Severe Acute Kidney Disease Is Associated With Worse Kidney Outcome Among Acute Kidney Injury Patients Ⅱ

Acute kidney disease (AKD) comprises acute kidney injury (AKI). However, whether the AKD staging system has prognostic values among AKI patients with different baseline estimated glomerular filtration (eGFR) remains a controversial issue. The algorithm-based approach was applied to identify AKI occurrence and to define different AKD stages. The risk ratio for major adverse kidney events (MAKE), including (1) eGFR decline> 35% from baseline, (2) initiation of dialysis, and (3) in-hospital mortality of different AKD subgroups were identified by multivariable logistic regression. Among the 4741 AKI patients identified from January 2015 to December 2018, AKD stages 1–3 after AKI was common (53% in the lower baseline eGFR group and 51% in the higher baseline eGFR group). In the logistic regression model adjusted for demographics and comorbidities at 1-year follow-up, AKD stages 1/2/3 (AKD stage 0 as the reference group) were associated with higher risks of MAKE (AKD stage: odds ratio, 95% confidence interval [95% CI], AKD 1: 1.85, 1.56–2.19; AKD 2: 3.43, 2.85–4.12; AKD 3: 10.41, 8.68–12.49). Regardless of baseline eGFR, staging criteria for AKD identified AKI patients who were at higher risk of kidney function decline, dialysis, and mortality. Post-AKI AKD patients with a severe stage need intensified care and timely intervention. 

A NEW HERB FOR CURE KIDNEY DISEASE

Acute kidney injury (AKI) occurs in approximately 10–15% of hospitalized patients; however, its incidence in the intensive care unit can be more than 50% of patients1. Previous meta-analyses and systematic reviews have demonstrated that AKI might significantly increase the risk and progression of chronic kidney disease (CKD) as well as the risk of end-stage kidney disease (ESKD) and even mortality2–6. However, a recent prospective study determined that AKI stages were not independently associated with adverse clinical outcomes after adjustment for multiple CKD risk factors7. Acute kidney disease (AKD), as an intermediary stage between AKI (abrupt deterioration of renal function within 7 days or less) and CKD (persistent renal function impairment or structural abnormalities for more than 90 days), indicates changing renal function that could be strongly associated with long-term renal outcomes and could serve as a valuable indicator of the appropriate time window for therapeutic interventions2,8,9. A large population-based observational study discovered a statistically signifcant association between AKD and major adverse kidney events (MAKEs) related to CKD progression, ESKD, and mortality10. With or without AKI, AKD was also found to be associated with adverse clinical outcomes among hospitalized patients11.

1 Division of Nephrology, Department of Internal Medicine, Shuang Ho Hospital, Taipei Medical University, No. 291, Zhongzheng Road, Zhonghe District, New Taipei City 235, Taiwan. 2Graduate Institute of Clinical Medicine, College of Medicine, Taipei Medical University, Taipei, Taiwan. 3TMU Research Center of Urology and Kidney, Taipei Medical University, Taipei, Taiwan. 4 Institute of Epidemiology and Preventive Medicine, College of Public Health, National Taiwan University, Taipei, Taiwan. 5Division of Nephrology, Department of Internal Medicine, School of Medicine, College of Medicine, Taipei Medical University, Taipei, Taiwan. 6Information Technology Ofce, Shuang Ho Hospital, Taipei Medical University, Taipei, Taiwan. 7Center for Neuropsychiatric Research, National Health Research Institutes, Miaoli County, Taiwan. 8Department of Anesthesiology, Shuang Ho Hospital, Taipei Medical University, New Taipei City, Taiwan. 9Department of Anesthesiology, School of Medicine, College of Medicine, Taipei Medical University, Taipei, Taiwan. 10 These authors contributed equally: Yih-Giun Cherng and Mai-Szu Wu. *Email: stainless1019@gmail.com; maiszuwu@gmail.com

Nevertheless, the disparate definitions of AKD might indicate variance in the acuity and severity of kidney disease among different subgroups of AKD patients12. Persistent kidney dysfunction for more than 7 days after AKI (post-AKI AKD) is associated with adverse renal and cardiovascular outcomes11,13–17. The 16th Acute Disease Quality Initiative (ADQI) recommends definitions and staging criteria for post-AKI AKD and renal recovery, but the clinical trajectories of renal function after AKI and the prognostic value of AKD staging remain largely unknown2,18. Therefore, to explore the evolutional pattern of renal function following AKI, we obtained all of the available serum creatinine (SCr) data from a single tertiary hospital to define the incidence and clinical outcomes of different stages of post-AKI AKD. 

Methods

Data source. 

We retrieved laboratory and administrative data from the health information system database of a single tertiary hospital to conduct a retrospective cohort study. The Joint Institutional Review Board of Taipei Medical University (TMU-JIRB-N201906062) approved the present study and waived the need for informed consent because all data had been de-identified. The present study was conducted by the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines19.

Cohort information. 

All available SCr data of patients older than 20 years from January 1, 2014, to December 31, 2019, were statistically analyzed. Patient SCr data from January 1, 2015, to December 31, 2018, were frst grouped into per capita data as an observational cohort. SCr data from January 1, 2014, to December 31, 2014, were all used as the source of the baseline SCr data pool (see "Definition and staging of AKI, AKD, and baseline eGFR classification"). SCr data and other laboratory data from January 1, 2019, to December 31, 2019, were included to ensure that the observational period of each patient with AKI identified by the algorithm would be equal to or more than 1 year. In addition, to explore the post-AKI prognostic value and clinical effect, we defined the date of AKI occurrence as the index date. Patients who did not develop AKI or who received their frst dialysis before the index date were excluded. Moreover, patients younger than 20 years and those with missing information regarding date of birth or sex were excluded. Patients without data after AKI, without data within 7–90 days after AKI, and those without data beyond 90 days after AKI were also excluded to delineate the evolutionary changes of renal function. Underlying comorbidities, such as hypertensive heart disease, diabetes mellitus, anemia, cerebrovascular disease, malignancy, chronic obstructive pulmonary disease, and digestive tract disease, were identified using ICD-9-CM codes if they had been used one or more times in inpatient diagnoses or one or more times for outpatient diagnoses within 1 year before the occurrence of AKI (determined according to the creatinine index [C1]). Furthermore, to evaluate the possible prognostic effects of electrolytes, we averaged all available serum sodium and potassium data within 3 months before the occurrence of AKI and used the mean values as bio-chemical factors in our model.

Definition and staging of AKI, AKD, and baseline eGFR classification. 

We defined AKI based on the 2014 NHS England "algorithm for detecting acute kidney injury (AKI) based on serum creatinine changes with time" (Fig. S1) to allow international comparisons with other studies on AKI electronic alert systems. The algorithm complied with the Kidney Disease: Improving Global Outcomes (KDIGO) guidelines regarding the definition of AKI based on SCr values. In this algorithm, C1 values served as the frst input data. The lowest SCr values within 0 to 7 days before C1 were defined as the baseline creatinine data (RV1). If RV1 was not available, then the median value of all the available SCr data within 8–365 days before the C1 values were obtained was used as the baseline SCr (RV2). The index SCr-to-baseline SCr ratio (C1/RV1 or C1/RV2) was calculated to define AKI. Once AKI was defined, the creatinine data with the maximum values between the 7th and 90th day after AKI were retrieved to calculate the AKD stages by using another algorithm (Fig. 1). The AKD staging algorithm was based on the 16th ADQI consensus2, which is congruent with AKI staging. The patients were then categorized according to the worst AKD stages into AKD stage 0 (reference group) and AKD stage 1–3 (comparison groups). Baseline eGFR classification was defined based on the estimated glomerular filtration rate (eGFR) derived from the reference SCr (RV1 or RV2), calculated using the Chronic Kidney Disease Epidemiology Collaboration equation20. To determine the clinical significance of AKI and AKD among the patients with different baseline eGFR, we categorized our cohort into two groups: "higher baseline eGFR group" and "lower baseline eGFR group" who had baseline eGFR ≥60 mL/min/1.73 m2 and baseline eGFR<60 mL/min/1.73 m2, respectively.

Statistical analysis. We compared baseline characteristics according to the baseline eGFR classification using a t-test for continuous variables and Ca hi-squared test for categorical variables. We used a logistic regression model to examine the prognostic effects of clinical indicators, including age, sex, comorbidities, baseline eGFR  

Figure 1. Algorithm for the staging of acute kidney disease. The algorithm for the staging of acute kidney disease (AKD), is based on the 16th Acute Disease Quality Initiative (ADQI) consensus, which is congruent with AKI staging. SCr serum creatinine, RV reference value. Patients with AKD stage 0 by this algorithm were considered as the reference group in the following statistical analysis. Conversion factors for serum creatinine in mg/dL to μmol/L,×88.4.

Ethics approval. 

The study was approved by the Joint Institutional Review Board of Taipei Medical University (TMU-JIRB-N201906062) and the need for informed consent was waived because all data had been de-identified.