Percutaneous Coronary Intervention versus Coronary-Artery Bypass Grafting for Severe Coronary Artery Disease
2009 · paper · Clinical trials
A randomized trial of 1800 patients with three-vessel or left main coronary disease, comparing stenting (PCI) with bypass surgery (CABG). At 12 months, major cardiac or cerebrovascular events were more frequent with PCI (17.8% vs 12.4%), mostly from repeat procedures; stroke was more frequent with CABG.
Checked: PDF from dea.lib.unideb.hu, fetched 5 Oct 2026. Page numbers are those of that PDF.
4 findings
Substance The number needed to treat for myocardial infarction is given as "71 for myocardial infarction", but the Table 3 MI rates (43/891 vs 28/849) give 1/difference ≈ 65.4. The other NNTs in the same paragraph (19, 14, 119, 60) match 1/difference from Table 3.
- p. 7 Results, first paragraph (NNT)
and 71 for myocardial infarction
- p. 9 Table 3, MI row
43/891 (4.8)
- p. 7 Primary outcome, NNT paragraph
71 for myocardial infarction
Only the authors can say what is right here.
Judge's note: Counts read from the Table 3 image. Using the same method that reproduces the paragraph's other NNTs, MI gives about 66, not 71. Even with the rounded rates (4.8 − 3.3) it is about 67. Neither the methods nor the text explain a different basis, such as Kaplan–Meier rates, so only the authors can say where 71 comes from.
Second check: another model agreed
Report a problem with this finding
Reference or label The Results text says "blood pressure of 130/80 mm Hg or higher was significantly larger in the PCI group", but the Table 1 row it cites is "Blood pressure ≥130/85 mm Hg — %" (68.9 vs 64.0, P=0.03).
- p. 4 Results, Study Participants, para 2
blood pressure of 130/80 mm Hg or higher was significantly larger in the PCI group
- p. 6 Table 1 row
Blood pressure ≥130/85 mm Hg — %
- p. 4 Results, Study Participants
blood pressure of 130/80 mm Hg or higher
- p. 4 Study Participants, para 2
blood pressure of 130/80
Only the authors can say what is right here.
Judge's note: Checked on the page 4 text and the page 6 Table 1 image. The text cites Table 1 for this imbalance, and the only significant blood-pressure row there uses ≥130/85 mm Hg. The paper reports no other blood-pressure variable. One of the two thresholds is mislabelled, and only the authors can say which is right.
Second check: another model agreed
Report a problem with this finding
Wording The Discussion says the PCI group had a high rate of "lesions characterized by bifurcation or trifurcation (>80%)", but Table 1 gives bifurcation 72.4% (PCI) and 73.3% (CABG), the Results give 72.8% overall, and 63.1% of PCI patients had a bifurcation or trifurcation treated.
- p. 9 Discussion, para 2
lesions characterized by bifurcation or trifurcation (>80%)
- p. 6 Table 1, Bifurcation row
72.4
- p. 4 Results
72.8% had a bifurcation lesion
- p. 9 Discussion, second paragraph
characterized by bifurcation or trifurcation (>80%)
Only the authors can say what is right here.
Judge's note: Checked Table 1 (page 6), the Results (page 4) and the Discussion (page 9). The only figure above 80% in the paper is the CABG registry's 80.8%, which is a different population. Trifurcations are not reported separately, so a lesion-level or combined definition cannot be ruled out. Only the authors can say what the >80% refers to.
Second check: another model agreed
Report a problem with this finding
Wording The Discussion says lesions were "totally occluded (>25%)", but Table 1 gives total occlusion 24.2% (PCI) and 22.2% (CABG), and the Results give 23.1% overall.
- p. 9 Discussion, para 2
or totally occluded (>25%)
- p. 6 Table 1, Total occlusion row
24.2
- p. 4 Results
total occlusion was identified in 23.1%
- p. 6 Table 1
Total occlusion — %
Only the authors can say what is right here.
Judge's note: Checked Table 1 (page 6), the Results (page 4) and the Discussion (page 9). No randomized figure exceeds 25%, and the Table 1 footnote says these are percentages of patients. The registry figure of 56.4% is a different population. The Discussion overstates the data.
Second check: another model agreed
Report a problem with this finding
3 minor (typos and the like; most checked by the judge only)
Wording The text says that of patients with diabetes "about one third required insulin". Table 1 gives insulin 9.9% of 25.6% (PCI) and 10.4% of 24.6% (CABG), which is about 39–42% of patients with diabetes.
- p. 4 Results, Study Participants
of whom about one third required insulin
- p. 6 Table 1
Requiring insulin
Judge's note: Values read from the Table 1 image and the page 4 text. The ratios are about 0.39 and 0.42, closer to two fifths than one third. The text says 'about', so this is a loose description rather than a wrong figure.
Second check: another model agreed
Report a problem with this finding
Wording The P value for 12-month death, stroke, or MI is "P=0.99" in Figure 2B but "0.98" in Table 3 (68/891 vs 65/849). Both captions say chi-square on binary rates.
- p. 8 Figure 2, Panel B
P=0.99
- p. 9 Table 3, Death, stroke, or MI row
0.98
- p. 9 Table 3, Death, stroke, or MI
68/891 (7.6)
- p. 8 Figure 2 panel B end labels
PCI 7.7 / CABG 7.6
Only the authors can say what is right here.
Judge's note: Read from the page 8 and page 9 images. The Figure 2 caption says P values come from the chi-square test on binary rates, and the Table 3 footnote names the same test for the same counts. An uncorrected chi-square gives z ≈ 0.019, P ≈ 0.985. The two printed values sit on either side of a rounding boundary and the difference changes no conclusion. The authors should say which value is intended. The label swap in this panel is handled under C007.
Second check: another model agreed
Report a problem with this finding
Typo Statistical Analysis names the primary end point as "adverse binary cardiac or cerebro-vascular events"; everywhere else it is "major adverse cardiac or cerebrovascular events".
- p. 3 Statistical Analysis
end point of adverse binary cardiac or cerebro-vascular events
Judge's note: Seen on the page 3 crop. The primary end point is called 'major adverse cardiac or cerebrovascular events' everywhere else in the paper (Results, Table 3, Figure 2D). Here 'major' is missing and 'binary' is in the wrong place. This is a wording slip with an obvious fix, and the meaning can still be recovered.
Second check: none (judge only)