Abstract
Background
Thrombosis remains a clinically important complication of acute promyelocytic leukemia (APL), coexisting with the prototypical hemorrhagic diathesis. Data from Gulf Cooperation Council (GCC) populations are limited in addition to data specifically about thrombosis, which is heterogenous in general.
Objectives
To describe the incidence, clinical characteristics, treatment, and outcomes of thrombotic events in patients with APL treated at a tertiary center in the GCC.
Methods
We conducted a retrospective cohort study of 75 consecutive patients with newly diagnosed APL managed at a single tertiary center. Baseline demographic variables (age, sex, regional background), hematologic parameters (hemoglobin, leukocyte, platelet counts), and coagulation indices (fibrinogen, activated partial thromboplastin time, prothrombin time, international normalized ratio, and D-dimer) were compared between patients with and without thrombosis. Thrombotic event type, timing, antithrombotic management, bleeding complications, and mortality (including early mortality within 30 days) were recorded.
Results
Eleven of 75 patients (14.7%) developed thrombosis. Baseline demographic characteristics, hemoglobin level, leukocyte count, platelet count, fibrinogen, and activated partial thromboplastin time were comparable between patients with and without thrombosis, whereas prothrombin time, international normalized ratio, and D-dimer levels were numerically higher in patients who developed thrombosis. Most thrombotic events (73%) occurred during active therapy, and the majority were venous (55%). Pulmonary embolism accounted for 36% of all thrombotic episodes, and catheter-associated thromboses represented a clinically relevant subset. Anticoagulation was initiated in 45% of patients with thrombosis and antiplatelet therapy in 18%. Bleeding episodes occurred in 36% of patients who experienced thrombosis. Thrombosis was associated with a 27% all-cause mortality.
Conclusions
In this GCC APL cohort, thrombosis, though less frequent than bleeding, emerged as a serious and clinically significant manifestation associated with substantial morbidity and mortality. These findings underscore the need for vigilant thrombo-hemorrhagic assessment and individualized antithrombotic strategies, particularly during induction and early consolidation therapy.
Introduction
The transformation of Acute Promyelocytic Leukemia APL outcomes with all-trans retinoic acid (ATRA) and arsenic trioxide (ATO) is one of modern hematology's major achievements, enabling chemotherapy-sparing, risk-adapted strategies and excellent survival in low- and intermediate-risk disease.1,2 Yet the disease remains defined by a labile coagulopathy in which both bleeding and thrombosis contribute to early morbidity and mortality. Contemporary series in the ATRA era generally report thrombotic incidences ranging from roughly 5% to 20%, with venous events predominating and pulmonary embolism a recurring concern The dynamic biology of differentiation, endothelial activation, and concurrent pro- and anti-coagulant perturbations make risk stratification difficult at the bedside, particularly when catheter placement and supportive care introduce additional triggers. While landmark protocols such as ATRA + ATO for non–high-risk APL and APML4 for broader risk categories have unified induction strategies, reliable clinical markers to identify patients at risk of thrombosis and to optimize anticoagulation without exacerbating bleeding are still lacking. Against this backdrop, we report our center's experience, detailing the incidence and characteristics of thrombotic events in APL, comparing baseline features between patients with and without thrombosis, and describing associated management approaches and early outcomes..3–6
Methods
Study Design and Setting
We retrospectively identified and analyzed 75 consecutive adult patients with confirmed APL treated at the National Center for Cancer Care and Research, Doha, Qatar, in the period between January 2015 and December 2024. As a tertiary referral center and offers government-funded, fully covered cancer treatment. Case identification and data abstraction were performed from departmental databases and source documentation.
Data Sources and Cohort Assembly
Three prespecified summary tables were used: (i) the overall cohort, (ii) a two-group comparison (thrombosis present vs absent), and (iii) a detailed table of the thrombosis subset. Patients were classified as thrombosis-positive if at least one objectively documented thrombotic event (as determined by imaging or clinician-adjudicated per record) was recorded at any time during induction or consolidation; all others were classified as thrombosis-negative.
Variables and Definitions
At diagnosis, the following baseline variables were abstracted: age, sex, regional background (Arab vs non-Arab), hemoglobin, leukocyte count, platelet count, prothrombin time (PT), international normalized ratio (INR), D-dimer, fibrinogen, and activated partial thromboplastin time (APTT).
Within the thrombosis subset, we recorded timing of the first event (at diagnosis vs during treatment), vascular territory (venous, arterial, or both), specific event types (eg, deep-vein thrombosis, pulmonary embolism), catheter-related thrombosis (CRT), organ-specific sites, bleeding events, use of anticoagulation and/or antiplatelet therapy, induction backbone (eg, ATRA + ATO, APML4-type), and early vital status at data cut-off.
Outcomes
The primary outcome was the occurrence of thrombotic events. Secondary descriptive outcomes included the distribution and timing of events, CRT frequency, concomitant bleeding, and patterns of antithrombotic management.
Statistical Analysis
Continuous variables are summarized as means (standard deviations); categorical variables as counts (percentages). Groupwise comparisons between thrombosis-positive and thrombosis-negative patients used two-sided tests exactly as reported in the source tables (p-values reproduced verbatim to avoid re-analysis of small strata or sparse cells). Because individual survival times and cause-specific hazards were not available, time-to-event modeling (Kaplan–Meier, Cox, or competing-risks) was not undertaken. No imputation was performed; analyses used only available data. Stata 17 used for analysis and cut of p0.05 for significance.
Ethics
The study was conducted in accordance with institutional policies for retrospective chart review; the need for informed consent was waived by the local review board due to the minimal risk and de-identification of the data. MRC Approval 01-25-130.
Results
The cohort comprised 75 patients. Males accounted for 80% (60/75), and Females accounted for 20% (15/75). The mean age of diagnosis in the overall cohort was 39.5 years in keeping with the youthful distribution typical of regional APL programs. Eleven patients experienced thrombosis, giving an incidence of 14.7%. Tables 1 and 2. ICU admission was required in 6/11 (54.5%) patients with thrombosis, reflecting substantial early clinical acuity (Figure 1).
Distribution of specific thrombotic events among patients with APL (n = 11). Bars represent the number of events reported for each thrombotic category, including DVT (Deep Venous Thrombosis), SVT (Superficial Venous Thrombosis), pulmonary embolism (PE), line-associated thrombosis, hepatic vein thrombosis, stroke, and other sites (renal, splenic, intracardiac).
Summary of Studies Prevalence of Thrombosis in APL.
Abbreviations: DVT = deep vein thrombosis; PE = pulmonary embolism; APL = acute promyelocytic leukemia.
Notes: Incidence is reported as the proportion of patients with any thrombotic event within each study cohort. Thrombosis categories (arterial vs venous) reflect the definitions used in the original publications. This table is adapted and updated from Choudry et al. 16
General Characteristics Whole APL Cohort.
Abbreviations: IQR = interquartile range; Hg = hemoglobin; WBC = white blood cell count; PT = prothrombin time; INR = international normalized ratio; PTT = partial thromboplastin time.
Notes: Continuous variables are presented as mean (± Standard Deviation); categorical variables as number (percentage). Laboratory values reflect measurements at diagnosis.
For group comparison, we compared APL patients with thrombotic events to those who did not experience thrombosis. Baseline characteristics did not materially differ between groups for most variables. Mean age was virtually identical (35.5 years in the thrombosis group and 35.4 years in the non-thrombosis group; p = 0.96). Although all patients with thrombosis in the summary table were male, the comparison with the non-thrombosis group, which included 23% females, yielded a borderline p-value (0.073) and should be interpreted cautiously, given the small number of thrombosis cases. Regional background (Arab, 40% vs 24%; p = 0.28) showed no significant association. Table 3
Comparison of APL with or Without Thrombosis.
Abbreviations: SD = standard deviation; Hg = hemoglobin; WBC = white blood cell count; PT = prothrombin time; INR = international normalized ratio; PTT = partial thromboplastin time; APL = acute promyelocytic leukemia.
Laboratory indices at diagnosis were broadly similar between groups. Mean hemoglobin (9.1 vs 9.0 g/dL; p = 0.91), leukocyte count (33.6 vs 26.3 × 10^9/L; p = 0.60), platelet count (27.0 vs 31.7 × 10^9/L; p = 0.66), fibrinogen (2.2 vs 2.0 g/L; p = 0.61), and partial thromboplastin time (28.6 vs 27.9 s; p = 0.63) did not differ significantly. By contrast, the thrombosis group showed numerically higher prothrombin times (16.7 vs 14.7 s; p = 0.096), international normalized ratios (1.5 vs 1.4; p = 0.091), and D-dimers (29.2 vs 16.9 mg/L FEU; p = 0.094), each with borderline statistical significance. Within the thrombosis subset, the timing was skewed toward events during therapy, with 73% occurring during treatment and 27.3% at diagnosis. Within treatment-associated events, thrombosis occurred predominantly during induction (6/11, 54.5%), with fewer events during consolidation (2/11, 18.2%). The vascular territory was predominantly venous (55%), with arterial events accounting for 18%, and combined venous-arterial events for 27%. Deep vein thrombosis and pulmonary embolism were the most commonly documented entities; pulmonary embolism was present in 36% of cases, and deep vein thrombosis in 55%. Catheter-related thrombosis was reported in 4 patients. Organ-specific thrombi were documented in the renal or splenic regions of 2 patients, while one patient had an intracardiac thrombus. Anticoagulation was administered in 45% of thrombosis cases, while 18% received antiplatelet therapy. Concomitant bleeding was documented in 36% of patients. All bleeding events recorded in the thrombosis subgroup were major events (eg, gastrointestinal bleeding, intracranial hemorrhage, pulmonary hemorrhage) that required withholding anticoagulation and initiation of massive transfusion protocols. Among the four patients in the thrombosis subgroup who developed major bleeding, three events occurred at presentation or during induction therapy in the setting of active disseminated intravascular coagulation (DIC), while one event occurred during the consolidation phase. Moreover, most of these patients required ICU-level care as they required hemodynamic and respiratory support.
Induction backbones among patients with thrombosis spanned APML4-like anthracycline-containing protocols, the Lo-Coco ATRA + ATO regimen, and PETHEMA risk-adapted approaches. outcomes showed 73% alive and 27% deceased. Table 4. Finally, Early outcomes were assessed using 30-day mortality. Overall, 30-day mortality was 12.0% (9/75), and was higher in patients with thrombosis (18.2%, 2/11) than in those without thrombosis (10.9%, 7/64). The cause of death was mostly refractory septic shock that is not responding to vasopressor and antimicrobials support.
Thrombotic Event Characteristics, Management, and Outcomes among Patients with Thrombosis (n = 11).
Abbreviations: APL, acute promyelocytic leukemia; DVT, deep vein thrombosis; SVT, superficial venous thrombosis; PE, pulmonary embolism.
Notes: Event-type categories are not mutually exclusive (a patient may have >1 event type/site). “Timing” and “vascular territory” refer to the first documented thrombotic event; other rows summarize any occurrence during the APL course.
Discussion
Hemorrhagic manifestations in APL are well recognized and extensively described; however, APL-associated thrombosis remains underappreciated. The data available on this complication is limited and reports a wide range of incidents. Moreover, proposed risk factors for thrombosis are inconsistent and not robust, making it difficult to distinguish patients at higher risk of bleeding from those at higher risk of thrombosis in the absence of reliable predictive scoring systems.17,18
APL typically presents with a striking thrombo-hemorrhagic diathesis, with up to 80%–90% of patients exhibiting severe mucocutaneous, intracerebral, or intrapulmonary bleeding and/or thromboembolic events. The diagnosis is established histopathologically, with two predominant morphologic variants described: hypergranular and hypogranular. APL is usually accompanied by laboratory evidence of a hyperinflammatory and coagulopathic state. Although the clinical picture was initially ascribed to a disseminated intravascular coagulation (DIC)-like process, it is now evident that the underlying hemostatic disturbance is more complex than consumptive coagulopathy alone, involving multiple, interrelated procoagulant, fibrinolytic, and inflammatory pathways. 16
In contrast to bleeding, the determinants of thrombotic risk in APL are not well established. A recent analysis of 83 patients with high-risk disease (WBC > 10 × 109/L) found that a higher WBC/D-dimer ratio together with a lower D-dimer/fibrinogen ratio independently correlated with the development of thrombosis. The WBC/D-dimer ratio was markedly elevated in those who experienced thrombotic events compared with patients who primarily had hemorrhagic complications. While leukocytosis remains the most consistently reported clinical correlation of thrombosis, the available cohorts are relatively small, and additional, clearly validated risk factors for thrombotic events in APL have yet to be defined. 14
In our single-center cohort, thrombosis occurred in 15% of patients—within the ∼5%–20% reported across modern APL series and reviews, with venous events predominating (notably PE and catheter-related thrombosis) and most events clustering during treatment rather than exclusively at diagnosis.
Furthermore, early death remains a key outcome in APL. In our cohort, 30-day mortality was 12.0% (9/75). Thirty-day mortality was higher among patients with thrombosis (18.2% [2/11]) compared with those without thrombosis (10.9% [7/64]). This observation should be interpreted cautiously given the small number of events and retrospective design; nevertheless, it highlights that thrombotic complications may contribute to early vulnerability and reinforces the need for careful risk–benefit balancing between thrombosis prevention/treatment and bleeding risk during the earliest phase of therapy.
The absence of statistically significant baseline discriminants between groups—despite numerically higher PT/INR and D-dimer among those who thrombosed—is directionally consistent with prior observations that global coagulation activation tracks with both bleeding and thrombosis in APL, yet routine indices at presentation have limited predictive value for thromboembolism3,4
From a practical perspective, these finding it could support trajectory-based monitoring of coagulation parameters (serial PT/INR, D-dimer) from presentation through early therapy and argue for systematic line-care bundles to mitigate catheter-related risk.
In our cohort, nearly half of the patients who developed thrombosis received anticoagulation, yet more than one-third experienced bleeding. This juxtaposition reflects the day-to-day equipoise clinicians navigate in APL, weighing the risks of hemorrhage against the harms of untreated thrombosis. Contemporary guidance supports an individualized approach, and when anticoagulation is deemed necessary, it should be accompanied by coordinated transfusion support and protocol-driven correction of coagulopathy during early induction.3,19 From a practical point of view, we apply international guidelines and expert opinion in order to manage the situation of risk bleeding and thrombosis at the same time. According to the ISTH Scientific and Standardization Committee (SSC) guidance for cancer-associated thrombosis in the setting of thrombocytopenia, therapeutic LMWH can be given at full dose when the platelet count is ≥50 × 109/L. When platelet counts fall between 25–50 × 109/L, a dose reduction to approximately 50% of the full dose or a prophylactic-dose regimen is advised. If the platelet count drops below 25 × 109/L, anticoagulation should be temporarily withheld and then restarted at full dose once platelets recover above 50 × 109/L 20
Multiple studies indicate catheter-related thrombosis (CRT) is over-represented in APL/AML and often peaks after induction, aligning with our higher venous/CRT burden and supporting line-care bundles (ultrasound-guided placement, securement, routine patency checks, early troubleshooting) and conservative placement strategies. AML-wide CRT models also implicate inflammatory/infectious triggers—reinforcing meticulous catheter stewardship during neutropenia 18 and our cohort's management commentary. 19
Some studies have examined more detailed parameters, including flow cytometry markers, and more recent work by Sabljic et al 21 reported statistically significant differences within specific APL subsets between patients who developed VTE and those who did not.
In a Chinese cohort, FLT3-ITD and WT1 mutations, CD15 expression, and the PAI-1 4G/4G polymorphism were identified as independent risk factors for thrombosis on multivariable analysis. Unfortunately, in our cohort, molecular abnormalities other than PML–RARA were not systematically evaluated; however, FLT3-ITD was documented in one patient who developed thrombosis. 22
Our regimen distribution reflects modern practice, with substantial ATRA + ATO use for non–high-risk disease and continued application of APML4-type strategies where appropriate; randomized and large cohort data support high survival and low relapse with ATRA + ATO, although our series is underpowered to compare TE risk across backbones and risk strata19,23,24
Operationally, our data supports a low threshold for imaging (eg, CTPA for new hypoxemia/pleuritic pain), trajectory-based labs (serial D-dimer, PT/INR rather than one-off cutoffs), and individualized anticoagulation with coordinated transfusion support to counter the parallel bleeding risk in APL.
Integration of molecular covariates (FLT3-ITD, WT1, PML-RARα isoform) may refine prediction; larger datasets (including AI-enabled, multicenter registries) can explore joint bleeding-thrombosis risk in real time. 22
Strengths and Limitations
This study has several strengths. First, it represents, to our knowledge, one of the few detailed descriptions of thrombotic complications in APL from the GCC region, based on a consecutive cohort treated at a single national tertiary cancer center with unified diagnostic pathways and access to modern ATRA/ATO- and anthracycline-based protocols. The use of a well-defined cohort over a 10-year period, systematic case identification from departmental databases, and inclusion of all objectively documented thrombotic events (imaging- or clinician-adjudicated) helped minimize selection bias and under-reporting. Detailed characterization of thrombotic events—including timing, vascular territory, catheter association, concomitant bleeding, and antithrombotic management—provides a clinically relevant picture that is directly applicable to real-world practice in similar settings.
However, important limitations must be acknowledged. The retrospective, single-center design, the sample size was modest, and the number of thrombotic events restricted statistical power, precluding robust multivariable modeling of risk factors and limiting our ability to detect small-to-moderate associations.
Future Directions
Despite major improvements in survival, thrombotic complications remain an important and often underappreciated contributor to early morbidity and mortality in APL. Future work should therefore focus on several key areas.
First, there is a clear need for large, multicenter retrospective and prospective cohorts dedicated to thrombotic events in APL. Such collaborations would provide sufficient power to precisely estimate the incidence of venous and arterial thrombosis across contemporary ATRA/ATO-based regimens and to validate putative risk factors, including leukocytosis and emerging laboratory ratios (eg WBC/D-dimer, D-dimer/fibrinogen).
Second, integrated risk-prediction models that distinguish hemorrhagic from thrombotic risk are urgently needed. Composite scores combining clinical features (age, comorbidities, catheter use), leukemic burden, and dynamic hemostatic markers could help identify patients who might benefit from intensified monitoring, early venous thromboembolism prophylaxis, or modified anticoagulation strategies, while minimizing bleeding risk.
Third, prospective studies of antithrombotic strategies tailored to APL are required. These should address the optimal timing, intensity, and duration of anticoagulation in the context of severe thrombocytopenia; the role of primary thromboprophylaxis in well-defined high-risk subgroups; and specific approaches to catheter-associated thrombosis. Harmonizing anticoagulation algorithms with platelet thresholds and transfusion practices, and formally evaluating their safety and efficacy, would help move beyond expert opinion.
Fourth, mechanistic studies exploring APL-specific pathways of thrombogenesis—including tissue factor expression, cancer procoagulant activity, inflammatory cytokines, and interactions with endothelium and platelets—may uncover novel therapeutic targets. Selected agents such as thrombomodulin, TF pathway inhibitors, or other modulators of coagulation and fibrinolysis warrant systematic evaluation in preclinical models and carefully designed early-phase trials.
Finally, implementation-focused research is needed to optimize real-world care pathways: rapid recognition of suspected APL, early initiation of ATRA and supportive transfusions, standardized coagulation monitoring, and clear escalation protocols for suspected thrombosis. By combining biological insight, robust risk stratification, and pragmatic clinical trials, future work has the potential to meaningfully reduce the burden of both thrombosis and bleeding and further improve outcomes in APL.
Conclusion
In this single-center cohort of 75 APL patients, thrombosis occurred in approximately one in seven patients, most often during treatment and primarily in the venous circulation, with pulmonary embolism common and catheter-related events not rare. Routine baseline variables did not clearly segregate risk, although higher prothrombin time, international normalized ratio, and D-dimer at diagnosis showed borderline elevation among patients who developed thrombosis. These findings support vigilant thrombo-hemorrhagic surveillance from presentation through induction, attention to catheter stewardship, and careful, individualized use of anticoagulation with concurrent bleeding precautions. Larger, time-resolved datasets that integrate protocol backbone, molecular covariates, and standardized imaging triggers are needed to refine prediction and management.
Footnotes
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.