Stopping Tenofovir After a Decade: What a Falling HBsAg Level Actually Licenses
A patient stable on tenofovir for ten years has a hepatitis B surface antigen level low enough to make him eligible for a structured stopping trial under a recently revised guideline. The disagreement is whether that eligibility is a reason to actually stop.
Tuan N., a 44-year-old man, has coached his son's youth soccer league every fall for six years, and jokes that the only appointment he's kept more consistently over the past decade is his twice-yearly hepatitis B labs. He was diagnosed with HBeAg-negative chronic hepatitis B fourteen years ago, likely acquired perinatally, and started tenofovir ten years ago after his ALT began rising; he has been fully virally suppressed with normal ALT on every check since. He has no cirrhosis, no family history of hepatocellular carcinoma, and has never missed a dose by his own report, confirmed by consistently undetectable viral load at every visit.
His most recent quantitative HBsAg came back at 380 IU/mL, and where that number falls is the whole argument. The 2025 AASLD/IDSA guideline conditionally recommends against stopping nucleos(t)ide analog therapy at all until HBsAg loss is achieved, and among the criteria it lists for the patients who might nonetheless be considered is an HBsAg below 100 IU/mL. He is nearly four times that. The meta-analytic data underneath that threshold (Hirode et al.) put the gap in plain terms — stopping at an end-of-treatment HBsAg under 100 IU/mL yielded HBsAg loss in 41.8 percent and virological relapse in 33.4 percent, while stopping at or above 100 IU/mL yielded loss in 4.6 percent and relapse in 72.1 percent. A separate, looser threshold of 1000 IU/mL has been proposed for non-Asian patients, and 380 sits comfortably inside that one; but he was likely infected perinatally, which places him in the subgroup the stricter figure was derived from. So the question isn't whether a decade of clean suppression has earned him a trial off therapy. It is whether a number that clears one threshold and fails the one that probably applies to him is a reason to act at all.
What a decade of stability is actually evidence for
My instinct is to leave a working regimen alone. He's been fully suppressed for ten years with zero toxicity signal, and off-therapy flares in HBeAg-negative disease are common, occasionally severe, and genuinely unpredictable in who gets a mild bump versus a real hepatitis flare. There's no problem here to solve by stopping.
I'd frame the decade differently, and I want to be honest about where his number sits before I do. At 380 he does not meet the under-100 criterion the 2025 guideline lists, and I won't pretend Hirode's 4.6 percent HBsAg-loss figure for the at-or-above-100 group is encouraging — it isn't. What I'd contest is treating that threshold as though it were derived for him. It comes overwhelmingly from Asian cohorts treated with entecavir; he is a decade into tenofovir, and Berg's randomized data in patients stopping tenofovir found seroclearance in 10 percent at one year and 19 percent at three, against none in those who continued.
'No problem to solve' treats indefinite therapy as risk-free, and it isn't — a decade more of daily medication, cost, and adherence burden in a 44-year-old man is a real cumulative cost, even without a toxicity event to point to. But I'll grant that on the guideline as written, the burden of proof is mine, not yours.
I don't think this comes down to whether the biomarker logic is sound — I think it comes down to whether he can actually deliver the monitoring the guideline assumes. The whole safety case for stopping rests on catching a flare early through frequent labs, and 'frequent' in this context means considerably more often than his current twice-yearly visits. Before we act on his HBsAg number, I'd want him to explicitly commit to monthly labs for the first six months off therapy — not as a formality, but as the actual precondition the guideline's own safety data assumes was in place.
No change made today. The team agreed his HBsAg level does not place him among the patients the current guideline would consider for withdrawal, and tenofovir was not discontinued at this visit — nor was a stopping trial scheduled.
Unresolved, deliberately: whether the under-100 threshold is the right one to hold him to, given that it was derived largely in Asian entecavir-treated cohorts while he is a long-term tenofovir patient. The infectious disease physician would recheck HBsAg in six months and reopen the question if it is falling; the hepatologist sees no reason to reopen anything while the guideline reads as it does. Neither position was overruled, and no monitoring commitment was sought from him today, since nothing was going to be stopped on the strength of this number.