BACKGROUND:

Complex cranial vault reconstruction (CCVR) often requires a large-volume transfusion of blood products. We implemented a series of improvement interventions to reduce blood donor exposures (BDE) and transfusion requirements in CCVR.

METHODS:

We implemented interventions over 4 epochs: (E1) reconstituted blood (1:1 ratio of donor-matched red blood cells and fresh-frozen plasma) for intraoperative transfusions, (E2) reconstituted blood plus postoperative transfusion guidelines, (E3) reconstituted blood plus intraoperative antifibrinolytics and postoperative guidelines, and (E4) fresh whole blood for intraoperative transfusion, antifibrinolytics, and postoperative guidelines. Primary outcomes, BDE, and total volume of blood products transfused are presented by using statistical process control charts, with statistical comparisons between each epoch and baseline data.

RESULTS:

We included 347 patients <72 months old who underwent CCVR between 2008 and 2016 (E1: n = 50; E2: n = 41; E3: n = 87; and E4: n = 169). They were compared with a baseline sample group of 138 patients who were managed between 2001 and 2006. Compared with our baseline group, patients in each epoch had a significant reduction in BDE (P = .02–<.0001). Conversely, compared with the baseline group, we observed an increase the volume of blood products transfused in E1 (P = .004), no difference in E2 (P = .6) or E3 (P = .46), and a reduction in the volume of blood products transfused in E4 (P < .0001).

CONCLUSIONS:

The implementation of sequential clinical improvement strategies resulted in a sustained reduction in BDE whereas only the use of whole blood resulted in a significant reduction in the total volume of blood products transfused in children undergoing CCVR.

The management of children who are undergoing complex cranial vault reconstruction (CCVR) requires a high degree of care coordination to achieve optimal outcomes. CCVR involves the use of a wide scalp dissection and multiple craniotomies, which place infants and toddlers at risk for substantial blood loss. Intraoperative blood loss routinely exceeds the circulating blood volume14  in young children, and replacement of losses may result in coagulopathy.5  This can lead to more transfusions and blood donor exposures (BDE), with the potential for increased transfusion-related complications. Therefore, a systematic approach to perioperative transfusions is essential.

Before 2008, we had no formal protocols to direct transfusion management in patients who were undergoing CCVR, and patient outcomes were not systematically evaluated. In 2008, we developed a local prospective registry to track the management and outcomes of patients who were undergoing CCVR within our institution and to monitor the impact of our improvement interventions. We began entering similar data into the multicenter Pediatric Craniofacial Surgery Perioperative Registry in 2012. Our group has used these registries to systematically evaluate the effects of a variety of interventions in our surgical population. We have demonstrated that monitoring central venous pressure is not an accurate measure of volume status in CCVR surgery,6  that reconstituted blood7  and antifibrinolytics3  can reduce blood loss and donor exposures, that the use of postoperative transfusion guidelines can lead to a more rational postoperative transfusion practice,8  and that whole blood provides some unique advantages in terms of limiting the incidence of perioperative coagulopathy in patients who are undergoing CCVR.9  All of these studies were conducted as retrospective case-control studies with specific inclusion and exclusion criteria. However, we have not analyzed the summative effects of our interventions over time in our entire craniofacial population nor have we evaluated the impact of these interventions on total perioperative blood transfusion rates. In the current report, we present the aggregated results of a series of stepwise improvement strategies and illustrate the impact of these interventions on BDE and total perioperative transfusion rates in our patient population over time.

This longitudinal quality improvement (QI) project was conducted at a major tertiary-care pediatric hospital that performs 30 to 50 CCVR surgeries in children aged ≤5 years annually. We focused on the care of patients aged ≤5 years because they represent the majority of children who undergo this surgery and are at risk for significant hemorrhage and transfusion.

Our inclusion criteria were an age of ≤72 months and CCVR surgery that included either fronto-orbital advancement or posterior cranial vault reconstruction. All procedures involved a craniotomy and were performed jointly by plastic surgeons and neurosurgeons. The first data source is a deidentified data set of patients who were managed from December 2001 through January 2006, which served as our historical control.5  These data were combined with data from 2 prospective, temporally sequential, and nonoverlapping registries. The prospective local registry contains data from April 2008 to June 2012 and has been described in previous publications.7,8  The prospective multicenter Pediatric Craniofacial Surgery Perioperative Registry was the source for patient data from June 2012 through September 2016. Additional details regarding this registry have been published elsewhere.1,10,11 

Setting and Stakeholders

Surgical management of the CCVR population at our institution has been consistent over the course of the interventions described below. All patients were managed by 1 of 3 surgeons in the baseline group while surgical management in epoch (E)1 to E4 included 2 surgeons from the baseline group as well as a third surgeon who joined the practice in 2011. Surgical management was consistent among the surgeons in this study, and there were no changes in the surgical technique (scalp blocks, surgical clips, etc) employed to specifically limit surgical blood loss.

Stakeholder participation evolved over time depending on the intervention. The roles of the anesthesiology team and surgical team have been consistent throughout. Like the surgical service, the anesthesiology team responsible for the care of these patients is limited to a small group of anesthesiologists who are familiar and comfortable with this surgical population. With the implementation of the reconstituted blood program, we identified stakeholders within the blood bank to assist with obtaining donor-matched packed red blood cells (PRBCs) and fresh-frozen plasma (FFP). These same stakeholders were instrumental in organizing our transition to the use of whole blood for intraoperative transfusions in this surgical population. Similarly, when developing postoperative transfusion guidelines for this surgical population, we identified key stakeholders in the ICU who could assist with the development and implementation of the transfusion guidelines.

Baseline Group: December 2001 to January 2006

This cohort contains patients who were managed before any improvement interventions between December 2001 and January 2006. There was no formal standardization of perioperative care at this time, and intraoperative blood loss was replaced by using crystalloid and blood component therapy. Typically, blood component therapy involves the transfusion of PRBCs first to treat anemia followed by hemostatic blood products (FFP, platelets, or cryoprecipitate). In an analysis of outcomes in this cohort, we found that 24% of patients were coagulopathic in the immediate postoperative period and intraoperative FFP administration was associated with a decreased likelihood of postoperative coagulopathy. We also identified that 74% of all patients received FFP at some point perioperatively.5 

E1: Reconstituted Blood for Intraoperative Blood Loss Replacement

On the basis of the findings in the preintervention cohort, our first plan-do-study-act (PDSA) cycle was to change our intraoperative blood loss replacement procedure from traditional component therapy to a prophylactic hemostatic resuscitation strategy in which FFP and PRBCs were administered in a 1:1 ratio to replace surgical blood loss. This was achieved by combining donor-matched FFP and PRBC units to make reconstituted blood. We hypothesized that this replacement strategy would prevent postoperative coagulopathy and in turn reduce postoperative transfusions. Additionally, we posited that the use of donor-matched PRBC and FFP units (ie, units derived from the same blood donor) would reduce total perioperative BDE.7  We partnered with our local American Red Cross to obtain donor-matched PRBC and FFP units. For each child, 2 sets of donor-matched FFP and PRBC units were prepared (1 U FFP and 1 U PRBCs each from 2 blood donors). Donor-matched units were not uniformly available for all patients in this cohort either because of ordering errors during program inception or in cases in which directed blood donations were arranged. However, all patients who were treated during this time period were included in our QI cohort, and the percentage of patients who received reconstituted blood is reported as a process measure.

E2: Implementation of Postoperative Transfusion Guidelines

In the course of transfusion outcome surveillance after our transition to reconstituted blood, we discovered that a number of children were reflexively being transfused FFP postoperatively on the basis of drain outputs without laboratory evidence of coagulopathy or hemodynamic instability. Having achieved improvements with the standardization of intraoperative transfusion management, we recognized that a similar level of scrutiny should be applied to the postoperative period. At our institution, all patients who undergo CCVR are admitted to the PICU postoperatively. We shared our postoperative transfusion findings with our PICU colleagues and focused our second PDSA cycle on the development and implementation of transfusion guidelines for postoperative management. The guideline consisted of a printed sheet that contained the mean ± 1.5 SDs (derived from local data) of the expected weight-based surgical drain output together with agreed on thresholds for PRBC, FFP, and platelet transfusions. Our intervention was simply to present this printed sheet to the bedside nurse and ICU team during our formal transfer of care to the ICU. This guidance was well received by the ICU team, which appreciated having concrete guidelines on which to base clinical decisions. These guidelines resulted in a 60% reduction in postoperative transfusions.8 

E3: Implementation of Intraoperative Antifibrinolytic Therapy

On the basis of literature demonstrating the efficacy of antifibrinolytics in reducing perioperative transfusions in pediatric CCVR and other pediatric major surgeries,1215  together with newly available pharmacokinetic data on this population,13,16  our third PDSA cycle was to incorporate antifibrinolytics into the routine management of our CCVR patients.3  Patients were managed with either aminocaproic acid (vast majority of patients) or tranexamic acid in addition to reconstituted blood and the postoperative transfusion guidelines. This intervention was implemented after discussions with our craniofacial surgeons, neurosurgeons, and pharmacists.

E4: Transition From Reconstituted Blood to Fresh Whole Blood

Our final PDSA cycle involved transitioning from reconstituted blood to leukoreduced fresh whole blood (FWB) (defined as cold storage ≤7 days) for the replacement of intraoperative blood loss. This change was based on evidence that revealed improved outcomes with FWB in children who underwent open-heart surgery, a population also at risk for significant perioperative transfusion.17,18  FWB has advantages in the setting of massive transfusions because it contains both soluble clotting factors and platelets, theoretically providing greater hemostatic activity. In an in vitro study of whole blood that was stored at 1 to 6°C by using thromboelastography and platelet aggregometry, it was found that normal coagulation was preserved for at least 11 days.19  There was already a system in place within our institution to obtain FWB for children undergoing cardiac surgery. To implement this test of change, we collaborated with our institution’s blood bank to make FWB available. FWB was administered with a platelet-sparing leukoreduction filter set, which allowed for the prevention of febrile reactions, HLA antigen alloimmunization, and cytomegalovirus transmission without taking away platelet function. Patients who were included in this epoch also received intraoperative antifibrinolytics and were managed with the postoperative transfusion guidelines described above. In cases in which FWB was not available, our intraoperative guidelines were updated to recommend replacement with traditional component therapy with early consideration of FFP administration because donor-matched FFP and PRBCs units were no longer available.

Process measures varied in each epoch. Compliance with specific blood product administration strategies was tracked with the expectation that a majority of patients would receive reconstituted blood in E1, E2, and E3 whereas a majority would receive FWB in E4. Similarly, we tracked compliance with antifibrinolytic therapy.

The effect of our interventions was tracked by using 2 transfusion outcomes: total perioperative BDE and total perioperative blood product transfusion. Total perioperative BDE included all unique exposures that occurred intraoperatively through postoperative day 4. Total perioperative blood product transfusion included any blood products that were transfused intraoperatively or in the first postoperative 48 hours. These composite outcomes reflect an array of clinical conditions, including blood loss, coagulation abnormalities, and hemodynamic stability.

Demographic data were captured for each improvement epoch. We characterized the presenting diagnosis on the basis of the number of synostotic cranial sutures and whether the child had a craniosynostosis syndrome (eg, Apert syndrome). A small subset of patients had other deformities (eg, frontonasal dysplasia and hypertelorism) that were included because they were treated with CCVR. We made comparisons between the baseline group and the patients in each epoch to assess for demographic differences. Process measures are presented as a percentage of patients receiving a specific therapy (reconstituted blood, antifibrinolytics, or FWB).

We report our primary outcomes using statistical process control (SPC) charts for total BDE and the mean total perioperative blood products transfused (mL/kg). Typically, we perform 4 to 5 CCVR cases monthly, so grouping patients by month or quarter creates group variability, which would make trending of outcomes challenging. To address this, we grouped sequential patients into subgroups of 10 (range: 7–11 on the basis of the number of patients in each respective epoch) and analyzed each subgroup using X-bar charts. X-bar charts reveal the average for a given outcome with control limits for each subgroup over time. Charts were created with SPC for Excel (BPI Consulting, LLC, Cypress, TX). The historical preintervention cohort (December 2001–January 2006) was used to generate the historical average and trial control limits for BDE and the total perioperative volume of blood products administered. Control limits were set at 3 σ. Using the rule of ≥8 consecutive points above or below the centerline to identify special cause variation,20  we adjusted the SPC charts when special cause variation was observed in the data.

Finally, we compared the mean BDE and the mean volume of perioperative blood products transfused (mL/kg) between the baseline group and the patients in each epoch to assess the impact of the interventions. Data were analyzed with JMP version 11.0 (SAS Institute, Inc, Cary, NC). Parametric or nonparametric tests were used for comparisons depending on the normality of data and the type of variables. χ2 or Fisher’s exact test was used for categorical data and the Student’s t test was used for comparisons with continuous data. A P value of <.05 was considered significant.

A total of 347 patients from April 2008 to September 2016 were analyzed together with a historical preintervention cohort of 138 patients. Patient characteristics are summarized in Table 1. Patients who were managed in each epoch were similar in age, weight, and sex with the exception that a greater proportion of patients in E4 were boys when compared with the baseline group. More patients were managed with a posterior approach in E2, E3, and E4 when compared with those in the baseline group and E1. The preoperative craniofacial diagnosis was similar to that in the baseline group with the exception that more patients were diagnosed with syndromic or ≥3 sutures in E4 compare with the baseline group. Finally, the incidence of a previous craniofacial surgery was similar between the baseline group and the patients in the intervention epochs.

TABLE 1

Characteristics of Patients Undergoing CCVR

VariableBaseline (n = 138)E1 (n = 50)PaE2 (n = 41)PaE3 (n = 87)PaE4 (n = 169)Pa
Age, mo, mean ± SD 16.9 ± 14.7 17.8 ± 16.2 .72 17 ± 13.9 .95 17.2 ± 16.2 .88 19 ± 16.3 .24 
Weight, kg, mean ± SD 10.2 ± 3.2 11.2 ± 4.7 .12 10.3 ± 3.5 .87 10.2 ± 4 .99 11.4 ± 9.8 .26 
Male sex, n (%) 74 (54) 29 (58) .59 25 (61) .4 40 (46) .26 112 (66) .02 
Diagnosis, n (%)   .26  .98  .1  .02 
 ≤2-suture synostosis 115 (83) 37 (74) — 34 (83) — 63 (72) — 120 (71) — 
 Syndromic or ≥3 sutures 19 (14) 12 (24) — 6 (15) — 22 (25) — 45 (27) — 
 Other 4 (3) 1 (2) — 1 (2) — 2 (2) — 4 (2) — 
Procedure category, n (%)   .37  .007  .006  <.001 
 Fronto-orbital advancement 118 (86) 40 (80) — 27 (66) — 61 (70) — 110 (65) — 
 Posterior cranial vault reconstruction 20 (14) 10 (20) — 14 (34) — 26 (30) — 59 (35) — 
Previous craniofacial surgery 32 (23) 14 (28) .5 10 (24) .87 24 (28) .46 39 (23) .98 
VariableBaseline (n = 138)E1 (n = 50)PaE2 (n = 41)PaE3 (n = 87)PaE4 (n = 169)Pa
Age, mo, mean ± SD 16.9 ± 14.7 17.8 ± 16.2 .72 17 ± 13.9 .95 17.2 ± 16.2 .88 19 ± 16.3 .24 
Weight, kg, mean ± SD 10.2 ± 3.2 11.2 ± 4.7 .12 10.3 ± 3.5 .87 10.2 ± 4 .99 11.4 ± 9.8 .26 
Male sex, n (%) 74 (54) 29 (58) .59 25 (61) .4 40 (46) .26 112 (66) .02 
Diagnosis, n (%)   .26  .98  .1  .02 
 ≤2-suture synostosis 115 (83) 37 (74) — 34 (83) — 63 (72) — 120 (71) — 
 Syndromic or ≥3 sutures 19 (14) 12 (24) — 6 (15) — 22 (25) — 45 (27) — 
 Other 4 (3) 1 (2) — 1 (2) — 2 (2) — 4 (2) — 
Procedure category, n (%)   .37  .007  .006  <.001 
 Fronto-orbital advancement 118 (86) 40 (80) — 27 (66) — 61 (70) — 110 (65) — 
 Posterior cranial vault reconstruction 20 (14) 10 (20) — 14 (34) — 26 (30) — 59 (35) — 
Previous craniofacial surgery 32 (23) 14 (28) .5 10 (24) .87 24 (28) .46 39 (23) .98 

—, not applicable.

a

Comparisons are presented between the baseline group and patients in each individual epoch by using Student’s t test for continuous variables or the χ2 test for categorical data.

We tracked compliance with our evolving transfusion strategy and the use of antifibinolytics, and the results are shown in Fig 1. Changes in blood management are reflected by the increase in the percentage of patients receiving reconstituted blood in E1 (88%), E2 (93%), and E3 (92%). Our ability to obtain and administer reconstituted blood was highly dependent on care teams in the plastic surgery and anesthesiology departments as well as on the blood bank. Patients were identified early, and appropriate orders were placed with sufficient time to allow for preparation of the reconstituted blood for the planned surgical procedures. Without this coordination, it is unlikely that we would have been able to achieve such a high rate of compliance. In E4, 76% of patients were managed with FWB. Compliance with FWB was reduced compared with our experience with reconstituted blood because this is a finite resource that is not always available in the blood bank. We have had to accept this as a limitation of this transfusion strategy, and when it is not available, we rely on component therapy for intraoperative transfusions. The intraoperative administration of antifibrinolytics was introduced in E3, and compliance was 97% and 93% in E3 and E4, respectively. The high compliance with this metric is the result of a standardization of care across the small group of anesthesiologists who routinely care for this surgical population.

FIGURE 1

Improvement intervention compliance with intraoperative transfusion strategy and antifibrinolytic administration over time.

FIGURE 1

Improvement intervention compliance with intraoperative transfusion strategy and antifibrinolytic administration over time.

Close modal

BDE

An SPC chart for the mean total BDE is displayed in Fig 2. Special cause variation was noted early in this improvement work. Beginning in E1, all points on the SPC chart were below the historical centerline, and a centerline shift was made in the chart. Interestingly, the average BDE appeared to have a fixed lower limit of ∼2 exposures in our population because subsequent interventions, including postoperative transfusion guidelines, antifibrinolytics, and FWB, did not result in a further reduction in the average BDE that achieved special cause. We compared the mean BDE in each epoch with the mean BDE in the baseline group, which was 4 exposures (Table 2). When comparing the baseline mean BDE (4) to that in each epoch, we found that with the initiation of the reconstituted blood protocol in E1, the mean BDE decreased to 3.2 (P = .02). In E2, the mean BDE decreased to 2.4 (P < .0001), and in E3, the mean BDE decreased to 2 (P < .0001). Finally, we observed our lowest mean BDE in E4 in which the average exposure was 1.7 (P < .0001).

FIGURE 2

Total BDE: SPC chart revealing patient BDE over time with control limits (3 SDs). Interventions were implemented over 4 epochs: RB (E1); RB and postoperative transfusion guidelines (E2); RB, transfusion guidelines, and antifibrinolytics (E3); and whole blood, transfusion guidelines, and antifibrinolytics (E4). LCL, lower control limit; RB, reconstituted blood; UCL, upper control limit.

FIGURE 2

Total BDE: SPC chart revealing patient BDE over time with control limits (3 SDs). Interventions were implemented over 4 epochs: RB (E1); RB and postoperative transfusion guidelines (E2); RB, transfusion guidelines, and antifibrinolytics (E3); and whole blood, transfusion guidelines, and antifibrinolytics (E4). LCL, lower control limit; RB, reconstituted blood; UCL, upper control limit.

Close modal
TABLE 2

Mean BDE During Intraoperative and Postoperative Care of Patients Undergoing CCVR by Epoch

VariableMean (95% CI)Difference in Mean (95% CI)P
Baseline 4 (3.7 to 4.4) n/a n/a 
E1 3.2 (2.6 to 3.8) −0.8 (−0.1 to −1.5) .02 
E2 2.4 (1.8 to 3) −1.7 (−1 to −2.4) <.0001 
E3 2 (1.8 to 2.2) −2 (−1.6 to −2.5) <.0001 
E4 1.7 (1.5 to 1.9) −2.3 (−1.9 to −2.7) <.0001 
VariableMean (95% CI)Difference in Mean (95% CI)P
Baseline 4 (3.7 to 4.4) n/a n/a 
E1 3.2 (2.6 to 3.8) −0.8 (−0.1 to −1.5) .02 
E2 2.4 (1.8 to 3) −1.7 (−1 to −2.4) <.0001 
E3 2 (1.8 to 2.2) −2 (−1.6 to −2.5) <.0001 
E4 1.7 (1.5 to 1.9) −2.3 (−1.9 to −2.7) <.0001 

Comparisons are presented between the baseline group and patients in each individual epoch by using Student’s t test. CI, confidence interval; n/a; not applicable.

Volume of Perioperative Blood Products Transfused

An SPC chart that reveals the total volume of blood products administered (mL/kg) during the perioperative period is shown in Fig 3. Compared with BDE, we observed a special cause reduction in the total volume of products late in our improvement work. Beginning toward the end of E3 and extending into E4, we observed consecutive data points below the centerline, and a centerline shift was introduced into the chart. The routine use of FWB units, which contain all the clotting factors that are needed for hemostasis, allowed for adequate resuscitation with lower total volumes of blood products when compare with combining PRBC units with hemostatic units (FFP, cryoprecipitate, or platelets). Additional comparisons between the baseline group and patients in each epoch are reported in Table 3. In E1, we observed a statistically significant increase in the total volume of blood products administered compared with that in the baseline group (84.2 vs 114.7 mL/kg; P = .004). We suspect that the observed increase in transfusion volumes was partly due to an intraoperative transfusion practice of targeting higher hemoglobin levels in anticipation of postoperative losses into surgical drains. These higher hemoglobin values were achieved by transfusing the remainder of partially transfused units of reconstituted blood. If there was no remainder, a lower hemoglobin concentration was accepted rather than transfusing from a new unit of blood. The intent of this practice was to transfuse blood that the patient had already been exposed to, targeting a higher hemoglobin concentration to buffer against anticipated postoperative losses to avoid subsequent transfusions. There was no difference in the volume of perioperative blood products transfused in E2 (84.2 vs 89.1 mL/kg; P = .85) or E3 (84.2 vs 79.7 mL/kg; P = .66). However, there was a significant reduction in the overall volume of blood products transfused in E4 with our transition to whole blood for intraoperative resuscitation (84.2 vs 52.9 mL/kg; P < .0001).

FIGURE 3

Total perioperative blood product administration: SPC chart revealing total perioperative blood products transfused in the first 48 hours of hospitalization with control limits (3 SDs). The perioperative blood products that are tracked include PRBCs, FFP, platelets, or cryoprecipitate. Interventions were implemented over 4 epochs: RB (E1); RB and postoperative transfusion guidelines (E2); RB, transfusion guidelines, and antifibrinolytics (E3); and whole blood, transfusion guidelines, and antifibrinolytics (E4). LCL, lower control limit; RB, reconstituted blood; UCL, upper control limit.

FIGURE 3

Total perioperative blood product administration: SPC chart revealing total perioperative blood products transfused in the first 48 hours of hospitalization with control limits (3 SDs). The perioperative blood products that are tracked include PRBCs, FFP, platelets, or cryoprecipitate. Interventions were implemented over 4 epochs: RB (E1); RB and postoperative transfusion guidelines (E2); RB, transfusion guidelines, and antifibrinolytics (E3); and whole blood, transfusion guidelines, and antifibrinolytics (E4). LCL, lower control limit; RB, reconstituted blood; UCL, upper control limit.

Close modal
TABLE 3

Mean Total Perioperative Blood Products Transfused (mL/kg) During Intraoperative and Postoperative Care of Patients Undergoing CCVR by Epoch

VariableMean (95% CI)Difference in Mean (95% CI)P
Baseline 84.2 (76.3 to 92) n/a n/a 
E1 114.7 (95.6 to 133.9) 30.5 (51.1 to 9.9) .004 
E2 89.1 (75.1 to 103.1) 4.9 (23.4 to −13.6) .6 
E3 79.7 (70.1 to 89.3) −4.5 (7.4 to −16.4) .46 
E4 52.9 (47.5 to 58.3) −31.2 (−21.8 to −40.7) <.0001 
VariableMean (95% CI)Difference in Mean (95% CI)P
Baseline 84.2 (76.3 to 92) n/a n/a 
E1 114.7 (95.6 to 133.9) 30.5 (51.1 to 9.9) .004 
E2 89.1 (75.1 to 103.1) 4.9 (23.4 to −13.6) .6 
E3 79.7 (70.1 to 89.3) −4.5 (7.4 to −16.4) .46 
E4 52.9 (47.5 to 58.3) −31.2 (−21.8 to −40.7) <.0001 

Comparisons are presented between the baseline group and patients in each individual epoch by using Student’s t test. CI, confidence interval; n/a, not applicable.

Given its association with significant perioperative transfusion requirements, CCVR represents a unique clinical scenario that is amenable to the implementation of sequential perioperative QI interventions that are designed to enhance patient safety and outcomes. In this study, we present a longitudinal view of our sequential interventions to improve perioperative blood management in CCVR and demonstrate meaningful and sustained reductions in our perioperative transfusion outcomes. By focusing on reducing BDE, we are improving patient safety by reducing patient exposure to the rare but significant risks of transfusion reactions, alloimmunization, and infection.21 

Caring for children who are undergoing CCVR is challenging given the potential rate and volume of blood loss that is associated with this surgical procedure. Careful preoperative preparation is critical to ensure optimal patient management and outcome. Historically at our institution, individual anesthesiologists determined the approach to managing intraoperative blood loss, whereas postoperative transfusion management was at the discretion of the attending pediatric intensivist. We did not look at our outcomes with any rigor, and care was not standardized. Beginning in 2008, we evaluated the management and outcomes for this population, and we observed a great deal of variability in terms of blood products used both intraoperatively and postoperatively as well as a significant incidence of coagulopathy in the immediate postoperative period. With this knowledge, we convened a small improvement group to focus on standardizing the care of this surgical population with the specific goal of optimizing perioperative blood management. In the years after, our incremental improvement efforts have resulted in a stepwise and sustained reduction in both the number of BDE incurred by these children perioperatively and the volume of blood products transfused. Although there were some differences in the procedural approach, with more patients undergoing fronto-orbital advancement as opposed to posterior vault reconstruction in the baseline group and E1 compared with E2, E3, and E4, the impact of a surgical approach on blood loss remains unclear, with some studies revealing that surgical approach is not an independent predictor of blood loss.5,22  However, it is possible that a lower rate of fronto-orbital advancement in E2, E3, and E4 may have influenced our findings of lower BDE and lower total perioperative transfusion volumes in E4 when compared with our baseline group.

When considering the outcome of BDE, our data suggest that 1 BDE is the lower limit for this surgery at our institution because all but 1 patient in our cohort had at least 1 transfusion. In CCVR, there was a well-documented inverse relationship between patient age and weight and transfusion volumes that are indexed to weight.2,5,22,23  Consistent with this, our lone transfusion-free patient was 5.5 years old and in the 93rd weight percentile of our data set. Because most patients who undergo this procedure are infants and small children, almost all patients require a transfusion at some point during their hospitalization. Despite the near universality of transfusion, we have demonstrated that through iterative improvement work, patient donor exposures could be reduced from an average of 4 BDE to an average of <2.

Our first intervention (routine use of reconstituted blood) marked the beginning of a sustained reduction in BDE that achieved special cause variation. The average BDE were incrementally reduced in each epoch, beginning with a 20% reduction in average BDE in E1 and ending with a 58% reduction in BDE in E4 compared with that in the baseline group. While there were incremental reductions in BDE with interventions introduced in E2, E3, and E4, these reductions did not reach special cause variations.

The other outcome we tracked was the impact of our interventions on the volume of perioperative blood products transfused. Unexpectedly, we noted an association between our use of reconstituted blood in E1 and a statistically significant increase in the total volume of perioperative blood products administered (114.7 vs 84.2 mL/kg), which did not result in special cause variation. As noted above, this may have been a result of targeting higher hemoglobin levels in the operating room. Additionally, an analysis of the outcomes of patients who received reconstituted blood in E1 revealed that many patients were transfused with hemostatic blood products (primarily FFP) to treat what was perceived to be a high surgical drain output as well as to treat mildly abnormal coagulation test results. These likely unnecessary transfusions may also have contributed to the increased transfusion volumes that were observed in E1. On the basis of this unexpected finding, we initiated a second PDSA cycle to develop a standardized postoperative transfusion protocol for this surgical population.

Despite evidence that supported the safety and efficacy of restrictive red blood cell transfusion thresholds in pediatric patients postsurgery,24  transfusion thresholds were not formalized at our institution or at most other US institutions.25  To remedy this, we partnered with relevant clinician stakeholders, including surgeons, intensivists, and anesthesiologists, and developed postoperative management guidelines that defined specific thresholds for red blood cell and hemostatic blood product transfusions. Our transfusion guidelines are presented during handoffs between the anesthesia team and the intensive care team. The form is then attached to the patient’s bed or included in the chart. This intervention was readily adopted by the ICU team, which expressed enthusiastic support for this type of postoperative guidance. With this intervention, we observed a statistically significant reduction in BDE and in total perioperative blood product transfusion volumes to a level similar to that in the baseline group (89.1 vs 84.2 mL/kg).

With implementation of the routine use of antifibrinolytics to help reduce intraoperative bleeding, compliance with administration was >90% in E3 and E4. This intervention was embraced by the small cohort of anesthesiologists who cared for this population, which is likely the reason for our high compliance. Interestingly, we did note a trend toward reduced volumes of blood products required toward the end of E3, but the average total volume of perioperative blood products administered was not significantly reduced when compared with our baseline data (79.7 vs 84.2 mL/kg). There is compelling evidence, including prospective randomized trials, that antifibrinolytics are effective in reducing blood loss and transfusions in craniofacial surgery and other pediatric surgical populations.3,12,13,26,27  We did not observe a significant reduction in the volume of blood products required with the routine use of antifibrinolytic therapy in E3. However, this should not lead the reader to question the role of antifibrinolytics in the management of this population. In general, the goal in QI initiatives is to drive the adoption of evidence-based practices rather than to establish evidence of efficacy. As a result, the subtle benefits in blood loss and transfusion requirements that were observed in controlled clinical trials may be muted in an analysis that includes all patients regardless of whether they received antifibrinolytic therapy.

The transfusion of FWB greatly simplifies blood loss replacement procedures in children who are undergoing CCVR. FWB provides red blood cells and all of the blood elements that are necessary for coagulation in an iso-oncotic solution, and we have previously demonstrated that the replacement of intraoperative blood loss with FWB is as effective as reconstituted blood in preventing the development of dilutional coagulopathy. The use of whole blood in massive transfusion scenarios is increasingly being revisited as a potentially superior strategy,2830  and our institution’s experience with cardiac surgery supports the targeted use of FWB for populations such as our CCVR population. For these reasons, we incorporated FWB to enhance our management of this specific surgical population, but we acknowledge that this resource may not currently be available in all hospital systems.

After implementation, FWB was administered in 76% of cases. Even with nonuniform FWB availability, we observed the lowest average BDE in E4 (1.7 vs 4), which was associated with a statistically significant reduction in total perioperative blood products administered compared with the preintervention cohort (52.9 vs 84.2 mL/kg). These findings suggest that the impact may have been greater if all patients were managed with FWB.

The strengths of this QI initiative include the large number of patients, a long duration of follow-up, the discrete interventions, and a systematic prospective data capture. Although we have previously described the outcomes of these interventions,1,3,511  the authors of those studies did not report on our “real world” experience of caring for this population because they excluded patients who were not receiving the designated interventions. The data we have presented provide insight into the summative effect of these sequential improvement efforts. We acknowledge that interventions that are effective at our institution might not be as effective or applicable at other institutions. However, our approach of iterative PDSA cycles using local registry data could be applied at other institutions.

We have presented the longitudinal results of our approach to improving transfusion outcomes and the safety of children who are undergoing CCVR. The approaches we describe represent a framework that can be applied in this surgical population at other institutions to optimize transfusion outcomes.

FUNDING: Supported by internal funds from the Children’s Hospital of Philadelphia Department of Anesthesiology and Critical Medicine as well as the Division of Plastic Surgery.

Dr Muhly led the analysis of the data and drafted the initial manuscript; Drs Tan and Hsu made a substantial contribution in data analysis and reviewed and revised the manuscript; Drs Sesok-Pizzini, Fiadjoe, Taylor, and Bartlett made substantial contributions to the conceptualization and design of the quality improvement pathway and reviewed and revised the manuscript; Dr Stricker conceptualized and designed the quality improvement pathway, led the development of data acquisition and analysis, and assisted in drafting the initial manuscript; and all authors approved the final manuscript as submitted.

1
Stricker
PA
,
Goobie
SM
,
Cladis
FP
, et al
;
Pediatric Craniofacial Collaborative Group
.
Perioperative outcomes and management in pediatric complex cranial vault reconstruction: a multicenter study from the Pediatric Craniofacial Collaborative Group
.
Anesthesiology
.
2017
;
126
(
2
):
276
287
2
White
N
,
Marcus
R
,
Dover
S
, et al
.
Predictors of blood loss in fronto-orbital advancement and remodeling
.
J Craniofac Surg
.
2009
;
20
(
2
):
378
381
3
Hsu
G
,
Taylor
JA
,
Fiadjoe
JE
, et al
.
Aminocaproic acid administration is associated with reduced perioperative blood loss and transfusion in pediatric craniofacial surgery
.
Acta Anaesthesiol Scand
.
2016
;
60
(
2
):
158
165
4
Oppenheimer
AJ
,
Ranganathan
K
,
Levi
B
, et al
.
Minimizing transfusions in primary cranial vault remodeling: the role of aminocaproic acid
.
J Craniofac Surg
.
2014
;
25
(
1
):
82
86
5
Stricker
PA
,
Shaw
TL
,
Desouza
DG
, et al
.
Blood loss, replacement, and associated morbidity in infants and children undergoing craniofacial surgery
.
Paediatr Anaesth
.
2010
;
20
(
2
):
150
159
6
Stricker
PA
,
Lin
EE
,
Fiadjoe
JE
, et al
.
Evaluation of central venous pressure monitoring in children undergoing craniofacial reconstruction surgery
.
Anesth Analg
.
2013
;
116
(
2
):
411
419
7
Stricker
PA
,
Fiadjoe
JE
,
Davis
AR
, et al
.
Reconstituted blood reduces blood donor exposures in children undergoing craniofacial reconstruction surgery
.
Paediatr Anaesth
.
2011
;
21
(
1
):
54
61
8
Stricker
PA
,
Fiadjoe
JE
,
Kilbaugh
TJ
, et al
.
Effect of transfusion guidelines on postoperative transfusion in children undergoing craniofacial reconstruction surgery
.
Pediatr Crit Care Med
.
2012
;
13
(
6
):
e357
e362
9
Thottathil
P
,
Sesok-Pizzini
D
,
Taylor
JA
,
Fiadjoe
JE
,
Vincent
A
,
Stricker
PA
.
Whole blood in pediatric craniofacial reconstruction surgery
.
J Craniofac Surg
.
2017
;
28
(
5
):
1175
1178
10
Goobie
SM
,
Cladis
FP
,
Glover
CD
, et al
;
Pediatric Craniofacial Collaborative Group
.
Safety of antifibrinolytics in cranial vault reconstructive surgery: a report from the pediatric craniofacial collaborative group
.
Paediatr Anaesth
.
2017
;
27
(
3
):
271
281
11
Thompson
DR
,
Zurakowski
D
,
Haberkern
CM
, et al
;
Pediatric Craniofacial Collaborative Group
.
Endoscopic versus open repair for craniosynostosis in infants using propensity score matching to compare outcomes: a multicenter study from the Pediatric Craniofacial Collaborative Group
.
Anesth Analg
.
2018
;
126
(
3
):
968
975
12
Dadure
C
,
Sauter
M
,
Bringuier
S
, et al
.
Intraoperative tranexamic acid reduces blood transfusion in children undergoing craniosynostosis surgery: a randomized double-blind study
.
Anesthesiology
.
2011
;
114
(
4
):
856
861
13
Goobie
SM
,
Meier
PM
,
Pereira
LM
, et al
.
Efficacy of tranexamic acid in pediatric craniosynostosis surgery: a double-blind, placebo-controlled trial
.
Anesthesiology
.
2011
;
114
(
4
):
862
871
14
Tzortzopoulou
A
,
Cepeda
MS
,
Schumann
R
,
Carr
DB
.
Antifibrinolytic agents for reducing blood loss in scoliosis surgery in children
.
Cochrane Database Syst Rev
.
2008
;(
3
):
CD006883
15
Chauhan
S
,
Das
SN
,
Bisoi
A
,
Kale
S
,
Kiran
U
.
Comparison of epsilon aminocaproic acid and tranexamic acid in pediatric cardiac surgery
.
J Cardiothorac Vasc Anesth
.
2004
;
18
(
2
):
141
143
16
Stricker
PA
,
Zuppa
AF
,
Fiadjoe
JE
, et al
.
Population pharmacokinetics of epsilon-aminocaproic acid in infants undergoing craniofacial reconstruction surgery
.
Br J Anaesth
.
2013
;
110
(
5
):
788
799
17
Manno
CS
,
Hedberg
KW
,
Kim
HC
, et al
.
Comparison of the hemostatic effects of fresh whole blood, stored whole blood, and components after open heart surgery in children
.
Blood
.
1991
;
77
(
5
):
930
936
18
Jobes
DR
,
Sesok-Pizzini
D
,
Friedman
D
.
Reduced transfusion requirement with use of fresh whole blood in pediatric cardiac surgical procedures
.
Ann Thorac Surg
.
2015
;
99
(
5
):
1706
1711
19
Jobes
D
,
Wolfe
Y
,
O’Neill
D
, et al
.
Toward a definition of “fresh” whole blood: an in vitro characterization of coagulation properties in refrigerated whole blood for transfusion
.
Transfusion
.
2011
;
51
(
1
):
43
51
20
Provost
LP
,
Murray
SK
.
The Health Care Data Guide: Learning From Data for Improvement
. 1st ed.
San Francisco, CA
:
Jossey-Bass
;
2011
21
Goodnough
LT
.
Risks of blood transfusion
.
Anesthesiol Clin North America
.
2005
;
23
(
2
):
241
252, v
22
Chow
I
,
Purnell
CA
,
Gosain
AK
.
Assessing the impact of blood loss in cranial vault remodeling: a risk assessment model using the 2012 to 2013 Pediatric National Surgical Quality Improvement Program data sets
.
Plast Reconstr Surg
.
2015
;
136
(
6
):
1249
1260
23
Goobie
SM
,
Zurakowski
D
,
Proctor
MR
, et al
.
Predictors of clinically significant postoperative events after open craniosynostosis surgery
.
Anesthesiology
.
2015
;
122
(
5
):
1021
1032
24
Rouette
J
,
Trottier
H
,
Ducruet
T
,
Beaunoyer
M
,
Lacroix
J
,
Tucci
M
;
Canadian Critical Care Trials Group
PALISI Network
.
Red blood cell transfusion threshold in postsurgical pediatric intensive care patients: a randomized clinical trial
.
Ann Surg
.
2010
;
251
(
3
):
421
427
25
Stricker
PA
,
Cladis
FP
,
Fiadjoe
JE
,
McCloskey
JJ
,
Maxwell
LG
.
Perioperative management of children undergoing craniofacial reconstruction surgery: a practice survey
.
Paediatr Anaesth
.
2011
;
21
(
10
):
1026
1035
26
Martin
K
,
Gertler
R
,
Sterner
A
, et al
.
Comparison of blood-sparing efficacy of ε-aminocaproic acid and tranexamic acid in newborns undergoing cardiac surgery
.
Thorac Cardiovasc Surg
.
2011
;
59
(
5
):
276
280
27
Alajmi
T
,
Saeed
H
,
Alfaryan
K
,
Alakeel
A
,
Alfaryan
T
.
Efficacy of tranexamic acid in reducing blood loss and blood transfusion in idiopathic scoliosis: a systematic review and meta-analysis
.
J Spine Surg
.
2017
;
3
(
4
):
531
540
28
Leeper
CM
,
Yazer
MH
,
Cladis
FP
,
Saladino
R
,
Triulzi
DJ
,
Gaines
BA
.
Use of uncrossmatched cold-stored whole blood in injured children with hemorrhagic shock
.
JAMA Pediatr
.
2018
;
172
(
5
):
491
492
29
Spinella
PC
,
Cap
AP
.
Whole blood: back to the future
.
Curr Opin Hematol
.
2016
;
23
(
6
):
536
542
30
Murdock
AD
,
Berséus
O
,
Hervig
T
,
Strandenes
G
,
Lunde
TH
.
Whole blood: the future of traumatic hemorrhagic shock resuscitation
.
Shock
.
2014
;
41
(
suppl 1
):
62
69

Competing Interests

POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.

FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.