Determining the Burden of Acute Kidney Injury Using Large Databases

In a recent review, Simon Sawhney, MBChB, University of Aberdeen United Kingdom, and Simon D. Fraser, DM, University of Southampton, United Kingdom, presented a review of recent large database studies focusing on the epidemiology of acute kidney injury (AKI); the researchers also addressed the advantages and drawbacks of using routine health data in observational research. The review focused on three themes: (1) the apparent increasing incidence of AKI; (2) the long-term prognosis of AKI; and (3) the role of large databases in quality of AKI care. The review was published in Advances in Chronic Kidney Disease [2017;24(4):194-204] on behalf of the National Kidney Foundation, Inc.

AKI accounts for 8% to 16% of hospital admission and a 4-fold increase in hospital mortality. Over the past decade, the term acute renal failure has been replaced with standardized AKI criteria that include small changes in creatinine and urine output, such as the Acute Kidney Injury Network (AKIN) and the Kidney Disease: Improving Global Outcomes (KDIGO) criteria.

The International Society of Nephrology has established the 0by25 initiative, which calls for prevention of all avoidable death from AKI worldwide by 2025. The society conducted a global study in 2014 to create a snapshot of the burden of KDIGO-based AKI across 289 centers and 72 countries. In that study, 7-day mortality was 10% to 12% in both high- and low-income countries. Key etiologic risk factors varied among countries, but included dehydration, shock, infection, sepsis, cardiac disease, and nephron-toxic medications.

Other studies have used International Classification of Diseases (ICD) coding of hospital episodes to quantify the incidence of AKI in various clinical settings. Results consistently demonstrate an increasing incidence of AKI over time. In the United States, Canada, and the United Kingdom, the rates of hospital coding of non–dialysis-requiring AKI have increased over a combined period of 25 years. Studies from general hospital settings have demonstrated a concurrent decline in AKI mortality.

According to the authors, increased coding of AKI may be due to changes in coding practice more generally or increased awareness of AKI, rather than reflecting a true change in AKI incidence. Further, in the context of decreased mortality rates, increased recognition of mild AKI may explain the decrease. “However,” they add, “it seems unlikely that an up to 10-fold increase in AKI incidence is attributable purely to changes in ICD coding.”

There have also been increases in AKI requiring renal replacement therapy (RRT) over the past 25 years in America and Europe in a variety of settings, including general hospital admissions, elective surgery, cardiac surgery, liver disease, coronary interventions, infectious diseases, obstetrics, sepsis, and stroke. Reasons for the increases are uncertain, but may include increased availability of RRT, increased recording of RRT use, financial incentives, or evolving trends in medical practice. However, the authors noted that in countries such as Denmark and the United Kingdom, with health care systems based on taxation, AKI-related RRT rates have also increased, casting doubt on the theory that financial incentives are the sole explanation.

There are few available data on the incidence of AKI based on creatinine change criteria. Results of three studies with definitions of AKI based on creatinine have reported an increase in AKI incidence in United States in databases from Kaiser Permanente, the Cleveland Clinic, and a national intensive care database that covers Australia and New Zealand. The changes in incidence were smaller than in studies based on code classification of AKI. Further, one study in the United States based on AKI criteria following myocardial infarction reported a decrease in AKI incidence and mortality; the decrease was attributed to improved awareness among clinicians, improved care and prevention of AKI, and may also have been related to case selection in the study.

Systemic reviews of more than 50 studies support linking AKI with poor long-term outcomes, including increased mortality, chronic kidney disease, and long-term RRT. Based on those studies, KDIGO AKI guidelines call for follow-up of patients following hospital discharge. In addition, there have been several studies designed to assess outcomes of AKI within subgroups of baseline kidney function, severity of AKI, time to renal recovery, and time to follow-up.

Results of those studies provide evidence of the association between greater mortality and adverse renal outcomes after AKI with baseline estimated glomerular filtration rate, previous episodes of AKI, the extent of renal recovery, and follow-up time. Stratification of those at increased risk can be informed by those key factors, allowing for sufficient monitoring of those at greatest risk.

In summary, the authors of the current review said, “Large database studies show that AKI pervades health care systems worldwide, with poor outcomes and an increasing use of RRT. Given the high incidence of AKI hospital survivors, prognostic risk stratification on AKI severity, baseline, recovery to baseline, and follow-up time may be helpful to improve post-AKI care, but such an approach would need careful evaluation. Temporal and regional comparisons of AKI incidence and prognosis may be biased by changes in AKI awareness, monitoring, recognition, and clinical practice. For this reason, researchers monitoring quality initiatives using observational data should be careful to capture all who may be affected by policy changes and ensure that both intended and unintended consequences are recognized.”

Takeaway Points

  1. A review of recent large database studies that have been designed to further understanding of the incidence and long-term prognosis of acute kidney injury (AKI).
  2. The incidence of AKI has been increasing over the past 25 years; studies assessing AKI incidence have been conducted based on International Classification of Diseases coding and on models based on creatinine change criteria.
  3. Key factors associated with AKI incidence include severity of AKI, baseline kidney function, and time to renal recovery, all of which should be considered as prognostic risk indicators.