Introduction

Teriparatide (Forteo; Eli Lilly and Company) [1], first approved in November 2002 by the United States (US) Food and Drug Administration (FDA), is a recombinant human parathyroid hormone analog indicated for the treatment of postmenopausal women with osteoporosis at high risk for fracture, and to increase bone mass in men with primary or hypogonadal osteoporosis at high risk for fracture. In July 2009, the indication for teriparatide was extended to include the treatment of osteoporosis associated with sustained systemic glucocorticoid therapy (glucocorticoid-induced osteoporosis) in both men and women.

In preclinical toxicology studies, teriparatide caused increases in bone mass and a dose-dependent increase in the incidence of osteosarcoma in rats [2]. Later animal studies identified a no-effect dose of 5 μg/kg/day in rats [3], and no osteosarcomas were observed in cynomolgus monkeys treated with this dose [4]. No cases of osteosarcoma emerged during clinical trials of teriparatide or in a 5-year posttreatment follow-up study, and few spontaneous cases of osteosarcoma have been reported in patients treated with teriparatide [1].

In 2009, the FDA required the implementation of a voluntary, prospective registry to estimate the incidence of osteosarcoma in patients receiving treatment with teriparatide as a condition for approval of teriparatide for the new glucocorticoid-induced osteoporosis indication [5]. The Forteo Patient Registry was established in the US in 2009 as a prospective voluntary registry to estimate the incidence of osteosarcoma in patients who have received treatment with teriparatide. This study is part of a broader postmarketing safety program evaluating the long-term safety of teriparatide using real-world data. Recently completed complementary studies have included the 15-year retrospective Osteosarcoma Surveillance Study [6], and a study using Medicare claims data to identify exposure and linkage with state cancer registries to identify osteosarcoma cases [7]. The methods of the Forteo Patient Registry have been described in detail elsewhere [8]. The aims of this paper are to present the final results of the Forteo Patient Registry in detail and to summarize the long-term safety evidence for teriparatide.

Methods

Study overview

Gilsenan et al. [8] described the methods of the Forteo Patient Registry in detail. Briefly, the Forteo Patient Registry was established in 2009 to estimate the incidence of osteosarcoma in US adult patients aged ≥ 18 years who have used teriparatide. Patients enrolled in the registry voluntarily between 2009 and 2019. The information from all enrolled participants was linked annually (2010–2019) with all participating cancer registries to ascertain any incident cases of osteosarcoma (i.e., cases diagnosed after the date of starting teriparatide and after study enrollment). Patients were considered exposed to teriparatide if they self-reported use of the drug at least once. Patients were classified as new users if they initiated teriparatide less than 3 months from the date of registration, recent users if the teriparatide start date was 3 to 6 months before registration, and past users if they initiated more than 6 months before registration. An osteosarcoma case was defined as histologically confirmed osteosarcoma that produced osseous matrix and fell within one of the categories identified using International Classification of Diseases for Oncology, Third Edition.

The study was approved by the RTI International (RTI) institutional review board and by local institutional review boards affiliated with participating state cancer registries when required.

Study size

The intent of the study was to estimate the incidence of osteosarcoma in adult patients who received treatment with teriparatide by observing 1.7 million patient-years of follow-up time. If achieved, when compared with the background rate at study initiation of 2.7 cases per million population per year (95% confidence interval [CI], 2.4–3.0; age-adjusted to the 2000 US standard population) [9], a threefold increase in the risk of osteosarcoma could be detected. A threefold risk increase would equate to an absolute risk increase of approximately one additional case per 185,000 patient-years observed.

Patient recruitment and enrollment

Patients were recruited primarily through pre-enrollment forms included in the teriparatide device packaging, treatment starter kits, and brochures distributed by physicians or nurses. Patients expressed interest in the registry via a completed pre-enrollment form. Eligible pre-enrolled patients were mailed a registration letter, registration form, informed consent form, and small compensation for their time in completing the forms. Registration was complete upon the patient’s return of the completed forms.

Cancer registry linkage

Cancer registries in all 50 states and the District of Columbia were invited to participate in the Forteo Patient Registry.

Link Plus (version 2.0), a probabilistic matching software program available from the Centers for Disease Control and Prevention (CDC), was used to develop a standard linkage algorithm that was used by all participating state cancer registries to ensure the annual data linkage was performed consistently [10]. Only incident osteosarcoma cases were considered reportable study outcomes.

During the annual linkage process, cumulative registration data from all enrolled patients—including name, date of birth, address, telephone number, last four digits of their Social Security number, race, and ethnicity—were sent to participating state cancer registries via a secure file transfer protocol. State cancer registries linked the registration data locally with all incident osteosarcoma cases in adults diagnosed since January 1, 2009, in their state cancer registry database.

Statistical analysis

The incidence rate of osteosarcoma was calculated as the number of new osteosarcoma cases following initiation of teriparatide treatment divided by the cumulative person-years at risk adjusted for mortality, and an exact 95% CI for the incidence rate was provided. The cumulative person-time at risk of osteosarcoma for registered patients was estimated by multiplying the number of new patients in the registry each year by the number of years since they first took the drug and adjusted for mortality using age-specific mortality rates from the National Center for Health Statistics. A standardized incidence ratio (SIR) of observed cases to the expected number of osteosarcoma cases adjusted to the age and sex of teriparatide users (3 per million person-years) and corresponding exact 95% CI was also calculated.

Adverse event reporting

The Forteo Patient Registry did not solicit adverse events (AEs) or product complaints; however, spontaneous reports of teriparatide AEs and/or product complaints from patients during the conduct of the study were reported to the teriparatide manufacturer as spontaneous AEs. Any osteosarcoma matches occurring after linkage with cancer registries were considered serious AEs and were reported to the teriparatide manufacturer.

Results

Participant characteristics

Among the 75,247 patients registered in the Forteo Patient Registry database, most were female (88.6%), non-Hispanic (94.3%), and white (93.9%), and the mean age was 69 years (Table 1). Overall, 70.4% of registrants in the Forteo Patient Registry were “new users,” defined as having a teriparatide start date less than 3 months from the date of registration; 18.4% of registered patients started teriparatide 3–6 months prior to completing registration, and 11.2% of registered patients started teriparatide more than 6 months before completing registration.

Table 1 Characteristics of patients included in the Forteo Patient Registry (n = 75,247)
Full size table

The overall sex distribution of patients who completed registration in the Forteo Patient Registry (89% female, 11% male) was similar to the sex distribution of the general teriparatide user population (91% female, 9% male) based on new patient start data [11]. Figure 1 displays the age distribution stratified by sex of the general teriparatide user population based on market research data, and Fig. 2 displays the distribution of age stratified by the sex of patients enrolled in the Forteo Patient Registry. As shown, the distribution of age by sex for those enrolled in the Forteo Patient Registry was generally similar to that of teriparatide users in the age category.

Fig. 1
figure 1

Age distribution among the general teriparatide population, stratified by sex (prevalent teriparatide users January 2009 through June 2018) (n = 75,247) (Adapted from IMS Health Prescription Data (LRx), January 2009–June 2018 data provided October 2018)

Full size image
Fig. 2
figure 2

Age distribution of registered patients, stratified by sex, as of June 30, 2019. N = 75,246. One patient was missing data for sex and was excluded from counts in this figure

Full size image

Distribution by method of pre-enrollment and by geography

The teriparatide packaging was the most common source of pre-enrollment forms received among registered patients, with other promotional materials the second most frequent source (Table 2). At least one patient from each participating state was included in the linkage file (Fig. 3), and the distribution of patients was consistent with the size of the population in each state. Although not all 50 states participated in the linkage, all patient data were provided to each of the participating state cancer registries in an effort to capture data on patients who may have moved across state lines after enrolling in the Forteo Patient Registry. States that did not participate in the study had a lower percentage of registered patients (i.e., < 1% or 1 to < 5% of registered patients). Approximately 95% of all enrollees contained in the 2019 data linkage file lived in one of the 42 states (including the District of Columbia) that participated in the 2019 linkage.

Table 2 Cumulative number of registered patients by mode of pre-enrollment
Full size table
Fig. 3
figure 3

Geographic distribution of patients included in the Forteo Patient Registry Linkage File (n = 75,247). (a) Approximately 5% of patients registered in the Forteo Patient Registry lived in one of the nine states where the cancer registry did not participate in the study. Geographic distribution based on state of residence at time of enrollment into the Forteo Patient Registry

Full size image

Linkage with cancer registries

For the 10th and final annual linkage, completed September 25, 2019, data for 75,247 patients enrolled in the Forteo Patient Registry were linked to the research databases for each of the 42 participating state cancer registries (covering 93% of the US population). These registries included information on 6180 cases of osteosarcoma in adults aged 18 years or older diagnosed since January 1, 2009. Owing to the typical lag time of 9 to 18 months between date of diagnosis and the date cases are available in the cancer registry database, not all cases for more recent diagnosis years were available. On average, registries reported being mostly complete for diagnosis years 2009–2016, almost complete for diagnosis year 2017, partially complete for diagnosis year 2018, and only 2% complete for diagnosis year 2019.

No matches with incident cases of osteosarcoma following registry enrollment were identified during any of the 10 annual linkages with the cancer registries. The cumulative person-time at risk for registered patients was estimated at 361,763 person-years after adjusting for mortality; therefore, the crude incidence rate was 0 (95% CI, 0–10.2) cases per million person-years. Compared with the expected number of cases adjusted to the background rate (3 cases per million person-years), the SIR was 0 (95% CI, 0–3.0).

Of note, one case was identified during the seventh annual linkage in 2016 for a patient who had been diagnosed with osteosarcoma before enrolling in the Forteo Patient Registry. Because this case did not qualify as newly diagnosed after study enrollment, it was not included as a reportable study outcome. However, this case was reported to the FDA as a spontaneously identified AE.

Discussion

The Forteo Patient Registry enrolled over 75,000 patients during a 10-year period and conducted annual linkages with individual cancer registries covering 93% of the US population. During this time, no incident cases of osteosarcoma were identified, although the likelihood of identifying one or more cases was somewhat limited given the lower than expected patient enrollment and cumulative person-years of follow-up. The fact that one case did match (despite not qualifying as a study outcome because the patient was diagnosed with osteosarcoma before enrolling in the Forteo Patient Registry) provides evidence that the linkage algorithm did work.

The Forteo Patient Registry is part of a broader postmarketing initiative to understand the long-term safety of teriparatide using real-world evidence. A complementary retrospective study, the Osteosarcoma Surveillance Study, identified incident cases of osteosarcoma through participating cancer registries in the US and evaluated potential teriparatide exposure through patient self-report or proxy report in telephone interviews [6]. Based on the background incidence rate for osteosarcoma, the expected number of cases among patients treated with teriparatide for this retrospective study was 4.17, and 3 cases of osteosarcoma were observed among 1173 patients interviewed. This study’s estimated standardized incidence ratio of expected to observed cases was 0.72 (90% CI, 0.20–1.86). In addition, a population-based comparative cohort study estimated the incidence rate ratio of osteosarcoma among US patients aged ≥ 65 years treated with teriparatide (identified via Medicare prescription claims) versus a matched comparator cohort [7]. The incidence rate in the comparator cohort was consistent with the background incidence rate among adults aged ≥ 65 years, and no osteosarcoma cases were observed in the teriparatide cohort, yielding an incidence rate ratio of 0 (95% CI, 0–3.2).

Based on the totality of evidence from the Forteo Patient Registry, the Osteosarcoma Surveillance Study, and the Medicare comparative cohort study, none of which showed an increased risk of osteosarcoma in patients treated with teriparatide; the FDA released Eli Lilly, the manufacturer of teriparatide, from its commitment to continue this monitoring program [6, 7]. Therefore, although the linkage of cumulative registration data from the Forteo Patient Registry had been planned to continue through 2024, the 2019 linkage was the final linkage of the study.

Some limitations of this study must be noted. First, the study did not have an internal comparator cohort, which was a key reason for initiating the Medicare comparative cohort study [7]. In addition, as with any voluntary registry, it is challenging to achieve sufficient participation, and the number of patients enrolled in the Forteo Patient Registry was insufficient to reach the target number of person-years at risk to evaluate the risk of developing osteosarcoma. Although no matches were found at the final cumulative 10th year of linkage, due to the low background rate of osteosarcoma, the CI of the final rate of osteosarcoma (0 to 10.2 cases per million person-years) is wide. Despite the imprecision, the upper bound of the 95% CI for zero cases is still a very low incidence. Importantly, no information on total duration of teriparatide treatment or other clinical information between time of registration and linkage with the cancer registries was collected. The simplified patient recruitment process required trade-offs between increasing participation and quantity of data collected. Such information would have been useful for subset analyses if the study had generated a higher incidence of osteosarcoma among teriparatide users than the general population, which was not the case. Further, due to the voluntary nature of the study, the group of patients who participated in the registry was not a random sample of the overall teriparatide patient population. It is possible but not likely that factors associated with the development of osteosarcoma could be differentially distributed between participants and nonparticipants. Thus, the incidence estimate based on participation may be biased. The analytic approach used in the study did not address the lag period between the date of cancer diagnosis and availability of patient data in the statewide cancer registries; the overall lag period between diagnosis and availability of data ranges from 9 to 18 months. Furthermore, given that the latency of osteosarcoma is unknown, this study does not address a possible extended latency period for development of osteosarcoma. Finally, due to the voluntary nature of the registry, generalizability to all patients treated with teriparatide should be made with caution. The comparison of available characteristics suggests that those who enrolled in the Forteo Patient Registry were similar to patients who were prescribed teriparatide; however, unmeasured characteristics that could potentially predispose (or not) patients to osteosarcoma could make results less generalizable.

Conclusions

The ability to draw conclusions about the incidence of osteosarcoma among patients participating in the registry was limited due to lower than expected patients enrolled and the smaller than expected cumulative amount of patient follow-up time, as well as the fact that no cases were identified. Given the SIR comparing the observed number of cases to the expected number of cases based on the adjusted background rate of 3 per million person-years, the findings suggest there could be somewhere between no increased risk of osteosarcoma up to a 3-fold increase.