Japanese Registry Assesses Stroke Rates for Off- vs. On-pump CABG, PCI

The overall incidence of stroke, regardless of the time of onset, is similar for patients undergoing minimally invasive off-pump coronary artery bypass graft (CABG) surgery or percutaneous coronary intervention (PCI), according to results from a large Japanese registry. The study, published online September 17, 2012, ahead of print in the American Journal of Cardiology, also found that compared with standard bypass surgery, off-pump CABG was associated with a lower rate of early stroke.

Researchers led by Akira Marui, MD, PhD, of Kyoto University Graduate School of Medicine, (Kyoto, Japan), studied 6,323 patients with multivessel and/or left main CAD enrolled in the multicenter CREDO-Kyoto PCI/CABG registry. Patients underwent revascularization via PCI (n = 3,877), conventional on-pump CABG (n = 1,381) or off-pump CABG (n = 1,065) from January 2000 through December 2002.

Strokes were classified as:

  • Early: within 24 hours after coronary revascularization
  • Delayed: after 24 hours and within 30 days
  • Late: after 30 days and up to a median follow up of 3.4 years

The incidence of stroke at 30 days was 0.23% for PCI, 2.5% for on-pump CABG, and 0.94% for off-pump CABG (P < 0.01). Thirty-day mortality was 0.85%, 2.2%, and 0.83%, respectively.

In propensity-matched analysis, no differences were seen in rates of early, delayed, and late stroke between the PCI and off-pump CABG groups (table 1).

Table 1. Stroke Occurrence: PCI vs. Off-pump CABG


Propensity Score-Adjusted Risk Ratio (95% CI)

P Value


0.65 (0.08-5.45)



0.36 (0.10-1.29)



0.81 (0.52-1.27)



In contrast, the incidence of early stroke after on-pump CABG was much higher than that after the off-pump procedure, with little difference in occurrence of delayed or late stroke (table 2).

Table 2. Stroke Occurrence: On-pump CABG vs. Off-pump CABG


Propensity Score-Adjusted Risk Ratio (95% CI)

P Value


7.22 (1.67-31.3)



1.66 (0.70-3.91)



1.18 (0.83-1.69)



Stroke history and A-fib were pre-operative risk factors for early, delayed, and late stroke. Additional risk factors included chronic kidney disease (OR 2.79; P = 0.02), for early stroke, malignancy (OR 3.39; P < 0.01), for delayed stroke, and hemodialysis (OR 2.99; P < 0.01), for late stroke.

Off-Pump Catching up to PCI?

According to the study authors, the 30-day stroke rates seen in the study are consistent with prior findings. In addition, since the early stroke rate after off-pump CABG was lower than that after on-pump CABG, it suggests the possibility that revascularization with the off-pump technique could “decrease the probability of perioperative stroke to a value comparable to that of PCI,” they write, especially with further improvements in postoperative management after off-pump CABG.

Dr. Marui and colleagues also point out that the 30-day mortality rate of 0.85% after PCI suggests that the occurrence of post-procedural stroke may not necessarily be associated with death. “Patients who developed a stroke who previously underwent longer cardiac catheterization procedures and those who had a greater use of contrast were more likely to have had the procedure for urgent reasons and to have had intra-aortic balloon counterpulsation, a procedure known to increase the risk of stroke,” they add.

The study authors also note that their data, like those from the New York State Angioplasty Registry, differ from several randomized controlled trials that have found no difference in morbidity and mortality between on-pump and off-pump CABG. They attribute the difference to selection bias, adding that “registry data, which include higher-risk patients than [randomized trials] can better demonstrate the advantages of [off-pump] compared to on-pump CABG.”

Not There Yet

In a telephone interview with TCTMD, James B. Hermiller Jr., MD, of the St. Vincent Heart Center of Indiana (Indianapolis, IN) said the registry data are “interesting” but nevertheless fraught with the inherent limitations of non-randomized studies.

“These data and other registries have demonstrated what we would intuitively think—less manipulation of the aorta leads to lower early stroke rates … but these are incremental data that I don’t think are compelling enough to suggest that everybody should be using [off-pump CABG] instead of on-pump CABG,” he said.

Dr. Hermiller added that he does not believe “the data are there yet to suggest that early stroke rates of off-pump CABG are going to be equal at this point to PCI. That’s a hypothesis and it’s not surprising that as there’s more experience with off-pump CABG and as surgeons become more facile with it, event rates are dropping. But we’ve known all along that early stroke rates are lower with PCI than with CABG, irrespective of how you do it, and that hasn’t changed based on this study. We would need adequately powered, good randomized studies to really suggest that off-pump is superior to on-pump CABG and we don’t have that.”



Marui A, Kimura T, Tanaka S, et al. Comparison of frequency of postoperative stroke in off-pump coronary artery bypass grafting versus on-pump coronary artery bypass grafting versus percutaneous coronary intervention. Am J Cardiol. 2012; Epub ahead of print.

Related Stories:

  • The study was supported in part by a grant from the Ministry of Health, Labor and Welfare, Tokyo, Japan, and an educational grant from the Research Institute for Production Development, Kyoto, Japan.
  • Drs. Marui and Hermiller report no relevant conflicts of interest.