Treatment patterns of prostate cancer with bone metastasis in Beijing: A real‐world study using data from an administrative claims database

Abstract Purpose To explore treatment patterns among patients with prostate cancer and bone metastasis and to compare clinical outcomes following use of different hormone therapies including combined androgen blockade (CAB), nonsteroidal antiandrogen (NSAA) monotherapy, and castration monotherapy. Methods We conducted a population‐based cohort study using data from the Urban Employee Basic Medical Insurance database (2011‐2014) in Beijing. We identified 475 patients with newly diagnosed bone metastatic prostate cancer with at least one prescription for hormone therapy and described their treatment patterns over a median follow‐up of 20.7 months. Cox proportional hazards model was used to compare time to chemotherapy initiation between patients starting on different hormone therapies. Results Hormone therapy and/or bisphosphonate therapy with zoledronic acid were the initial treatments in the majority of patients (87.8%); chemotherapy, radiotherapy, and surgery were usually given later in the treatment pathway. CAB was the most common hormone treatment (73.7%). For time to chemotherapy initiation, hazard ratios (95% confidence intervals) were 2.43 (1.08‐5.44) for NSAA alone vs CAB and 1.29 (0.78‐2.13) for castration alone vs CAB. Conclusions Our findings show that while a wide range of therapies are used to treat patients with prostate cancer and bone metastasis in Beijing, hormone therapy and bisphosphonate therapy are the most commonly prescribed, and use of CAB was seen to be advantageous in delaying time to chemotherapy initiation over NSAA monotherapy. Future studies should explore longer‐term treatment patterns, including use of newly approved treatments.


KEYWORDS
bone metastasis, claims database, combined androgen blockade, drug utilization, pharmacoepidemiology, prostate cancer, treatment pattern

| INTRODUCTION
China has seen a rapid increase in the incidence and prevalence 1 of prostate cancer in recent decades (ranking sixth in all male cancers), which is likely due to the ageing population and changes in diagnostic practices. 2,3 In the absence of early diagnosis and primary treatment, prostate cancer commonly progresses to an advanced stage, frequently with bone metastases that have a profound impact on patients' quality of life and place a great burden on healthcare resources. 4 We have previously shown that bone metastasis in prostate cancer patients in China are common (33.2% in 2011), and higher than in other countries, and related treatment costs have increased. 1 Continuous androgen deprivation therapy (ADT) as the main hormone therapy is the standard of care for patients with metastatic prostate cancer, providing symptom relief and delaying disease progression. 5 However, ADT can lead to loss of bone mineral density and an increased risk of osteoporotic fragility fractures. The combined effects of ADT and bone metastasis can result in skeletal complications such as pain, ineffective haematopoiesis, and skeletalrelated events (SREs). 6 Thus, the management of patients with bone metastatic prostate cancer often requires use of bone-targeted agents (BTAs) such as bisphosphonates, denosumab, and radium-233, 4,7-9 as well as palliative radiation, bone surgery, and pain relief. After initial ADT treatment, most patients eventually become unresponsive to castration 10 -metastatic castration-resistant prostate cancer (mCRPC). First-line treatment for this patient population has been docetaxel-based chemotherapy although a few newer agents have been approved in recent years. 11,12 There are three main types of hormone therapy for patients with prostate cancer. 13 The first type is antiandrogen alone, usually a nonsteroidal antiandrogen (NSAA) such as bicalutamide, flutamide, and nilutamide. The second type is castration alone, including luteinizing hormone-releasing hormone (LHRH) agonists or antagonists (ie, medical castration) and bilateral orchiectomy (ie, surgical castration).
The third type is medical or surgical castration combined with an NSAA-combined androgen blockade (CAB). Comparative effectiveness studies of the three methods have shown mixed findings, and international guidelines are conflicting. [14][15][16][17][18][19][20] In China, there are limited data regarding the treatment pathways of patients with prostate cancer and bone metastasis in routine clinical practice, yet this is important knowledge to gain to evaluate whether patients are currently receiving the best available medical treatment. Furthermore, little is also known about the effectiveness of CAB compared with hormone monotherapies. We therefore aimed to explore treatment patterns among this patient population, including frequencies and sequences of therapies used and a comparison of different hormone therapies on the clinical outcomes of time to chemotherapy initiation and SRE occurrence. The study was set in Beijing, which has the nation's most advanced health care service.

| Study design and data source
We conducted a population-based cohort study using data from a large medical claims database in Beijing-the Urban Employee Basic Medical Insurance (UEBMI) database. The UEBMI is one of the three main national health insurance schemes in China and is a mandatory program covering urban and retired employees. 21 The UEBMI contains

KEY POINTS
• We used a claims database to summarize the observed treatment patterns of prostate cancer with bone metastasis in real-world clinical practice in Beijing; thus, our results will have better generalizability compared with those from clinical trials.
• Hormone therapy and bisphosphonate therapy were the most commonly prescribed treatments for prostate cancer with bone metastasis, in accordance with clinical guidelines.
• Chemotherapy was seldom used as initial therapy, but when used, the most common regimens were docetaxel and estramustine. Radiotherapy and surgery were used less frequently, but when used were often given in a late phase of the treatment pathway.
• Combined androgen blockade potentially has an advantage in delaying chemotherapy initiation over

| Study population
We included patients meeting all the following criteria between 1 July

| Measurements
We evaluated seven classes of treatments according to the 2014 version of the Chinese Urology Association (CUA) guideline for diagnosis and treatment of prostate cancer and following clinical expert consultation: surgery, radiotherapy, hormone therapy, chemotherapy, bisphosphonate therapy, pain treatment, and traditional Chinese medicine (TCM). 20 Hormone therapy was classed into three main groups: NSAA alone, castration alone, or CAB. If NSAA treatment period was less than 30 days and the interval between start of LHRH and NSAA treatments was less than 14 days, the patient was included in the castration alone group rather than the CAB group. This was because all patients with planned use of LHRH agonists should be prescribed short-term antiandrogens to prevent disease flares. 16 SREs were defined as clinical manifestations of pathological fracture, spinal cord compression, hypercalcemia, or surgery involving bone. Comorbidities were identified using ICD-10 codes and/or Chinese text in the medical claim records (any time from 1 January 2011 to the index date) and included hypertension, cardiovascular diseases, cerebrovascular disease, type 2 diabetes, and respiratory diseases (asthma and chronic obstructive pulmonary disease). Drug prescriptions were identified using Anatomical Therapeutic Chemical classification codes, and medical procedures were identified using a unified coding system developed by the data owner.

| Treatment sequence and clinical outcomes
In addition to identifying the initial treatment prescribed, we described the sequence of treatment over time, in particular noting the first three distinct treatments and the time interval between adjacent treatments. A Sankey plot was produced to visualize the treatment pathways. To compare clinical outcomes after treatment with CAB, NSAA alone, or castration alone, we followed patients from the first date of the respective hormone therapies to identify the occurrence of SREs and the initiation of chemotherapy (as a proxy for disease progression). To do this, we conducted two time-to-event analyses, one for each outcome. In these two analyses, patients who did not receive any of the three exposure treatments of interest (eg, patients who only received oestrogens as hormone therapy) and those who already had an outcome event (chemotherapy) before the relevant exposure were excluded. Patients without an outcome of interest were censored when they were last known to be event-free during the observation period.       Table 3 shows the numbers of patients prescribed specific hormone therapy and chemotherapy agents, and the number of medical/hospital visits made for use of each of these therapies.

| Commonly used drugs
Among the hormone therapies, NSAAs (predominantly bicalutamide followed by flutamide) and LHRHs (mostly leuprorelin or goserelin followed by triptorelin) were the most commonly prescribed drugs.
Among chemotherapy drugs, docetaxel was the most commonly prescribed, followed by estramustine, platinum-based agents, and fluorouracil. to radiotherapy was 171 days, and to surgery was 226 days.

| Time to initiation of chemotherapy and first SRE
Adjusted HRs and median progression time from first hormone therapy to initiation of chemotherapy/first SRE are presented in Table 4. Direct adjusted survival curves are shown in Figure 2. After adjusting for covariates, patients treated with NSAA alone had more than double the likelihood of subsequently receiving chemotherapy compared with those treated with CAB (HR, 2.43; 95% CI, 1.08-5.44; P = .031), while no significant difference was found between other hormone therapy treatment groups for chemotherapy initiation.
In the time to first SRE analysis, the median progression time to SRE was not reached in each group and there was no evidence of difference in time to first SRE between hormone groups.

| DISCUSSION
In this population-based cohort study, we have shown that hormone therapy and bisphosphonate therapy with zoledronic acid are the most commonly used for patients in Beijing with prostate cancer and bone metastasis, with the majority of patients starting treatment with one or both of these therapies, in accordance with current guidelines. 19,25 We have also shown that time to chemotherapy initiation was longer  93.5% of patients received castration in our study, the proportion is higher than that reported by Flaig et al (65%). Among them, medical castration (85.9%) was a more common form of castration than surgery (7.6%), though they appear equally effective with orchiectomy proved even safer recently. 27,28 Possible reason is that medical castration is potentially reversible and avoids the physical and psychological discomfort associated with orchiectomy. 29 Bone-targeting agents reduce bone reabsorption by primarily targeting osteoclasts. 4 They are important in the treatment of patients with bone metastatic prostate cancer, and if not used, approximately half of patients will experience one or more SREs within 2 years. 30 Data from clinical trials suggest that conventional bone-targeting treatments, such as bisphosphonate therapy with zoledronic acid, which was prescribed to most patients in our study, does not increase overall survival. 9,30 Radium-233-one of several newer agents available-has, however, been shown to prolong survival in addition to reducing symptomatic bone complications. 31 Forty-two percent of patients in our study were prescribed chemotherapy, which is very similar to that reported in a claims database study in the United States (45.4%). 32 Docetaxel is the standard of care in patient with CRPC and was the most commonly used chemotherapy drug in our study cohort, followed by estramustine, which is no longer recommended by the American guideline 19 but still retained in the Chinese guidelines. 20,25 Since 2015, the use of chemotherapy together with ADT as initial therapy for mHSPC patients has been supported by clinical guidelines. 5,33 In our study, we found that chemotherapy was rarely prescribed as the initial therapy and therefore was likely a sign of disease progression, as used as a secondary endpoint in a recent clinical trial. 11 Radiotherapy and surgery were less frequently used and usually given late in the treatment pathway. Similar treatment patterns were observed by Seal et al in a hospital database study. 34 However, for the specific drugs used in each class of treatment, patterns found in our study differed from that in a US population, 35   evidence is required, especially from prospective randomized trials, to be confident about a survival advantage with CAB. According to ASCO and CUA, NSAA alone may be discussed as an alternative in patients at certain stage, while EAU is more neutral to its use and NCCN does not recommend its use. 5,17,19,20 The results of our study moderately add to the evidence favouring CAB over NSAA alone but not over castration alone on delaying chemotherapy initiation in Chinese bone metastatic prostate cancer patients. We did not find any differences in occurrence of SREs between CAB and hormone monotherapies, although it should be noted that it is BTAs rather than hormone therapies that specifically aim to reduce these outcomes.
Our study has several strengths. To our knowledge, it is the first to look at treatment patterns among patients with prostate cancer and bone metastasis in China and thus provides valuable clinically important information describing the medical management of these patients.
Without data such as ours to use as a benchmark, it would be difficult to describe use and effectiveness of newer approved medications as they become approved and prescribed in China, through both future descriptive and analytical epidemiological studies. We used a large database representative of urban employees in Beijing enabling a reasonable sample size to be acquired. Our study also has some limitations. Firstly, the UEBMI database does not provide results of mortality data or detailed clinical and pathologic information; hence, we were unable to directly identify patients with mCRPC, and we used time to chemotherapy initiation and first SRE as surrogate outcome measures. Secondly, the sample size might still be insufficient to detect small to moderate differences, and the observation period was short leading to many censored observations in the survival analysis especially for the SRE endpoint, which to some extent undermines the observed results.
Finally, Beijing is one of the few first tier cities in China with better health care resources; therefore, treatment patterns identified in this study cannot necessarily be generalized to other regions of China.

| CONCLUSIONS
Our findings show that while a wide range of therapies are used to

ETHICS STATEMENT
This study was approved by the Ethical Review Board of Peking University Health Science Center (No. IRB00001052-16027-Exempt), and informed consent of participant was exempted.