Risk of fracture among lung cancer survivors by cancer treatment modality: A population-based retrospective cohort study Authors: Seonghye Kim, Kyung-Do Han, Jinhyung Jung, Dong Woog Yoon, Jong Ho Cho, Dong Wook Shin Published: 2026 ABSTRACT AbstractIntroductionEvidence on fracture incidence among lung cancer (LC) survivors remains limited. We assessed fracture risk in LC survivors who underwent curative-intent surgery, stratified by treatment modality and postoperative time.MethodsThis retrospective, population-based cohort study used the Korean National Health Insurance Service database (2009–2022), including LC survivors (n = 33,292) and 1:3 age- and sex-matched controls (n = 99,876). Competing risk analyses were performed with all-cause mortality as a competing risk.ResultsLC survivors had an elevated risk of fracture compared with non-cancer controls across the overall follow-up period. Vertebral fracture risk was most pronounced within the first year after surgery (cause-specific hazard ratios [csHRs] 1.81, 95% confidence interval [CI] 1.45–2.26). Hip fracture risk within the first year was similarly directed but did not reach significance in the overall cohort (csHR 1.40, 95% CI 0.98–1.98), whereas it was significantly elevated in the screening subgroup after additional adjustment for health behaviors and body mass index (csHR 1.83, 95% CI 1.11–4.43). During the overall follow-up period, the risk of any fracture was elevated among LC survivors who received chemotherapy (csHR 1.32, 95% CI 1.21–1.45), radiation therapy (csHR 1.78, 95% CI 1.43–2.23), or chemoradiotherapy (csHR 1.73, 95% CI 1.50–1.99), compared with controls.ConclusionCompared with non-cancer controls, LC survivors had a generally higher risk of fracture at various sites, which varied by time since surgery and by cancer treatment modality. These findings suggest that osteoporosis screening and multimodal fracture prevention strategies should be considered for LC survivors.Keywords: Lung cancer, Cancer survivor, Fracture, Chemotherapy, Radiation therapy, Chemoradiotherapy Introduction Lung cancer (LC) has been the primary cause of both cancer incidence and mortality globally. [1] As survival rates for LC continue to rise, there is a growing need to address comorbid conditions, such as osteoporosis and fracture. [2] Although the risk of osteoporosis and fractures has been extensively assessed in breast, prostate, and hematological malignancies, [3], [4], [5], [6], [7] the literature on osteoporosis and fracture risk among LC survivors remains limited. A US-based cross-sectional study utilizing the National Health and Nutrition Examination Survey database indicated a higher prevalence of osteoporosis and fractures in LC survivors relative to a control group matched for demographic variables. [8] Regarding fracture risk, a French single-center study of 289 LC survivors reported a cumulative incidence of osteoporotic vertebral fracture of 12.5% over a median follow-up of 29.7 months, with an incidence rate of 4 per 100 patient-years. [9] Recently, a UK study utilizing the Clinical Practice Research Datalink demonstrated that LC survivors experienced a 1.4-fold increased risk of any fracture compared to non-cancer controls. This elevated risk peaked 1–2 years after LC diagnosis, subsequently declined, but remained higher than that observed in controls. Nonetheless, these studies had several limitations, including their cross-sectional design, [8] absence of a control group, [9] and failure to separately report vertebral and hip fractures. [9] Additionally, no prior study has systematically stratified fracture risk across the full spectrum of cancer treatment modalities, including surgery alone, chemotherapy (ChT), radiation therapy (RT), and chemoradiotherapy (CRT) (Supplementary Table S1). [3], [8], [9] Given the methodological constraints of previous research, the present study sought to evaluate the risks of any fracture, as well as vertebral and hip fractures, among LC survivors who underwent curative-intent surgery and their matched non-cancer controls, stratified by cancer treatment modalities and follow-up duration after surgery. Materials and methods Data source We conducted a population-based retrospective cohort study utilizing data from the Korean National Health Insurance System (KNHIS) database, collected between 2009 and 2022. Approximately 97% of Koreans are enrolled in the KNHIS, while the remaining 3% are covered by Medical Aid. The KNHIS database contains health information including medical claims based on International Classification of Disease, Tenth Revision (ICD-10) codes, demographic data, insurance qualifications, and mortality information. Koreans aged ≥40 years, or employed and self-employed individuals who pay insurance premiums regardless of age, are eligible for national general health screenings every two years. Consequently, information such as anthropometric measurements, laboratory test results, and health behavioral characteristics can be extracted from the KNHIS database. [10] Previous epidemiological studies have examined the risk of fractures among survivors of stomach, thyroid, and prostate cancer using the KNHIS database. [11], [12], [13] Study population We initially identified 52,516 patients who underwent curative-intent surgery for LC within the first year after LC diagnosis between 2010 and 2017. LC survivors were identified by an operational definition incorporating both the ICD-10 codes (C33, C34) and the Rare and Intractable Disease (RID) registration codes (V193). Previous reports indicate that using both the ICD-10 and RID registration codes to define cancer incidence improves accuracy compared to relying solely on the ICD-10 codes of primary diagnosis. [14] This increased accuracy is likely attributable to the RID registration system, which offers cancer-related medical expense reductions to patients newly diagnosed with cancer for at least five years. [14], [15] The index date was defined as the date of surgery for LC. We excluded LC survivors who 1) had been diagnosed with cancers other than LC (n = 15,272), 2) were younger than 20 years old (n = 29), or 3) had prevalent fractures (n = 3923) at the index date (Fig. 1). Flow diagram illustrating enrollment of study participants (all participants). Abbreviations: LC, lung cancer; N, number. To establish matched controls, we initially obtained a control pool of 1,135,922 participants without LC from the KNHIS database, who were matched by sex and birth year. We applied the identical exclusion criteria used for LC survivors to this control pool. For the control group, the surgery date assigned to the corresponding LC survivors served as the index date. Subsequently, a 1:3 exact matching was performed based on age and sex, resulting in the final inclusion of 33,292 LC survivors and 99,876 matched controls in the study (Fig. 1). Additionally, we defined screening groups comprising LC survivors and non-cancer controls who participated in the national health screening program within two years prior to the index date. This approach enabled us to incorporate health screening data, including anthropometric measures and health behavioral characteristics, to enhance the adjustment for potential confounders. Ultimately, 22,773 LC survivors and 57,728 non-cancer controls were included in this analysis. LC survivors were classified into four groups according to the cancer treatment modalities received: surgery only, surgery with ChT, surgery with RT, and surgery with CRT. Study outcomes and follow-up We identified incident any, vertebral, and hip fractures using the ICD-10 codes from administrative data (Supplementary Table S2), a methodology that has been validated in prior studies investigating fracture risk among cancer survivors and other patient populations. [11], [12], [13], [16], [17], [18], [19] ICD-10 codes for pathological fractures specific to neoplastic diseases (e.g., fractures due to bone metastasis; ICD-10 codes: M49.5, M84.5) were not considered in our operational definition of fracture. Follow-up for all participants continued from the index date until the occurrence of a fracture, death, emigration, or December 31, 2022, whichever occurred first. Covariates Age was presented both as a continuous variable (mean ± standard deviation) and as a categorical variable with three groups: 20–49 years, 50–69 years, and ≥ 70 years. Income levels were organized into four groups according to monthly insurance premiums. [10] Residence was classified as either rural or urban. Comorbidities, such as hypertension (HTN), diabetes mellitus (DM), dyslipidemia (DL), and end-stage renal disease (ESRD), were identified via corresponding ICD-10 codes, RID registration codes, or prescription claims. In the screening groups, data regarding smoking status, alcohol consumption, regular exercise, and body mass index (BMI) were collected. Smoking status was categorized into three groups: never-smokers, ex-smokers, and current smokers. Alcohol consumption was divided into three categories: none, mild–moderate drinkers, and heavy drinkers, according to the reference values of 30 g/day for males and 20 g/day for females. [20] Regular exercise was defined according to both exercise frequency and intensity. BMI was divided into five categories using the Asia-Pacific criteria: < 18.5, 18.5–23, 23–25, 25–30, ≥ 30 kg/m2. [21] Statistical analysis Descriptive statistics were performed using Student's t-test for analysis of continuous variables and the chi-square test for categorical variables. Competing risk analyses evaluated the risks of any fracture, vertebral fracture, and hip fracture with reporting of cause-specific hazard ratios (csHRs), according to cancer treatment modalities and the duration after surgery, while considering all-cause mortality as a competing risk. [22], [23] Additionally, this study assessed the risk of site-specific fractures based on postoperative duration: across overall follow-up, during the first year, and one year or more following surgery. One year was chosen as the threshold for subgroup analysis, informed by evidence from a UK cohort study indicating the peak risk of fracture occurred one to two years after LC diagnosis. [3] We adjusted for age, sex, income, place of residence, and comorbidities (HTN, DM, DL, and ESRD) as potential confounders. [24] In analyses of the screening groups, additional adjustments included BMI and health behavioral factors, specifically smoking status, alcohol consumption, and regular exercise. [24] To further explore potential modifiers, we performed stratified analyses by age and sex. All statistical analyses were performed using SAS version 9.4 (SAS Institute Inc., Cary, NC, USA), with a P-value <0.05 considered as the threshold for statistical significance. Ethical statement The study received approval from the Institutional Review Board (IRB) of Samsung Medical Center (IRB file No. SMC 2024–05-113), and the requirement for informed consent was waived as only de-identified data were used in accordance with confidentiality guidelines. Results Baseline characteristics of participants Among both LC survivors and matched controls, 64.1% were men and the mean age was 63.6 ± 10.1 years. LC survivors had higher proportions of individuals with high income, urban residency, and comorbidities (HTN, DM, or DL) compared to matched controls. The majority of LC survivors received surgery only (64.0%), while the remainder underwent surgery with ChT (22.3%), surgery with RT (3.0%), or surgery with CRT (10.6%) as treatment for LC (Table 1). Table 1Baseline characteristics of study participants (all participants).CharacteristicsControl(n = 99,876)Lung cancer survivors(n = 33,292)P valueSex, n (%)1.000Men63,996 (64.1)21,332 (64.1)Women35,880 (35.9)11,960 (35.9)Age in years (mean ± SD)63.6 ± 10.163.6 ± 10.11.000Age group in years, n (%)1.00020–498490 (8.5)2830 (8.5)50–6960,420 (60.5)20,140 (60.5)≥7030,966 (31.0)10,322 (31.0)Income level, n (%)<0.001Q1 (lowest 25%)22,556 (22.6)6677 (20.1)Q217,312 (17.3)5231 (15.7)Q323,504 (23.5)7709 (23.2)Q4 (highest 25%)36,504 (36.6)13,675 (41.1)Place of residence, n (%)0.002Rural54,998 (55.1)18,008 (54.1)Urban44,878 (44.9)15,284 (45.9)Comorbidities, n (%)Hypertensiona40,528 (40.6)15,204 (45.7)<0.001Diabetes mellitusb15,089 (15.1)6031 (18.1)<0.001Dyslipidemiac24,381 (24.4)10,554 (31.7)<0.001End-stage renal diseased359 (0.4)139 (0.4)0.133Cancer treatment modalities, n (%)Surgery21,316 (64.0)Surgery + chemotherapy7433 (22.3)Surgery + radiation therapy1002 (3.0)Surgery + chemoradiotherapy3541 (10.6)Mean follow-up duration, years (mean ± SD)Any fracture7.7 ± 2.86.3 ± 3.4<0.001Vertebral fracture8.0 ± 2.76.5 ± 3.3<0.001Hip fracture8.0 ± 2.66.5 ± 3.3<0.001SD, standard deviation; Q, quartile. Baseline characteristics of study participants (all participants). SD, standard deviation; Q, quartile. Risk of fracture during the overall follow-up period Over a mean follow-up period of 6.3 ± 3.4 years in LC survivors and 7.7 ± 2.8 years in matched controls, there were 2527 and 7930 incident fractures, respectively. The cumulative incidence of fractures was displayed based on LC status and cancer treatment modalities (Fig. 2). Cumulative incidence functions for fracture considering all-cause mortality as a competing event (overall follow-up period). (a) Incidence of any fracture, (b) incidence of vertebral fracture, (c) incidence of hip fracture, (d) incidence of any fracture by treatment modality, (e) incidence of vertebral fracture by treatment modality, (f) incidence of hip fracture by treatment modality. Abbreviations: ChT, chemotherapy; CRT, chemoradiotherapy; F/U, follow-up; LC, lung cancer; N, number; RT, radiation therapy; S, surgery. LC survivors exhibited a higher risk of fracture than matched controls. The greatest risk was associated with vertebral fractures (csHR 1.44, 95% CI 1.33–1.56), followed by hip fractures (csHR 1.29, 95% CI 1.15–1.44), and then any fracture (csHR 1.19, 95% CI 1.13–1.24) (Table 2). Table 2Fracture risk by follow-up duration (all participants).Any fractureVertebral fractureHip fractureEventIRacsHRb(95% CI)EventIRacsHRb(95% CI)EventIRacsHRb(95% CI)Overall follow-up periodBy LC statusControl793010.251.00 (Ref.)22152.781.00 (Ref.)12921.611.00 (Ref.)LC252712.081.19 (1.13–1.24)8333.871.44 (1.33–1.56)4131.901.29 (1.15–1.44)By modality of cancer treatmentControl793010.251.00 (Ref.)22152.781.00 (Ref.)12921.611.00 (Ref.)S only174611.951.10 (1.04–1.16)5413.591.22 (1.11–1.34)2851.881.15 (1.01–1.31)S + ChT50411.621.32 (1.21–1.45)1804.051.86 (1.60–2.17)781.741.51 (1.20–1.90)S + RT7918.161.78 (1.43–2.23)306.702.32 (1.62–3.33)153.301.90 (1.14–3.16)S + CRT19812.821.73 (1.50–1.99)825.193.04 (2.44–3.80)352.192.62 (1.87–3.67)Within 1 year after surgeryBy LC statusControl8999.081.00 (Ref.)2202.221.00 (Ref.)1051.061.00 (Ref.)LC3179.981.11 (0.98–1.26)1273.991.81 (1.45–2.26)451.411.40 (0.98–1.98)By modality of cancer treatmentControl8999.081.00 (Ref.)2202.221.00 (Ref.)1051.061.00 (Ref.)S only21410.451.08 (0.93–1.25)823.991.63 (1.26–2.10)281.361.25 (0.82–1.90)S + ChT547.570.96 (0.73–1.26)223.081.72 (1.11–2.68)70.981.12 (0.52–2.42)S + RT1921.292.24 (1.42–3.53)77.823.04 (1.43–6.45)55.573.92 (1.59–9.63)S + CRT309.141.34 (0.93–1.92)164.873.37 (2.02–5.63)51.522.12 (0.86–5.22)1 year or more after surgeryBy LC statusControl703110.421.00 (Ref.)19702.841.00 (Ref.)11611.661.00 (Ref.)LC221012.451.20 (1.14–1.26)7013.841.39 (1.28, 1.52)3611.971.28 (1.14–1.45)By modality of cancer treatmentControl703110.421.00 (Ref.)19702.841.00 (Ref.)11611.661.00 (Ref.)S only153212.201.10 (1.04–1.16)4553.521.17 (1.06–1.30)2531.951.15 (1.01–1.32)S + ChT45012.421.39 (1.26–1.52)1574.231.89 (1.60–2.22)701.871.57 (1.23–2.00)S + RT6017.351.67 (1.30–2.16)236.452.20 (1.46–3.32)102.781.55 (0.84–2.88)S + CRT16813.821.82 (1.56–2.12)665.293.01 (2.36–3.86)282.222.56 (1.76–3.74)ChT, chemotherapy; CI, confidence interval; CRT, chemoradiotherapy; csHR, cause-specific hazard ratio; LC, lung cancer; IR, incidence rate; PY, person-years; Ref., reference; RT, radiation therapy; S, surgery.aIncidence rate reported per 1000 person-years.bAnalyses were adjusted for age, sex, income, place of residence, and comorbidities (hypertension, diabetes mellitus, dyslipidemia, end-stage renal disease). Fracture risk by follow-up duration (all participants). ChT, chemotherapy; CI, confidence interval; CRT, chemoradiotherapy; csHR, cause-specific hazard ratio; LC, lung cancer; IR, incidence rate; PY, person-years; Ref., reference; RT, radiation therapy; S, surgery. Incidence rate reported per 1000 person-years. Analyses were adjusted for age, sex, income, place of residence, and comorbidities (hypertension, diabetes mellitus, dyslipidemia, end-stage renal disease). The risk of fracture differed according to cancer treatment modality. For any fracture, the greatest risk was seen in LC survivors receiving surgery with RT (csHR 1.78, 95% CI 1.43–2.23), followed by surgery with CRT (csHR 1.73, 95% CI 1.50–1.99), surgery with ChT (csHR 1.32, 95% CI 1.21–1.45), and surgery only (csHR 1.10, 95% CI 1.04–1.16). For vertebral or hip fracture, the greatest risk was observed in LC survivors treated with surgery and CRT (csHR 3.04, 95% CI 2.44–3.80 for vertebral fracture; csHR 2.62, 95% CI 1.87–3.67 for hip fracture) (Table 2). Risk of fracture within 1 year and 1 year or more after surgery The risk of any fracture was elevated in LC survivors compared to non-cancer controls, although statistical significance was present only one year or more after surgery (csHR 1.20, 95% CI 1.14–1.26). By contrast, the risks of vertebral fracture (csHR 1.81, 95% CI 1.45–2.26) and hip fracture (csHR 1.40, 95% CI 0.98–1.98) were greater during the first year following surgery than in later periods (Table 2). During the first year following surgery, the risk of fracture was significantly elevated among LC survivors who underwent surgery with RT (csHR 2.24, 95% CI 1.42–3.53 for any fracture; csHR 3.04, 95% CI 1.43–6.45 for vertebral fracture; csHR 3.92, 95% CI 1.59–9.63 for hip fracture). Among individuals who underwent surgery with CRT, vertebral fracture risk (csHR 3.37, 95% CI 2.02–5.63) was significantly elevated, whereas the association with any fracture (csHR 1.34, 95% CI 0.93–1.92) and hip fracture (csHR 2.12, 95% CI 0.86–5.22) were not significant (Table 2). At one year or more post-surgery, the elevated risk of fracture persisted among LC survivors who had undergone surgery with RT (csHR 1.67, 95% CI 1.30–2.16 for any fracture; csHR 2.20, 95% CI 1.46–3.32 for vertebral fracture; csHR 1.55, 95% CI 0.84–2.88 for hip fracture) or CRT (csHR 1.82, 95% CI 1.56–2.12 for any fracture; csHR 3.01, 95% CI 2.36–3.86 for vertebral fracture; csHR 2.56, 95% CI 1.76–3.74 for hip fracture) (Table 2). Risk of fracture by stratified analyses Stratified analyses demonstrated that the elevated fracture risk in men was more pronounced for any fracture (csHR 1.29, 95% CI 1.21–1.38, n = 1159; p for interaction <0.001) and vertebral fracture (csHR 1.72, 95% CI 1.53–1.92, n = 427; p for interaction <0.001), whereas the sex difference was not significant for hip fracture (csHR 1.39, 95% CI 1.20–1.60, n = 249; p for interaction 0.116) (Fig. 3, Supplementary Table S3). Fracture risk among lung cancer survivors stratified by age and sex. (a) Overall follow-up period; (b) within 1 year after surgery; (c) 1 year or more after surgery. Abbreviations: CI, confidence interval. Analyses were adjusted for income, place of residence, and comorbidities (hypertension, diabetes mellitus, dyslipidemia, end-stage renal disease). Furthermore, among young LC survivors aged 20–49 years, the age interaction was significant only for vertebral fracture (csHR 4.91, 95% CI 2.53–9.53, n = 21; p for interaction <0.001), whereas the interaction was not significant for any fracture (csHR 1.46, 95% CI 1.14–1.86, n = 92; p for interaction 0.223) or hip fracture (csHR 1.33, 95% CI 0.35–5.02, n = 3; p for interaction 0.182) (Fig. 3, Supplementary Table S3). Risk of fracture in the screening groups The baseline characteristics of the screening groups are detailed in the Supplementary Table S4. In this subgroup analysis, LC survivors also exhibited an elevated risk of fracture compared with non-cancer participants over the entire follow-up period. The risk was greatest for hip fracture (csHR 1.44, 95% CI 1.24–1.67), followed by vertebral fracture (csHR 1.42, 95% CI 1.29–1.57), and any fracture (csHR 1.21, 95% CI 1.14–1.28). Regarding fracture risk by cancer treatment modality, for any fracture, the greatest risk was seen in LC survivors receiving surgery with CRT (csHR 1.83, 95% CI 1.55–2.15), followed by surgery with RT (csHR 1.48, 95% CI 1.08–2.01), surgery with ChT (csHR 1.38, 95% CI 1.24–1.53), and surgery only (csHR 1.12, 95% CI 1.05–1.19). For vertebral and hip fracture, the greatest risk was again observed in LC survivors treated with surgery and CRT (csHR 3.32, 95% CI 2.58–4.26 for vertebral fracture; csHR 2.90, 95% CI 1.90–4.43 for hip fracture) (Supplementary Table S5). During the first year following surgery, the risk of any fracture (csHR 1.15, 95% CI 0.98–1.35) did not differ significantly from that of non-cancer participants, although it was elevated one year or more after surgery (csHR 1.22, 95% CI 1.15–1.29). The risks of vertebral fracture (csHR 1.82, 95% CI 1.37–2.43) and hip fracture (csHR 1.83, 95% CI 1.11–3.02) were greater during the first year following surgery than in later periods (Supplementary Table S5). Within the first year post-surgery, the risk of fracture was significantly elevated among individuals who underwent surgery with CRT (csHR 1.67, 95% CI 1.12–2.48 for any fracture; csHR 4.04, 95% CI 2.30–7.11 for vertebral fracture; csHR 3.35, 95% CI 1.18–9.55 for hip fracture), whereas the risk among those who received surgery with ChT or RT was not statistically significant. At one year or more post-surgery, the elevated fracture risk persisted among LC survivors who had undergone surgery with CRT (csHR 1.86, 95% CI 1.55–2.23 for any fracture; csHR 3.21, 95% CI 2.42–4.25 for vertebral fracture; csHR 2.89, 95% CI 1.82–4.59 for hip fracture) or ChT (csHR 1.45, 95% CI 1.29–1.62 for any fracture; csHR 1.92, 95% CI 1.58–2.32 for vertebral fracture; csHR 1.76, 95% CI 1.30–2.38 for hip fracture). For surgery with RT, the risk was elevated for hip fracture (csHR 2.22, 95% CI 1.10–4.47), borderline for any fracture (csHR 1.41, 95% CI 1.00–2.00), and not significant for vertebral fracture (csHR 1.56, 95% CI 0.86–2.83) (Supplementary Table S5). Discussion We observed an elevated risk of any, vertebral, and hip fractures among LC survivors compared to non-cancer participants across the entire follow-up period. This risk was most pronounced for vertebral and hip fractures within the first postoperative year; notably, the elevated hip fracture risk in the overall cohort attained statistical significance only after adjustment for health behavioral characteristics and BMI, and subsequently declined. Beyond the first year, however, the risk of any fracture increased again. Additionally, during the overall follow-up period, the risk of any fracture was substantially higher among LC survivors who underwent surgery combined with ChT, RT or CRT, compared with controls. LC survivors who underwent curative-intent surgery may have been predisposed to a higher fracture risk through several contributing pathways. The extent of surgical resection and the presence of pre-existing pulmonary disease may contribute to extended hospital stays or reductions in physical activity, thereby increasing susceptibility to weight loss, sarcopenia, and falls. [7] A recent umbrella review reported that the prevalence of sarcopenia among LC survivors ranged between 42.8% and 45.0%. [25] Although we adjusted for age, sex, BMI, smoking status, alcohol consumption, and exercise habits, residual confounding by unmeasured variables (such as peri- or postoperative malnutrition or sarcopenia) cannot be excluded. Our study further demonstrated that patients receiving ChT, RT or CRT exhibited a higher fracture risk than non-cancer controls, a finding that may be consistent with the concept of cancer treatment-induced bone loss, although alternative explanations warrant consideration. Notably, the association between RT and vertebral fracture risk was not significant in the screening subgroup, which may in part reflect residual confounding. Several biological mechanisms have been proposed to link cancer treatment with bone loss. [26], [27], [28], [29] Direct irradiation of the vertebrae has been proposed to suppress osteogenesis and promote osteoclastogenesis, potentially increasing the risk of vertebral fracture. [30], [31] Hip fracture risk was also elevated in the combination treatment groups. This finding is unlikely to reflect direct irradiation, since the femoral neck lies outside the thoracic radiation field, and may instead point to systemic mechanisms. A prior Danish cohort study reported an increased fracture risk among cancer survivors who received ChT (HR 1.31, 95% CI, 1.09–1.57 at <5 years after cancer diagnosis; HR 1.22, 95% CI, 0.99–1.51 at 5 years after cancer diagnosis) than those who did not. [5] Several anti-cancer agents used in LC therapy (such as cyclophosphamide, ifosfamide, and 5-fluorouracil) have also been associated with reductions in bone mineral density (BMD), possibly through myelosuppression. [26], [27], [28], [32] Cisplatin-induced hypogonadism has similarly been implicated in bone loss among male LC survivors, [33] and glucocorticoids, frequently used to manage treatment-related side effects or to enhance the efficacy of anticancer therapy, may further disrupt bone metabolism by suppressing osteoblastic differentiation and stimulating osteoclast activity at high doses. [32] Nonetheless, these treatment-related biological mechanisms cannot fully account for the observed associations, and stage- or indication-related confounding must also be considered. The potential for overestimating the fracture risk attributable to ChT, RT, or CRT should therefore be acknowledged. This consideration is particularly important, as patients with advanced-stage cancer often undergo multimodal therapies and frequently exhibit weight loss, sarcopenia or worse performance status already at the time of LC diagnosis, which may independently elevate fracture risk regardless of subsequent treatment. [34] The risk of vertebral fracture was greatest within the first year following surgery. This pattern may relate to bone loss during active treatment, although differential ascertainment likely also contributes, as LC survivors undergo frequent postoperative computed tomography (CT) surveillance that routinely visualizes the thoracic spine, increasing the likelihood of incidental detection relative to non-cancer controls. [35], [36], [37] The pattern for hip fracture may differ: risk may rise in the early postoperative period, when reduced pulmonary function and diminished physical activity promote sarcopenia and frailty, and then decline after one year as these functions recover. A Korean prospective cohort study of 620 LC survivors found a marked decline in pulmonary function two weeks after surgery, followed by improvements at six months and one year. [38] Consistently, a non-randomized Korean trial documented reduced physical activity soon after surgery, which improved over time in both groups and more so among participants in a perioperative exercise program using a wearable device. [39] As functional capacity and physical activity recover over the first postoperative year, fracture risk at various sites may rise again through a greater likelihood of falls during more strenuous activity. In the present study, we identified interactions between age and sex that influenced fracture risk among LC survivors. The higher relative risk of vertebral fracture in men and those under 50 years of age may be attributed to the inherently low baseline fracture risk within these subgroups. [40] Additionally, the lower rates of regular BMD testing and anti-osteoporotic medication use among men or younger individuals, compared to women or older adults, may further contribute to these observations. However, several of these subgroup estimates were based on a small number of fracture events, resulting in wide confidence intervals, and such estimates should be interpreted with appropriate caution rather than emphasized as robust subgroup effects. Our findings highlight the importance of comprehensive management strategies, incorporating early diagnosis and proactive treatment for reduced bone health, with particular attention to LC survivors who received combination treatments. However, the 2013 National Comprehensive Cancer Network task force report primarily addressed bone health among breast and prostate cancer survivors treated with endocrine therapies. [41] While a recent clinical practice guideline proposed an algorithm for the use of bone-targeted therapy in the context of bone metastases and myeloma bone disease, [42] additional research on multimodal interventions, such as physical activity, nutritional support, or pharmacological treatment, is required to decrease fracture risk and enhance the quality of life in LC survivors. We have several limitations to address. First, we could not obtain data regarding cancer stage at diagnosis, histological subtype, recurrence, specific chemotherapeutic regimen, or cumulative doses and field of irradiation, because these were not available from the KNHIS database. Second, potential confounders could not be comprehensively adjusted for, including glucocorticoid and anti-osteoporotic drug use, which were unavailable in our dataset and may partly act as mediators or reflect confounding by indication rather than simple baseline confounders. Third, the incidence of fracture may be underestimated due to limitations inherent to claims data. Radiological confirmation of incident fractures was not feasible, as we could not access X-ray or CT scans. Moreover, the number of fracture events was limited in certain subgroups, and the corresponding estimates should therefore be interpreted with caution. Finally, our findings may not be generalizable to other ethnic groups, since this study was conducted only including the Korean population. Conclusion LC survivors experienced consistently elevated fracture risks at various sites compared to non-cancer controls, with especially high risks observed in the first year after surgery, although the elevated hip fracture risk attained significance only after adjustment for health behavioral factors and BMI. Fracture risk was higher among LC survivors treated with surgery combined with ChT, RT or CRT, although the association between RT and vertebral fracture risk was not significant in the screening subgroup. Early evaluation of BMD and active implementation of multimodal interventions may help reduce fracture burden and improve quality of life for LC survivors. Dislosures Seonghye Kim, Kyung-Do Han, Jinhyung Jung, Dong Woog Yoon, Jong Ho Cho, and Dong Wook Shin confirm that they have no conflicts of interest related to the materials or methods employed in this research or the results presented in this article. CRediT authorship contribution statement Seonghye Kim: Writing – review & editing, Writing – original draft, Visualization, Methodology, Investigation, Conceptualization. Kyung-Do Han: Writing – review & editing, Visualization, Software, Methodology, Formal analysis, Data curation, Conceptualization. Jinhyung Jung: Writing – review & editing, Visualization, Software, Formal analysis, Data curation. Dong Woog Yoon: Writing – review & editing. Jong Ho Cho: Writing – review & editing, Validation, Methodology, Investigation, Conceptualization. Dong Wook Shin: Writing – review & editing, Validation, Supervision, Project administration, Methodology, Investigation, Conceptualization. Ethics approval and consent to participate This study was reviewed and approved by the institutional review board (IRB) of Samsung Medical Center (IRB file no. SMC 2024–05-113). Only de-identified data governed by confidentiality regulations were utilized, and the requirement for informed consent was waived accordingly. Funding No funding was received from other organizations for this work. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Footnotes Supplementary data to this article can be found online at https://doi.org/10.1016/j.jbo.2026.100791. Contributor Information Jong Ho Cho, Email: jongho.cho@gmail.com. Dong Wook Shin, Email: dwshin.md@gmail.com. Supplementary data Supplementary material References Associated Data Supplementary Materials Supplementary material