Re-exploration for bleeding and long-term survival after adult cardiac surgery: a meta-analysis of reconstructed time-to-event data

Background: Postoperative bleeding requiring re-exploration is a serious complication that occurs in 2.8–4.6% of patients undergoing cardiac surgery. Re-exploration has previously been associated with a higher risk of short-term mortality. However, a comprehensive analysis of long-term outcomes after re-exploration for bleeding has not been published. Materials and methods: The authors performed a systematic, three databases search to identify studies reporting long-term outcomes in patients who required re-exploration for bleeding after cardiac surgery compared to patients who did not, with at least 1-year of follow-up. Long-term survival was the primary outcome. Secondary outcomes were operative mortality, myocardial infarction, stroke, renal and respiratory complications, and hospital length of stay. Random-effects models was used. Individual patient survival data was extracted from available survival curves and reconstructed using restricted mean survival time. Results: Six studies totaling 135 456 patients were included. The average follow-up was 5.5 years. In the individual patient data, patients who required re-exploration had a significantly higher risk of death compared with patients who did not [hazard ratio (HR): 1.21; 95% CI: 1.14–1.27; P<0.001], which was confirmed by the study-level survival analysis (HR: 1.32; 95% CI: 1.12–1.56; P<0.01). Re-exploration was also associated with a higher risk of operative mortality [odds ratio (OR): 5.25, 95% CI: 4.74–5.82, P<0.0001], stroke (OR: 2.05, 95% CI: 1.72–2.43, P<0.0001), renal (OR: 4.13, 95% CI: 3.43–4.39 P<0.0001) respiratory complications (OR: 3.91, 95% CI: 2.96–5.17, P<0.0001), longer hospital length of stay (mean difference: 2.69, 95% CI: 1.68–3.69, P<0.0001), and myocardial infarction (OR: 1.85, 95% CI: 1.30–2.65, P=0.0007). Conclusion: Postoperative bleeding requiring re-exploration is associated with lower long-term survival and increased risk of short-term adverse events including operative mortality, stroke, renal and respiratory complications, and longer hospital length of stay. To improve both short-term and long-term outcomes, strategies to prevent the need for re-exploration are necessary.


Introduction
Postoperative bleeding requiring re-exploration is a serious but infrequent complication that occurs in 2.8-4.6% of patients undergoing cardiac surgery [1] .The three most frequently reported sites of bleeding are the body of grafts (20.2%), the sternum (17.0%), and vascular sutures (12.5%) [2] .
While it is known that re-exploration for bleeding is associated with higher risk of operative mortality after cardiac surgery, its association with long-term outcomes in patients undergoing cardiac surgery remains unclear, and published studies that assessed the topic have had contradictory findings [4][5][6][7][8][9] .
Herein, we performed a systematic review and meta-analysis with the aim of assessing the long-term impact of postoperative bleeding requiring re-exploration among patients undergoing cardiac surgery.

Material and methods
Ethical approval was waived as no human subjects or animals were involved.This meta-analysis was registered with the Research Registry UIN and PROSPERO, both databases for systematic review protocols.This work is reported in line with Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) (Supplemental Digital Content 1, http:// links.lww.com/JS9/C716,Supplemental Digital Content 2, http:// links.lww.com/JS9/C717) and Assessing the Methodological Quality of Systematic Reviews (AMSTAR) (Supplemental Digital Content 3, http://links.lww.com/JS9/C718)Guidelines [10,11] .

Search strategy
A medical librarian performed a comprehensive literature search to identify studies reporting long-term outcomes in patients who required re-exploration for bleeding after cardiac surgery versus patients who did not, with at least 1-year of follow-up.

Study selection and data extraction
Studies were screened by two different reviewers and discrepancies were resolved by the senior author.A first round of screening based on title and abstract content was performed.Studies were considered for inclusion if they were written in English and compared long-term outcomes in patients who required re-exploration for bleeding after cardiac surgery versus patient who did not, with at least 1-year follow-up.Studies not reporting long-term outcomes, abstracts, case reports, commentaries, editorials, expert opinions, conference presentations, and animal studies were excluded.For the second round of screening, the full text of the selected studies was pulled.References lists were also reviewed for relevant studies not initially captured.
Two investigators independently performed data extraction, while its accuracy was verified by the corresponding author.From each study, the following variables were extracted: study characteristics (publication year, country, sample size, study design, mean follow-up, and type of surgery) as well as patient demographics and clinical data [age, sex, BMI, smoking status, hypertension, diabetes, chronic obstructive pulmonary disease (COPD), prior cerebrovascular accident (CVA), prior myocardial infarction (MI), cardiopulmonary bypass, and cross-clamp time].The quality of the included studies was assessed using the Newcastle-Ottawa Scale for observational studies (Supplementary Table 2, Supplemental Digital Content 4, http://links.lww.com/JS9/C719).

Outcomes
The primary outcome was long-term all-cause mortality.Secondary outcomes were operative all-cause mortality, stroke, renal, and respiratory complications, MI, and hospital length of stay.The specific definition of the endpoints is provided in Supplementary Table 3 (Supplemental Digital Content 4, http:// links.lww.com/JS9/C719).

Statistical analysis
Categorical variables were analyzed using odds ratio (OR), incidence rate ratio (IRR), and 95% CI.An OR and IRR greater than 1 indicated that an outcome was more frequently present in the re-explored arm.Continuous variables were analyzed using mean difference (MD) and 95% CI.A MD lower than zero corresponded to larger values in the re-explored arm.
Clinical heterogeneity between studies was assessed by random effects models.Results were displayed using forest plots.Between-study statistical heterogeneity was assessed with the Cochran Q statistic and by estimating I 2 .High heterogeneity was confirmed with a significance level of P < 0.10 and I 2 of at least 50% or more.Subgroup analyses of the primary outcome were performed based on types of surgery, criteria for re-exploration and single versus multicenter studies to investigate heterogeneity.

Reconstruction of individual patient survival data
The methods described by Wei et al. were used to reconstruct individual patient data from the Kaplan-Meier curves of all eligible studies for the long-term outcome [12,13] .Raster and Vector images of the Kaplan-Meier survival curves were preprocessed and digitized, so that the values reflecting specific timepoints and their corresponding survival/mortality information could be extracted.Where additional information (e.g.number-at-risk tables or total number of events) was available, it was used to further calibrate the accuracy of the time-to-events.Departures from monotonicity were detected using isotonic regression and corrected with a pool-adjacent-violators algorithm [12,13] .To confirm the quality of the timing of failure events captured, we thoroughly checked the consistency with the reported survival or mortality data provided in the original publications.

HIGHLIGHTS
• This is the first meta-analysis that analyzes long-term mortality after re-exploration for bleeding.• Patients who required re-exploration had significantly higher rates of long-term, all-cause mortality compared to patients who did not.• Patients requiring re-exploration had significantly higher rates of operative mortality, stroke, renal complications, respiratory complications, longer hospital length of stay, and myocardial infarction when compared with patients who did not require re-exploration for bleeding.
Meta-analysis of reconstructed data -One-stage survival meta-analysis The Kaplan-Meier method was used to calculate the overall survival.The Cox proportional hazards regression model was used to assess differences between groups.For these Cox models, the proportional hazards assumption was verified by plotting scaled Schoenfeld residuals, log-log survival plots, and predicted versus observed survival functions.The survival curves were plotted using the Kaplan-Meier product limit method and hazard ratios (HRs) as well as 95% CIs for each group were calculated.

Sensitivity analyses
Leave-one-out analysis for the primary outcome was performed to assess the robustness of the obtained estimate.A funnel plot was performed to assess publication bias.Meta-regression was used to explore the effects of sex

Results
A total of 2583 studies were retrieved from the systematic search, six of which met the inclusion criteria.The PRISMA flow diagram outlining the study selection process is provided in Supplementary Figure 1 (Supplemental Digital Content 4, http:// links.lww.com/JS9/C719).

Study characteristics
The included studies were published between 2021 and 2023 and were all observational.

Two-stage survival meta-analysis
Figure 5 shows the forest plot for long-term all-cause mortality.Patients who required re-exploration had significantly higher rates of long-term all-cause mortality compared to patients who did not (HR: 1.32, 95% CI: 1.12-1.56,P < 0.01).The subgroup analyses was qualitatively consistent with the primary analysis (Supplementary Figures 9-11, Supplemental Digital Content 4, http://links.lww.com/JS9/C719).

Secondary outcomes
Table 2 summarizes other key results of the meta-analysis including the investigated secondary outcomes and their respective effect estimates.
While risk factors and early outcomes of patients undergoing reexploration have been studied, research on the mid-term and longterm outcomes is limited and mixed [7,9] .Three of the six published studies found no significant difference in long-term mortality between patients who did and did not undergo re-exploration, while the remaining reported higher short-term mortality among patients who underwent re-exploration [4,6,9] .The studies that reported no significant difference between re-exploration and no-re-exploration had longer follow-up (Heimisdottir: mean 4.6 years, Marteisson: median 7.6 years, Stroo: median 9.7 years) [4,6,9] .
Our study has corroborated previous short-term results, bringing new evidence concerning long-term outcomes of reexploration for bleeding after cardiac surgery.
Our data suggests that to reduce morbidity and mortality after cardiac surgery, critical preventive and procedural measures need to be established in order to avoid re-exploration.A comprehensive approach should be considered addressing preoperative, intraoperative, and postoperative practices [14] .Preoperatively, it is important to identify patients with high risk of bleeding by conducting a thorough preoperative history [15] .Risk assessment tools like FMT and other preoperative risk assessment tools could be used to aid assessment of bleeding risk [12] .Intraoperatively, preventive measures such as acute normovolemic hemodilution, intraoperative cell salvage, ultrafiltration, pharmacological therapies targeting hemostasis, coagulation, and fibrinolysis could help reduce bleeding [15] .Intraoperative checklists may also be useful to predict postoperative bleeding [15] .Postoperatively, avoiding iatrogenic blood loss and utilizing re-exploration only when necessary are essential [15] .In fact, conservative management for stable patients with significant but noncritical bleeding might be a safe option, thereby reducing postoperative morbidity and hospital stay [16] .
The unfavorable outcomes after re-exploration for bleeding are likely multifactorial.Identifying a single underlying cause for these poor outcomes is challenging.Further research is needed to better understand the mechanisms involved while continuing to explore strategies for improving patient outcomes.

Study strengths and limitations
This is the first meta-analysis to examine the relationship of reexploration for bleeding after cardiac surgery and long-term outcomes.Moreover, we assessed seven different outcomes and performed different sensitivity analyses including a meta-regression of nine different preoperative factors.
However, this work has the intrinsic limitations typical of observational studies, including the risk of methodological heterogeneity, residual confounders, and ecologic fallacy of metaregression.Most importantly, criteria for re-exploration and the definition of secondary outcomes were heterogenous and/or poorly defined across studies Table 3.

Conclusion
Re-exploration for bleeding after cardiac surgery is associated with a higher risk of long-term mortality, operative mortality, and morbidity after cardiac surgery.These findings highlight the substantial impact of re-exploration for bleeding on both short-term and long-term outcomes following adult cardiac surgery.

Figure 1 .
Figure 1.Pooled Kaplan-Meier curve showing the cumulative risk of all-cause mortality following re-exploration for bleeding versus no re-exploration.HR, hazard ratio.

Figure 2 .
Figure 2. Pooled Kaplan-Meier curve showing the cumulative risk of all-cause mortality following re-exploration for bleeding versus no re-exploration at 30 day follow-up.HR, hazard ratio.

Figure 3 .
Figure 3. Pooled Kaplan-Meier curve showing the cumulative risk of all-cause mortality following re-exploration for bleeding versus no re-exploration at 1-year follow-up.HR, hazard ratio.

Figure 4 .
Figure 4. Pooled Kaplan-Meier curve showing the cumulative risk of all-cause mortality following re-exploration for bleeding versus no re-exploration beyond 1-year follow-up.HR, hazard ratio.

Figure 5 .
Figure 5. Forest plot for the primary endpoint (long-term all-cause mortality).HR, hazard ratio.

Table 4 (
Supplemental Digital Content 4, http://links.lww.com/JS9/C719).Age ranged from 64.9 to 70.3 years; the percentage of female patients ranged from 13.9 to 31.3%; the percentage of actively smoking patients ranged from 13.4 to 73.1%; the percentage of patients with hypertension ranged from 36.9 to 88.5%; the percentage of patients with diabetes ranged from 11.2 to 85.3%; the percentage of patients with COPD ranged from 3.5 to 12.2%; the percentage of patients with prior CVA ranged from 3.6 to 23.6%; the percentage of patients with prior MI ranged from 0.9 to 34.6%; the cardiopulmonary bypass time ranged from 87 to 132 min; the crossclamp time ranged from 46 to 67 min.Criteria used for reexploration are summarized in Supplementary Table5(Supplemental Digital Content 4, http://links.lww.com/JS9/C719).
The included studies originated from United States, Sweden, Poland, Iceland, Denmark, or the Netherlands.A total of 135 456 patients were included in the final analysis.The number of patients in each study ranged from 2060 to 48 060 with a median sample size of 16 579.5[interquartilerange(IQR): 30 529 [10 824-41 353]) (Table1).

Table 1
Summary of included studies.

Table 2
Outcomes summary.