J Nephrol DOI 10.1007/s40620-015-0181-5
ORIGINAL ARTICLE
Chronic kidney disease progression is mainly associated with non-recovery of acute kidney injury Eric D’hoore • Nathalie Neirynck • Eva Schepers Raymond Vanholder • Francis Verbeke • Mira Van Thielen • Wim Van Biesen
•
Received: 5 October 2014 / Accepted: 6 February 2015 Ó Italian Society of Nephrology 2015
Abstract Background Identifying individuals who are at increased risk for accelerated progressive chronic kidney disease (CKD) and who might benefit from preventive interventions is an important challenge. Methods The present observational study evaluated the effect of an episode of Acute Kidney Injury (AKI) on the evolution of the renal trajectory in a cohort of 311 ambulatory CKD patients. We analyzed the evolution of eGFR in this cohort within a 5-year time window around an AKI episode. The mean of the available eGFR-values over a 6 month period was calculated once at the start and once at the end of the 5-year period. Slow and fast CKD progression were defined as a decrease by respectively B or [1 category of 15 ml/min/1.73 m2 over the 5-year time window. The influence of AKI on progression status was analyzed. Results Median eGFR decline over the 5 year period was 11, 22 and 6 ml/min/1.73 m2 in the total, AKI and no AKI group respectively. AKI occurred in 44/72 versus 50/239 of fast versus slow progressors (odds ratio: 5.9, 95 % confidence interval: 3.4–10.5). An incomplete recovery of eGFR after an AKI episode (median in overall, fast progressors, slow progressors 11, 20 and 4 ml/min/1.73 m2 respectively) was the major component for the overall loss of renal function over the 5-year window. Our data failed to provide
Electronic supplementary material The online version of this article (doi:10.1007/s40620-015-0181-5) contains supplementary material, which is available to authorized users. E. D’hoore (&) N. Neirynck E. Schepers R. Vanholder F. Verbeke M. Van Thielen W. Van Biesen Nephrology Division, Department of Internal Medicine, Ghent University Hospital, 0K12IA, De Pintelaan 185, Ghent, Belgium e-mail: [email protected]
evidence that the CKD progression became more accelerated once kidney function was stabilized after the AKI episode. Conclusions Incomplete recovery of AKI was related with accelerated CKD-progression. Episodes of AKI were not associated with an accelerated decline of kidney function once the AKI episode had resolved. In the group without AKI episode, the progression was similar to that of the general population without CKD. Keywords Acute kidney injury (AKI) Cohort study Evolution renal function Epidemiology Renal function trajectory
Introduction Chronic kidney disease (CKD) is an increasing global health concern, with a reported prevalence of up to 13 % worldwide for stages 1-4. It is associated with increased morbidity, mortality and higher socio-economic costs [1– 7]. Identifying individuals who are at increased risk for accelerated progression of CKD and who might benefit from preventive intervention is therefore an important challenge [2, 8, 9]. For a long time, progression of CKD was considered to happen in a more or less stable (linear) fashion, but recent findings have challenged this view. Over the last years, and in contrast to previous paradigms, episodes of acute kidney injury (AKI) have been associated with a higher odds for CKD and end stage renal disease (ESRD), whereas on the other hand, pre-existing CKD has been found to predispose also for AKI [10–18]. Some authors suggested that distinction between AKI and CKD is artificial, as at least in some patients both conditions could be considered as an integrated clinical syndrome of
123
J Nephrol
chronically diminished glomerular filtration rate (GFR), with intermittent acute deteriorations [19, 20]. The association between AKI and chronic progression of the estimated glomerular filtration rate (eGFR) trajectory in CKD patients is blurred by the fact that most studies in this area started from patients who actually did have an episode of AKI and/or contain a mix of patients with and without preexisting CKD. Only few data exist on the progression of CKD expressed in eGFR, and the impact of having AKI generally was not evaluated [21, 22]. In the current study we intended to explore whether an episode of AKI impacts on the rate of decline of renal function in a cohort of patients with chronic kidney disease.
Subjects and methods Study population Consecutive adult patients with CKD attending the nephrology outpatient clinic of a university hospital (UZGent) between January 2011 and September 2012 were included in this cohort study. Medical history and comorbidities were registered in a dedicated database. All available serum creatinine values from each individual patient measured between December 2001 and April 2013 were registered. As in our institution serum creatinines were determined with an IDMS traceable method, eGFR was calculated according to the CKD-EPI formula [23]. AKI episodes were defined as a sudden loss of eGFR by C25 % confirmed by at least two measurements versus the last stable eGFR before the AKI (eGFR-preAKI) [24]. Age, gender, underlying kidney disease and comorbidities at inclusion were noted. Underlying kidney diseases were classified into three types: 1/diabetic nephropathy or nephroangiosclerosis, 2/glomerular disease, 3/others: composed of interstitial disease, autosomal dominant polycystic kidney disease (ADPKD) and miscellaneous. The registered comorbidities were: 1/cardiovascular disease (defined as a
entire population (n =576)
history of an event related to coronary, cerebral or peripheral artery disease), 2/heart failure (defined as hospitalization for cardiac decompensation), 3/diabetes (treatment with glycaemia lowering drugs or a history of diabetes). Each AKI episode was categorized based on presumed underlying aetiology into one of the five following groups: 1/renal hypoperfusion; 2/intrinsic renal causes including: glomerulonephritis, acute interstitial nephritis, vasculitis, pre-eclampsia, cast nephropathy, amyloidosis and cholesterol embolism; 3/toxic, drug induced or pigment nephropathy; 4/post-renal; and 5/unknown. Patients were excluded for further analysis if insufficient data were available to accurately define eGFR at the different time points. Also presence of creatinine shifts following a nephrectomy during the observation period was an exclusion criterion. Of the original cohort of 576 patients, 265 conformed with one of the exclusion criteria, leaving 311 patients for final analysis (Fig. 1). Measurements and calculations The evolution of eGFR was analyzed over a 5-year time window. eGFR at the start (eGFRinitial) and at the end (eGFRend) of each 5-year window was calculated as the mean of the available eGFRs over the 6 months period at the beginning and at the end of the 5 years. For patients without any AKI episode, the evolution over the 5 most recent years was analysed. For patients with at least one AKI episode during the entire follow-up, the 5-year time window was defined as the period which approximated 2.5 years before (eGFRinitial) until 2.5 years after (eGFRend) the lowest eGFR value of the most recent AKI episode (eGFR-AKI), with a minimum of at least 1 year before or after AKI. The last stable eGFR value before and the first stable eGFR value after the AKI episode were labelled eGFR-preAKI and eGFR-postAKI respectively (Fig. 2). In order to minimize the influence of eGFR fluctuations due to day by day variation and analytical imprecision of
exclusion: n= 265 - available follow-up less than 5 year: n= 161 - no serum creatinine values available at the predefined time points (in figure 2): n= 82 - less than 4 creatinine values available: n= 10 - RRT needed during all measured serum creatinine values: n= 8
study population (n =311)
Fig. 1 Flow chart. RRT renal replacement therapy
123
- nephrectomy during follow-up: n= 4
J Nephrol
eGFR
eGFR-preAKI
eGFR-postAKI eGFRinitial
eGFRend
eGFR-AKI
Time 2.5 years before AKI
2.5 years after AKI
Fig. 2 Illustration of parameters analysed in conjunction with the AKI episode. eGFR estimated glomerular filtration rate, AKI episode acute kidney injury episode, eGFRinitial and eGFRend mean eGFR of available eGFR values over a 6 month period respectively before and after the AKI episode, eGFR-preAKI and eGFR-postAKI respectively last/first stable eGFR value before/after the AKI episode, eGFR-AKI: lowest eGFR value during AKI
creatinine measurement, each eGFR value was also categorized into eight strata: 1: C91 ml/min/1.73 m2, 2A: 90 to 76 ml/min/1.73 m2, 2B: 75 to 61 ml/min/1.73 m2, 3A: 60 to 46 ml/min/1.73 m2, 3B: 45 to 31 ml/min/ 1.73 m2, 4: 30 to 16 ml/min/1.73 m2, 5A: B15 ml/min/ 1.73 m2 without renal replacement therapy (RRT), 5B: RRT. Based on this categorisation, a patient was defined as ‘‘fast progressor’’ if there was a decrease by [1 category over the 5-year time window and as ‘‘slow progressor’’ for all other patients. Statistics For the statistical analysis, SPSS statistics 22 (SPSS Inc., www.spss.com) was used. For categorical values, Chi square analysis was used. All continuous data were nonparametrically distributed. As statistical tests the Mann– Whitney U Test, Pearson Chi-square, Kruskal–Wallis test, related-samples Friedman’s two-way analysis were applied to evaluate the baseline clinical variables between the different groups. The association between the different clinical characteristics as independent variables and the primary outcome, being CKD progression, was analyzed by univariate logistic regression analysis. In multivariate logistic regression analysis, all variables that were significant in univariate analysis were analyzed together using a backward method (model 1: eGFR loss by AKI, cardiovascular disease, glomerular disease, heart failure). In a second model age and gender were added to the variables included in model 1.
Results Study population All demographic data of the overall cohort are displayed in Table 1, as well as the separate data for fast and slow progressors with their AKI status. The median decline in eGFR over the 5-year window was 11, 22 and 6 ml/min/ 1.73 m2 in the total, AKI and no AKI group respectively. In patients without an AKI episode, the median rate of renal function decline was comparable to that of a healthy population [25, 26]. Ninety-four patients (30 %) had an AKI episode, with renal hypoperfusion as the predominant cause (63 %). Twenty-three percent of the study population was defined as fast progressor ([1 category loss over 5 year), with a median eGFR decline over the 5-year period of 31 ml/min/1.73 m2 versus 5 ml/min/1.73 m2 in the slow progressors. The prevalence of AKI differed significantly between both groups: 44/72 versus 50/239 respectively in fast progressors versus slow progressors. (odds ratio: 5.9, 95 % confidence interval: 3.4– 10.5) (Table 2; Fig. 3). There were significantly more serum creatinine measurements in the AKI group (median: 52) than in the non-AKI group (median: 22) (P-value\ 0.001), this was mainly due to the additional measurements performed during the actual AKI. In multivariate logistic regression analysis with conditions significant in univariate analysis, fast versus slow progression was independently associated with incomplete recovery of eGFR during an AKI episode, and by having underling cardiovascular and glomerular kidney disease (Table 3).
123
J Nephrol Table 1 characteristics of study population and of the 4 subgroups: fast versus slow progressor with AKI and no AKI Study population (n = 311)
Fast progressor ([1 category loss over 5 year), n = 72
Slow progressor (B1 category loss over 5 year), n = 239
AKI n = 44
AKI n = 50
No AKI n = 28
P-value
No AKI n = 189
Age (year)
66 [52–75]
67 [60 to 75]
63 [50 to 73]
73 [62 to 81]
64 [49 to 74]
0.001*
Gender (% Men)
60
61
75
52
59
0.259
eGFRinitial (ml/min/1.73 m2)
55 [40 to 78]
78 [61 to 86]
53 [41 to 64]
51 [35 to 68]
53 [39 to 81]
\0.001*
cat decline DeGFRinitial to eGFRend
1 [0 to 1]
3 [2 to 4]
2 [2]
1 [0 to 1]
0 [0 to 1]
\0.001*
eGFR decline DeGFRinitial to eGFRend (ml/min/1.73 m2/5y) (Fig. 3)
11 [1 to 22]
39 [29 to 50]
25 [23 to 30]
14 [2 to 20]
5 [0 to 12]
\0.001*
Number of serum creatinine measurements
28 [16 to 49]
59 [37 to 87]
36 [18 to 68]
50 [31 to 79]
20 [12 to 31]
\0.001*
Cardiovascular disease
35
44
61
54
25
\0.001*
Heart failure
13
26
25
26
5
\0.001*
DM
33
32
50
34
31
0.245
DM ? vascular
41
39
61
44
38
0.143
Glomerular
20
14
7
18
24
0.119
Interstitial/ADPKD/miscellaneous
39
48
32
38
38
0.562
Comorbidities (at inclusion) (%)
Underlying kidney disease (%)
Aetiology (%) of AKI episodes Renal hypoperfusion
63
55
–
70
–
0.122
Renal
15
20
–
10
–
0.155
Toxic/drug induced/pigment nephropathy
7
7
–
8
–
0.828
Post-renal
7
11
–
4
–
0.175
Unknown
7
7
–
8
–
0.828
AKI episode eGFR-preAKI (ml/min/1.73m2)
50 [36 to 71]
57 [38 to 83]
–
47 [35 to 65]
–
0.058
eGFR-AKI (ml/min/1.73m2)
20 [12 to 30]
21 [12 to 31]
–
18 [12 to 29]
–
0.601
eGFR-postAKI (ml/min/1.73m2)
40 [27 to 54]
39 [27 to 51]
–
40 [27 to 58]
–
0.239
Time to stabilisation from eGFR-AKI to eGFR-postAKI (days)
48 [9 to 142]
22 [2 to 86]
–
104 [19 to 190]
–
0.001*
Decline (DeGFR-preAKI–eGFR-AKI) (ml/min/1.73 m2)
28 [19 to 43]
33 [22 to 53]
–
24 [17 to 41]
–
0.039*
Percent decline ([DeGFR-preAKI–eGFR-AKI] X100/eGFR-preAKI) (%)
57 [45 to 72]
59 [45 to 73]
–
54 [43 to 72]
–
0.299
Severity of AKI
Recovery of AKI Cat loss by AKI (DeGFR-preAKI–eGFR-postAKI)
1 [0 to 1]
1.0 [1 to 3]
–
0 [0 to 1]
–
\0.001*
eGFR loss by AKI (DeGFR-preAKI–eGFRpostAKI) (ml/min/1.73 m2)
11 [0 to 22]
20 [6 to 35]
–
4 [-1 to 12]
–
\0.001*
Percent definitive loss due to AKI ([DeGFRpreAKI–eGFR-postAKI] X100/eGFR-preAKI) (%)
19 [0 to 42]
38 [10 to 56]
–
11 [-3 to 30]
–
\0.001*
Median [interquartile range], cat category, AKI acute kidney injury, DM diabetes mellitus, ADPKD autosomal dominant polycystic kidney disease, D difference * Statistical significant P-value (\0.05)
123
J Nephrol Table 2 Fast and slow progression in AKI and no AKI patients Study population
Fast progressor
Slow progressor
Total
AKI
44 (14 %)
50 (16 %)
94 (30 %)
no AKI
28 (9 %)
189 (61 %)
217 (70 %)
Total
72 (23 %)
239 (77 %)
311 (100 %)
Data are absolute numbers of patients and (percentage between brackets). The odds ratio between fast and slow progressor for the occurrence of AKI: 5.9 (95 % confidence interval 3.4 to 10.5)
the AKI X length of AKI), and the projected median was 49 ml/min/1.73 m2 (34–70). This differs significantly from the real eGFR-postAKI [median: 40 ml/min/1.73 m2 (27–54)] (P-value\0.001). Hence, AKI does accelerate the trajectory of progression during the AKI. As a consequence, incomplete recovery of renal function after the AKI episode was the major determinant for the overall loss of renal function over the evaluated 5 years (Tables 3, 4). In contrast, the occurrence of an AKI episode was not associated with an acceleration of the rate of kidney function decline in the period after the AKI was resolved (Table 4). The results remained similar when fast versus slow progression were defined as a loss by an absolute value of respectively more or less than 10 ml/min/1.73 m2 over a 5-year window (data not shown).
Discussion
Fig. 3 Boxplot of eGFR decline over the 5-year period according to fast and slow progressors in AKI versus no AKI. Overall P-value (Kruskal–Wallis) \0.001, post hoc analysis (Mann–Whitney U): all P-values among subgroups B0.001)
Characteristics of AKI episodes in fast versus slow progressors In the subgroup with AKI, the eGFR decline before, during and after AKI was compared between fast and slow progressors. Although the severity of AKI, when expressed as the absolute decrease of eGFR during the AKI episode (DeGFR-preAKI–eGFR-AKI), was higher in fast versus slow progressors, this was not significant when it was expressed in percentage as proportion to baseline ([DeGFRpreAKI–eGFR-AKI] X100/eGFR–preAKI). In contrast, loss of renal function after AKI recovery, defined as the quantity of eGFR lost due to an AKI episode, was more substantial in fast versus slow progressors, both when expressed as absolute (DeGFR-preAKI – eGFR-postAKI) or as percentage ([DeGFR-preAKI – eGFR-AKI] X100/ eGFR-preAKI) decline (both P-values \0.001) (Table 1). In the AKI subgroup a theoretical eGFR-postAKI was calculated at the endpoint of the AKI episode as if no AKI would have occurred. The theoretical eGFR-postAKI was extrapolated from eGFR-preAKI–(slope of eGFR before
This study evaluated the influence of an episode of AKI versus no AKI on the progression of renal insufficiency in a cohort of consecutive patients with pre-existing CKD. The three major findings are: 1/an episode of AKI was strongly associated with accelerated CKD-progression, even after adjustment for other conditions and comorbidities (Tables 2, 3). 2/The accelerated loss of renal function in the fast progressor subgroup was mainly associated to incomplete recovery of the AKI episode, whereas the mean deterioration rate of eGFR after the AKI episode was not more pronounced than before the AKI episode (Table 4). 3/In the group without an AKI episode, the rate of progression of kidney function deterioration over the 5-year period was comparable to that of the general population. In recent years, AKI has been associated with a higher risk of progression of CKD and ESRD [10–17]. However, so far the underlying mechanisms have not been elucidated. Most studies include only patients who actually have experienced an episode of AKI, making a comparison with the evolution of CKD in patients without AKI difficult. Our data indicate that AKI patients have a much higher risk (5.9 odds) for fast progression of their underlying CKD (Table 2). However, an AKI episode was not associated with a faster decline of renal function once kidney function had stabilized post AKI. (Table 4) This is in line with the findings reported earlier in a study on intensive care unit (ICU) patients with AKI, which also found that the accelerated CKD-progression could mostly be attributed to incomplete recovery of renal function during the AKI episode [13]. Severity of AKI appeared to play a less important role, the absolute decline in eGFR during the AKI period (DeGFR-preAKI–eGFR-AKI) was only significantly more prominent in the fast versus slow progressor group when expressed as absolute differences, but not when
123
J Nephrol Table 3 Model 1 and 2 of multivariate logistic regression for prediction of progression status of the total study population Study population
Progression status (fast: [1 category loss over 5 year = 1, slow = 0) Odds ratio
95 % CI
Model 1 Constant
0.15
eGFR loss by AKI (ml/min/1.73 m2)
1.09
1.06–1.13*
Cardiovascular disease (1 = yes, 0 = no)
2.38
1.29–4.39*
0.24
0.08–0.79*
Glomerular (1 = yes, 0 = no) Heart failure (1 = yes, 0 = no) Model 2
not significant
Constant
0.53
Age (year)
0.98
0.96–1.00
Gender (1 = female, 0 = Men)
0.64
0.33–1.24
eGFR loss by AKI (ml/min/1.73 m2)
1.10
1.06–1.12*
Cardiovascular disease (1 = yes, 0 = no)
2.29
1.06–4.96*
Glomerular (1 = yes, 0 = no)
0.21
0.06–0.68*
Heart failure (1 = yes, 0 = no)
1.99
0.82–4.82
AKI acute kidney injury, eGFR loss by AKI estimated glomerular filtration rate loss due to an AKI episode (for the patients with no AKI this is zero), 95 % CI the 95 percentage confidence interval around the odds ratio. All registered parameters where included in the model, after stepwise backward method only these three parameters were significant. R2 of the model 1: 0.21, R2 of model 2: 0.22. The univariate logistic regression models of all individual parameters are displayed in supplementary Table 1 * Statistical significant P-value (\0.05)
expressed as percentages ([DeGFR-preAKI – eGFR-AKI] X100/eGFR-preAKI) (Table 1). These results are conform with the study of Macedo et al. [27], but discrepant with some other reports [11, 15, 28]. The design of these studies was however different from our study. All these studies started with an in-hospital AKI event [11, 15, 27, 28], used ICD9 codes for detection of AKI instead of the currently used RIFLE or AKIN criteria [15, 28] and/or excluded CKD patients [15, 28]. The period of recovery time taken into account was very short (less than 90 days after discharge), so that a false impression of recovery can be
created by temporary muscle mass loss or overhydration [11, 28]. They finally also contained no clear-cut definition of baseline kidney function and/or had the latter value calculated based on only one value of serum creatinine [11, 15, 27, 28]. The present study was designed in an attempt to overcome these pitfalls. The overall decrease in kidney function in this CKD study population [median 11 ml/min/1.73 m2 over a 5-year window (Table 2)] was twice as fast as the loss reported in the general population (5 ml/min/1.73 m2 over a 5 year window) [25, 26]. However, in the non-AKI group the median decline was only 6 ml/min/1.73 m2 over the 5 year period, very comparable to the general population. Thus, the current data suggests that kidney function of many CKD patients does not progress much faster than that of the general population, except if they suffer from an intercurrent AKI episode. Our data also strongly underscore the necessity to prevent AKI in CKD patients, by avoiding all known acute renal risk factors (such as nephrotoxic drugs, radio contrast or dehydration). There is a trend that the more prominent the progression before the AKI-episode is, the more evident the loss during the AKI-episode will be (Table 4). However by using linear regression analysis we could not show significance (Pvalue 0.122). It would be very interesting to analyze this finding in a larger group of AKI patients. Measurements of serum creatinine shortly before AKI were sometimes unavailable. As a consequence, it cannot be excluded that in some cases progression of CKD was defined as an AKI-episode. The median time to stabilization of renal function after AKI in our study was 48 days, but there was a large variability. Also in other studies, a large deviation in stabilization time ranging from 7 days up to one year was observed [27, 29]. Remarkably, recovery time appeared to be longer in those with slow versus fast progression (Table 1). To the best of our knowledge, this is the first study using a 5-year time window equally distributed before and after an AKI episode and focusing on an established CKD population. In most other studies the starting point was not CKD, but an in-hospital AKI episode [11, 15, 27, 28]. Therefore, it was by definition mostly impossible to
Table 4 Decline of eGFR before, during and after AKI in all patients with an AKI episode eGFR decline median [interquartile range] (ml/min/ 1.73 m2)/median time (year)
Before AKI (DeGFRinitial – eGFR-preAKI)
During AKI (DeGFRpreAKI–eGFRpostAKI)
After AKI (DeGFRpostAKI–eGFRend)
Related-samples Friedman’s two-way analysis: P-value
AKI population
5 [0–13]/2.5
11 [0–22]
1 [2–6] /2.5
0.009*
Fast progressor
10 [0–22]/2.5
20 [6–35]
2 [0–11]/2.5
0.006*
1 [-2 to 5]/2.5
0.487
Slow progressor
3 [-1 to 9]/2.5
AKI acute kidney injury, D difference * Statistical significant P-value (\0.05)
123
4 [-1 to 12]
J Nephrol
calculate baseline and endpoint eGFR as an average of several values over a stable time frame, as done in our study. Finally, by starting from in-hospital AKI, most studies selected patients with more severe AKI episodes, often occurring during an ICU stay. In contrast in our study, AKI was defined referring to a pre-defined drop in eGFR, whenever it occurred, which offers more chances to include all possible causes of acute kidney injury, thus conforming more with real-life circumstances. The strength of the present study is the long observation period before and after the reference AKI episode. Using a mean of 6 months of eGFR values to define eGFRinitial and eGFRend and categorizing the renal function minimized the risk for misclassification of fast versus slow progressors. Changing the definition of progression to a loss of more than 10 ml/min/1.73 m2 over a 5-year window, did not alter the conclusions. However, ascertainment bias cannot be excluded. This is related to the likelihood that sicker patients are at higher risk of having received more nephrotoxic agents while at the same time having had more serum creatinine measurements, increasing the probability to detect AKI.
Conclusion Progression of renal insufficiency after an AKI episode in a population with known CKD is mostly associated to incomplete recovery of renal function after the AKI insult, whereas rate of progression once kidney function has recovered from the AKI episode does not accelerate. The majority of CKD patients who do not develop an AKI episode seem to have a rather stable kidney function and in general show not a faster deterioration of renal function than the general population. Conflict of interest On behalf of all authors, the corresponding author states that there is no conflict of interest. Ethical approval Collection of samples and patient medical data was approved by the ethical committee of the Ghent University Hospital.
References 1. Meguid El Nahas A, Bello AK (2005) Chronic kidney disease: the global challenge. Lancet 365(9456):331–340. doi:10.1016/ S0140-6736(05)17789-7 2. Goldstein SL, Jaber BL, Faubel S, Chawla LS, Acute Kidney Injury Advisory Group of American Society of N (2013) AKI transition of care: a potential opportunity to detect and prevent CKD. Clin J Am Soc Nephrol CJASN 8(3):476–483. doi:10. 2215/CJN.12101112
3. Turin TC, Coresh J, Tonelli M, Stevens PE, de Jong PE, Farmer CK, Matsushita K, Hemmelgarn BR (2013) Change in the estimated glomerular filtration rate over time and risk of all-cause mortality. Kidney Int 83(4):684–691. doi:10.1038/ki.2012.443 4. Perkins RM, Bucaloiu ID, Kirchner HL, Ashouian N, Hartle JE, Yahya T (2011) GFR decline and mortality risk among patients with chronic kidney disease. Clin J Am Soc Nephrol CJASN 6(8):1879–1886. doi:10.2215/CJN.00470111 5. Pereira MB, Zanetta DM, Abdulkader RC (2012) The real importance of pre-existing comorbidities on long-term mortality after acute kidney injury. PLoS One 7(10):e47746. doi:10.1371/ journal.pone.0047746 6. Rosansky SJ (2012) Renal function trajectory is more important than chronic kidney disease stage for managing patients with chronic kidney disease. Am J Nephrol 36(1):1–10. doi:10.1159/ 000339327 7. Lebiedz P, Knickel L, Engelbertz C, Luders F, Gebauer K, Berdel WE, Waltenberger J, Reinecke H (2014) Impact of preexisting chronic kidney disease on acute and long-term outcome of critically ill patients on a medical intensive care unit. J Nephrol 27(1):73–80. doi:10.1007/s40620-013-0016-1 8. Collins GS, Omar O, Shanyinde M, Yu LM (2013) A systematic review finds prediction models for chronic kidney disease were poorly reported and often developed using inappropriate methods. J Clin Epidemiol 66(3):268–277. doi:10.1016/j.jclinepi. 2012.06.020 9. Obrador GT, Mahdavi-Mazdeh M, Collins AJ, Global Kidney Disease Prevention N (2011) Establishing the global kidney disease prevention network (KDPN): a position statement from the National Kidney Foundation. Am J Kidney Dis Off J Nat Kidney Found 57(3):361–370. doi:10.1053/j.ajkd.2010.12.006 10. Turin TC, Coresh J, Tonelli M, Stevens PE, de Jong PE, Farmer CK, Matsushita K, Hemmelgarn BR (2012) Short-term change in kidney function and risk of end-stage renal disease. Nephrol Dial Transplant Off Publ Euro Dial Trans Assoc Euro Renal Assoc 27(10):3835–3843. doi:10.1093/ndt/gfs263 11. Bucaloiu ID, Kirchner HL, Norfolk ER, Hartle JE 2nd, Perkins RM (2012) Increased risk of death and de novo chronic kidney disease following reversible acute kidney injury. Kidney Int 81(5):477–485. doi:10.1038/ki.2011.405 12. Lameire NH, Bagga A, Cruz D, De Maeseneer J, Endre Z, Kellum JA, Liu KD, Mehta RL, Pannu N, Van Biesen W, Vanholder R (2013) Acute kidney injury: an increasing global concern. Lancet 382(9887):170–179. doi:10.1016/S0140-6736(13)60647-9 13. Van Berendoncks AM, Elseviers MM, Lins RL, Group SS (2010) Outcome of acute kidney injury with different treatment options: long-term follow-up. Clin J Am Soc Nephrol CJASN 5(10):1755–1762. doi:10.2215/CJN.00770110 14. Wald R, Quinn RR, Luo J, Li P, Scales DC, Mamdani MM, Ray JG, University of Toronto Acute Kidney Injury Research G (2009) Chronic dialysis and death among survivors of acute kidney injury requiring dialysis. JAMA J Am Med Assoc 302(11):1179–1185. doi:10.1001/jama.2009.1322 15. Chawla LS, Amdur RL, Amodeo S, Kimmel PL, Palant CE (2011) The severity of acute kidney injury predicts progression to chronic kidney disease. Kidney Int 79(12):1361–1369. doi:10. 1038/ki.2011.42 16. Belayev LY, Palevsky PM (2014) The link between acute kidney injury and chronic kidney disease. Curr Opin Nephrol Hypertens 23(2):149–154. doi:10.1097/01.mnh.0000441051.36783.f3 17. Hsu CY (2012) Yes, AKI truly leads to CKD. J Am Soc Nephrol JASN 23(6):967–969. doi:10.1681/ASN.2012030222 18. Basile C (2008) The long-term prognosis of acute kidney injury: acute renal failure as a cause of chronic kidney disease. J Nephrol 21(5):657–662
123
J Nephrol 19. Ronco C, Rosner MH (2012) Acute kidney injury and residual renal function. Crit Care 16(4):144. doi:10.1186/cc11426 20. Chawla LS, Kimmel PL (2012) Acute kidney injury and chronic kidney disease: an integrated clinical syndrome. Kidney Int 82(5):516–524. doi:10.1038/ki.2012.208 21. Hoefield RA, Kalra PA, Lane B, O’Donoghue DJ, Foley RN, Middleton RJ (2013) Associations of baseline characteristics with evolution of eGFR in a referred chronic kidney disease cohort. QJM Mon J Assoc Phys 106(10):915–924. doi:10.1093/qjmed/ hct115 22. Zhang AH, Tam P, LeBlanc D, Zhong H, Chan CT, Bargman JM, Oreopoulos DG (2009) Natural history of CKD stage 4 and 5 patients following referral to renal management clinic. Int Urol Nephrol 41(4):977–982. doi:10.1007/s11255-009-9604-3 23. Levey AS, Stevens LA, Schmid CH, Zhang YL, Castro AF 3rd, Feldman HI, Kusek JW, Eggers P, Van Lente F, Greene T, Coresh J, Ckd EPI (2009) A new equation to estimate glomerular filtration rate. Ann Intern Med 150(9):604–612 24. Bellomo R, Ronco C, Kellum JA, Mehta RL, Palevsky P, Acute Dialysis Quality Initiative Workgroup (2004) Acute renal failure– definition, outcome measures, animal models, fluid therapy and information technology needs: the Second International
123
25.
26.
27.
28.
29.
Consensus Conference of the Acute Dialysis Quality Initiative (ADQI) Group. Crit Care 8(4):R204–R212. doi:10.1186/cc2872 Davies DF, Shock NW (1950) Age changes in glomerular filtration rate, effective renal plasma flow, and tubular excretory capacity in adult males. J Clin Investig 29(5):496–507. doi:10. 1172/JCI102286 Rowe JW, Andres R, Tobin JD, Norris AH, Shock NW (1976) The effect of age on creatinine clearance in men: a cross-sectional and longitudinal study. J Gerontol 31(2):155–163 Macedo E, Zanetta DM, Abdulkader RC (2012) Long-term follow-up of patients after acute kidney injury: patterns of renal functional recovery. PLoS One 7(5):e36388. doi:10.1371/journal. pone.0036388 Jones J, Holmen J, De Graauw J, Jovanovich A, Thornton S, Chonchol M (2012) Association of complete recovery from acute kidney injury with incident CKD stage 3 and all-cause mortality. Am J Kidney Dis Off J Nat Kidney Found 60(3):402–408. doi:10. 1053/j.ajkd.2012.03.014 Magden K, Yildirim I, Kutu M, Ozdemir M, Peynir S, Altas A, Yildiz G, Hur E (2013) Recovery process in patients followed-up due to acute kidney injury. Hippokratia 17(3):239–242
ORIGINAL ARTICLE
Chronic kidney disease progression is mainly associated with non-recovery of acute kidney injury Eric D’hoore • Nathalie Neirynck • Eva Schepers Raymond Vanholder • Francis Verbeke • Mira Van Thielen • Wim Van Biesen
•
Received: 5 October 2014 / Accepted: 6 February 2015 Ó Italian Society of Nephrology 2015
Abstract Background Identifying individuals who are at increased risk for accelerated progressive chronic kidney disease (CKD) and who might benefit from preventive interventions is an important challenge. Methods The present observational study evaluated the effect of an episode of Acute Kidney Injury (AKI) on the evolution of the renal trajectory in a cohort of 311 ambulatory CKD patients. We analyzed the evolution of eGFR in this cohort within a 5-year time window around an AKI episode. The mean of the available eGFR-values over a 6 month period was calculated once at the start and once at the end of the 5-year period. Slow and fast CKD progression were defined as a decrease by respectively B or [1 category of 15 ml/min/1.73 m2 over the 5-year time window. The influence of AKI on progression status was analyzed. Results Median eGFR decline over the 5 year period was 11, 22 and 6 ml/min/1.73 m2 in the total, AKI and no AKI group respectively. AKI occurred in 44/72 versus 50/239 of fast versus slow progressors (odds ratio: 5.9, 95 % confidence interval: 3.4–10.5). An incomplete recovery of eGFR after an AKI episode (median in overall, fast progressors, slow progressors 11, 20 and 4 ml/min/1.73 m2 respectively) was the major component for the overall loss of renal function over the 5-year window. Our data failed to provide
Electronic supplementary material The online version of this article (doi:10.1007/s40620-015-0181-5) contains supplementary material, which is available to authorized users. E. D’hoore (&) N. Neirynck E. Schepers R. Vanholder F. Verbeke M. Van Thielen W. Van Biesen Nephrology Division, Department of Internal Medicine, Ghent University Hospital, 0K12IA, De Pintelaan 185, Ghent, Belgium e-mail: [email protected]
evidence that the CKD progression became more accelerated once kidney function was stabilized after the AKI episode. Conclusions Incomplete recovery of AKI was related with accelerated CKD-progression. Episodes of AKI were not associated with an accelerated decline of kidney function once the AKI episode had resolved. In the group without AKI episode, the progression was similar to that of the general population without CKD. Keywords Acute kidney injury (AKI) Cohort study Evolution renal function Epidemiology Renal function trajectory
Introduction Chronic kidney disease (CKD) is an increasing global health concern, with a reported prevalence of up to 13 % worldwide for stages 1-4. It is associated with increased morbidity, mortality and higher socio-economic costs [1– 7]. Identifying individuals who are at increased risk for accelerated progression of CKD and who might benefit from preventive intervention is therefore an important challenge [2, 8, 9]. For a long time, progression of CKD was considered to happen in a more or less stable (linear) fashion, but recent findings have challenged this view. Over the last years, and in contrast to previous paradigms, episodes of acute kidney injury (AKI) have been associated with a higher odds for CKD and end stage renal disease (ESRD), whereas on the other hand, pre-existing CKD has been found to predispose also for AKI [10–18]. Some authors suggested that distinction between AKI and CKD is artificial, as at least in some patients both conditions could be considered as an integrated clinical syndrome of
123
J Nephrol
chronically diminished glomerular filtration rate (GFR), with intermittent acute deteriorations [19, 20]. The association between AKI and chronic progression of the estimated glomerular filtration rate (eGFR) trajectory in CKD patients is blurred by the fact that most studies in this area started from patients who actually did have an episode of AKI and/or contain a mix of patients with and without preexisting CKD. Only few data exist on the progression of CKD expressed in eGFR, and the impact of having AKI generally was not evaluated [21, 22]. In the current study we intended to explore whether an episode of AKI impacts on the rate of decline of renal function in a cohort of patients with chronic kidney disease.
Subjects and methods Study population Consecutive adult patients with CKD attending the nephrology outpatient clinic of a university hospital (UZGent) between January 2011 and September 2012 were included in this cohort study. Medical history and comorbidities were registered in a dedicated database. All available serum creatinine values from each individual patient measured between December 2001 and April 2013 were registered. As in our institution serum creatinines were determined with an IDMS traceable method, eGFR was calculated according to the CKD-EPI formula [23]. AKI episodes were defined as a sudden loss of eGFR by C25 % confirmed by at least two measurements versus the last stable eGFR before the AKI (eGFR-preAKI) [24]. Age, gender, underlying kidney disease and comorbidities at inclusion were noted. Underlying kidney diseases were classified into three types: 1/diabetic nephropathy or nephroangiosclerosis, 2/glomerular disease, 3/others: composed of interstitial disease, autosomal dominant polycystic kidney disease (ADPKD) and miscellaneous. The registered comorbidities were: 1/cardiovascular disease (defined as a
entire population (n =576)
history of an event related to coronary, cerebral or peripheral artery disease), 2/heart failure (defined as hospitalization for cardiac decompensation), 3/diabetes (treatment with glycaemia lowering drugs or a history of diabetes). Each AKI episode was categorized based on presumed underlying aetiology into one of the five following groups: 1/renal hypoperfusion; 2/intrinsic renal causes including: glomerulonephritis, acute interstitial nephritis, vasculitis, pre-eclampsia, cast nephropathy, amyloidosis and cholesterol embolism; 3/toxic, drug induced or pigment nephropathy; 4/post-renal; and 5/unknown. Patients were excluded for further analysis if insufficient data were available to accurately define eGFR at the different time points. Also presence of creatinine shifts following a nephrectomy during the observation period was an exclusion criterion. Of the original cohort of 576 patients, 265 conformed with one of the exclusion criteria, leaving 311 patients for final analysis (Fig. 1). Measurements and calculations The evolution of eGFR was analyzed over a 5-year time window. eGFR at the start (eGFRinitial) and at the end (eGFRend) of each 5-year window was calculated as the mean of the available eGFRs over the 6 months period at the beginning and at the end of the 5 years. For patients without any AKI episode, the evolution over the 5 most recent years was analysed. For patients with at least one AKI episode during the entire follow-up, the 5-year time window was defined as the period which approximated 2.5 years before (eGFRinitial) until 2.5 years after (eGFRend) the lowest eGFR value of the most recent AKI episode (eGFR-AKI), with a minimum of at least 1 year before or after AKI. The last stable eGFR value before and the first stable eGFR value after the AKI episode were labelled eGFR-preAKI and eGFR-postAKI respectively (Fig. 2). In order to minimize the influence of eGFR fluctuations due to day by day variation and analytical imprecision of
exclusion: n= 265 - available follow-up less than 5 year: n= 161 - no serum creatinine values available at the predefined time points (in figure 2): n= 82 - less than 4 creatinine values available: n= 10 - RRT needed during all measured serum creatinine values: n= 8
study population (n =311)
Fig. 1 Flow chart. RRT renal replacement therapy
123
- nephrectomy during follow-up: n= 4
J Nephrol
eGFR
eGFR-preAKI
eGFR-postAKI eGFRinitial
eGFRend
eGFR-AKI
Time 2.5 years before AKI
2.5 years after AKI
Fig. 2 Illustration of parameters analysed in conjunction with the AKI episode. eGFR estimated glomerular filtration rate, AKI episode acute kidney injury episode, eGFRinitial and eGFRend mean eGFR of available eGFR values over a 6 month period respectively before and after the AKI episode, eGFR-preAKI and eGFR-postAKI respectively last/first stable eGFR value before/after the AKI episode, eGFR-AKI: lowest eGFR value during AKI
creatinine measurement, each eGFR value was also categorized into eight strata: 1: C91 ml/min/1.73 m2, 2A: 90 to 76 ml/min/1.73 m2, 2B: 75 to 61 ml/min/1.73 m2, 3A: 60 to 46 ml/min/1.73 m2, 3B: 45 to 31 ml/min/ 1.73 m2, 4: 30 to 16 ml/min/1.73 m2, 5A: B15 ml/min/ 1.73 m2 without renal replacement therapy (RRT), 5B: RRT. Based on this categorisation, a patient was defined as ‘‘fast progressor’’ if there was a decrease by [1 category over the 5-year time window and as ‘‘slow progressor’’ for all other patients. Statistics For the statistical analysis, SPSS statistics 22 (SPSS Inc., www.spss.com) was used. For categorical values, Chi square analysis was used. All continuous data were nonparametrically distributed. As statistical tests the Mann– Whitney U Test, Pearson Chi-square, Kruskal–Wallis test, related-samples Friedman’s two-way analysis were applied to evaluate the baseline clinical variables between the different groups. The association between the different clinical characteristics as independent variables and the primary outcome, being CKD progression, was analyzed by univariate logistic regression analysis. In multivariate logistic regression analysis, all variables that were significant in univariate analysis were analyzed together using a backward method (model 1: eGFR loss by AKI, cardiovascular disease, glomerular disease, heart failure). In a second model age and gender were added to the variables included in model 1.
Results Study population All demographic data of the overall cohort are displayed in Table 1, as well as the separate data for fast and slow progressors with their AKI status. The median decline in eGFR over the 5-year window was 11, 22 and 6 ml/min/ 1.73 m2 in the total, AKI and no AKI group respectively. In patients without an AKI episode, the median rate of renal function decline was comparable to that of a healthy population [25, 26]. Ninety-four patients (30 %) had an AKI episode, with renal hypoperfusion as the predominant cause (63 %). Twenty-three percent of the study population was defined as fast progressor ([1 category loss over 5 year), with a median eGFR decline over the 5-year period of 31 ml/min/1.73 m2 versus 5 ml/min/1.73 m2 in the slow progressors. The prevalence of AKI differed significantly between both groups: 44/72 versus 50/239 respectively in fast progressors versus slow progressors. (odds ratio: 5.9, 95 % confidence interval: 3.4– 10.5) (Table 2; Fig. 3). There were significantly more serum creatinine measurements in the AKI group (median: 52) than in the non-AKI group (median: 22) (P-value\ 0.001), this was mainly due to the additional measurements performed during the actual AKI. In multivariate logistic regression analysis with conditions significant in univariate analysis, fast versus slow progression was independently associated with incomplete recovery of eGFR during an AKI episode, and by having underling cardiovascular and glomerular kidney disease (Table 3).
123
J Nephrol Table 1 characteristics of study population and of the 4 subgroups: fast versus slow progressor with AKI and no AKI Study population (n = 311)
Fast progressor ([1 category loss over 5 year), n = 72
Slow progressor (B1 category loss over 5 year), n = 239
AKI n = 44
AKI n = 50
No AKI n = 28
P-value
No AKI n = 189
Age (year)
66 [52–75]
67 [60 to 75]
63 [50 to 73]
73 [62 to 81]
64 [49 to 74]
0.001*
Gender (% Men)
60
61
75
52
59
0.259
eGFRinitial (ml/min/1.73 m2)
55 [40 to 78]
78 [61 to 86]
53 [41 to 64]
51 [35 to 68]
53 [39 to 81]
\0.001*
cat decline DeGFRinitial to eGFRend
1 [0 to 1]
3 [2 to 4]
2 [2]
1 [0 to 1]
0 [0 to 1]
\0.001*
eGFR decline DeGFRinitial to eGFRend (ml/min/1.73 m2/5y) (Fig. 3)
11 [1 to 22]
39 [29 to 50]
25 [23 to 30]
14 [2 to 20]
5 [0 to 12]
\0.001*
Number of serum creatinine measurements
28 [16 to 49]
59 [37 to 87]
36 [18 to 68]
50 [31 to 79]
20 [12 to 31]
\0.001*
Cardiovascular disease
35
44
61
54
25
\0.001*
Heart failure
13
26
25
26
5
\0.001*
DM
33
32
50
34
31
0.245
DM ? vascular
41
39
61
44
38
0.143
Glomerular
20
14
7
18
24
0.119
Interstitial/ADPKD/miscellaneous
39
48
32
38
38
0.562
Comorbidities (at inclusion) (%)
Underlying kidney disease (%)
Aetiology (%) of AKI episodes Renal hypoperfusion
63
55
–
70
–
0.122
Renal
15
20
–
10
–
0.155
Toxic/drug induced/pigment nephropathy
7
7
–
8
–
0.828
Post-renal
7
11
–
4
–
0.175
Unknown
7
7
–
8
–
0.828
AKI episode eGFR-preAKI (ml/min/1.73m2)
50 [36 to 71]
57 [38 to 83]
–
47 [35 to 65]
–
0.058
eGFR-AKI (ml/min/1.73m2)
20 [12 to 30]
21 [12 to 31]
–
18 [12 to 29]
–
0.601
eGFR-postAKI (ml/min/1.73m2)
40 [27 to 54]
39 [27 to 51]
–
40 [27 to 58]
–
0.239
Time to stabilisation from eGFR-AKI to eGFR-postAKI (days)
48 [9 to 142]
22 [2 to 86]
–
104 [19 to 190]
–
0.001*
Decline (DeGFR-preAKI–eGFR-AKI) (ml/min/1.73 m2)
28 [19 to 43]
33 [22 to 53]
–
24 [17 to 41]
–
0.039*
Percent decline ([DeGFR-preAKI–eGFR-AKI] X100/eGFR-preAKI) (%)
57 [45 to 72]
59 [45 to 73]
–
54 [43 to 72]
–
0.299
Severity of AKI
Recovery of AKI Cat loss by AKI (DeGFR-preAKI–eGFR-postAKI)
1 [0 to 1]
1.0 [1 to 3]
–
0 [0 to 1]
–
\0.001*
eGFR loss by AKI (DeGFR-preAKI–eGFRpostAKI) (ml/min/1.73 m2)
11 [0 to 22]
20 [6 to 35]
–
4 [-1 to 12]
–
\0.001*
Percent definitive loss due to AKI ([DeGFRpreAKI–eGFR-postAKI] X100/eGFR-preAKI) (%)
19 [0 to 42]
38 [10 to 56]
–
11 [-3 to 30]
–
\0.001*
Median [interquartile range], cat category, AKI acute kidney injury, DM diabetes mellitus, ADPKD autosomal dominant polycystic kidney disease, D difference * Statistical significant P-value (\0.05)
123
J Nephrol Table 2 Fast and slow progression in AKI and no AKI patients Study population
Fast progressor
Slow progressor
Total
AKI
44 (14 %)
50 (16 %)
94 (30 %)
no AKI
28 (9 %)
189 (61 %)
217 (70 %)
Total
72 (23 %)
239 (77 %)
311 (100 %)
Data are absolute numbers of patients and (percentage between brackets). The odds ratio between fast and slow progressor for the occurrence of AKI: 5.9 (95 % confidence interval 3.4 to 10.5)
the AKI X length of AKI), and the projected median was 49 ml/min/1.73 m2 (34–70). This differs significantly from the real eGFR-postAKI [median: 40 ml/min/1.73 m2 (27–54)] (P-value\0.001). Hence, AKI does accelerate the trajectory of progression during the AKI. As a consequence, incomplete recovery of renal function after the AKI episode was the major determinant for the overall loss of renal function over the evaluated 5 years (Tables 3, 4). In contrast, the occurrence of an AKI episode was not associated with an acceleration of the rate of kidney function decline in the period after the AKI was resolved (Table 4). The results remained similar when fast versus slow progression were defined as a loss by an absolute value of respectively more or less than 10 ml/min/1.73 m2 over a 5-year window (data not shown).
Discussion
Fig. 3 Boxplot of eGFR decline over the 5-year period according to fast and slow progressors in AKI versus no AKI. Overall P-value (Kruskal–Wallis) \0.001, post hoc analysis (Mann–Whitney U): all P-values among subgroups B0.001)
Characteristics of AKI episodes in fast versus slow progressors In the subgroup with AKI, the eGFR decline before, during and after AKI was compared between fast and slow progressors. Although the severity of AKI, when expressed as the absolute decrease of eGFR during the AKI episode (DeGFR-preAKI–eGFR-AKI), was higher in fast versus slow progressors, this was not significant when it was expressed in percentage as proportion to baseline ([DeGFRpreAKI–eGFR-AKI] X100/eGFR–preAKI). In contrast, loss of renal function after AKI recovery, defined as the quantity of eGFR lost due to an AKI episode, was more substantial in fast versus slow progressors, both when expressed as absolute (DeGFR-preAKI – eGFR-postAKI) or as percentage ([DeGFR-preAKI – eGFR-AKI] X100/ eGFR-preAKI) decline (both P-values \0.001) (Table 1). In the AKI subgroup a theoretical eGFR-postAKI was calculated at the endpoint of the AKI episode as if no AKI would have occurred. The theoretical eGFR-postAKI was extrapolated from eGFR-preAKI–(slope of eGFR before
This study evaluated the influence of an episode of AKI versus no AKI on the progression of renal insufficiency in a cohort of consecutive patients with pre-existing CKD. The three major findings are: 1/an episode of AKI was strongly associated with accelerated CKD-progression, even after adjustment for other conditions and comorbidities (Tables 2, 3). 2/The accelerated loss of renal function in the fast progressor subgroup was mainly associated to incomplete recovery of the AKI episode, whereas the mean deterioration rate of eGFR after the AKI episode was not more pronounced than before the AKI episode (Table 4). 3/In the group without an AKI episode, the rate of progression of kidney function deterioration over the 5-year period was comparable to that of the general population. In recent years, AKI has been associated with a higher risk of progression of CKD and ESRD [10–17]. However, so far the underlying mechanisms have not been elucidated. Most studies include only patients who actually have experienced an episode of AKI, making a comparison with the evolution of CKD in patients without AKI difficult. Our data indicate that AKI patients have a much higher risk (5.9 odds) for fast progression of their underlying CKD (Table 2). However, an AKI episode was not associated with a faster decline of renal function once kidney function had stabilized post AKI. (Table 4) This is in line with the findings reported earlier in a study on intensive care unit (ICU) patients with AKI, which also found that the accelerated CKD-progression could mostly be attributed to incomplete recovery of renal function during the AKI episode [13]. Severity of AKI appeared to play a less important role, the absolute decline in eGFR during the AKI period (DeGFR-preAKI–eGFR-AKI) was only significantly more prominent in the fast versus slow progressor group when expressed as absolute differences, but not when
123
J Nephrol Table 3 Model 1 and 2 of multivariate logistic regression for prediction of progression status of the total study population Study population
Progression status (fast: [1 category loss over 5 year = 1, slow = 0) Odds ratio
95 % CI
Model 1 Constant
0.15
eGFR loss by AKI (ml/min/1.73 m2)
1.09
1.06–1.13*
Cardiovascular disease (1 = yes, 0 = no)
2.38
1.29–4.39*
0.24
0.08–0.79*
Glomerular (1 = yes, 0 = no) Heart failure (1 = yes, 0 = no) Model 2
not significant
Constant
0.53
Age (year)
0.98
0.96–1.00
Gender (1 = female, 0 = Men)
0.64
0.33–1.24
eGFR loss by AKI (ml/min/1.73 m2)
1.10
1.06–1.12*
Cardiovascular disease (1 = yes, 0 = no)
2.29
1.06–4.96*
Glomerular (1 = yes, 0 = no)
0.21
0.06–0.68*
Heart failure (1 = yes, 0 = no)
1.99
0.82–4.82
AKI acute kidney injury, eGFR loss by AKI estimated glomerular filtration rate loss due to an AKI episode (for the patients with no AKI this is zero), 95 % CI the 95 percentage confidence interval around the odds ratio. All registered parameters where included in the model, after stepwise backward method only these three parameters were significant. R2 of the model 1: 0.21, R2 of model 2: 0.22. The univariate logistic regression models of all individual parameters are displayed in supplementary Table 1 * Statistical significant P-value (\0.05)
expressed as percentages ([DeGFR-preAKI – eGFR-AKI] X100/eGFR-preAKI) (Table 1). These results are conform with the study of Macedo et al. [27], but discrepant with some other reports [11, 15, 28]. The design of these studies was however different from our study. All these studies started with an in-hospital AKI event [11, 15, 27, 28], used ICD9 codes for detection of AKI instead of the currently used RIFLE or AKIN criteria [15, 28] and/or excluded CKD patients [15, 28]. The period of recovery time taken into account was very short (less than 90 days after discharge), so that a false impression of recovery can be
created by temporary muscle mass loss or overhydration [11, 28]. They finally also contained no clear-cut definition of baseline kidney function and/or had the latter value calculated based on only one value of serum creatinine [11, 15, 27, 28]. The present study was designed in an attempt to overcome these pitfalls. The overall decrease in kidney function in this CKD study population [median 11 ml/min/1.73 m2 over a 5-year window (Table 2)] was twice as fast as the loss reported in the general population (5 ml/min/1.73 m2 over a 5 year window) [25, 26]. However, in the non-AKI group the median decline was only 6 ml/min/1.73 m2 over the 5 year period, very comparable to the general population. Thus, the current data suggests that kidney function of many CKD patients does not progress much faster than that of the general population, except if they suffer from an intercurrent AKI episode. Our data also strongly underscore the necessity to prevent AKI in CKD patients, by avoiding all known acute renal risk factors (such as nephrotoxic drugs, radio contrast or dehydration). There is a trend that the more prominent the progression before the AKI-episode is, the more evident the loss during the AKI-episode will be (Table 4). However by using linear regression analysis we could not show significance (Pvalue 0.122). It would be very interesting to analyze this finding in a larger group of AKI patients. Measurements of serum creatinine shortly before AKI were sometimes unavailable. As a consequence, it cannot be excluded that in some cases progression of CKD was defined as an AKI-episode. The median time to stabilization of renal function after AKI in our study was 48 days, but there was a large variability. Also in other studies, a large deviation in stabilization time ranging from 7 days up to one year was observed [27, 29]. Remarkably, recovery time appeared to be longer in those with slow versus fast progression (Table 1). To the best of our knowledge, this is the first study using a 5-year time window equally distributed before and after an AKI episode and focusing on an established CKD population. In most other studies the starting point was not CKD, but an in-hospital AKI episode [11, 15, 27, 28]. Therefore, it was by definition mostly impossible to
Table 4 Decline of eGFR before, during and after AKI in all patients with an AKI episode eGFR decline median [interquartile range] (ml/min/ 1.73 m2)/median time (year)
Before AKI (DeGFRinitial – eGFR-preAKI)
During AKI (DeGFRpreAKI–eGFRpostAKI)
After AKI (DeGFRpostAKI–eGFRend)
Related-samples Friedman’s two-way analysis: P-value
AKI population
5 [0–13]/2.5
11 [0–22]
1 [2–6] /2.5
0.009*
Fast progressor
10 [0–22]/2.5
20 [6–35]
2 [0–11]/2.5
0.006*
1 [-2 to 5]/2.5
0.487
Slow progressor
3 [-1 to 9]/2.5
AKI acute kidney injury, D difference * Statistical significant P-value (\0.05)
123
4 [-1 to 12]
J Nephrol
calculate baseline and endpoint eGFR as an average of several values over a stable time frame, as done in our study. Finally, by starting from in-hospital AKI, most studies selected patients with more severe AKI episodes, often occurring during an ICU stay. In contrast in our study, AKI was defined referring to a pre-defined drop in eGFR, whenever it occurred, which offers more chances to include all possible causes of acute kidney injury, thus conforming more with real-life circumstances. The strength of the present study is the long observation period before and after the reference AKI episode. Using a mean of 6 months of eGFR values to define eGFRinitial and eGFRend and categorizing the renal function minimized the risk for misclassification of fast versus slow progressors. Changing the definition of progression to a loss of more than 10 ml/min/1.73 m2 over a 5-year window, did not alter the conclusions. However, ascertainment bias cannot be excluded. This is related to the likelihood that sicker patients are at higher risk of having received more nephrotoxic agents while at the same time having had more serum creatinine measurements, increasing the probability to detect AKI.
Conclusion Progression of renal insufficiency after an AKI episode in a population with known CKD is mostly associated to incomplete recovery of renal function after the AKI insult, whereas rate of progression once kidney function has recovered from the AKI episode does not accelerate. The majority of CKD patients who do not develop an AKI episode seem to have a rather stable kidney function and in general show not a faster deterioration of renal function than the general population. Conflict of interest On behalf of all authors, the corresponding author states that there is no conflict of interest. Ethical approval Collection of samples and patient medical data was approved by the ethical committee of the Ghent University Hospital.
References 1. Meguid El Nahas A, Bello AK (2005) Chronic kidney disease: the global challenge. Lancet 365(9456):331–340. doi:10.1016/ S0140-6736(05)17789-7 2. Goldstein SL, Jaber BL, Faubel S, Chawla LS, Acute Kidney Injury Advisory Group of American Society of N (2013) AKI transition of care: a potential opportunity to detect and prevent CKD. Clin J Am Soc Nephrol CJASN 8(3):476–483. doi:10. 2215/CJN.12101112
3. Turin TC, Coresh J, Tonelli M, Stevens PE, de Jong PE, Farmer CK, Matsushita K, Hemmelgarn BR (2013) Change in the estimated glomerular filtration rate over time and risk of all-cause mortality. Kidney Int 83(4):684–691. doi:10.1038/ki.2012.443 4. Perkins RM, Bucaloiu ID, Kirchner HL, Ashouian N, Hartle JE, Yahya T (2011) GFR decline and mortality risk among patients with chronic kidney disease. Clin J Am Soc Nephrol CJASN 6(8):1879–1886. doi:10.2215/CJN.00470111 5. Pereira MB, Zanetta DM, Abdulkader RC (2012) The real importance of pre-existing comorbidities on long-term mortality after acute kidney injury. PLoS One 7(10):e47746. doi:10.1371/ journal.pone.0047746 6. Rosansky SJ (2012) Renal function trajectory is more important than chronic kidney disease stage for managing patients with chronic kidney disease. Am J Nephrol 36(1):1–10. doi:10.1159/ 000339327 7. Lebiedz P, Knickel L, Engelbertz C, Luders F, Gebauer K, Berdel WE, Waltenberger J, Reinecke H (2014) Impact of preexisting chronic kidney disease on acute and long-term outcome of critically ill patients on a medical intensive care unit. J Nephrol 27(1):73–80. doi:10.1007/s40620-013-0016-1 8. Collins GS, Omar O, Shanyinde M, Yu LM (2013) A systematic review finds prediction models for chronic kidney disease were poorly reported and often developed using inappropriate methods. J Clin Epidemiol 66(3):268–277. doi:10.1016/j.jclinepi. 2012.06.020 9. Obrador GT, Mahdavi-Mazdeh M, Collins AJ, Global Kidney Disease Prevention N (2011) Establishing the global kidney disease prevention network (KDPN): a position statement from the National Kidney Foundation. Am J Kidney Dis Off J Nat Kidney Found 57(3):361–370. doi:10.1053/j.ajkd.2010.12.006 10. Turin TC, Coresh J, Tonelli M, Stevens PE, de Jong PE, Farmer CK, Matsushita K, Hemmelgarn BR (2012) Short-term change in kidney function and risk of end-stage renal disease. Nephrol Dial Transplant Off Publ Euro Dial Trans Assoc Euro Renal Assoc 27(10):3835–3843. doi:10.1093/ndt/gfs263 11. Bucaloiu ID, Kirchner HL, Norfolk ER, Hartle JE 2nd, Perkins RM (2012) Increased risk of death and de novo chronic kidney disease following reversible acute kidney injury. Kidney Int 81(5):477–485. doi:10.1038/ki.2011.405 12. Lameire NH, Bagga A, Cruz D, De Maeseneer J, Endre Z, Kellum JA, Liu KD, Mehta RL, Pannu N, Van Biesen W, Vanholder R (2013) Acute kidney injury: an increasing global concern. Lancet 382(9887):170–179. doi:10.1016/S0140-6736(13)60647-9 13. Van Berendoncks AM, Elseviers MM, Lins RL, Group SS (2010) Outcome of acute kidney injury with different treatment options: long-term follow-up. Clin J Am Soc Nephrol CJASN 5(10):1755–1762. doi:10.2215/CJN.00770110 14. Wald R, Quinn RR, Luo J, Li P, Scales DC, Mamdani MM, Ray JG, University of Toronto Acute Kidney Injury Research G (2009) Chronic dialysis and death among survivors of acute kidney injury requiring dialysis. JAMA J Am Med Assoc 302(11):1179–1185. doi:10.1001/jama.2009.1322 15. Chawla LS, Amdur RL, Amodeo S, Kimmel PL, Palant CE (2011) The severity of acute kidney injury predicts progression to chronic kidney disease. Kidney Int 79(12):1361–1369. doi:10. 1038/ki.2011.42 16. Belayev LY, Palevsky PM (2014) The link between acute kidney injury and chronic kidney disease. Curr Opin Nephrol Hypertens 23(2):149–154. doi:10.1097/01.mnh.0000441051.36783.f3 17. Hsu CY (2012) Yes, AKI truly leads to CKD. J Am Soc Nephrol JASN 23(6):967–969. doi:10.1681/ASN.2012030222 18. Basile C (2008) The long-term prognosis of acute kidney injury: acute renal failure as a cause of chronic kidney disease. J Nephrol 21(5):657–662
123
J Nephrol 19. Ronco C, Rosner MH (2012) Acute kidney injury and residual renal function. Crit Care 16(4):144. doi:10.1186/cc11426 20. Chawla LS, Kimmel PL (2012) Acute kidney injury and chronic kidney disease: an integrated clinical syndrome. Kidney Int 82(5):516–524. doi:10.1038/ki.2012.208 21. Hoefield RA, Kalra PA, Lane B, O’Donoghue DJ, Foley RN, Middleton RJ (2013) Associations of baseline characteristics with evolution of eGFR in a referred chronic kidney disease cohort. QJM Mon J Assoc Phys 106(10):915–924. doi:10.1093/qjmed/ hct115 22. Zhang AH, Tam P, LeBlanc D, Zhong H, Chan CT, Bargman JM, Oreopoulos DG (2009) Natural history of CKD stage 4 and 5 patients following referral to renal management clinic. Int Urol Nephrol 41(4):977–982. doi:10.1007/s11255-009-9604-3 23. Levey AS, Stevens LA, Schmid CH, Zhang YL, Castro AF 3rd, Feldman HI, Kusek JW, Eggers P, Van Lente F, Greene T, Coresh J, Ckd EPI (2009) A new equation to estimate glomerular filtration rate. Ann Intern Med 150(9):604–612 24. Bellomo R, Ronco C, Kellum JA, Mehta RL, Palevsky P, Acute Dialysis Quality Initiative Workgroup (2004) Acute renal failure– definition, outcome measures, animal models, fluid therapy and information technology needs: the Second International
123
25.
26.
27.
28.
29.
Consensus Conference of the Acute Dialysis Quality Initiative (ADQI) Group. Crit Care 8(4):R204–R212. doi:10.1186/cc2872 Davies DF, Shock NW (1950) Age changes in glomerular filtration rate, effective renal plasma flow, and tubular excretory capacity in adult males. J Clin Investig 29(5):496–507. doi:10. 1172/JCI102286 Rowe JW, Andres R, Tobin JD, Norris AH, Shock NW (1976) The effect of age on creatinine clearance in men: a cross-sectional and longitudinal study. J Gerontol 31(2):155–163 Macedo E, Zanetta DM, Abdulkader RC (2012) Long-term follow-up of patients after acute kidney injury: patterns of renal functional recovery. PLoS One 7(5):e36388. doi:10.1371/journal. pone.0036388 Jones J, Holmen J, De Graauw J, Jovanovich A, Thornton S, Chonchol M (2012) Association of complete recovery from acute kidney injury with incident CKD stage 3 and all-cause mortality. Am J Kidney Dis Off J Nat Kidney Found 60(3):402–408. doi:10. 1053/j.ajkd.2012.03.014 Magden K, Yildirim I, Kutu M, Ozdemir M, Peynir S, Altas A, Yildiz G, Hur E (2013) Recovery process in patients followed-up due to acute kidney injury. Hippokratia 17(3):239–242
