Prothrombin complex concentrate versus fresh frozen plasma in patients undergoing cardiac surgery: a systematic review and meta-analysis of randomized controlled trials

Introduction

The role of cardiopulmonary bypass (CPB) in cardiac surgery is vital, as it helps maintain the patient's cardiovascular and respiratory functions during procedures that require cessation of the heart. However, the use of cardiopulmonary bypass poses significant risks, one of which is the risk of coagulopathic bleeding. Excessive bleeding and transfusions have been described as poor prognostic factors in patients undergoing cardiac surgery, leading to adverse perioperative outcomes such as extended length of hospital stay, increased expenditure, and even mortality (1).

Coagulopathy in cardiac surgery has been attributed to multiple factors, ranging from thrombocytopenia, impaired thrombin generation, fibrinogen depletion, and excessive consumption of coagulation factors (2, 3). Moreover, the use of CPB machines requires a priming solution, usually a crystalloid, which can result in plasma expansion, hemodilution, and worsening coagulopathy (4–6). CPB use has been found to cause a 40- 50% reduction in clotting factors and an over-50% reduction in thrombin-generating capacity, contributing to coagulopathy (7–9). Considering the significant risk of bleeding associated with the use of CPB, it is of utmost importance to quickly reverse the coagulopathy following cardiac surgery.

Fresh frozen plasma is currently the mainstay of management for coagulopathic bleeding, as evidenced by its use in over 15% of cardiac surgeries in the US (10). However, its use is associated with increased hemodilution and increased red blood cell (RBC) transfusion volume and carries significant risks of transfusion reactions, including transfusion-related acute lung injury (TRALI) and transfusion-associated circulatory overload (TACO) (10). Due to the adverse reactions associated with FFP, prothrombin complex concentrate has gained considerable consideration as a safer and more efficacious alternative (10).

Most formulations of prothrombin complex concentrates contain all the vitamin K-dependent clotting factors (II, VII, IX, X) in small amounts of factor VII (3). Moreover, some formulations contain protein C, protein S, antithrombin, and/or heparin (3). Moreover, PCC has been shown to effectively reduce the International Normalized Ratio (INR) in patients requiring reversal of Vitamin K Antagonist therapy, making it a suitable treatment option for managing coagulopathy in cardiac surgery.

Despite numerous trials assessing the effectiveness and safety of PCC vs. FFP in the management of bleeding in cardiac surgery, significant equipoise remains regarding the efficacy and safety profile of PCC compared to FFP due to the lack of sufficiently large trials. In an attempt to provide high-quality evidence to bridge the knowledge gap, we performed a systematic review and meta-analysis of randomized controlled trials comparing the efficacy and safety of prothrombin complex concentrate with fresh frozen plasma in the management of bleeding after cardiac surgery.

Two prior systematic reviews have compared PCC to FFP in cardiac surgery. The Cochrane review by Hayes et al. (29) included only 3 RCTs (n = 284) and concluded that evidence was insufficient to determine comparative effects. The meta-analysis by Viana et al. (11) included 4 RCTs (n = 554) but did not perform a GRADE certainty assessment or pool observational studies alongside RCTs. Our study provides the following incremental contributions: (1) inclusion of the recently published FARES-II trial [32), n = 420], representing a significant increase in patient data over Viana et al.; (2) restriction to RCTs only, eliminating potential confounding from observational designs; (3) comprehensive GRADE certainty assessment for all outcomes (see Table 1); (4) leave-one-out sensitivity analyses with influence diagnostics (see Figures 13–16); (5) explicit handling of imprecision using optimal information size criteria; and (6) updated risk of bias assessment using ROB-2 (see Table 2 for detailed domain-level assessments). These methodological refinements provide the most current and rigorous synthesis available to date.

Table 1. Summary of findings (GRADE assessment).

CI, confidence interval; FP, frozen plasma; ICU, intensive care unit; MD, mean difference; PCC, prothrombin complex concentrate; RBC, red blood cells; RCT, randomized controlled trials; RR, risk ratio.

The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).

Explanations.

There was some concern for bias in Smith et al. (12) due to randomization assignments being disclosed to investigators.

There is significant heterogeneity in the pooled analysis.

Confidence interval crosses null value.

Table 2. Risk of bias summary for randomized studies (RoB 2).

The randomization assignments were disclosed to study staff.

Methods

This systematic review and meta-analysis were conducted in accordance with the Cochrane Handbook for Systematic Reviews of Interventions and reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement (14, 15).

Eligibility criteria

Only studies meeting the following eligibility criteria were included in this meta-analysis: (1) randomized controlled trials; (2) comparison of prothrombin complex concentrate (PCC) (any formulation or dose) with fresh frozen plasma (FFP) as the primary hemostatic intervention; (3) enrollment of adult patients (≥18 years) undergoing cardiac surgery; and (4) reporting of at least one clinical endpoint of interest. Studies were excluded if they: (1) had an observational design; (2) lacked a control group; (3) included overlapping patient populations reported in multiple publications; (4) involved non-cardiac surgery; (5) reported no endpoint of interest; or (6) allowed clinician-driven rescue co-administration of PCC and FFP in either study arm. Studies permitting protocol-defined co-administration were retained, with such interventions analyzed according to the intention-to-treat principle. Studies were included regardless of whether coagulopathy was objectively confirmed, provided the authors defined clinically significant bleeding requiring hemostatic intervention.

Search strategy and data extraction

We systematically searched PubMed, Embase, and the Cochrane Central Register of Controlled Trials from inception to May 2025 with the search terms ‘cardiac surgical procedures’, ‘cardiopulmonary bypass’, ‘prothrombin complex concentrate’, and ‘fresh frozen plasma’. The full search strategy can be found in the Supplementary Data Sheet 1 (see Figure 1).

References from prior systematic reviews and meta-analyses were also searched manually for additional studies. Two authors (S.G.J.F. and J.Y.B.N.) independently extracted data using predefined search criteria and quality assessment. Data extraction was performed using a spreadsheet by two authors; any conflicts were resolved by consensus, including a third author (M.O.A.). The prospective meta-analysis was registered on PROSPERO on June 11, 2025, under the protocol number CRD420251066886. An attempt was made to obtain stratified data from the authors of the included studies, but to no avail.

Endpoints

Primary efficacy endpoints included reoperation for bleeding, 24-hour chest tube drainage, and the number of red blood cell units transfused. Safety outcomes included incidence of adverse events, serious adverse events, all-cause mortality, thromboembolic events (including stroke and transient ischemic attacks), and acute kidney injury. Resource utilization outcomes were also assessed, namely the length of ICU and hospital stay. The primary endpoint definitions are presented in Table 3.

Table 3. Definition of terms.

Quality assessment

We evaluated the risk of bias of all included studies using version 2 of the Cochrane Risk of Bias assessment tool (16).

This tool assesses bias across five domains: (1) randomization process, (2) deviations from intended interventions, (3) missing outcome data, (4) outcome measurement, and (5) selective reporting. For each domain, studies were rated as “low risk,” “some concerns,” or “high risk” based on predefined signaling questions. Two independent authors (SGJF and JYBN) conducted the assessments and resolved disagreements by consensus. The overall risk of bias for each study was determined by the highest level of bias identified across domains. Results are presented in Table 2.

Statistical analyses

When studies reported medians and interquartile ranges (IQR) without evidence of skewness, means and standard deviations were estimated using the method of Wan et al. (17) Skewness was assessed by comparing (median - minimum) to (maximum - median); if these differed by >20% of the range, data were considered skewed and not converted. This applied to the Smith et al. (12) study for chest tube drainage. Heterogeneity was assessed using I² statistics and the Cochrane Q test; substantial heterogeneity was defined as I² > 50% with p < 0.10 for the Q test (Higgins JPT, BMJ. 2003;327:557–560). I² values of 25%–49% were considered ‘low heterogeneity,’ 50%–74% ‘moderate,’ and ≥75% ‘considerable.’ A random effects model was used. Pooled analysis followed an intention-to-treat protocol.

We used Review Manager 8.13.0 (Cochrane Center, The Cochrane Collaboration, Denmark) for statistical analysis (18). For continuous endpoints analysis, medians and interquartile ranges were used to estimate means and standard deviations, respectively, when no significant evidence of skewness was found (14, 19–21).

We performed a leave-one-out sensitivity analysis using R (meta package) to ensure that the results were not dependent on a single study. We also analyzed all endpoints using the fixed-effects model to ensure that the results were not dependent on the choice of model.

We present these results in Figures 13–16.

We employed the DerSimonian-Laird estimator for between-study variance (τ²), as implemented in Review Manager. For outcomes with three or fewer studies, we present both random-effects and fixed-effects estimates in Table 4; no substantive differences were observed. For the primary outcomes, we also calculated confidence intervals using the Hartung-Knapp-Sidik-Jonkman adjustment for sensitivity.

Table 4. Baseline characteristics of included studies (PCC/FFP).

Certainty of evidence assessment

The Grading of Recommendations, Assessment, Development, and Evaluation (GRADE) framework was used to assess the certainty of evidence for each outcome in our meta-analysis. GRADE evaluates evidence across five key domains, with the potential to downgrade or upgrade the certainty rating: risk of bias, inconsistency, indirectness, imprecision, and publication bias.

We used the following GRADE assessment criteria: We utilized the GRADEPro GDT software, and the results were reported in Table 1. The GRADE Working Group grades of evidence are as follows:

• Inconsistency: Downgraded by one level if I² > 50% and p < 0.10 for the Q test, or if point estimates varied substantially in direction or magnitude across studies. Downgraded by two levels if I² > 75% with overlapping confidence intervals that failed to identify a consistent effect. Example: Chest tube drainage (I² = 69%) was downgraded one level for inconsistency.
• Indirectness: Downgraded by one level if the population, intervention, comparator, or outcome differed from the review question. For this review, no indirectness was identified as all studies directly compared PCC to FFP in cardiac surgery patients.
• Imprecision: Downgraded by one level if the total sample size was less than the optimal information size (OIS) of 800 patients for binary outcomes (based on a minimally important difference of 25% relative risk reduction with α=0.05, β=0.20). For continuous outcomes, OIS was calculated using a standardized mean difference of 0.2. Downgraded by two levels if the confidence interval included both clinically important benefit and harm AND the total sample size was <400. Example: RBC units (n = 571 < OIS 800, CI crosses null) was downgraded one level for imprecision.
• Publication bias: Not assessed due to <10 studies (funnel plots have low power); however, we qualitatively discuss this limitation and the possibility of unpublished negative trials could not be excluded.
• Large effect: Upgrade by one level for RR <0.5 or >2.0 with no plausible confounders; upgrade by two levels for RR <0.2 or >5.0. No outcomes met these thresholds.
• Plausible confounding: No upgrade applied as all included studies were RCTs.
• Dose-response gradient: Not applicable to our study

• - High certainty: We are confident that the true effect lies close to that of the estimate of the effect.
• - Moderate certainty: We are moderately confident in the effect estimate; the true effect is likely close to the estimate, though it may differ substantially.
• - Low certainty: Our confidence in the effect estimate is limited; the true effect may differ substantially from the estimate.
• - Very low certainty: We have very little confidence in the effect estimate; the true effect is likely to differ substantially from the estimate.

Results

Study selection and baseline characteristics

The initial search yielded 271 results. After removing duplicates and ineligible studies identified through title and abstract review, 10 studies remained and were fully reviewed against the inclusion criteria. Five randomized controlled trials comprising 711 patients were included (Figure 2).

A total of 363 (51.06%) patients were randomized to PCC and 348 (48.94%) to the FFP group. Characteristics of all included studies are summarized in Table 4. The mean age of participants across the studies ranged from 63 to 71. The overall proportion of female patients in our study was 30.31%. There were no significant differences in baseline characteristics of included studies (Table 4).

Pooled analyses of all included studies

Regarding transfusion strategies and bleeding, patients in the PCC group had a possibly lower 24-hour chest tube drainage volume in comparison to the FFP group (MD = −143.59 mL; 95% CI −278.76 to −8.41; p = 0.04; I² = 69%; low certainty; Figure 3). Patients in the PCC group equally experienced lower rates of reoperation for bleeding (5.83% vs. 8.53% RR, 0.69; 95% CI 0.48–0.98; p = 0.04; I² = 0%; moderate certainty; Figure 2). Moreover, patients in the PCC group had a lower amount of RBC units transfusions following surgery (MD = −0.93; 95% CI −1.83 to −0.02; p = 0.05; I² = 0%; moderate certainty; Figure 4).

Overall, there was no statistically significant difference both groups regarding adverse events (87.53% vs. 89.36%; RR, 0.99; 95% CI 0.95–1.04; p = 0.64; I² = 0%; Figure 5) and serious adverse events (32.59% vs. 40.33%; RR, 0.84; 95% CI 0.37–2.08; p = 0.59; I² = 48%; Figure 6). There was equally no statistically significant difference between the two groups in terms of all cause mortality (3.20% vs. 3.96%; RR, 0.81; 95% CI 0.57–1.15; p = 0.16; I² = 0% low certainty; Figure 7), Stroke or Transient Ischemic Attack (2.62% vs. 3.65%; RR, 0.79; 95% CI 0.24–2.60; p = 0.57; I² = 0%; low certainty; Figure 8) and Acute Kidney Injury (14.86% vs. 19.20%; RR, 0.86; 95% CI 0.43–1.70; p = 0.53; I² = 38%; very low certainty; Figure 9). Moreso, there was no statistically significant difference between the two groups in terms of Length of ICU stay (MD = −0.17; 95% CI −0.79–0.44; p = 0.43; I² = 0%; low certainty; Figure 10) and length of hospital stay (MD = −0.56; 95% CI −1.73–0.61; p = 0.22; I² = 13%; low certainty; Figure 11). Conversely, low level evidence revealed that the use of PCC was associated with a 22% increase in thromboembolic events (9.32% vs. 7.62%; RR, 1.22; 95% CI 0.64–2.33; p = 0.40; I² = 0%; low certainty; Figure 12) however this finding was not statistically significant and we observed a wide confidence interval.

Sensitivity analyses

We conducted a leave-one-out analysis with a few selected outcomes, and this yielded the following variations:

• All-Cause Mortality: The leave-one-out analysis shows that no single study unduly influences the pooled estimate. The point estimates for each study omitted range from approximately RR 0.76–0.86, all with confidence intervals that cross the null and substantially overlap with the pooled estimate (RR 0.81). The I² remains 0% across all iterations. This indicates robust findings that are not dependent on individual studies (Figure 13).
• Chest Tube Drainage: Omitting Karkouti et al. (32) (the largest and most precise study) shifts the pooled estimate from −143.6 mL to approximately −118.5 mL and widens the confidence interval such that it crosses the null. Conversely, omitting Smith et al. (12) shifts the estimate to approximately −189.4 mL. This demonstrates that the overall effect is sensitive to the inclusion of the largest trial (32) and that the heterogeneity (I² = 69%) is influenced by the divergent findings of Smith et al. (12) (Figure 14).
• Serious Adverse Events: The leave-one-out analysis shows considerable instability in the pooled estimate. Omitting Demeyere 2010 (a small study with an extremely wide confidence interval) changes the point estimate modestly (RR from 0.84 to approximately 0.86) but reduces heterogeneity (I² drops from 48% to 0%). Omitting Karkouti et al. (32) slightly shifts the point estimate and reduces the I²–32%. The wide confidence intervals across all iterations (e.g., 0.65–1.14, 0.57–1.28) suggest that the pooled estimate for serious adverse events is not robust and should be interpreted with caution (Figure 15).
• Thromboembolic Events: The leave-one-out analysis shows stable point estimates ranging from RR 1.10–1.32, all with wide confidence intervals that cross the null. Omitting Karkouti et al. (31) results in RR 1.10 (95% CI 0.52–2.33), while omitting Karkouti et al. (32) results in RR 1.32 (95% CI 0.58–3.01). The I² remains 0% across all iterations. This indicates that the finding of no statistically significant difference is robust, though the wide confidence intervals underscore the imprecision of the estimate (Figure 16).

Quality assessment

The Risk of Bias 2 (ROB-2) tool was used for the quality assessment. No studies were considered at high risk for bias, as described in Table 2. However, there was some concern for bias in the Smith et al. (12) study owing to disclosure of randomization assignment to study staff. A funnel plot analysis could not be conducted because we had fewer than 10 studies; hence, publication bias could not be assessed.

Discussion

Summary of findings

In this comprehensive meta-analysis of randomized controlled trials, prothrombin complex concentrate (PCC), compared with fresh frozen plasma (FFP), was associated with several clinically relevant benefits in adult patients undergoing cardiac surgery. PCC probably reduces the risk of reoperation for bleeding by approximately 31% (RR 0.69, 95% CI 0.48–0.98; moderate certainty) and likely results in a slight reduction in chest tube drainage volume (MD −143.6 mL, 95% CI −278.8 to −8.4; low certainty). PCC also likely reduces postoperative red blood cell (RBC) transfusion requirements by nearly one unit (MD −0.93 units, 95% CI −1.83 to −0.02; moderate certainty).

Regarding safety outcomes, PCC is also likely to result in a slight reduction in all-cause mortality (RR 0.81, 95% CI 0.57–1.15; low certainty), but the confidence interval crossed the null.

Regarding safety, PCC use may increase the risk of thromboembolic events by 22% (RR 1.22, 95% CI 0.64–2.33; low certainty). However, the wide confidence interval and low event rates preclude firm conclusions. The evidence suggests that PCC results in a slight reduction in stroke or TIA (RR 0.79, 95% CI 0.24–2.60; low certainty) and PCC may reduce/have little to no effect on Acute kidney injury, but the evidence is uncertain (RR 0.86, 95% CI 0.43–1.70; very low certainty).

Concerning resource utilization, PCC resulted in a slight reduction in length of ICU stay (MD −0.17 days, 95% CI −0.79–0.44; low certainty) and probably resulted in little to no difference in the length of hospital stay (MD −0.56 days, 95% CI −1.73–0.61; low certainty). Overall, the evidence suggests that PCC offers efficacy advantages in bleeding control and transfusion reduction without an increase in thromboembolic events. However, the certainty for most safety outcomes remains low.

Comparison with previous studies

Our findings align with and extend the growing body of literature comparing PCC and FFP in cardiac surgery. A previous meta-analysis, primarily dominated by observational data, suggested that PCC is associated with reduced transfusion requirements and improved hemostatic control compared with FFP, without an apparent increase in thromboembolic complications (11). This study reported a mean reduction in 24-hour chest tube drainage of approximately 148 mL and a decrease in RBC transfusion by nearly one unit with PCC use — estimates remarkably similar to those observed in our analysis.

Our efficacy results also align with earlier observational cohorts in which PCC appeared more beneficial among patients with severe coagulopathy or massive transfusion requirements. Cappabianca et al. reported that PCC reduced postoperative bleeding and transfusion needs after cardiac surgery, albeit with a possible increased risk of acute kidney injury (22). Similarly, Arnekian et al. observed markedly lower blood loss in patients treated with PCC compared with after cardiac surgery (23). The magnitude of benefit observed in such studies, however, often exceeded that reported in our pooled RCTs, suggesting that patient selection and confounding may have amplified the apparent effect.

Our results also contrast with the neutral findings of some RCTs, which failed to demonstrate significant differences in transfusion volumes and chest tube drainage (12, 13). These discrepancies may be explained by their limited sample sizes, underpowering for clinical endpoints, and variability in PCC dosing regimens.

Regarding safety, our pooled estimates are consistent with prior meta-analyses in showing no statistically significant increase in thromboembolic events with PCC. Nevertheless, the wide confidence intervals and low event rates underscore the persistent uncertainty in this domain. Observational studies have reported increased and unchanged thrombotic risk, often influenced by differences in patient comorbidity profiles, perioperative anticoagulation management, and the timing of PCC administration (24, 25). Importantly, our analysis suggests that PCC may increase thromboembolic events, but further adequately powered trials are warranted to confirm this.

The logistical benefits of smaller infusion volumes may explain the observed reductions in bleeding and transfusion requirements in clinical practice.

Contributions compared to prior meta-analysis

Our findings are largely consistent with Viana et al. (11) regarding the direction of effect for bleeding-related outcomes (reoperation for bleeding: RR 0.69 vs. 0.68; chest tube drainage: MD −144 mL vs. −148 mL). However, our study provides several substantive incremental contributions that extend beyond mere replication:

Importantly, we include the recently published FARES-II trial [(32), n = 420], increasing the total sample size from 554 (Viana et al.,) to 711 patients—a 28% increase in the number of patients. This enhances precision for all outcomes, particularly for rare events such as thromboembolic complications and mortality, where larger sample sizes are critical for reliable estimation.

Second, we conducted a comprehensive GRADE certainty assessment (Table 1), which was not included in the meta-analysis published by Viana et al. This assessment reveals that most safety outcomes—including thromboembolic events (LOW certainty), all-cause mortality (LOW certainty), and acute kidney injury (VERY LOW certainty)—are supported by low- or very-low-quality evidence. This finding significantly tempers clinical recommendations. While Viana et al. concluded that PCC “can be used without increased thromboembolic risk,” our GRADE-informed interpretation suggests that this conclusion overstates the certainty of the evidence, as the wide confidence interval (RR 1.22, 95% CI 0.64–2.33) does not rule out clinically important harm. This represents a meaningful advance in evidence synthesis, as it provides clinicians with a more nuanced and cautious interpretation of the available data.

Third, we performed comprehensive leave-one-out sensitivity analyses with influence diagnostics (Figures 13–16), which were not reported by Viana et al. These analyses reveal that the pooled estimate for chest tube drainage is influenced by the largest trial (32) and that heterogeneity is driven in part by dosing differences (Smith, (12), used fixed low-dose PCC). This provides important context for interpreting the results and identifies directions for future research.

Fourth, we conducted updated risk-of-bias assessments using the ROB-2 tool (Table 2) with domain-level reporting, whereas Viana et al. used the older Cochrane risk-of-bias tool. This provides a more contemporary and granular assessment of methodological quality.

Fifth, we explicitly addressed imprecision using optimal information size (OIS) criteria, revealing that the current evidence base (n = 711) remains underpowered for several outcomes, including thromboembolic events (OIS ≈ 800). This underscores the need for future adequately powered clinical trials on this topic.

Last but not least, we incorporated economic and resource considerations, acknowledging the substantial cost differential between PCC and FFP and the implications for clinical decision-making, particularly in resource-limited settings.

Heterogeneity and its implications

The substantial heterogeneity observed for chest tube drainage (I² = 69%) warrants careful consideration. Several factors may have contributed to this wide heterogeneity. First, PCC dosing regimens differed across studies: Smith et al. (12) used a fixed low dose of 1000 IU; Karkouti et al. (31, 32) used weight-based dosing (15–25 IU/kg); and Green et al. (13) used variable dosing based on INR. Fixed low-dose PCC may be insufficient to achieve hemostasis in larger patients, potentially explaining why Smith et al. reported a point estimate favoring FFP. Indeed, the leave-one-out analysis (Figure 14) shows that omitting Smith et al. (12) shifts the pooled estimate from −143.6 mL to −189.4 mL with reduced heterogeneity (I² from 69% to 44%), suggesting that the divergent dosing regimen in this study contributed to the overall heterogeneity. Second, bleeding definitions and reasons for transfusion varied: Karkouti et al. used objective criteria (chest tube output thresholds), while Smith et al. and Green et al. relied on clinician judgment. We employed a random-effects model (DerSimonian-Laird) to account for this between-study variation, which provides more conservative estimates than a fixed-effects model. Despite this heterogeneity, the direction of effect consistently favored PCC in four of five studies, and the pooled estimate remained statistically significant. However, the heterogeneity reduces our confidence in the precise magnitude of the effect, and clinicians should interpret the pooled estimate with caution.

Clinical implications

The consistent reduction in postoperative bleeding and RBC transfusion requirements with PCC supports its consideration as an alternative to FFP in selected patients undergoing cardiac surgery with cardiopulmonary bypass.

From a logistical perspective, PCC offers significant advantages over FFP. It can be reconstituted and administered rapidly, avoiding the delays inherent to FFP thawing and cross-matching, which may be critical in ongoing surgical bleeding. The smaller infusion volume of PCC also reduces the risk of transfusion-associated circulatory overload (TACO). It may be particularly advantageous in patients with impaired ventricular function or those at risk of fluid overload (26, 27).

The magnitude of the benefit—approximately one fewer RBC unit transfused—may appear modest at the individual level but could translate into substantial cumulative savings in blood product use at the institutional or health system level. This has implications for resource management, especially in low— and middle-income countries or centers where blood product availability is constrained.

The potential for thromboembolic complications remains a key consideration. Although our pooled estimates did not demonstrate a statistically significant increase in events, the wide confidence intervals and low event rates mean that a clinically relevant risk cannot be excluded. Clinicians should therefore exercise caution in patients with a high baseline thromboembolic risk, such as those with mechanical prosthetic valves or recent coronary stenting.

Beyond statistical significance, it is important to consider whether the observed benefits of PCC translate into meaningful improvements for patients. The reduction in chest tube drainage of approximately 144 mL (95% CI −278.76 to −8.41 mL; p = 0.04) is modest, and the wide confidence interval ranges from a potentially meaningful reduction to a negligible effect. Moreover, the substantial heterogeneity (I² = 69%) and low certainty of this estimate warrant cautious interpretation. Similarly, the reduction in RBC transfusion requirements of approximately 0.93 units (95% CI −1.83 to −0.02 units; p = 0.05) is borderline statistically significant, with the confidence interval barely excluding the null effect and the upper bound approaching zero, indicating considerable uncertainty in the precise magnitude of this benefit. The reduction in reoperation for bleeding (5.83% vs. 8.53%; RR 0.69, 95% CI 0.48–0.98; p = 0.04) also demonstrated marginal significance, with the confidence interval narrowly excluding the null.

Economic implications

Prothrombin Complex has been reported to be way more expensive than fresh frozen plasma by several studies.

A study comparing the cost of PCC to that of FFP in patients who underwent heart transplantation revealed that PCC cost about $4,949 (median; IQR: 4,428–6,509), as compared to $ 3,677 (2,726–5,211) for patients who did not receive PCC, representing a $1,272 difference (28). This stark difference indicates that, despite its proven efficacy and safety, the cost may remain a significant barrier to the routine use of PCC for all cardiac surgery patients.

Particularly, in resource-limited settings where PCC may not be available, FFP may remain the only viable option. Even in well-resourced settings, the modest absolute benefits and uncertain safety profile do not support routine PCC use in all cardiac surgery patients.

Conclusion

This meta-analysis of 711 patients suggests that prothrombin complex concentrate, compared with fresh frozen plasma, may confer benefits in the management of coagulopathic bleeding in adult patients undergoing cardiac surgery. We also found no strong evidence of greater harm with respect to mortality or thrombosis, though the latter remains a concern requiring further study. There were no statistically significant differences in the safety endpoints such as all-cause mortality, thromboembolic events, stroke or transient ischemic attack, and acute kidney injury. However, overall certainty was very low to low for these outcomes, suggesting a need for future high-quality studies to evaluate the adverse effects of prothrombin complex concentrate. Altogether, our findings can guide clinical practice and decision-making in cardiac surgery, though the evidence is not strong.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher's note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fanes.2026.1843947/full#supplementary-material

Supplementary material Supplementary material at the source

Footnotes

Edited by: Stefano Turi, San Raffaele Hospital (IRCCS), Italy

Reviewed by: Aleena Ihtasham, King Edward Medical University, Pakistan

Reza Widianto Sudjud, Padjadjaran University, Indonesia