Factors predicting need for intervention and mortality in patients with spontaneous rectus sheath hematoma Authors: Süleyman Utku Çelik, Mehmet Mert Demircioğlu, Yasin Gülap, Mehmet Bahadır Demir, Güneş Görgülüer Batur, Emre Alp, Sacit Altuğ Kesikli Published: 2026 ABSTRACT BACKGROUNDSpontaneous rectus sheath hematoma (RSH) is a potentially life-threatening cause of acute abdominal pain, particularly in elderly and anticoagulated patients. While most cases are managed conservatively, a significant proportion requires interventional procedures, and mortality remains substantial. The aim of the study was to identify the predictors of interventional procedure requirements and 90-day mortality in patients with spontaneous RSH.METHODSThis retrospective study included consecutive adult patients diagnosed with spontaneous RSH at a tertiary care center between January 2017 and December 2023. Clinical, laboratory, and imaging data were collected. Multivariate logistic regression was used to identify independent predictors of interventional radiology treatment and 90-day all-cause mortality.RESULTSThe study consisted of 59 patients with a median age of 72 years, and 77.9% were females. Interventional radiology procedures were performed in 27.1% of cases, and 90-day mortality was 22.0%. Hematoma volume was independently associated with the need for intervention (odds ratio [OR]: 1.002; p=0.004). In a subgroup of patients undergoing computed tomography angiography (n=36), contrast extravasation was also a significant predictor (OR: 17.460; p=0.009). Independent predictors of 90-day mortality were advanced age (OR: 1.088; p=0.045) and greater hemoglobin decline (OR: 1.086; p=0.008).CONCLUSIONHematoma volume and contrast extravasation are significant predictors of the need for interventional treatment, while advanced age and hemoglobin decline predict 90-day mortality in spontaneous RSH. These findings support the use of objective radiologic and laboratory parameters for early risk stratification. Prospective multicenter validation is warranted to guide standardized management strategies in this high-risk population. INTRODUCTION Rectus sheath hematoma (RSH) is an increasingly recognized clinical entity that poses significant diagnostic and therapeutic challenges. While historically considered rare, the incidence of spontaneous RSH has risen in parallel with the increased use of anticoagulant and antiplatelet therapies, particularly among elderly patients with multiple comorbidities.[1,2] This condition results from bleeding into the rectus sheath space, most commonly due to rupture of the epigastric vessels or direct muscle injury, and often presents with acute abdominal pain and a palpable mass.[3] The clinical significance of RSH lies in its potential to mimic acute abdominal emergencies and its association with substantial morbidity, including hemorrhagic shock and mortality rates reaching 22% in high-risk populations.[4,5] Current understanding of RSH management remains largely based on retrospective studies and small case series, revealing critical gaps in evidence-based decision-making. While conservative management with blood product support and anticoagulation reversal is effective for the majority of patients, approximately 20–30% require invasive interventions such as angiographic embolization or surgical exploration.[6,7] Identifying patients at risk of conservative management failure who may require urgent intervention remains a persistent clinical challenge.[8,9] Recent efforts, such as the scoring model by Contrella et al.,[2] have proposed radiological and clinical predictors – hematoma volume, transfusion requirement, active extravasation – but lack prospective validation and widespread clinical adoption. Despite technical success rates approaching 100%, previous studies have also reported mortality rates ranging from 14% to 44% in patients undergoing embolization for spontaneous RSH, underscoring the limited physiologic reserve of this patient population.[10] Early identification of patients likely to require intervention may help reduce mortality by enabling timely and appropriate management, while also preventing unnecessary procedures in low-risk individuals. Importantly, once minimal hemodynamic deterioration is observed, delaying intervention in these patients may worsen outcomes. Earlier treatment may allow for the use of less invasive strategies, thereby reducing procedure-related morbidity and improving overall prognosis.[10,11] This study was designed to address these knowledge gaps by analyzing a consecutive cohort of patients with spontaneous RSH managed at a tertiary care center. We aimed to identify factors associated with the need for interventional radiology procedures and to determine independent predictors of 90-day mortality in patients with spontaneous RSH. MATERIALS AND METHODS Study Design and Patients The study was conducted in accordance with the Helsinki Declaration. Following the approval of the Institutional Ethics Committee of Gulhane Scientific Research Ethics Committee (date: 10.09.2024, decision number: 2024-391), a retrospective study was conducted to evaluate patients diagnosed with spontaneous RSH between January 2017 and December 2023 at a tertiary care center. All adult patients were identified from the hospital electronic database. Patients aged 18 years or older who were diagnosed with spontaneous RSH and either presented to the emergency department or were inpatients in other departments and consulted by general surgery during the study period were included. Patients were excluded if the hematoma was iatrogenic or traumatic in origin or if their data were incomplete. Data Collection and Variables Patient data were retrospectively collected from hospital medical records. The variables included demographic characteristics, comorbidities, use of anticoagulant or antiplatelet medications, type of hospital admission, laboratory values (hemoglobin, hematocrit, platelet count, prothrombin time, international normalized ratio, creatinine), diagnostic methods, hematoma volume, blood transfusion status, treatment strategies, length of hospital stay, and clinical outcomes. Definitions Hematoma volume was calculated using the ABC/2 method, defined as maximal length × maximal width × maximal thickness divided by 2.[10] When abdominal computed tomography (CT) was performed, the hematoma was classified according to the Berná classification.[12] 90-day mortality was defined as death from any cause within 90 days of the index admission. 90-day readmission was defined as any hospital readmission occurring within 90 days after discharge from the index hospitalization. Patient Follow-up and Outcomes All patients were followed for a period of 90 days after discharge to assess survival and clinical outcomes. The primary outcome of the study was 90-day mortality. The secondary outcome was the requirement for interventional procedures during hospitalization. Statistical Analysis Normal distribution of continuous variables was evaluated with quantile-quantile plots, histograms, and the Shapiro–Wilk test. Continuous variables were presented as median values with minimum and maximum ranges. Categorical variables were expressed as counts and percentages. The Wilcoxon Mann–Whitney U test was used for comparison of continuous variables, while the Pearson χ2 test or Fisher’s exact test was employed for categorical variables. Multivariate logistic regression analysis was performed using the backward stepwise method based on the likelihood ratio to identify independent predictors of 90-day mortality and the need for interventional procedures. All variables with a p<0.20 in the univariate analysis and all clinically significant variables regardless of statistical significance were included in the multivariate models. The linearity assumption between continuous independent variables and the logit of the dependent variable was evaluated using the Box-Tidwell procedure,[13] and multicollinearity was assessed using the variance inflation factor and tolerance statistics. The goodness-of-fit of the final multivariate models was assessed using the Hosmer-Lemeshow test. Adjusted odds ratios (ORs) and 95% confidence intervals (CIs) were reported for statistically significant predictors. A two-tailed p<0.05 was considered statistically significant. All statistical analyses were conducted using the Statistical Package for the Social Sciences version 26.0 (IBM Corp., Armonk, NY, USA), and data visualizations were created using GraphPad Prism version 8.0 (GraphPad Software, San Diego, CA, USA). RESULTS Patient Characteristics A total of 59 patients diagnosed with spontaneous RSH were included in the study. The median age was 72 years (range: 18–94), and 77.9% were female. During hospitalization, 16 patients underwent interventional radiology procedures (27.1%), of whom 3 required repeat embolization. Surgical intervention was required in only one patient due to persistent hemodynamic instability and failure to achieve hemorrhage control with radiological intervention. Surgical exploration successfully achieved hemostasis and resulted in stabilization of the patient’s condition. The overall 90-day mortality rate was 22.0%. Demographic, laboratory, and clinical characteristics of the patients are presented in Table 1. Table 1Demographic, clinical, and laboratory characteristics of patients with spontaneous rectus sheath hematoman=59 (%)Age, years72 (18–94)Female gender, n (%)46 (77.9)Hypertension, n (%)45 (76.3)Coronary artery disease, n (%)19 (32.2)Diabetes mellitus, n (%)20 (33.9)COPD, n (%)19 (32.2)Anticoagulation, n (%)41 (69.5)Antiplatelet agents, n (%)33 (55.9)Admission types, n (%) Emergency department admission40 (67.8) Inpatient-onset events19 (32.2)Hemoglobin (at admission), g/dL11.1 (7.2–17.0)Hemoglobin decline rate, %18.2 (0–61.7)Hematocrit (at admission), %34.5 (22.9–50.0)Hematocrit decline rate, %20.3 (0–59.5)Platelets,×109/L237 (16–579)Prothrombin time, s13.0 (8.5–144.3)INR1.1 (0.9–14.2)Creatinine, mg/dL0.98 (0.53–4.66)Diagnostic procedures, n (%) Only US23 (39.0) US and CT36 (61.0)CT classification, n (%)* Berna I12 (33.3) Berna II6 (16.7) Berna III18 (50.0)Extravasation on CTA, n (%)*15 (41.7)Hematoma volume, cm3144 (10–10927)Blood transfusion, n (%)28 (47.5)Red blood cell transfusion, units0 (0–27)Fresh frozen plasma transfusion, units0 (0–35)Platelet transfusion, units0 (0–11)Management, n (%) Conservative treatment43 (72.9) Interventional procedure16 (27.1)Surgical treatment, n (%)1 (1.7)Intensive care unit stay, n (%)19 (32.2)Length of hospital stay, day7 (0–58)90-day readmission, n (%)6 (10.2)90-day mortality, n (%)13 (22.0)CT: Computed tomography; CTA: Computed tomography angiography; COPD: Chronic obstructive pulmonary disease; US: Ultrasound. *Only patients who underwent CT were analyzed (n=36). Demographic, clinical, and laboratory characteristics of patients with spontaneous rectus sheath hematoma CT: Computed tomography; CTA: Computed tomography angiography; COPD: Chronic obstructive pulmonary disease; US: Ultrasound. *Only patients who underwent CT were analyzed (n=36). Predictors of Interventional Procedure Requirement Comparative analysis revealed that patients requiring interventional management had significantly lower admission hemoglobin (9.9 vs. 11.5 g/dL, p=0.038) and hematocrit levels (29.3% vs. 36.8%, p=0.032), greater hemoglobin (25.3% vs. 10.7%, p=0.002) and hematocrit decline rates (27.8% vs. 11.8%, p=0.004), and higher hematoma volumes (1070 vs. 49 cm3, p<0.001). In addition, Berna Type III classification and contrast extravasation on CT angiography (CTA) were significantly more common among these patients (p=0.007 and p<0.001, respectively). Blood transfusion, intensive care unit (ICU) admission, and length of hospital stay were also significantly more prevalent in the interventional group (p=0.002, p=0.016, and p<0.001, respectively) (Table 2). Table 2Comparison of patients’ characteristics between those managed conservatively and those who underwent interventional procedures for spontaneous rectus sheath hematomaConservative management (n=43) (%)Interventional procedure (n=16) (%)pAge, years71 (18–94)73.5 (21–88)0.573Female gender, n (%)35 (81.4)11 (68.8)0.311Hypertension, n (%)32 (74.4)13 (81.3)0.738Coronary artery disease, n (%)14 (32.6)5 (31.3)0.924Diabetes mellitus, n (%)13 (30.2)7 (43.8)0.329COPD, n (%)15 (34.9)4 (25.0)0.470Anticoagulation, n (%)27 (62.8)14 (87.5)0.111Antiplatelet agents, n (%)22 (51.2)11 (68.8)0.226Admission types, n (%)0.247 Emergency department admission31 (72.1)9 (56.3) Inpatient-onset events12 (27.9)7 (43.7)Hemoglobin (at admission), g/dL11.5 (7.4–17.0)9.9 (7.2–14.1)0.038Hemoglobin decline rate, %10.7 (0–61.7)25.3 (13.2–52.3)0.002Hematocrit (at admission), %36.8 (22.9–50.0)29.3 (23.5–42.1)0.032Hematocrit decline rate, %11.8 (0–59.5)27.8 (14.6–48.5)0.004Platelets,×109/L239 (16–479)229 (100–579)0.805Prothrombin time, s13.2 (8.5–144.3)13.1 (8.6–105.8)0.580INR1.1 (0.9–14.2)1.2 (0.9–9.7)0.429Creatinine, mg/dL1.01 (0.53–4.66)0.91 (0.54–3.93)0.727Diagnostic procedures, n (%)0.052 Only US20 (46.5)3 (18.7) US and CT23 (53.5)13 (81.3)CT classification, n (%)*0.007 Berna I11 (47.8)1 (7.7) Berna II5 (21.7)1 (7.7) Berna III7 (30.4)11 (84.6)Extravasation on CTA, n (%)*4 (17.4)11 (84.6)<0.001Hematoma volume, cm349 (10–1566)1070 (38–10927)<0.001Blood transfusion, n (%)15 (34.9)13 (81.3)0.002Red blood cell transfusion, units0 (0–27)3.5 (0–10)<0.001Fresh frozen plasma transfu-sion, units0 (0–35)2 (0–14)<0.001Platelet transfusion, units0 (0–11)0 (0–3)0.480Intensive care unit stay, n (%)10 (23.3)9 (56.3)0.016Length of hospital stay, day4 (0–50)15 (1–58)<0.00190-day readmission, n (%)4 (9.3)2 (12.5)0.65890-day mortality, n (%)8 (18.6)5 (31.3)0.311CT: Computed tomography; CTA: Computed tomography angiography; COPD: Chronic obstructive pulmonary disease; US: Ultrasound. *Only patients who underwent CT were analyzed (n=36). Comparison of patients’ characteristics between those managed conservatively and those who underwent interventional procedures for spontaneous rectus sheath hematoma CT: Computed tomography; CTA: Computed tomography angiography; COPD: Chronic obstructive pulmonary disease; US: Ultrasound. *Only patients who underwent CT were analyzed (n=36). In the multivariate analysis, hematoma volume was independently associated with the need for interventional procedures (OR: 1.002; 95% CI: 1.001–1.003; p=0.004), whereas blood transfusion was retained in the final model but did not reach statistical significance (OR: 5.158; 95% CI: 0.940–28.316; p=0.059) (Figure 1). Multivariate analysis for predictors of the need for interventional procedures and 90-day mortality in patients with spontaneous rectus sheath hematoma. In a predefined computed tomography angiography subgroup, contrast extravasation emerged as an additional independent predictor (odds ratio: 17.460; 95% confidence interval: 2.065–147.615). CI: Confidence interval. Furthermore, a subgroup analysis of 36 patients who underwent evaluation with CTA revealed that, in addition to hematoma volume, the presence of contrast extravasation on imaging was also an independent predictor of interventional radiologic management (OR: 17.460; 95% CI: 2.065–147.615; p=0.009). Predictors of 90-Day Mortality A total of 7 patients (11.9%) died within 30 days, and 13 patients (22.0%) experienced 90-day mortality. Table 3 shows univariate analysis of demographic and clinical factors associated with 90-day mortality. Patients experiencing mortality tended to be significantly older (81 vs. 69 years; p=0.002). They also had higher rates of inpatient-onset hematomas (p=0.018), greater hemoglobin and hematocrit decline rates (p=0.005 and p=0.003, respectively), elevated serum creatinine levels (p=0.019), and higher hematoma volumes (p=0.043). The mortality group had significantly more frequent blood product transfusions and ICU admissions (p=0.016 and p=0.018, respectively). Table 3Comparative analysis of patient characteristics in terms of 90-day mortality among patients with spontaneous rectus sheath hematomaAlive at 90 days(n=46) (%)90-day mortality(n=13) (%)pAge, years69 (18–94)81 (40–88)0.002Female gender, n (%)36 (78.3)10 (76.9)1.000Hypertension, n (%)33 (71.7)12 (92.3)0.159Coronary artery disease, n (%)15 (32.6)4 (30.8)1.000Diabetes mellitus, n (%)15 (32.6)5 (38.5)0.746COPD, n (%)14 (30.4)5 (38.5)0.738Anticoagulation, n (%)29 (63.0)12 (92.3)0.049Antiplatelet agents, n (%)25 (54.3)8 (61.5)0.645Admission types, n (%)0.018 Emergency department admission35 (76.1)5 (38.5) Inpatient-onset events11 (23.9)8 (61.5)Hemoglobin (at admission), g/dL11.7 (7.2–17.0)10.7 (7.8–14.1)0.104Hemoglobin decline rate, %15.0 (0–51.4)30.4 (0–61.7)0.005Hematocrit (at admission), %36.2 (22.9–50.0)29.1 (23.5–42.1)0.122Hematocrit decline rate, %15.6 (0–53.0)32.5 (0–59.5)0.003Platelets, ×109/L229 (100–579)255 (16–517)0.674Prothrombin time, s13 (8.5–144.3)13.7 (9.4–80.5)0.712INR1.1 (0.9–14.2)1.2 (0.9–7.7)0.937Creatinine, mg/dL0.91 (0.53–4.34)1.41 (0.78–4.66)0.019Diagnostic procedures, n (%)0.548 Only US17 (37.0)6 (46.2) US and CT29 (63.0)7 (53.8)CT classification, n (%)*0.109 Berna I12 (41.4)0 (0) Berna II4 (13.8)2 (28.6) Berna III13 (44.8)5 (71.4)Extravasation on CTA, n (%)*11 (37.9)4 (57.1)0.418Hematoma volume, cm377.7 (10–10927)424 (10–3443)0.043Blood transfusion, n (%)18 (39.1)10 (76.9)0.016Red blood cell transfu-sion, units0 (0–10)2 (0–27)0.015Fresh frozen plasma transfusion, units0 (0–14)1 (0–35)0.094Platelet transfusion, units0 (0–0)0 (0–11)0.007Management, n (%)0.311 Conservative treatment35 (76.1)8 (61.5) Interventional procedure11 (23.9)5 (38.5)Intensive care unit stay, n (%)11 (23.9)8 (61.5)0.018Length of hospital stay, day5 (0–58)11 (0–50)0.04190-day readmission, n (%)4 (8.7)2 (15.4)0.605CT: Computed tomography; CTA: Computed tomography angiography; COPD: Chronic obstructive pulmonary disease; US: Ultrasound. *Only patients who underwent CT were analyzed (n=36). Comparative analysis of patient characteristics in terms of 90-day mortality among patients with spontaneous rectus sheath hematoma CT: Computed tomography; CTA: Computed tomography angiography; COPD: Chronic obstructive pulmonary disease; US: Ultrasound. *Only patients who underwent CT were analyzed (n=36). In multivariate analysis, two independent predictors of 90-day mortality were identified: advanced age (OR: 1.088; 95% CI: 1.002–1.182; p=0.045) and greater hemoglobin decline (OR: 1.086; 95% CI: 1.021–1.154; p=0.008). Admission hemoglobin level showed a borderline association with mortality (OR: 0.699; 95% CI: 0.458–1.066; p=0.096) (Fig. 1). DISCUSSION This study investigated factors predicting the need for intervention and 90-day mortality in patients diagnosed with spontaneous RSH at a tertiary care center. Our analysis revealed that hematoma volume is a significant independent predictor for requiring interventional procedures, while advanced age and hemoglobin decline independently predict 90-day mortality. These findings underscore the complex clinical profile of spontaneous RSH, highlighting distinct risk factors associated with procedural necessity versus overall survival in this vulnerable patient population. The need for angiographic embolization in our cohort (27.1%) was consistent with rates reported in previous studies, typically ranging from 20% to 30%.[6,7] Our multivariate analysis for the entire cohort identified hematoma volume as a significant independent predictor for requiring intervention (OR 1.002/cm3 increase, p=0.004). Blood transfusion showed a borderline association, likely reflecting its common use in patients already deemed unstable or having large bleeds, rather than being an independent driver for the decision to intervene. Furthermore, a specific subgroup analysis focusing on the patients who underwent CTA revealed that, alongside hematoma volume, the presence of contrast extravasation on imaging was also an independent predictor of the need for interventional radiology management. This finding is consistent with a previous report by Contrella et al.,[2] which identified extravasation on CTA, larger hematoma volume, transfusion of ≥4 units of packed red blood cells, and a rate of hemoglobin decrease (>0.25 g/dL/h) as key predictors of conservative management failure in patients with spontaneous RSH. In addition, our results also align with those of Porras Fimbres et al.,[14] who conducted a multivariate analysis in an independent cohort and similarly found hematoma volume and contrast extravasation on CTA to be the two most significant independent predictors of failure of conservative management. Other parameters, such as hemoglobin nadir and transfusion requirement, were not retained in their final model. Clinically, these results suggest that both a large hematoma volume and the presence of active contrast extravasation on CTA serve as crucial indicators of likely failure of conservative management, warranting early consultation with interventional radiology. Although the majority of RSH cases are managed conservatively, mortality rates remain non-negligible, ranging between 14% and 44% in previous series involving interventional treatment.[2,3,5,10] The 90-day mortality rate observed in our cohort (22%) is in line with existing literature and reinforces the concept that RSH, particularly in frail and elderly patients, often serves as a marker of systemic decompensation rather than an isolated event. Baekgaard et al.[4] documented an in-hospital mortality rate of 22%, with only 27% of deaths clearly linked to hemorrhagic complications. Similarly, Sheth et al.[3] reported a 14.8% overall mortality, none of which were directly ascribed to RSH. Jawhari et al.[10] noted a 30-day mortality rate of 44%, yet only one-third of these deaths were directly related to persistent or recurrent bleeding; the remainder resulted from secondary complications such as sepsis or cardiovascular events, reflecting the role of RSH as a marker of systemic deterioration. Contrella et al.[2] observed a 19% 30-day mortality in embolized patients, though most deaths were associated with underlying comorbidities rather than the bleeding episode itself. Collectively, these findings highlight that adverse outcomes in RSH are often driven by multisystem compromise rather than isolated hemorrhage. This broader pathophysiological context is further supported by our findings, in which advanced age and greater hemoglobin decline emerged as independent predictors of 90-day mortality in multivariate analysis. These results are consistent with those of Baekgaard et al.,[4] who similarly identified age and hemodynamic instability as major contributors to mortality risk. Departing from prior studies, this study targeted 90-day all-cause mortality, enabling a more comprehensive understanding of the determinants of patient outcomes. While 30-day or in-hospital mortality primarily reflects early procedural or bleeding-related events, a 90-day timeframe captures delayed complications such as infection, deconditioning, or organ failure – particularly relevant in older adults. While the extended follow-up period may introduce confounding from unrelated comorbidities, it offers a more comprehensive understanding of patient vulnerability and underscores the importance of holistic management beyond initial hemostasis. This study has several strengths that enhance the robustness and clinical applicability of its findings, foremost among them the inclusion of a consecutive patient cohort over a 7-year period at a tertiary care center, which supports internal validity, reduces selection bias, and ensures the representativeness of real-world clinical practice. Moreover, the 90-day follow-up allowed for a more comprehensive assessment of outcomes beyond the immediate phase of hemorrhagic control. However, it is not without limitations. The retrospective design inherently limits causal inference and introduces potential for missing data, misclassification, and unmeasured confounding – particularly in transfusion practices and intervention timing, which were not standardized. CTA was not uniformly applied, possibly introducing verification bias, and hematoma volume estimation through the ABC/2 method, although practical, may be imprecise in irregularly shaped hemorrhages. Finally, as a single-center study, external generalizability remains limited. Future studies should aim to validate these findings in prospective, multicenter cohorts and facilitate the development of standardized, evidence-based risk stratification models incorporating hematoma volume, contrast extravasation, and dynamic clinical parameters. Integration of bedside assessment tools—such as serial hemoglobin monitoring and point-of-care ultrasonography – may further enhance early identification of patients at risk of deterioration. In parallel, investigations assessing the feasibility, cost-effectiveness, and clinical impact of incorporating routine volumetric analysis into emergency imaging protocols are warranted. Finally, emphasis should be placed on evaluating long-term outcomes, including post-discharge functional status, quality of life, and the role of geriatric assessment in guiding comprehensive care strategies for this high-risk population. CONCLUSION This study identifies hematoma volume and contrast extravasation as significant predictors of the need for interventional treatment, while advanced age and hemoglobin decline independently predict 90-day mortality in patients with spontaneous RSH. These findings underscore the importance of integrating objective radiological and laboratory parameters into early clinical assessment to guide timely intervention and prognostication. Prospective validation and the development of standardized management pathways are warranted to improve risk stratification and optimize outcomes in this complex clinical entity.