Unscheduled bleeding and endometrial cancer in women on postmenopausal hormone replacement therapy and their matched controls: protocol for a descriptive cohort study using the Orchid-e database

This paper outlines the protocol for a descriptive cohort study using the ORCHID-e database to investigate the diagnostic accuracy of unscheduled bleeding for endometrial cancer in postmenopausal women taking hormone replacement therapy compared to matched controls, aiming to inform guidelines on unnecessary referrals and invasive testing.

The Gist

The Big Picture: A "Traffic Light" Problem

Imagine your body is a car. When a woman reaches menopause (usually between ages 45 and 55), the engine stops making a specific fuel called estrogen. Without this fuel, the car might start sputtering, shaking, or overheating (symptoms like hot flashes, mood swings, and sleep trouble).

To fix this, many women use Hormone Replacement Therapy (HRT), which is like adding a synthetic fuel back into the tank to smooth out the ride.

The Problem:
Sometimes, when you add this new fuel, the car makes a weird noise: unscheduled bleeding.

• The Good News: This noise is often just the engine adjusting to the new fuel. It's a common side effect, especially when you first start taking the medicine.
• The Scary News: That same noise can also be the "check engine" light for womb cancer.

Currently, the rulebook (medical guidelines) says: "If you hear this noise after menopause, you must immediately go to the specialist for a deep inspection."

The Dilemma:
Because so many women are now taking HRT, and because HRT often causes this "noise" as a side effect, doctors are sending thousands of women for scary, invasive, and expensive cancer tests. Many of these women don't have cancer; they just have a reaction to their medicine. It's like calling a tow truck every time your car makes a weird noise, even if it's just a loose belt.

The Goal:
This study wants to figure out: Is the current rulebook too strict for women on HRT? They want to know if we can create a new, smarter rulebook that tells doctors when to panic and when to just wait and see.

---

How They Are Going to Find the Answer

The researchers are acting like detectives using a massive, anonymous digital library called the ORCHID-e database. This library contains the medical records of over 34 million people in England, but no one's name is attached to the files.

1. The Setup: The "Twin" Study

They aren't just looking at women on HRT. They are setting up a match-up game:

• Group A (The Drivers): Women aged 40+ who just started taking HRT.
• Group B (The Passengers): Women of the exact same age who are not taking HRT.

They pair these women up like twins. This ensures that if Group A has more bleeding than Group B, it's likely because of the HRT, not because Group A is older or lives in a different neighborhood.

2. The Investigation: What Are They Tracking?

The researchers will follow these women over time (from 2015 to 2025) and watch for three specific events:

1. The Noise: Did the woman go to her GP complaining of unscheduled bleeding?
2. The Tow Truck: Did the doctor immediately send her for an urgent cancer test (a referral)?
3. The Diagnosis: Did she actually have womb cancer found within a year?

3. The "Traffic Light" Analysis

They are trying to build a new traffic light system for doctors:

• Red Light: If a woman on HRT bleeds, does she really need a cancer test immediately? Or is it just a side effect?
• Yellow Light: Maybe we should wait 6 months to see if the bleeding stops before sending her for tests.
• Green Light: If she has other risk factors (like being very overweight or having a family history of cancer), maybe the rules should be stricter.

They will look at different "types" of drivers:

• Does it matter if she is 45 or 75?
• Does it matter if she has been on HRT for 1 month or 5 years?
• Does the type of HRT (pills vs. patches) change the risk?

Why This Matters to You

Right now, the medical system is playing it safe by over-testing. This causes stress, pain, and wasted money for women who are perfectly healthy.

By using this massive database, the researchers hope to prove that not all bleeding is an emergency. They want to give doctors a "cheat sheet" so they can:

• Stop sending women on HRT for unnecessary, scary tests.
• Still catch the real cancer cases quickly.
• Create guidelines that actually fit the modern reality of how many women are taking HRT.

The Bottom Line

Think of this study as a quality control check on the medical rulebook. The researchers are asking: "Are we treating every 'check engine' light as a total engine failure, when sometimes it's just a loose belt?"

If they find the answer, it could save thousands of women from unnecessary anxiety and invasive procedures, while still keeping everyone safe from cancer.

---

Note: This document is a "protocol," which means it is the plan for the study, not the final results. It's the blueprint before the house is built.

Technical Summary

Based on the provided document, here is a detailed technical summary of the study protocol. Note that as this is a protocol paper (a plan for research), it outlines the study design and objectives but does not yet contain final results.

1. Problem Statement

• Clinical Context: Menopause often requires Hormone Replacement Therapy (HRT) to manage symptoms. Since 2015, HRT usage in the UK has increased significantly, with nearly 20% of women aged 45–55 taking it by 2024.
• The Conflict: Unscheduled bleeding (postmenopausal bleeding) is a primary symptom of endometrial (womb) cancer. Current guidelines mandate an urgent referral for suspected cancer for any woman experiencing this. However, unscheduled bleeding is also a common, benign side effect of HRT, particularly in the first few months of use or after regimen changes.
• The Gap: Current guidelines were established when HRT usage was lower. There is a lack of contemporary data to determine if these guidelines are causing unnecessary, invasive, and distressing investigations (e.g., hysteroscopy) for women on HRT who are not at high risk of cancer. Furthermore, the risk profile for women starting HRT at older ages (60+) or on higher doses is poorly understood.
• Objective: To determine if current referral pathways are appropriate for women on HRT by quantifying the incidence of unscheduled bleeding, subsequent referrals, and cancer diagnoses in this specific population compared to non-users.

2. Methodology

• Study Design: A matched descriptive cohort study using a retrospective analysis of anonymized electronic health records (EHR).
• Data Source: The ORCHID-e database, containing ~34.3 million anonymized EHRs from 1,186 GP practices across England, designed to be representative of regions, ethnicity, and socio-economic status.
• Study Period: January 1, 2015 – March 2025.
• Study Population:
  • Inclusion: Women aged ≥40 years, registered at a GP practice for at least one year.
  • Exposed Group (HRT Users): Women initiating a prescription for an oestrogen-containing HRT product (with or without progestogen) during the study period.
    • Index Date: Date of the first new HRT prescription.
    • Exclusion: Any oestrogen HRT prescription in the 12 months prior to the index date (to ensure "new user" status).
  • Control Group (Unexposed): Women of the same age who have not initiated HRT.
  • Matching: Each exposed woman is matched to up to four unexposed women based on age at the index date. The unexposed group's index date is set to the matched exposed woman's index date.
  • Exclusion Criteria: Prior diagnosis of endometrial cancer, history of hysterectomy, or prior oestrogen HRT use.
• Exposure Definition: Initiation of HRT containing oestrogen. Progestogen prescriptions (including IUS) are extracted for description and sensitivity analysis but are not strictly required for entry, acknowledging potential under-recording of IUS in primary care.
• Outcomes:
  1. Unscheduled Bleeding: Defined by a specific clinical code for postmenopausal bleeding.
  2. Urgent Referral: Defined as an urgent referral for suspected gynaecological cancer occurring within 7 days of a bleeding code.
  3. Endometrial Cancer Diagnosis: Confirmed diagnosis within one year of the referral/bleeding event.
• Statistical Analysis:
  • Calculation of incidence proportions and adjusted hazard ratios (HR) for bleeding, referrals, and cancer diagnoses.
  • Stratification by time since index (6-month intervals), age, calendar year, socio-demographics (IMD, ethnicity, rurality), and HRT type (oral/transdermal, sequential/combined).
  • Diagnostic Accuracy: Calculation of sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) of unscheduled bleeding as a predictor of cancer for both HRT users and non-users.

3. Key Contributions (Planned)

• Evidence-Based Guidelines: The study aims to provide the first large-scale, contemporary data on the relationship between HRT use, unscheduled bleeding, and endometrial cancer risk in the UK.
• Refinement of Referral Pathways: By quantifying the "false positive" rate of bleeding in HRT users, the study will inform whether current urgent referral guidelines need to be stratified based on HRT status, timing of initiation, and regimen type.
• Diagnostic Accuracy Metrics: It will establish the predictive value of unscheduled bleeding for cancer specifically within the HRT population, a group currently under-represented in diagnostic accuracy studies.
• Subgroup Analysis: The study will investigate risk variations in older women (60+), those on higher doses, and those with specific risk factors (e.g., obesity, diabetes, Lynch syndrome).

4. Results

• Status: No results are available yet.
• Feasibility Estimates: The authors estimate the cohort will include approximately 400,000 new HRT users. Based on historical data, they anticipate identifying:
  • ~343 endometrial cancer diagnoses.
  • ~3,811 urgent referrals.
  • ~7,622 consultations for postmenopausal bleeding.
• These projected numbers are deemed sufficient to calculate statistically robust proportions and hazard ratios for various subgroups.

5. Significance

• Clinical Impact: The findings could prevent thousands of women from undergoing unnecessary, invasive, and anxiety-inducing cancer investigations (hysteroscopy) if the data shows that bleeding in HRT users has a low predictive value for cancer. Conversely, it could refine guidelines to ensure high-risk women are not missed.
• Public Health: With HRT usage rising, optimizing the diagnostic pathway is crucial for efficient resource allocation within the NHS and reducing patient burden.
• Policy: The study is funded by the NIHR Policy Research Programme and involves Patient and Public Involvement and Engagement (PPIE), ensuring the research questions and dissemination strategies align with patient experiences and healthcare policy needs.
• Limitations Acknowledged: The study relies on primary care data, meaning private prescriptions or secondary care diagnoses might be missed. It also cannot perfectly distinguish between partial and total hysterectomies or track dose changes over time with absolute precision.

Conclusion: This protocol outlines a rigorous, large-scale observational study designed to resolve a critical uncertainty in menopause management: balancing the need to detect endometrial cancer early against the risk of over-investigating benign HRT side effects. The results are expected to directly influence national clinical guidelines for postmenopausal bleeding.