During elective surgery, the decision to transfuse a patient to correct anemia or an unstable hemodynamic state is often determined on a case by case basis according to personal and hospital transfusion practice. There is wide variability in these practices. Literature suggests that a restrictive transfusion practice is equivalent, if not better, than a liberal practice in terms of morbidity and mortality. However, these data may not be generalizable to specific head, neck, and spine surgeries.

At the head and neck department of Montpellier University Hospital, the hemorrhagic risk has not been precisely measured for certain surgical procedures. Thus hemorrhagic risk tends to be overestimated and unnecessary blood tests tend to be ordered.

In our single center, retrospective, observational study the investigators will determine the hemorrhagic risk related to each elective surgical procedure performed in our center. The investigators will also analyze the transfusion triggers used, and will compare them to national and international recommendations. The results of this study will inform transfusion and laboratory practice for patients undergoing head, neck, or spine surgery.

Our single center, retrospective, observational study will analyze the transfusion practices at the Montpellier University Hospital during the perioperative period (defined as 24 hours before surgery to 48 hours after surgery).

Aim of the study The aim of the study is to identify the transfusion rate for each surgical procedure (transfused patients / total number of procedures performed).

The study will also allow us:

• to identify the hemorrhagic risk related to each surgical procedure performed;
• to define the triggers for and factors determining transfusions of PRBC and others bloods products, according to the surgical procedure and to the patient's caracteristicus;
• to assessee the quality of transfusion practices according to the national and international recommendations;
• to evaluate the accuracy of the preoperative bleeding risk estimation;
• to check the traceability of the transfusions in our department;
• to evaluate the length of hospital stay (LOS) and the 30-days outcome of transfused patients

Starting from these objectives, a model of transfusion management could be defined.

Also, requests of pre-transfusion blood test could be limited, reducing the costs associated with transfusion management.

Definition of the population This is an observational monocentric retrospective study. All the patients undergoing to a scheduled surgical procedure at the Montpellier University Hospital, head and neck department, between 01-01-2015 and 31-12-2017.

The scheduled surgery will be defined as performed at least 48 hours after the indication given by the surgeon.

The only exclusion criteria will be age < 18 years and emergency surgery patients.

Data circuit definition All data will come from our computerized medical records. A complete list of transfused patients for this period will be requested to the hospital transfusion center.

Each patient will be renamed with a progressive numeric code and the database will be saved in a password-protected hospital computer. The data will be read and analyzed exclusively on Montpellier University Hospital's computers by people

Data collection

The following data will be collected for each patient:

Demographic data

• Name, hospital ID
• Age, sex, weight, height, BMI, ASA classification
• Reason for hospitalization, type and day of surgery
• Emergency surgery during the hospitalisation
• Duration of anesthesia
• Existing co-morbidities (especially liver and coagulation diseases, anemia, chronic inflammatory bowel diseases and diseases influencing the level of hemoglobin and coagulation)
• Anti-aggregating treatments, personal anticoagulants at risk of haemorrhage

Biological data

• Basal Hb, Ht, INR, ACT, aPTT, platelets count, fibrinogen, albumin, liver function, lactate and SvO2
• Hb, Ht, INR, ACT, aPTT, platelets count, fibrinogen, albumin, liver function, lactate and SvO2 at the beginning of the surgery, before (period of 30 min) transfusion PRBC and at the end of procedure

Parameters and drugs

• Vital signs, use of vasoactive drugs at the beginning of the surgery, before (period of 30 min) transfusion PRBC and at the end of procedure
• Cardiac events or other incidents before (period of 30 min) transfusion
• Estimated blood loss
• Volume of crystalloids, colloids, labile blood products administered
• Estimated loss of volume until the end of the surgery (blood loss, Urine, other liquids)
• Input-output balance
• Use of: tranexamic acid, PPSB, fibrinogen, Factor VIIa, cryoprecipitate, factor XIII, cell saver
• Traceability of the transfusion

Follow-up

• Blood transfusions during the follow-up period and reason for transfusion
• ICU stay (duration, complications, hours of mechanical ventilation)
• LOS at the hospital
• Mortality at the hospital
• Patient 30-day survival

The end of the follow-up is defined by the hospital discharge or by the 30th postoperative day.

Statistical management of data

This study is observational retrospective, no intervention of the patient will take place, only descriptive statistical methods will be used.

Future development This database could be the basis for a second study, comparing data from not-transfused patients undergoing to the same surgery.

These results could be also at the origin of a prospective randomized controlled study concerning the transfusion practice in our center.

Inclusion Criteria:

• scheduled surgery between 1st Jan 2015 and 31st Dec 2017
• age > 18

Exclusion Criteria:

• Emergency surgery
• Age < 18

All patients who have had scheduled surgery at the Gui de Chauliac Hospital between 01-01-2015 and 31-12-2017

NC