Alcohol dependence ICD-10 codes: F10.2x explained

In this article▾

1. What is the alcohol dependence ICD-10 code?
2. What do the complication digits actually mean?
3. Why is a CIWA score not enough?
4. Common pitfalls
5. How does F10.20 support a detox authorization?
6. What is the difference between F10.20 and F10.10 on a claim?
7. When does alcohol dependence become in remission?
8. Which alcohol dependence codes get denied most?
9. How does the F10 code connect to your level of care?
10. What does an auditor look for on F10.2 claims?
11. How the withdrawal documentation gap can be automated
12. FAQ

TL;DR: The alcohol dependence ICD-10 family is F10.2, and the digits after it decide what your claim can support. F10.20 is the plain version. The withdrawal codes underneath it are what actually carry a detox or residential authorization, and most facilities document the withdrawal in a flowsheet instead of a diagnosis.

What this Blog covers

• What is the alcohol dependence ICD-10 code
• What the complication digits actually mean
• Why a CIWA score isn't enough
• Common pitfalls
• Difference between F10.20 and F10.10 on a claim
• Which alcohol dependence codes get denied most

Your client came in shaking. Their CIWA peaked at 18 overnight, nursing gave scheduled lorazepam, and by day three they were sleeping and eating. Textbook managed withdrawal, handled well.

Now open the chart and look at the diagnosis field. There's a fair chance it says F10.20, alcohol dependence uncomplicated, which describes a stable dependent person and says nothing about the last seventy-two hours.

That gap isn't a knowledge problem. Everyone involved knew what was happening. It's that withdrawal got recorded where clinicians record observations and never made it to where the diagnosis lives.

What is the alcohol dependence ICD-10 code?

F10.2, with a fourth and sometimes fifth digit describing what else is going on. F10.20 is uncomplicated, F10.21 is in remission, and the F10.23 series covers withdrawal. Intoxication and induced conditions have their own digits.

The .2 tells anyone reading that this is the severe end of the F10 family (American Psychological Association Services). Everything after it describes the clinical event you were managing.

What do the complication digits actually mean?

They say what the dependence is doing right now. The fourth digit sets the category, the fifth adds severity within it.

F10.239 deserves a note. It exists for cases where withdrawal happened but nobody characterized it, and it sits below the three codes above it while still beating F10.20. If your team uses it often, the gap is in what gets written rather than in what gets known.

Why is a CIWA score not enough?

Because it's data, not a diagnosis, and the distinction matters to everyone who reads the chart after you.

A flowsheet showing CIWA 18 tells a reader that someone measured something. It doesn't say who concluded the client was in withdrawal, what they thought was driving it, or what they did about it. Those are clinical judgments, and clinical judgments belong in your note.

Here's the sentence that closes the gap:

"Alcohol dependence with withdrawal. CIWA peaked at 18 on day one, treated with scheduled lorazepam per protocol, tapering by day three. No delirium or perceptual disturbance observed."

Thirty seconds. It supports F10.230, it rules out the two more severe codes explicitly, and it means nobody calls you in six weeks asking what happened.

Compare it to the version that usually appears, which is "client monitored for withdrawal, tolerating well." True, kind, and it supports nothing.

Common pitfalls

Letting the nursing record carry the diagnosis. Vitals and scores are collected diligently and live in a part of the chart that doesn't feed the diagnosis field. Someone has to move the conclusion across, and that someone is you.

Coding both abuse and dependence. When one chart documents both for the same substance, only the dependence code is assigned (American Medical Coding). This happens most often when a psychiatric evaluation says abuse in the history and dependence in the assessment. Code the assessment.

Coding a complication you didn't document. F10.231 asserts delirium. If the chart doesn't show delirium, that code is a compliance exposure rather than a stronger claim.

Leaving the admission code in place all the way to discharge. By day five your client isn't in withdrawal anymore. A code frozen at admission describes a person who no longer exists.

Assuming remission because things are going well. More on that next.

How does F10.20 support a detox authorization?

On its own, weakly. F10.20 says the patient is dependent on alcohol. It does not say they are in withdrawal, at risk of seizure, or requiring medication and monitoring, which are the facts a payer uses to approve a detox day.

Withdrawal changes the argument. F10.230 tells a reviewer the patient is actively withdrawing under your care, and F10.231 tells them the withdrawal produced delirium. Those codes describe an acute medical event, and acute medical events are what medically monitored settings exist for.

Insurers denied 19% of in-network marketplace claims in 2024, and only 5% of those denials were formally categorized as medical necessity while 36% were logged as "other" (KFF). Do not read the small medical necessity number as reassurance. The reason field is a payer's categorization choice, and weak clinical justification hides across several of those buckets.

The documentation test is whether a stranger reading the chart could tell the patient was withdrawing without seeing the vitals. A CIWA score alone fails that test. A physician assessment naming withdrawal, describing its severity, and recording the response to treatment passes it.

What is the difference between F10.20 and F10.10 on a claim?

One digit, and a substantial difference in what the claim can carry. F10.10 is abuse and F10.20 is dependence, so the second describes a more severe pattern and supports more intensive care.

The coding hierarchy settles the case where both appear in one chart. When a record documents abuse and dependence for the same substance, only the dependence code goes on the claim (American Medical Coding). Sending both is not thorough, it is contradictory, and a reviewer will read it as a chart that cannot decide.

This comes up constantly in psychiatric evaluations, where the history section says abuse and the assessment says dependence. The history is describing a past pattern. The assessment is the diagnosis. Code the assessment.

When does alcohol dependence become in remission?

When you say so in writing. The ICD-10-CM guidelines hold that remission codes cannot be assigned without clear documentation in the record by the provider, specifying that the disorder is in remission (ACDIS).

Nobody downstream can build remission out of good attendance and clean screens, however convincing that pattern looks from a spreadsheet. That's a protection, not an obstacle. Remission is a judgment about a whole person and the code set says it's yours to make.

Remission also differs from a history-of code, which describes something no longer treated or monitored (ACDIS). Anyone in your continuing care group is being monitored, so history-of does not apply to them.

There's a consequence in both directions worth sitting with. Coding remission early can undercut the necessity of services your client still needs. Coding it late misrepresents someone who has genuinely recovered, and that record follows them.

Which alcohol dependence codes get denied most?

The unspecified ones, and the ones that outrun their documentation. F10.20 on a claim for a level of care that implies acuity is the most common pattern, because the code is accurate and insufficient at the same time.

Run that check on the F10.20 claims specifically, since they are the ones most likely to have been coded by default. The claim goes out, clears every automated check, and gets read weeks later by someone deciding whether the days you provided were justified. That review sits inside the wider revenue cycle your organization runs.

If you want to see how those denials land on the other side of the building, our note on what a medical necessity denial actually means walks through it.

The pattern that usually appears is a set of detox and residential days billed under an uncomplicated code while the chart describes an actively withdrawing patient.

Fewer than 1% of denied claims were appealed in 2024, and 66% of the appeals filed were upheld in the insurer's favor (KFF). The economics point hard at prevention. An appeal is expensive, slow, and usually unsuccessful, while a documentation fix is none of those.

How does the F10 code connect to your level of care?

Directly, in the eyes of a utilization reviewer. The ASAM criteria that most payers use to evaluate residential and detox authorizations weigh withdrawal potential heavily, and your diagnosis code is the shorthand version of that assessment.

A patient coded F10.230 has withdrawal in their diagnosis. A patient coded F10.20 does not, whatever the flowsheet says. When a reviewer is deciding whether a medically monitored level of care was appropriate, the code that arrives with the claim frames the whole conversation.

None of that is a coding rule. It is how the code gets read once it reaches someone deciding whether to pay for a bed, which is the only reading that determines your revenue.

What does an auditor look for on F10.2 claims?

Consistency between the code, the note, and the orders. An auditor is not checking whether F10.231 is a valid code. They are checking whether the patient described in the chart is the patient described by the code.

Three questions come up repeatedly. Does a provider note the name of the withdrawal, or does it appear only in nursing documentation? Do the medication orders match the severity coded? Does the diagnosis change across the stay as the clinical picture changes, or does it sit frozen from admission to discharge?

That third one catches more facilities than the other two combined. A diagnosis that never moves suggests it was set once at intake and never revisited, which is usually exactly what happened.

How the withdrawal documentation gap can be automated

The failure here is almost never clinical. Withdrawal gets recognized, treated well, and written in the wrong place, so the diagnosis on the claim describes a calmer patient than the one your staff actually managed.

Supa runs inside the systems you already use. It reads the flowsheet and the assessment together, so a chart with rising CIWA scores and an uncomplicated dependence code surfaces while the record is still open. It flags diagnoses that haven't moved across an admission, and codes that no longer match the clinical picture on the page.

You keep making the call. What disappears is the version where you managed a complicated withdrawal beautifully and the chart described a quiet week.

If your clinicians are answering records requests about admissions from last quarter, this is where that starts. Book a demo and we will pull a few charts with you.

FAQ

Q: What is the ICD-10 code for alcohol dependence?

A: F10.20 is alcohol dependence uncomplicated. The wider F10.2 family adds digits for remission, intoxication, withdrawal, and induced conditions, so F10.230 covers dependence with uncomplicated withdrawal and F10.21 covers dependence in remission.

Q: What is the difference between F10.20 and F10.230?

A: F10.20 says the patient is dependent. F10.230 says they are dependent and actively withdrawing. The second describes an acute event that justifies medical monitoring, which makes it a stronger code for detox and residential claims when the chart supports it.

Q: Can I code alcohol dependence and alcohol abuse together?

A: No. When one chart documents both for the same substance, the coding hierarchy assigns only the dependence code. Reporting both reads as an internal contradiction and invites review rather than demonstrating thoroughness.

Q: How long does someone need to be sober before I can code F10.21?

A: There is no coder-facing time threshold. Remission is coded when the provider documents that the disorder is in remission. Duration is a clinical judgment the provider makes and records, not a rule a billing team can apply on its own.

Q: Is F10.239 an acceptable code to bill?

A: It is valid, and it is the weakest of the withdrawal codes because it leaves severity uncharacterized. If your team uses it often, the underlying problem is usually that notes record withdrawal without describing whether delirium or perceptual disturbance was present.

Q: Why do my F10.20 claims deny more than my F10.23 claims?

A: Because F10.20 describes a stable dependent patient, which is a hard fit with detox or residential intensity. Reviewers approve acuity, and the withdrawal codes describe acuity while the uncomplicated code does not.