Perioperative Outcomes in Factor XI Deficiency: A Systematic Review

ABSTRACT

Background

Congenital factor XI (FXI) deficiency has a variable bleeding phenotype, and FXI activity alone does not predict perioperative bleeding. Management practices remain heterogeneous, and available evidence derives from case reports, small case series, and retrospective cohorts. This systematic review evaluated perioperative bleeding outcomes and management strategies in FXI‐deficient patients undergoing surgical procedures.

Methods

We systematically searched five databases (2016–2024). Two reviewers independently screened studies, extracted data, and assessed methodological quality. Because of heterogeneity among studies, cohort data were synthesized qualitatively. Case reports were summarized descriptively, and an exploratory framework combining FXI category and fibrinolytic risk of the procedure was constructed to illustrate patterns in bleeding outcomes.

Results

Twenty‐one studies met inclusion criteria: four retrospective cohorts and seventeen case reports, describing more than 590 procedures. Cohort bleeding rates ranged from 0% to 19%, with major bleeding ≤ 7%. FXI activity showed no consistent association with perioperative bleeding. Procedural factors and personal bleeding history are more often aligned with outcomes. Most case reports involved severe FXI deficiency and used FFP‐based prophylaxis, often with antifibrinolytics. Bleeding events in case reports were infrequent. In the exploratory two‐dimensional visualization, bleeding appeared more common among patients with FXI ≤ 10 IU/dL undergoing high‐fibrinolytic procedures; however, event numbers were small and prophylaxis was widely used, limiting interpretation.

Conclusions

FXI levels alone appear insufficient to predict bleeding risk in FXI‐deficient patients. A multifactorial approach, integrating surgical risk and haemostatic strategy, may be required, and further study is needed to optimize preoperative management strategies for FXI‐deficient patients.

Clinical trial registration

Clinical trial registration was not applicable for this study.

1. Introduction

Factor XI (FXI) deficiency is a rare inherited bleeding disorder characterized by a highly variable and often unpredictable bleeding phenotype, particularly in the perioperative setting. Although FXI plays a key role in amplification of thrombin generation through activation of factor IX, bleeding risk in FXI deficiency does not correlate reliably with baseline factor activity [1, 2]. As a result, patients with similar FXI levels may experience markedly different bleeding outcomes when exposed to surgical or traumatic challenges.

FXI deficiency is classically inherited in an autosomal recessive pattern and affects approximately 1 in 1,000,000 individuals in the general population, with higher prevalence in certain populations such as Ashkenazi Jews due to founder mutations [2, 3]. Clinically, FXI deficiency is commonly categorized as severe when FXI activity is < 20 IU/dL and as partial or mild when activity ranges from 20 IU/dL to the lower limit of normal (generally 60–65 IU/dL) [1]. However, this biochemical classification has limited utility for predicting bleeding risk. Unlike haemophilia A or B, spontaneous bleeding is uncommon, and haemorrhagic events typically occur in association with trauma, surgery, or procedures involving areas with high fibrinolytic activity, such as the oral cavity, nasopharynx, and genitourinary tract [4, 5, 6]. The variability in bleeding phenotype exists even among individuals with comparable FXI levels or within the same family, underscoring the contribution of additional modifiers beyond FXI level alone [1, 4, 5, 6, 7]. Among rare bleeding disorders, FXI deficiency is therefore distinguished by its particularly poor correlation between factor activity levels and clinical bleeding severity [8].

Perioperative management of patients with FXI deficiency is not well‐defined, and no standardized guidelines currently exist [5]. Prohemostatic treatment options include fresh frozen plasma (FFP), FXI concentrates, recombinant activated factor VII (rFVIIa), and antifibrinolytic agents such as tranexamic acid (TXA) [5, 9]. FXI concentrate is effective but carries a recognized thromboembolic risk and is not widely available [5]. Plasma‐based therapy may result in volume overload, whereas TXA is frequently used as an adjunctive or standalone option for minor procedures [10, 11, 12].

Given this uncertainty and the publication of several large cohort studies in recent years, an updated synthesis of perioperative bleeding outcomes and management strategies in congenital FXI deficiency is warranted. We therefore conducted a systematic review to evaluate the relationship between FXI activity, surgical context, and perioperative haemostatic interventions—including FFP, rFVIIa, and TXA—and bleeding outcomes, with the aim of informing clinical decision‐making and guiding future evidence‐based approaches to perioperative care in FXI‐deficient patients.

2. Methods

2.1. Study Design

We conducted a systematic review to evaluate perioperative bleeding risk, prophylactic strategies, and haemostatic outcomes in patients with congenital FXI deficiency undergoing surgical or obstetric procedures. We followed the Preferred Reporting Items for Systematic Reviews and Meta‐Analyses (PRISMA 2020) guidelines, with a structured search, dual screening, and predefined inclusion criteria.

2.2. Search Strategy

The reviewers performed a comprehensive electronic search across MEDLINE (Ovid), Embase, Web of Science, Scopus, and the Cochrane Central Register of Controlled Trials (CENTRAL), covering publications from January 2016 to November 2024 to extend the evidence summarized in a prior systematic review that included studies published up to 2016 [13]. The review combined Medical Subject Headings (MeSH) and free‐text terms related to ‘Factor XI deficiency,’ ‘haemophilia C,’ ‘perioperative,’ ‘surgery,’ ‘bleeding,’ and ‘management,’ using Boolean operators (AND/OR) and restricting the search to English‐language human studies. Reviewers manually screened the reference lists of included papers and relevant reviews to identify additional eligible studies. Supplementary Appendix 1 provides the complete search strategy for each database.

2.3. Eligibility Criteria

We included studies that (i) involved patients with congenital FXI deficiency, (ii) described perioperative bleeding outcomes, and (iii) reported perioperative management, including the use of FFP, rFVIIa, TXA, FXI concentrate, or Therapeutic plasma Exchange (TPE). Eligible study designs comprised randomized studies, prospective studies, retrospective cohort studies, case series, and case reports. Studies were eligible irrespective of patient age.

We excluded studies that (a) were published before 2016, (b) were reviews, expert opinions, or conference abstracts without original data, (c) focused on acquired FXI deficiency, or (d) lacked documentation of both treatment strategy and bleeding outcome.

2.4. Definitions

We classified bleeding outcomes as major, clinically relevant non‐major bleeding, or none, using International Society on Thrombosis and Haemostasis (ISTH) criteria when reported, or study‐specific definitions when standardized criteria were not applied [14, 15].

We categorized procedures as major or minor based on surgical extent [16] and as areas with high‐fibrinolytic (e.g., oral, nasopharyngeal, urologic, obstetric, gynecologic) or low‐fibrinolytic (e.g., cardiac, orthopedic, general) activities based on anatomical site and prior observational literature and was used as a descriptive framework rather than a validated risk stratification system [17].

Severe FXI deficiency is conventionally defined as FXI activity < 20 IU/dL; however, for exploratory descriptive comparisons, FXI activity was further stratified into three clinically meaningful categories: < 10 IU/dL, 10–20 IU/dL, and > 20 IU/dL. These categories were selected to reflect increasing severity within accepted clinical definitions and to provide a pragmatic, hypothesis‐generating framework rather than to define strict treatment thresholds.

2.5. Study Selection

Following the search, we imported all retrieved citations into Covidence (Veritas Health Innovation, Melbourne, Australia), removed duplicates, and screened records for relevance. Two reviewers (Z.H. and R.M.) independently assessed titles, abstracts, and full texts against inclusion criteria, resolving discrepancies by consensus or adjudication by a third reviewer (D.M.) when required. Figure 1 summarizes the study selection process and reasons for exclusion.

FIGURE 1.

PRISMA flow diagram of study selection process.

2.6. Data Extraction

We extracted data using a standardized, pilot‐tested form. Extracted variables included study design and country; number of patients and procedures; demographic characteristics (age, sex, FXI level, bleeding history); procedure type (major or minor surgery; anatomical site with high or low fibrinolytic activity); perioperative management (type, dose, and timing of FFP, rFVIIa, TXA, FXI concentrate, or TPE); bleeding outcomes (major, clinically relevant non‐major bleeding, or none); and reported thrombotic events.

Data was extracted at the procedure level when reported. In retrospective cohort studies, procedures were stratified according to reported FXI activity and surgical characteristics rather than by individual patient. For case reports, individual patient‐level data were captured. When outcomes or management details were ambiguously reported, both reviewers independently reviewed the source articles and resolved discrepancies by consensus using a conservative classification approach.

2.7. Quality and Risk‐of‐Bias Assessment

We assessed study quality using design‐specific frameworks. For cohort and case‐control studies, we applied the Newcastle–Ottawa Scale (NOS) [18] to evaluate participant selection, comparability, and outcome assessment, rating studies as low (< 6), moderate [6–7.5], or high quality (> 7.5). We appraised case reports using the CARE guidelines to assess completeness of patient description, intervention detail, and outcome documentation [19].

We further evaluated risk of bias using the Joanna Briggs Institute (JBI) [20] critical appraisal checklists, applying the appropriate version for each study design. Assessed domains included patient selection, exposure and outcome measurement, and adequacy of follow‐up, with each item rated as ‘Yes,’ ‘No,’ or ‘Unclear.’ Two reviewers independently completed all assessments and resolved disagreements by consensus. In accordance with JBI guidance, we did not assign composite scores; instead, qualitative judgments informed the overall risk‐of‐bias appraisal.

2.8. Data Synthesis and Analysis

Because of the heterogeneity in study designs and outcome definitions, we synthesized data qualitatively and summarized findings narratively.

For individual‐level data from case reports and small series, we organized FXI levels and bleeding outcomes within an exploratory two‐dimensional framework that combined FXI activity category with the fibrinolytic risk of the surgical site to illustrate general trends. Given the limited sample size and variability in outcome reporting, we did not perform formal hypothesis testing or between‐group comparisons.

3. Results

3.1. Study Selection

Figure 1 (PRISMA flow diagram) presents the study selection process.

3.2. Study Characteristics

Twenty‐one studies met the inclusion criteria, comprising four retrospective cohort studies [21, 22, 23, 24] and seventeen individual case reports [25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41]. Two cohort studies reporting exclusively on patients with mild FXI deficiency (FXI > 20 IU/dL) were identified but excluded from the primary synthesis, as their populations do not reflect the management challenges specific to severe deficiency. The cohort studies originated from the United States and Israel and evaluated between 3 and 256 procedures across high‐ and low‐fibrinolytic surgical sites, obstetric deliveries, cardiac operations, and mixed surgical settings. Reported baseline FXI activity ranged from severe deficiency (< 10–20 IU/dL) to mild deficiency (> 20 IU/dL), with one obstetric cohort including a broader FXI activity range due to genetic inclusion criteria, as described in Table 1. Reported major bleeding rates ranged from 0% to 7%, although definitions varied among studies. Use of prophylactic FFP, TXA, rFVIIa, or TPE also varied, and several studies identified personal bleeding history—rather than FXI level alone—as a predictor of perioperative bleeding. Table 1 summarizes procedural characteristics, prophylaxis strategies, and bleeding outcomes from the retrospective studies.

TABLE 1.

Characteristics of retrospective cohort studies on perioperative management in factor XI deficiency.

Study (Year)	Country	N Procedures	FXI range (IU/dL)	Bleeding outcomes	Perioperative management	Key insight
Wu et al., 2024 [21]	USA	256	1–18 (median = 3; all severe)	Total: ≈19%2 Major: ≈7%2	FFP ± TXA (≈60% of procedures)2	Major bleeds occurred predominantly without prophylaxis. FFP ± TXA significantly reduced bleeding (p < 0.05). Neuraxial anesthesia used safely. No thrombotic events.
Handa et al., 2023 † [22]	USA	298 (252 obstetric + 46 surgical)	2–190 (median 50; 12% severe) †	Total: 13% (33/252) Major: 0% *	FFP in 8.3%; TXA in 2.8%; RBC transfusion in bleeding cases only	Personal bleeding history strongest predictor (OR 5.92, p = 0.001). Higher FXI associated with lower risk (OR 0.72/10 IU/dL, p = 0.05). Neuraxial anesthesia safe in 174 procedures including 5 with FXI < 10 IU/dL.
Salomon et al., 2019 [24]	Israel	12	< 9 (all severe)	Total: 0% Major: 0%	Single low‐dose rFVIIa (10–15 µg/kg at wound closure) + TXA (1 g pre‐op then q6h ×3–5 days)	Low dose rFVIIa + TXA achieved complete haemostatic control without plasma use or thrombosis. Thrombin generation remained below normal (non‐thrombogenic).
Shalabi et al., 2020 [23] (FXI subgroup only)	Israel	3	< 20 (all severe)	Total: 0% Major: 0%	FFP to target FXI ≈30–40 IU/dL pre‐/intra‐op + TXA	Major cardiac surgery is feasible in severe FXI deficiency with moderate FXI targets and TXA support. No bleeding or thrombotic events.

Note: 1 Major bleeds defined per study‐specific or ISTH/ACOG criteria (≥ 1000 mL blood loss, transfusion, re‐operation, or hemodynamic instability).

2 Wu et al. [21] values derived from Figure 1 of the original publication (visual estimation).

Abbreviations: AVR, aortic valve replacement; CABG, coronary artery bypass graft; CPB, cardiopulmonary bypass; FFP, fresh frozen plasma; OR, odds ratio; RBC, red blood cells; rFVIIa, recombinant activated factor VII; TXA, tranexamic acid.

^*Handa et al. reported clinically significant bleeding in 13% of procedures; none met strict major‐bleed criteria (transfusion, re‐operation, hemodynamic instability).

^†Handa et al. cohort included patients across the full FXI activity spectrum (median 50 IU/dL, range 2–190 IU/dL) due to inclusion of genetically confirmed FXI variants; only 12% had severe deficiency (< 20 IU/dL).

The seventeen case reports published between 2016 and 2024 described individual‐level data across cardiac, abdominal, orthopedic, neurosurgical, urologic, gynecologic, and oral procedures. Most cases involved severe FXI deficiency (< 10 IU/dL) and used prophylactic FFP, antifibrinolytics, FXI concentrate or TPE. Major bleeding occurs in a minority of cases with prophylaxis. Table 2 presents individual patient characteristics, management approaches, and perioperative outcomes.

TABLE 2.

Summary of individual case reports on perioperative management in congenital FXI deficiency (2016–2024).

Author (Year)	Age (y) / FXI (IU/dL) / Severity	Procedure (Site)	Perioperative management	Bleeding outcome	Fibrinolytic risk †
Pagano MB et al., 2016 [40]	65 / 7 / Severe	Total hip replacement (orthopedic)	TPE pre‐op; FFP 3 U/day POD1–7; EACA peri‐op	No bleeding	Low
Hou X et al., 2016 [34]	50 / 0.1 / Severe	Radical gastrectomy (abdominal)	FFP 800 mL pre‐op + 200 mL intra‐op + 400 mL/day ×3 post‐op	No bleeding	Low
Edahiro Y et al., 2016 [41]	83 / <1 / Severe	Burr‐hole craniostomy (neurosurgical)	FFP 4 U (∼10 mL/kg) pre‐op	No bleeding	Low
Petroulaki A et al., 2017 [29]	73 / 2.9 / Severe	Aortic valve replacement (cardiac)	FXI concentrate (Hemoleven 1000 U pre‐op + POD3); TXA intra‐op; FFP + RBC + platelets intra‐op	No bleeding	Low
McCarthy ML et al., 2018 [39] (Procedure 1)	74 / 52 / Mild	Total thyroidectomy (neck)	No prophylaxis (FXI deficiency undiagnosed); FFP given post‐operatively for bleed	Yes—requiring cauterization	Low
McCarthy ML et al., 2018 [39] (Procedure 2)	74 / 52 / Mild	Wide local excision of melanoma (skin)	FFP prophylactically post‐op (based on Procedure 1 diagnosis)	No bleeding	Low
Fitzsimons MG et al., 2018 [38]	62 / <1 / Severe	Redo aortic root replacement (cardiac)	FFP 15 U pre‐op; FFP 10 U + platelets intra‐op; EACA 10 g bolus then 2 g/h ×72 h post‐op	No bleeding	Low
Kondo H et al., 2019 [26]	82 / 3 / Severe	Laparoscopic ileocolic resection (abdominal)	FFP 8 U pre‐op (4 U/day ×2 days)	No bleeding	Low
Ince ME et al., 2020 [36]	68 / 11.4 / Moderate	CABG ×3 (cardiac)	FFP 8 U pre‐op + 6 U intra‐op; TXA 30 mg/kg bolus + infusion throughout	No bleeding	Low
Serikyaku H et al., 2021 [30]	66 / <1 / Severe	Open‐door laminoplasty C3–C6 (spine)	FFP 4 U pre‐op + 4 U intra‐op (8 U total)	No bleeding	Low
Honda S et al., 2023 [31]	79 / 1.3 / Severe (+ FVIII deficiency)	CABG (cardiac)	Recombinant FVIII continuous infusion + FFP 1440 mL intra‐op; FFP 480 mL/day POD1–7	No bleeding	Low
Kianian S et al., 2023 [37]	75 / <2 / Severe (FFP allergy)	CABG (cardiac)	FFP 5 U (slow infusion with premedication); pre‐op FXI raised to 25 IU/dL	No bleeding	Low
Kato T et al., 2023 [28]	43 / <3 / Severe	Multiple tooth extractions—2 sessions (oral)	FFP 2 U/day ×2 pre‐op each session; TXA 1 g IV (maxillary session); local haemostasis	No bleeding	High
Gurcan M et al., 2023 [32] (TURBT 2—index)	42 / 3.7 / Severe	TURBT (urologic)	No prophylaxis; FFP 6 U + RBC 2 U given post‐operatively after diagnosis	Yes—requiring return to OR	High
Gurcan M et al., 2023 [32] (TURBT 3—with prophylaxis)	42 / 3.7 / Severe	TURBT—recurrence (urologic)	FFP 2 U ×3 daily from day before surgery through POD3	No bleeding	High
Beomonte Zobel L et al., 2024 [33]	49 / 9 / Severe	Laparoscopic sleeve gastrectomy (abdominal)	Virus‐inactivated FFP 10 mL/kg from 1 h pre‐op through POD4; enoxaparin POD1–25	No bleeding	Low
Supthut H et al., 2024 [25] (Surgery 1—2018)	64 / 6 / Severe	Total hip arthroplasty (orthopedic)	FFP 6 U pre‐op + 4 U 1 h pre‐op; TXA 1 g pre‐op + 1 g over 8 h; FFP 1 U post‐op	Yes—significant bleeding	Low
Supthut H et al., 2024 [25] (Surgery 2—2019)	64 / 6 / Severe	Resection of hip ossifications (orthopedic)	FFP 6 U over 3 days pre‐op; TXA	No bleeding	Low
Supthut H et al., 2024 [25] (Surgery 3—2022)	64 / 6 / Severe	Resection of hip ossifications (orthopedic)	FFP 6 U (2 U pre‐op day before + 4 U 1 h pre‐op); TXA	No bleeding	Low
Tian Y et al., 2024 [35]	74 / 0.7 / Severe	Minimally invasive esophagectomy (thoraco‐abdominal)	FFP 15 mL/kg/day from 1 week pre‐op through POD10; APTT + FXI:C monitored daily	No bleeding	Low
Martínez‐López P et al., 2024 [27]	25 / 40 / Mild	Laparoscopic salpingectomy (gynecologic)	TXA 1 g IV + FFP 250 mL pre‐op; ROTEM‐guided	No bleeding	High

Note: Supthut et al. [25] contributed 3 separate surgical encounters for 1 patient. Gurcan et al. [32] contributed 2 TURBT procedures (index surgery without prophylaxis; repeat surgery with prophylaxis). McCarthy et al. [39] contributed 2 procedures (thyroidectomy with post‐op bleeding; melanoma excision without bleeding). Total: 21 procedures across 17 patients.

Abbreviations: CABG, coronary artery bypass graft; EACA, ε‐aminocaproic acid; FFP, fresh frozen plasma; ITA, internal thoracic artery; OR, operating room; POD, postoperative day; RBC, red blood cells; ROTEM, rotational thromboelastometry; TPE, therapeutic plasma exchange; TURBT, transurethral resection of bladder tumour; TXA, tranexamic acid.

^†High fibrinolytic risk: oral, nasopharyngeal, urologic, and gynecologic/obstetric sites. All other sites classified as low fibrinolytic risk.

3.3. Quality Assessment and Risk of Bias

Design‐specific quality appraisal indicated variable methodological quality across the included studies. Four studies met high‐quality criteria, fifteen met moderate‐quality criteria, and two met low‐quality criteria. Among cohort studies, differences in comparability, follow‐up completeness, and outcome reporting contributed to variation in scoring.

Domain‐level assessment with the Joanna Briggs Institute (JBI) tools identified a low‐to‐moderate risk of bias across study designs. Most cohort studies provided adequate information on patient selection, exposure measurement, and outcome ascertainment. Salomon 2019 met low‐risk criteria, whereas Shalabi 2020, Handa 2023 and Wu 2024 met moderate‐risk criteria due to incomplete follow‐up and reporting limitations.

Case reports met moderate‐to‐high reporting standards and were rated as low risk of bias, with missing elements typically involving timelines, diagnostic details, or full dosing information. Tables S1 and S2 present study‐level quality assessments and domain‐specific risk‐of‐bias evaluations.

3.4. Cohort and Institutional Studies

Four retrospective institutional cohorts published between 2017 and 2024 reported approximately 323 procedures performed in patients with congenital FXI deficiency (Table 2). These studies included surgical and obstetric settings across high‐ and low‐fibrinolytic sites. Overall bleeding rates ranged from 0% to approximately 19%, and major bleeding was ≤ 7%.

3.4.1. Individual Cohorts Contributed the Following

Wu et al. [21]: 256 mixed surgical procedures in patients with severe deficiency (FXI 1–18 IU/dL, median ≈3); bleeding ≈19% (≈7% major); lower bleeding rates in procedures receiving FFP with or without TXA.

Handa et al. [22]: 252 obstetric and 46 surgical procedures; clinically significant bleeding in 13%, predominantly postpartum haemorrhage. Only 8 of 33 bleeding events occurred in patients with FXI < 20 IU/dL.

Salomon et al. [24]: 12 major procedures in severe deficiency; no bleeding with low dose rFVIIa plus TXA.

Shalabi et al. [23]: three cardiac surgeries managed with FFP plus TXA; no bleeding events.

Haemostatic management strategies varied and included FFP, TXA or epsilon‐aminocaproic acid (EACA), rFVIIa, and TPE. Several cohorts reported that prophylaxis decisions were based on procedural considerations or prior bleeding history rather than FXI activity alone. No thrombotic events were identified. Both Wu and Handa reported the use of neuraxial anesthesia in selected patients with low FXI activity and no prior bleeding history without adverse haemorrhagic outcomes, including five patients with FXI < 10 IU/dL in the Handa cohort.

3.4.2. Predictors of Bleeding

Two cohorts evaluated factors associated with bleeding. Handa et al. reported an association between personal bleeding history and obstetric bleeding (OR 5.92, p = 0.001), and an association between higher FXI activity and lower bleeding risk (OR 0.72 per 10 IU/dL, p = 0.05). Procedural factors, including cesarean delivery and high‐fibrinolytic sites, were also associated with bleeding (OR 2.83, p = 0.04).

Wu et al. [21] reported lower bleeding rates in procedures that received FFP with or without TXA, while FXI activity alone was not independently associated with bleeding. Other cohorts did not identify consistent associations between FXI activity and bleeding.

3.5. Case Reports

Seventeen case reports describing individual patients with congenital FXI deficiency were identified (Table 2). Procedures included cardiac, abdominal, orthopedic, neurosurgical, urologic, gynecologic, oral and spine surgeries. Most patients (88%) had severe deficiency (FXI < 20 IU/dL).

Clinicians primarily used FFP for perioperative management, typically administering 10–20 mL/kg before surgery with additional dosing when indicated. FXI concentrate, TPE, and antifibrinolytics (TXA or ε‐aminocaproic acid) appeared in selected cases. No patients received rFVIIa.

Three patients (18%) experienced significant bleeding, which resolved after additional plasma or antifibrinolytic therapy. No thrombotic events or transfusion‐related complications were reported. Patients who achieved FXI activity around 25–40 IU/dL or normalized APTT before surgery generally did not bleed, including during high fibrinolytic procedures.

3.6. Exploratory Analysis: FXI Level and Procedural Fibrinolytic Risk

In an exploratory analysis of individual‐level case report data, perioperative bleeding events were examined according to combined FXI activity category (≤ 10 IU/dL, 10–20 IU/dL, > 20 IU/dL) and surgical fibrinolytic risk (Figure 2). Bleeding events were more frequently observed among patients with FXI ≤ 10 IU/dL undergoing high‐fibrinolytic procedures. Two of the three bleeding events occurred in patients who did not receive pre‐operative prophylaxis, one a transurethral resection of a bladder (high fibrinolytic risk) and one a total thyroidectomy (low fibrinolytic risk), and in both cases, subsequent procedures performed after diagnosis and with appropriate prophylaxis were completed without significant bleeding. No high‐fibrinolytic procedures were reported in patients with FXI activity between 10 and 20 IU/dL. A single bleeding event occurred in the > 20 IU/dL, low‐fibrinolytic category, in a patient with genetically confirmed FXI deficiency and a documented bleeding phenotype undergoing thyroid surgery.

FIGURE 2.

Perioperative bleeding (intraoperative and/or postoperative) across categories defined by baseline FXI activity (≤ 10 IU/dL, 10–20 IU/dL, and > 20 IU/dL) and surgical fibrinolytic risk (high vs. low). Data represent 21 procedures across 17 patients; N above each bar indicates procedures per category. Significant bleeding was defined as bleeding explicitly described as excessive, abnormal, or requiring unplanned intervention; blood loss within the expected range for the procedure was classified as no significant bleeding. No high‐fibrinolytic procedures were reported in patients with FXI activity between 10 and 20 IU/dL and are therefore not shown. The single bleeding event observed in the > 20 IU/dL, low‐fibrinolytic category occurred in a patient with genetically confirmed FXI deficiency and a strong personal bleeding history undergoing thyroid surgery, despite a baseline FXI activity of 52 IU/dL.

4. Discussion

This systematic review shows that FXI activity alone does not reliably predict perioperative bleeding in congenital FXI deficiency. Across more than 590 procedures, we found bleeding at all FXI levels, including in some patients with activity > 20 IU/dL—while many patients with severe deficiency (< 20 IU/dL) underwent surgery without complications. Surgical context and a personal bleeding history appeared to align more consistently with bleeding risk than FXI activity alone.

Perioperative management strategies varied across studies but relied primarily on FFP‐based replacement and antifibrinolytic therapy. These approaches associated with successful haemostasis control in a broad range of procedures, including major cardiac and abdominal operations. In individual case reports, bleeding events were infrequent and usually mild, suggesting that individualized procedure‐specific prophylaxis can support safe perioperative care even in severe FXI deficiency. Our exploratory visualization suggested that patients with very low FXI activity (< 10 IU/dL) undergoing procedures in areas with high fibrinolytic activity may experience more bleeding events. However, the limited number of cases and the widespread use of perioperative haemostatic prophylaxis preclude definitive conclusions.

Our findings are broadly consistent with previously proposed procedure‐based classification frameworks for perioperative management in FXI deficiency. Our data support a tiered approach: antifibrinolytics alone appeared sufficient for minor procedures at high‐fibrinolytic sites in patients without a significant bleeding history, while FFP‐based replacement was predominantly used for major cardiac and abdominal operations [17]. Salomon et al. [24] demonstrated that single low dose rFVIIa combined with TXA achieved complete haemostatic control across 12 major elective procedures without plasma use, and Wu et al. [21] reported that absence of prophylaxis was not associated with major bleeding across procedural subtypes, suggesting that routine replacement may not be necessary for all procedure types. These observations support individualized, procedure‐stratified prophylaxis planning based on surgical bleeding risk and personal bleeding history rather than FXI activity threshold alone. While the central finding that FXI activity poorly predicts perioperative bleeding is consistent with prior literature, this review extends the evidence base to the most recent decade, incorporates individual‐level case data alongside institutional cohorts, and provides an exploratory procedure‐fibrinolytic risk framework that may inform prospective classification efforts. The single bleeding event observed among patients with FXI activity >20 IU/dL occurred in an individual with genetically confirmed FXI deficiency and a documented bleeding phenotype undergoing thyroid surgery, despite a baseline FXI activity of 52 IU/dL. This observation highlights the recognized discordance between FXI activity and bleeding phenotype and reinforces the limitations of using FXI activity alone to guide perioperative risk assessment.

The selection of haemostatic agent should incorporate treatment‐related risk. FFP carries risks of volume overload, allergic reactions, and infection transmission, Wu et al. [21] reported a 16% adverse event rate from haemostatic management in their cohort. FXI concentrate, while effective, carries a thrombotic risk of up to 5–7% and remains unavailable in several countries [13, 42]. Although not directly addressed in the studies included in this review, published literature suggests that rFVIIa at low doses combined with antifibrinolytics may represent a safer alternative in patients with null mutations or confirmed FXI inhibitors, in whom standard replacement therapy may be ineffective; genotyping and inhibitor screening should be considered in patients with severe deficiency, particularly those with an inadequate response to prior FFP therapy [1, 6, 43, 44].

This review has several limitations. Case reports are subject to publication bias, and heterogeneity in reporting limited direct comparisons across cohorts. Many cohorts reported more procedures than unique patients, indicating that some individuals underwent multiple operations; however, procedure‐level reporting prevented systematic assessment of intra‐patient variability. In the few cases where repeated procedures could be inferred, outcomes differed between operations, suggesting that procedural factors and perioperative management may modulate risk beyond baseline FXI activity. TPE was used in a small number of procedures, but inhibitor status was not reported, preventing clarification of its intended role. Additionally, sparse data within certain FXI activity strata, particularly the absence of observations in the 10–20 IU/dL range for high‐fibrinolytic procedures, reflect gaps in the literature rather than an absence of bleeding risk. Despite these constraints, integrating institutional cohort data with detailed patient‐level reports provides a more comprehensive perspective on perioperative risk than any single study.

Recent publications support our findings. Investigators have safely performed neuraxial anesthesia in obstetric patients with FXI activity ≥ 30 IU/dL when no bleeding history was present, and clinicians have successfully managed major procedures with rFVIIa or TXA even at low FXI levels [45]. Although pharmacologic FXI inhibition is biologically distinct from congenital FXI deficiency, the development of FXI‐ and FXIa‐directed anticoagulants has renewed interest in the role of FXI in perioperative haemostasis [46]. As more patients receive these agents, clinicians increasingly face perioperative management questions, affecting an estimated one in four individuals on long‐term anticoagulation annually [47]. Recent perioperative guidance for patients treated with anti‐FXI(a) therapies emphasizes individualized assessment based on procedural bleeding risk rather than laboratory thresholds alone [48]. Although indirect, this broader experience reinforces the central conclusion of our review: FXI activity alone is insufficient to guide perioperative decision‐making.

Clinically, our findings support a multifactorial, patient‐centred approach to perioperative management in FXI deficiency. Consideration of surgical bleeding risk, personal bleeding phenotype, and the selective use of antifibrinolytics or FFP‐based replacement appears more appropriate than reliance on FXI activity thresholds alone.

Future research should prioritize prospective data collection through disease‐specific registries, collaborative multicentre efforts, and the adoption of standardized outcome definitions. These approaches would help clarify bleeding predictors and inform prophylactic strategies across diverse surgical contexts. Given the rarity of congenital FXI deficiency, randomized controlled trials may be impractical. However, comparative observational studies, real‐world data, and expert‐driven methodologies, including clinician surveys, Delphi panels, and consensus guidelines, could meaningfully advance clinical practice. Additionally, the growing interest in FXI as a target for anticoagulation has renewed attention on its complex role in haemostasis.

In summary, FXI activity does not reliably predict perioperative bleeding. Successful perioperative care should integrate procedure type, individual bleeding history, and tailored haemostatic planning to support safe outcomes in FXI‐deficient patients.

Funding

The authors have nothing to report.

Ethics Statement

Ethics approval was not required for this study as it is a systematic review of previously published data.

Consent

Patient consent was not required for this study as it is based exclusively on published literature and does not involve direct patient participation.

Conflicts of Interest

The authors stated that they had no interests which might be perceived as posing a conflict or bias.

Permission to reproduce material

No permission was required to reproduce material from other sources.

Supporting information

Supporting Information: hae70321‐sup‐0001‐SuppMat.docx

Data Availability Statement

Data sharing is not applicable to this article as no new datasets were generated or analysed. All data supporting the findings of this study are derived from published literature and are cited within the article and supplementary material.