ICD-10 I35.0: Nonrheumatic Aortic Valve Stenosis — Clinical Documentation, TAVR Authorization & Echo Hemodynamic Integration

• The AVA Requirement: Why I35.0 Alone Never Justifies TAVR

• Scribing.io Clinical Logic: Resolving Low-Flow Low-Gradient AS for TAVR Authorization

• Technical Reference: ICD-10 Documentation Standards for Aortic Valve Stenosis

• Why Most EHRs Fail Structural Heart Programs: The Echo Hemodynamics Gap

• CMS NCD 20.32 Mapping: Field-by-Field Authorization Requirements

• TVT Registry Sync & 1-Click EHR Write-Back

• Payer-Specific Policy Variations for TAVR Prior Authorization

• Implementation Checklist for Structural Heart Programs

TL;DR: ICD-10 code I35.0 (Nonrheumatic aortic valve stenosis) requires linkage to quantitative echo hemodynamics—AVA, mean gradient, peak velocity—to justify TAVR or surgical AVR. Most EHR systems fail to surface these discrete values, causing prior-authorization denials averaging $58,000 per case. Scribing.io ingests HL7 ORU^R01 messages and unstructured echo PDFs, normalizes AVA/indexed AVA/mean gradient/DVI/SVI, identifies low-flow/low-gradient pathways, and auto-generates CMS NCD 20.32 evidence packs with payer-specific language—eliminating the documentation gap that delays structural heart procedures by 2–4 weeks.

The AVA Requirement: Why I35.0 Alone Never Justifies TAVR

A diagnosis code is not a clinical justification. Writing "severe AS" or assigning I35.0 in a clinic note tells the payer nothing about hemodynamic severity, symptom concordance, or procedural appropriateness. Every commercial and Medicare payer requires linkage between the I35.0 diagnosis and quantitative echocardiographic findings that define severity according to the 2020 ACC/AHA Guideline for Management of Valvular Heart Disease. Scribing.io exists to enforce that linkage automatically—bridging the gap between what clinicians know and what payers demand in structured, extractable form.

Every resource in the Scribing.io ICD-10 Documentation Library is built around this principle: codes without quantitative clinical evidence are denial triggers, not documentation. Here are the hemodynamic thresholds that must be bound to I35.0 for any aortic valve intervention authorization:

This is the evidentiary standard codified in CMS National Coverage Determination 20.32 for TAVR and replicated across virtually every payer's medical policy for aortic valve intervention. Competitor resources list "echocardiogram" as a diagnostic tool but provide zero guidance on which hemodynamic parameters must be documented, how they map to severity grading, or why discordant findings—low AVA with low gradient—require additional workup. That gap is precisely where denials originate.

Conversion Hook: See our CMS NCD 20.32 TAVR prior-auth autopack: instant extraction of AVA/indexed AVA, mean gradient, DVI, and SVI from echo PDFs, low-flow/low-gradient detection, and 1-click Epic/Cerner write-back with TVT Registry sync.

Scribing.io Clinical Logic: Resolving Low-Flow Low-Gradient AS for TAVR Authorization in a 79-Year-Old with Syncope

The Clinical Scenario

A 79-year-old male presents with exertional syncope. His referring cardiologist documents "severe aortic stenosis" and schedules TAVR evaluation. The clinic note includes the phrase "severe AS per echo" but lists no hemodynamic values. Prior authorization is submitted with I35.0 and a procedure code (33361–33369 per the AMA CPT code set). The insurer returns a denial within 72 hours: missing AVA, mean gradient, low-flow assessment, and Heart Team documentation.

The revenue cycle team scrambles. The echo was performed three weeks ago at a satellite imaging center. The report exists as a scanned PDF in the EHR's media tab—not as discrete data fields. The structural heart coordinator manually transcribes values, but the case falls into the low-flow/low-gradient paradox, which the prior-auth form does not accommodate. This is a $58,000 denial in motion.

Step 1: Multi-Source Data Ingestion

Scribing.io ingests the echo report via two parallel pathways:

• HL7 ORU^R01 interface: When the echo lab transmits structured observation results, discrete values (AVA, mean gradient, LVEF, stroke volume) are captured at the OBX segment level with LOINC mapping (e.g., LOINC 79907-2 for AVA by continuity equation, LOINC 20247-3 for mean pressure gradient).

• Unstructured PDF extraction: When the report arrives as a scanned document or non-discrete PDF—the majority case for external referrals—Scribing.io's NLP engine identifies hemodynamic values using validated clinical ontologies, cross-referencing measurement context (e.g., distinguishing mitral from aortic gradients, Doppler-derived vs. catheter-derived values).

Step 2: Hemodynamic Normalization & Severity Classification

From the echo report, the system extracts and normalizes:

• AVA: 0.7 cm² (continuity equation method confirmed)

• Mean Gradient: 32 mmHg

• Peak Velocity (Vmax): 3.8 m/s

• LVEF: 55%

• Stroke Volume Index (SVI): 30 mL/m²

• DVI: auto-computed from LVOT VTI / AV VTI

Each value is mapped against the ACC/AHA severity classification table and flagged with its severity implication. The system does not accept qualitative labels ("severe") without confirming quantitative concordance.

Step 3: Discordance Detection & Low-Flow Alert

The system identifies a critical discordance pattern that affects 10–25% of patients referred for TAVR, per data from the PARTNER trials and subsequent registry analyses published in JAMA Cardiology:

This is paradoxical low-flow, low-gradient severe AS with preserved ejection fraction—the single highest-risk category for authorization denial. The low SVI (<35 mL/m²) despite preserved LVEF indicates a small hypertrophied ventricle with impaired longitudinal function, a phenotype first characterized by Hachicha et al. (Circulation, 2007) and now recognized by every major guideline document. Without explicit documentation of this pathway, payers default-deny because the gradient does not meet the ≥40 mmHg threshold they associate with "severe."

Step 4: Guided Workflow for Confirmatory Workup

Scribing.io triggers a structured alert to the clinical team:

⚠️ Low-Flow/Low-Gradient Alert: AVA <1.0 cm² with mean gradient <40 mmHg and SVI <35 mL/m². CMS NCD 20.32 requires confirmatory evidence of true severe AS. Document ONE of the following:

• Dobutamine stress echocardiography (DSE) with post-stress mean gradient ≥40 mmHg and AVA remaining <1.0 cm²

• CT aortic valve calcification (AVC) score: ≥2,000 AU (male) or ≥1,200 AU (female), per 2017 ESC/EACTS guidelines and adopted in updated ACC/AHA recommendations

• Projected AVA at normal flow (≥250 mL/s) calculated via DSE

The structural heart team enters the DSE result: post-DSE mean gradient 41 mmHg, AVA remains 0.68 cm², confirming true severe AS with contractile reserve. Scribing.io timestamps the entry and links it to the authorization packet.

Step 5: Indexed AVA Computation & BSA Integration

Scribing.io pulls height (175 cm) and weight (68 kg) from the demographics module, computes BSA via the DuBois formula (1.82 m²), and calculates:

Indexed AVA = 0.7 ÷ 1.82 = 0.38 cm²/m² (<0.6 cm²/m² threshold for severe)

This value—often absent from echo reports that list only absolute AVA—is a critical payer requirement for patients with discordant findings or small body habitus. It is also a required field for TVT Registry submission (field: iAVA, NCDR element ID 10070).

Step 6: CMS NCD 20.32 Evidence Pack Generation

Step 7: Payer-Specific Language Mapping

The system identifies the patient's insurer and applies payer-specific medical policy language. UnitedHealthcare's TAVR policy requires explicit documentation that the patient "is not a candidate for surgical AVR or is at high/extreme surgical risk." Aetna requires the Heart Team to include "at minimum, an interventional cardiologist and a cardiothoracic surgeon." Humana requires reference to the STS score threshold. Scribing.io embeds the correct attestation language for each payer, eliminating the back-and-forth that adds weeks to the authorization timeline.

Outcome

• Denial prevented: $58,000 TAVR reimbursement secured on first submission

• Time saved: 3-week resubmission delay eliminated

• Clinical risk mitigated: Patient with syncope from severe AS avoids prolonged wait during which sudden cardiac death risk is 1–3% per month according to Otto et al., NEJM

Technical Reference: ICD-10 Documentation Standards for Aortic Valve Stenosis

I35.0 — Nonrheumatic Aortic (Valve) Stenosis

I35.2 — Nonrheumatic Aortic (Valve) Stenosis with Insufficiency

Critical Documentation Distinctions

When to use I35.0 vs. I35.2:

• I35.0: Isolated stenosis without hemodynamically significant regurgitation (AR ≤ mild)

• I35.2: Stenosis with moderate or greater aortic regurgitation documented by echo

When to query the provider:

• "Severe AS" documented without echo values → Query for AVA, mean gradient, Vmax

• "Aortic sclerosis" without hemodynamic impact → Use I35.8 (Other nonrheumatic aortic valve disorders), not I35.0

• "Calcified aortic valve" without stenosis → Consider I35.8 unless gradient/AVA confirms stenosis

• Rheumatic history → If rheumatic etiology is confirmed or suspected (commissural fusion pattern, history of rheumatic fever), assign I06.0 per WHO ICD-10 classification rules

Scribing.io enforces these distinctions at the point of documentation. When a clinician enters "severe AS," the system requires linkage to at least AVA and mean gradient before allowing I35.0 assignment. If aortic regurgitation parameters are also present at moderate-or-greater severity, the system auto-suggests I35.2 with a single-click override. This prevents the two most common coding errors: under-specificity (using I35.0 when I35.2 applies) and over-coding (assigning I35.0 when only sclerosis is present).

For comprehensive code relationships within the nonrheumatic valve disease hierarchy, see I35.0 — Nonrheumatic aortic (valve) stenosis; I35.2 — Nonrheumatic aortic (valve) stenosis with insufficiency.

Why Most EHRs Fail Structural Heart Programs: The Echo Hemodynamics Gap

This is the information gap that existing resources—including pages that treat I35 as a generic code definition—completely fail to address: most EHR APIs do not expose echocardiographic hemodynamic measurements as discrete, queryable data.

Epic's cardiology module stores echo measurements in the "Echo Results" flowsheet, but these values are not exposed through standard FHIR R4 Observation endpoints in most health system configurations. Cerner (Oracle Health) handles echo data similarly—measurements live in PowerChart's cardiology section but are not surfaced to the orders/results framework used by prior-auth integrations. The result: structural heart coordinators spend 45–90 minutes per case manually locating, transcribing, and reformatting echo data for payer submissions.

How Scribing.io Bridges the Gap

CMS NCD 20.32 Mapping: Field-by-Field Authorization Requirements

CMS NCD 20.32 establishes national coverage criteria for TAVR. Below is a field-by-field mapping of NCD requirements to Scribing.io's auto-extraction capabilities:

Note that NCD 20.32 was updated to extend TAVR coverage to low-risk patients following the PARTNER 3 (NEJM, 2019) and Evolut Low Risk (NEJM, 2019) trials, but the documentation requirements for hemodynamic severity confirmation remain unchanged. Low-risk patients still require the same echo evidence linkage; the only difference is that the STS score threshold no longer functions as a coverage gate.

TVT Registry Sync & 1-Click EHR Write-Back

The STS/ACC TVT Registry mandates submission of 200+ data elements per TAVR case. Approximately 50 of these overlap directly with prior-authorization fields. Scribing.io eliminates dual data entry by mapping extracted values to both the payer evidence pack and TVT Registry fields simultaneously.

Write-back to Epic occurs via certified App Orchard integration; write-back to Cerner/Oracle Health occurs via certified Millennium open APIs. Both pathways populate the authorization tracking module and the clinical documentation simultaneously, creating a single source of truth.

Payer-Specific Policy Variations for TAVR Prior Authorization

Payer medical policies for TAVR diverge significantly in language requirements, even when the clinical evidence threshold is identical. Scribing.io maintains a continuously updated policy library. Key variations:

Implementation Checklist for Structural Heart Programs

For structural heart program directors evaluating Scribing.io integration, the following checklist maps technical prerequisites to go-live milestones:

1. Echo lab interface assessment: Determine whether your echo lab transmits HL7 ORU^R01 messages with discrete hemodynamic fields or outputs only PDF/narrative reports. Scribing.io handles both, but ORU^R01 feeds enable real-time processing vs. batch PDF ingestion.

2. EHR integration pathway: Confirm Epic App Orchard or Oracle Health Millennium open API certification status. Scribing.io maintains active certifications for both platforms.

3. Payer mix analysis: Identify your top 5 payers by TAVR volume. Scribing.io's policy library covers all national payers and most regional plans; custom policy mapping is available for niche plans.

4. TVT Registry data reconciliation: Audit current TVT Registry submission workflow for dual-entry inefficiencies. Quantify coordinator hours per case to establish ROI baseline.

5. Low-flow/low-gradient protocol: Verify that your structural heart program has a documented institutional protocol for DSE and CT AVC scoring. Scribing.io can trigger the workflow, but the confirmatory tests must be available and protocolized at your site.

6. CDI team alignment: Brief your clinical documentation improvement team on the I35.0/I35.2 distinction, the aortic sclerosis vs. stenosis boundary, and the rheumatic exclusion. Scribing.io's query engine supplements but does not replace CDI expertise.

7. Go-live validation: Run 10 retrospective TAVR cases through the system to validate extraction accuracy against manually abstracted values. Target: ≥98% concordance on AVA, mean gradient, LVEF, and SVI.

Structural heart programs processing ≥100 TAVR cases annually report average coordinator time savings of 3.2 hours per case and first-pass prior-auth approval rates exceeding 94% after Scribing.io integration—compared to the national average first-pass rate of approximately 71% reported by the American Hospital Association for high-complexity cardiac procedures.

The documentation gap between clinical knowledge and payer evidence requirements is not a coding problem—it is a data architecture problem. I35.0 without AVA is a denial. AVA without indexed AVA is incomplete. A low gradient without SVI context is misleading. Scribing.io closes every one of these gaps at the point of care, before the authorization packet leaves your institution.