A 79-year-old man, who was retired farmer from rural Shandong Province, came to the emergency department in October 2025 with 4 days of spiking fevers up to 39.5°C and rigors. Over the preceding month, he had experienced 4 syncopal episodes; each time he regained consciousness within minutes but had headache and dizziness afterward. He spent most of his days tending a backyard garden and had a long history of drinking ≈60 mL of liquor daily. He kept no animals and had not traveled recently.

On arrival, he looked pale. His temperature was 38.3°C, pulse 92, and blood pressure 136/78 mm Hg. The conjunctivae were blanched. Heart sounds were regular without murmurs, the lungs were clear, and the abdomen was soft with no organomegaly. Neurologic examination and electrocardiogram results were unremarkable. Cranial computed tomography showed only mild left anterior cerebral artery stenosis, insufficient to explain his syncopal episodes. Chest and abdominal computed tomography showed scattered lymphadenopathy in the anterior mediastinum and both inguinal regions (largest node 1.4 cm).

Laboratory studies revealed pancytopenia and multiorgan dysfunction: white cells 3,690 cells/μL (reference range 3,500–9,500 cells/μL); hemoglobin 8.5 g/dL (reference 13.0–17.5 g/dL); mean corpuscular volume 106.8 fL (reference 82–100 fL); platelets 84,000/µL (reference 125,000–350,000/μL); alanine aminotransferase 131 IU/L (reference 9–50 IU/L); aspartate aminotransferase 103 IU/L (reference 15–40 IU/L); lactate dehydrogenase 442 IU/L (reference 120–250 IU/L); creatinine 159 μmol/L (reference 57–111 μmol/L); activated partial thromboplastin time 49.8 seconds (reference 24.0–39.0 seconds); D-dimer 1.92 mg/L (reference 0.00–1.00 mg/L); and ferritin 757 ng/mL (reference 13–150 ng/mL) with low total iron binding capacity. Tests for brucellosis, HIV, bunyavirus, hepatitis, cytomegalovirus, and SARS-CoV-2 were all negative.

Bone marrow aspirate showed 23% ring sideroblasts, and next-generation sequencing (NGS) detected an SF3B1 mutation (c.1986C>A, variant allele frequency 50.63%) with a 45,X,-Y karyotype, establishing a diagnosis of myelodysplastic syndrome (MDS) SF3B1 (6). However, MDS alone could not account for the fevers, the syncopal episodes, or the lymphadenopathy.

Figure 1

Figure 1. Bone marrow morphology in samples from a 79-year-old man with scrub typhus and myelodysplastic syndrome, Shandong Province, China, 2025. A) High-power view showing a macrophage with abundant cytoplasm filled with...

Of particular note on the bone marrow aspirate were clusters of large macrophages whose cytoplasm was packed with fine eosinophilic granules (Figure 1). Those cells were morphologically indistinguishable from the typhoid cells long regarded as a hallmark of Salmonella enterica serovar Typhi bone marrow infection (3,4). Hemophagocytosis of red cells and platelets also was evident; the patient’s platelets dropped from 84,000/µL to 60,000/µL over just a few days.

The morphologic findings were strongly suggestive of enteric fever. However, the patient had no abdominal pain, no rose spots, and no diarrhea. Blood and bone marrow cultures grew nothing, and the Widal test was negative, leaving bone marrow morphology suggestive of typhoid fever but clinical manifestations that did not support this diagnosis.

Figure 2

Figure 2. Eschar in the popliteal fossa of a 79-year-old man with scrub typhus and myelodysplastic syndrome, Shandong Province, China, 2025. A 5-mm eschar with central black necrosis and surrounding erythema was...

We sent the bone marrow aspirate for targeted NGS. The result, available on hospital day 6, was unexpected: Orientia tsutsugamushi was identified (2,7). (The O. tsutsugamushi targeted NGS sequence data have since been deposited in the National Center for Biotechnology Inofrmation Sequence Read Archive [BioProject accession no. PRJNA1470902; BioSample accession no. SAMN60430499]). A repeat physical examination of the patient found a 5-mm eschar with a dark necrotic center in the left popliteal fossa, an area easily overlooked was (Figure 2).The patient, when asked again, recalled no insect bite but confirmed he gardened barefoot and in shorts daily.

We administered intravenous doxycycline (100 mg 2/×d) (8). The patient’s fever broke within 48 hours. By day 7, his platelets had climbed to 152,000/µL, and transaminases had normalized (alanine aminotransferase 27 IU/L, aspartate aminotransferase 23 IU/L). At 3-week follow-up, the syncopal episodes had not recurred, and he began treatment for MDS.

Several aspects of this patient’s course deserve comment. First and foremost, a meticulous head-to-toe skin examination, with deliberate inspection of concealed areas including the axillae, groin, and popliteal fossae, might have revealed the eschar at admission and spared the patient a prolonged diagnostic workup. Shandong Province is a recognized endemic area for scrub typhus, and this simple clinical step remains the most accessible and cost-effective path to early diagnosis (4,7). The expanding geographic range of scrub typhus across China makes a thorough and systematic search for eschars all the more critical in the evaluation of febrile patients (3,5).

The bone marrow typhoid cells nearly led to a wrong diagnosis. Such distinctive macrophages have been considered relatively specific for enteric fever (3,4). That O. tsutsugamushi can produce morphologically identical cells expands the known spectrum of bone marrow pathology in scrub typhus. When such cells are encountered but microbiologic evidence of Salmonella infection is lacking, alternative intracellular infections, including scrub typhus, should be considered.

This patient exhibited features consistent with secondary hemophagocytic lymphohistiocytosis: spiking fevers, progressive pancytopenia, hyperferritinemia, elevated lactate dehydrogenase, and hemophagocytosis on bone marrow aspirate. In a systematic review by Naoi et al. (9), 30 patients with scrub typhus–associated hemophagocytic lymphohistiocytosis were analyzed; the overall mortality was 6.7%, and approximately half recovered with anti-rickettsial antibiotics alone. That pattern mirrors our patient’s course. The rapid normalization of platelet counts after doxycycline treatment (from 60,000/µL to 152,000/µL within 7 days) strongly suggests that the acute pancytopenia was driven primarily by scrub typhus–associated hemophagocytosis, rather than by MDS (10).

For diagnostic confirmation, specific PCR targeting of the 47 kDa gene of O. tsutsugamushi offers high sensitivity and is more widely available than sequencing-based methods (11). Where scrub typhus is clinically suspected, 47 kDa PCR should be the first-line molecular test. Targeted NGS retains value when the causative pathogen is not suspected, as in this patient.