In Reply

We thank Drs. Zhang and Xue for their interest in our article¹  and appreciate the opportunity to answer their questions. First, we agree that preoperative coronary angiography with contrast media may be related to the development of postoperative acute kidney injury (AKI).²  However, as shown in table 1 in our article,¹  there was no significant difference in the number of patients who underwent coronary angiography less than or equal to 7 days before surgery between the two groups. Additionally, our previous study showed that the timing of the coronary angiography, the type of contrast agent, or the amount of contrast agent did not affect the development of postoperative AKI in patients who underwent off-pump coronary artery bypass surgery.³  Thus, this is not likely to have had an impact on our results.

Second, although we agree that perioperative anemia and transfusion are important risk factors for postoperative AKI,⁴  we do not agree that “perioperative hemoglobin levels were not provided.” We presented the preoperative hematocrit level, the intraoperative lowest hematocrit level, and the number of packed red blood cell units transfused during or after surgery in tables 1 and 2 in our article,¹  all of which were similar between groups. As we stated in our article, we also performed adjustment for preoperative hematocrit using logistic regression.

Third, we agree that adjustment of creatinine for fluid balance has been proposed for a more accurate assessment of AKI.⁵  However, in our study, we did not calculate corrected serum creatinine for fluid balance due to inaccurate fluid balance information including incorrect counts of bleeding. Therefore, it is possible that the positive fluid balance may influence the diagnosis of postoperative AKI, and the differences in fluid balance between the two groups influenced our results. However, as shown in table 2 in our article,¹  although there was no significant difference between the two groups, slightly higher postoperative weight gain, which indicates more positive fluid balance, was shown in the control group. Given that the effects of adjustment are larger in patients with more positive fluid balance, the incidence of AKI in the control group may be more masked than that in the albumin group.

Finally, we also agree on the point that several confounding postoperative events could affect our results. Therefore, as the authors suggested, we performed additional analyses to assess the effects of adjustment for postoperative confounding variables (except postoperative hypoalbuminemia and sepsis due to the possibility of mediator and no occurrence, respectively). After additional adjustment for postoperative low cardiac output syndrome and anemia using logistic regression, similar protective effects of albumin treatment were found (odds ratio, 0.409; 95% CI, 0.195 to 0.859; P = 0.018 and odds ratio, 0.411; 95% CI, 0.196 to 0.862; P = 0.019, respectively).