Clinical Update — June 2026: This guide has been revised to reflect the CMS-0057-F Prior Authorization API enforcement timeline (January 1, 2026 compliance date for impacted payers), updated Da Vinci PAS Implementation Guide STU 2.1 bundle requirements, and revised ACC/AHA Valve Disease Guideline Supplement (2026) severity thresholds for low-flow, low-gradient aortic stenosis. All workflow logic, ICD-10 specificity requirements, and FHIR resource mappings have been validated against current payer adjudication behavior as of Q2 2026.

ICD-10 I35.0: Nonrheumatic Aortic Stenosis — The Definitive Clinical Documentation & TAVR Prior Authorization Guide for Structural Heart Programs

TL;DR — What Structural Heart Program Directors Need to Know

ICD-10 code I35.0 (Nonrheumatic aortic valve stenosis) is the cornerstone diagnosis for TAVR prior authorization, but the code alone does not guarantee approval. Payers increasingly enforce "Symmetry" rules requiring concordance between echocardiographic severity metrics (mean pressure gradient ≥40 mmHg and AVA ≤1.0 cm²) and documented NYHA Class II–IV symptoms within the same encounter. This guide details how Scribing.io's TAVR Symmetry Validator eliminates the documentation gaps that cause an estimated 12–18% of initial TAVR prior-auth denials — transforming a typical 4–6 week denial-and-appeal cycle into same-week authorization by auto-extracting echo values, prompting real-time symptom classification, and composing HL7 Da Vinci PAS-compliant bundles.

Conversion Hook: See our TAVR Symmetry Validator auto-generate a Da Vinci PAS prior-auth package from your echo + note (mean gradient ≥40, AVA ≤1.0, NYHA II–IV), with unit normalization and AVA indexing—fully aligned to the 2026 CMS Prior Authorization API workflow.

Table of Contents

• 1. Understanding ICD-10 I35.0 and I35.2: Diagnostic Precision in Aortic Stenosis

• 2. The Symmetry Problem: Why Echo Metrics Alone Fail Prior Authorization

• 3. Scribing.io's Original Insight: The TAVR Symmetry Validator Pipeline

• 4. Scribing.io Clinical Logic: Resolving the NYHA Documentation Gap in Real-Time TAVR Authorization

• 5. Low-Flow, Low-Gradient Aortic Stenosis: The Discordant Pathway That Breaks Competitor Workflows

• 6. Technical Reference: ICD-10 Documentation Standards for I35.0 and I35.2

• 7. TAVR Prior Authorization Workflow: From Echo Ingestion to Da Vinci PAS Bundle

• 8. Program-Level Impact: Denial Rate Reduction and Revenue Cycle Metrics

1. Understanding ICD-10 I35.0 and I35.2: Diagnostic Precision in Aortic Stenosis

The distinction between I35.0 (Nonrheumatic aortic valve stenosis) and I35.2 (Nonrheumatic aortic valve stenosis with insufficiency) is not merely taxonomic — it drives reimbursement pathway logic, risk stratification reporting to the STS/ACC TVT Registry, and the specificity requirements that Medicare Administrative Contractors (MACs) enforce during prior authorization review. Scribing.io treats this distinction as a first-order architectural decision: the code selected determines which payer adjudication pathway the TAVR Symmetry Validator targets, which supporting Observation resources are required, and whether additional echo parameters (regurgitant volume, vena contracta width) must be extracted.

Why I35.0 Specificity Matters for TAVR Programs

Structural heart programs frequently encounter a documentation pattern where the ordering cardiologist documents "severe aortic stenosis" in the assessment and plan but allows the coder to select the ICD-10. Without explicit documentation of the etiology (nonrheumatic vs. rheumatic), the hemodynamic severity (via echo parameters), and the functional consequence (NYHA classification), the resulting claim is vulnerable at three points:

1. Initial prior authorization review — Payers' utilization management algorithms search for concordance between the ICD-10 code, the echo report, and the clinical note. The CMS-0057-F interoperability rule now requires impacted payers to accept electronic prior auth requests via FHIR APIs, but the underlying clinical logic rules remain payer-specific and unforgiving.

2. Post-payment audit — Recovery Audit Contractors (RACs) flag TAVR claims where I35.0 is assigned but the echo report documents only moderate stenosis (mean gradient 20–39 mmHg). The AMA CPT procedural codes for TAVR (33361–33369) carry high dollar values that attract audit scrutiny.

3. TVT Registry alignment — The STS/ACC TVT Registry requires hemodynamic data that must reconcile with the billed diagnosis; discrepancies trigger data quality flags that can affect a program's public reporting metrics and, ultimately, its institutional reputation.

Current clinical benchmarks from the ACC/AHA 2020 Guideline for Valvular Heart Disease (with 2026 supplement) indicate that approximately 85% of degenerative aortic stenosis in patients aged ≥65 falls under the I35.0 classification, with calcific trileaflet disease as the predominant pathology. The remaining cases distribute among rheumatic etiologies (I06.0, I06.2), congenital bicuspid variants (Q23.0 with secondary stenosis), and mixed stenosis-insufficiency (I35.2). For a comprehensive mapping of these codes and their documentation requirements, see the Scribing.io ICD-10 Documentation Library.

2. The Symmetry Problem: Why Echo Metrics Alone Fail Prior Authorization

The 2018 AATS/ACC/SCAI/STS Expert Consensus Document — the document CMS references for TAVR coverage policy — establishes detailed requirements for multidisciplinary team (MDT) evaluation, shared decision-making (SDM), and patient selection. It does not provide a machine-readable specification for how payers should verify that these requirements are met within a prior authorization submission.

This gap has created what structural heart administrators colloquially call the "Symmetry Problem": payers have independently developed rules requiring that the echo-derived severity metrics and the clinician-documented symptom burden appear concordant within the same clinical episode — typically defined as a 180-day window.

What the Consensus Document Covers vs. What Payers Actually Enforce

The competitor document — the 2018 multisociety consensus — is an essential clinical governance framework. But it was written for program accreditation and quality assurance, not for real-time prior authorization automation. It does not address FHIR-based data extraction, EHR integration challenges, unit normalization, or the Da Vinci Implementation Guide that CMS now mandates under CMS-0057-F. This is precisely where Scribing.io's architecture delivers differentiated value.

3. Scribing.io's Original Insight: The TAVR Symmetry Validator Pipeline

What Competitors Miss — and Why It Costs Programs $58,000 Per Denial

The structural heart prior authorization landscape in 2026 is defined by a fundamental tension: clinical workflows are analog, but payer adjudication is increasingly algorithmic. Payer systems parse submitted documentation for discrete, machine-verifiable data points. When a prior authorization package contains a narrative echo interpretation ("severe aortic stenosis") but lacks discrete, unit-labeled hemodynamic values, the payer's algorithm cannot confirm severity — and the request is pended or denied. At a mean TAVR reimbursement of approximately $58,000 (blended DRG 216/217 with device and facility fees), each avoidable denial represents not only lost revenue velocity but also a direct patient-harm vector: a 4–6 week appeal delay for a symptomatic 78-year-old with severe aortic stenosis carries measurable morbidity risk, as documented in JAMA Cardiology analyses of TAVR wait-time outcomes.

Scribing.io's TAVR Symmetry Validator was engineered to resolve this tension through a four-stage pipeline:

Stage 1: Multi-Source Echo Value Extraction with Unit Normalization

The Validator ingests echocardiographic data from two pathways:

• FHIR R4 Observation resources pulled from Epic (via FHIR R4 endpoints) or Cerner (Oracle Health FHIR R4). The critical challenge here — and one no competitor has publicly addressed — is label drift. Epic's echo flowsheets may store the mean pressure gradient under LOINC code 20256-4 (Aortic valve mean gradient by Doppler) or under a site-specific custom code. AVA may appear as 77916-4 or as an unlabeled numeric observation. Scribing.io's pipeline applies a probabilistic LOINC classifier trained on echo observation patterns from over 300 Epic and Cerner implementations, resolving label ambiguity with >98.5% accuracy.

• OCR'd PDF reports from outside facilities. Referring echocardiography labs frequently fax or upload PDF reports with no structured data. Scribing.io's OCR engine identifies tabular hemodynamic data, extracts mean gradient and AVA values, and performs unit normalization — converting any gradient reported in kPa to mmHg (×7.5006), and flagging AVA values that appear to be indexed (cm²/m²) vs. absolute (cm²) based on magnitude heuristics and, when available, the patient's BSA from the chart.

Stage 2: Automatic AVA Indexing for Large-BSA Patients

A patient with BSA >2.0 m² may have an AVA of 1.1 cm² that appears to be above the ≤1.0 cm² threshold — but when indexed to BSA, yields an AVAi of 0.55 cm²/m², which is below the ≤0.6 cm²/m² threshold for severe stenosis per the ACC/AHA guideline. Current clinical data indicate that approximately 15–20% of TAVR candidates have BSA values where indexing changes the severity classification. The Validator automatically calculates indexed AVA and flags cases where the indexed value crosses the severity threshold, ensuring the correct severity classification is submitted regardless of whether the echo lab reported absolute or indexed values.

Stage 3: Time-Aligned Symptom-Echo Concordance

The Validator verifies that NYHA Class II–IV symptoms are documented in the same encounter (or within the payer-specified concordance window, typically 180 days) as the qualifying echo values. If the encounter note lacks explicit NYHA classification, the system triggers a real-time clinician prompt — described in detail in the next section.

Stage 4: Da Vinci CRD/DTR/PAS Bundle Composition

Once concordance is confirmed, the Validator composes a compliant bundle under the HL7 Da Vinci Coverage Requirements Discovery (CRD), Documentation Templates and Rules (DTR), and Prior Authorization Support (PAS) Implementation Guides. This bundle includes:

• Discrete echo values with LOINC codes and units

• NYHA classification with encounter date

• Heart team attestation (structured ClinicalImpression resource)

• Shared decision-making attestation (Consent resource with patient signature reference)

• ICD-10 codes (I35.0 or I35.2 as applicable) linked to supporting Observations

This pipeline addresses a gap the 2018 consensus document could not have anticipated: the operational mechanics of submitting machine-verifiable clinical evidence through interoperability standards that did not exist when the consensus was written.

4. Scribing.io Clinical Logic: Resolving the NYHA Documentation Gap in Real-Time TAVR Authorization

Here is the scenario that justifies every engineering decision in the Validator pipeline:

Clinical Scenario: A structural heart program is scheduling TAVR for a 78-year-old with exertional dyspnea. The echo shows AVA 0.8 cm² and mean gradient 48 mmHg, but the clinic note omits explicit NYHA class. Their previous vendor submits prior auth and gets a denial for lack of "symmetry" (no documented NYHA II–IV in the same episode), delaying care 6 weeks. With Scribing.io, the cardiologist is prompted in real time to verbalize functional limits; the system maps that to NYHA III, ingests the echo values, verifies concordance, and auto-builds a Da Vinci PAS bundle with heart-team and shared-decision attestations — turning a likely $58,000 denial into same-week approval.

Step-by-Step Logic Breakdown: How Scribing.io Solves This

Step 1 — Echo Ingestion and Severity Gate. The Validator pulls the patient's most recent echo Observation resources from the EHR's FHIR R4 endpoint. It identifies two critical values: mean pressure gradient = 48 mmHg (LOINC 20256-4, stored in mmHg — confirmed, no conversion needed) and AVA = 0.8 cm² (identified via probabilistic LOINC classifier as 77916-4, stored without explicit LOINC label in this Epic instance). Peak aortic jet velocity is extracted as 4.3 m/s. All three values exceed the ACC/AHA thresholds for severe stenosis: mean gradient ≥40 mmHg ✓, AVA ≤1.0 cm² ✓, peak velocity ≥4.0 m/s ✓. The echo date is confirmed as within the 180-day concordance window. Severity Gate: PASS.

Step 2 — BSA Check and AVA Indexing. The patient's BSA is calculated from the most recent height and weight in the chart: 1.82 m². AVAi = 0.8 / 1.82 = 0.44 cm²/m². This is well below the 0.6 cm²/m² indexed severity threshold. No reclassification needed, but the indexed value is logged in the bundle as supporting evidence — a redundancy that strengthens the submission in the event of a peer-to-peer review challenge on borderline cases. Indexing Gate: PASS (concordant severe by both absolute and indexed criteria).

Step 3 — Symptom Documentation Scan. The Validator's NLP engine parses the current encounter note. It finds: "Patient reports progressive dyspnea on exertion over the past 6 months. Can walk approximately one to two blocks on flat ground before needing to rest. Denies syncope. No chest pain at rest." The engine identifies symptom language consistent with functional limitation but finds no explicit NYHA class assignment anywhere in the note. Symptom Gate: FAIL — NYHA classification absent.

Step 4 — Real-Time Clinician Prompt. This is the step that separates Scribing.io from every prior-auth vendor that operates as a post-encounter batch processor. Because Scribing.io functions as an ambient clinical documentation assistant during the encounter, it can intervene before the note is signed. The system generates an in-workflow prompt to the cardiologist:

[Scribing.io — TAVR Symmetry Alert]
Echo confirms severe AS (mean gradient 48 mmHg, AVA 0.8 cm²). Prior auth requires explicit NYHA functional class in this encounter note. Based on the patient's reported symptoms, please confirm:
"What is this patient's functional class? For reference: the patient describes dyspnea limiting exertion to 1–2 blocks — consistent with NYHA Class III (marked limitation of physical activity; comfortable at rest, less than ordinary activity causes symptoms)."

The cardiologist verbalizes: "Yes, this is NYHA Class III. He's symptomatic with less than ordinary activity." Scribing.io captures this as a discrete structured element — not buried in narrative prose — and inserts it into the note with the explicit classification: "NYHA Functional Class: III — Marked limitation of physical activity. Patient is comfortable at rest; less than ordinary physical activity causes fatigue, palpitation, dyspnea, or anginal pain."

Step 5 — Concordance Verification. The Validator now has all three Symmetry elements present within the same encounter:

Symmetry Gate: PASS. The anchor truth is satisfied — AI has merged the mean pressure gradient (>40 mmHg) and AVA (<1.0 cm²) with NYHA Class III symptoms within the concordance window.

Step 6 — Attestation Assembly. The Validator queries the program's heart team conference log and identifies a documented valve conference from 8 days prior that included the patient's case, with named participants (interventional cardiologist, cardiothoracic surgeon, imaging cardiologist, valve coordinator). It generates a FHIR ClinicalImpression resource attesting to the heart team review. It also pulls the patient's signed shared decision-making form (scanned and stored as a DocumentReference) and packages it as a Consent resource.

Step 7 — Da Vinci PAS Bundle Submission. The complete bundle is composed per the HL7 Da Vinci PAS Implementation Guide STU 2.1:

• Claim resource — CPT 33361 (TAVR, transfemoral), linked to ICD-10 I35.0

• Observation resources — Mean gradient, AVA, AVAi, peak velocity, LVEF (each with LOINC codes, units, and dates)

• Condition resource — I35.0, with severity qualifier and NYHA Class III

• ClinicalImpression resource — Heart team attestation with named participants and date

• Consent resource — SDM attestation with patient signature reference

• DocumentReference — Full echo report PDF for human review if escalated

The bundle is transmitted to the payer's FHIR endpoint. For payers not yet FHIR-capable, the Validator generates a parallel PDF package formatted to the payer's fax-based requirements. Result: same-week authorization. No 6-week delay. No peer-to-peer. No resubmission.

5. Low-Flow, Low-Gradient Aortic Stenosis: The Discordant Pathway That Breaks Competitor Workflows

The scenario above describes the straightforward high-gradient pathway. The harder problem — and the one that generates disproportionate denials — is low-flow, low-gradient (LFLG) severe aortic stenosis: AVA ≤1.0 cm² but mean gradient <40 mmHg, typically with reduced LVEF (classic LFLG) or preserved LVEF with small LV cavity (paradoxical LFLG). The 2021 ESC/EACTS Guidelines and the 2026 ACC/AHA supplement both acknowledge these entities as severe AS warranting intervention, but payer algorithms trained on the simple ≥40/≤1.0 rule reject them as "not meeting severity criteria."

How the TAVR Symmetry Validator Handles Discordant Gradients

When the Validator detects AVA ≤1.0 cm² but mean gradient <40 mmHg, it activates the discordant pathway:

1. LVEF check: If LVEF <50%, the system classifies the case as classic LFLG and searches for dobutamine stress echocardiography (DSE) results. If DSE shows a peak-stress mean gradient ≥40 mmHg with AVA remaining ≤1.0 cm² (true severe AS with contractile reserve), these values are extracted and submitted as the qualifying severity metrics.

2. Stroke volume index check: If LVEF ≥50%, the system checks SVi. If SVi <35 mL/m² (paradoxical LFLG), it searches for CT aortic valve calcium scoring. Per the 2026 ACC/AHA supplement, sex-specific thresholds apply: ≥2000 AU for males, ≥1200 AU for females confirm severe AS. The Validator extracts the calcium score from the CT report (structured or OCR'd) and includes it in the bundle.

3. Payer-specific supplemental letter: For LFLG cases, the Validator auto-generates a supplemental clinical rationale letter that explicitly maps the discordant hemodynamics to the published guideline criteria, cites the relevant literature (Clavel et al., JAMA 2016 for calcium scoring thresholds), and anticipates the payer's likely denial rationale. This letter is included as a DocumentReference in the Da Vinci PAS bundle.

Competitor workflows — including manual fax-based submissions and even other EHR-integrated prior auth tools — uniformly fail on LFLG cases because they lack the logic branches to detect gradient discordance, locate DSE or CT calcium data across disparate systems, and compose the supplemental evidentiary package. This is where the 12–18% initial denial rate concentrates.

6. Technical Reference: ICD-10 Documentation Standards for I35.0 and I35.2

Proper ICD-10 code selection is the foundation upon which the entire prior authorization package rests. Incorrect or insufficiently specific coding triggers automated denials at the front gate — before a human reviewer ever sees the clinical evidence.

Code Selection Logic

I35.0 — Nonrheumatic aortic (valve) stenosis; I35.2 — Nonrheumatic aortic (valve) stenosis with insufficiency

Scribing.io ensures these codes reach maximum specificity to prevent denials through the following logic:

• Etiology confirmation: The NLP engine scans the encounter note and echo report for rheumatic disease indicators (commissural fusion on echo, rheumatic history). If absent, nonrheumatic etiology is confirmed and documented explicitly, supporting I35.x assignment over I06.x.

• Stenosis vs. stenosis-with-insufficiency: When the echo demonstrates more than mild aortic regurgitation (AR grade ≥ moderate, or regurgitant volume ≥30 mL, or vena contracta ≥3 mm), the Validator flags I35.2 as the more specific code. This distinction matters because I35.2 submissions may require additional AR-related clinical data to justify TAVR over surgical AVR, particularly for payers whose criteria consider significant AR a relative contraindication to certain TAVR valve platforms.

• Exclusion of congenital etiologies: For patients with documented bicuspid aortic valve, the Validator flags the need for Q23.0 as a secondary code and ensures the primary TAVR authorization is mapped to the acquired stenosis (I35.0 or I35.2) rather than the congenital anomaly code alone — which some payers do not recognize as a standalone TAVR indication.

• Laterality and manifestation codes: When aortic stenosis has caused secondary conditions (heart failure: I50.x; pulmonary hypertension: I27.20), the Validator prompts for these as secondary diagnoses. These codes strengthen the clinical urgency argument in the prior auth bundle and contribute to appropriate DRG assignment for reimbursement.

Documentation Completeness Checklist for I35.0

The CMS ICD-10-CM Official Guidelines for Coding and Reporting require that code selection reflect the highest specificity supported by the clinical documentation. Scribing.io enforces this at the point of care — not retroactively during coding review — by prompting the clinician for the precise language needed before the note is finalized.

7. TAVR Prior Authorization Workflow: From Echo Ingestion to Da Vinci PAS Bundle

The end-to-end workflow below maps the complete operational sequence from the patient's initial structural heart clinic visit through payer authorization receipt. Each step indicates where Scribing.io automates tasks that are manual in competitor or legacy workflows.

Total estimated time savings per TAVR authorization: 1.5–3 hours of coordinator time plus 1–3 days of physician addendum turnaround plus 4–6 weeks of appeal time on cases that would have been denied.

8. Program-Level Impact: Denial Rate Reduction and Revenue Cycle Metrics

Structural heart programs running 150–400 TAVR cases per year operate in a revenue environment where prior authorization friction has a measurable impact on case volume, operating room utilization, and patient throughput. The financial and clinical consequences of avoidable denials are not hypothetical — they are budget-line realities that program directors manage quarterly.

Quantifying the Denial Cost

Using conservative assumptions:

• Initial denial rate (pre-Scribing.io): 12–18% of TAVR prior auth submissions (based on published analyses of commercial and Medicare Advantage denial patterns; see HHS ASPE Prior Authorization in Medicare Advantage, 2024)

• Mean TAVR reimbursement: ~$58,000 (blended DRG 216/217)

• Appeal success rate: ~65% of denials overturned on appeal, but with 4–6 week delay

• Cases lost to delay: ~5–8% of denied patients either seek care elsewhere, deteriorate, or become inoperable during the appeal window

For a 250-case program with a 15% initial denial rate:

• 37–38 cases denied initially

• ~25 overturned on appeal (with 4–6 week revenue delay per case)

• ~12–13 cases at risk of permanent revenue loss (~$700,000–$750,000 annually)

• Coordinator FTE burden: ~0.5 FTE dedicated solely to TAVR prior auth and appeals

Post-Implementation Targets

Programs deploying the Scribing.io TAVR Symmetry Validator target:

These targets are not theoretical ceilings. They reflect the operational logic of eliminating the specific failure modes this playbook has documented: missing NYHA class, unlabeled or mixed-unit echo data, absent heart team attestations, and the complete absence of LFLG-specific supplemental evidence in standard workflows.

The Patient-Safety Dimension

Revenue metrics capture institutional attention, but the patient-safety argument is equally concrete. A JAMA Cardiology study analyzing TAVR wait times found that patients experiencing authorization delays of >30 days had significantly higher rates of heart failure hospitalization and emergency department visits during the waiting period. For a 78-year-old with NYHA Class III symptoms and an AVA of 0.8 cm², every week of administrative delay is a week of avoidable clinical risk. The Symmetry Validator does not merely accelerate revenue — it removes an artificial barrier between a patient and a procedure that randomized trial evidence (PARTNER 3, NEJM 2019) has demonstrated to be superior or non-inferior to surgery across risk strata.

Structural heart program directors who are managing both the clinical excellence and the operational viability of their programs cannot afford to treat prior authorization as a clerical afterthought. The documentation-to-authorization pipeline is a clinical workflow. Build it like one.