	Cronbach’s α = 0.96 in this study
Physical function	Toronto Extremity Salvage Score (TESS) Japanese version	32 items	Total scores ranged from 0 to 100, with higher scores indicating higher levels of physical functioning	Cronbach’s α = 0.93 in this study
Sleep disturbance	Patient-Reported Outcome Measurement Information System (PROMIS) Japanese version	8 items	Total scores were converted into T-scores using a conversion table	Cronbach’s α = 0.89 in this study
Fatigue	Patient-Reported Outcome Measurement Information System (PROMIS) Japanese version	8 items	Total scores were converted into T-scores using a conversion table	Cronbach’s α = 0.96 in this study
Anxiety and depression	The Japanese version of the Hospital Anxiety and Depression Scale	7 items for anxiety; 7 items for depression	Score range (0–21); 0–7 = none, 8–10 = need to reevaluate, ≧11 = clinical anxiety or depression	Anxiety: Cronbach’s α = 0.79 Depression: Cronbach’s α = 0.57
Social support	Japanese version of Zimet’s perceived social support scale	Total 12 items; 4 items for every three factors: “support from family,” “support from the significant other,” and “support from friends”	The mean score was calculated for each subscale Higher scores indicate greater social support	Cronbach’s α = 0.93 in this study

Interview survey

Semistructured interviews lasting approximately 60 minutes were conducted either online or in person, according to the participants’ preference. In-person interviews were held in a private room at the first author’s office. All interviews were conducted by the first author and audio-recorded with participants’ consent. The interview guide included the following prompts: (1) “Tell me about the pain in the affected limb” and (2) “Tell me about your perception of the pain in the affected limb.”

Analysis

Quantitative data analysis

Statistical analyses were performed using SPSS version 30. Descriptive statistics summarized participants’ demographic characteristics, disease-related variables, and assessment scale scores. Differences in demographic characteristics, pain intensity, and pain interference were examined by conducting the Mann–Whitney U test or the Kruskal–Wallis test, as appropriate. Spearman’s rank correlation coefficients were calculated to examine relationships among the eight variables in the hypothesized model (pain intensity, pain interference, physical Functioning, fatigue, sleep disturbance, anxiety, depression, and social support), as well as variables showing significant bivariate differences (educational status and age at diagnosis). Multiple regression analyses were conducted to identify pain intensity and pain interference predictors. Given the limited sample size, variable reduction was applied to derive the final regression models, with selection based on the F-statistic and clinical interpretability.

Qualitative data analysis

The analysis of interview data aimed to identify (1) the characteristics of pain experienced by survivors and (2) pain-related experiences that complemented the multiple regression analysis model. After statistically analyzing quantitative data, interview transcripts were analyzed through Braun and Clarke’s reflective thematic analysis (TA), following a six-phase process with NVivo version 15.³⁰ Reflective TA is not limited to a single approach and can be flexibly selected according to research objectives. In this study, inductive coding was used to clarify pain characteristics, while deductive coding was utilized to clarify the explanation of the multiple regression analysis model.

Results

Demographic data

Of the 100 potential participants identified by coinvestigators across four oncology hospitals, 67 survivors consented to participate. Questionnaires were mailed to all 67 individuals, and 64 completed responses were received, yielding a response rate of 95.5%. Among the respondents, 33 participants (51.5%) expressed willingness to participate in interviews, and 26 (40.6% of respondents) ultimately completed the interviews.

The study sample comprised 38 male (59.4%) and 26 female (40.6%) participants, with a mean age of 27.5 years. The mean age at diagnosis was 16.3 years, and the mean time since diagnosis was 11.2 years. Educational attainment was categorized into two groups: 52 participants had post-high school education, and 12 had a high school or junior high school diploma. At the time of the survey, 46 participants were employed, and 18 were married (Table 2).

Table 2.

Pain Intensity and Pain Interference with Participants’ Characteristics

	n	%	Pain intensity					Pain interference
			Mean (SD)	p	Within normal	Mild	Moderate	Mean (SD)	p	Within normal	Mild	Moderate
Characteristic					≦55	55–60	60–70			≦55	55–60	60–70
All participants	64	100.0			46.2 (10)		50			7	7	51.4 (8.1)	36	20	8
Interviewees	26	40.6	46.4 (10)		20	4	2	51.1 (8.5)		16	6	4
Noninterviewees	38	59.4	46.0 (10)		30	3	5	51.6 (7.8)		20	14	4
Gender
Male	38	59.4	46.4 (10.1)	0.961^a	30	3	5	50.8 (8.0)	0.245^a	25	9	4
Female	26	30.6	45.8 (10.0)	0.961^a	20	4	2	52.3 (8.1)	0.245^a	11	11	4
Education
High school diploma	12	18.8	54.2 (9.8)	0.003 ^a	6	3	3	57.4 (6.6)	0.005 ^a	3	5	4
Post-high school	52	81.2	44.3 (9.2)	0.003 ^a	44	4	4	50.1 (7.8)	0.005 ^a	33	15	4
Employment
Yes	46	71.8	46.5 (9.9)	0.225^b	35	6	5	46.5 (9.9)	0.005 ^b	27	13	6
No	9	14.1	49.4 (11.2)		7	—	2	49.4 (11.2)		2	5	2
Students	9	14.1	41.7 (8.7)		8	1	—	41.7 (8.7)		7	2	—
Age at diagnosis
7–14	24		43.4 (9.2)	0.010 ^b	21	—	3	51.1 (8.1)	0.027 ^b	12	10	2
15–19	26		44.5 (9.4)		21	4	1	49.3 (7.5)		19	5	2
20–29	14		54.2 (8.8)		8	3	3	56.3 (7.5)		5	5	4
Duration since diagnosis
3–5	13		51.2 (11.2)	0.157^b	7	3	3	53.2 (8.7)	0.139^b	5	5	3
6–9	12		45.2 (10.1)		10	1	1	53.9 (7.0)		5	5	2
10–20	39		44.3 (9.2)		33	3	3	50.1 (8.0)		26	10	3
Therapy
Operations, chemotherapy, and rehabilitation	52	81.3	45.5 (12.0)		41	7	4	51.4 (10.0)		29	18	5
Operations and chemotherapy	8	12.5	49.6 (11.8)		6	—	2	52.3 (10.4)		4	2	2
Operations and rehabilitation	2	3.1	59.9 (7.1)		1	—	1	58.3 (5.4)		1	—	1
Chemotherapy	1	1.6	36.3 (−)		1	—	—	41.1 (−)		1	—	—
Operations	1	1.6	36.3 (−)		1	—	—	41.1 (−)		1	—	—
Surgery
Endoprosthesis replacement	50	78.1	46.5 (9.7)		39	6	5	52.3 (7.9)		26	17	7
Autologous/artificial bone grafting	6	9.4	46.0 (9.7)		4	—	2	49.4 (9.6)		4	1	1
Amputation	4	6.2	47.5 (10.8)		3	1	—	51.2 (8.1)		2	2	—
Tumor resection only	3	4.7	42.5 (13.6)		3	—	—	45.5 (7.7)		3	—	—
Rotationplasty	1	1.6	36.3 (−)		1	—	—	41.1 (−)		1	—	—
Tumor site
Femur	37	57.8	45.6 (9.7)		30	4	3	52.5 (7.6)		25	9	5
Tibia	21	32.8	48.2 (11.0)		14	3	4	52.6 (7.6)		10	8	3
Others	3	4.7	47.5 (9.8)		2	1	—	51.7 (9.5)		2	1	—
No answer	3	4.7	38.9 (2.3)		3	—	—	54.3 (5.1)		1	2	—
Metastasis/recurrence
Yes	10	15.6	49.1 (10.6)		7	1	2	52.0 (6.8)		6	3	1
No	54	84.4	45.4 (9.8)		43	5	5	51.3 (8.4)		30	16	7

P-values < 0.05 are shown in bold.

Mann–Whitney U test.

Kruskal–Wallis’s test.

The interview sample included 26 participants (16 male [61.5%] and 10 females [38.5%]). Interviews were conducted online or in person and lasted an average of 63 minutes. The mean age of interviewees was 28.2 years, the mean age at diagnosis was 16.7 years, and the mean time since diagnosis was 11.5 years. Regarding treatment history, 20 participants underwent endoprosthesis replacement, two underwent amputation, two underwent tumor resection only, and two underwent autologous or artificial bone grafting. At the time of the interview, 21 participants were employed, two were students, and three were unemployed.

Characteristics of pain in the affected limb

The mean and median scores for pain intensity and pain interference within the 7 days before the survey were as follows: pain intensity, 46.2 (SD = 10, range 36.3–64.9); pain interference, 51.4 (SD = 8.1, range 41.1–68.7). Participants who responded “None at all” to all three items on the pain Intensity scale and all six items on the pain Interference scale (n = 22) were classified as having no pain in the past 7 days. Accordingly, 42 participants (65.6%) reported experiencing some pain or pain interference. Comparisons of scores between interviewees and noninterviewees showed no distinctive pattern for the interviewee group. Pain intensity differed significantly by education status and age at diagnosis, whereas pain interference differed significantly by education status, employment status, and age at diagnosis (Table 2). Qualitative analysis of interviews identified four themes characterizing survivors’ pain in the affected limb: endoprosthesis-related pain, phantom limb pain, secondary pain from strain on the legs, and weather-related pain (Table 3).

Table 3.

Characteristics of Pain Determined from Interviews

Theme	Subtheme	Examples
Pain-related endoprosthesis	Knee pain around artificial knee joints	In my case, it’s my knee that causes pain. Above the knee, below the thigh. The muscles around the knee often hurt. (Male, 22 years old, PIT 36.3, PIF 41.1)
	Knee pain around artificial knee joints	Pain occurs around the knee plate or the back of the knee. I think it is at the joint. (Male, 29 years old, PIT 58.5, PIF 55.6)
	Pain at the connection between the artificial knee joint and the autologous bone	It usually hurts the next day when I walk long distances. As for where the pain comes from, I have an artificial joint in my lower leg, and my tibia is an artificial joint. (Male, 26 years old, PIT 51.4, PIF 55.6)
		The artificial joints are inserted in the center of the knee, so the burden is placed on the knee, and the knee is a little creaky or painful. (Female, 26 years old, PIT 36.3, PI 56.6)
	Pain with flexion of the knee	Sometimes I get pain when I stand up suddenly, or when I know I’m going to stand up, but I also get pain when I stand up suddenly. (Male, 27 years old, PIT 54.8, PIF 52.2)
	Pain with flexion of the knee	I walk while extending my knee. I received such rehabilitation and was instructed to walk. So, when I had to return from a bent knee to an extended knee, I felt lots of pain. (Male, 27 years old, PIT 52.1, PIF 58.6)
	Pain caused by an artificial knee joint not fitting properly	I just replaced them this summer, so right after replacing them, they were a bit uncomfortable. It’s more like they were slipping and causing pain or not fitting my feet properly. They still haven’t fully adjusted to my feet yet, so there’s some pain. (Male, 29 years old, PIT 58.5, PIF 55.6)
Pain caused by amputation	Pain caused by a misfitting prosthesis leg	The pain from my prosthetic leg is pretty serious. When it feels like it doesn’t fit right today, I worry, “Is it going to hurt? What if it does...?” Right now, it’s worse than the phantom limb pain... (Male, 31 years old, PIT 40.2, PIF 48.6)
	Phantom limb pain	The phantom pain never goes away. Even now, when I am slightly conscious, I sometimes subtly feel phantom limbs, though it is not pain. I am still a little concerned about it. (Male, 31 years old, PIT 40.2, PIF 48.6)
	Phantom limb pain	Suddenly, I get phantom limb pain. It’s like a needle has been inserted into my residual area. If I rub it a little, it goes away, but I never know when it’s coming. Sometimes, it happens when I’m going to bed or when I’m walking outside. (Female, 31 years old, PIT, 54.8, PIF 59.5)
Secondary pain from strain on the legs	Body weight-related pain	If you are carrying a larger load than your muscles can support, or if the load is more than your own body weight, the damage is great. I think there is an appropriate weight for one’s own physique. When I start to feel that I am gaining weight, my knee joints start to feel pain or strain. (Male, 26 years old, PIT 36.3, PTF 41.1)
	Hip and ankle joint pain	If I walk with the image of rotating my knee, it relieves some knee pain, and I can use it for a longer period. But since it is almost impossible to be aware of it, the hip joint hurts instead. (Male, 33 years old, PIT 58.5, PIF 64.8)
	Pain associated with imbalance	I sometimes have pain in other parts of my body due to my body trying to keep balance. I have heard that the burden is placed on the diagonal. In my case, I have a lot of burden on my left ankle. (Female, 24 years old, PIT 36.3, PIF 41.1)
	Pain associated with imbalance	I also had some pain in my lower back. The leg with the artificial joint is longer than the other one. If I walk all the time as it is, my lower back sometimes hurts. (Female, 25 years old, PIT 54.8, PIF 60.4)
Weather-related pain		Sometimes, when the weather gets bad or when it’s cold, I get unrelieved pain, and then I get pain even when I walk. (Male, 27 years old, PIT 43.1, PIF 52.2)
Weather-related pain		My knee usually hurts a little when there is a sudden change in temperature or when it gets cold. (Female, 26 years old, PIT 36.3, PIF 56.6)

PIT, pain intensity; PIF, pain interference.

Variables related to pain intensity and interference

Factors related to pain-including physical functioning (TESS), fatigue, sleep disturbance, anxiety, depression, and social support-were measured. Scores were calculated separately for interviewees and noninterviewees, with no distinctive pattern observed in the interviewee group. Variable scores for all participants, interviewees, and noninterviewees are presented in Table 4.

Table 4.

Descriptive Statistics of Related Variables

	All participants (n = 64)		Interviewees (n = 26)		Noninterviewees (n = 38)
Variable	Mean (SD)	Range	Mean (SD)	Range	Mean (SD)	Range
Age at the time of the survey	27.5 (5.2)	18–39	28.2 (4.5)	18–37	27.0 (5.5)	18–39
Age at diagnosis	16.3 (5.0)	7–29	16.7 (4.9)	7–28	16.0 (5.1)	7–29
Years from diagnosis	11.2 (4.7)	3–20	11.5 (4.4)	3–17	11.0 (5.0)	3–20
TESS	84.0 (11.4)	57.1–100.0	86.5 (11.5)	57.1–100.0	82.3 (11.2)	57.1–100.0
Fatigue	49.4 (11.1)	33.1–77.8	47.5 (13.3)	33.1–77.8	50.1 (9.5)	33.1–68.7
Sleep disturbance	49.8 (10.0)	30.5–71.8	49.4 (10.0)	30.5–71.8	50.4 (8.5)	30.5–68.7
Anxiety	6.0 (4.4)	0.0–20.0	6.3 (4.5)	0.0–18.0	6.0 (4.4)	0.0–20.0
Depression	5.8 (3.8)	1.0–17.0	5.6 (4.3)	1.0–17.0	6.0 (3.7)	1.0–17.0
Social support	5.8 (1.1)	3.0–7.0	5.8 (1.1)	3.0–7.0	5.9 (1.1)	3.0–7.0

TESS, Toronto Extremity Salvage Score.

Multiple regression analysis with pain intensity and pain interference

Tables 5 and 6 present the correlation and multiple regression analyses. Pain intensity was positively correlated with pain interference, fatigue, sleep disturbance, depression, and age at diagnosis, and negatively correlated with physical functioning and education status. Multiple regression analysis (adjusted R² = 0.56) showed that pain interference (β = 0.49, p < 0.001), physical functioning (β = −0.21, p = 0.035), social support (β = −0.16, p = 0.039), and age at diagnosis (β = 0.20, p = 0.003) significantly predicted pain intensity.

Table 5.

Factors Related to Pain Intensity

	n	r	Forced entry model			Variable reduction model
Variable	n	r	Standardization β	t-value	p	Standardization β	t-value	p
Pain interference	64	0.703^**	0.489	4.264	<0.001	0.493	4.887	<0.001
TESS	64	−0.453^**	−0.186	−1.885	0.065	−0.208	−2.159	0.035
Fatigue	63	0.329^**	−0.073	−0.570	0.571
Sleep disturbance	63	0.281^*	0.082	0.682	0.499
Anxiety	64	0.188	−0.198	−1.485	0.144	−0.125	−1.281	0.206
Depression	63	0.348^**	0.109	0.785	0.436
Social support	64	−0.179	−0.152	−1.523	0.134	−0.195	−2.110	0.039
Education status	64	−0.380^**	−0.088	−0.887	0.379
Age at diagnosis	63	0.385^**	0.258	2.525	0.015	0.300	3.123	0.003

P-values < 0.05 are shown in bold.

r, Spearman’s rank correlation coefficient, **p < 0.01, *p < 0.05.

Forced entry model β: Standardized partial regression coefficient by multiple linear regression analysis (n = 60, R² = 0.550).

Variable reduction model β: Standardized partial regression coefficient by multiple linear regression analysis (n = 60, R² = 0.564).

TESS, Toronto Extremity Salvage Score.

Table 6.

Factors Related to Pain Interference

	N	r	Forced entry model			Variable reduction model
Variable	N	r	Standardization β	t-value	p	Standardization β	t-value	p
Pain intensity	64	0.703^**	0.537	4.654	<0.001	0.560	5.395	<0.001
TESS	64	−0.380^**	−0.135	−1.283	0.205	−0.130	−1.288	0.203
Fatigue	63	0.484^**	0.368	2.935	0.005	0.355	3.316	0.002
Sleep disturbance	63	0.192	−0.153	−1.222	0.227	−0.157	−1.469	0.147
Anxiety	64	0.224	0.049	0.346	0.730
Depression	63	0.279^*	−0.094	−0.646	0.521
Social support	64	−0.148	−0.047	−0.437	0.664
Education status	64	−0.357^**	−0.116	−1.116	0.270
Age at diagnosis	63	0.273	−0.025	−0.220	0.827

P-values < 0.05 are shown in bold.

r, Spearman’s rank correlation coefficient, **p < 0.01, *p < 0.05.

Forced entry model β: Standardized partial regression coefficient by multiple linear regression analysis (n = 60, R² = 0.505).

Variable reduction model β: Standardized partial regression coefficient by multiple linear regression analysis (n = 60, R² = 0.533).

TESS, Toronto Extremity Salvage Score.

Pain interference was positively correlated with pain intensity, fatigue and depression, and negatively correlated with physical functioning and education status. Multiple regression analysis (adjusted R² = 0.53) indicated that pain intensity (β = 0.56, p < 0.001) and fatigue (β = 0.36, p = 0.002) significantly predicted pain interference.

Interview results complemented by multiple regression analysis findings

Based on the multiple regression analysis results, we identified three themes: “Physical Function and Pain Intensity,” “Social Support and Pain Intensity,” and “Fatigue and Pain Interference” (Table 7).

Table 7.

Interview Results Complemented by Multiple Regression Analysis Findings

Theme	Subtheme	Examples
Physical Function and Pain Intensity	Pain caused by stiff artificial knee joints	Whether I sit for extended periods or stand for long stretches, my knees stiffen up, and bending them can sometimes be extremely painful. (Female, 24 years old, PIT 60.4, PIF 58.5)
		Bending my knee once causes pain and a catching sensation. I think it’s the bending part of the artificial knee joint. Sometimes it suddenly happens where I can’t straighten it because of the catching sensation, and if I force it straight, it catches again and I can’t bend it...（Male, 33 years old, PIT 36.3, PIF 48.6)
		If I keep my leg stretched out like this, it won’t bend anymore, and if I keep it bent, it won’t straighten out. It’s really painful. (Female, 18 years old, PIT 40.2, PIF 41.1)
	Pain due to the restricted mobility of an artificial knee joint	I think it probably feels like the joint movement becomes less smooth. Also, if I don’t consciously move my knee by bending and straightening it every day, it gradually gets harder to bend. The pain at that time is painful. (Male, 27 years old, PIT 52.1, PIF 58.6)
		The knee joint’s range of motion is still limited to about 90 degrees. So, when I relax my strength and it suddenly buckles, if my foot twists slightly in that position, it kind of jams. If my leg bends beyond what I expect or beyond my actual range of motion, I get sudden, sharp pain. (Male, 26 years old, PIT 36.3, PTF 41.1)
		When I try to straighten my left leg... it just won’t straighten, you know? It’s like when you touch something hot, and your leg jerks reflexively—that’s the worst part. And when I can’t walk because of it... that’s absolutely the worst. (Male, 29 years old, PIT 58.5, PIF 55.6)
Social Support and Pain Intensity	Preventing pain with support	Saying “I’m tired” is easier to brush off with “Let’s take a break” than saying “It hurts.” If you say, “It hurts,” people ask “What happened? Are you okay?” But even then, no one can really do anything about it. I don’t know how to fix it myself either. So, I just tell myself, “Resting will make it better. (Female, 18 years old, PIT 40.2, PIF 41.1)
		If it really is painful and I can’t walk, or if I just need a little break, I’ll say so without hesitation. I only say it to people who will accommodate me. (Male, 27 years old, PIT 36.3, PTF 41.1)
		Long-time friends understand me, so they give me breaks. With people I’ve just met, I can’t really say anything, so I often end up apologizing early and cutting things short to head home. (Female, 25 years old, PIT 54.8, PIF 60.4)
	Efforts to understand pain	To others, I just look perfectly normal. I’m not even using a cane. But even though I look fine, I have to make sure to tell people—even if I hate it—that my knees are actually painful, prone to hurting, and fragile. Otherwise, I’ll just suffer myself. Well, it’s not easy to say, though. (Female, 25 years old, PIT 40.2, PIF 57.6)
	Efforts to understand pain	I persevered for about two or three years, but the pain made it difficult to continue. When I consulted my supervisor, they adjusted my duties, so I no longer have to do it. My supervisor’s understanding was incredibly important. (Male, 27 years old, PIT 43.1, PIF 52.2)
Fatigue and Pain Interference	Take a break to prevent fatigue and pain	When pain occurs, if you keep moving without resting, it can worsen or intensify. So I try to rest as soon as I feel any pain. (Male, 27 years old, PIT 54.8, PIF 52.2)
	Take a break to prevent fatigue and pain	I’ve gotten to the point where I can pretty much sense when something might start feeling painful, so I try not to stay outside for too long. (Male, 33 years old, PIT 58.5, PIF 64.8)
	Fatigue and pain interfere with the daily plan	After walking for a long time, my feet start to pain, so sometimes I have to decide to leave a little earlier... Even when I’m hanging out with friends. (Female, 25 years old, PIT 54.8, PIF 60.4)
	Fatigue and pain interfere with the daily plan	When I go out, I think about doing this and that, going here and there, visiting that shop too. But after doing a few things, I often end up thinking, “I’m tired, so I’ll skip this,” or “It hurts, so I’ll head home early.” My plans might change. (Female, 25 years old, PIT 40.2, PIF 57.6)
	Fatigue as a precursor of pain	When fatigue or tiredness sets in, I sometimes realize my knees are weaker than I imagined. That’s when I start stumbling more often. Like when going downstairs, I’ll suddenly feel like I’m about to miss a step... (Male, 26 years old, PIT 36.3, PTF 41.1)
		When I walk too much without thinking and get tired, I sometimes feel a faint sensation like pain—almost like a warning sign saying, “Here it comes.” For example, in the past, I’d feel a sharp pain when standing up after sitting for a long time. It was like a little marker telling me, “Pain might be coming...” (Male, 27 years old, PIT 36.3, PTF 41.1)
		While I didn’t feel it at all during work, sometimes when I get home and think “I’m tired,” I realize it might actually be pain. (Male, 31 years old, PIT 40.2, PIF 48.6)

PIT, pain intensity; PIF, pain interference.

Discussion

In this study, low physical function, older age at diagnosis, and perceived lack of social support were associated with pain intensity, while fatigue severity was associated with pain interference. Pain intensity and pain interference were strongly related. To our knowledge, this is the first study to investigate self-reported pain characteristics in the affected limb of long-term osteosarcoma survivors, including pain intensity, pain interference, and associated factors.

The prevalence of pain among survivors was 65%, exceeding the 40% reported in previous studies.^4,13,31 However, mean PROMIS pain intensity and interference scores were 46.2 and 51.4, respectively. Although direct comparisons with prior studies of malignant bone tumor survivors are not possible, the general US population mean of 50 suggests that many participants remained within the normal range.²⁴ Nonetheless, individual variability indicated considerable differences in both pain intensity and interference within the cohort. Fifty participants had undergone endoprosthesis replacement in this study. Complications of endoprosthesis are defined as soft-tissue failures, aseptic loosening, structural failures, infection, and tumor progression.³² It is unclear whether complications occurred in this study’s participants. However, they reported the pain around knee endoprosthesis in interviews, suggesting a possible association with some form of complication. In addition, survivors’ lived experiences suggested that pain may occur within recurring contextual patterns. Future research requires to identify potential causes of pain around knee endoprosthesis.

The pain intensity model identified low physical function, older age at diagnosis, and perceived lack of social support as significant predictors. Pain interference was primarily exacerbated by severe fatigue. Older age at diagnosis was associated with greater pain intensity,^2,4,33 and lower social support correlated with higher pain, which may reduce from support by family, friends, and health care orividers.^34,35 Interviews suggested that empathetic support, understanding of pain, and daily life adjustments may alleviate pain. Malignant bone tumors such as osteosarcoma are rare cancers, so the number of survivors is relatively small, making it difficult to understand their pain. Therefore, they worked hard to gain that understanding. Mindfulness intervention studies suggest that family support may reduce postoperative pain in children.³⁶ Young adult survivors who are financially independent may have less access to family support compared with childhood and may experience higher pain intensity. For AYA survivors, peer support has been shown to reduce isolation and promote coping.³⁷ Facilitating peer networking and providing guidance on self-care may mitigate pain experiences.

The mean TESS score was 84.0, consistent with prior studies (82.9–89.0),^38,39 and low physical function has been linked to pain.^3,4,40 Exercise interventions have been found to improve walking function, and physical performance⁴¹ and may reduce pain. In this study, several participants also reported that moderate leg activity helped improve their pain. In addition, although interview responses included descriptions of becoming easily fatigued during daily activities and rescheduling plans to avoid pain, the term “fatigue” itself was not explicitly mentioned. This suggests that participants’ fatigue awareness or recognition may be limited. However, some participants often described fatigue as “feeling tired” preceding pain, which can contribute to mental health difficulties⁴² and affect role functioning.⁴⁰ Interventions targeting fatigue, including exercise, are therefore important.⁴³ The relatively low average pain scores suggest that survivors may be effectively managing their pain,⁴⁴ highlighting the need for further research on self-care and pain management strategies.

Several limitations should be noted. The cross-sectional design precluded assessment of temporal changes and causal inference. Despite being the largest self-reported sample of long-term osteosarcoma survivors in Japan, the sample size was insufficient for robust statistical analyses, and nonparticipant data were unavailable, potentially underestimating pain prevalence. Subjective self-report introduces possible response bias. Routine analgesic use was not assessed, which may have influenced reported pain intensity. In addition, orthopedic data, including tumor resection status and prosthesis details, were limited. Future studies should address these factors and evaluate analgesic use to clarify pain mechanisms and management.

Conclusion

This study demonstrated that pain among long-term osteosarcoma survivors is influenced by multiple biopsychosocial factors. Lower physical function, older age at diagnosis, and perceived lack of social support were associated with higher pain intensity, while greater fatigue severity was associated with pain interference. Although average pain scores were within the normal range, substantial individual variability indicated heterogeneous pain experiences. Qualitative findings described prosthesis-related pain and the importance of understanding and support from others. These results imply the need for comprehensive pain management in AYA cancer survivors, incorporating interventions to improve physical function, reduce fatigue, and enhance social support. Future studies are needed to examine the effectiveness of these interventions. Health care providers should recognize the characteristics of pain in AYA osteosarcoma survivors and its long-term impact on daily life.

Authors’ Contributions

S.K.: Conceptualization (lead), Methodology (lead), investigation (lead), formal analysis (lead), writing—original draft (lead), and writing—review and editing (equal). M.O.: Conceptualization (supporting), methodology (supporting), formal analysis (supporting), writing—original draft (supporting), and writing—review and editing (equal). S.Y.: Conceptualization (supporting), methodology (supporting), formal analysis (supporting), writing—original draft (supporting), and writing—review and editing (equal). S.I.: Investigation (supporting) and writing—review and editing (equal). K.A.: Investigation (supporting) and writing—review and editing (equal). K.H.: Investigation (supporting) and writing—review and editing (equal). T.G.: Investigation (supporting) and writing—review and editing (equal). T.Y.: Investigation (supporting) and writing—review and editing (equal).

Disclaimer

Portions of this study were presented at a scientific meeting; this article represents the first full publication.

Ethics Approval and Consent to Participate

Ethical approval for this study was obtained from the Ethics Committee of Tokyo Medical and Dental University (Currently Institute of Science Tokyo) School of Medicine under a batch review (M2021-186). Following this, permission was granted by the director of the collaborating institution. Participants were informed in writing about the study’s purpose and the protection of personal information. Consent was obtained by allowing participants to indicate agreement in the questionnaire survey and by explaining the study’s purpose in writing, online, or in person during interviews. Participants provided signed consent either manually or through an electronic signature.

Footnotes

Acknowledgment

The authors sincerely thank the survivors of lower limb osteosarcoma who participated in this study. The authors also express gratitude to Dr. Mitsue Maru (Vice Superintendent and Director of Nursing, Yodogawa Christian Hospital) for providing expert opinion and advice during the development of research plans and data analysis, as well as Professor Iori Sato (Institute of Science Tokyo) for providing expert opinion and advice during data analysis and reviewing and editing the article as a supervisor.

Author Disclosure Statement

No competing financial interests exist.

Funding Information

This work was supported by a grant-in-aid of the YAMAJI FUMIKO NURSING RESEARCH FUND 2020 and JST SPRING, Grant Number JPMJSP2120.

References

1. Anghelescu

, Steen

, Wu

, et al. Prospective study of neuropathic pain after definitive surgery for extremity osteosarcoma in a pediatric population. Pediatr Blood Cancer 2017;64(3); doi: 10.1002/pbc.26162

2. Allen

, Niel

, Guo

, et al. Psychosocial factors and psychological interventions: Implications for chronic post-surgical pain in pediatric patients with osteosarcoma. J Clin Psychol Med Settings 2021;28(3):468–476; doi: 10.1007/s10880-020-09748-y

3. Furtado

, Grimer

, Cool

, et al. Physical functioning, pain and quality of life after amputation for musculoskeletal tumours: A national survey. Bone Joint J 20152015;97-B(9):1284–1290; doi: 10.1302/0301-620X.97B9.35192

4. Katsumoto

, Maru

, Yonemoto

, et al. Uncertainty in young adult survivors of childhood and adolescent cancer with lower-extremity bone tumors in Japan. J Adolesc Young Adult Oncol 2019;8(3):291–296; doi: 10.1089/jayao.2018.0120

5. Treede

, Rief

, Barke

, et al. A classification of chronic pain for ICD-11. Pain 2015;156(6):1003–1007; doi: 10.1097/j.pain.0000000000000160

6. Taylor

, Pooley

. Sarcoma survivors’ perspectives on their body image and functional quality of life post-resection/limb salvage surgery. Eur J Cancer Care (Engl) 2017;26(6):e12667; doi: 10.1111/ecc.12667

7. Ogura

, Kawano

, Shimose

, et al. Osteosarcoma in Japan: Report from the bone and soft tissue tumor registry 2006–2022. Jpn J Clin Oncol 2025;55(9):1046–1053; doi: 10.1093/jjco/hyaf094

8.Foundation for Promotion of Cancer Research. Cancer statistics in Japan 2024. Tokyo; 2024.

9. Yamamoto

, Araki

, Tsuchiya

. Joint-preservation surgery for bone sarcoma in adolescents and young adults. Int J Clin Oncol 2023;28(1):12–27; doi: 10.1007/s10147-022-02154-4

10.

10. Han

, Bi

, Xu

, et al. Amputation versus limb-salvage surgery in patients with osteosarcoma: A meta-analysis. World J Surg 2016;40(8):2016–2027; doi: 10.1007/s00268-016-3500-7

11.

11. Husson

, Zebrack

. Psychometric evaluation of an adolescent and young adult module of the impact of cancer instrument. J Adolesc Young Adult Oncol 2017;6(1):159–170; doi: 10.1089/jayao.2016.0038

12.

12. Fauske

, Bondevik

, Bruland

ØS

, et al. Negative and positive consequences of cancer treatment experienced by long-term osteosarcoma survivors: A qualitative study. Anticancer Res 2015;35(11):6081–6090.

13.

13. Heinemann

, Hoffmann

, Hardes

, et al. Pain in survivors of Ewing sarcoma: Prevalence, associated factors and prediction of recurrence. Pediatr Blood Cancer 2021;68(3):e28801; doi: 10.1002/pbc.28801

14.

14. Kumar

, Sharma

, Shantanu

, et al. Limb salvage strategy amendment for a better future in the era of bone cancer therapy: A cross-sectional study in North India. Cureus 2023;15(7):e41768; doi: 10.7759/cureus.41768

15.

15. Kuffler

. Coping with phantom limb pain. Mol Neurobiol 2018;55(1):70–84; doi: 10.1007/s12035-017-0718-9

16.

16. Wong

, Svingos

, Heaton

. Behavioral therapy for acute post-surgical phantom limb pain in a young adult with osteosarcoma. Pain Pract 2019;19(5):571–572; doi: 10.1111/papr.12768

17.

17. DeMoss

, Ramsey

, Karlson

. Phantom limb pain in pediatric oncology. Front Neurol 2018;9:219; doi: 10.3389/fneur.2018.00219

18.

18. Raja

, Carr

, Cohen

, et al. The revised International Association for the Study of Pain definition of pain: Concepts, challenges, and compromises. Pain 2020;161(9):1976–1982; doi: 10.1097/j.pain.0000000000001939

19.

19. Alberts

, Leisenring

, Whitton

, et al. Characterization of chronic pain, pain interference, and daily pain experiences in adult survivors of childhood cancer: A report from the Childhood Cancer Survivor Study. Pain 2024;165(11):2530–2543; doi: 10.1097/j.pain.0000000000003284

20.

20. Bevers

, Watts

, Kishino

, et al. The biopsychosocial model of the assessment, prevention, and treatment of chronic pain. US Neurol 2016;12:98–104; doi: 10.17925/USN.2016.12.02.98

21.

21. Karlson

, Alberts

, Liu

, et al. Longitudinal pain and pain interference in long-term survivors of childhood cancer: A report from the Childhood Cancer Survivor Study. Cancer 2020;126(12):2915–2923; doi: 10.1002/cncr.32853

22.

22. Wilson

. Integrating the concept of pain interference into pain management. Pain Manag Nurs 2014;15(2):499–505; doi: 10.1016/j.pmn.2011.06.004

23.

23. Edelen

, Rodriguez

, Herman

, et al. Crosswalking the patient-reported outcomes measurement information system physical function, pain interference, and pain intensity scores to the Roland-Morris disability questionnaire and the Oswestry disability index. Arch Phys Med Rehabil 2021;102(7):1317–1323; doi: 10.1016/j.apmr.2021.02.014

24.

24. Cella

, Choi

, Garcia

, et al. Setting standards for severity of common symptoms in oncology using the PROMIS item banks and expert judgment. Qual Life Res 2014;23(10):2651–2661; doi: 10.1007/s11136-014-0732-6

25.

25.Health Measures. PROMIS score cut points: PROMIS Reference Populations. HealthMeasures.net. Available from: https://www.healthmeasures.net/score-and-interpret/interpret-scores/promis/reference-populations?utm_source=chatgpt.com [Last accessed: December 23, 2025].

26.

26. Davis

, Wright

, Williams

, et al. Development of a measure of physical function for patients with bone and soft tissue sarcoma. Qual Life Res 1996;5(5):508–516; doi: 10.1007/BF00540024

27.

27. Ogura

, Uehara

, Akiyama

, et al. Cross-cultural adaptation and validation of the Japanese version of the Toronto Extremity Salvage Score (TESS) for patients with malignant musculoskeletal tumors in the lower extremities. J Orthop Sci 2015;20(6):1098–1105; doi: 10.1007/s00776-015-0767-8

28.

28. Hatta

, Higashi

, Yashiro

, et al. A validation of the hospital anxiety and depression scale [in Japanese]. Jpn J Psycho Som Med 1998;38:309–315; doi: 10.15064/jjpm.38.5_309

29.

29. Iwasa

, Gondo

, Masui

, et al. Reliability and validity of the Japanese version of the Social Support Scale—Study targeting middle-aged and older adults -[in Japanese]. J Health Welf Stat 2007;54(6):26–33.

30.

30. Braun

, Clarke

. Thematic analysis: A practical guide. Sage: London; 2022.

31.

31. Patton

, Forster

, Forbes

, et al. Characterizing pain in long-term survivors of childhood cancer. Support Care Cancer 2022;30(1):295–303; doi: 10.1007/s00520-021-06386-4

32.

32. Henderson

, Groundland

, Pala

, et al. Failure mode classification for tumor endoprostheses: Retrospective review of five institutions and a literature review. J Bone Joint Surg Am 2011;93(5):418–429; doi: 10.2106/JBJS.J.00834

33.

33. Ramos

, Mierke

, Filler

, et al. Outpatient opioid prescribing habits in pediatric patients with bone sarcomas after undergoing primary tumor resection. J Pediatr Orthop 2022;42(5):e501–e506; doi: 10.1097/bpo.0000000000002109

34.

34. Corey

, Haase

, Azzouz

, et al. Social support and symptom distress in adolescents/young adults with cancer. J Pediatr Oncol Nurs 2008;25(5):275–284; doi: 10.1177/1043454208321117

35.

35. Pennant

, Lee

, Holm

, et al. The role of social support in adolescent/young adults coping with cancer treatment. Children (Basel) 2019;7(1):2; doi: 10.3390/children7010002

36.

36. Donovan

, Martin

, Seidman

, et al. A Mobile-based mindfulness and social support program for adolescents and young adults with sarcoma: Development and pilot testing. JMIR Mhealth Uhealth 2019;7(3):e10921; doi: 10.2196/10921

37.

37. Morris

, Lepore

, Wilson

, et al. Adopting a survivor identity after cancer in a peer support context. J Cancer Surviv 2014;8(3):427–436; doi: 10.1007/s11764-014-0355-5

38.

38. Morri

, Bekkering

, Cotti

, et al. Cross-cultural validation of the Italian version of the BT-DUX: A subjective measure of health-related quality of life in patients who underwent surgery for lower extremity malignant bone tumour. Cancers (Basel) 2020;12(8):2015; doi: 10.3390/cancers12082015

39.

39. Morgan

, Friedly

, Amtmann

, et al. Cross-sectional assessment of factors related to pain intensity and pain interference in lower limb prosthesis users. Arch Phys Med Rehabil 2017;98(1):105–113; doi: 10.1016/j.apmr.2016.09.118

40.

40. Jin

, Hong

, Rhea

, et al. PROMIS Physical Function and Pain Interference scores correlate with the lower extremity Toronto Extremity Salvage Score. JB JS Open Access 2023;8(3):e23; doi: 10.2106/JBJS.OA.23.00011

41.

41. Basteck

, Guder

, Dirksen

, et al. Effects of an exercise intervention on gait function in young survivors of osteosarcoma with megaendoprosthesis of the lower extremity—Results from the pilot randomized controlled trial proGAIT. Curr Oncol 2022;29(10):7754–7767; doi: 10.3390/curroncol29100613

42.

42. Saravanan

, Bajaj

, Matthews

, et al. Fatigue and depressive mood in chronic low back pain. Pain Manag Nurs 2024;25(3):225–230; doi: 10.1016/j.pmn.2024.02.003

43.

43. Almeida

, Lima

, Martins

. Understanding quality of life’s challenges in sarcoma patients: A mixed-methods study. Eur J Oncol Nurs 2024;70:102585; doi: 10.1016/j.ejon.2024.102585

44.

44. Katsumoto

, Okamitsu

, Yago

. A qualitative literature review: Impact on life resulting from changes in physical function as perceived by adolescent young adult survivors of malignant bone tumors [in Japanese]. Bull Tokyo Healthc Univ 2024;19:119–125.