Until fairly recently, follow-up studies of survivors of critical care have focused, primarily, on short-term outcomes, such as 28-day mortality or length of hospital stay. This is of particular relevance when one considers the effects of AKI. Rather than the outcome being a simple binary endpoint of dialysis or not, or survival or not, the picture is much more complex and worrying than that. If one considers some of the economic burden of the treatment of AKI, for example, this is considerable. Indeed, observational studies demonstrate a significant deterioration of renal function in patients that have survived an episode of AKI with an apparent initial resolution, measured as eGFR. Follow-up from a retrospective observational cohort in 2012 showed progression of renal impairment by 90 days post-index AKI [
7]. Furthermore, a prospective observation study of a cohort of 226 survivors of AKI patients treated with RRT found that 14 % had CKD at the end of the 5-year follow-up period [
5] The long-term effects on renal function are common even where “renal recovery”, as determined by return of eGFR to pre-morbid levels, has occurred. Persistent microalbuminuria for periods as long as 4 years post-AKI have been observed [
31]. The effect of an episode of AKI was elegantly demonstrated by Kolonko et al., who observed that the long-term risk for graft loss was significantly higher among the group of kidneys recovered from donors with AKI than those without (27.8 % vs 7.1 %;
P = 0.02) [
32]. Interestingly, the need for long-term dialysis is also related to AKI. The RENAL study follow-up demonstrated 5.4 % long-term requirement for maintenance dialysis after an episode of AKI requiring RRT, with Schiffl and Fischer reporting a similar incidence of 5 % [
5,
31]. A pooled analysis by Goldberg and Dennen found a wide range of results, depending on patient population, but an average long-term dialysis requirement of 12.5 % in survivors over follow-up periods of 1 to 10 years [
33]. In keeping with a long-term risk for chronic RRT, patients surviving AKI also have an increased rate of both adverse cardiovascular and cerebrovascular events. In a large, matched cohort of over 4000 patients, the incidence of coronary events were 19.8 and 10.3 per 1000 person years in the AKI and the non-AKI group, respectively (HR 1.67, 95 % CI 1.36–2.04), and the incidence of stroke was both higher (HR 1.25;
P = 0.037) and of increased severity in the AKI group [
34]. Following on from these observations, it would seem likely that quality of life (QOL) indices also show a relationship with AKI. A prospective, matched cohort study compared long-term outcome and QOL in patients with AKI requiring RRT compared to matched controls using the EuroQOL-SD and Short-Form (SF)-36 before ICU admission, then at 3 months, 1 year and 4 years after ICU discharge. Surprisingly, despite a higher severity of illness and 28.6 % remaining dialysis-dependent in the AKI-RRT group, the QOL scores were comparable to non-AKI RRT patients. Interestingly, both groups had worse QOL scores than the general public. As expected, the lowest QOL level was at 3 months and then improved, although still below baseline, at 1 and 4 years [
35]. These results were mirrored by data from the FINNAKI study, with non-significant differences in EuroQOL-SD measurement at 6 months between AKI and non-AKI ICU patients [
36]. However, despite these observations, it is known that in the USA an episode of AKI confers an OR of 2 for the requirement of transfer from hospital to a long-term care facility [
37].
As discussed, an episode of AKI carries a significant short-term mortality risk depending on the severity of AKI, and this relationship also holds for longer term outcomes. In an extended follow up of the RENAL study, there were 810 survivors form the original RENAL study at day 90, of which 32 % died during the 4-year follow-up period, reflecting an overall mortality rate of 62 % in the study cohort [
31]. In keeping with these results, the FINNAKI study reported 6-month mortality at 35.5 % for AKI patients and 16.5 % for non-AKI patients admitted with a critical illness [
36]. A recent prospective cohort evaluation of 2010 ICU patients in a tertiary centre revealed an adjusted 10-year mortality risk of 1.26 (1.0–1.6). After propensity matching, even stage 1 AKI was associated with decreased 10-year survival (
P = 0.036) [
2].
These long-term outcomes for AKI repeatedly demonstrate high mortality rates; for example, Morgera and Hans, in 979 RRT-requiring AKI patients: 14 % 5-year survival [
13]. Interestingly, careful analysis of mortality outcome has revealed a biphasic pattern [
38]: an initial high rate of mortality in the immediate post-AKI setting followed by a steady, but increased, mortality rate against matched non-AKI controls.
It must be remembered that AKI is not a clinical diagnosis but a clinical syndrome characterized by a cluster of observations. The causes of AKI are numerous, and therefore it comes as no surprise that the varying aetiology of AKI may also affect long-term outcome. This is seen in a single-centre analysis of mortality after AKI at 6 months and 5 years. Only a small increase in mortality from 6 months to 5 years was seen (10 %), potentially explained by a higher than usual contribution due to intoxication and rhabdomyolysis. Regardless, the total all-cause 5-year mortality was high at 65 % and consistent with other long-term outcome AKI studies [
39]. Finally of note is the fact that the development of advanced CKD, dialysis dependence, and mortality are competitive endpoints, as many patients surviving AKI do not live long enough to develop advanced CKD requiring dialysis, for example.