As of this morning, every claim with a date of service on or after October 1, 2024 is judged against the FY2025 ICD-10-CM code set. CMS and the National Center for Health Statistics published the update in June, and it brings 252 new codes, 36 deletions and 13 revisions. Compared with the FY2024 update, which added 395 codes, this is a smaller year. It still contains enough deletions to fill a rejection report if the practice is still coding from last year's favorites.
We reviewed the addenda with the practices we support over the past month. Below is what we think a working coder, biller or practice manager needs to know today, and what to check between now and Friday. The short version: the additions are useful and mostly optional, the deletions are few and mandatory, and the date-of-service logic in your billing system matters more this week than the codes themselves.
Key takeaways
- The FY2025 set applies by date of service, not submission date. September visits keep FY2024 codes no matter when the claim goes out.
- The additions most practices will use are the obesity class codes (E66.811 to E66.813), hypoglycemia levels (E16.A1 to E16.A3), 63 lymphoma-in-remission codes, and neurology additions including KCNQ2-related epilepsy and serotonin syndrome.
- All 36 deletions are expansions: the old code becomes a parent and the claim needs one of the new children.
- Confirm three systems loaded the set (EHR, practice management, clearinghouse), fix favorites and order sets, and read the rejection report daily through mid-October.
The additions most practices will use
The new codes cluster in a few chapters. The ones that come up in ordinary primary care and internal medicine work are below, with the codes a coder will actually type.
| Area | New codes | What changes | Who uses them |
|---|---|---|---|
| Obesity by class | E66.811 (class 1), E66.812 (class 2), E66.813 (class 3) | The diagnosis code now carries the severity the BMI Z code only implied | Primary care, endocrinology, bariatric programs |
| Hypoglycemia by level | E16.A1 (level 1), E16.A2 (level 2), E16.A3 (level 3) | Distinguishes a mild reading from a severe event with altered mental status | Endocrinology, primary care, urgent care |
| Lymphoma in remission | 63 new codes with a fifth character A, for example C81.0A to C81.9A for Hodgkin lymphoma | Ends the choice between an active cancer code and a personal history code | Oncology, primary care managing survivors |
| KCNQ2-related epilepsy | G40.84 with children G40.841 to G40.844 | Specifies intractable or not, with or without status epilepticus | Neurology, pediatrics |
| Serotonin syndrome | G90.81 | A specific code for a diagnosis that previously used a general autonomic disorder code | Emergency, hospital medicine, psychiatry |
The obesity codes are the ones we expect to matter most in primary care. Practices that manage weight or bill for obesity counseling should switch to the class codes where the documentation supports the class. The class has to be in the note, not inferred from the BMI by the coder; a BMI of 41 with no statement of class does not support E66.813 on its own. That is a template change for providers, and today is the day to make it.
The full addenda also touches injury, external cause and Z codes. Specialty practices should read their own chapter; the specialty societies have published summaries over the last several weeks, and the CDC's addenda file lists every change with the old and new descriptions side by side.
The deletions that cause this week's rejections
Deletions in ICD-10-CM are almost always expansions: a code that used to be billable becomes a parent and gains more specific children. The billable code disappears and the claim rejects at the clearinghouse or the payer with a message like "diagnosis code invalid for date of service". Every one of the 36 deletions works this way, and the fix is the same each time: find the parent on the addenda, pick the child that matches the documentation, and update the favorites list so the provider does not choose the old code tomorrow.
Two things make this worse than it needs to be. Some EHRs continue to display a deleted code in search results for weeks after the effective date, because the code remains valid for older encounters. And some providers type codes from memory. Neither is fixed by the vendor update alone; both are fixed by a one-page sheet per provider showing the codes they used that are gone and the replacement for each.
Who this affects, and how much
| Role | What changes today | Effort this week |
|---|---|---|
| Providers | New, more specific options for obesity, hypoglycemia, lymphoma remission and several neurological conditions; some familiar codes gone; class must be documented | 15 minutes to review the updated favorites list and the one-page sheet |
| Coders | Addenda review against the practice's top 200 diagnoses; order set and charge rule check | Two to four hours |
| Billers | Daily rejection review for invalid diagnosis codes; corrections resubmitted within 48 hours | 30 minutes a day for two weeks |
| Practice manager | Confirm the PM system, the EHR and the clearinghouse all loaded the FY2025 set | One hour, today |
| Quality or care management lead | Add new codes to registry criteria so patients are not dropped from panels | One hour this month |
Date of service is what matters, not submission date
A visit on September 30 coded on October 2 uses FY2024 codes. A visit on October 1 uses FY2025 codes. Practice management systems handle this with date-of-service logic, but only if the old codes were end-dated rather than deleted from the system. If your update removed the September codes entirely, you will see rejections on last week's visits that have nothing to do with the new set, and the fix is to restore the old codes with an end date of September 30. Check one late-September claim today to be sure.
The same logic runs the other way. If a provider documents a September 27 visit today and picks E66.813 from the new favorites list, the claim rejects because the code did not exist on the date of service. Late documentation across the boundary is a small but reliable source of first-week rejections, and the biller should know to look for it.
A worked example
A 62-year-old established patient is seen today for follow-up of type 2 diabetes with a documented hypoglycemic episode last week (glucose 52 mg/dL, no assistance required) and a BMI of 41, with the provider documenting class 3 obesity. Before today the encounter would have carried E11.649 (type 2 diabetes with hypoglycemia without coma), E66.01 (morbid obesity due to excess calories) and Z68.41 (BMI 40.0 to 44.9, adult). Under FY2025 the coder adds E16.A2, hypoglycemia level 2, because a glucose below 54 mg/dL without assistance meets the level 2 definition, and reports E66.813 for class 3 obesity alongside the BMI code. The claim tells the payer more, and it supports the chronic care management and obesity counseling work the practice is doing. If the provider had not written the class, the coder stays with E66.01 and queries the provider for next time.
What this means for quality reporting and risk adjustment
Practices in Medicare Advantage risk contracts or ACOs should look at the additions with their risk adjustment coder. More specific codes do not automatically change hierarchical condition category assignments, and the CMS-HCC model has its own mapping tables that lag the code set. Don't assume a new code carries a new HCC until the mapping is published. Do assume that a payer auditing 2025 charts will expect the specific code where the documentation supported it, and that a chart documenting class 3 obesity coded as E66.01 will look like a coding gap in a retrospective review.
Quality registries have the same issue in reverse. A diabetes or hypertension registry built on a list of diagnosis codes will not pick up patients coded with a new child code until the list is updated. The quality lead should add the new codes to the registry criteria this month, before the fourth-quarter measure rates are pulled.
Questions we hear
Do we need to recode encounters from last week?
No. September dates of service use the FY2024 code set regardless of when the claim goes out. Recoding them to FY2025 codes would create rejections, not prevent them.
Are payers ready on day one?
Medicare and most large commercial payers load the update on time. Smaller plans and some Medicaid managed care organizations occasionally lag by a week or two and reject valid new codes. If a new code rejects at one payer and not others, hold that payer's claims for a few days and call, rather than changing the code to something less accurate. Keep a note of which payer and which code; the same plan often lags again in January.
How is a "revised" code different from a deleted one?
A revised code keeps its number and gets a new description. The claim does not reject, but the code may no longer mean what your provider thinks it means. Review the 13 revisions with the same care as the deletions, and update the description text in the favorites list so the provider sees the new wording.
What to do this week
- Confirm the FY2025 set is loaded in the EHR, the practice management system and the clearinghouse. Three systems, three confirmations, today.
- Run the practice's top 200 diagnoses against the deletion list. Replace every deleted code in favorites, superbills and order mappings.
- Send each provider a one-page sheet: the deleted codes they used and the replacement for each, plus the new obesity and hypoglycemia codes with the documentation each needs.
- Check one September claim submitted this week to confirm date-of-service logic is working in both directions.
- Read the rejection report every morning through October 15. Sort by reason. Fix the source of each invalid-code rejection, not only the claim.
- Add the new codes your practice will use to favorites and, where relevant, to care management and quality registries.
- If you want a coder to run the addenda against your actual charge data, ask about a scoped review through the Revelrex medical coding service; the code-set walkthrough is also part of the winter session of our RCM training courses.
