Relation of Serum Creatinine Twitches and Outcomes among STEMI Patients Open Access

Acute kidney injury (AKI) is a common complication in different acute conditions and in particular among patients with ST-segment-elevation myocardial infraction (STEMI) [1]. Several studies found STEMI patients to be more prone to renal complications due to their hemodynamic instability, contrast media exposure, and broad use of nephrotoxic medications, which are part of common practice for cardiac remodeling prevention [2, 3]. From a wide array of in-hospital compilations, AKI is notoriously associated with adverse outcomes and mortality as well as with prolonged hospitalizations [4, 5]. However, AKI is based on a predefined general definition without adjustment for the specific primary etiology. The common definition used in practice is the rise in serum creatinine (sCr) by 0.3 mg/dL, as defined by the KDIGO guidelines [6]. Using this general cutoff with creatinine might lead to delayed diagnosis and limit the benefit in early treatment [6‒8]. As a result, great efforts are being taken to allow early diagnosis of renal injury; these include a continuous search for risk factors or early renal injury markers [9‒11]. Yet none of these newly devised markers have been introduced into routine use or shown superiority and thus this gap between renal injury occurrence and diagnosis remains.

Recently, some studies offered the use of minor creatinine changes, that do not meet the criteria for AKI diagnosis, as a marker for renal injury and adverse outcomes. The studies targeted minute changes of between 0.1 and 0.4 mg/dL in creatinine, which were associated with long-term outcomes and mortality in specific subset populations [12‒15]. However, most of these studies examined relatively small cohorts and focused on surgical patients. Large-scale data regarding the prognostic value of minor variations in creatinine are scarce although it is worth mentioning a study by Newsome et al. [16] that studied a retrospective cohort of elderly patients with acute coronary syndrome and found that even minor increments in creatinine values (0.1 mg/dL) result in higher risk for end-stage renal disease progression and mortality. In addition, a small prospective cohort tracked patients with heart failure exacerbation and found that small changes in serum creatinine are associated with elevated risk for 6-month mortality and re-admissions [14]. We aimed to assess the effects of minor creatinine changes (“creatinine twitches”) in a large cohort of STEMI patients. Validating their prognostic utility and offering a new cutoff value for future use could hopefully lead to earlier and more accurate prognostication in this at-risk population.

We performed a retrospective analysis of a prospective database, which included all patients that were admitted to the cardiac intensive care unit (CICU) of a tertiary referral hospital with the diagnosis of acute STEMI between January 2008 and January 2022. The hospital (Tel Aviv Sourasky Medical Center) has 24/7 primary percutaneous coronary intervention services and all patients with STEMI are admitted to the hospital’s CICU. Overall, our medical center is one of the largest in Israel, with more than 1,500 hospitalization beds and over 100,000 admissions per year. Patients had to meet the following criteria to be included in the study cohort: (1) included in the main database, (2) at least 3 creatinine measurements, (3) without chronic dialysis use, and (4) have available access to their medical records. Of note, this database was also used for analysis of previous renal outcomes in previously published reports [17].

The diagnosis of STEMI was established with typical chest pain, appropriate electrocardiographic changes, and serial elevation of serum cardiac biomarkers [18]. The diagnosis was verified before inclusion to the database. PPCI was carried out in patients within 12 h from the onset of pain or 12–24 h from symptoms onset given ongoing significant pain. Statins, renin/angiotensin blockers, and β-blockers were started in all patients unless contraindicated. Following PCI, 0.9% saline was given for 12 h at a rate of 1 mL/kg/h or lower if patients had an overt heart failure.

Baseline demographic and medical history, treatment characteristics, and laboratory results were included in the database for all patients (n = 2,940). Left ventricular ejection fraction was assessed within the first 48 h by a bedside echocardiography in all cases. In-hospital mortality and complications occurring during the acute hospitalization were obtained using computerized reports. Complications included AKI, cardiogenic shock, intra-aortic balloon pump placement, ventricular tachycardia/fibrillation episodes, severe bleeding (requiring blood transfusion), and stent thrombosis. One-year mortality was examined using the National Population Registry of the Israel Ministry of the Interior, which is updated on a weekly basis.

The study was conducted according to the Declaration of Helsinki and approved by the Tel Aviv Sourasky Medical Center review board (TLV-16-0224). Informed consent was obtained from all subjects involved in the database.

Serum creatinine is regularly evaluated in our center upon patient admission prior to primary PCI and at least once daily at each morning during hospitalization. Estimated glomerular filtration rate (eGFR) was calculated using the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation and a presenting eGFR of under 60 mL/min/1.73 m² was determined for chronic kidney disease (CKD) definition [19]. AKI was defined according to the “Kidney Disease: Improving Global Outcomes (KDIGO)” criteria [20]. In summary, an increase in serum creatinine by >0.3 mg/dL in 2 days or by >50% in 1 week defines AKI. Minor serum creatinine changes (sCr twitches) were defined as a change from baseline to peak in-hospital levels that is higher than 0.1 mg/dL and lower than 0.3 mg/dL. These values were chosen to account for changes below the threshold for AKI definition yet within the laboratory detection ability.

For our analyses, the cohort was divided into 3 groups based on in-hospital kidney function, and they included AKI, sCr twitches, and stable creatinine. Patients’ characteristics and outcomes were compared each time between two groups, with emphasis on creatinine twitches. Continuous variables were presented as median (interquartile range) given that all had a non-normal distribution (assessed with the Kolmogorov-Smirnov test) and compared by Mann-Whitney U tests. Categorical variables were presented as sum (percentage from total) and compared by chi-square tests. Mortality was presented as a total rate in 30 days and 1 year and as a time-dependent variable by Kaplan-Meier curves. Cox multivariate regression analysis was used to assess the prognostic value of creatinine twitches. All variables included in this model had a variance inflation factor lower than 2 between them.

Receiver operating characteristic curve analysis was used to evaluate the predictive ability of in-hospital raise in creatinine for 1-year mortality. The optimal predictive cutoff for creatinine raise was chosen by calculating a Youden’s index (J index) for each result. Sensitivity, specificity, positive predictive value, and negative predictive value were calculated for the prediction of 1-year mortality by the new cutoff. We performed additional sub-analyses. First, subjects with creatinine twitches were divided to those who improved to within 0.1 mg/dL from first creatinine and those who did not. Second, the cohort was divided based on the baseline eGFR, using the accepted cutoff values for GFR categories of chronic kidney disease [21] – i.e., >90 mL/min, 90–60, 60–45, <45. All analyses were performed with SPSS software version 28.0 (SPSS Inc., Chicago, IL, USA), with p value <0.05 determined for significance.

During our study period, a total of 2,940 patients were admitted to the coronary intensive care unit and included in the database; of them, 7 were on chronic dialysis and excluded from our study. Overall, 2,128 patients (72%) had stable levels of sCr during their hospitalization, 551 (19%) patients were found to have creatinine twitches during hospitalization, and 254 patients (9%) were diagnosed with AKI. Baseline characteristics are shown in Table 1 and compared against stable creatinine or AKI. Subjects with creatinine twitches were older, had higher prevalence of hypertension and atrial fibrillation compared to those with stable creatinine. They also presented with longer time to reperfusion, diffuse coronary disease and suffered hemodynamic instability or malignant arrhythmias. These changes were vice versa when subjects with creatinine twitches were compared to those with AKI.

Compared to the stable group, sCr twitches were associated with prolonged hospital stay (median 4 [4, 5] vs. 5 [4‒6] days, Fig. 1), although it was shorter than the AKI group (median 7 [5‒11] days, Fig. 1). In addition, subjects with sCr twitches had higher mortality rates compared to the stable group in both 30 days (1% vs. 2.5%) and 1 year (1.6% vs. 4.4%) periods, as seen in Figure 2. sCr twitches were found to have higher hazard for 1-year mortality compared with stable creatinine (HR 2.76, 95% CI: 1.64–4.65, p < 0.001, Fig. 3). This association remained significant in multivariate cox regression model (HR 1.87, 95% CI: 1.09–3.2, p = 0.023, Table 2), alongside older age, ionotropic use, and the presence of AKI or CKD. Similar to 1-year mortality, sCr twitches had higher hazard for 30-day mortality (HR 1.90, 95% CI: 1.12–3.23, p = 0.016), although it did not remain statistically significant in multivariate analysis (HR 1.52, 95% CI: 0.96–2.96, p = 0.09).

When attempting to evaluate the predictive value of sCr twitch, change between peak to first sCr had an AUC of 0.780 (95% CI: 0.73–0.83) for the prediction of 1-year mortality (Fig. 4). Based on receiver operating characteristic analysis, the optimal creatinine increase as a cutoff to predict 1-year mortality was 0.12, with a J-index of 0.50. This cutoff had a sensitivity of 71%, specificity of 79%, positive predictive value of 14%, and negative predictive value of 98%.

We assessed the effect of twitch resolution throughout CICU stay. Twitches resolved prior to discharge (decreased to a value of up to 0.1 from first value) in 362 patients (66%). Patients with resolved twitches still had a higher hazard for mortality compared with stable creatinine (HR 2.44, 95% CI: 1.31–4.55, p = 0.005). However, in multivariate cox analysis, only patients with unresolved twitches continue to have a higher hazard for mortality (Table 3).

In a second subgroup analysis, we divided the cohort based on baseline eGFR values (online suppl. Table S1; for all online suppl. material, see https://doi.org/10.1159/000545523). sCr twitch was associated with a higher hazard of mortality in all groups except those with eGFR >90 mL/min. In multivariate analysis, adjusted for age, sex, and requirement of ionotropic medications, sCr twitch remained associated with mortality in patients with baseline eGFR <45 mL/min (HR 2.76, 95% CI: 1.05–7.92, p = 0.02), with a trend in eGFR 60–90 mL/min. Of note, rates of 1-year mortality were significantly fewer in patients with eGFR >45 than those without (3% vs. 22%), possibly affecting the strength of this sub-analysis. To further evaluate the interactions between baseline variables and the above associations, we performed a subgroup interaction analysis (online suppl. Table S2). A higher hazard for mortality in patients with sCr twitches consisted of most subgroups, other than in patients younger than age 65, those without hypertension, and with past myocardial infraction. Finally, we analyzed the prognostic utility of additional levels of AKI (online suppl. Table S3). Hazard for 1-year mortality increased with the increasing severity of AKI.

In this study, we demonstrate that small elevations of sCr (twitches), while not fulfilling the consensus criteria for AKI, were independently associated with an increased risk for adverse outcomes, including prolonged hospitalization and mortality. Resolution of these twitches and return to baseline were also found associated with mortality yet not upon multivariate regression. We also found a high prognostic utility for creatinine increments, with a new cutoff for 1-year mortality prediction among this unique patient population.

AKI is a common ICU complication that is associated with adverse outcomes in all patients, particularly those with STEMI [1, 14, 22, 23]. Clinical AKI is diagnosed when renal damage and dysfunction reach a threshold sufficient for creatinine rise above 0.3 mg/dL [6]. Such level of renal damage or dysfunction becomes evident only after significant structure and function of nephrons are affected (loss of renal functional reserve). It is estimated that relatively high amount of kidney function may be lost before significant creatinine rise is observed [7, 24]. Furthermore, serum creatinine half-life is quite long and its variations over time are slow. As a result, steep rise in serum creatinine that merits AKI definition may be a sign of ongoing renal damage for hours or days [8, 25]. In any case of a subtle and transient decrease in eGFR, creatinine may not rise in a significant manner and therefore may not accurately mirror renal function. Recent data suggested that even a slight increase in creatinine, although still under the critical value of 0.3 mg/dL, is associated with adverse outcomes [26, 27].

Previous reports found small changes in postoperative serum creatinine levels to be associated with increased mortality in subsets of patients. An example of such report can be seen in Lassnig et al. work, where minor creatinine changes following cardiac surgery were associated with mortality [15]. However, in this report, the level of creatinine elevation was up to 0.5 mg/dL, not within the consensus criteria used today. Another study in post-cardiac surgery patients described that an increase in creatinine that did not meet current AKI criteria was independently associated with 30-day all-cause mortality (OR 3.93, 95% CI: 1.68–9.22, p < 0.01). This report is consistent with our findings, yet their augmented effect may be attributed to a greater initial renal insult (bypass cardiac surgery) compared to our STEMI patients [28]. It is also worth mentioning that in this report, the effect was evident mainly in patients without preexisting renal disease, whereas in our report the effect of subclinical AKI was independent of chronic kidney disease.

A study by Smith et al. [14] targeted heart failure patients with minor changes in creatinine (0.1–0.5 mg/dL). The study describes a stepwise increase in mortality risk with every 0.1 rise in sCr, reaching 1.91 for 0.3 mg/dL, similar to our results. In addition, the authors attempt to establish an optimal cutoff for prediction of adverse outcomes and found similar values to our cohort with optimal values yielding 75% sensitivity and 79% specificity. Another study in patients after cardiac arrest also described a 0.19 mg/dl increase as an optimal cutoff value for predicting mortality and neurological damage [29] (p. 24). This great resemblance between the results suggests that the mechanism involved in minor renal injury for STEMI patients might be more similar to advanced heart failure than to cardiac surgery.

Although our study was not aimed to evaluate the pathophysiologic processes and mechanisms, we hypothesize global tissue hypoperfusion may be the root cause for serum creatinine elevation. Accordingly, STEMI patients tend to present with hemodynamic fluctuations related to the acuity of their presentation and thus may increase susceptibility to both contrast-induced AKI and direct damage related to renal hypoperfusion [30]. In our cohort, patients with creatinine twitches had longer symptom duration, ionotropic use, and malignant arrhythmias, all previously shown to be associated with increased risk for AKI in STEMI patients [31].

Our findings bear some important clinical implications. Using present criteria and practicing small subclinical changes in sCr are often neglected. Given our findings, minor changes in sCr (0.1–0.3 mg/dL, sCr twitch) may be used to identify susceptible patients in relatively early stage of organ dysfunction or tissue hypoperfusion. These novel findings should highlight the importance to further investigate serum creatinine as a sensitive and early prognostic marker for this patient population. The Acute Dialysis Quality Initiative (ADKI) proposed a new perspective regarding the diagnosis of AKI that relies on novel biomarkers with no elevation in sCr or reduced urinary output [32]. Similarly, the entity of “subclinical AKI” has recently been introduced and used for patients who display a slight increase in creatinine, which is below the critical value of 0.3 mg/dL [26]. In a practical sense, these creatinine twitches can be used as an early marker for renal injury alongside novel biomarkers that are being evaluated and together form a bundle for early renal injury detection [33] and become sufficient to trigger interventions and protective strategies for the kidneys.

Our study has several limitations; first, being a single center study, the generalizability of our results may be limited. Second, creatinine levels were the only measurement used to assess renal function and urine output was not available to analysis, a fact that may cause an underestimation of renal injury. In addition, mortality cause was unavailable and could have added another interesting aspect to our results. Moreover, given the retrospective nature of our analysis some data were not available for analysis including medication changes and specifically nephrotoxic agents; the treatment plan for each patient was guided by the CICU protocols, but a degree of variance can be assumed. Further studies with prospective follow-up and additional kidney injury biomarkers are needed to establish the clinical significance of creatinine twitches on other long-term outcomes. Another important connection that should be addressed in future studies, but has lacked sufficient evidence so far, is the relationship between body mass index and cardiorenal outcomes. While obesity is well established as a risk factor for cardiovascular disease and the progression of renal disease, the interplay between them has been seldom studied and warrants specific attention due to the fact that high body mass index is associated with hemodynamic changes, renal workload and therefore can predispose renal injury in this subset of patients [34, 35, 36].

In conclusion, sCr twitches are common among STEMI patients and correlate with elevated 30-day and 1-year mortality. Even these minor changes warrant renal protective strategies for early detection and treatment.

Study protocol was reviewed and approved by the institutional Ethics Committee, Approval No. TLV-16-0224. Written informed consent was obtained by writing from all participants. The study was conducted in accordance with the standards as laid down in the 1964 Declaration of Helsinki and its later amendments and approved by the Tel Aviv Sourasky Medical Center review board (TLV-16-0224).

On behalf of all authors, the corresponding author states that there is no conflict of interest.

No funding was received for this study; all authors have nothing to declare.

Shir Frydman: writing – original draft and data curation; Ophir Freund: statistical analysis and writing – original draft; Lior Zornitzki: writing – review and editing; Nevo Barel: data curation; Shmuel Banai: methodology and writing – review and editing; and Yacov Shacham: conceptualization, methodology, and writing – review and editing.

The data that support the findings of this study are not publicly available due to containing information that could compromise the privacy of research participants but are available from the corresponding author, S.F., upon reasonable request.