Abstract

Background. Recent research has shown cystatin C to predict mortality and cardiovascular morbidity independent of renal function. The aim of this study was to evaluate the prognostic value of cystatin C on mortality in adult general ICU patients with acute kidney injury (AKI). We later expanded the study and included patients without signs of AKI.

Methods. A total of 845 ICU patients were analysed for cystatin C and classified according to the RIFLE criteria. Of these, 271 patients with either creatinine >150 μmol/l, urea >25 or anuria/oliguria entered the AKI cohort. The remaining 562 patients entered the non-AKI cohort. Both cohorts were divided into quartiles according to cystatin C at entry. In the non-AKI cohort, we split the highest cystatin C quartile into two. The relationship between the different cystatin C quartiles and mortality in patients with and without AKI was estimated by hazard ratios (HR) derived from the Cox proportional hazards regression model.

Results. A relationship between cystatin C and mortality was found in patients with and without AKI, being stronger in patients without AKI. In AKI patients, the HR comparing cystatin C above and below the median more than doubled from the second year on compared to the first year follow-up. After exclusion of patients in the non-AKI cohort with ‘potential AKI’ (creatinine >100 μmol/l or urea > 20 mmol/l), the correlation between cystatin C levels and risk of death was strengthened.

Conclusions. Cystatin C is correlated with mortality independently of renal function measured by creatinine in patients entering the general ICU.

Introduction

Patients with acute kidney injury (AKI) requiring renal replacement therapy (RRT) in the intensive care unit (ICU) have high mortality [ 1–3 ] and morbidity [ 4,5 ]. Even though slightly raised serum levels are associated with elevated mortality [ 6,7 ], the changes in serum creatinine occur with a delay after the AKI has been sustained, and recent research has focused on novel biomarkers for the early detection of AKI [ 8–11 ].

One of the proposed biomarkers is serum cystatin C. Cystatin C is a 13-kD endogenous cysteine proteinase housekeeping protein claimed to be produced at a constant rate by all nucleated cells and filtered freely at the glomerulus and reabsorbed and catabolized but not secreted by the tubules. Although proposed to be a superior marker of early AKI, recent research has indicated that cystatin C is not independent of age, gender, height, weight, smoking and C-reactive protein (CRP) [ 12 ]. Furthermore, studies show cystatin C to be an independent predictor of mortality for different groups of patients [ 13,14 ].

The aim of the present study was to test the hypothesis that cystatin C has an ability to predict mortality of ICU patients with AKI. We later expanded the study and tested the same hypothesis on patients without AKI.

Materials and methods

This study was approved by the Karolinska Institute Ethics Committee, and is therefore in compliance with the Helsinki Declaration. Informed consent was waived by the Ethics Committee as no extra blood samples were taken and all registry data were handled on a group level.

Serum cystatin C was determined by turbidimetry using reagents from Gentian AS (Moss, Norway) on a Hitachi 911 autoanalyser (Roche Diagnostics, Mannheim, Germany).

Study design

This was a prospective cohort study of hospitalized patients in the ICU with AKI. The project was expanded and retrospectively included a second cohort of ICU patients without signs of AKI. The study was conducted in the general, multidisciplinary ICU at the Karolinska Hospital, a major university hospital in Stockholm. The hospital is the trauma referral centre of the greater Stockholm area (1.8 million inhabitants). The departments admitting to the ICU include surgery, urology, gynaecology, obstetrics, otorhinolaryngology and internal medicine. Thoracic and neurosurgical cases are admitted to the unit if these conditions are part of a major trauma. The general ICU is a 10-bed unit treating around 1000 patients per year. About 50% of the patients stay >24 h, and ∼40% are treated with mechanical ventilation.

All consecutive ICU patients (with AKI) were eligible for prospective enrolment as of June 2003 until November 2007. They were screened for the presence of creatinine >150 μmol/l, urea >25 mmol/l or anuria/oliguria (<800 ml/24 h or <30 ml/h for 6 h). The Acute Dialysis Quality Initiative (ADQI) developed the RIFLE classification of AKI [ 15,16 ]. The acronym RIFLE defines three grades of increasing severity of acute renal failure (ARF) (risk, injury and failure, respectively, R, I, and F) and two outcome variables (loss and end-stage renal disease, L and E). A clever feature of the RIFLE classification is that the three grades of severity of renal dysfunction are based on an individual change in serum creatinine, reflecting changes in glomerular filtration rate (GFR) or duration and severity of decline in urine output from the baseline. We excluded patients with RIFLE L or E. The AKI patients were stratified into quartiles according to their cystatin C level. A total of 271 patients fulfilled the AKI criteria.

The retrospective control (non-AKI) cohort was made up of ICU patients between June 2006 and November 2007 with a cystatin C measurement, not fulfilling the above-mentioned renal criteria. Since June 2006, all ICU patients staying >24 h have a serum cystatin C measured. We used the first serum cystatin C measurement available in these patients. The non-AKI patients were then stratified into quartiles according to their cystatin C level. The highest quartile was split into two, giving us five levels of cystatin C (please see the ‘Statistical analyses’ section for the reasons for this). The non-AKI patients were also classified as having a potential AKI if their creatinine exceeded 100 μmol/l or if urea was >20 mmol/l. The non-AKI cohort consisted of 562 patients, with 124 of these classified as having a potential AKI.

Data collection

Research nurses screened the patients at bedside and recorded clinical data including baseline demographics, coexisting conditions and the above-mentioned renal variables. At the time of enrolment, we recorded illness severity scores, the Acute Physiology and Chronic Health Evaluation II (APACHE II) and Sequential Organ Failure Assessment (SOFA). The RIFLE score was calculated at entry as well as the need for inotropes, mechanical ventilation and RRT. Furthermore, we measured cystatin C levels at inclusion. The main ICU diagnosis was recorded.

The non-AKI cohort was added from June 2006. Cystatin C, creatinine, urea (all three measured at ICU admission), APACHE II score and main ICU diagnoses were recorded prospectively, although the data were added retrospectively to this study.

Follow-up

The unique 10-digit national registration number assigned to every Swedish resident ensures the identification and follow-up of the patients in the ICU. We used this number to record outcome measurements including ICU death, as well as mortality at 30, 60 and 180 days. Post-180-day long-term mortality was also recorded.

Statistical analyses

Each cohort member contributed person-years from the entry date until the date of death or the end of November 2007, whichever occurred first. The associations between quartiles of cystatin C and risk of death were estimated by hazard ratios (HR) derived from the Cox proportional hazards regression model with adjustment for age and ICU diagnosis. In the non-AKI cohort, we split the highest quartile into two, in order to analyse if the higher mortality in this quartile came from a few patients with extreme serum cystatin C values. Additional analyses were made with adjustment for RIFLE category and APACHE II score. In the non-AKI cohort, further sensitivity analyses were made by excluding patients with potential kidney injury (all patients with creatinine >100 μmol/l or urea >20 mmol/l), or stratifying by 55 years of age.

Kaplan–Meier cumulative survival curves were plotted for cystatin C quartiles (with the fourth quartile in the non-AKI group split into two) and creatinine quartiles with a follow-up through 2007. The log-rank test was used to examine the difference of survival curves between the groups, not shown in the figures as this is shown in the HR displayed in the tables.

Results

Descriptive data regarding the AKI cohort are shown in Table 1 . More than 90% of the AKI patients were included within 5 days of ICU admission. The last 10%, 25 patients, were included from Day 6 to Day 18.

Table 1

Demographic data of the AKI cohort

 AKI, frequency (%) 
Sex  
 Women 94 (34.81) 
 Men 176 (65.19) 
 Missing observations 
Comorbidities  
 Heart failure 40/264 (15.15) 
 Missing observations 
 Diabetes 50/265 (18.87) 
 Missing observations 
 Hypertension 84/265 (31.70) 
 Missing observations 
Main ICU diagnosis  
 Sepsis/infections 87 (32.10) 
 Circulatory failure 20 (7.38) 
 Respiratory failure 46 (16.97) 
 Acute surgery 44 (16.24) 
 Trauma 29 (10.70) 
 Neurology/neurosurgery 5 (1.85) 
 Obstetrics/gynaecology 2 (0.74) 
 Other diagnoses 38 (14.02) 
RRT in the ICU  
 No RRT 159 (58.67) 
 RRT 112 (41.33) 
Inotropes in the ICU  
 Inotropes 209/264 (78.87) 
 Missing observations 
Ventilator in the ICU  
 Ventilator 190/271 (70.11) 
Death within 30 days 91 (33.58) 
Death between Days 31 and 60 6 (2.21) 
Death between Days 61 and 180 12 (4.43) 
RIFLE on inclusion  
 RIFLE 0 18 (6.64) 
 RIFLE R 45 (16.61) 
 RIFLE I 81 (29.89) 
 RIFLE F 127 (46.86) 
 AKI, mean (std dev) 
Mean Apache II score 24.49 (8.09) 
Mean SOFA score 10.45 (3.69) 
Mean creatinine (μmol/l) 249.45 (122.29) 
Mean urea (mmol/l) 24.19 (26.91) 
Mean cystatin C (mg/l) 2.6 (1.10) 
Mean urinary output (ml/h) a 236.78 (408.81) 
Mean age at entry (years) 61.55 (16.00) 
Total 271 (100) 
 AKI, frequency (%) 
Sex  
 Women 94 (34.81) 
 Men 176 (65.19) 
 Missing observations 
Comorbidities  
 Heart failure 40/264 (15.15) 
 Missing observations 
 Diabetes 50/265 (18.87) 
 Missing observations 
 Hypertension 84/265 (31.70) 
 Missing observations 
Main ICU diagnosis  
 Sepsis/infections 87 (32.10) 
 Circulatory failure 20 (7.38) 
 Respiratory failure 46 (16.97) 
 Acute surgery 44 (16.24) 
 Trauma 29 (10.70) 
 Neurology/neurosurgery 5 (1.85) 
 Obstetrics/gynaecology 2 (0.74) 
 Other diagnoses 38 (14.02) 
RRT in the ICU  
 No RRT 159 (58.67) 
 RRT 112 (41.33) 
Inotropes in the ICU  
 Inotropes 209/264 (78.87) 
 Missing observations 
Ventilator in the ICU  
 Ventilator 190/271 (70.11) 
Death within 30 days 91 (33.58) 
Death between Days 31 and 60 6 (2.21) 
Death between Days 61 and 180 12 (4.43) 
RIFLE on inclusion  
 RIFLE 0 18 (6.64) 
 RIFLE R 45 (16.61) 
 RIFLE I 81 (29.89) 
 RIFLE F 127 (46.86) 
 AKI, mean (std dev) 
Mean Apache II score 24.49 (8.09) 
Mean SOFA score 10.45 (3.69) 
Mean creatinine (μmol/l) 249.45 (122.29) 
Mean urea (mmol/l) 24.19 (26.91) 
Mean cystatin C (mg/l) 2.6 (1.10) 
Mean urinary output (ml/h) a 236.78 (408.81) 
Mean age at entry (years) 61.55 (16.00) 
Total 271 (100) 

a Mean urinary output during the last 6 h prior to inclusion in the AKI cohort.

AKI cohort

Figure 1 A and B show the Kaplan–Meier survival curves of the AKI cohort when it is divided into quartiles according to cystatin C (Figure 1 A) and serum creatinine at entry (Figure 1 B). As seen in these figures, the two highest cystatin C levels seem to be associated with higher mortality compared to the lower levels. In contrast, note the lack of distinction in survival between quartiles of serum creatinine. Table 3 details the HR of the quartiles of cystatin C; the highest quartile of cystatin C reaches statistical significance. However, after adding adjustment for AKI as measured by the RIFLE classification, the point estimates for cystatin C are lowered.

Fig. 1

( A ) Kaplan–Meier survival curve of the AKI cohort by cystatin C level. ( B ) Kaplan–Meier survival curve of the AKI cohort by the creatinine level. ( C ) Kaplan–Meier survival curve of the AKI cohort by the RIFLE classification level.

Fig. 1

( A ) Kaplan–Meier survival curve of the AKI cohort by cystatin C level. ( B ) Kaplan–Meier survival curve of the AKI cohort by the creatinine level. ( C ) Kaplan–Meier survival curve of the AKI cohort by the RIFLE classification level.

Table 2

Demographic data of the non-AKI cohort

 Non-AKI, frequency (%) 
Sex  
 Women 187 (35.02) 
 Men 347 (64.98) 
 Missing observations 28 
Main ICU diagnosis  
 Sepsis/infections 64 (11.39) 
 Circulatory failure 19 (3.38) 
 Respiratory failure 95 (16.90) 
 Acute surgery 61 (10.85) 
 Trauma 217 (38.61) 
 Neurology/neurosurgery 23 (4.09) 
 Obstetrics/gynaecology 10 (1.78) 
 Other diagnoses 73 (12.99) 
RRT in the ICU  
 No RRT 562 (100) 
 RRT 0 (0) 
Death within 30 days 63 (11.21) 
Death between Days 31 and 60 10 (1.78) 
Death between Days 61 and 180 11 (1.96) 
 Non-AKI, mean (std dev) 
Mean Apache II score 15.80 (8.22) 
Mean creatinine (μmol/l) 80.13 (25.53) 
Mean urea (mmol/l) 6.27 (4.60) 
Mean cystatin C (mg/l) 1.12 (0.59) 
Mean age at entry (years) 50.09 (19.38) 
Total 562 (100) 
 Non-AKI, frequency (%) 
Sex  
 Women 187 (35.02) 
 Men 347 (64.98) 
 Missing observations 28 
Main ICU diagnosis  
 Sepsis/infections 64 (11.39) 
 Circulatory failure 19 (3.38) 
 Respiratory failure 95 (16.90) 
 Acute surgery 61 (10.85) 
 Trauma 217 (38.61) 
 Neurology/neurosurgery 23 (4.09) 
 Obstetrics/gynaecology 10 (1.78) 
 Other diagnoses 73 (12.99) 
RRT in the ICU  
 No RRT 562 (100) 
 RRT 0 (0) 
Death within 30 days 63 (11.21) 
Death between Days 31 and 60 10 (1.78) 
Death between Days 61 and 180 11 (1.96) 
 Non-AKI, mean (std dev) 
Mean Apache II score 15.80 (8.22) 
Mean creatinine (μmol/l) 80.13 (25.53) 
Mean urea (mmol/l) 6.27 (4.60) 
Mean cystatin C (mg/l) 1.12 (0.59) 
Mean age at entry (years) 50.09 (19.38) 
Total 562 (100) 
Table 3

Relative risk of all-cause mortality in AKI patients

Variable (ref)  HR a,b  HR a,c  HR a,d 
Level of cystatin C    
 Q1 (≤1.85) 1.0 1.0 1.0 
 Q2 (1.86–2.35) 0.93 (0.52–1.66) 0.92 (0.44–1.90) 0.69 (0.38–1.25) 
 Q3 (2.36–3.19) 1.56 (0.93–2.65) 1.81 (0.98–3.34) 1.23 (0.72–2.09) 
 Q4 (>3.20) 1.74 (1.02–2.97) 1.55 (0.79–3.03) 1.41 (0.82–2.42) 
Variable (ref)  HR a,b  HR a,c  HR a,d 
Level of cystatin C    
 Q1 (≤1.85) 1.0 1.0 1.0 
 Q2 (1.86–2.35) 0.93 (0.52–1.66) 0.92 (0.44–1.90) 0.69 (0.38–1.25) 
 Q3 (2.36–3.19) 1.56 (0.93–2.65) 1.81 (0.98–3.34) 1.23 (0.72–2.09) 
 Q4 (>3.20) 1.74 (1.02–2.97) 1.55 (0.79–3.03) 1.41 (0.82–2.42) 

a Hazard ratios (HR) presented with 95% confidence intervals.

b HR adjusted for age, ICU diagnosis.

c HR adjusted for age, ICU diagnosis and RIFLE.

d HR adjusted for age, ICU diagnosis, RIFLE and APACHE II score.

The longer long-term impact—after Day 500 and on—of high cystatin C on the AKI cohort, shown in the survival curve of Figure 1 A, is also highlighted by the risk estimates from the Cox model. Comparing cystatin C above and below the median (median cystatin C: 2.35) yields an HR of 1.62 [95% confidence interval (CI) 1.10–2.38] during the first year of the follow-up when adjusted for age and ICU diagnosis. From the second year (Day 366 and onwards), the same comparison produced a risk estimate of 5.01 (CI 1.31 –19.21). After adjusting for the RIFLE and APACHE II score, we found the HR of having a cystatin C above the median during Year 1 to be 1.46 (CI 0.99–2.16) and from the second year and onwards 5.29 (CI 1.37–20.39).

Figure 1 C details the survival of the AKI cohort stratified into four groups according to the RIFLE classification. Note that this figure includes AKI patients that did not fulfil the RIFLE criteria, as well as risk, injury and failure patients.

We found no correlation between cystatin C levels (above versus below median) and need for RRT in the ICU. Out of the 112 patients on RRT in the AKI group, 57 (49.1%) patients had a cystatin C level above, and 55 patients (50.9%) below the median.

Non-AKI cohort

Table 2 shows the demographics of the non-AKI cohort. These patients are younger, have lower illness severity and lower mortality as compared to their AKI counterparts. The HR for risk of death are shown in Table 4 . Note the higher point estimates when we restrict the cohort by excluding the 124 patients with potential AKI. (Please see the ‘Materials and methods’ section for exclusion criteria. Also observe the extreme HR for the highest cystatin C levels when we restrict the analysis to patients under the age of 55 years.)

Table 4

Relative risk of all-cause mortality in non-AKI patients

Variable (ref)  HR a,b  HR a,c  HR a,d 
Level of cystatin C    
 Q1 (≤0.70) 1.0 1.0 1.0 
 Q2 (0.71–0.91) 1.78 (0.66–4.77) 1.81 (0.61–5.38) 1.39 (0.37–5.28) 
 Q3 (0.92–1.40) 2.22 (0.86–5.75) 3.01 (1.06–8.53) 1.20 (0.28–5.04) 
 Q4 split into lower (1.41–1.78) 3.81 (1.45–10.05) 5.53 (1.80–16.99) 6.15 (1.58–24.01) 
Q4 split into higher (>1.78) 5.74 (2.20–14.96) 5.95 (1.83–19.36) 12.60 (3.25–48.87) 
Variable (ref)  HR a,b  HR a,c  HR a,d 
Level of cystatin C    
 Q1 (≤0.70) 1.0 1.0 1.0 
 Q2 (0.71–0.91) 1.78 (0.66–4.77) 1.81 (0.61–5.38) 1.39 (0.37–5.28) 
 Q3 (0.92–1.40) 2.22 (0.86–5.75) 3.01 (1.06–8.53) 1.20 (0.28–5.04) 
 Q4 split into lower (1.41–1.78) 3.81 (1.45–10.05) 5.53 (1.80–16.99) 6.15 (1.58–24.01) 
Q4 split into higher (>1.78) 5.74 (2.20–14.96) 5.95 (1.83–19.36) 12.60 (3.25–48.87) 

a Hazard ratios (HR) presented with 95% confidence intervals.

b HR adjusted for age, ICU diagnosis.

c HR adjusted for age, ICU diagnosis and restricted to the 438 patients without potential kidney injury.

d HR adjusted for age, ICU diagnosis and restricted to the 325 patients under the age of 55 years.

Figure 2 and Table 4 show survival in the non-AKI cohort by the cystatin C level. The higher the cystatin C levels, the higher the mortality.

Fig. 2

Kaplan–Meier survival curve of the non-AKI cohort by the cystatin C level.

Fig. 2

Kaplan–Meier survival curve of the non-AKI cohort by the cystatin C level.

Discussion

The present study found that cystatin C is correlated with mortality in ICU patients both with and without clinical signs of AKI.

This finding raises one important question : the association between cystatin C and mortality—is it renal dependent or not? Do patients enter the ICU with differing baseline risks, dependent on or independent of GFR—and is this measured by their baseline cystatin C? We have tried to approach this issue in our study by ‘removing’ or isolating the immediate impact of the AKI on mortality. First, we do this by adjusting for RIFLE in our multivariable analysis of the AKI cohort; and although not reaching significance, the point estimates for higher cystatin C remain elevated. Secondly, we look at the mortality of the AKI patients 1 year after their inclusion in the cohort. Indeed, the bearing of elevated (above or below the median) cystatin C on mortality during the follow-up from Day 366 and onwards is more than twice as strong as during the first year of follow-up. Thirdly, by analysing the patients in the non-AKI cohort—and further restricting our data by excluding patients with potential kidney injury—we found that cystatin C still was correlated with mortality. The association of AKI—and a following lowered GFR, leading to an acute elevation of cystatin C—has been shown [ 17,18 ], and the relationship between AKI and mortality is well known [ 19 ]. The AKI effect on cystatin C levels is apparent in our study, where mean cystatin C was more than twice as high in the AKI cohort compared to the non-AKI patients. We conclude that removal of the effect on mortality of the AKI leads to a reduction of the distortion on the cystatin C signal. This in turn leaves us with the baseline effects of the patients’ cystatin C on mortality.

So, why is elevated cystatin C correlated with mortality, even long-term mortality? Are we dealing with a measurement of tumours, inflammation, apoptosis or cellular destruction? Or is cystatin C merely gauging GFR, indicating mild kidney failure? Does mild kidney failure predict mortality?

There are studies indicating that mild or moderate impairment of GFR—‘passing under the creatinine radar’—is an independent risk factor for mortality. Van Biesen and co-workers show that mild renal failure is associated with cardiovascular death within an apparent healthy population, and the detrimental effect starts at a GFR of 90 ml/min/ 1.73 m 2 [ 20 ]. In a study on elderly, lower predicted GFR was an independent predictor of mortality [ 21 ].

Now, we cannot be sure that the baseline cystatin C levels seen in our patients actually reflect baseline GFR. However, several studies do indicate that cystatin C detects mild renal dysfunction (i.e. GFR >60 ml/min/1.73 m 2 ) better than creatinine [ 22,23 ]. Herget-Rosenthal et al . have shown that increased cystatin C levels preceded a rise in creatinine by up to 1.5 days in ICU patients developing ARF [ 17 ]. Theoretically, this can be explained by the molecular difference between cystatin C and creatinine. Cystatin C is a positively charged, middle-sized molecule (13 kDa). Creatinine on the other hand is smaller (113 Da) and electroneutral. Early changes in glomerular filter porosity and charge during certain physiological (pregnancy) and pathological (diabetes mellitus) conditions may impair glomerular filtration of middle-sized molecules without affecting the passage of smaller molecules like creatinine [ 24,25 ].

A number of investigations have demonstrated a correlation between cystatin C and mortality. In patients with acute heart failure, cystatin C was a predictor of mortality, even in patients with normal creatinine [ 26 ]. It has been demonstrated to predict heart failure per se [ 27 ]. One study detailed an association of the cystatin C level with all-cause and cardiovascular mortality that was as strong as or even stronger than that of iothalamate GFR with these outcomes in stage 3 or 4 chronic kidney disease. Regarding patients who planned for heart surgery, GFR estimated by cystatin C, but not GFR estimated from creatinine, was correlated with death [ 28 ]. Niccoli and co-workers described an association of cystatin C with extent of coronary atherosclerosis in patients with creatinine-derived GFR of >90 ml/min/1.73 m 2 [ 29 ].

Whilst other studies have demonstrated a relationship between kidney disease and mortality [ 30,31 ] that does not guarantee that cystatin C reflects kidney function in our cohorts. To look at the basics, there can be three explanations for an increased baseline cystatin C (the non-AKI effect): (1) increased cellular production, (2) leakage of intracellular cystatin C or (3) reduced filtration in the kidneys, indicating renal damage, not measured by the usual instruments.

Support of theory no. 3 has been mentioned above, but we have also alluded to the fact that cystatin C could reflect another pathogenic state affecting long-term outcome. Although proposed to be an ideal endogenous GFR marker, conclusions whether cystatin C is superior to creatinine in detecting a wide range of GFR are difficult to draw [ 32 ]. Adding to the confusion is the fact that very few studies use accepted reference methods for GFR measurements [ 33 ], and methods to analyse cystatin C and creatinine vary. Complicating the matter further are the reports that cystatin C, but not creatinine, is elevated in patients with HIV [ 34 ]. In two large studies of over 8000 and 1200 patients, respectively [ 12 , 35 ], many factors (such as age, sex, weight, smoking) were associated with cystatin C after adjustment for kidney function.

The question whether elevated cystatin C is an indicator of something other than renal damage—or if cystatin C has a detrimental biologic effect per se —remains to be answered.

There are strengths and weaknesses of this study. In all single-centre studies, there are inherent limitations as results may not be replicable in other settings. Another drawback is the fact that the two cohorts have differing follow-up times, even though they overlap. The later inclusion of the retrospective non-AKI cohort (obviously) led to a shorter follow-up time; they were only followed up for up to 500 days. This limits our ability to draw conclusions on the longer long-term mortality effects of cystatin C. Moreover, we have more data on the AKI patients, and even though these two cohorts are clearly different, an increased transparency in the non-AKI patients would be present if we had pre-morbidity data and information on the need for inotropic drugs and so forth. Luckily, we do have data on age, APACHE II score and ICU diagnoses. Another limitation is that we cannot score the non-AKI patients according to the RIFLE criteria. Naturally, we assume that these patients have a better kidney function than their AKI counterparts, but some of them could be fit into the R category. We have tried to shed some light on this matter by the exclusion of the non-AKI patients with ‘potential’ AKI (please see the ‘Materials and methods’ section for exclusion criteria), but a further limitation is the fact that as AKI by RIFLE criteria also includes urinary output, we could still theoretically have AKI patients dwelling even in this selection of patients. The detail that serum cystatin C was measured at inclusion in the prospective AKI cohort (in >90% of the cases, this meant within 5 days) and at ICU admission (most often meaning within 24 h) for the non-AKI cohort could be seen as a problem, but the fact of the matter is that this probably reflects a clinical reality, where tests can be taken at different time points. Another drawback is that true baseline kidney function is also unknown for these two cohorts. The confounding of inflammation—probably seen in our trauma patients, corticosteroid treatment—in some of the sepsis patients cannot be ruled out, but as we adjust for these ICU diagnoses when we analyse the correlation between cystatin C and risk of death, we find it unlikely to be significant. Another potential confounder is cancer. However, looking at the strong relationship between cystatin C and risk of death in the non-AKI cohort, where the mean age is only 50 years, we believe this confounding to be improbable. Our inability to access information on cause of death in these patients is a limitation as well. This is because the Swedish death register has a lag time of 2 years. Since previous studies imply a correlation between high cystatin C and cardiovascular death, this is a limitation as well as the basis of a potential future study.

Conclusions

Cystatin C levels are correlated with mortality in ICU patients with and without AKI. The distortion created by the AKI does not hinder us from finding an effect, where baseline cystatin C is concurrent with long-term mortality. In the patients without AKI, the cystatin C effect behaves in a dose-dependent fashion. The question remains: are we measuring some hitherto unknown cystatin C-related pathogenic state—affecting long-term outcome—or are we measuring mildly lowered GFR? Regardless, clinical interpretation of renal function by cystatin C alone should be made with caution.

The authors wish to thank Ola Friman, Åsa Bengtsson and all nurses at the Karolinska central ICU for helping them collect data for this study. This work was performed in the following institutions: Department of Anaesthesiology, Surgical Services and Intensive Care Medicine, Karolinska University Hospital, Solna, Sweden, and the Department of Physiology and Pharmacology, Karolinska Institutet, Stockholm, Sweden.

Conflict of interest statement . None. Bell, Martling and Löfberg designed the study. Bell, Granath, Ekbom, Mårtensson and Ekbom analysed the results, with Granath and Bell focusing on the statistical analyses. All authors read the paper on numerous occasions, with Bell writing most of it.

References

1
Bell
M
Liljestam
E
Granath
F
, et al.  . 
Optimal follow-up time after continuous renal replacement therapy in actual renal failure patients stratified with the RIFLE criteria
Nephrol Dial Transplant
 , 
2005
, vol. 
20
 (pg. 
354
-
360
)
2
Korkeila
M
Ruokonen
E
Takala
J
Costs of care, long-term prognosis and quality of life in patients requiring renal replacement therapy during intensive care
Intensive Care Med
 , 
2000
, vol. 
26
 (pg. 
1824
-
1831
)
3
Vivino
G
Antonelli
M
Moro
ML
, et al.  . 
Risk factors for acute renal failure in trauma patients
Intensive Care Med
 , 
1998
, vol. 
24
 (pg. 
808
-
814
)
4
Bell
M
Granath
F
Schon
S
, et al.  . 
Continuous renal replacement therapy is associated with less chronic renal failure than intermittent haemodialysis after acute renal failure
Intensive Care Med
 , 
2007
, vol. 
33
 (pg. 
773
-
780
)
5
Uchino
S
Bellomo
R
Kellum
JA
, et al.  . 
Patient and kidney survival by dialysis modality in critically ill patients with acute kidney injury
Int J Artif Organs
 , 
2007
, vol. 
30
 (pg. 
281
-
292
)
6
Lassnigg
A
Schmidlin
D
Mouhieddine
M
, et al.  . 
Minimal changes of serum creatinine predict prognosis in patients after cardiothoracic surgery: a prospective cohort study
J Am Soc Nephrol
 , 
2004
, vol. 
15
 (pg. 
1597
-
1605
)
7
Chertow
GM
Burdick
E
Honour
M
, et al.  . 
Acute kidney injury, mortality, length of stay, and costs in hospitalized patients
J Am Soc Nephrol
 , 
2005
, vol. 
16
 (pg. 
3365
-
3370
)
8
Han
WK
Bailly
V
Abichandani
R
, et al.  . 
Kidney injury molecule-1 (KIM-1): a novel biomarker for human renal proximal tubule injury
Kidney Int
 , 
2002
, vol. 
62
 (pg. 
237
-
244
)
9
Mishra
J
Dent
C
Tarabishi
R
, et al.  . 
Neutrophil gelatinase-associated lipocalin (NGAL) as a biomarker for acute renal injury after cardiac surgery
Lancet
 , 
2005
, vol. 
365
 (pg. 
1231
-
1238
)
10
Parikh
CR
Abraham
E
Ancukiewicz
M
, et al.  . 
Urine IL-18 is an early diagnostic marker for acute kidney injury and predicts mortality in the intensive care unit
J Am Soc Nephrol
 , 
2005
, vol. 
16
 (pg. 
3046
-
3052
)
11
Liangos
O
Perianayagam
MC
Vaidya
VS
, et al.  . 
Urinary N -acetyl-beta-(d)-glucosaminidase activity and kidney injury molecule-1 level are associated with adverse outcomes in acute renal failure
J Am Soc Nephrol
 , 
2007
, vol. 
18
 (pg. 
904
-
912
)
12
Knight
EL
Verhave
JC
Spiegelman
D
, et al.  . 
Factors influencing serum cystatin C levels other than renal function and the impact on renal function measurement
Kidney Int
 , 
2004
, vol. 
65
 (pg. 
1416
-
1421
)
13
Shlipak
MG
Sarnak
MJ
Katz
R
, et al.  . 
Cystatin C and the risk of death and cardiovascular events among elderly persons
N Engl J Med
 , 
2005
, vol. 
352
 (pg. 
2049
-
2060
)
14
Ni
L
Lu
J
Hou
LB
, et al.  . 
Cystatin C, associated with hemorrhagic and ischemic stroke, is a strong predictor of the risk of cardiovascular events and death in Chinese
Stroke
 , 
2007
, vol. 
38
 (pg. 
3287
-
3288
)
15
Bellomo
R
Kellum
JA
Ronco
C
Defining acute renal failure: physiological principles
Intensive Care Med
 , 
2004
, vol. 
30
 (pg. 
33
-
37
)
16
Kellum
JA
Levin
N
Bouman
C
, et al.  . 
Developing a consensus classification system for acute renal failure
Curr Opin Crit Care
 , 
2002
, vol. 
8
 (pg. 
509
-
514
)
17
Herget-Rosenthal
S
Marggraf
G
Husing
J
, et al.  . 
Early detection of acute renal failure by serum cystatin C
Kidney Int
 , 
2004
, vol. 
66
 (pg. 
1115
-
1122
)
18
Villa
P
Jimenez
M
Soriano
MC
, et al.  . 
Serum cystatin C concentration as a marker of acute renal dysfunction in critically ill patients
Crit Care
 , 
2005
, vol. 
9
 (pg. 
R139
-
R143
)
19
Ricci
Z
Cruz
D
Ronco
C
The RIFLE criteria and mortality in acute kidney injury: a systematic review
Kidney Int
 , 
2008
, vol. 
73
 (pg. 
538
-
546
)
20
Van Biesen
W
De Bacquer
D
Verbeke
F
, et al.  . 
The glomerular filtration rate in an apparently healthy population and its relation with cardiovascular mortality during 10 years
Eur Heart J
 , 
2007
, vol. 
28
 (pg. 
478
-
483
)
21
Maaravi
Y
Bursztyn
M
Hammerman-Rozenberg
R
, et al.  . 
Moderate renal insufficiency at 70 years predicts mortality
Q J Med
 , 
2006
, vol. 
99
 (pg. 
97
-
102
)
22
Christensson
A
Ekberg
J
Grubb
A
, et al.  . 
Serum cystatin C is a more sensitive and more accurate marker of glomerular filtration rate than enzymatic measurements of creatinine in renal transplantation
Nephron Physiol
 , 
2003
, vol. 
94
 (pg. 
19
-
27
)
23
Hoek
FJ
Kemperman
FA
Krediet
RT
A comparison between cystatin C, plasma creatinine and the Cockcroft and Gault formula for the estimation of glomerular filtration rate
Nephrol Dial Transplant
 , 
2003
, vol. 
18
 (pg. 
2024
-
2031
)
24
Roberts
M
Lindheimer
MD
Davison
JM
Altered glomerular permselectivity to neutral dextrans and heteroporous membrane modeling in human pregnancy
Am J Physiol
 , 
1996
, vol. 
270
 (pg. 
F338
-
F343
)
25
Oberbauer
R
Nenov
V
Weidekamm
C
, et al.  . 
Reduction in mean glomerular pore size coincides with the development of large shunt pores in patients with diabetic nephropathy
Exp Nephrol
 , 
2001
, vol. 
9
 (pg. 
49
-
53
)
26
Lassus
J
Harjola
VP
Sund
R
, et al.  . 
Prognostic value of cystatin C in acute heart failure in relation to other markers of renal function and NT-proBNP
Eur Heart J
 , 
2007
, vol. 
28
 (pg. 
1841
-
1847
)
27
Sarnak
MJ
Katz
R
Stehman-Breen
CO
, et al.  . 
Cystatin C concentration as a risk factor for heart failure in older adults
Ann Intern Med
 , 
2005
, vol. 
142
 (pg. 
497
-
505
)
28
Ledoux
D
Monchi
M
Chapelle
JP
, et al.  . 
Cystatin C blood level as a risk factor for death after heart surgery
Eur Heart J
 , 
2007
, vol. 
28
 (pg. 
1848
-
1853
)
29
Niccoli
G
Conte
M
Della Bona
R
, et al.  . 
Cystatin C is associated with an increased coronary atherosclerotic burden and a stable plaque phenotype in patients with ischemic heart disease and normal glomerular filtration rate
Atherosclerosis
 , 
2008
, vol. 
198
 (pg. 
373
-
380
)
30
Go
AS
Chertow
GM
Fan
D
, et al.  . 
Chronic kidney disease and the risks of death, cardiovascular events, and hospitalization
N Engl J Med
 , 
2004
, vol. 
351
 (pg. 
1296
-
1305
)
31
Anavekar
NS
McMurray
JJ
Velazquez
EJ
, et al.  . 
Relation between renal dysfunction and cardiovascular outcomes after myocardial infarction
N Engl J Med
 , 
2004
, vol. 
351
 (pg. 
1285
-
1295
)
32
Zahran
A
El-Husseini
A
Shoker
A
Can cystatin C replace creatinine to estimate glomerular filtration rate? A literature review
Am J Nephrol
 , 
2007
, vol. 
27
 (pg. 
197
-
205
)
33
Blaufox
MD
Aurell
M
Bubeck
B
, et al.  . 
Report of the Radionuclides in Nephrorology Committee on Renal Clearance
J Nucl Med
 , 
1996
, vol. 
37
 (pg. 
1883
-
1890
)
34
Odden
MC
Scherzer
R
Bacchetti
P
, et al.  . 
Cystatin C level as a marker of kidney function in human immunodeficiency virus infection: the FRAM study
Arch Intern Med
 , 
2007
, vol. 
167
 (pg. 
2213
-
2219
)
35
Wasen
E
Isoaho
R
Mattila
K
, et al.  . 
Serum cystatin C in the aged: relationships with health status
Am J Kidney Dis
 , 
2003
, vol. 
42
 (pg. 
36
-
43
)

Comments

0 Comments