- Email: [email protected]
Predictive factors of intraoperative cell salvage during pediatric scoliosis surgery. Cell saver during scoliosis surgery in children
Accepted Manuscript Title: Predictive factors of intraoperative Cell Salvage during Pediatric Scoliosis Surgery Author: Daphn´e Michelet Florence Julien-Marsollier Julie Hilly Thierno Diallo Christophe Vidal Souhayl Dahmani PII: DOI: Reference:
S2352-5568(16)30227-2 http://dx.doi.org/doi:10.1016/j.accpm.2017.03.003 ACCPM 252
To appear in: Received date: Accepted date:
29-11-2016 17-3-2017
Please cite this article as: Daphn´e MicheletFlorence Julien-MarsollierJulie HillyThierno DialloChristophe VidalSouhayl Dahmani Predictive factors of intraoperative Cell Salvage during Pediatric Scoliosis Surgery (2017), http://dx.doi.org/10.1016/j.accpm.2017.03.003 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Predictive factors of intraoperative Cell Salvage during Pediatric Scoliosis Surgery. Cell Saver during scoliosis surgery in children
Cell Saver during scoliosis surgery in children # ♥,
Julie Hilly *
#♥
, Thierno Diallo *
#♥
,
cr
Daphné Michelet * # ♥, Florence Julien-Marsollier * Christophe Vidal ** # ♥, and SouhaylDahmani * #♥.
ip t
Predictive factors of intraoperative Cell Salvage during Pediatric Scoliosis Surgery.
♥DHU
an
us
* Department of anesthesia and Intensive care Robert Debré University Hospital. Paris, France. ** Department of pediatric orthopedic surgery, Robert Debré University Hospital. Paris, France. # Paris Diderot University (Paris VII).PRES Paris Sorbonne Cité. Paris. France PROTECT. INSERM U1141.Robert Debré University Hospital. Paris. France.
Ac
ce pt
ed
M
Corresponding author: Professor Souhayl Dahmani. Department of anesthesiology and Intensive Care. Robert Debré Hospital, 48 Boulevard Sérurier, 75019 Paris, France Phone: 00 33 1 40 03 41 83 Fax : 00 33 1 40 03 20 00 e-mail: [email protected]
1
Page 1 of 16
Ac
ce pt
ed
M
an
us
cr
ip t
Abstract Introduction: blood saving strategy during spinal surgery in children often includes recombinant erythropoietin (rEPO) and antifibrinolytic therapapy (AFT). The aim of this study was to investigate the efficacy of intraoperative blood salvage in decreasing homologous blood transfusion. Material and Methods: using the prospective data from patients operated during a one year period for scoliosis correction, we calculate the predictable hematocrit at day postoperative 1 without the use of blood salvage and compare it to the target hematocrit transfusion according to patient’s status. Predictors analyzed were: age, weight, surgical indication, Cobb’s angle, ASA status, preoperative hemoglobin, number of level fused, sacral fusion and thoracoplasty. Statistical analyses were performed using a classification tree analysis. Results: this study included 147 patients. Blood salvage was estimated avoiding homologous blood transfusion in 17 patients. Predictors of the efficacy of blood salvage were: neuromuscular indications, number of level fused and BMI. Blood salvage was found totally ineffective in: patients with no neuromuscular diseases with either: surgeries interesting < 13 levels fused or surgeries interesting > 13 levels with a preoperative BMI ≥ 21. In all other cases, blood salvage can decrease homologous transfusion. The model exhibited 97 % of accurate for the prediction if the inefficacy of blood salvage. The AUCROC of the model was 0.93 [95 % confidence interval 0.9 to 0.99] and the overall validation was 60.1 % of explained variability. Conclusion: The present study indicates that blood salvage is ineffective under certain circumstances. More studies are mandatory to confirm these results.
2
Page 2 of 16
Introduction
an
us
cr
ip t
Spinal surgery in children is a high risk procedure associated with intraoperative and postoperative complications (1). Ten percent of patients have neuromuscular diseases or cerebral palsy, thus increasing postoperative cardiovascular, respiratory and infectious complications (1-4). Spinal surgery is typically associated with extensive blood loss and perioperative transfusion (5). Moreover transfusion therapy has been shown to induce many specific complications such as bacterial infections, acute immunological incompatibility reactions and long term immunological effects (infections, tumors recurrence)(6,7). During the last decade, the use of blood saving strategies such as recombinant Erythropoietin (rEPO) and Antifibrinolytic therapy (AFT), resulted in a reduction of perioperative autologous blood transfusion (8-11). Consequently, the use of autologous blood donation has been completely withdrawal from current practice and blood salvage during the intraoperative period is contested given the few number of transfused patients. For example, a recent adult trial showed the ineffective effect of cell-saver in reducing autologous blood transfusion during major orthopedic surgeries while increasing costs(12). In the present study, we evaluated the efficacy of intraoperative blood salvage in preventing homologous blood transfusion in patients scheduled for scoliosis correction surgery managed with preoperative rEPO and intraoperative AFT. Material and Methods
ce pt
ed
M
The study consists on a prospective analysis of perioperative data of patients operated in our institution from January 2012 to December 2012. This study was approved by our institutional review board (Comité d’Evaluation de l’Ethique des projets de Recherche Biomédicale (CEERB) Paris Nord; # 12-021). Informed written consent was obtained from all patients (when allowed by their neurological status) and/or parents. All patients undergoing scoliosis surgery during the above quoted period were included in this study. Data used for the present manuscript were previously published for the purpose of analyzing the predictors of homologous transfusion (13).
Ac
Perioperative anesthesia and blood saving management Preoperative anesthesia and surgical preparation included standardized recombinant Erythropoietin (rEPO, Eprex®, JANSSEN-CILAG, Issy-Les-Moulineaux, France) treatment. Hemoglobin concentration was determined 24 hours before each scheduled date of injection, and a value strictly below 15 g/dl was followed by recombinant Erythropoietin injection. This consists in subcutaneous administration of 600 U/kg of rEPO weekly, started 4 weeks before the scheduled date of surgery with the last injection 48 hours before the date of surgery. Erythropoietin therapy was systematically associated with iron supplementation. Antifibrinolytic agent administration begins after the induction; it consists in tranexamic acid (Exacyl®, Sanofi-Aventis, Paris, France) given by a bolus of 50 mg/kg, plus a continuous intraoperative administration of 10 mg/kg/h. At skin closure, this drug was discontinued in order to respect its maximal dose recommended in France (4 g per day). Intraoperative blood salvage (Haemonetics Cell Saver 5, Haemonetics Corporation, Braintree, MA, USA) was used in all patients with an automatic salvage program in order to ensure blood transfusion with optimal hematocrit level (55 %). The cell-saver system was primed with 100 ml of normal saline solution and salvage process was initiated when at least 1000 ml of blood loss was available in the collector boil. According to our local protocols anesthesia was standardized. Transfusion protocol follows
3
Page 3 of 16
ip t
our national recommendations1. Intraoperative and postoperative hemoglobin concentration were maintained above 8 g.dl-1 (hematocrit = 0.24) for idiopathic or congenital scoliosis. This targeted concentration was increased to 10 g.dl-1 (hematocrit = 0.30) in case of neuromuscular scoliosis or in patients with ASA status > 2. This was achieved by autologous transfusion when available (intraoperative blood salvage) and homologous transfusion if necessary. Fresh frozen plasma and platelets were administered if perioperative prothrombin time was below 50 % and platelet count was below 50 000 elements.mm-3, respectively. Finally, all patients were given intravenous iron therapy (3 mg.kg-1) twice during the postoperative period (during 48 hours).
ed
M
an
us
cr
Surgical management (14-17) Hybrid constructs that included sublaminar bands in the thoracic spine, auto stable pediculo-supralaminar claws (Instinct, Zimmer Spine, Bordeaux, France) at the two proximal levels and pedicular screws (Shiraz, Zimmer spine, Bordeaux, France) at lumbar levels were placed into patients undergoing posterior correction and fusion for adolescent idiopathic scoliosis. Fusion levels were selected according to the same criteria during the entire study period. The 2 rods were adjusted to the desired sagittal profile and connected with 2 transverse connectors before placement. The rods were first introduced in the pedicle screws, and lumbar correction was performed. Thoracoplasty was performed by the posterior method according to patient’s cosmetic concern. Spinal cord function was monitored by means of somatosensory/motorevoked potentials. Patients with neuromuscular deformity were operated with the same technique, but the pelvic fixation was performed using intrasacral rods, as described in the modified Jackson’s technique. Patients with congenital scoliosis (hemivertebrea or unilateral unsegmented bar) were treated by circumferential convexe epiphysiodesis, combining posterior instrumented arthrodesis (2 levels above and 2 levels below) and anterior epiphysiodesis.
Ac
ce pt
Data collected Data collected and analyzed consist on: age, weight, body mass index (BMI), indication of spinal surgery (idiopathic scoliosis, neuromuscular scoliosis: neuromuscular disease or cerebral palsy, congenital scoliosis: including hemivertebrea or unilateral unsegmented bar), Cobb’s angle, ASA status, preoperative and postoperative hemoglobin concentration, techniques of blood saving (rEPO, AFT), number of levels fused, thoracoplasty, sacral fusion, intraoperative volumes of colloids and crystalloids administered and volumes of homologous and autologous (blood salvage) blood transfused during the perioperative period including: the intraoperative period and the first postoperative day. Concerning the efficiency of cell-saver, the observed rate of autologous transfusion produced by blood salvage could not be considered as an indicator of the efficacy of this technique because blood salvage was systematically transfused when available even if not indicated by the level of intraoperative hemoglobin (some patients with a hematocrit = 12 g.dl-1 might receive blood salvage). Consequently, we calculated the predictive final hematocrit at postoperative day 1 if blood salvage was not used (Htf- : appendix 1) and compared it to the target hematocrit for each patient (0.30 for patient ASA status > 2 and neuromuscular scoliosis, and 0.26 for idiopathic scoliosis and patients with an ASA status ≤ 2). When the difference between the estimated final hematocrit if blood salvage was not used (Htf-) and the target hematocrit was < 0, the cell-saver was considered to have avoid autologous 1
http://www.has-sante.fr/portail/upload/docs/application/pdf/201502/transfusion_de_globules_rouges_homologues_-_anesthesie_reanimation_chirurgie_urgence__fiche_de_synthese.pdf
4
Page 4 of 16
transfusion.
Ac
ce pt
ed
M
an
us
cr
ip t
Statistical analysis The primary outcome of this study was to determine the factors associated with the prevention of homologous transfusion (Htf- - the targeted hematocrit according to recommendations < 0). In order to determine predictors of blood salvage efficacy a classification tree analysis (CTA) was used (18-22). Basically, prediction in medical statistics rely on multivariate analysis however this methodology is not in close from the “clinical reality” while it usually gives a probability of the studied outcome. Alternatively, the tree analysis allows us to build a classification trees made by nodes computed and optimized in order to clearly classify the population according to the influence of entered parameters. Using such a statistical tool, one might expect identifying one or more subpopulation without any efficacy of blood salvage. CTA relies on the dichotomization of the population according to the predictors entered in the model. Each dichotomization (node) built two child groups. This is performed by decreasing the variability of the child groups obtained by the partition of the population. In a practical way, the predictor that will define the node is the one which will decrease the variability in a greater extend. Each child group will go further dichotomization according to the same methodology. Predictors used in the analysis were: age, BMI, indication of spinal surgery (idiopathic scoliosis versus neuromuscular scoliosis), Cobb’s angle, ASA status, preoperative hemoglobin concentration, techniques of blood saving (rEPO, AFT), number of levels fused, thoracoplasty and sacral fusion. Concerning the interpretation of the results, one can easily read the tree by focusing on the terminal nodes which constitute “groups”. Going up in the tree allows defining these groups according to the presence or absence of risk factors defined in the tree ramification. The validity of the model was studied using the percentage of the variability explained by the model (derived from the misclassification risk of the CTA model), the area under the curve of the received operator characteristics of the model (AUCROC) and by the cross validation of the model. A tenfold cross validation was used in the current study. This process divides the study population into 10 random subgroups. The first tree is based on all of the cases except those in the first sub sample; the second tree is based on all of the cases except those in the second subsample fold, and so on. For each tree, misclassification risk is estimated by applying the tree to the subsample excluded in generating it. The overall misclassification risk is the average of the risks for all of the trees. Results of CART analysis are displayed as a tree, with each node containing the predictor factor, the number of patients within the node, the number of patients with an efficacy of blood salvage and their percentage within the subpopulation and those with no efficacy of blood salvage (with also their number and percentage within the subpopulation). Results
One hundred and forty seven patients had surgical correction of scoliosis during the study period and all procedures were made during the same operative session. Description of the population is displayed in table 1. Seventy nine percent of surgeries were idiopathic scoliosis and 11.6 % were neuromuscular scoliosis. Concerning the preoperative blood saving technique used, 93.2 % of patients were given preoperative rEPO and all received intraoperative AFT. Ten patients refused preoperative rEPO (median haemoglobin level in those patients 12.6 g.dl-1 with rages from 11.3 to 13.5 g.dl-1). Concerning the perioperative red cell transfusion (table 2), 27 patients (18.3 %) were transfused during the entire study period (transfusion rate of 76.5 % in neuromuscular scoliosis versus 10.7 % in non-neuromuscular scoliosis), most of them during the intraoperative period (25 patients). Only 4 patients were transfused during PACU stay and
5
Page 5 of 16
us
cr
ip t
none after discharge to the surgical ward. Cell-Saver was prepared for all patients but blood salvage was available for only 54 patients (36.5 %). Blood salvage was found efficient in decreasing homologous blood transfusion in 17 patients (7.5 %). Inclusion of the several studied factors in the CART analysis shed the light on some of them that hierarchically influence the efficacy of blood salvage as illustrate by figure 1. According to the results, 6 nodes were formed with 4 terminal nodes (nodes 1,4,5 and 6). Parameters that found to categorize patients were: the indication of scoliosis correction (neuromuscular versus non-neuromuscular), number of level fused (< 13 versus ≥ 13) and BMI (< 21 versus ≥ 21)(table 3 and figure 1). Interactions of these factors on the outcome of this study are summarized in table 3. According to the misclassification risk of the model, it allows to correctly classify 13 patients of the 17 cases (76.5 %) with an efficacy of blood salvage and 126 on the overall 130 cases (97 %) where blood salvage was ineffective. The area under the curve of the model was 0.93 [95 % confidence interval 0.9 to 0.99]. The cross-validation of the CTA model found 60.1 % of the variability of the data explained by the derived model and 59.2 % by the cross-validation. Finally, the model could identify a subpopulation of patients (49.7 % of the overall population), in which blood salvage was ineffective in decreasing blood transfusion (node 4 and 6).
Ac
ce pt
ed
M
an
Discussion In the present study we found that in a population of pediatric patients undergoing scoliosis surgery with preoperative blood saving strategies including perioperative rEPO and intraoperative AFT, blood salvage is estimated to be totally ineffective in preventing homologous blood transfusion in 50 % of cases (74 patients). We took as independent predictors the efficacy of blood salvage: neuromuscular indications, number of level fused and BMI. The model exhibited high accuracy with 76.5 % and 97 % of accurate prediction for the efficacy and inefficacy of blood salvage, respectively. The AUCROC of the model was0.93 [95 % confidence interval 0.9 to 0.99] and the ten-fold cross-validation of the model found close result than the overall validation (59.2 % versus 60.1 % of explained variability). Surgical indication for spine surgery in our sample concerned 65.5% of idiopathic scoliosis, and overall transfusion rate was 18.4 % which is similar to other previous publications using preoperative rEPO (9). Neuromuscular scoliosis was a strong predictor of the efficacy of blood salvage in reducing homologous blood transfusion. This result is not surprising given the strong association previously found in the same cohort, and in others, between neuromuscular scoliosis and homologous blood transfusion (2,9,23-26). As extensively discussed elsewhere, this might be, at least in part, related to the decrease of bone mineralization and osteoporosis, previously described in this population (27-31). In addition, the increased target of hematocrit in neuromuscular patients might have also favored the efficacy of blood salvage in this particular population. The number of level fused was also found as a predictor of the efficacy of blood salvage. This result was predictable given its association with blood losses and the need for homologous transfusion, even in patients prepared with rEPO, as recently published (13). Finally, Body Mass Index was also found as a modest predictor of blood salvage efficacy. This could be explained, as previously described, by the association between a decreased BMI and the decrease of bone mineralization, which might have participated to the increase of blood loss during the perioperative period (28-31). Our results corroborate previous studies aiming to identify factors associated with an efficacy of blood salvage during scoliosis and spine surgeries. These studies have also found this technique to be effective when facing surgeries with important blood losses (more than 500 to 1000 ml)(32-35). However, in these studies there was no use of preoperative rEPO and intraoperative AFT (32-35). Recombinant Erythropoietin increases preoperative hemoglobin
6
Page 6 of 16
Ac
ce pt
ed
M
an
us
cr
ip t
and allow delaying transfusion even when facing important perioperative blood loss. This is supported by the results showing that 36.5 % of patients in whom cell-saver produced autologous blood in spite of its low efficacy in decreasing blood transfusion (7.5 % of cases). In addition, all patients received intraoperative AFT, which has probably contributed to the decreased of blood losses and limited the possibility of obtaining autologous blood. Regarding the key question of using or not blood salvage during pediatric scoliosis correction, current results suggest that physicians could dispense with this blood saving strategy when either (a) patients have no neuromuscular diseases and have surgeries interesting < 13 levels fused; or (b) have no neuromuscular diseases, have surgeries interesting > 13 levels with a preoperative BMI ≥ 21. In all other cases, blood salvage was estimated avoiding homologous blood transfusion in 7.9 % to 75.5 % of cases and is recommended in those patients. Interestingly, BMI, found as a predictor of blood salvage efficacy (nodes 5 and 6), was the sole factor that can be improved preoperatively. Although further researches are needed to assess the role of this factor, one can expect decreasing the need of blood salvage when improving patient’s nutritional status, especially when expected level fused exceed 13 (nodes 5and 6). Our study suffers one major limitation, because it is a simulation study thus not based on direct observed effects of cell-saver on homologous transfusion. However, the restrictions in computing cases with an efficacy of blood salvage were very demanding given that a strict respect of recommendations was observed. This is not the case in practice while patients with hemoglobin values close to the recommended values are not transfused (for example a 7.5 g/dl of hemoglobin instead of 8 g/dl recommended in ASA I and II patients). Moreover, this indirect methodology can be considered as a valuable alternative to retrospective studies usually performed to assess the efficacy of blood salvage during scoliosis surgery in children that suffer from major selection biases (patients with an expected more blood loss are almost indicated for the use of cell-saver). A randomized controlled study is the most adequate method to respond to this question as performed in adult patient during hip and knee surgeries (12). However, given the low incidence of blood transfusion achieved using current bloodsaving strategies (rEPO and AFT); this might necessities a large sample size (800 patients per study arm would be necessary to found a 5 % decrease in homologous blood transfusion with an alpha risk of 5 % and a power of 80 %).The internal validity of the derived model seems very high given the AUC and the ten-fold validation highlighted by our analyses. However, given the specificity of our blood saving protocol (rEPO and AFT), our results might not be generalized in case these techniques are not used. In conclusion, our study found that neuromuscular scoliosis, number of level fused and preoperative BMI were predictors of blood salvage efficacy during pediatric scoliosis surgery when concomitantly used with preoperative rEPO and intraoperative AFT. Our study allows us to determinate a subpopulation of patients with no additional risk of homologous blood transfusion when intraoperative blood salvage was not used. However, more studies are needed to assess the external validity of these results. Funding: Support was provided solely from institutional and departmental sources. Authors declare no conflict of interest.
7
Page 7 of 16
Author contribution
Ac
ce pt
ed
M
an
us
cr
ip t
D. Michelet: conceptualized and designed the study, collected data, corrected the manuscript and approved the final manuscript as submitted. F. Julien-Marsollier: verified statistics, corrected the manuscript and approved the final manuscript as submitted. J. Hilly: verified statistics, corrected the manuscript and approved the final manuscript as submitted. T. Diallo: collected data, corrected the manuscript and approved the final manuscript as submitted. C. Vidal: verified statistics, corrected the manuscript and approved the final manuscript as submitted. S. Dahmani : conceptualized and designed the study, designed the data collection instruments, carried out the initial analyses and verified statistics, drafted the initial manuscript, corrected the manuscript and approved the final manuscript as submitted.
8
Page 8 of 16
Legend of tables and figures Figure 1: diagram of the classification and tree analysis. Each node contain the total number of patients, the studied factor and the proportion of patients with an efficacy (BC+) or inefficacy (BC-) of intraoperative blood salvage. Table 1: Description of the population. BMI : body mass index, rEPO : recombinant erythropoietin, AFT : Antifibrinolytic therapy.
us
cr
ip t
Table 2: Perioperative transfusion and fluid intakes. NR : not reported. EBV : estimated blood volume, ERCV: estimated red cell volume, ERCT: estimated red cell transfused, ERCL: estimated red cell lost and ERCT: estimated red cell deficit. iHt : preoperative hematocrit level, fHt : hematocrit at postoperative day 1. Hematocrit of homologous blood was set to 0.55 according to our blood banking. Hematocrit of autologous blood produced by blood salvage was set to 0.55 according to the manufacturer. Table 3: classification table of estimated efficacy of intraoperative blood salvage. Nodes refer to the terminal nodes displayed in figure 1. BMI : Body mass index (kg/m²).
Ac
ce pt
ed
M
an
Appendix 1: Derived parameters when using cell-saver (+) and when cee-saver was not used. EBV : estimated blood volume, ERCV: estimated red cell volume, ERCT: estimated red cell transfused, ERCL: estimated red cell lost and ERCT: estimated red cell deficit. Hti : preoperative hematocrit level, Htf : hematocrit at postoperative day 1. Hematocrit of homologous blood was set to 0.55 according to our blood banking. Hematocrit of autologous blood produced by blood salvage was set to 0.55 according to the manufacturer.
9
Page 9 of 16
References
2.
3.
10.
11.
12.
13.
14.
15.
an
M
9.
ed
8.
ce pt
6. 7.
Ac
5.
us
cr
4.
Wazeka AN, DiMaio MF, Boachie-Adjei O. Outcome of pediatric patients with severe restrictive lung disease following reconstructive spine surgery. Spine (Phila Pa 1976) 2004;29:528-34; discussion 35. Barsdorf AI, Sproule DM, Kaufmann P. Scoliosis surgery in children with neuromuscular disease: findings from the US National Inpatient Sample, 1997 to 2003. Arch Neurol 2010;67:231-5. Benson ER, Thomson JD, Smith BG, Banta JV. Results and morbidity in a consecutive series of patients undergoing spinal fusion for neuromuscular scoliosis. Spine (Phila Pa 1976) 1998;23:2308-17; discussion 18. Maesani M, Doit C, Lorrot M, Vitoux C, Hilly J, Michelet D, Vidal C, JulienMarsollier F, Ilharreborde B, Mazda K, Bonacorsi S, Dahmani S. Surgical Site Infections in Pediatric Spine Surgery: Comparative Microbiology of Patients with Idiopathic and Nonidiopathic Etiologies of Spine Deformity. Pediatr Infect Dis J 2016;35:66-70. Guay J, de Moerloose P, Lasne D. Minimizing perioperative blood loss and transfusions in children. Can J Anaesth 2006;53:S59-67. Eder AF. Improving safety for young blood donors. Transfus Med Rev 2012;26:14-26. Eder AF, Chambers LA. Noninfectious complications of blood transfusion. Arch Pathol Lab Med 2007;131:708-18. Tzortzopoulou A, Cepeda MS, Schumann R, Carr DB. Antifibrinolytic agents for reducing blood loss in scoliosis surgery in children. Cochrane Database Syst Rev 2008:CD006883. Vitale MG, Privitera DM, Matsumoto H, Gomez JA, Waters LM, Hyman JE, Roye DP, Jr. Efficacy of preoperative erythropoietin administration in pediatric neuromuscular scoliosis patients. Spine (Phila Pa 1976) 2007;32:2662-7. Vitale MG, Roye BD, Ruchelsman DE, Roye DP, Jr. Preoperative use of recombinant human erythropoietin in pediatric orthopedics: a decision model for long-term outcomes. Spine J 2007;7:292-300. Vitale MG, Stazzone EJ, Gelijns AC, Moskowitz AJ, Roye DP, Jr. The effectiveness of preoperative erythropoietin in averting allogenic blood transfusion among children undergoing scoliosis surgery. J Pediatr Orthop B 1998;7:203-9. So-Osman C, Nelissen RG, Koopman-van Gemert AW, Kluyver E, Poll RG, Onstenk R, Van Hilten JA, Jansen-Werkhoven TM, van den Hout WB, Brand R, Brand A. Patient blood management in elective total hip- and knee-replacement surgery (Part 1): a randomized controlled trial on erythropoietin and blood salvage as transfusion alternatives using a restrictive transfusion policy in erythropoietin-eligible patients. Anesthesiology 2014;120:839-51. Dupuis C, Michelet D, Hilly J, Diallo T, Vidal C, Delivet H, Nivoche Y, Mazda K, Dahmani S. Predictive factors for homologous transfusion during paediatric scoliosis surgery. Anaesthesia Critical Care & Pain Medicine 2015;34:327-32. Ilharreborde B, Even J, Lefevre Y, Fitoussi F, Presedo A, Pennecot GF, Mazda K. Hybrid constructs for tridimensional correction of the thoracic spine in adolescent idiopathic scoliosis: a comparative analysis of universal clamps versus hooks. Spine (Phila Pa 1976) 2010;35:306-14. Ilharreborde B, Even J, Lefevre Y, Fitoussi F, Presedo A, Souchet P, Pennecot GF, Mazda K. How to determine the upper level of instrumentation in Lenke types 1 and 2 adolescent idiopathic scoliosis: a prospective study of 132 patients. J Pediatr Orthop 2008;28:733-9.
ip t
1.
10
Page 10 of 16
22. 23.
24.
25.
26.
27.
28.
29. 30.
31.
32.
ip t
cr
us
21.
an
20.
M
19.
ed
18.
ce pt
17.
Ilharreborde B, Hoffmann E, Tavakoli S, Queinnec S, Fitoussi F, Presedo A, Pennecot GF, Mazda K. Intrasacral rod fixation for pediatric long spinal fusion: results of a prospective study with a minimum 5-year follow-up. J Pediatr Orthop 2009;29:594601. Ilharreborde B, Morel E, Fitoussi F, Presedo A, Souchet P, Pennecot GF, Mazda K. [How to determine the upper level of instrumentation in thoracic adolescent idiopathic scoliosis: a prospective study of 103 patients]. Rev Chir Orthop Reparatrice Appar Mot 2008;94:481-9. Lemon SC, Roy J, Clark MA, Friedmann PD, Rakowski W. Classification and regression tree analysis in public health: methodological review and comparison with logistic regression. Ann Behav Med 2003;26:172-81. Philips Z, Ginnelly L, Sculpher M, Claxton K, Golder S, Riemsma R, Woolacoot N, Glanville J. Review of guidelines for good practice in decision-analytic modelling in health technology assessment. Health Technol Assess 2004;8:iii-iv, ix-xi, 1-158. Loh WY. Classification and regression trees. Wiley Interdisciplinary Reviews: Data Mining and Knowledge Discovery 2011;1:14-23. Marshall RJ. The use of classification and regression trees in clinical epidemiology. Journal of Clinical Epidemiology 2001;54:603-9. Siciliano R, Mola F. Multivariate data analysis and modeling through classification and regression trees. Computational Statistics & Data Analysis 2000;32:285-301. Lenoir B, Merckx P, Paugam-Burtz C, Dauzac C, Agostini MM, Guigui P, Mantz J. Individual probability of allogeneic erythrocyte transfusion in elective spine surgery: the predictive model of transfusion in spine surgery. Anesthesiology 2009;110:105060. Powell ETt, Krengel WF, 3rd, King HA, Lagrone MO. Comparison of same-day sequential anterior and posterior spinal fusion with delayed two-stage anterior and posterior spinal fusion. Spine (Phila Pa 1976) 1994;19:1256-9. Ozturk C, Aydinli U, Vural R, Sehirlioglu A, Mutlu M. Simultaneous versus sequential one-stage combined anterior and posterior spinal surgery for spinal infections (outcomes and complications). Int Orthop 2007;31:363-6. Yang HL, Chen KW, Wang GL, Lu J, Ji YM, Liu JY, Wu GZ, Gu Y, Sun ZY. Preoperative transarterial embolization for treatment of primary sacral tumors. J Clin Neurosci 2010;17:1280-5. Zheng F, Cammisa FP, Jr., Sandhu HS, Girardi FP, Khan SN. Factors predicting hospital stay, operative time, blood loss, and transfusion in patients undergoing revision posterior lumbar spine decompression, fusion, and segmental instrumentation. Spine (Phila Pa 1976) 2002;27:818-24. Bui T, Shapiro F. Posterior spinal fusion to sacrum in non-ambulatory hypotonic neuromuscular patients: sacral rod/bone graft onlay method. J Child Orthop 2014;8:229-36. Hasler CC. Operative treatment for spinal deformities in cerebral palsy. J Child Orthop 2013;7:419-23. Jeon YK, Shin MJ, Shin YB, Kim CR, Kim SJ, Ko HY, Kim IJ. Effect of increased axial rotation angle on bone mineral density measurements of the lumbar spine. Spine J 2014. Joyce NC, Hache LP, Clemens PR. Bone health and associated metabolic complications in neuromuscular diseases. Phys Med Rehabil Clin N Am 2012;23:77399. Carey PA, Schoenfeld AJ, Cordill RD, Tompkins BJ, Caskey PM. A comparison of cell salvage strategies in posterior spinal fusion for adolescent idiopathic scoliosis. J
Ac
16.
11
Page 11 of 16
ip t
cr us an M ed
35.
ce pt
34.
Spinal Disord Tech 2015;28:1-4. Liang J, Shen J, Chua S, Fan Y, Zhai J, Feng B, Cai S, Li Z, Xue X. Does intraoperative cell salvage system effectively decrease the need for allogeneic transfusions in scoliotic patients undergoing posterior spinal fusion? A prospective randomized study. Eur Spine J 2015;24:270-5. Miao YL, Ma HS, Guo WZ, Wu JG, Liu Y, Shi WZ, Wang XP, Mi WD, Fang WW. The efficacy and cost-effectiveness of cell saver use in instrumented posterior correction and fusion surgery for scoliosis in school-aged children and adolescents. PLoS One 2014;9:e92997. Chanda A, Smith DR, Nanda A. Autotransfusion by cell saver technique in surgery of lumbar and thoracic spinal fusion with instrumentation. J Neurosurg 2002;96:298-303.
Ac
33.
12
Page 12 of 16
Table 1: Description of the population. BMI : body mass index, rEPO : recombinant erythropoietin, AFT : Antifibrinolytic therapy.
M
an
us
cr
ip t
Results 15 ± 3 50 ± 14 19 ± 4 19 (12.9 %) 58 ± 16 116 (78.9 %) 8 (5.4 %) 9 (6.1 %) 17 (11.6 %) 14 (9.5 %) 17 (11.6 %) / 37 (27.2 %) / 87 (63.3 %) 137 (93.2 %) 147 (100 %) 13 ± 3 11 (7.5 %) 39 (26.5 %) 227 ± 65
Ac
ce pt
ed
Factors Age (years) Weight(Kg) BMI (Kg/m²) BMI < 3d percentile Cobb’s Angle (degrees) Idiopathic scoliosis N (%) Myopathy N (%) Cerebral Palasy scoliosis N (%) Neuromuscular scoliosis N (%) Congenital scoliosis N (%) ASA 1/2/3 N (%) rEPO N (%) AFT N (%) Level fused (N) Sacral Fusion N (%) Gibbectomy N (%) Duration of Surgery (minutes)
13
Page 13 of 16
Table 2 : Perioperative transfusion and fluid intakes. NR : not reported. EBV : estimated blood volume, ERCV: estimated red cell volume, ERCT: estimated red cell transfused, ERCL: estimated red cell lost and ERCT: estimated red cell deficit. iHt : preoperative hematocrit level, fHt : hematocrit at postoperative day 1. Hematocrit of homologous blood
Postoperative (Day 1) 10.8 ± 1.3 NR 0±0 17 ± 160 4 (2.7 %) 95 ± 265 27 (18.4 %) 3753 ±1058 1663 ±522 123 ±235 438 ±221 562 ±321
M
an
us
Intraoperative 14.6 ± 1.3 17 ± 10 10 ± 10 129 ± 240 ml 54 (36.5 %) 77 ± 213 25 (17 %) NR NR NR NR NR
Ac
ce pt
ed
Factors Hb (g/dl) Crystalloids (ml/Kg/h) Colloids (ml/Kg) Cell Saver (ml) Cell Saver (%) Homologous (ml) Homologous N (%) Total Homologous Transfusion (ml) Total Homologous Transfusion (%) EBV(ml) ERCV(ml) ERCT(ml) ERCL(ml) ERCD(ml)
cr
blood salvage was set to 0.55 according to the manufacturer.
ip t
was set to 0.55 according to our blood banking. Hematocrit of autologous blood produced by
14
Page 14 of 16
Table 3: classification table of estimated efficacy of intraoperative blood salvage. Nodes refer to the terminal nodes displayed in figure 1. BMI : Body mass index (kg/m²).
< 21 ≥ 21
ip t
≥ 13 ≥ 13 < 13
Predictive probability of Blood salvage efficacy 75.5 % 7.9 % 0% 0%
cr
BMI
ce pt
ed
M
an
us
Level fused
Ac
Node 1 Node 5 Node 6 Node 4
Neuromuscular Scoliosis Yes No No No
15
Page 15 of 16
16
Page 16 of 16
ed
ce pt
Ac
us
an
M
cr
ip t
S2352-5568(16)30227-2 http://dx.doi.org/doi:10.1016/j.accpm.2017.03.003 ACCPM 252
To appear in: Received date: Accepted date:
29-11-2016 17-3-2017
Please cite this article as: Daphn´e MicheletFlorence Julien-MarsollierJulie HillyThierno DialloChristophe VidalSouhayl Dahmani Predictive factors of intraoperative Cell Salvage during Pediatric Scoliosis Surgery (2017), http://dx.doi.org/10.1016/j.accpm.2017.03.003 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Predictive factors of intraoperative Cell Salvage during Pediatric Scoliosis Surgery. Cell Saver during scoliosis surgery in children
Cell Saver during scoliosis surgery in children # ♥,
Julie Hilly *
#♥
, Thierno Diallo *
#♥
,
cr
Daphné Michelet * # ♥, Florence Julien-Marsollier * Christophe Vidal ** # ♥, and SouhaylDahmani * #♥.
ip t
Predictive factors of intraoperative Cell Salvage during Pediatric Scoliosis Surgery.
♥DHU
an
us
* Department of anesthesia and Intensive care Robert Debré University Hospital. Paris, France. ** Department of pediatric orthopedic surgery, Robert Debré University Hospital. Paris, France. # Paris Diderot University (Paris VII).PRES Paris Sorbonne Cité. Paris. France PROTECT. INSERM U1141.Robert Debré University Hospital. Paris. France.
Ac
ce pt
ed
M
Corresponding author: Professor Souhayl Dahmani. Department of anesthesiology and Intensive Care. Robert Debré Hospital, 48 Boulevard Sérurier, 75019 Paris, France Phone: 00 33 1 40 03 41 83 Fax : 00 33 1 40 03 20 00 e-mail: [email protected]
1
Page 1 of 16
Ac
ce pt
ed
M
an
us
cr
ip t
Abstract Introduction: blood saving strategy during spinal surgery in children often includes recombinant erythropoietin (rEPO) and antifibrinolytic therapapy (AFT). The aim of this study was to investigate the efficacy of intraoperative blood salvage in decreasing homologous blood transfusion. Material and Methods: using the prospective data from patients operated during a one year period for scoliosis correction, we calculate the predictable hematocrit at day postoperative 1 without the use of blood salvage and compare it to the target hematocrit transfusion according to patient’s status. Predictors analyzed were: age, weight, surgical indication, Cobb’s angle, ASA status, preoperative hemoglobin, number of level fused, sacral fusion and thoracoplasty. Statistical analyses were performed using a classification tree analysis. Results: this study included 147 patients. Blood salvage was estimated avoiding homologous blood transfusion in 17 patients. Predictors of the efficacy of blood salvage were: neuromuscular indications, number of level fused and BMI. Blood salvage was found totally ineffective in: patients with no neuromuscular diseases with either: surgeries interesting < 13 levels fused or surgeries interesting > 13 levels with a preoperative BMI ≥ 21. In all other cases, blood salvage can decrease homologous transfusion. The model exhibited 97 % of accurate for the prediction if the inefficacy of blood salvage. The AUCROC of the model was 0.93 [95 % confidence interval 0.9 to 0.99] and the overall validation was 60.1 % of explained variability. Conclusion: The present study indicates that blood salvage is ineffective under certain circumstances. More studies are mandatory to confirm these results.
2
Page 2 of 16
Introduction
an
us
cr
ip t
Spinal surgery in children is a high risk procedure associated with intraoperative and postoperative complications (1). Ten percent of patients have neuromuscular diseases or cerebral palsy, thus increasing postoperative cardiovascular, respiratory and infectious complications (1-4). Spinal surgery is typically associated with extensive blood loss and perioperative transfusion (5). Moreover transfusion therapy has been shown to induce many specific complications such as bacterial infections, acute immunological incompatibility reactions and long term immunological effects (infections, tumors recurrence)(6,7). During the last decade, the use of blood saving strategies such as recombinant Erythropoietin (rEPO) and Antifibrinolytic therapy (AFT), resulted in a reduction of perioperative autologous blood transfusion (8-11). Consequently, the use of autologous blood donation has been completely withdrawal from current practice and blood salvage during the intraoperative period is contested given the few number of transfused patients. For example, a recent adult trial showed the ineffective effect of cell-saver in reducing autologous blood transfusion during major orthopedic surgeries while increasing costs(12). In the present study, we evaluated the efficacy of intraoperative blood salvage in preventing homologous blood transfusion in patients scheduled for scoliosis correction surgery managed with preoperative rEPO and intraoperative AFT. Material and Methods
ce pt
ed
M
The study consists on a prospective analysis of perioperative data of patients operated in our institution from January 2012 to December 2012. This study was approved by our institutional review board (Comité d’Evaluation de l’Ethique des projets de Recherche Biomédicale (CEERB) Paris Nord; # 12-021). Informed written consent was obtained from all patients (when allowed by their neurological status) and/or parents. All patients undergoing scoliosis surgery during the above quoted period were included in this study. Data used for the present manuscript were previously published for the purpose of analyzing the predictors of homologous transfusion (13).
Ac
Perioperative anesthesia and blood saving management Preoperative anesthesia and surgical preparation included standardized recombinant Erythropoietin (rEPO, Eprex®, JANSSEN-CILAG, Issy-Les-Moulineaux, France) treatment. Hemoglobin concentration was determined 24 hours before each scheduled date of injection, and a value strictly below 15 g/dl was followed by recombinant Erythropoietin injection. This consists in subcutaneous administration of 600 U/kg of rEPO weekly, started 4 weeks before the scheduled date of surgery with the last injection 48 hours before the date of surgery. Erythropoietin therapy was systematically associated with iron supplementation. Antifibrinolytic agent administration begins after the induction; it consists in tranexamic acid (Exacyl®, Sanofi-Aventis, Paris, France) given by a bolus of 50 mg/kg, plus a continuous intraoperative administration of 10 mg/kg/h. At skin closure, this drug was discontinued in order to respect its maximal dose recommended in France (4 g per day). Intraoperative blood salvage (Haemonetics Cell Saver 5, Haemonetics Corporation, Braintree, MA, USA) was used in all patients with an automatic salvage program in order to ensure blood transfusion with optimal hematocrit level (55 %). The cell-saver system was primed with 100 ml of normal saline solution and salvage process was initiated when at least 1000 ml of blood loss was available in the collector boil. According to our local protocols anesthesia was standardized. Transfusion protocol follows
3
Page 3 of 16
ip t
our national recommendations1. Intraoperative and postoperative hemoglobin concentration were maintained above 8 g.dl-1 (hematocrit = 0.24) for idiopathic or congenital scoliosis. This targeted concentration was increased to 10 g.dl-1 (hematocrit = 0.30) in case of neuromuscular scoliosis or in patients with ASA status > 2. This was achieved by autologous transfusion when available (intraoperative blood salvage) and homologous transfusion if necessary. Fresh frozen plasma and platelets were administered if perioperative prothrombin time was below 50 % and platelet count was below 50 000 elements.mm-3, respectively. Finally, all patients were given intravenous iron therapy (3 mg.kg-1) twice during the postoperative period (during 48 hours).
ed
M
an
us
cr
Surgical management (14-17) Hybrid constructs that included sublaminar bands in the thoracic spine, auto stable pediculo-supralaminar claws (Instinct, Zimmer Spine, Bordeaux, France) at the two proximal levels and pedicular screws (Shiraz, Zimmer spine, Bordeaux, France) at lumbar levels were placed into patients undergoing posterior correction and fusion for adolescent idiopathic scoliosis. Fusion levels were selected according to the same criteria during the entire study period. The 2 rods were adjusted to the desired sagittal profile and connected with 2 transverse connectors before placement. The rods were first introduced in the pedicle screws, and lumbar correction was performed. Thoracoplasty was performed by the posterior method according to patient’s cosmetic concern. Spinal cord function was monitored by means of somatosensory/motorevoked potentials. Patients with neuromuscular deformity were operated with the same technique, but the pelvic fixation was performed using intrasacral rods, as described in the modified Jackson’s technique. Patients with congenital scoliosis (hemivertebrea or unilateral unsegmented bar) were treated by circumferential convexe epiphysiodesis, combining posterior instrumented arthrodesis (2 levels above and 2 levels below) and anterior epiphysiodesis.
Ac
ce pt
Data collected Data collected and analyzed consist on: age, weight, body mass index (BMI), indication of spinal surgery (idiopathic scoliosis, neuromuscular scoliosis: neuromuscular disease or cerebral palsy, congenital scoliosis: including hemivertebrea or unilateral unsegmented bar), Cobb’s angle, ASA status, preoperative and postoperative hemoglobin concentration, techniques of blood saving (rEPO, AFT), number of levels fused, thoracoplasty, sacral fusion, intraoperative volumes of colloids and crystalloids administered and volumes of homologous and autologous (blood salvage) blood transfused during the perioperative period including: the intraoperative period and the first postoperative day. Concerning the efficiency of cell-saver, the observed rate of autologous transfusion produced by blood salvage could not be considered as an indicator of the efficacy of this technique because blood salvage was systematically transfused when available even if not indicated by the level of intraoperative hemoglobin (some patients with a hematocrit = 12 g.dl-1 might receive blood salvage). Consequently, we calculated the predictive final hematocrit at postoperative day 1 if blood salvage was not used (Htf- : appendix 1) and compared it to the target hematocrit for each patient (0.30 for patient ASA status > 2 and neuromuscular scoliosis, and 0.26 for idiopathic scoliosis and patients with an ASA status ≤ 2). When the difference between the estimated final hematocrit if blood salvage was not used (Htf-) and the target hematocrit was < 0, the cell-saver was considered to have avoid autologous 1
http://www.has-sante.fr/portail/upload/docs/application/pdf/201502/transfusion_de_globules_rouges_homologues_-_anesthesie_reanimation_chirurgie_urgence__fiche_de_synthese.pdf
4
Page 4 of 16
transfusion.
Ac
ce pt
ed
M
an
us
cr
ip t
Statistical analysis The primary outcome of this study was to determine the factors associated with the prevention of homologous transfusion (Htf- - the targeted hematocrit according to recommendations < 0). In order to determine predictors of blood salvage efficacy a classification tree analysis (CTA) was used (18-22). Basically, prediction in medical statistics rely on multivariate analysis however this methodology is not in close from the “clinical reality” while it usually gives a probability of the studied outcome. Alternatively, the tree analysis allows us to build a classification trees made by nodes computed and optimized in order to clearly classify the population according to the influence of entered parameters. Using such a statistical tool, one might expect identifying one or more subpopulation without any efficacy of blood salvage. CTA relies on the dichotomization of the population according to the predictors entered in the model. Each dichotomization (node) built two child groups. This is performed by decreasing the variability of the child groups obtained by the partition of the population. In a practical way, the predictor that will define the node is the one which will decrease the variability in a greater extend. Each child group will go further dichotomization according to the same methodology. Predictors used in the analysis were: age, BMI, indication of spinal surgery (idiopathic scoliosis versus neuromuscular scoliosis), Cobb’s angle, ASA status, preoperative hemoglobin concentration, techniques of blood saving (rEPO, AFT), number of levels fused, thoracoplasty and sacral fusion. Concerning the interpretation of the results, one can easily read the tree by focusing on the terminal nodes which constitute “groups”. Going up in the tree allows defining these groups according to the presence or absence of risk factors defined in the tree ramification. The validity of the model was studied using the percentage of the variability explained by the model (derived from the misclassification risk of the CTA model), the area under the curve of the received operator characteristics of the model (AUCROC) and by the cross validation of the model. A tenfold cross validation was used in the current study. This process divides the study population into 10 random subgroups. The first tree is based on all of the cases except those in the first sub sample; the second tree is based on all of the cases except those in the second subsample fold, and so on. For each tree, misclassification risk is estimated by applying the tree to the subsample excluded in generating it. The overall misclassification risk is the average of the risks for all of the trees. Results of CART analysis are displayed as a tree, with each node containing the predictor factor, the number of patients within the node, the number of patients with an efficacy of blood salvage and their percentage within the subpopulation and those with no efficacy of blood salvage (with also their number and percentage within the subpopulation). Results
One hundred and forty seven patients had surgical correction of scoliosis during the study period and all procedures were made during the same operative session. Description of the population is displayed in table 1. Seventy nine percent of surgeries were idiopathic scoliosis and 11.6 % were neuromuscular scoliosis. Concerning the preoperative blood saving technique used, 93.2 % of patients were given preoperative rEPO and all received intraoperative AFT. Ten patients refused preoperative rEPO (median haemoglobin level in those patients 12.6 g.dl-1 with rages from 11.3 to 13.5 g.dl-1). Concerning the perioperative red cell transfusion (table 2), 27 patients (18.3 %) were transfused during the entire study period (transfusion rate of 76.5 % in neuromuscular scoliosis versus 10.7 % in non-neuromuscular scoliosis), most of them during the intraoperative period (25 patients). Only 4 patients were transfused during PACU stay and
5
Page 5 of 16
us
cr
ip t
none after discharge to the surgical ward. Cell-Saver was prepared for all patients but blood salvage was available for only 54 patients (36.5 %). Blood salvage was found efficient in decreasing homologous blood transfusion in 17 patients (7.5 %). Inclusion of the several studied factors in the CART analysis shed the light on some of them that hierarchically influence the efficacy of blood salvage as illustrate by figure 1. According to the results, 6 nodes were formed with 4 terminal nodes (nodes 1,4,5 and 6). Parameters that found to categorize patients were: the indication of scoliosis correction (neuromuscular versus non-neuromuscular), number of level fused (< 13 versus ≥ 13) and BMI (< 21 versus ≥ 21)(table 3 and figure 1). Interactions of these factors on the outcome of this study are summarized in table 3. According to the misclassification risk of the model, it allows to correctly classify 13 patients of the 17 cases (76.5 %) with an efficacy of blood salvage and 126 on the overall 130 cases (97 %) where blood salvage was ineffective. The area under the curve of the model was 0.93 [95 % confidence interval 0.9 to 0.99]. The cross-validation of the CTA model found 60.1 % of the variability of the data explained by the derived model and 59.2 % by the cross-validation. Finally, the model could identify a subpopulation of patients (49.7 % of the overall population), in which blood salvage was ineffective in decreasing blood transfusion (node 4 and 6).
Ac
ce pt
ed
M
an
Discussion In the present study we found that in a population of pediatric patients undergoing scoliosis surgery with preoperative blood saving strategies including perioperative rEPO and intraoperative AFT, blood salvage is estimated to be totally ineffective in preventing homologous blood transfusion in 50 % of cases (74 patients). We took as independent predictors the efficacy of blood salvage: neuromuscular indications, number of level fused and BMI. The model exhibited high accuracy with 76.5 % and 97 % of accurate prediction for the efficacy and inefficacy of blood salvage, respectively. The AUCROC of the model was0.93 [95 % confidence interval 0.9 to 0.99] and the ten-fold cross-validation of the model found close result than the overall validation (59.2 % versus 60.1 % of explained variability). Surgical indication for spine surgery in our sample concerned 65.5% of idiopathic scoliosis, and overall transfusion rate was 18.4 % which is similar to other previous publications using preoperative rEPO (9). Neuromuscular scoliosis was a strong predictor of the efficacy of blood salvage in reducing homologous blood transfusion. This result is not surprising given the strong association previously found in the same cohort, and in others, between neuromuscular scoliosis and homologous blood transfusion (2,9,23-26). As extensively discussed elsewhere, this might be, at least in part, related to the decrease of bone mineralization and osteoporosis, previously described in this population (27-31). In addition, the increased target of hematocrit in neuromuscular patients might have also favored the efficacy of blood salvage in this particular population. The number of level fused was also found as a predictor of the efficacy of blood salvage. This result was predictable given its association with blood losses and the need for homologous transfusion, even in patients prepared with rEPO, as recently published (13). Finally, Body Mass Index was also found as a modest predictor of blood salvage efficacy. This could be explained, as previously described, by the association between a decreased BMI and the decrease of bone mineralization, which might have participated to the increase of blood loss during the perioperative period (28-31). Our results corroborate previous studies aiming to identify factors associated with an efficacy of blood salvage during scoliosis and spine surgeries. These studies have also found this technique to be effective when facing surgeries with important blood losses (more than 500 to 1000 ml)(32-35). However, in these studies there was no use of preoperative rEPO and intraoperative AFT (32-35). Recombinant Erythropoietin increases preoperative hemoglobin
6
Page 6 of 16
Ac
ce pt
ed
M
an
us
cr
ip t
and allow delaying transfusion even when facing important perioperative blood loss. This is supported by the results showing that 36.5 % of patients in whom cell-saver produced autologous blood in spite of its low efficacy in decreasing blood transfusion (7.5 % of cases). In addition, all patients received intraoperative AFT, which has probably contributed to the decreased of blood losses and limited the possibility of obtaining autologous blood. Regarding the key question of using or not blood salvage during pediatric scoliosis correction, current results suggest that physicians could dispense with this blood saving strategy when either (a) patients have no neuromuscular diseases and have surgeries interesting < 13 levels fused; or (b) have no neuromuscular diseases, have surgeries interesting > 13 levels with a preoperative BMI ≥ 21. In all other cases, blood salvage was estimated avoiding homologous blood transfusion in 7.9 % to 75.5 % of cases and is recommended in those patients. Interestingly, BMI, found as a predictor of blood salvage efficacy (nodes 5 and 6), was the sole factor that can be improved preoperatively. Although further researches are needed to assess the role of this factor, one can expect decreasing the need of blood salvage when improving patient’s nutritional status, especially when expected level fused exceed 13 (nodes 5and 6). Our study suffers one major limitation, because it is a simulation study thus not based on direct observed effects of cell-saver on homologous transfusion. However, the restrictions in computing cases with an efficacy of blood salvage were very demanding given that a strict respect of recommendations was observed. This is not the case in practice while patients with hemoglobin values close to the recommended values are not transfused (for example a 7.5 g/dl of hemoglobin instead of 8 g/dl recommended in ASA I and II patients). Moreover, this indirect methodology can be considered as a valuable alternative to retrospective studies usually performed to assess the efficacy of blood salvage during scoliosis surgery in children that suffer from major selection biases (patients with an expected more blood loss are almost indicated for the use of cell-saver). A randomized controlled study is the most adequate method to respond to this question as performed in adult patient during hip and knee surgeries (12). However, given the low incidence of blood transfusion achieved using current bloodsaving strategies (rEPO and AFT); this might necessities a large sample size (800 patients per study arm would be necessary to found a 5 % decrease in homologous blood transfusion with an alpha risk of 5 % and a power of 80 %).The internal validity of the derived model seems very high given the AUC and the ten-fold validation highlighted by our analyses. However, given the specificity of our blood saving protocol (rEPO and AFT), our results might not be generalized in case these techniques are not used. In conclusion, our study found that neuromuscular scoliosis, number of level fused and preoperative BMI were predictors of blood salvage efficacy during pediatric scoliosis surgery when concomitantly used with preoperative rEPO and intraoperative AFT. Our study allows us to determinate a subpopulation of patients with no additional risk of homologous blood transfusion when intraoperative blood salvage was not used. However, more studies are needed to assess the external validity of these results. Funding: Support was provided solely from institutional and departmental sources. Authors declare no conflict of interest.
7
Page 7 of 16
Author contribution
Ac
ce pt
ed
M
an
us
cr
ip t
D. Michelet: conceptualized and designed the study, collected data, corrected the manuscript and approved the final manuscript as submitted. F. Julien-Marsollier: verified statistics, corrected the manuscript and approved the final manuscript as submitted. J. Hilly: verified statistics, corrected the manuscript and approved the final manuscript as submitted. T. Diallo: collected data, corrected the manuscript and approved the final manuscript as submitted. C. Vidal: verified statistics, corrected the manuscript and approved the final manuscript as submitted. S. Dahmani : conceptualized and designed the study, designed the data collection instruments, carried out the initial analyses and verified statistics, drafted the initial manuscript, corrected the manuscript and approved the final manuscript as submitted.
8
Page 8 of 16
Legend of tables and figures Figure 1: diagram of the classification and tree analysis. Each node contain the total number of patients, the studied factor and the proportion of patients with an efficacy (BC+) or inefficacy (BC-) of intraoperative blood salvage. Table 1: Description of the population. BMI : body mass index, rEPO : recombinant erythropoietin, AFT : Antifibrinolytic therapy.
us
cr
ip t
Table 2: Perioperative transfusion and fluid intakes. NR : not reported. EBV : estimated blood volume, ERCV: estimated red cell volume, ERCT: estimated red cell transfused, ERCL: estimated red cell lost and ERCT: estimated red cell deficit. iHt : preoperative hematocrit level, fHt : hematocrit at postoperative day 1. Hematocrit of homologous blood was set to 0.55 according to our blood banking. Hematocrit of autologous blood produced by blood salvage was set to 0.55 according to the manufacturer. Table 3: classification table of estimated efficacy of intraoperative blood salvage. Nodes refer to the terminal nodes displayed in figure 1. BMI : Body mass index (kg/m²).
Ac
ce pt
ed
M
an
Appendix 1: Derived parameters when using cell-saver (+) and when cee-saver was not used. EBV : estimated blood volume, ERCV: estimated red cell volume, ERCT: estimated red cell transfused, ERCL: estimated red cell lost and ERCT: estimated red cell deficit. Hti : preoperative hematocrit level, Htf : hematocrit at postoperative day 1. Hematocrit of homologous blood was set to 0.55 according to our blood banking. Hematocrit of autologous blood produced by blood salvage was set to 0.55 according to the manufacturer.
9
Page 9 of 16
References
2.
3.
10.
11.
12.
13.
14.
15.
an
M
9.
ed
8.
ce pt
6. 7.
Ac
5.
us
cr
4.
Wazeka AN, DiMaio MF, Boachie-Adjei O. Outcome of pediatric patients with severe restrictive lung disease following reconstructive spine surgery. Spine (Phila Pa 1976) 2004;29:528-34; discussion 35. Barsdorf AI, Sproule DM, Kaufmann P. Scoliosis surgery in children with neuromuscular disease: findings from the US National Inpatient Sample, 1997 to 2003. Arch Neurol 2010;67:231-5. Benson ER, Thomson JD, Smith BG, Banta JV. Results and morbidity in a consecutive series of patients undergoing spinal fusion for neuromuscular scoliosis. Spine (Phila Pa 1976) 1998;23:2308-17; discussion 18. Maesani M, Doit C, Lorrot M, Vitoux C, Hilly J, Michelet D, Vidal C, JulienMarsollier F, Ilharreborde B, Mazda K, Bonacorsi S, Dahmani S. Surgical Site Infections in Pediatric Spine Surgery: Comparative Microbiology of Patients with Idiopathic and Nonidiopathic Etiologies of Spine Deformity. Pediatr Infect Dis J 2016;35:66-70. Guay J, de Moerloose P, Lasne D. Minimizing perioperative blood loss and transfusions in children. Can J Anaesth 2006;53:S59-67. Eder AF. Improving safety for young blood donors. Transfus Med Rev 2012;26:14-26. Eder AF, Chambers LA. Noninfectious complications of blood transfusion. Arch Pathol Lab Med 2007;131:708-18. Tzortzopoulou A, Cepeda MS, Schumann R, Carr DB. Antifibrinolytic agents for reducing blood loss in scoliosis surgery in children. Cochrane Database Syst Rev 2008:CD006883. Vitale MG, Privitera DM, Matsumoto H, Gomez JA, Waters LM, Hyman JE, Roye DP, Jr. Efficacy of preoperative erythropoietin administration in pediatric neuromuscular scoliosis patients. Spine (Phila Pa 1976) 2007;32:2662-7. Vitale MG, Roye BD, Ruchelsman DE, Roye DP, Jr. Preoperative use of recombinant human erythropoietin in pediatric orthopedics: a decision model for long-term outcomes. Spine J 2007;7:292-300. Vitale MG, Stazzone EJ, Gelijns AC, Moskowitz AJ, Roye DP, Jr. The effectiveness of preoperative erythropoietin in averting allogenic blood transfusion among children undergoing scoliosis surgery. J Pediatr Orthop B 1998;7:203-9. So-Osman C, Nelissen RG, Koopman-van Gemert AW, Kluyver E, Poll RG, Onstenk R, Van Hilten JA, Jansen-Werkhoven TM, van den Hout WB, Brand R, Brand A. Patient blood management in elective total hip- and knee-replacement surgery (Part 1): a randomized controlled trial on erythropoietin and blood salvage as transfusion alternatives using a restrictive transfusion policy in erythropoietin-eligible patients. Anesthesiology 2014;120:839-51. Dupuis C, Michelet D, Hilly J, Diallo T, Vidal C, Delivet H, Nivoche Y, Mazda K, Dahmani S. Predictive factors for homologous transfusion during paediatric scoliosis surgery. Anaesthesia Critical Care & Pain Medicine 2015;34:327-32. Ilharreborde B, Even J, Lefevre Y, Fitoussi F, Presedo A, Pennecot GF, Mazda K. Hybrid constructs for tridimensional correction of the thoracic spine in adolescent idiopathic scoliosis: a comparative analysis of universal clamps versus hooks. Spine (Phila Pa 1976) 2010;35:306-14. Ilharreborde B, Even J, Lefevre Y, Fitoussi F, Presedo A, Souchet P, Pennecot GF, Mazda K. How to determine the upper level of instrumentation in Lenke types 1 and 2 adolescent idiopathic scoliosis: a prospective study of 132 patients. J Pediatr Orthop 2008;28:733-9.
ip t
1.
10
Page 10 of 16
22. 23.
24.
25.
26.
27.
28.
29. 30.
31.
32.
ip t
cr
us
21.
an
20.
M
19.
ed
18.
ce pt
17.
Ilharreborde B, Hoffmann E, Tavakoli S, Queinnec S, Fitoussi F, Presedo A, Pennecot GF, Mazda K. Intrasacral rod fixation for pediatric long spinal fusion: results of a prospective study with a minimum 5-year follow-up. J Pediatr Orthop 2009;29:594601. Ilharreborde B, Morel E, Fitoussi F, Presedo A, Souchet P, Pennecot GF, Mazda K. [How to determine the upper level of instrumentation in thoracic adolescent idiopathic scoliosis: a prospective study of 103 patients]. Rev Chir Orthop Reparatrice Appar Mot 2008;94:481-9. Lemon SC, Roy J, Clark MA, Friedmann PD, Rakowski W. Classification and regression tree analysis in public health: methodological review and comparison with logistic regression. Ann Behav Med 2003;26:172-81. Philips Z, Ginnelly L, Sculpher M, Claxton K, Golder S, Riemsma R, Woolacoot N, Glanville J. Review of guidelines for good practice in decision-analytic modelling in health technology assessment. Health Technol Assess 2004;8:iii-iv, ix-xi, 1-158. Loh WY. Classification and regression trees. Wiley Interdisciplinary Reviews: Data Mining and Knowledge Discovery 2011;1:14-23. Marshall RJ. The use of classification and regression trees in clinical epidemiology. Journal of Clinical Epidemiology 2001;54:603-9. Siciliano R, Mola F. Multivariate data analysis and modeling through classification and regression trees. Computational Statistics & Data Analysis 2000;32:285-301. Lenoir B, Merckx P, Paugam-Burtz C, Dauzac C, Agostini MM, Guigui P, Mantz J. Individual probability of allogeneic erythrocyte transfusion in elective spine surgery: the predictive model of transfusion in spine surgery. Anesthesiology 2009;110:105060. Powell ETt, Krengel WF, 3rd, King HA, Lagrone MO. Comparison of same-day sequential anterior and posterior spinal fusion with delayed two-stage anterior and posterior spinal fusion. Spine (Phila Pa 1976) 1994;19:1256-9. Ozturk C, Aydinli U, Vural R, Sehirlioglu A, Mutlu M. Simultaneous versus sequential one-stage combined anterior and posterior spinal surgery for spinal infections (outcomes and complications). Int Orthop 2007;31:363-6. Yang HL, Chen KW, Wang GL, Lu J, Ji YM, Liu JY, Wu GZ, Gu Y, Sun ZY. Preoperative transarterial embolization for treatment of primary sacral tumors. J Clin Neurosci 2010;17:1280-5. Zheng F, Cammisa FP, Jr., Sandhu HS, Girardi FP, Khan SN. Factors predicting hospital stay, operative time, blood loss, and transfusion in patients undergoing revision posterior lumbar spine decompression, fusion, and segmental instrumentation. Spine (Phila Pa 1976) 2002;27:818-24. Bui T, Shapiro F. Posterior spinal fusion to sacrum in non-ambulatory hypotonic neuromuscular patients: sacral rod/bone graft onlay method. J Child Orthop 2014;8:229-36. Hasler CC. Operative treatment for spinal deformities in cerebral palsy. J Child Orthop 2013;7:419-23. Jeon YK, Shin MJ, Shin YB, Kim CR, Kim SJ, Ko HY, Kim IJ. Effect of increased axial rotation angle on bone mineral density measurements of the lumbar spine. Spine J 2014. Joyce NC, Hache LP, Clemens PR. Bone health and associated metabolic complications in neuromuscular diseases. Phys Med Rehabil Clin N Am 2012;23:77399. Carey PA, Schoenfeld AJ, Cordill RD, Tompkins BJ, Caskey PM. A comparison of cell salvage strategies in posterior spinal fusion for adolescent idiopathic scoliosis. J
Ac
16.
11
Page 11 of 16
ip t
cr us an M ed
35.
ce pt
34.
Spinal Disord Tech 2015;28:1-4. Liang J, Shen J, Chua S, Fan Y, Zhai J, Feng B, Cai S, Li Z, Xue X. Does intraoperative cell salvage system effectively decrease the need for allogeneic transfusions in scoliotic patients undergoing posterior spinal fusion? A prospective randomized study. Eur Spine J 2015;24:270-5. Miao YL, Ma HS, Guo WZ, Wu JG, Liu Y, Shi WZ, Wang XP, Mi WD, Fang WW. The efficacy and cost-effectiveness of cell saver use in instrumented posterior correction and fusion surgery for scoliosis in school-aged children and adolescents. PLoS One 2014;9:e92997. Chanda A, Smith DR, Nanda A. Autotransfusion by cell saver technique in surgery of lumbar and thoracic spinal fusion with instrumentation. J Neurosurg 2002;96:298-303.
Ac
33.
12
Page 12 of 16
Table 1: Description of the population. BMI : body mass index, rEPO : recombinant erythropoietin, AFT : Antifibrinolytic therapy.
M
an
us
cr
ip t
Results 15 ± 3 50 ± 14 19 ± 4 19 (12.9 %) 58 ± 16 116 (78.9 %) 8 (5.4 %) 9 (6.1 %) 17 (11.6 %) 14 (9.5 %) 17 (11.6 %) / 37 (27.2 %) / 87 (63.3 %) 137 (93.2 %) 147 (100 %) 13 ± 3 11 (7.5 %) 39 (26.5 %) 227 ± 65
Ac
ce pt
ed
Factors Age (years) Weight(Kg) BMI (Kg/m²) BMI < 3d percentile Cobb’s Angle (degrees) Idiopathic scoliosis N (%) Myopathy N (%) Cerebral Palasy scoliosis N (%) Neuromuscular scoliosis N (%) Congenital scoliosis N (%) ASA 1/2/3 N (%) rEPO N (%) AFT N (%) Level fused (N) Sacral Fusion N (%) Gibbectomy N (%) Duration of Surgery (minutes)
13
Page 13 of 16
Table 2 : Perioperative transfusion and fluid intakes. NR : not reported. EBV : estimated blood volume, ERCV: estimated red cell volume, ERCT: estimated red cell transfused, ERCL: estimated red cell lost and ERCT: estimated red cell deficit. iHt : preoperative hematocrit level, fHt : hematocrit at postoperative day 1. Hematocrit of homologous blood
Postoperative (Day 1) 10.8 ± 1.3 NR 0±0 17 ± 160 4 (2.7 %) 95 ± 265 27 (18.4 %) 3753 ±1058 1663 ±522 123 ±235 438 ±221 562 ±321
M
an
us
Intraoperative 14.6 ± 1.3 17 ± 10 10 ± 10 129 ± 240 ml 54 (36.5 %) 77 ± 213 25 (17 %) NR NR NR NR NR
Ac
ce pt
ed
Factors Hb (g/dl) Crystalloids (ml/Kg/h) Colloids (ml/Kg) Cell Saver (ml) Cell Saver (%) Homologous (ml) Homologous N (%) Total Homologous Transfusion (ml) Total Homologous Transfusion (%) EBV(ml) ERCV(ml) ERCT(ml) ERCL(ml) ERCD(ml)
cr
blood salvage was set to 0.55 according to the manufacturer.
ip t
was set to 0.55 according to our blood banking. Hematocrit of autologous blood produced by
14
Page 14 of 16
Table 3: classification table of estimated efficacy of intraoperative blood salvage. Nodes refer to the terminal nodes displayed in figure 1. BMI : Body mass index (kg/m²).
< 21 ≥ 21
ip t
≥ 13 ≥ 13 < 13
Predictive probability of Blood salvage efficacy 75.5 % 7.9 % 0% 0%
cr
BMI
ce pt
ed
M
an
us
Level fused
Ac
Node 1 Node 5 Node 6 Node 4
Neuromuscular Scoliosis Yes No No No
15
Page 15 of 16
16
Page 16 of 16
ed
ce pt
Ac
us
an
M
cr
ip t