Adult-Onset Familial HLH Shows Distinct Genetics and Poorer Survival in China

Familial hemophagocytic lymphohistiocytosis (FHLH) has long been framed as a disease of infancy and early childhood, but a nationwide retrospective cohort study from China suggests that when the inherited disorder first announces itself in adulthood, it behaves differently in ways that matter for diagnosis and survival. Researchers assembled 223 genetically confirmed patients from 27 centers across 27 provinces, municipalities, and autonomous regions, and found that adult-onset disease—defined as onset at age 18 or older—accounted for 54 cases, or 24.2% of the cohort. These adult patients carried a distinct sequence-level variant spectrum, waited longer for a diagnosis, underwent allogeneic hematopoietic stem cell transplantation (allo-HSCT) less often, and died more frequently than patients whose disease began before adulthood.

FHLH is an inherited disorder of immune regulation caused by defects in cytotoxic lymphocyte function. Pathogenic variants in genes such as PRF1, UNC13D, STX11, and STXBP2 impair perforin-mediated killing, cytotoxic granule priming, or granule exocytosis, unleashing uncontrolled immune activation and life-threatening hyperinflammation. Because the disease typically presents in infancy or early childhood, its diagnostic and therapeutic framework—including HLH-directed immunochemotherapy followed by allo-HSCT for patients needing definitive immune reconstitution—has historically been shaped by pediatric practice. Adult presentations, by contrast, often occur in the setting of Epstein–Barr virus (EBV) infection, lymphoma, other severe infections, or inflammatory disorders, and these acquired triggers can lead clinicians to classify the illness as secondary HLH even when an inherited cytotoxic pathway defect underlies it.

The new study, published in Biomarker Research, screened institutional HLH registries, inpatient records, genetic testing databases, and transplantation records for patients diagnosed between 2013 and 2025. Eligible patients met the HLH-2004 diagnostic criteria and had evidence of an inherited FHLH-associated disorder, primarily biallelic pathogenic or likely pathogenic variants in PRF1, UNC13D, STX11, or STXBP2. Patients with only monoallelic variants, variants of uncertain significance without supportive evidence, insufficient clinical data, or alternative genetic diagnoses were excluded. Genetic testing methods varied by center and included targeted next-generation sequencing panels, whole-exome and whole-genome sequencing, Sanger sequencing, and copy-number analysis, with variants classified according to ACMG/AMP guidelines.

The cohort included 117 males and 106 females, with a median age at onset of 7.3 years and a range stretching from 5 days to 67.4 years. Adult-onset patients faced a markedly longer diagnostic delay than younger patients—a median of 3.0 months versus 1.1 months, a difference the authors suggest reflects difficulty recognizing inherited disease in adults. Fever was the most common initial manifestation, occurring in 192 patients, while CNS symptoms at onset were the primary manifestation in 22 patients. A family history of similar phenotypes or unexplained early death was documented in only 16 of 214 evaluable patients, underscoring how easily inherited disease can escape notice when it surfaces late.

Trigger documentation, available for 51 of the 54 adult-onset patients, revealed that 32 had at least one identifiable trigger. EBV was the predominant one, documented in 27 of 51 adults (52.9%), with other infections in 6 and lymphoma in 4, in overlapping categories. Even after excluding lymphoma-associated cases, EBV remained the most frequent trigger at 51.1%. Three of the four lymphoma-associated cases were also EBV-positive. The authors emphasize the practical implication: an identifiable infectious or malignant trigger can coexist with inherited FHLH, and its presence should not end the search for a germline cause.

Genetically, UNC13D and PRF1 dominated the cohort, together accounting for 93.7% of cases, while STXBP2 and STX11 variants were rare. Compared with pediatric and adolescent patients, adults had higher proportions of PRF1 variants (57.4% vs. 38.5%) and homozygous genotypes (33.3% vs. 17.5%), and non-truncating variants—particularly missense variants—were enriched in adult-onset disease. In the PRF1/UNC13D analysis set of 192 patients, greater non-truncating variant burden was associated with adult onset in univariable analysis (odds ratio per unit increase, 1.95; 95% CI, 1.17–3.39). However, joint adjustment for gene and zygosity attenuated the association (adjusted OR, 1.58; 95% CI, 0.94–2.75), leaving uncertainty about whether the link stands independently of these genetic features. Unadjusted sensitivity analyses restricted to confirmed biallelic pathogenic cases showed stronger and statistically significant associations, but the authors caution that their sequence-based classification reflects predicted consequences, not measured residual protein function.

Central nervous system involvement proved common across the age spectrum, documented in 102 of 220 evaluable patients (46.4%). PRF1-mutated patients had the highest frequency of CNS involvement at 58.1%, significantly exceeding patients with non-PRF1 genotypes and those with UNC13D variants. CNS involvement itself did not differ by age at onset, occurring in 48.1% of adults and 45.8% of younger patients. An exploratory survival comparison found no statistically significant difference in overall survival by CNS involvement (log-rank P = 0.285), though the authors note that the composite definition combined heterogeneous clinical, cerebrospinal fluid, and imaging findings, and that survival does not capture long-term neurologic morbidity.

Transplantation patterns diverged sharply. Of 214 patients with known transplantation status, 139 (65.0%) underwent allo-HSCT, but adults were transplanted less frequently than younger patients (49.1% vs. 70.2%), had a shorter interval from diagnosis to transplantation (median 2.7 vs. 4.7 months), and were far less often in complete response before transplantation (15.4% vs. 46.8%). Among adults treated with etoposide- or teniposide-containing regimens, only 7 of 46 achieved complete response, with 28 achieving partial response and 11 progressive disease. These figures suggest that adults reached transplantation with less well-controlled disease, a factor the authors flag as clinically consequential.

Survival outcomes reinforced the divergence. Overall survival was evaluable in 194 patients, with 44 deaths and median follow-up of 48.1 months; median survival was not reached overall, with 1-, 3-, and 5-year estimates of 80.2%, 75.0%, and 75.0%. Adults had markedly lower 3-year survival than pediatric and adolescent patients (51.2% vs. 82.2%). Adult onset remained associated with higher mortality after multivariable adjustment for diagnostic delay, causal gene, and time-dependent allo-HSCT (adjusted hazard ratio, 3.74; 95% CI, 1.96–7.12). Post-transplant survival was also poorer in adults: among 131 recipients with evaluable data, 3-year post-HSCT survival was 60.8% in adults versus 92.1% in younger patients (HR, 5.48; 95% CI, 2.11–14.28), and the association persisted after exploratory adjustment for pre-transplant complete response. Of 17 recorded post-transplant deaths, six were infection-related and 11 stemmed from other transplant-associated complications; none was attributed to HLH relapse.

The authors are careful about the limits of their findings. The retrospective design, center-specific testing practices, and incomplete records permit residual confounding; survival data were missing for 29 patients; trigger documentation was more complete in adults; and subgroup comparisons were exploratory and not specifically powered. The time-dependent allo-HSCT estimate was imprecise, and the authors stress that this nonsignificant result should not be read as evidence that transplantation lacks benefit. Still, the central message stands: adult-onset FHLH in this cohort combined non-truncating variant enrichment, frequent EBV-associated presentations, and poorer overall and post-transplant survival. The findings support considering inherited disease in adults with HLH even when an acquired trigger is identified, with appropriate germline testing, evaluation for EBV and lymphoma, and particular attention to disease control and transplant-related risks before transplantation.

Subject of Research: Age-related differences in genetics, triggers, transplantation, and survival in familial hemophagocytic lymphohistiocytosis

Article Title: Adult-onset familial hemophagocytic lymphohistiocytosis has distinct genetic architecture and outcomes: a nationwide multicenter cohort study in China

Article References: Li, Y., Sun, Y., Bai, J., Yang, R., Liu, R., Cao, J., Liao, M., Wang, Y., Qi, J., Luo, Z., Wang, D., Yang, H., Ji, M., He, Y., Liu, S., Kong, W., Li, X., Han, Y., Suo, T., … Wang, Z. (2026). Adult-onset familial hemophagocytic lymphohistiocytosis has distinct genetic architecture and outcomes: a nationwide multicenter cohort study in China. Biomarker Research, 14(1), Article 149. https://doi.org/10.1186/s40364-026-01026-7

Image Credits: AI Generated

DOI: 10.1186/s40364-026-01026-7

Keywords: familial hemophagocytic lymphohistiocytosis, adult-onset disease, PRF1, UNC13D, EBV, allogeneic stem cell transplantation, CNS involvement, genetic variants, HLH-2004 criteria, retrospective cohort, survival outcomes, China