Case presentation

A 37-year-old woman had a remote history of a cesarean section complicated by chronic intermittent infection of the Pfannenstiel surgical wound, characterized by recurrent local discharge and multiple courses of empiric antibiotic therapy through self-medication, without definitive resolution. Her additional medical history included prior episodes of post-COVID-19 bronchospasm and recent odontalgia in the days preceding admission, treated with nonsteroidal anti-inflammatory drugs. The patient presented with a subacute history of general malaise, asthenia, adynamia, and fatigue associated with pain in the right lower limb. During the initial evaluation, superficial venous thrombosis of the small saphenous vein was documented in the setting of significant hematologic cytopenias. On admission, laboratory studies showed severe leukopenia of 1,100 cells/mm³, anemia, and marked thrombocytopenia of 23,000 cells/mm³, accompanied by elevated inflammatory markers. The key laboratory and oxygenation findings during the clinical course are summarized in Table 1.

Parameter | Admission/diagnosis | Respiratory deterioration/HFNC | During HFNC and methylprednisolone | Follow-up | Reference range | Unit
White blood cell count | 1,100 | 61,500 | 12,800 | 1,200 | 5,000-10,000 | cells/mm³
Platelet count | 23,000 | 31,000 | 42,000 | 35,000 | 150,000-400,000 | cells/mm³
Absolute neutrophil count | 748 | 30,135 | 1,152 | 240 | 2,500-8,000 | cells/mm³
PaO₂/FiO₂ ratio | 320 | 150 | 180 | 350 | >300 | ratio
ROX index | Not applicable | 4 | 8 | 14 | Not applicable | ratio
C-reactive protein | 260 | 97 | 88 | 45 | 0-6 | mg/L
Procalcitonin | 1.54 | 0.29 | 0.12 | 0.1 | <0.5 | ng/mL
Fibrinogen | 344 | 233 | 228 | 169 | 180-350 | mg/dL
International normalized ratio | 1.3 | 1.26 | 1.24 | 1.02 | 0.9-1.3 | ratio
D-dimer | 39 | 14 | 12 | 12 | 0-0.5 | mg/L

During the initial clinical course, the patient developed methicillin-sensitive _Staphylococcus aureus _bacteremia and severe pneumonia, for which targeted antimicrobial therapy was initiated. Concurrently with the hematologic workup, she experienced a rapid reversal of her leukocyte profile, progressing from severe leukopenia to marked hyperleukocytosis, with a peak white blood cell count of 61,500 cells/mm³. Diagnostic evaluation confirmed APL through detection of the PML-RARA rearrangement. After confirmation of APL, funduscopic examination revealed bilateral Roth spots, with macular involvement of the right eye (Figure 1).

Figure 1: Fundoscopic examination revealing bilateral Roth spots.(A) Right eye, posterior pole, showing multiple rounded retinal hemorrhages with pale centers, consistent with Roth spots, including macular involvement. (B) Right eye, peripheral retina, showing additional hemorrhagic lesions, illustrating the multifocal distribution of the retinopathy. (C) Left eye, showing a peripheral hemorrhagic lesion with sparing of the macular region. Both right-eye images were included because they show complementary retinal regions, documenting both posterior pole involvement and peripheral extension of the hemorrhagic lesions. These findings were interpreted as hemorrhagic retinopathy secondary to the underlying hematologic disorder.

Given the history of S. aureus bacteremia, echocardiographic evaluation was performed and showed no evidence of vegetations or other findings suggestive of infective endocarditis. Therefore, the retinal findings were interpreted primarily in the context of leukemia, anemia, and severe thrombocytopenia. Following diagnostic confirmation, differentiation therapy with ATRA was initiated at a dose of 90 mg/day according to internationally accepted treatment protocols in a setting where arsenic trioxide (ATO) was unavailable. Given the progressive hyperleukocytosis, hydroxyurea at 1,500 mg/day was added as a cytoreductive measure during the induction phase. Daunorubicin 120 mg was also administered according to the institutional protocol. In parallel, intensive, protocolized, and closely monitored transfusion support was implemented to manage consumptive coagulopathy and the high hemorrhagic risk characteristic of APL during the initial phase of the disease. In addition, dexamethasone was administered at a total daily dose of 32 mg/day as prophylaxis for DS.

Despite steroid prophylaxis, 48 hours after initiation of ATRA, the patient developed rapidly progressive respiratory deterioration characterized by worsening hypoxemia, diffuse bilateral pulmonary infiltrates (Figure 2A), a marked increase in oxygen requirements, and fluid retention with weight gain exceeding 5 kg. At that time, the PaO₂/FiO₂ ratio was 150, compatible with moderate hypoxemic respiratory failure and a clinical profile associated with a higher risk of intubation. The clinical presentation posed a complex diagnostic challenge that included sepsis-related ARDS, severe pneumonia, pulmonary leukostasis, opportunistic viral or fungal infection, diffuse alveolar hemorrhage, TRALI, TACO, drug-induced lung injury, and cardiogenic pulmonary edema. These possibilities were evaluated according to the clinical course, microbiologic findings, hemodynamic status, transfusion history, radiographic pattern, echocardiographic findings, inflammatory biomarker trends, and response to therapy. Active infectious progression was considered less likely because respiratory samples showed no bacterial growth, and inflammatory biomarkers decreased during the respiratory course. Cardiogenic pulmonary edema and TACO were considered less likely because there was no clinical or echocardiographic evidence of cardiac failure or predominant volume overload. Diffuse alveolar hemorrhage was considered in the setting of thrombocytopenia and coagulopathy, but there was no hemoptysis or clinical evidence of active pulmonary bleeding. Overall, the close temporal relationship with ATRA initiation, marked hyperleukocytosis, fluid retention, predominant pulmonary involvement, and the absence of a more likely alternative cause supported pulmonary DS as the predominant mechanism of respiratory deterioration.

Figure 2: Serial chest radiographs during pulmonary DS.(A) Chest radiograph obtained 48 hours after ATRA initiation, showing diffuse bilateral alveolar-interstitial opacities, predominantly in the perihilar and basal lung fields, consistent with severe pulmonary involvement in the setting of acute hypoxemic respiratory deterioration. (B) Follow-up radiograph after 48 hours of HFNC support, showing persistent bilateral alveolar-interstitial opacities, which prompted escalation to pulse methylprednisolone therapy. (C) Follow-up radiograph obtained after three days of pulse methylprednisolone therapy and continued HFNC support, showing marked radiologic improvement with significant resolution of the bilateral pulmonary opacities.ATRA, all-trans retinoic acid; DS, differentiation syndrome; HFNC, high-flow nasal cannula

Given the progression of respiratory failure, HFNC support was initiated, reaching an FiO₂ of up to 80% and a flow rate of 60 L/min, combined with awake prone positioning. The respiratory rate-oxygenation (ROX) index was 4, suggesting a higher risk of intubation. However, invasive mechanical ventilation was not immediately pursued because the patient remained hemodynamically stable, awake, cooperative, and without altered mental status, severe ventilatory failure, or clinical signs of respiratory exhaustion. Although close monitoring was maintained because of the high risk of deterioration, the absence of immediate standard indications for intubation and the need to minimize invasive procedure-related risks in the setting of severe thrombocytopenia and coagulopathy supported a trial of HFNC. After 48 hours of HFNC support, the ROX index increased to 8, compatible with partial functional improvement, supporting the decision to continue noninvasive respiratory support under close surveillance. However, given the persistence of radiographic pulmonary involvement (Figure 2B), severe pulmonary-predominant DS despite dexamethasone prophylaxis, marked hyperleukocytosis, and concern for ongoing inflammatory lung injury related to cytokine-mediated endothelial activation and capillary leak, corticosteroid therapy was intensified using pulse methylprednisolone at a dose of 1 g/day for three days. This decision was made as an individualized rescue strategy rather than as a standard replacement for dexamethasone-based management, in the context of rapidly progressive hypoxemic respiratory failure and the need to avoid invasive mechanical ventilation if clinically feasible.

After initiation of pulse methylprednisolone therapy, the patient showed a favorable clinical and oxygenation response, with improvement in the PaO₂/FiO₂ ratio from 150 to 350, progressive reduction in oxygen requirements, and respiratory stabilization, thereby avoiding the need for invasive mechanical ventilation. Follow-up chest imaging showed significant improvement in the bilateral pulmonary infiltrates (Figure 2C), consistent with the favorable clinical course and response to pulse methylprednisolone therapy, HFNC support, and awake prone positioning. The patient subsequently experienced progressive respiratory improvement, allowing gradual withdrawal of HFNC support and transition to low-flow oxygen therapy. However, her subsequent course was complicated by profound neutropenia, with a nadir absolute neutrophil count of 240 cells/mm³, and persistent severe thrombocytopenia, with a platelet count of 35,000 cells/mm³, in the context of the underlying disease and daunorubicin-associated myelotoxicity. For this reason, she required ongoing intensive transfusion support and treatment with filgrastim 300 mcg every 24 hours. At the time of intensive care unit discharge, the patient had sustained respiratory improvement, had been weaned from HFNC support, and remained stable on low-flow oxygen therapy, with a PaO₂/FiO₂ ratio of 350 and a ROX index of 14. She was ultimately discharged from the intensive care unit for continued specialized hematologic management.

At the four-month outpatient follow-up, laboratory assessment showed hematologic recovery, with values within the normal range. Molecular monitoring of bone marrow was performed using quantitative reverse transcription-polymerase chain reaction for t(15;17) PML-RARA, showing no detectable molecular expression, with a ratio of 0.00000 and an expression rate of 0.000%. These findings were compatible with a complete molecular response and supported a favorable clinical and hematologic status at follow-up.