Washington Post Reports Pentagon Undercounted US Iran War Deaths by Four, Hegseth Calls It a Lie

Since Operation Epic Fury began Feb. 28, the Pentagon's public casualty database has listed 18 U.S. military deaths in the Iran war, a number that stood unchanged as of Friday, Sept. 18. That's the number reporters and families have relied on for months. Now that number itself is in dispute.

The Washington Post reported Friday that at least 22 U.S. troops have died since the war began, four more than the Pentagon's Defense Casualty Analysis System shows. The Post's count relies on five unidentified U.S. officials. A sixth official told the paper the real number is 23.

Defense Secretary Pete Hegseth didn't wait long to respond. On X, he called the report "DISGUSTING and FAKE. A complete LIE. Shame on the Washington Post — they're worse than Iranian state media," according to Hindustan Times. The Pentagon, per the Post's own account, declined to answer questions about the discrepancy when asked directly.

What the Pentagon's database actually shows

The public accounting isn't a simple yes-or-no ledger. Of the 18 deaths currently listed, 14 fall under Operation Epic Fury itself, according to the Jerusalem Post's account of the Washington Post reporting. Four more are filed under a separate category called Overseas Operations, tied to the July ceasefire between the U.S. and Iran that later collapsed.

Two additional deaths are on the Pentagon's public record but categorized as unrelated to the Iran war entirely, even though both occurred at bases Iranian forces had attacked. Sgt. Devin Seibel died May 31 at Erbil Air Base in Iraq in what the Pentagon called a training-related incident, tied to Operation Inherent Resolve against ISIS, according to Times Now. Maj. Sorffly Davius died March 6 at Camp Buehring, Kuwait, of what the military listed as an unspecified medical emergency during Operation Spartan Shield.

Neither death appears in DCAS as Iran-war-related. Whether that categorization is accurate, or whether it understates the war's toll, is exactly what the Washington Post's officials are disputing.

The contractor question and the benefits stakes

Beyond the four uniformed personnel, one official told the Post that three U.S. contractors based in the Centcom region have also died since the conflict started, according to Times Now's account. Contractor deaths typically don't appear in military casualty databases at all, which is a separate transparency gap from the DCAS dispute itself.

The Jerusalem Post noted the stakes go beyond a headline number. How a service member's death gets classified determines whether their family receives survivor benefits tied to death in the line of duty. Officials who spoke to the Post reportedly expressed frustration that additional deaths weren't reflected in the public portal families and journalists check.

The case for skepticism

Hegseth's denial rests on unnamed officials at the Washington Post, not a named Pentagon spokesperson, a document release, or an on-record briefing. No service member's name has been attached to the four disputed deaths. The military's practice of separating combat deaths from unrelated deaths at the same base may seem unclear, but it isn't automatically evidence of a cover-up. A training accident or a medical emergency happening on a base under Iranian fire requires judgment about where to draw the line.

The Pentagon had a clean opportunity to settle the question and didn't take it. The Post said the department declined to address the discrepancy when asked directly, according to Reuters' account of the story carried by Internazionale. Declining to explain a four-to-five death gap in a public casualty database while calling the reporting a lie on social media isn't the same as producing records that resolve it.

None of the disputed individuals have been named publicly. The Pentagon has not released updated DCAS figures, corrected the database, or explained the categorization of the Seibel and Davius deaths beyond its original press releases. Until it does, the gap between 18, 22, and 23 remains an open question resting on anonymous sourcing on one side and a flat denial on the other.