2.1

Enrollment

Candidates were screened prior to diagnostic coronary angiography or scheduled PCI. HbA1c was measured using a point-of-care device (DCA 2000+, Bayer Corporation). Those providing consent were taught to measure and record blood glucose and to inject subcutaneous insulin or saline with a pen device. For those undergoing scheduled qualifying PCI, a 2-day run-in phase allowed coordinators to determine candidates’ ability to self-inject insulin and to self-monitor blood and record blood glucose. Those undergoing diagnostic angiography and possible PCI were required to complete an abbreviated run-in phase requiring patients to demonstrate one successful glucose test and one successful injection of saline. Candidates unable to comply were not enrolled in order to reduce the frequency of noncompliance or dropout in those assigned to insulin.

Central randomization by telephone assigned subjects in a 1:1 ratio using permuted blocks stratified by vessel size, planned versus ad hoc PCI, and participating site. The study nurse called patients assigned to the insulin group every 2 to 3 days for the first 10 days to support adherence to protocol.

2.2

Treatment arms

Subjects assigned insulin were initially treated with NPH insulin (Novolin NPH, Novo Nordisk A/S, Bagsvaerd, Denmark, 100 IU/ml) 10 units at bedtime and then followed a titration algorithm to achieve fasting capillary glucose levels of 4.0–5.4 mmol/l. They were required to measure and record their capillary glucose levels at least twice daily during titration and at least once daily thereafter. Insulin was continued for 6 months following the PCI procedure by protocol, although patients were given the option to continue insulin indefinitely. Hypoglycemia was addressed by reducing oral agents and/or the dose of insulin. Follow-up visits were scheduled at 6, 12, 18, 24, and 52 weeks, and telephone follow-up at 3, 9, 15, and 21 weeks following PCI. Insulin dose, weight, symptoms and signs of hypoglycemia, and all interval diary entries were reviewed and recorded at each visit.

Subjects assigned usual care were advised to continue their oral diabetes medications, measure glucose levels at least daily, and discuss any changes in glucose-lowering therapy with their physician. Visits occurred at 12, 24, and 52 weeks post-PCI. If the 12-week A1c was greater than 8.4%, the dose of one oral agent was increased or (if maximal doses were already prescribed) another agent was added (metformin was not used if the serum creatinine was ≥130 μmol/l). All subjects received counseling regarding diet and activity.

2.3

PCI procedure

PCI procedures employed approved devices and standard antiplatelet and antithrombotic regimens. Abciximab was recommended. For practical reasons and to reflect routine practice, the use of both bare metal stents (BMS) and DES was allowed in this study. Patients underwent systematic angiography and IVUS at baseline and follow-up using standard techniques.

2.4

Follow-up coronary imaging

Patients were considered eligible for angiographic follow-up if no adverse events [e.g., abrupt closure, re-PCI, coronary artery bypass surgery (CABG)] occurred within the first 7 days post-PCI. Patients undergoing angiography before 3 months who did not exhibit angiographic restenosis by visual assessment were requested to have repeat angiography at 6 months with IVUS.

2.5

IVUS and angiography

IVUS was performed using motorized pullback at 0.5 mm/s beginning at an easily identifiable distal landmark. We employed blinded quantitative analyses of angiographic and IVUS images at a core laboratory (G.B.J.M., M.R.) using standard methods .

2.6

Study endpoints

The primary outcome was absolute change in “normalized” plaque volume within the stented segment as assessed by IVUS . Relative change in plaque volume (“percent” volume change) was also calculated. Other outcomes included (a) plaque volume in adjacent 5-mm length proximal and distal per-stent segments were also performed, (b) angiographic in-stent late loss (difference of baseline and follow-up minimal luminal diameters), and (c) clinical safety at 12 months.

2.7

Statistical methods

We estimated that 240 participants would be required to demonstrate a 25% between-group difference in IVUS-measured “normalized” plaque volume at 6 months, with a power of 80% and alpha of 0.05, assuming a mean intimal hyperplasia volume and standard deviation volume of 35 mm ³ and 22 mm ³ , respectively, and a 20% rate of unsuccessful interventions and nonadherence. However, the study was terminated after 78 patients were enrolled due to protracted recruitment.

All analyses were performed using SAS statistical software (version 8.12, SAS Institute Inc., Cary, NC, USA). An intent-to-treat approach was used. Categorical variables were described using frequencies, while continuous variables were reported as mean (with 95% confidence intervals) and standard deviations. The effect of allocation to the insulin group (i.e., the treatment group) on the primary outcome was analyzed using a generalized linear regression model in which the dependent variable was the 6-month change in this outcome and in which the independent variables were treatment allocation, age, gender, and baseline HbA1c to account for any between-group differences at the time of randomization. All analyses were confirmed using nonparametric methods; parametric results are reported.

2.1

Enrollment

Candidates were screened prior to diagnostic coronary angiography or scheduled PCI. HbA1c was measured using a point-of-care device (DCA 2000+, Bayer Corporation). Those providing consent were taught to measure and record blood glucose and to inject subcutaneous insulin or saline with a pen device. For those undergoing scheduled qualifying PCI, a 2-day run-in phase allowed coordinators to determine candidates’ ability to self-inject insulin and to self-monitor blood and record blood glucose. Those undergoing diagnostic angiography and possible PCI were required to complete an abbreviated run-in phase requiring patients to demonstrate one successful glucose test and one successful injection of saline. Candidates unable to comply were not enrolled in order to reduce the frequency of noncompliance or dropout in those assigned to insulin.

Central randomization by telephone assigned subjects in a 1:1 ratio using permuted blocks stratified by vessel size, planned versus ad hoc PCI, and participating site. The study nurse called patients assigned to the insulin group every 2 to 3 days for the first 10 days to support adherence to protocol.

2.2

Treatment arms

Subjects assigned insulin were initially treated with NPH insulin (Novolin NPH, Novo Nordisk A/S, Bagsvaerd, Denmark, 100 IU/ml) 10 units at bedtime and then followed a titration algorithm to achieve fasting capillary glucose levels of 4.0–5.4 mmol/l. They were required to measure and record their capillary glucose levels at least twice daily during titration and at least once daily thereafter. Insulin was continued for 6 months following the PCI procedure by protocol, although patients were given the option to continue insulin indefinitely. Hypoglycemia was addressed by reducing oral agents and/or the dose of insulin. Follow-up visits were scheduled at 6, 12, 18, 24, and 52 weeks, and telephone follow-up at 3, 9, 15, and 21 weeks following PCI. Insulin dose, weight, symptoms and signs of hypoglycemia, and all interval diary entries were reviewed and recorded at each visit.

Subjects assigned usual care were advised to continue their oral diabetes medications, measure glucose levels at least daily, and discuss any changes in glucose-lowering therapy with their physician. Visits occurred at 12, 24, and 52 weeks post-PCI. If the 12-week A1c was greater than 8.4%, the dose of one oral agent was increased or (if maximal doses were already prescribed) another agent was added (metformin was not used if the serum creatinine was ≥130 μmol/l). All subjects received counseling regarding diet and activity.

2.3

PCI procedure

PCI procedures employed approved devices and standard antiplatelet and antithrombotic regimens. Abciximab was recommended. For practical reasons and to reflect routine practice, the use of both bare metal stents (BMS) and DES was allowed in this study. Patients underwent systematic angiography and IVUS at baseline and follow-up using standard techniques.

2.4

Follow-up coronary imaging

Patients were considered eligible for angiographic follow-up if no adverse events [e.g., abrupt closure, re-PCI, coronary artery bypass surgery (CABG)] occurred within the first 7 days post-PCI. Patients undergoing angiography before 3 months who did not exhibit angiographic restenosis by visual assessment were requested to have repeat angiography at 6 months with IVUS.

2.5

IVUS and angiography

IVUS was performed using motorized pullback at 0.5 mm/s beginning at an easily identifiable distal landmark. We employed blinded quantitative analyses of angiographic and IVUS images at a core laboratory (G.B.J.M., M.R.) using standard methods .

2.6

Study endpoints

The primary outcome was absolute change in “normalized” plaque volume within the stented segment as assessed by IVUS . Relative change in plaque volume (“percent” volume change) was also calculated. Other outcomes included (a) plaque volume in adjacent 5-mm length proximal and distal per-stent segments were also performed, (b) angiographic in-stent late loss (difference of baseline and follow-up minimal luminal diameters), and (c) clinical safety at 12 months.

2.7

Statistical methods

We estimated that 240 participants would be required to demonstrate a 25% between-group difference in IVUS-measured “normalized” plaque volume at 6 months, with a power of 80% and alpha of 0.05, assuming a mean intimal hyperplasia volume and standard deviation volume of 35 mm ³ and 22 mm ³ , respectively, and a 20% rate of unsuccessful interventions and nonadherence. However, the study was terminated after 78 patients were enrolled due to protracted recruitment.

All analyses were performed using SAS statistical software (version 8.12, SAS Institute Inc., Cary, NC, USA). An intent-to-treat approach was used. Categorical variables were described using frequencies, while continuous variables were reported as mean (with 95% confidence intervals) and standard deviations. The effect of allocation to the insulin group (i.e., the treatment group) on the primary outcome was analyzed using a generalized linear regression model in which the dependent variable was the 6-month change in this outcome and in which the independent variables were treatment allocation, age, gender, and baseline HbA1c to account for any between-group differences at the time of randomization. All analyses were confirmed using nonparametric methods; parametric results are reported.