Acute kidney injury (AKI) is a sudden loss of kidney function that, from start to finish, occurs in less than 7 days. It is well known from experimental models that, depending on the severity of AKI, some tubule cells are irreversibly lost and replaced by renal progenitor cells. Tubules regenerating after AKI may fail to differentiate and exhibit profibrotic paracrine activity before they become atrophic, so these mechanisms of loss and maladaptive repair imply post-AKI chronic kidney disease (CKD) and a reduction of kidney lifespan (1, 2). In addition, clinical data suggest that AKI at any stage is an independent risk factor for CKD and end-stage CKD (ESCKD) (3). Although the connection between AKI and CKD is well established, it was not until 2017 that the Acute Disease Quality Initiative (ADQI) reached a consensus and defined acute kidney disease (AKD) as disease developing in the period between 7 and 90 days after AKI, which led to the design of studies focusing on interventions in this period, with the aim of preventing CKD after AKI (1).

The incidence of AKI following cardiac surgery (CSA-AKI) is up to 30%, and 2%–5% of patients require renal replacement therapy (RRT) during an AKI episode. CSA-AKI increases the risk of death during admission, which can increase to 50% when there is a need for RRT (4). Given the high incidence of AKI in this controlled scenario, studies have focused on the incidence of de novo CKD [defined as eGFR < 60 mL/min] after cardiac surgery. In 2017, Legouis D et al. studied a cohort of 4,791 patients and found that the risk of CKD was higher in patients who had experienced CSA-AKI than in the AKI-free population (5).

Despite the link between AKI and CKD, information about AKI (even for those patients with a need for RRT) is not always provided in the discharge documentation, which makes it difficult for primary care doctors to improve their kidney function follow-up. This issue was recently reviewed by Ostermann et al. (6). Among the AKI patients who received RRT in intensive care units (ICUs) in the UK, the development of AKI and the need for RRT were mentioned in 85% and 82% of critical care discharge letters, respectively, and the monitoring of kidney function post discharge was recommended in only 36.3% of hospital discharge summaries (6).

Providing clinicians with tools to identify patients at risk of CKD after AKI should be a key priority. With this in mind, Legouis D et al. developed a prediction score for de novo CKD (defined as e GFR< 60 mL/min) 1 year after cardiac surgery that was found to have fair accuracy in a validation cohort [receiver operating characteristic curve (AUC) value of 0.78]. The score comprises preoperative eGFR by Modification of Diet in Renal Disease formula (MDRD) < 80 mL/min (1 point), age > 65 years (1 point), transplant or aortic surgery (2 points), aortic clamping time > 50 minutes (1 point), and AKI stage one (1 point) and AKI stage 2 or 3 (2 points) (7).

With the aim of including all clinically meaningful renal endpoints in AKI clinical trials, the concept of major adverse kidney events (MAKE) was introduced. This composite endpoint comprises persistently impaired renal function (i.e., a 25% or greater decline in eGFR), new hemodialysis, and death. It has been proposed as a way to improve the capacity to understand AKI and provide a means of comparing different interventions (8).

The aim of our study was to assess the incidence of MAKE 1 year after cardiac surgery and its risk factors and, as a secondary objective, to evaluate the utility of the score developed by Legouis D et al. (CSA-CKD score) in the prediction of MAKE 1 year after surgery, and in so doing to shed light on potential tools for the identification of at-risk patients that require particular follow-up.

We conducted a unicentric retrospective study of patients admitted to Hospital Clínic de Barcelona for cardiac surgery with cardiopulmonary bypass (CPB) from January 2015 to December 2015, with a 1-year follow-up after the intervention. The inclusion criteria were patients over 18 years old who had undergone cardiac surgery [i.e., valve substitution (VS), coronary artery bypass graft (CABG), or a combination of both procedures] and who were in need of a CPB. Patients with chronic kidney diseases at any stage were included. However, patients who were already undergoing chronic dialysis therapy, renal transplant recipients, and those who had had an AKI immediately prior to surgery were not included in the study. In addition, patients who had undergone emergent surgeries, intra-aortic balloon pump (IABP) users, patients who died during surgery or admission, and patients with endocarditis were excluded. The Ethics Committee of our institution approved the study (Reg. HCB/2019/0959)

In this retrospective unicentric study, we evaluated the risk of MAKE after CSA-AKI in a Spanish cohort, and the utility of the CSA-CKD score in the prediction of MAKE after discharge. Any-stage CSA-AKI was the only variable associated with the outcome when analyzed in a multivariate analysis with baseline characteristics of the patients. The CSA-CKD score had acceptable discrimination (AUC 0.737) for the prediction of CKD (eGFR < 60), but the AUC decreased to 0.682 for the prediction of MAKE after 1 year.

GFR generally declines at a rate of 1 mL/min/year (9), but in our cohort we observed median declines of 3 mL/min/year and 8 mL/min/year in patients who did not and did experience CSA-AKI, respectively. Patients who undergo cardiac surgery are at an increased risk of losing kidney function, probably because of their comorbidities (for example, we found that a high percentage of patients who underwent cardiac surgery had diabetes and were hypertensive), but this risk is significantly increased when CSA-AKI occurs (p = 0.024). In that regard, Reyden et al. studied a cohort of 29,330 patients who underwent primary isolated CABG in Sweden, with a mean follow-up period of 4.3 years, and found that the risk of end-stage chronic kidney disease (ESCKD) was significantly increased for any-CSA-AKI stage compared with non-CSA-AKI patients, also when stratified by preoperative renal function (10).

Previous studies have focused on the risk of CKD (an eGFR of < 60 mL/min) in this population 1 year after cardiac surgery, but recent evidence shows that defining worsened renal function as a decline of ≤ 25% in eGFR can help to identify patients that can develop CKD in later years, or patients who already have CKD and whose episodes of CSA-AKI could accelerate the decline of their renal function (8). Legouis et al. studied a cohort of 4,791 patients and observed that patients without pre-existing CKD (regardless of their AKI stage) were associated with a risk of de novo CKD after fully recovering from an AKI episode after cardiac surgery, and, based on this finding, they developed a CSA-CKD score to identify at-risk patients (5, 7). It is important to note that excluding patients with an eGFR of < 60 mL/min prevents clinicians from identifying patients who can rapidly progress to ESKD and who may benefit from nephrology follow-up. This is particularly important in cardiac surgery as the percentage of patients with pre-existing CKD is increasing, alongside increased rates of patient comorbidity. For instance, in our cohort almost 20% of the patients had pre-existing CKD. Another study, conducted by Ishami et al., included 29,388 individuals who underwent cardiac surgery. They found that a creatinine increase, defined as either none (0%) or as class I (1%-24%), II (25%-49%), III (50%-99%), or IV (100%) was associated, in a graded manner, with an increased risk of incident CKD, CKD stage progression, and mortality (11). This study also gives more weight to the categories of CKD than to the percentage of GFR decline itself. To our knowledge, the present study is the first that focuses on the impact of CSA-AKI on MAKE, with a special focus on the relative reduction of eGFR in line with current AKI research.

Interestingly, the risk of MAKE in our cohort was not associated with age or sex. This is always a major concern when studying eGFR decline, because the CKD-EPI formula includes not only sCR levels but also age and sex (12). Moreover, we did not find differences in the risk of MAKE between the diabetic and non-diabetic populations, which could be explained by the high comorbidity of the whole cohort, which had a median Charlson Comorbidity Index of 4.

Providing information about AKI episodes is key not only to attempts to change the natural history of AKI to CDK transition, but also to the introduction of strategies that identify patients at increased risk to determine which patients may benefit from a nephrology or primary-care follow-up. In that regard, patients in which sCr levels do not return to baseline levels at discharge could be considered candidates for specialist follow-up. However, we must take into account that hyperfiltration after AKI, changes in distribution volume, and loss of muscle mass during long hospital admissions may also decrease creatinine values, and therefore that a large percentage of patients could be lost (13, 14). Interestingly, low sCr levels have been associated with higher mortality rates as a result of malnutrition. On the contrary, when using cystatine C, a biomarker that is independent of muscle metabolism, there is a linear rather than a U–shaped association between eGFR and adverse events (13). The use of cystatin-C may not always be possible, but the measurement of creatinine clearance could be a way to identify patients with persistent kidney dysfunction after CSA-AKI (15). Studies of biomarkers in AKI have mainly been conducted by intensivists and have focused on short-term outcomes. In this field, only a secondary analysis of the Sapphire study for NephroCheck^® ([TIMP-2]×[IGFBP7]), known as the cell cycle arrest biomarker, showed that a result of >2 was equivalent to AKI stage progression on the risk of ESKD or death at 9 months (16).

Tools such as the CSA-CKD scoring system developed by Legouis et al. show promise as simple ways to identify patients at risk of kidney disease progression (7). In our study, we first tried to assess if the score had fair discrimination for CKD, as was first described in its original study. We found that the AUC value for CKD in patients without pre-existing CKD was 0.737 (95% CI 0.657–0.817), similar to the validation cohort of the original study (0.78 [95% CI 0.72–0.83]). On the other hand, when analyzing AUC for MAKE in the overall population, the AUC value decreased to 0.682 (95% CI 0.611–0.754). In our study we used the CKD-EPI formula, since it is currently the formula with the most international endorsement. Legouis et al. used the MDRD formula for the estimation of basal GFR in patients without CKD, but it has been proven that this formula has worse precision for eGFRs of 60–90 mL/min, and in that scoring system patients received 1 point for eGFR < 80 ml/min. We believe that multicenter studies are needed to create a new scoring system that focuses on MAKE and uses CKD-EPI as the formula for eGFR estimation (17, 18).

However, after patients at risk of MAKE have been identified, there is still no robust data about the benefits of a specific nephrology follow-up compared to standard care. The first randomized controlled trial investigating this was published in 2021 (19). Patients who survived severe AKI stage 2 or 3 were enrolled and randomized to receive either comprehensive or standard care for 12 months. The comprehensive group comprised a multidisciplinary team that included nephrologists, nurses, nutritionists, and pharmacists. The primary outcome was feasibility and the secondary outcomes included incidence of MAKE, renal function, and albuminuria rate at 12 months. They accomplished the primary feasibility outcomes; for the secondary outcomes they found statistically significant differences only in albuminuria rate. However, blood pressure was better controlled in the comprehensive group.

Our study has some limitations. First, the albuminuria rate was not assessed because data were not available. It is known that post-AKI proteinuria is associated with kidney disease progression and, even in patients without changes in eGFR at 1 year, it is considered a sequela of AKI (20). Second, owing to the nature of this study, almost 7% of the original cohort were lost to follow-up and therefore could not be included in the final analysis. Third, the lack of association between CSA-AKI and new AKI episodes could also be due to the nature of the study, as only 14 new AKI episodes were registered because of the short and retrospective follow-up. Finally, this is a unicentric retrospective study that provides information about the increased risk of MAKE after CSA-AKI, but multicentric and prospective studies are needed to confirm our results and create a scoring system that tries to identify patients at risk of MAKE.

In conclusion, based on our results, any-stage CSA-AKI is associated with MA; however, the development of further scoring systems that help clinicians to identify at-risk patients is needed so that appropriate patient follow-up can be provided.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The Ethics Committee of our institution approved the study (Reg. HCB/2019/0959). Written informed consent from the patient or patient’s legal guardian/next of kin was not required to participate in this study in accordance with the national legislation and the institutional requirements.

Research idea and study designs: AM and EP. Data acquisition: AM, VE, and AL. Data analysis and interpretation: AM. Supervision and mentorship: EP. IR, PM, CI, MB, ES, LQ, MC, GP, and EQ contributed important intellectual content during manuscript drafting. All authors contributed to the article and approved the submitted version.