Acute kidney injury

Late-onset acute kidney injury—subacute or more of the same?

Fujii and colleagues define 'subacute' kidney injury (s-AKI) as AKI that takes >7 days to develop, timed from admission or lowest creatinine measurement after admission. Although s-AKI is unlikely to be a distinct syndrome from AKI, the association with increased mortality highlights the need to monitor patient creatinine levels.

Consensus definitions of acute kidney injury (AKI) have evolved over the past 12 years. Initially, AKI was characterized by a rapid decline in glomerular filtration rate (GFR), identified by increased levels of creatinine within 7 days or oliguria, but GFR was later removed as a criterion. Now, AKI criteria state that it must develop within 48 h of a precipitating event and includes an absolute increase in creatinine levels (26.5 µmol/l) to deal with equivalent injury at different baseline creatinine values, as even small increases in creatinine are associated with increased mortality.1 The KDIGO (Kidney Disease Improving Global Outcomes) definition of AKI brought the RIFLE (Risk, Injury, Failure, Loss and ESRD [end-stage renal disease]) and AKIN (Acute Kidney Injury Network) definitions together as the two identify different patients in the mildest category of AKI who have increased mortality risk.2 The ADQI (Acute Dialysis Quality Initiative) group has now recommended the inclusion of biomarkers of kidney damage into the definition,3 as it is clear from retrospective4 and prospective5 studies that patients with increased levels of biomarkers of kidney damage, such as urinary neutrophil gelatinase-associated lipocalin (NGAL) and kidney injury molecule 1 (KIM-1), but without increased levels of creatinine (termed as biomarker-positive, creatinine-negative), are also at increased risk of mortality and dialysis. Thus, AKI is best viewed as a continuum of injury of variable duration and severity.6

Fujii and colleagues7 describe the epidemiology of a cohort of patients, designated as having 'subacute' kidney injury (s-AKI), who developed AKI after admission to hospital or after an arbitrary baseline value over 8–90 days. In a retrospective study of 56,567 patients admitted to a single teaching hospital in Japan, 1.1% of admissions had s-AKI, compared with 11.0% who developed AKI within 7 days (defined by the RIFLE criteria). Patients with s-AKI had less severe kidney injury than those diagnosed with AKI, as evidenced by levels of serum creatinine, but demonstrated increased rates of in-hospital mortality compared with patients without AKI (shown by multivariable regression analysis). Patients with AKI were at even greater risk of in-hospital mortality when compared with patients with s-AKI. When s-AKI was categorized according to severity, (R, I, or F of the RIFLE criteria), a linear correlation with mortality was observed. No data were provided on the duration of s-AKI, which is probably a more critical link with mortality than incidence.8

Credit: © ronstik/iStock/Thinkstock

What does expansion of the detection window for AKI beyond 7 days do? Limiting increases in creatinine levels to within 48 h favours identification of AKI following severe injury, such as after cardiopulmonary bypass or renal transplantation. Expanding the window to 7 days enables the identification of AKI that develops after the accumulation of less serious injuries, which incrementally or together decrease GFR, for example after repeated nephrotoxic drug exposure or the cytokine storm of sepsis. As such, removing the restriction of AKI to within 48 h after a potential initiating event is logical. However, the majority of causes of AKI (including sepsis, ischaemia–reperfusion injury, and toxins) do not cause AKI to only be detectable after 7 days. Longer time frames for the development of AKI are not consistently observed and the frequency of s-AKI (1.1%) was much lower than that of AKI (11.0%), casting doubt on the importance of stratifying AKI beyond 7 days as suggested by Fujii and colleagues.7

Rapidly progressive glomerulonephritis (RPGN), for example in Goodpasture syndrome or antineutrophil cytoplasmic antibody (ANCA)-associated disease, produces progressive glomerular obliteration and often irreversible renal failure within days to weeks. RPGN is part of the differential diagnosis of AKI.9 However, the onset of an episode of glomerulonephritis or other genuinely progressive conditions is difficult, if not impossible, to define and defining an appropriate window for exclusion is, therefore, difficult. Consistent with such projections, the retrospective analysis conducted by Fujii and colleagues7 showed that patients with s-AKI were less likely than those patients with AKI to be admitted to emergency and surgical departments, and more likely to be admitted by other medical departments, such as oncology and psychiatry, as might be expected if patients with s-AKI mostly present with toxicant and sepsis-induced AKI. Even in the nephrology department, where glomerulonephritis-induced renal failure is commonly found, AKI incidence predominated over s-AKI. As this retrospective study was essentially data-mining, the causes of AKI and s-AKI in these patients were not reported.

The longer the duration allowed for detection, the more difficult it is to link development of AKI—defined either by a functional change or by levels of biomarkers of kidney damage—with cause. Although setting data-mining parameters for creatinine change to beyond 7 days enables 'discovery' of an apparently unrecognized cohort, the question is whether this cohort is really different to one already defined by the KDIGO criteria, for which the start time is ignored or allowed to shift. The cut-off of 7 days, and skewed distribution of hospital length of stay, causes the frequency distribution to skew to 8 days. Is an 8-day episode (or a 10-day or 12-day episode) really different from a 2-day episode when both are in the same RIFLE class? After all, the assumption is that injury has started at the beginning of the window period, not just in the few days during which time the increase in creatinine level has become apparent. Furthermore, why would an episode of AKI occurring at day 89 be the same as one occurring at day 8?

The measurement of baseline creatinine level within 1–12 months of admission was available in only 43% of patients with s-AKI; the remainder were evaluated using backcalculation assuming a GFR of 75 ml/min/1.73 m2, a pragmatic approach that systematically overestimates the incidence of AKI.10 This approach might explain the lower mortality risk (8.0% versus 17.5%) and lower dialysis risk (0.2% versus 2.2%) associated with s-AKI than with AKI.7 Alternatively, the later AKI develops after admission, the less likely it is to follow the same cause that precipitated admission, or perhaps AKI developing and managed within-hospital has a lower mortality and dialysis risk than AKI initiated before hospital admission.

Thus, the evidence suggests that the designation of patients in this study as 'subacute' is inappropriate. The researchers in this study have captured the 9% of patients with AKI (1.1%/[11.0%+1.1%]) who develop AKI over a period longer than 7 days after admission7 and more-detailed analysis will probably determine that the majority of these cases were caused by nephrotoxins or sepsis. A few of the patients might have RPGN. 'Subacute' implies an ongoing disease process over a prolonged period, for example, subacute bacterial endocarditis with ongoing infection and its sequelae. However, little evidence demonstrates that the patients identified as having s-AKI had an ongoing AKI process. We are currently unable to time most critical renal injury events. For example, in toxicant-AKI following a known nephrotoxin given more than once, it is impossible to know whether it is the first, last or the cumulative dose of nephrotoxin that precipitates AKI. Earlier identification of these patients by measuring biomarkers of kidney damage will enable prompter intervention or cessation of an offending agent and perhaps even identify the phase and mechanism of renal cellular injury. Although identification of patients with late-onset AKI is important, as all patients with AKI have an increased mortality risk, identifying these patients as having s-AKI does not add pathophysiological insight and might falsely imply that continuing progressive renal injury is common. Identification of the duration of AKI and long-term follow-up monitoring of these patients is required to understand the implications of late-onset AKI. What this study does highlight is that, regardless of the defining window, all patients with AKI need to be identified, the cause established and the condition treated.


  1. 1

    Chertow, G. M., Burdick, E., Honour, M., Bonventre, J. V. & Bates, D. W. Acute kidney injury, mortality, length of stay, and costs in hospitalized patients. J. Am. Soc. Nephrol. 16, 3365–3370 (2005).

    Article  PubMed  Google Scholar 

  2. 2

    KDIGO. Clinical practice guideline for acute kidney injury. Section 2: AKI definition. Kidney Int. Suppl. 2, 19–36 (2012).

  3. 3

    Murray, P. T. et al. Current use of biomarkers in acute kidney injury: report and summary of recommendations from the 10th Acute Dialysis Quality Initiative consensus conference. Kidney Int.

  4. 4

    Haase, M. et al. The outcome of neutrophil gelatinase-associated lipocalin-positive subclinical acute kidney injury a multicenter pooled analysis of prospective studies. J. Am. Coll. Cardiol. 57, 1752–1761 (2011).

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  5. 5

    Nickolas, T. L. et al. Diagnostic and prognostic stratification in the emergency department using urinary biomarkers of nephron damage: a multicenter prospective cohort study. J. Am. Coll. Cardiol. 59, 246–255 (2012).

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  6. 6

    Nejat, M. et al. Some biomarkers of acute kidney injury are increased in pre-renal acute injury. Kidney Int. 81, 1254–1262 (2012).

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  7. 7

    Fujii, T., Uchino, S., Takinami, M. & Bellomo, R. Subacute kidney injury in hospitalized patients. Clin. J. Am. Soc. Nephrol.

  8. 8

    Brown, J. R., Kramer, R. S., Coca, S. G. & Parikh, C. R. Duration of acute kidney injury impacts long-term survival after cardiac surgery. Ann. Thorac. Surg. 90, 1142–1148 (2010).

    Article  PubMed  Google Scholar 

  9. 9

    Endre, Z. H. et al. Differential diagnosis of AKI in clinical practice by functional and damage biomarkers: workgroup statements from the tenth Acute Dialysis Quality Initiative Consensus Conference. Contrib. Nephrol. 182, 30–44 (2013).

    Article  PubMed  Google Scholar 

  10. 10

    Pickering, J. W. & Endre, Z. H. Back-calculating baseline creatinine with MDRD misclassifies acute kidney injury in the intensive care unit. Clin. J. Am. Soc. Nephrol. 5, 1165–1173 (2010).

    CAS  Article  PubMed  PubMed Central  Google Scholar 

Download references

Author information



Corresponding author

Correspondence to Zoltán H. Endre.

Ethics declarations

Competing interests

The authors declare no competing financial interests.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Endre, Z., Pickering, J. Late-onset acute kidney injury—subacute or more of the same?. Nat Rev Nephrol 10, 133–134 (2014).

Download citation


Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing