Long-term Safety Follow-up of Subjects With Giant Cell Tumor of Bone Treated With Denosumab in Study 20062004: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Long-term Safety Follow-up of Subjects With Giant Cell Tumor of Bone Treated With Denosumab in Study 20062004 (NCT03301857), a non-randomized phase 4 study. The primary outcome is Number of Participants Experiencing Adverse Events (AEs) of Interest (EOI).

2. Clinical Question

The study addressed long-term safety among people previously treated with denosumab for giant cell tumor of bone. Its focus was not a randomized test of denosumab against an assigned control. Rather, follow-up continued for participants who remained in Study 20062004 at its conclusion, with their subsequent treatment or follow-up status distinguishing the groups.

The registry also lists disease progression or recurrence and receipt of interventions for giant cell tumor of bone among secondary outcomes. These outcomes could describe what happened during follow-up, but they do not by themselves establish that denosumab caused or prevented an event. The posted record does not report numerical results for the primary endpoint or those secondary outcomes.

3. Trial Design

This completed phase 4 study, NCT03301857, was a long-term safety follow-up for participants with giant cell tumor of bone treated in Study 20062004. The registry lists a planned enrollment of 85 and describes a parallel, non-randomized design with no masking. The sponsor was Amgen. The listed start date is 2017-11-13, and the primary completion date is 2023-07-27.

The follow-up included two groups: participants exposed to the investigational product and participants not exposed to it. The registry describes continued denosumab treatment for participants still receiving it when Study 20062004 ended, and ongoing safety visits for participants who had completed treatment. The latter visits were scheduled every 6 months and could take place by telephone or in person. For those undergoing retreatment, the described denosumab regimen was 120 mg subcutaneously on Days 1, 8, 15 and 28, then every 4 weeks subsequently.

These groups were not formed by random assignment. Differences between them may therefore reflect who continued, stopped, or restarted treatment, as well as other differences not described in the registry record. The comparison is useful as a description of safety outcomes within the follow-up cohort, but it is not a like-for-like control comparison.

The primary safety outcome was the number of participants experiencing adverse events of interest over a timeframe of up to approximately 5 years. Other listed outcomes included treatment-emergent adverse events, disease progression or recurrence, and receipt of giant cell tumor of bone interventions. The record does not report a detailed statistical analysis plan, interim analyses, or a sample-size justification.

4. Participant Flow and Accounting

The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.

Before assignment: Participants in the study were enrolled upon completing study 20062004.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: FAS included all enrolled participants (from study 20062004) who provided informed consent and had a non-missing enrolment date in this study.

5. Baseline Characteristics

The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.

Population: Full analysis set (FAS) included all enrolled participants (from study 20062004) who provided informed consent and had a non-missing enrolment date in this study.

6. Primary Endpoint

The registry lists the primary outcome as Number of Participants Experiencing Adverse Events (AEs) of Interest (EOI), measured over Up to approximately 5 years.

The primary endpoint counts participants who experienced adverse events of interest, rather than counting every event. The events included signs and symptoms of osteonecrosis of the jaw, malignancy including malignancy in giant cell tumor of bone, atypical femoral fracture, hypocalcemia, hypercalcemia after treatment discontinuation, and pregnancy or lactation when occurring during treatment or within 5 months of the last denosumab dose.

The registry applies specific time rules to some outcomes. Hypocalcemia includes treatment-emergent adverse events occurring after 30 days following the last dose of the investigational product. Other events of interest are counted from signing informed consent through the end of the study, described as approximately 5 years. Osteonecrosis of the jaw and atypical femoral fracture events were adjudicated by independent reviewers.

A participant count is a clinically interpretable way to describe how many people experienced at least one event in a safety follow-up study. It does not convey the number of repeated events, their duration or severity, or the time each participant remained at risk. The endpoint is therefore a defined safety summary, not a complete account of each person’s clinical course. The registry record does not report the primary endpoint result.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The registry presents outcomes primarily as participant counts. For a binary safety outcome, each participant is classified according to whether the specified event occurred during the defined observation period. The resulting count answers how many participants had at least one qualifying event; it does not automatically count repeated events separately. This distinction matters because a participant with one event and a participant with several events can each contribute one person to a participant-count endpoint.

What does a participant count tell the reader?

A count is a direct description of the observed cohort, but interpretation depends on the population and observation window. The primary endpoint covers adverse events of interest over up to approximately 5 years, with different event-specific rules. Other adverse-event summaries have their own definitions. The registry record should therefore be read according to each outcome’s stated definition rather than treating all adverse-event counts as interchangeable.

Counts also need context about how many people were followed and for how long. The record lists 51 exposed participants and 34 not exposed for the adverse-event groups, but does not report person-time or detailed follow-up duration by group. A simple count does not account for potentially unequal time at risk. Nor does it distinguish events that occurred once from recurrent events unless a separate event-count measure is reported.

Why does non-randomization matter?

In a randomized trial, chance assignment is intended to make comparison groups similar on average at the outset. Here, allocation is listed as non-randomized. Participants’ treatment status during the extension may reflect their prior treatment course or circumstances at the end of Study 20062004. The record does not give enough information to determine which factors shaped group membership. As a result, differences in event counts can reflect differences between participants or in observation, as well as any possible treatment-related difference.

Masking is listed as none. In general, masking can reduce the influence of knowledge about treatment assignment on some aspects of care, reporting, or assessment. Its absence does not invalidate recorded outcomes, but it is relevant when considering subjective event reporting and the interpretation of group differences. For osteonecrosis of the jaw and atypical femoral fracture, the record states that events were adjudicated by independent reviewers, which describes an additional review process for those events.

What can be said about precision and statistical testing?

The posted results include safety counts and selected serious-event terms, but no inferential analysis results are reported in the registry record. There are no reported p-values, confidence intervals, or adjusted comparisons in the record. A confidence interval is a statistical range intended to convey uncertainty around an estimate under a specified model. Without a reported estimate and its interval, readers cannot assess the precision of a comparative effect from these entries. A difference between observed counts alone is not proof of a treatment effect.

The record mentions planning illustrations based on the observed effect and assumptions including a two-sided significance level of 0.05, equal allocation, and a normal approximation. These are planning calculations, not evidence that the study was designed with those assumptions or that a particular power was achieved. Such illustrations do not represent observed power. The registry record does not report a sample-size justification for the study’s own design, so the planning assumptions should not be mistaken for the original protocol’s rationale.

How should event percentages and categories be read?

The serious adverse-event terms are reported as percentages. A percentage tied to an event term is not necessarily the same measure as the participant count for all serious adverse events. The registry record does not specify in the listed summary whether event terms can overlap within participants, nor does it provide a detailed event-level breakdown. The terms and percentages should be understood as the reported listing rather than used to reconstruct unreported counts.

The category “adverse events of interest” is also narrower than the full set of adverse events. Its definition includes selected events and specified timing conditions, while the record separately reports serious and other adverse-event summaries. A person could potentially be represented in more than one safety category, depending on the definitions; the record does not describe person-level overlap. For these reasons, adding categories or interpreting them as mutually exclusive would not be justified.

Overall, the available methods information supports descriptive reading of the listed counts and definitions. It does not support reconstruction of an unreported statistical model, correction for confounding, or calculation of a precise treatment contrast. The registry record does not report an analysis plan detailed enough to answer those questions.

10. Multiplicity, Missing Data and Analysis Populations

The registry lists a primary safety endpoint and secondary outcomes that include treatment-emergent adverse events, disease progression or recurrence, and receipt of giant cell tumor of bone interventions. It does not report how statistical testing across multiple outcomes was planned or whether any adjustment for multiple comparisons was used. Without that information, the possibility of chance findings across several outcomes cannot be quantified from the record. No inferential outcome analyses are reported in the registry record.

Missing-data handling is not reported in the registry record. The primary analysis population is defined as participants from Study 20062004 who enrolled in this follow-up, provided informed consent, and had a non-missing enrolment date. This eligibility definition identifies who could be included, but it does not explain how incomplete follow-up, missing event information, withdrawal, or unavailable visits were handled. The record also does not describe whether missing outcomes were treated as absent events or addressed by another method.

The lack of reported missing-data procedures limits interpretation of the safety counts, especially if completeness of follow-up differed across groups. The record does not provide enough detail to assess whether all participants contributed the same observation time or whether participants lost to follow-up differed in their event risk. Those uncertainties are distinct from the stated outcome definitions and should not be resolved by assumption.

Although the record includes the term “FAS” for the defined population, the detailed handling of analysis populations beyond that definition is not reported in the registry record. No protocol-based multiplicity strategy, imputation method, or interim-analysis plan is reported.

11. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 5% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: femur fracture (3.9%), anaemia (2.0%), anal fistula (2.0%), gastrointestinal obstruction (2.0%), medical device site infection (2.0%).

The adverse-event listing distinguishes serious adverse events, other adverse events, and deaths by exposure group. It reports 8 participants with serious adverse events among 51 exposed, compared with 0 among 34 not exposed. The corresponding other-adverse-event counts are 37 and 0. Death counts are 0 and 2, respectively. These are descriptive counts, and the registry record does not provide a causal attribution for them.

The listed serious-event terms are femur fracture at 3.9%, anaemia at 2.0%, anal fistula at 2.0%, gastrointestinal obstruction at 2.0%, medical device site infection at 2.0%, and sepsis at 2.0%. These term-level percentages should not be treated as a complete account of each participant’s safety experience. The record does not provide event timing, recurrence, participant-level overlap, or a detailed accounting of exposure duration for these terms.

Seriousness and relationship to treatment are different concepts: a serious event is not automatically caused by the intervention. The group differences also cannot be interpreted as a controlled comparison because allocation was non-randomized. The primary adverse-events-of-interest outcome has no reported numerical result in the registry record, so it should not be inferred from these separate safety summaries.

12. What the Result Does, and Does Not, Mean

The record establishes that a long-term follow-up study was completed and reports selected adverse-event summaries for participants grouped by exposure status. It also defines a primary safety outcome focused on the number of participants experiencing specified events over up to approximately 5 years. This provides a framework for describing safety monitoring in a population previously treated for giant cell tumor of bone.

The posted counts do not establish whether denosumab caused any listed event, nor do they establish that one group had greater or lesser risk. Participants were not randomized to the groups, and the record does not report adjustment for differences between them, the timing of events, or group-specific follow-up duration. The observed counts therefore describe what was recorded in the groups rather than an unbiased causal effect.

The absence of a reported numerical result for the primary endpoint is important. The event-of-interest definition cannot be used to infer how many participants experienced those events. Likewise, the record lists disease progression or recurrence and receipt of interventions as secondary outcomes but does not report their numerical results. The available information cannot answer whether those outcomes differed by group.

In general, statistical uncertainty concerns how much observed estimates might vary across comparable studies. Here, no comparative estimate with a confidence interval or p-value is reported, so the precision of a group contrast is not available. The event counts should not be converted into unreported percentages, relative effects, or significance claims.

The results are best understood as a registry-level description of safety summaries in a non-randomized extension follow-up. They are not a randomized efficacy comparison, and they do not provide a complete account of risks or benefits. The record’s definitions and group counts are informative, while important details needed for causal or precise comparative interpretation are not reported.

13. Why This Trial Matters Statistically

Long-term follow-up is statistically distinct from a short, fixed treatment comparison. Safety events may arise at different times, and participants may contribute different amounts of observation. A participant count can be useful for describing how many people had an event, but it does not by itself account for variation in follow-up time or repeated events. The registry’s event-specific definitions and follow-up framework are therefore central to interpreting its summaries.

This study also illustrates the limits of comparisons formed by treatment exposure rather than random assignment. In a follow-up cohort, treatment continuation and safety monitoring occur within the context of prior study participation. Without randomization or reported adjustment, group counts cannot isolate the effect of treatment from other differences between participants. The group labels must be read as exposure classifications, not as equivalent trial arms.

The contrast between detailed endpoint definitions and absent primary numerical results is another important statistical lesson. A clear endpoint definition tells readers what was intended to be counted, but it cannot substitute for the observed result. Similarly, lists of serious-event terms do not replace a complete analysis of the primary safety outcome, disease outcomes, missingness, or uncertainty.

For clinical researchers, the record emphasizes the value of reporting denominators, event definitions, time at risk, completeness of follow-up, and analysis methods alongside safety counts. Without those elements, readers can describe the posted data but cannot make a reliable causal comparison or assess precision. This is especially consequential in a non-randomized, long-term safety study, where treatment history and follow-up patterns may differ across groups.

14. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

The central limitation is that the registry record does not report the numerical result for the primary endpoint, the number of participants experiencing adverse events of interest. It also does not report numerical results for the listed disease progression or recurrence outcome or the outcome measuring receipt of interventions. This prevents direct assessment of the study’s main stated safety question and its disease-related secondary questions.

The design was non-randomized and unmasked. Group membership may be related to prior treatment and other characteristics, but the record does not describe those characteristics or how they were accounted for. No adjusted comparative estimate is reported. The safety counts therefore should not be interpreted as causal contrasts.

Follow-up completeness and time at risk by group are not reported. The registry describes six-month safety visits for participants who had completed denosumab treatment, but does not provide group-specific follow-up duration or explain how missing visits and incomplete outcome data were handled. This makes it difficult to determine whether the opportunities to record events were comparable.

Event-level detail is limited. The record lists selected serious-event terms and percentages, but does not provide timing, recurrence, person-level overlap, or a complete breakdown that would clarify how those summaries relate to the participant-level categories. Multiplicity handling, interim analyses, and the study’s statistical analysis plan are not reported in the registry record. These limitations restrict both the precision and scope of conclusions that can be drawn from the posted summaries.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Long-term Safety Follow-up of Subjects With Giant Cell Tumor of Bone Treated With Denosumab in Study 20062004 — NCT03301857.
• Registry record last updated 2024-05-20; status last verified 2024-05.
• Results first posted 2024-05-20.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

18. Record Summary

NCT03301857 is a completed phase 4, parallel, non-randomized, unmasked long-term safety follow-up study of participants with giant cell tumor of bone previously treated in Study 20062004. The registry lists enrollment of 85, a start date of 2017-11-13, and a primary completion date of 2023-07-27.

The study followed participants who remained in the earlier study at its completion. The primary endpoint was the number of participants experiencing adverse events of interest over up to approximately 5 years. The record also lists treatment-emergent adverse events, disease progression or recurrence, and receipt of interventions as outcomes.

Safety summaries list 51 participants exposed to the investigational product and 34 not exposed, with serious adverse-event counts of 8 and 0, other adverse-event counts of 37 and 0, and death counts of 0 and 2. The numerical result for the primary endpoint and results for the listed disease outcomes are not reported in the registry record. Because the groups were not randomized, the counts are descriptive and do not establish a treatment effect.