Cost-Effectiveness Analysis of Denosumab in the Prevention of Skeletal-Related Events Among Patients With Breast Cancer With Bone Metastasis in India

Abstract

PURPOSE

Denosumab is clinically superior to zoledronic acid (ZA) for preventing and delaying time to first and subsequent skeletal-related events (SREs) among patients with breast cancer (BC) with bone metastases. We evaluated the cost and health benefits of denosumab and ZA (once every 4 weeks and once every 12 weeks) among four different molecular subtypes of BC with bone metastases in India.

MATERIALS AND METHODS

A Markov model was developed in Microsoft Excel to estimate lifetime health consequences and resulting costs among cohort of 1,000 patients with BC with bone metastasis, for three intervention scenarios, namely denosumab (once every 4 weeks), ZA (once every 4 weeks), and ZA (once every 12 weeks). The health outcomes were measured in terms of SREs averted and quality-adjusted life-years (QALYs) gained. The cost of each intervention scenario was measured using both the health system and the patient's perspectives. Indirect costs because of lost productivity were not included. The future costs and outcomes were discounted at the standard rate of 3%.

RESULTS

Over a lifetime, the incremental number of SREs averted with use of denosumab once every 4 weeks (compared with ZA once every 4 weeks and once every 12 weeks) among patients with luminal A, luminal B, human epidermal growth factor receptor 2–enriched, and triple negative breast cancer were estimated as 0.39, 0.26, 0.25, and 0.19, respectively. The number of QALYs lived were slightly higher in the denosumab arm (1.45-2.80) compared with ZA once every 4 weeks and once every 12 weeks arms (1.44-2.78). However, denosumab once every 4 weeks was not found to be a cost-effective alternative for either of the four molecular subtypes of breast cancer. ZA once every 12 weeks was found to be a cost-effective option with an average cost-effectiveness ratio ranging between ₹68,254 and ₹73,636.

CONCLUSION

ZA once every 12 weeks is the cost-effective treatment option for BC with bone metastases in India. The present study findings hold significance for standard treatment guidelines under India’s government-funded health insurance program.

ZA 12-weekly is the cost-effective treatment option compared with denosumab for patients with metastatic breast cancer in India.

INTRODUCTION

Breast cancer (BC) is the most common cancer among women in India, with more than 178,361 incident cases in 2020.³ Approximately 20%-25% of cases present with upfront metastatic disease, categorized into molecular subtypes: luminal A, luminal B, human epidermal growth factor receptor 2 (HER2)–enriched, and triple-negative breast cancer (TNBC).⁴ Advanced breast cancer eventually develops bone metastases, increasing the risk of skeletal-related events (SREs). SREs, which include pathologic fracture, hypercalcemia of malignancy, spinal cord compression, orthopedic surgical intervention, and palliative bone radiation, have a deleterious effect on patients' health-related quality of life (HRQoL) and higher medical costs.⁵

CONTEXT

• Key Objective

• How do the cost and health outcomes of denosumab (administered once every 4 weeks) and zoledronic acid (ZA) administered once every 4 weeks and once every 12 weeks differ among various molecular subtypes of breast cancer (BC) with bone metastases in India?

• Knowledge Generated

• Our study demonstrates that while denosumab offers marginal clinical advantages, particularly in averting skeletal-related events and enhancing quality-adjusted life-years, ZA once every 12 weeks is a cost-effective treatment option for managing skeletal related events among diverse molecular subtypes of patients with BC with bone metastases in India.

• Relevance

• Our study findings can be useful to inform payers regarding formulary inclusion in the existing health benefit packages under Ayushman Bharat Pradhan Mantri Jan Arogya Yojana, and further develop standard treatment guidelines for the treatment of BC with bone metastasis in India.

The introduction of bone-modifying agents (BMAs) has revolutionized the management of SREs in patients suffering from BC with bone metastases. Several studies show that BMAs reduce SREs, fractures, hypercalcemia, pain medication use, prevent osteoporotic fractures, and reduce the need for bone-directed radiation therapy and/or surgery.⁶ There are two primary classes of BMAs, namely bisphosphonates, for example, zoledronic acid (ZA), and RANK ligand inhibitor denosumab. The recommended dosage of denosumab is 120 mg given once every 4 weeks via a subcutaneous route, while ZA 4 mg is given once every 4 weeks or 12 weeks via an intravenous route, for a minimum duration of 2 years, which may be extended depending upon the clinical profile of the patient.^7,8 The BMAs are given along with antineoplastic therapy in patients with BC with bone metastases.

Recently published evidence from large randomized controlled trials suggests that denosumab is clinically superior to ZA for preventing and delaying time to first and subsequent SREs among patients with BC with bone metastases.^6,8-10 In addition, denosumab is administered subcutaneously and is safe for patients with deranged renal function tests. However, lack of significant differences in the pain control, HRQoL, overall survival (OS), and higher cost of denosumab may influence treatment choice in favor of ZA, for resource-limited settings.¹¹

Published evidence on cost-effectiveness indicates that denosumab may not be cost-effective compared with ZA without a patient access scheme.⁶ Although few studies found denosumab to be more cost-effective than ZA once every 4 weeks in prostate cancer, BC, and other solid tumors,^12,13 others have reported ZA to be a more cost-effective option than denosumab.^14-16

Considering the heterogeneity in existing evidence and significant variations in demographic, socioeconomic, and clinical profile of Indian patients, these findings cannot be generalized to India. Moreover, there are certain methodological limitations in the previously published evidence. First, none of these studies quantified the improvement in terms of lifetime gain in quality-adjusted life-years (QALYs). Second, no study has considered different molecular subtypes of BC, which is a major limitation, as the findings cannot be generalized to all BC subtypes owing to differences in disease burden, disease progression, OS, and HRQoL. To address these gaps, we evaluated the cost-effectiveness of denosumab once every 4 weeks and ZA (once every 4 weeks and once every 12 weeks) among four molecular subtypes of patients with BC with bone metastases, over the lifetime horizon.

MATERIALS AND METHODS

Evaluation Approach

A Markov model was developed in Microsoft Excel to estimate lifetime health consequences and resulting costs among a cohort of 1,000 treatment-naive patients with BC with bone metastasis, for three intervention scenarios, namely denosumab (once every 4 weeks), ZA (once every 4 weeks), and ZA (once every 12 weeks). The cost-effectiveness analyses were undertaken for four BC subtypes, that is, patients with luminal A, luminal B, HER2-enriched, and TNBC in the Indian context.

The model consists of 12 mutually exclusive health states (Fig 1). The transition of health states in the model is detailed in Appendix. The model, with a monthly cycle length, starts at age 50 years, which is consistent with the median age of diagnosis of BC with bone metastases in India.¹⁷ The health outcomes were measured in terms of SREs averted, life-years gained, and QALYs gained. The cost of each intervention scenario was measured over the lifetime using a disaggregated societal perspective, which included health system costs and direct patient costs. Indirect costs because of lost productivity were not included in the present analysis. The future costs and outcomes were discounted at the standard rate of 3% per year as recommended by the Indian reference case, developed by Health Technology Assessment in India.¹⁸

FIG 1.

Model structure of relevant health states. SRE, skeletal-related event.

Intervention and Comparator

Denosumab once every 4 weeks was considered as an intervention arm, while ZA once every 4 weeks and ZA once every 12 weeks were considered as comparators. The recommended dosage of denosumab is 120 mg once every 4 weeks via subcutaneous route. Similarly, ZA is administered in a dose of 4 mg once every 4 weeks or once every 12 weeks via intravenous route. The treatment in either arm is given for a maximum duration of 2 years.⁷

Valuation of Consequences

The probability of occurrence of first SRE (at different time points) among treatment-naïve patients with BC with bone metastases on denosumab once every 4 weeks and ZA once every 4 weeks was obtained using Kaplan-Meier (KM) curves given in a trial by Stopeck et al.⁹ The KM curves were digitized using Engauge software.¹⁹ Parametric curves were fitted assuming the exponential, weibull, log-logistic, log-normal, generalized gamma, and gompertz distributions. The best-fitting distribution was chosen on the basis of statistical information criteria, visual inspection of the curve, and clinical plausibility.²⁰ Transition probabilities from first SRE to subsequent SREs for denosumab and ZA once every 4 weeks were estimated directly using average monthly incidence rates given in published literature,⁹ because of lack of availability of KM curves for subsequent SREs. Additionally, probabilities of first and subsequent SREs for patients with BC with bone metastases on ZA once every 4 weeks and once every 12 weeks were assumed to be the same, on the basis of a statistically insignificant difference in incidence rates between the two treatment arms.⁸

BC–specific mortality rates for all molecular subtypes were derived from published Indian literature.²¹ Furthermore, the proportion of patients having various fractures such as lower-limb, upper-limb, and rib fractures was obtained using expert opinion. The life tables of the Registrar General of India were used to obtain age-specific all-cause mortality rates.²² Published evidence in the form of rates were converted to monthly transition probabilities using standard methods of conversion.²³ Table 1 summarizes the input parameters for estimating the effectiveness of BMAs along with its data sources.

TABLE 1.

Monthly Input Parameters Used in Estimating the Effectiveness of Treatment Arms Included in the Model

Parameters		Transition Probabilities			Source
SRE	Treatment Arms	Base Value	Lower Limit	Upper Limit
First pathologic fracture	Denosumab (120 mg once every 4 weeks)	0.1831	0.1648	0.2014	Stopeck et al9
	ZA (4 mg once every 4 weeks)	0.0206	0.0185	0.0226	Stopeck et al9
	ZA (4 mg once every 12 weeks)	0.0206	0.0185	0.0226	Assumption based on Himelstein et al8
First bone radiation	Denosumab (120 mg once every 4 weeks)	0.0066	0.0060	0.0073	Stopeck et al9
	ZA (4 mg once every 4 weeks)	0.0097	0.0087	0.0107	Stopeck et al9
	ZA (4 mg once every 12 weeks)	0.0097	0.0087	0.0107	Assumption based on Himelstein et al8
First spinal cord compression	Denosumab (120 mg once every 4 weeks)	0.0006	0.0006	0.0007	Stopeck et al9
	ZA (4 mg once every 4 weeks)	0.0005	0.0004	0.0005	Stopeck et al9
	ZA (4 mg once every 12 weeks)	0.0005	0.0004	0.0005	Assumption based on Himelstein et al8
First bone surgery	Denosumab (120 mg once every 4 weeks)	0.0009	0.0008	0.0010	Stopeck et al9
	ZA (4 mg once every 4 weeks)	0.0006	0.0005	0.0006	Stopeck et al9
	ZA (4 mg once every 12 weeks)	0.0006	0.0005	0.0006	Assumption based on Himelstein et al8
Subsequent pathologic fracture	Denosumab (120 mg once every 4 weeks)	0.0473	0.03784	0.05676	Stopeck et al9
	ZA (4 mg once every 4 weeks)	0.0500	0.0015	0.0025
	ZA (4 mg once every 12 weeks)	0.0500	0.0015	0.0025	Assumption based on Himelstein et al8
Subsequent bone radiation	Denosumab (120 mg once every 4 weeks)	0.0183	0.01464	0.02196	Stopeck et al9
	ZA (4 mg once every 4 weeks)	0.025	0.02	0.03
	ZA (4 mg once every 12 weeks)	0.025	0.02	0.03	Assumption based on Himelstein et al8
Subsequent spinal cord compression	Denosumab (120 mg once every 4 weeks)	0.0002	0.0016	0.0024	Stopeck et al9
	ZA (4 mg once every 4 weeks)	0.0015	0.0012	0.0018
	ZA (4 mg once every 12 weeks)	0.0015	0.0012	0.0018	Assumption based on Himelstein et al8
Subsequent bone surgery	Denosumab (120 mg once every 4 weeks)	0.0027	0.00216	0.00324	Stopeck et al9
	ZA (4 mg once every 4 weeks)	0.0017	0.00136	0.00204
	ZA (4 mg once every 12 weeks)	0.0017	0.00136	0.00204	Assumption based on Himelstein et al8
Disease-specific mortality	Luminal A	0.164694	0.1152861	0.21410	Gogia et al21
	Luminal B	0.025816	0.018071	0.0335
	HER2-enriched	0.037776	0.026443	0.04910
	TNBC	0.028740	0.020118	0.03736
All-cause mortality	50-54 years	0.000720	0.000504	0.000936	Sample Registration System 201722
	55-59 years	0.001055	0.000739	0.001372
	60-64 years	0.001634	0.001144	0.002124
	65-69 years	0.003882	0.002717	0.005046

Abbreviations: HER2, human epidermal growth factor receptor 2; SRE, skeletal-related event; TNBC, triple-negative breast cancer; ZA, zoledronic acid.

Estimation of HRQoL

To assess HRQoL among patients with BC with bone metastases in each health state of the model, mean HRQoL scores for both nonhospitalized and hospitalized patients with BC with bone metastases were obtained from the National Cancer Database for Cost and Quality of Life (CaDCQol).¹⁸ These scores were derived using Indian tariff values.²⁴ The HRQoL scores for different types of SREs considered in the model were estimated using these base mean HRQoL scores among patients with BC with bone metastases and disutility values reported by Matza et al,²⁵ for all types of SREs, that is, pathologic fracture (lower limb, rib, upper limb, spinal cord compression [with and without paralysis]), radiation therapy, and bone surgery (Table 2).

TABLE 2.

Skeletal-Related Event–Specific Mean Utility Scores Along With 95% CIs

Utility Values
Parameter	Disutility Score	Estimated Utility Score (hospitalized cases)	Estimated Utility Score (nonhospitalized cases)
Base value (MBC with no SRE)18	—	0.634 (0.6 to 0.669)	0.716 (0.705 to 0.728)
Spinal cord compression without paralysis24	–0.22	0.414 (0.3312 to 0.4968)	0.496 (0.3968 to 0.5952)
Spinal cord compression with paralysis24	–0.32	0.314 (0.2512 to 0.3768)	0.396 (0.3168 to 0.4752)
Fracture of the lower limb24	–0.06	0.574 (0.4592 to 0.6888)	0.656 (0.5248 to 0.7872)
Fracture of the rib24	–0.03	0.604 (0.4832 to 0.7248)	0.686 (0.5488 to 0.8232)
Fracture of the upper limb24	–0.04	0.594 (0.4752 to 0.7128)	0.676 (0.5408 to 0.8112)
Radiation therapy24	–0.06	0.574 (0.4592 to 0.6888)	0.656 (0.5248 to 0.7872)
Bone surgery24	–0.07	0.564 (0.4512 to 0.6768)	0.646 (0.5168 to 0.7752)

Abbreviations: MBC, metastatic breast cancer; SRE, skeletal-related event.

Valuation of Costs

To estimate the cost of management of SREs, the provider payment rates under India's national health insurance program—Ayushman Bharat Pradhan Mantri Jan Arogya Yojana (AB PM-JAY)—were used. These rates cover the cost of prehospitalization diagnostic workup, any hospitalization cost, procedure cost, and posthospitalization medicine cost up to 15 days of discharge. Additionally, direct nonmedical expenditure inclusive of traveling, boarding, lodging, food, informal payments, and so on were derived from CaDCQol Indian database.¹⁸ Reimbursement rates and direct nonmedical expenditure were included in all the three treatment arms, that is, denosumab once every 4 weeks, ZA once every 4 weeks, and ZA once every 12 weeks with respect to all four subtypes of BC with bone metastases. The treatment protocol for the management of each SRE used in computation of end results was derived from clinical guidelines²⁶ and expert opinion (Table 3).

TABLE 3.

Model Input Cost Parameters

Input Cost Estimates	Unit Cost (₹)			Source
	Base Value	Lower Limit	Upper Limit
Package cost for management of SRE subtypes
Cost per lower-limb fracture	11,000	5,500	16,500	Package costs under AB PM-JAY27,a
Cost per upper-limb fracture	11,000	5,500	16,500
Cost per rib fracture	11,000	5,500	16,500
Cost per spinal cord compression (with or without paralysis)	40,000	20,000	60,000
Cost per bone surgery	14,900	7,450	22,350
Cost per radiotherapy (five fractions)	11,000	5,500	16,500
Cost per radiotherapy (10 fractions)	10,000	5,000	15,000
Direct nonmedical out-of-pocket expenditure
Direct nonmedical (OPD)	2,116.44	1,058.22	3,174.66	CaDCQol18
Direct nonmedical (hospitalization)	3,934.24	1,967.12	5,901.36
Package costs for bone-targeting drugs
Denosumab 120 mg (once every 4 weeks)	18,000	9,000	27,000	Package costs under AB PM-JAY27
ZA 4 mg (once every 4 weeks)	3,500	1,750	5,250
ZA 4 mg (once every 12 weeks)	1,166.67	583.33	1,750
Miscellaneous costs
Cost per OPG	202	101	103	CGHS28
Cost per calcium 500 mg and vitamin D tablet	0.196	0.098	0.294	RMSC29

Abbreviations: AB PM-JAY, Ayushman Bharat Jan Arogya Yojana; CaDCQol, National Cancer Database for Cost and Quality of Life; CGHS, Central Government Health Scheme; OPD, outpatient department; OPG, orthopantomography; RMSC, Rajasthan Medical Services Corporation; SRE, skeletal-related event; ZA, zoledronic acid.

^aPackage costs inclusive of cost of drug administration, outpatient consultation, diagnostics, and routine follow-up.

Sensitivity Analysis

A probabilistic sensitivity analysis was conducted to test parameter uncertainty.³⁰ We used gamma distribution for cost parameters and beta distribution for effectiveness parameters and utility scores. For the remaining parameters, we used uniform distribution to simulate random values. Upper and lower bounds were computed using standard error provided in the published literature or calculated using primary data. Wherever upper and lower bounds were not available, a 20% variation on either side of the base estimate was assumed for transition probabilities, 30% for the risk of mortality, and 50% for cost parameters. We used Monte Carlo method for simulating the results, and the number of iterations were restricted to 10,000 times. The median was computed along with the 2.5th and 97.5th percentiles to estimate the 95% CI. The per capita gross domestic product (GDP) of India of ₹172,000 ($2,075 in US dollars [USD]) for the year 2022-2023 was used to compare incremental cost-effectiveness ratios (ICERs) to make recommendations about cost-effectiveness.³¹

RESULTS

Costs

Over the lifetime, a patient with luminal A BC with bone metastases was found to incur a societal cost of ₹523,051, ₹205,005, and ₹1,038,746 with ZA once every 4 weeks, ZA once every 12 weeks, and denosumab once every 4 weeks, respectively. Similarly, the lifetime costs incurred by a patient with luminal B, HER2-enriched, and TNBC with bone metastases with ZA once every 4 weeks, ZA once every 12 weeks, and denosumab once every 4 weeks, respectively, are reported in Tables 4 and 5.

TABLE 4.

Incremental Cost per SRE Averted for a Patient With BC With Bone Metastases Receiving ZA (once every 4 weeks and once every 12 weeks) and Denosumab (once every 4 weeks) in India

Breast Cancer Subtype	Intervention Strategy	Lifetime Cost (₹)	Mean SREs	Incremental Cost per SRE Averted (₹)	Status	ACER per SRE Averted (₹)	Status
Luminal A	ZA 4 mg (once every 12 weeks)	205,005 (131,770-314,728)	2.78 (2.05-3.62)	—	ND	73,528	Lowest cost per SRE averted
	Denosumab 120 mg (once every 4 weeks)	1,038,746 (557,131-1,768,332)	2.39 (1.79-3.14)	2,141,223	ND	—	—
	ZA 4 mg (once every 4 weeks)	523,051 (330,768-820,010)	2.78 (2.05-3.62)	—	D	—	—
Luminal B	ZA 4 mg (once every 12 weeks)	139,259 (88,905-223,041)	1.86 (1.36-2.48)	—	ND	74,513	Lowest cost per SRE averted
	Denosumab 120 mg (once every 4 weeks)	715,897 (379,238-1,271,565)	1.60 (1.15-2.17)	2,205,877	ND	—	—
	ZA 4 mg (once every 4 weeks)	367,404 (214,538-563,570)	1.86 (1.36-2.48)	—	D	—	—
HER2-enriched	ZA 4 mg (once every 12 weeks)	129,621 (79,478-203,780)	1.72 (1.20-2.41)	—	ND	75,110	Lowest cost per SRE averted
	Denosumab 120 mg (once every 4 weeks)	651,752 (349,736-1,138,341)	1.47 (1.03-2.01)	2,088,909	ND	—	—
	ZA 4 mg (once every 4 weeks)	335,648 (198,106-543,391)	1.72 (1.20-2.41)	—	D	—	—
TNBC	ZA 4 mg (once every 12 weeks)	98,551 (61,180-164,175)	1.29 (0.92-1.85)	—	ND	76,269	Lowest cost per SRE averted
	Denosumab 120 mg (once every 4 weeks)	491,409 (259,189-918,092)	1.09 (0.76-1.56)	2,020,323	ND	—	—
	ZA 4 mg (once every 4 weeks)	254,771 (152,181-427,067)	1.29 (0.92-1.85)	—	D	—	—

Abbreviations: ACER, average cost-effectiveness ratio; D, dominated; HER2, human epidermal growth factor receptor 2; MBC, metastatic breast cancer; ND, nondominated; SRE, skeletal-related event; TNBC, triple-negative breast cancer; ZA, zoledronic acid.

TABLE 5.

Incremental Cost per QALY Gained by a Patient With BC With Bone Metastases Receiving ZA (once every 4 weeks and once every 12 weeks) and Denosumab (once every 4 weeks) in India

Breast Cancer Subtype	Intervention Strategy	Lifetime Cost (in ₹) Along With 95% CI	QALY	ICER per QALY Gained	Status	ACER per QALY Gained (₹)	Status
Luminal A	ZA 4 mg (once every 12 weeks)	205,005 (131,770-314,728)	2.78 (2.17-3.50)		ND	73,636	Cost-effective
	Denosumab 120 mg (once every 4 weeks)	1,038,746 (557,131-1,768,332)	2.80 (2.18-3.53)	42,221,466	ND	—	Not cost-effective
	ZA 4 mg (once every 4 weeks)	523,051 (330,768-820,010)	2.78 (2.17-3.50)		D	—	—
Luminal B	ZA 4 mg (once every 12 weeks)	139,259 (88,905-223,041)	1.98 (1.50-2.55)		ND	70,294	Cost-effective
	Denosumab 120 mg (once every 4 weeks)	715,897 (379,238-1,271,565)	1.99 (1.50-2.59)	30,891,364	ND	—	Not cost-effective
	ZA 4 mg (once every 4 weeks)	367,404 (214,538-563,570)	1.98 (1.50-2.55)	—	D	—	—
HER2-enriched	ZA 4 mg (once every 12 weeks)	129,621 (79,478-203,780)	1.83 (1.37-2.45)		ND	70,694	Cost-effective
	Denosumab 120 mg (once every 4 weeks)	651,752 (349,736-1,138,341)	1.84 (1.37-2.47)	45,442,097	ND	—	Not cost-effective
	ZA 4 mg (once every 4 weeks)	335,648 (198,106-543,391)	1.83 (1.37-2.45)		D	—	—
TNBC	ZA 4 mg (once every 12 weeks)	98,551 (61,180-164,175)	1.44 (1.06-1.96)		ND	68,254	Cost-effective
	Denosumab 120 mg (once every 4 weeks)	491,409 (259,189-918,092)	1.45 (0.76-1.56)	44,791,297	ND	—	Not cost-effective
	ZA 4 mg (once every 4 weeks)	254,771 (152,181-427,067)	1.44 (1.06-1.96)		D	—	—

Abbreviations: ACER, average cost-effectiveness ratio; D, dominated; HER2, human epidermal growth factor receptor 2; ICER, incremental cost-effectiveness ratio; MBC, metastatic breast cancer; ND, nondominated; QALY, quality-adjusted life-year; SRE, skeletal-related event; TNBC, triple-negative breast cancer; ZA, zoledronic acid.

Health Outcomes

The number of SREs were found to be less in the denosumab arm (2.39) compared with ZA once every 4 weeks and once every 12 weeks (2.78) among patients with luminal A BC with bone metastases. Similarly, in case of luminal B and HER2-neu patients receiving ZA (once every 4 weeks and once every 12 weeks), the mean number of SREs was estimated as 1.86 (compared with 1.60 in the denosumab arm) and 1.72 (compared with 1.47 in the denosumab arm), respectively. The patients with TNBC experienced least number of SREs (1.29 in ZA and 1.09 in the denosumab arm).

Furthermore, it was found that the number of QALYs lived per patient with BC with bone metastases were slightly higher in the denosumab arm (ranging between 1.45 and 2.80) compared with ZA once every 4 weeks and once every 12 weeks arms (ranging between 1.44 and 2.78). The number of QALYs lived (denosumab v ZA) were found to be highest among luminal A BC with bone metastases cases (2.80 v 2.78), followed by luminal B (1.99 v 1.98), HER2-enriched (1.84 v 1.83), and TNBC (1.45 v 1.44; Table 5).

Cost-Effectiveness

Among the three interventions, denosumab once every 4 weeks and ZA once every 12 weeks were nondominated interventions. However, ZA once every 4 weeks was dominated by ZA once every 12 weeks in all four molecular subtypes. The incremental cost per QALY gained by a patient with BC with bone metastases receiving denosumab once every 4 weeks varied from ₹30,891,364 to ₹45,442,097 relative to ZA once every 12 weeks, which is not cost-effective at the current willingness to pay threshold of one-time per capita GDP of India (₹172,000 or 2,075 USD). Therefore, ZA once every 12 weeks was found to be a cost-effective alternative, with an average cost-effectiveness ratio (average cost-effectiveness ratio [ACER] per QALY gained) varying from ₹68,254 to ₹73,636, which is 39.4%-42.5% of per capita GDP of India (Table 5).

Cost per SRE Averted

The incremental cost per SRE averted in the denosumab once every 4 weeks treatment arm relative to ZA once every 12 weeks among different molecular subtypes of BC with bone metastases ranged from ₹2,020,323 to ₹2,205,877, while the average cost per SRE averted in ZA once every 12 weeks lies in the range of ₹73,528-₹76,269 (Table 4).

DISCUSSION

Bone metastases affect 60%-80% of BC cases, posing a higher risk for young Indian patients because of early age of onset.^32,33 The treatment goal in BC with bone metastases is to enhance quality of life with minimal toxicity.³⁴ This becomes even more important for patients with bone metastases in view of their prolonged survival in the setting of an incurable disease.³⁵ Systemic therapy is the mainstay of treatment with supportive drugs for symptomatic relief and palliation.

Three FDA-approved agents—pamidronate, ZA, and denosumab—aim to prevent SREs in BC with bone metastases. Previous evidence has shown that ZA was more cost-effective than pamidronate.^36,37 Denosumab, a newer agent, outperforms ZA in preventing or delaying SREs, offering renal safety and the convenience of subcutaneous administration. Our analysis also showed that denosumab once every 4 weeks has the lowest SRE incidence (1.09-2.39) compared with ZA once every 12 or once every 4 weeks (1.29-2.78). SREs were most prominent in the luminal A subgroup, followed by luminal B, aligning with the higher incidence of bone metastases in hormone-positive disease.^35,38

Given that the drug costs for denosumab are substantially higher than ZA, wide adoption of this new treatment should be contingent upon evidence on both clinical effectiveness and cost-effectiveness.³⁹ We found that ZA once every 12 weeks is a cost-effective option with ACER ranging between ₹68,254 and ₹73,636 compared with denosumab. The use of ZA once every 12 weeks resulted in the highest lifetime cost in the luminal A subgroup, which is attributed to the higher incidence of bone metastases and prolonged survival observed in this subgroup.^35,40

Traditionally, bisphosphonates were administered once every 4 weeks to align with chemotherapy schedules. Although systemic reviews support the noninferiority of ZA once every 12 weeks compared with ZA once every 4 weeks, robust data for de-escalating denosumab to once every 12 weeks are not as conclusive. Trials for metastatic breast and prostate cancer suggest that de-escalating denosumab to once every 12 weeks is a reasonable option.^41,42 Given the significant burden of managing BC with bone metastases, adopting ZA once every 12 weeks improves patient convenience and compliance, and reduces the financial strain on the health system. To optimize the health benefits of denosumab, particularly its superiority in decreasing SREs at a reduced economic burden, country-specific data and trials for de-escalating denosumab to once every 12 weeks are needed. To the best of our knowledge, the present analysis is the first study to estimate the incremental cost per QALY gained with use of denosumab once every 4 weeks compared with ZA once every 4 weeks and once every 12 weeks among patients with BC with bone metastases belonging to different molecular subtypes that is, luminal A, luminal B, TNBC, and HER2-enriched.

The findings from our model align with existing clinical and epidemiologic evidence for both denosumab and ZA treatment arms. Our estimated median OS for luminal A, luminal B, HER2-enriched, and TNBC (41, 27, 26, and 19 months, respectively) is consistent with published estimates.^21,43,44 Studies by Khadkban et al⁴³ and Parikh et al⁴⁴ reported OS ranging from 21.2 to 31.4 months and 9 to 22 months, respectively, while Gogia et al²¹ reported separate survival rates with respect to different types of molecular subtypes, ranging from 17.5 (95% CI, 11 to 19.2) months to 41.4 (95% CI, 30 to 52.8). Our results align closely with this real-world evidence in the Indian context (Fig 2).

FIG 2.

Comparison of median overall survival (in months) with published evidence.

Additionally, we found that the denosumab once every 4 weeks arm has a lower mean number of SREs at 24 months (ranging between 0.58 and 0.74) compared with ZA once every 4 weeks and ZA once every 12 weeks treatment arms (ranging between 0.69 and 0.89). Previous studies showed a relative difference of 16% and 18% in the mean number of SREs between denosumab (once every 4 weeks) and ZA (once every 4 weeks),⁹ and denosumab and ZA once every 12 weeks,⁵ respectively, which aligns with our estimated relative difference ranging from 16.01% to 18.26% between these treatment arms.

Our study estimated lower QALYs in individual treatment arms compared with existing literature, likely because of the use of lower utility scores derived from Indian data.^12,13 However, the incremental gain in QALYs with denosumab aligns with published evidence. The mean QALYs gained at 12 months in the denosumab once every 4 weeks treatment arm ranged from 0.43 to 0.56 compared with ZA once every 4 weeks and ZA once every 12 weeks (ranging between 0.42 and 0.54). Previous studies also reported QALYs gained with denosumab ranging from 0.39 to 0.88,⁴⁵ compared with ZA once every 4 weeks and ZA once every 12 weeks treatment arms, ranging from 0.38 to 0.90.¹³

Considering the significant disease and economic burden that cancer imposes, several government-funded health insurance schemes have been implemented in India. The AB PM-JAY also includes several cancer health benefit packages (HBPs). Both ZA (once every 4 weeks) and denosumab (once every 4 weeks) are part of the PM-JAY package. Our study has important policy implications. First, from the point of view of reimbursement rates set up under HBP 2.0, denosumab is not cost-effective compared with ZA (once every 4 weeks/once every 12 weeks) because of its higher cost (2.5-5 times higher than ZA once every 4 weeks/once every 12 weeks) with limited incremental health benefits. Second, ZA once every 12 weeks is the most cost-effective option, with an ACER ranging from ₹69,444 to ₹75,578 depending on the breast cancer subtype, falling within the acceptable threshold of one-time per capita GDP. Therefore, clinical guidelines by the Indian Council of Medical Research and National Health Authority should consider recommending ZA once every 12 weeks as the primary treatment choice. Additionally, reconsidering the inclusion of denosumab under HBP in PM-JAY is suggested.

We would like to highlight a few strengths of our analysis. First, to our knowledge, this is the first study from India that has compared all three treatment interventions, namely ZA once every 4 weeks, ZA once every 12 weeks, and denosumab once every 4 weeks. Moreover, none of the previous studies compared the cost-effectiveness of denosumab and ZA in all four molecular subtypes of BC with bone metastases. Second, the data for determining costs were based on country-specific prices (rates listed by AB PM-JAY). Similarly, utility scores were obtained from a nationally representative study, thus further strengthening the generalizability of findings. Third, the lifetime horizon was used to account for health benefits and costs for a better understanding, while previous studies estimated costs and consequences for a period of 1 or 2 years. Fourth, we used month-wise transition probabilities for first SREs, which were obtained using KM curves given in published literature.⁸ Finally, we have used country-specific disease mortality rates with respect to each molecular subtype.²¹

We would like to mention some limitations of our study. First, indirect costs because of loss of productivity were not considered in the present analysis. This was in agreement with Indian Health Technology Assessment guidelines.¹⁸ Second, the probabilities of incidence of subsequent SREs were assumed to be same as that of first SREs because of lack of availability of SRE-specific incidence rates in the subsequent stage. Furthermore, we relied on global estimates for transition probabilities because of lack of country-specific estimates. Third, we did not assess bisphosphonate-associated nephrotoxicity and discontinuation. Fourth, our findings are specific to patients with BC with bone metastases, limiting generalization to other conditions such as bone metastases from other solid tumors, multiple myeloma, or osteoporosis. Finally, active oncology costs, including systemic therapy or local therapy in various subgroups, were not included because of significant cost variations, which could have altered lifetime costs and diluted the assessment of cost and efficacy related to BMAs.

In conclusion, we have found ZA once every 12 weeks to be a cost-effective alternative among all four molecular subtypes of BC with bone metastases. These findings can be used to inform payers regarding formulary inclusion in the existing HBPs under PM-JAY and further develop standard treatment guidelines for the treatment of BC with bone metastasis in India.

APPENDIX. TRANSITION OF HEALTH STATES

The model starts with a hypothetical cohort of 1,000 treatment-naïve women age 50 years with breast cancer (BC) with bone metastases. This is consistent with the median age of diagnosis of BC with bone metastases in India. All individuals in the cohort enter the model in no skeletal-related event (SRE) health state. It was assumed that the individuals in a no SRE health state may develop either of the four types of first SRE (pathologic fracture, spinal cord compression, radiation to bone, or bone surgery) or die due to natural or disease-related causes or remain in the no SRE health state. Furthermore, it was assumed that there is a probability of moving from each of these first SREs to subsequent SREs irrespective of the type of first SRE that had occurred previously among these patients.

The patients in the first SRE health state were further assumed to either recover from the existing health state and move to no SRE with history of (h/o) SRE or die due to both all-cause and disease-specific causes. The individuals in any of the subsequent SRE health states were assumed to either recover and move to no SRE with h/o SRE health state or may die due to natural and disease-specific causes. It was also assumed that there was no probability of remaining in the same health state for patients in the first SRE and subsequent SRE health states. The patients in the no SRE with h/o SRE health state were assumed to have a probability to either develop subsequent SREs or die due to natural and disease-specific causes or may also remain in the same health state. The disease-specific and all-cause mortality were absorbing health states in the model (Fig 1).

DATA SHARING STATEMENT

The data sets and analysis will be available upon request. The study investigators retain ownership of their data. Any requests for access to data should be made directly to study investigator.

AUTHOR CONTRIBUTIONS

Conception and design: Raina Wadhwa, Nidhi Gupta, Jyoti Dixit, Pankaj Malhotra, Shankar Prinja

Administrative support: Nidhi Gupta, Jyoti Dixit, Shankar Prinja

Provision of study materials or patients: Nidhi Gupta, Jyoti Dixit

Collection and assembly of data: Raina Wadhwa, Nidhi Gupta, Jyoti Dixit, Pankaj Malhotra, PVM Lakshmi, Shankar Prinja

Data analysis and interpretation: Raina Wadhwa, Nidhi Gupta, Jyoti Dixit, Pankaj Malhotra, PVM Lakshmi, Shankar Prinja

Manuscript writing: All authors

Final approval of manuscript: All authors

Accountable for all aspects of the work: All authors

AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated unless otherwise noted. Relationships are self-held unless noted. I = Immediate Family Member, Inst = My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about ASCO's conflict of interest policy, please refer to www.asco.org/rwc or ascopubs.org/go/authors/author-center.

Open Payments is a public database containing information reported by companies about payments made to US-licensed physicians (Open Payments).

No potential conflicts of interest were reported.