1. General FDA Position and Guidance

The FDA strongly encourages sponsors to identify and submit uses of RWE/RWD to support regulatory decisions regarding product safety and effectiveness. Sponsors are directed to include such information in their submission cover letters and to consult relevant FDA guidance documents. However, the FDA maintains high standards for the inclusion of RWE/RWD in Prescribing Information (PI), emphasizing methodological rigor, data quality, and pre-specified protocols and statistical analysis plans (SAPs) ^{1234561011121819}.

---

2. Examples Where RWE/RWD Was Considered or Used as Supportive Evidence

A. DUVYZAT (Givinostat Hydrochloride)

• Use of RWE/RWD: The applicant submitted an integrated analysis of long-term efficacy using natural history data (RWE) as confirmatory evidence alongside a pivotal randomized controlled trial for Duchenne muscular dystrophy (DMD).
• FDA Assessment: The FDA acknowledged the value of RWE as confirmatory evidence, especially in rare diseases, but cited several limitations:
  • Lack of a pre-specified protocol and SAP agreed upon with the FDA.
  • Heterogeneity in disease progression and lack of comparability between treated and control groups.
  • Potential for residual confounding and differences in follow-up.
• Outcome: The RWE was considered exploratory and supportive, not definitive. The FDA recommended that such analyses be interpreted with caution and not included as substantial evidence in the PI ^71314232528.

B. FABRAZYME (Agalsidase Beta)

• Use of RWE/RWD: The FDA reviewed a matched analysis using registry data (treated patients) and an external natural history study (untreated patients) to supplement evidence from a Phase 4 trial.
• FDA Assessment: The RWE was considered "low-level evidence" due to:
  • Use of a non-contemporaneous external comparison group.
  • Selection bias, differences in clinical site location, calendar time, and age of diagnosis.
  • High patient attrition and missing data.
• Outcome: The RWE was accepted as supporting evidence but not as standalone or primary evidence for efficacy in the PI ^1516222430.

C. SKYCLARYS (Omaveloxolone)

• Use of RWE/RWD: The FDA reviewed a natural history study as confirmatory evidence for efficacy in Friedreich's ataxia.
• FDA Assessment: The analysis was post hoc, lacked a pre-specified protocol/SAP, and suffered from limitations typical of external control studies (e.g., uncontrolled confounding, selection bias).
• Outcome: The RWE was considered supportive but not sufficient for inclusion as primary evidence in the PI ⁸¹⁷²⁶.

D. PROGRAF (Tacrolimus)

• Use of RWE/RWD: Astellas Pharma sought to add a new indication based on RWE from a retrospective analysis of the Scientific Registry of Transplant Recipients.
• FDA Assessment: The FDA considered RWE as supplemental but not as the primary basis for establishing efficacy for labeling changes, requiring more robust clinical trial data for new indications ²⁷.

---

3. Examples Where RWE/RWD Was Submitted but Not Included in Prescribing Information

A. XPOVIO (Selinexor)

• Use of RWE/RWD: The sponsor submitted analyses using retrospectively collected electronic health record (EHR) data.
• FDA Assessment: The FDA did not accept this RWD for inclusion in the PI due to:
  • Lack of a pre-specified protocol and SAP submitted prior to the study.
  • Substantial differences in inclusion/exclusion criteria between the clinical trial and RWD cohorts.
  • Concerns about selection bias, misclassification, and confounding.
• Outcome: The RWE was not accepted for labeling due to methodological and procedural concerns ²⁰.

B. LITFULO (Ritlecitinib Tosylate)

• Use of RWE/RWD: The sponsor proposed using RWE to supplement Phase 2b/3 data for longer-term safety.
• FDA Assessment: The FDA responded to procedural questions but did not indicate that RWE would be included in the PI. Acceptance was conditional and context-dependent ⁹.

C. XARELTO (Rivaroxaban)

• Use of RWE/RWD: The FDA discussed the potential for including RWE in benefit-risk assessments.
• FDA Assessment: The review highlighted the need for further assessment and consensus on the appropriateness of including RWE in PI, with no specific example of inclusion ²⁹.

D. DOJOLVI (Triheptanoin)

• Use of RWE/RWD: The FDA addressed the use of published literature and real-world studies not conducted by or for the applicant.
• FDA Assessment: The agency clarified that reliance on such data requires a 505(b)(2) pathway and rigorous justification ³¹.

---

4. Summary of FDA Rationales for Exclusion or Limited Use of RWE/RWD in Prescribing Information

Across multiple cases, the FDA has cited the following reasons for not including RWE/RWD as primary evidence in Prescribing Information:

• Lack of pre-specified protocols and SAPs reviewed by the FDA prior to study conduct ^{78913141720232528}.
• Non-comparability of groups due to selection bias, differences in patient populations, time period, or geography ^{7131415162224252830}.
• Potential for residual confounding and bias inherent in observational data ^{71314151617222324252830}.
• Exploratory or post hoc nature of analyses without pre-specified plans ^{7813141720232528}.
• High attrition or missing data impacting validity ^1516222430.
• Regulatory pathway requirements (e.g., 505(b)(2) for reliance on external data) ³¹.
• Need for further methodological development and consensus on appropriateness ²⁹.

---

5. Conflicting Interpretations

There is general agreement across sources that while RWE/RWD is increasingly considered in regulatory submissions and may be referenced as supportive or confirmatory evidence (especially in rare diseases), the FDA rarely includes it as primary evidence in Prescribing Information due to methodological concerns. Some sources note that details may be redacted or missing, and thus there could be additional examples not captured in the available documents ³⁵.

---

Conclusion

RWE/RWD is most often used as supportive or confirmatory evidence in FDA drug reviews, particularly for rare diseases or when traditional trials are challenging. However, inclusion in Prescribing Information as primary evidence is rare and subject to strict methodological standards. The FDA typically cites concerns about study design, comparability, lack of pre-specified protocols, and potential for bias as reasons for exclusion.

---

References

^{12345678910111213141516171819202122232425262728293031}
(Numbers correspond to the merged citations from all provided answers.)