During the conduction of a recent analysis, an error in the coding of the rule-in path of the ESC 0/3-h algorithm in the main database was discovered. This coding error affected the diagnostic measures for the positive predictive value and the specificity to detect acute myocardial infarction. In summary, the coding error resulted in an incorrect high relative increase of troponin comparing the baseline and 3-h value, while all other analyses were not affected.
This coding error occurred at the beginning of the analysis in the main database; it occurred for both groups, patients with and without diabetes, similarly. Since the erroneous variable provides a relative change, the comparison of both groups and the interpretation of data in this particular analysis of patients with diabetes is largely unaffected. No other analyses were affected by this coding error. The overall message as well as the conclusions drawn based on the results discussed in the article do not change.
The authors have corrected the absolute numbers for the specificity and the positive predictive value of the rule-in pathway of the ESC 0/3-h algorithm in the abstract and main text of the article, Table 2, and Fig. 2 as specified below.
In the abstract, “0/3-h (AD 1.03 [95% CI –7.63, 9.7])” has been replaced with “0/3-h (AD 8.82 [95% CI –1.08, 18.73]).”
In the results, the following text has been corrected:
“Specificity for rule-in was significantly decreased in patients with DM, while the PPV was similar” has been replaced with “Specificity and the PPV for rule-in were similar for patients with DM and those without DM.”
“DM (497 of 513 [96.9%] vs. 2,852 of 2,910, [98.0%]), and the targets for NPV and PPV were not met” has been replaced with “DM (422 of 513 [82.3%] vs. 2,612 of 2,910 [90.1%]), and the target for NPV was not met.”
“PPV 31.0% [95% CI 26.3, 36.0], specificity 35.7% [95% CI 31.0, 40.6]” has been replaced with “PPV 38.4% [95% CI 31.5, 45.7], specificity 70.4% [95% CI 65.6, 74.8].”
“Therefore, to improve rule-in and reach the targeted PPV of 75%, we tested individual cutoffs for optimization of rule-in” has been replaced with “Therefore, to improve rule-in, we tested individual cutoffs for optimization of rule-in.”
“AD 11.03 [95% CI 5.91, 16.14], P < 0.001” has been replaced with “AD 10.5 [95% CI 3.33, 17.67], P = 0.004.”
In the discussion, the text “However, our predefined target of a PPV of 75.0% was solely achieved by applying substantially increased hs-cTnI cutoffs. In this regard, the need for this dramatic increase could only in part be explained by the higher baseline values of hs-cTnI” has been replaced with “However, our predefined target of a PPV of 75.0% was solely achieved by the ESC 0/3-h algorithm. In this regard, the need for increased cutoffs could in part be explained by the higher baseline values of hs-cTnI.”
In Table 2, the data for ESC 0/3-h algorithm specificity for rule-in, PPV for rule-in, and proportion of patients for rule-in, respectively, were corrected as follows:
Patients without DM: “90.2 (89.0, 91.4)” to “95.3 (94.4, 96.1),” “63.7 (60.0, 67.4)” to “73.3 (68.9, 77.4),” and “662 (22.7)” to “431 (14.8).”
Patients with DM: “85.5 (82.0, 89.1)” to “96.2 (93.8, 97.9),” “64.8 (56.8, 72.2)” to “82.1 (72.3, 89.6),” and “159 (31.0)” to “84 (16.4).”
In Table 2, the absolute difference and P values, respectively, were corrected from “–4.4 (–8.2, 0.6)” to “0.89 (–1.32, 3.1)” and “0.01” to “0.51” (for specificity for rule-in), and “1.03 (–7.63, 9.7)” to “8.82 (–1.08, 18.73)” and “0.88” to “0.12” (for PPV for rule-in).
Fig. 2 was replaced with the figure above.
The authors apologize for the errors.
The online version of the article (https://doi.org/10.2337/dc19-1327) has been updated with the correct data.
