After obtaining Faculty of Medicine, Ain-Shams University ethical committee approval (FMASU R16/2021), informed consent was taken from 60 patients, ASA physical status II, aged 21–65 years, known to have type 2 diabetes mellitus, with pre-operative FBG ˂ 350 mg/dl, scheduled to undergo elective laparotomy surgeries; expected to exceed 2 hours duration in this randomized study at Ain-Shams University Hospitals. Randomization was done using computer-generated random number tables with sealed opaque envelopes.

Pre-operative history taking, physical examination were done and investigations included; complete blood count, the coagulation profile, liver and kidney function tests, FBG, s.k, glycosylated hemoglobin (HbA1c) and electrocardiography (ECG). During the pre-anesthetic visit, patients were counseled to continue their non-insulin injectables and oral hypoglycemic drugs (insulin secretagogues, metformin, thiazolidinediones and dipeptidyl peptidase − 4 “DPP-4”), and to stop the sodium glucose co-transporter-2 (SGLT-2) inhibitors the day before surgery (Salpeter, 2010; Joshi et al., 2010; Umpierrez et al., 2013; Handelsman et al., 2016). Patients treated with insulin were counseled to; reduce their basal insulin (glargine or detemir) dose by 25% the evening before or the morning of surgery, if twice daily dosing. Neutral protamine Hagedorn (NPH) insulin and premixed formulations to be reduced; by 20% the evening before surgery and by 50% the morning of surgery (Likavec et al., 2006; Rosenblatt et al., 2012).

For all patients, the CBG was measured pre-operative in the ward every 2 hours by glucometer, (Accu-Chek Performa and Accu-Chek Performa glucose strips; error of measurement is ±15% of the measured glucose values when compared with standard laboratory values), with shifting to sliding scale using rapid acting SC insulin injection at least every 4 hours if the CBG > 140 mg/dl. If CBG > 400 mg/dl; continuous rapid acting insulin infusion was started, with ruling out diabetic ketoacidosis and hyperglycemic hyperosmolar Syndrome (Stephen, 2016).

Exclusion criteria

Patients’ refusal, patients taking steroids, patients with active infection, diabetic ketoacidosis, hyperglycemic hyperosmolar syndrome, s. K ˂3.5 mEq/L, HbA1c > 8.5% (Raju et al., 2009), patient’s baseline CBG measured in the induction room (0 time) ˂180 mg/dl, and the presence of acetone in urine for CBG ˃300 mg/dl in the induction room (Fig. 1).

Fig. 1

Flow chart for patients’ enrollment and allocation

In the induction room, patients had a 20G IV cannula inserted for IV insulin injection and insulin or normal saline (NS) infusion.

Preparation of the study drugs

The rapid acting insulin used was Monocomponent Human Insulin, biosynthetic r-DNA origin-Human Actrapid®(manufactured by the Egyptian Drug trading Company, under license from Novo Nordisk Production SAS. Chartres, France).

In a 1 ml (100 units) insulin syringe, ten international units (IU) (0.1 ml) of rapid acting insulin were added to 0.9 ml NS (1 IU insulin/ 0.1 ml).

In a 10 ml syringe, the calculated insulin units for each patient were taken from the prepared inulin syringe, and then NS was added to have a 10 ml volume.

In a 50 ml syringe installed in an infusion pump, 50 IU (0.5 ml) of rapid acting insulin were added to 49.5 ml of NS (1 IU of insulin/ 1 ml).

Another 50 ml syringe installed in an infusion pump, containing NS only was prepared.

Patients were then divided into 2 equal groups of 30 patients each

Group (BII): Bolus-Infusion approach of rapidly acting crystalline insulin. The patient’s baseline CBG (0 min) was divided by 100 (Elizabeth et al., 2017). The calculated insulin units were given from the prepared 10 ml syringe IV over 10 minutes (1 ml/min). Then, the calculated insulin units were given as an IV infusion/hour (Duggan et al., 2017) by the 50 ml syringe prepared in the syringe pump. The CBG was measured every 30 minutes and in the PACU, and the insulin infusion rate was adjusted by dividing the measured CBG by 100. The insulin infusion was stopped when the CBG was ˂180 mg/dl. Supplemental IV insulin bolus and infusion; was given if the CBG re- increased ≥180 mg/dl after a period of intra-operative glycemic control, and stoppage of the insulin infusion.

Control group (SSI): The sliding scale approach of rapidly acting crystalline insulin was used according to the patient’s CBG value; 4 IU of insulin were given for CBG range (180–250 mg/dl), 6 IU of insulin were given for CBG range (251–300 mg/dl), 8 IU of insulin were given for CBG range (301–350 mg/dl) and 10 IU of insulin were given for CBG range (351–400 mg/dl) (Duggan et al., 2017). The required insulin units were given from the prepared 10 ml syringe IV over 10 minutes (1 ml/min). The patient’s baseline CBG (0 min) was then divided by 100, and the calculated value was IV infused/hour, by the 50 ml syringe prepared in the syringe pump containing NS only. The CBG was measured every 30 minutes and in the PACU. The NS infusion rate was adjusted according to the measured CBG divided by 100. The NS infusion was stopped when the CBG was ˂180 mg/dl.

If hypoglycemia (CBG < 100 mg/dl) occurred, 10–20 g of hypertonic dextrose (10%) were IV given. CBG measurement was repeated after 15 min with additional dextrose given to maintain the CBG > 100 mg/dl.

The anesthetic technique

On arrival to the operating room, pulse oximetry and non-invasive blood pressure (NIBP) were applied to the patients. An 18 G IV line was secured. Under complete aseptic conditions and local skin infiltration, thoracic epidural catheter was inserted at T10–11 or T11–12 intervertebral space. Patients were co-loaded with IV 10 ml/kg Ringer’s solution. Patients were positioned supine after epidural catheter fixation. Five leads ECG monitor was applied. The epidural was activated by 5 ml of 0.25% bupivacaine after checking the blood pressure of the patient. Maintenance of the epidural analgesia was done with 5 ml/hr. of 0.25% bupivacaine, readjusted according to the patient’s hemodynamics.

Induction of general anesthesia and endo-tracheal intubation was done. Central venous line, arterial canulation, and a urinary catheter were secured according to the type of operation. Patients were mechanically ventilated and capnography was attached to the breathing circuit. After completion of surgery and tracheal extubation, patients were transferred to the PACU.

Primary outcome

Efficacy of the BII approach, in terms of percentage of the operative time with the target CBG range (140–180 mg/dl).

Secondary outcomes 1.

Extra insulin units given to the patients in both groups: Defined as the sum of the insulin units given all over the operation, after the initial insulin units given.

2.

Supplemental IV insulin units given to the patients in the BII approach: Defined as the insulin bolus and infusion units given, for re-increase of the CBG ≥ 180 mg/dl, after a period of intra-operative control in the target CBG range and stoppage of the insulin infusion.

3.

Total insulin IU given to the patients in both groups.

4.

Number of patients who developed hypoglycemia.

5.

Peri-operative changes in serum potassium; measured in the PACU and compared to the baseline pre-operative laboratory level.

Statistical analysis

Sample size calculation was done by the G power program, setting alpha error at 0.05 and power at 80%, assuming a statistically significant difference with large effect size (d˃0.6) between the 2 interventions. Based on these data, a sample size of at least 26 patients per group was needed.

Data were analyzed using Statistical Package for Social Science (SPSS) version 21.0. Chicago, Illinois, USA. Quantitative data were expressed as mean ± standard deviation. Qualitative data were expressed as count. The independent-samples t-test was used to compare between means in the 2 groups. Chi square test was used to compare proportions between two qualitative parameters. A two-way mixed ANOVA was conducted, to detect if there is any significant effect of the two approaches on mean s. K concentration over time (pre-operative versus post-operative). P value < 0.05 was considered statistically significant.