Letter: Beyond Altitude—Referral Filters and Diagnostic Gaps in Young Stroke

Dear Editor,

We read with interest this two-center comparison of the first-ever ischemic stroke in adults aged 18–50 years from Peking Union Medical College Hospital (44 m) and Tibet Autonomous Region People’s Hospital (3,650 m) (Tang et al., 2026), which highlights a higher proportion of small-vessel occlusion (SVO) and “undetermined” subtypes at altitude, alongside more frequent hyperhemoglobinemia and elevated D-dimer. The dataset is valuable; however, several structural issues make the causal leap—from “altitude-associated phenotype” to “chronic hypoxia-driven SVO pathogenesis”—difficult to sustain without further qualification.

First, the comparison risks conflated altitude exposure with referral-filter and diagnostic-resolution gradients. A national tertiary hub versus a regional main hospital inevitably differs in whom the system selects (complex/rare etiologies referred out; unstable/severe unable to transfer) and in how exhaustively “other/undetermined” etiologies can be ruled in/out. In Tang et al.’s Table 2 (Tang et al., 2026), “other determined etiology” falls from 45.6% to 18.9% (driven largely by vasculitis/APS/tumor-related strokes), while “undetermined” rose from 12.1% to 39.2% at altitude. When the diagnostic work-up depth is asymmetrical, TOAST buckets—especially SVO and undetermined—can shift for reasons unrelated to pathophysiology (Adams et al., 1993).

Second, “hyperhemoglobinemia” and D-dimer positivity are presented as altitude-related prothrombotic signals; however, the article does not standardize time from onset to sampling nor account for dehydration/transport-related hemoconcentration or fluid status, all of which can inflate Hb/Hct early after admission. Without these anchors, the 11.8% versus 0% contrast is fragile to interpret as “erythrocytosis of acclimatization” (Ortiz-Prado et al., 2022).

Third, the logistic model labels the grouping “high altitude” while adjusting for ethnicity, but the Tibetan versus Han distribution is nearly complete between sites; ethnicity here proxies language, diet, care-seeking path, and possibly genetic adaptation. Treating it as a simple covariate may absorb—rather than isolate—what is truly altitude-related (Goldstein et al., 2001).

To strengthen inference, we suggest (1) defining altitude exposure by habitual residence duration/residence altitude rather than admitting-city elevation; (2) reporting core work-up completion rates (CTA/MRA, prolonged ECG monitoring, minimum embolic-source screen) to bound “undetermined”; (3) a sensitivity analysis merging SVO + undetermined or applying stricter lacune criteria; and (4) anchoring Hb/Hct to time window + hydration/renal markers. These steps would help determine whether SVO enrichment reflects a genuine hypoxic microangiopathy or, in part, a byproduct of referral and resolution bias.

Respectfully, we believe that the authors’ data are an important starting point for a much-needed high-altitude stroke phenotype map; making these limitations explicit will only increase the study’s utility for clinicians working on the Plateau.

Authors’ Contributions

W.L. contributed to the conception and design of the critique. D. H. performed the literature review and drafted the article. J.D. provided critical revisions and supervised the entire process. All authors contributed equally to the interpretation of the target article and approved the final version of this letter.