Sleep apnea grants at the Board rose sharply after 2023 as remands eased

The Board of Veterans' Appeals granted sleep apnea claims in 33.2% of its decisions in 2024 and 2025, up from 21.3% in 2018 and 2019. The grant share rose by 11.8 percentage points, and in 2026 so far it stands at 36.6%.

The figures come from 59,712 Board decisions on sleep apnea from 2017 to 2026. Across all years, the Board granted 26.4%, denied 21.8% and remanded 50.3%. Counting only decisions on the merits, it granted 54.7%.

Share of decisions each year. 2026 is a partial year.

A slow climb, then a step

For most of the period, about half of sleep apnea appeals went back for more development. Remands were 45.3% in 2017, climbed to 55.7% in 2022 and held at 55.5% in 2023. They then fell, to 45.7% in 2025 and 40.7% in 2026.

The grant share moved slowly at first: 17.3% in 2017, 21.7% in 2021 and 22.8% in 2022. The larger step came between 2023, at 25.8%, and 2024, at 32.6%. On the merits alone, the Board granted 44.5% of claims in 2018 and 2019 and 65.0% in 2024 and 2025. The data does not show why the rate changed.

Opinions and secondary theories travel with grants

The evidence in the record lines up with outcomes. When the record had a positive nexus opinion, the Board granted on the merits in 86.9% of decisions; without one, 33.0%. With a private medical opinion, the rate was 72.3%, against 35.7% without. A treating physician opinion showed 73.5% versus 53.4%, and a favorable VA exam 72.6% versus 51.6%.

Claims made as secondary to a service-connected condition under 38 CFR 3.310 were granted on the merits in 68.0% of decisions, compared with 42.7% for other claims. In the Board's stated reasons, "current diagnosis established" appeared in 2,554 grants and "caused by the service-connected disability" in 1,150. "Service connection established" followed in 1,046.

In the recordNot in the record

Positive nexus opinion

86.9%

33.0%

Private medical opinion

72.3%

35.7%

Claimed as secondary

68.0%

42.7%

Buddy statement

76.0%

52.0%

Merits decisions only (granted or denied). Associations, not causes.

Not every pattern is intuitive. A negative nexus opinion made almost no difference: 54.8% with one, 54.5% without. Records with a lay statement were granted less often, 51.4% versus 70.0%, while records with a buddy statement were granted more often, 76.0% versus 52.0%. These are associations, not causes. Records differ in many ways at once, and a single piece of evidence rarely explains an outcome.

The Board applied the benefit of the doubt under 38 CFR 3.102 in 78.2% of grants.

Why claims are denied

Denials center on what happened in service. The most common reason was "no in-service complaints, treatment or diagnosis," cited in 2,414 denials. Next came "va examiner: less likely than not related to service," in 2,313, and "service records negative or silent," in 1,790.

Two of the three top reasons concern the service record itself. That fits the gap between direct and secondary claims: a secondary theory does not depend on in-service symptoms in the same way.

Read the decisions

Recent examples, with the Board's reasons as summarized by AI:

• 26005118, granted in 2026: obesity found to be a substantial factor in causing or aggravating OSA.
• 26005125, granted in 2026: service treatment records show complaints of sleep problems during deployment.
• A26040512, granted in 2026: private medical opinion linked weight gain to service-connected knee/ankle conditions.
• A26040473, denied in 2026: no presumptive service connection (no in-service onset or relation to TERA).
• A26040482, denied in 2026.
• A26040502, denied in 2026: vA sleep study negative for obstructive sleep apnea.

For representatives

• Where service records are silent, consider whether a secondary theory under 38 CFR 3.310 fits the facts; secondary claims were granted on the merits in 68.0% of decisions, against 42.7% for others.
• A positive nexus opinion is the strongest association in the data: 86.9% granted on the merits with one, 33.0% without.
• An unfavorable VA exam is not the end of a claim. Records with one were still granted on the merits in 53.1% of decisions, and private or treating opinions appear often in grants.