Intraoperative blood glucose concentration abnormalities are associated with increased perioperative morbidity and mortality (1-4). Severe intraoperative hyperglycemia (BGC ≥ 200 mg/dl) in patients undergoing craniotomy for urgent/emergent craniotomy after traumatic brain injury complicates 15% of the cases and is associated with higher in-hospital mortality. Intraoperative use of dexamethasone during craniotomy is also known to induce an increase in blood glucose concentration.

The importance of blood glucose concentration in neurosurgical patients is witnessed by the effects of tight blood glucose control on incidence of infections and neurological outcome . Currently available evidence suggest that, in neurosurgical patients, perioperative BGC values should be within the 80-180 mg/dl range .

Data on the prevalence of severe intraoperative hyper (blood glucose concentration >180 mg/dl) and hypoglycemia (blood glucose concentration <80 mg/dl) in patients undergoing craniotomy for supra or infratentorial surgery as elective or emergency procedure are lacking as it is not known whether in these patients intraoperative severe hyperglycemia relates to an increased incidence of postoperative infections is unknown.

Aim of this prospective observational study -in patients undergoing craniotomy for supra or infratentorial surgery as elective or emergency procedure- was to test the hypothesis that severe intraoperative hyperglycemia (blood glucose concentration ≥180mg/dl) is associated with an increased incidence of infections within the first postoperative week (pneumonia, sepsis, urinary and wound and cerebral infections). We also recorded the prevalence of severe intraoperative hyper and hypoglycemia (blood glucose concentration<80 mg/dl) in recruited patients.

METHODS Study Design This prospective observational cohort study, in adult patients who underwent craniotomy for supra or infratentorial surgery as elective or emergency procedure, was approved by the University of Rome "Sapienza", Policlinico Umberto I hospital, Institutional Review Board (2665, 28/3/2013). Recruited patients were categorized according the type of procedure into 4 groups: primary brain tumors (gliomas and meningiomas); metastatic tumors; neurovascular lesions (intracranial hemorrhage, subarachnoid hemorrhage, arterovenous malformation); traumatic brain injury (14-16). In each group "elective" and "emergency" procedures were recorded apart. Because of the observational study design, perioperative clinical management was run by the attending anesthesiologist. Standard of care guidelines do not include the use of insulin infusion should full nutritional supply is not provided.

Patients population Adults ≥18 years of age, who underwent elective or emergency craniotomy for brain lesions in the supra or infratentorial space were prospectively recruited. In all patients intraoperative blood glucose concentration was measured twice: at the beginning of the procedure (immediately after endotracheal intubation) and at the end of surgery (immediately before extubation) on arterial whole blood by blood gas analysis (Instrumentation Laboratory, BGA analyzer GEM 4000, UK). Patients with one or both BGC>180 mg/dl were categorized as having "severe intraoperative hyperglycemia".

Intraoperative Setting All patients received normal saline (NaCl 0.9%) as pre and intraoperative fluid infusion. There was no fixed protocol for the anesthetic management and the anesthetic regime was determined by the attending anesthesiologist. Anesthesia was maintained with either halogenated inhalational anaesthetics (desflurane or sevoflurane) or with continuous propofol infusion and fentanyl as per clinical needs. Intraoperative monitoring was accomplished with 3 leads ECG, SpO2, end tidal CO2, invasive arterial pressure, core body temperature and urine output. Patients were normoventilated (PaCO2 35-40 mmHg) and mild hyperventilation was used to facilitate brain relaxation only when necessary . There was no dedicated protocol for treatment of intraoperative hyperglycemia and intraoperative insulin use was discouraged unless the attending anesthesiologist considered it necessary. Recorded demographic and clinical characteristics included: age, gender, localization of the brain lesion (supra or infratentorial), type of surgery (primary brain tumors, metastatic brain tumors, neurovascular, traumatic brain), setting of surgery (elective or emergent), duration of surgery, history of diabetes mellitus (and the use of chronic oral hypoglycemic agents or insulin), associated cardiac, vascular, renal or hepatic morbidity the intraoperative use of mannitol and steroids. In the postoperative period, all patients were kept fasting for the first 24 hours, subsequently full calories load -by self alimentation or enteral/parenteral nutrition- was warranted. As per local protocol, in the postoperative period, insulin was administered only in patients with blood glucose concentration> 180 mg/dl.

Definitions Hyperglycemia was defined as BGC ≥ 180 mg/dL. Hypoglycemia was defined as glucose≤ 80 mg/dL. Postoperative infections (pneumonia, blood stream-sepsis, urinary, surgical site/wound and cerebral infections) diagnosed according to Center for Disease Control and Prevention criteria within 7 days after surgery . The onset and type of infection was determined by 2 consultants for infective disease.

Outcomes The primary outcome was the prevalence of postoperative infections in normoglycemic patients and in those that presented at one intraoperative blood glucose concentration ≥ 180 mg/dL.

Secondary outcomes measures were: incidence of severe intraoperative hyper (blood glucose concentration ≥180 mg/dL) and hypo (blood glucose concentration <80 mg/dL) glycemia in patients undergone craniotomy; the relationship between the setting of surgery (elective or emergency), history of diabetes mellitus and intraoperative use of mannitol or steroids and severe intraoperative hyperglycemia.

Statistical Analysis For sample size calculation, we hypothesize that incidence of post operative infections would be 30% higher in patients that had a severe intraoperative hyperglycemia. We estimated that 28 patients should be recruited in each study group to detect a 30% difference in the incidence of postoperative infections. A difference in the incidence of intraoperative hypo and hyper glycemia in the different subgroups of patients undergone craniotomy was calculated using the Chi square test.

Inclusion Criteria:

• patients undergoing craniotomy

Exclusion Criteria:

• age <18 years

at least 28 patients in each group