ORIGINAL ARTICLE

Renal replacement therapy in the ICU: comparison of clinicalfeatures and outcomes of patients with acute kidney injuryand dialysis-dependent end-stage renal disease

Turkay Akbas • Sait Karakurt • Serhan Tuglular

Received: 4 December 2013 / Accepted: 31 August 2014

� Japanese Society of Nephrology 2014

Summary

Background The goal of this study is to study clinical

features and outcomes of the patients who had renal

replacement therapy (RRT) in the intensive care unit (ICU)

between 2000 and 2007.

Methods We retrospectively studied 222 patients.

Results Overall ICU mortality and invasive mechanical

ventilation (IMV) rates were 58.1 and 61.3 %. The mean

APACHE II score was 27.6 ± 8.3. Chronic dialysis (CD)

patients formed 45.5 % of the study population. Acute

kidney injury (AKI) patients had higher rates of IMV (73

vs. 51.5 %, p = 0.002), cancer (27.8 vs. 7.9 %, p B 0.001)

and mortality (67.8 vs. 50.5 %, p = 0.010) than CD

patients. AKI patients with normal kidney function (NKF)

before ICU admission had poorer prognosis than acute-on-

chronic kidney disease (CKD) and CD patients (78.6, 51

and 50.5 %, respectively, p B 0.001). Multivariate analysis

showed that IMV (OR, 14.8; 95 % CI, 5.47–40.05;

p B 0.001) and having NKF before hospitalization (OR,

2.8; 95 % CI, 1.04–7.37; p = 0.041) were predictors of

overall ICU mortality. Additionally, IMV is found as a

prognostic factor for both AKI (OR, 18.7; 95 % CI,

4.48–77.72; p B 0.001) and CD patients (OR, 8.14; 95 %

CI, 2.01–33.04; p = 0.003), but APACHE II score is

meaningful only for CD patients (OR, 1.13; 95 % CI,

1.02–1.26; p = 0.024). The areas under the ROC curves

for APACHE II score were 0.52 (95 % CI, 0.39–0.66) for

AKI and 0.78 (95 % CI, 0.55–0.89) for CD patients.

Conclusion The observed ICU mortality among patients

requiring RRT is high and IMV is associated with mor-

tality. AKI patients have increased mortality compared to

CD patients. AKI patients with past NKF have poorer

prognosis than acute-on-CKD and CD patients.

Keywords Acute kidney injury � End-stage kidney

disease � Intensive care unit � Renal replacement therapy

Introduction

Renal replacement therapy (RRT) is a commonly used

therapy modality among critically ill patients. The aim of

dialysis is, except for intoxication and hypervolemia,

generally to support renal function in those patients who

develop acute kidney injury (AKI) during their intensive

care unit (ICU) stay or to do regular dialysis of those

patients on chronic dialysis (CD) programs admitted to the

ICU. End-stage renal disease (ESRD) patients consist of

1.3–5.8 % of all admissions to the ICU [1–4] and it was

displayed that 2–20 % of ESRD patients require ICU

admission every year [3, 5]. The main causes of ICU

admission in ESRD patients are infection, bleeding and

cardiovascular events [1, 5–7]. Mortality rate among these

patients is reported to be between 9 and 54 % [1, 3–10].

AKI occurs in 1–50 % of all critically ill patients,

depending on the population studied and criteria used to

T. Akbas (&)

Department of Internal Medicine and Critical Care Unit, School

of Medicine, Marmara University, Istanbul, Turkey

e-mail: turkayakbas@yahoo.com

T. Akbas

Fevzi Cakmak Mahallesi, Mimar Sinan Caddesi, No:41 Ust

Kaynarca, Pendik, Istanbul 34899, Turkey

S. Karakurt

Department of Pulmonary and Critical Care Medicine, School of

Medicine, Marmara University, Istanbul, Turkey

S. Tuglular

Department of Nephrology, School of Medicine, Marmara

University, Istanbul, Turkey

123

Clin Exp Nephrol

DOI 10.1007/s10157-014-1028-4

define it [8, 11–13]. The rate of RRT among AKI patients

is reported to be between 11 and 70 % [8, 11, 12, 14, 15].

The main reason for development of AKI is sepsis which is

followed by major surgery [11, 14, 15]. Generally, AKI is a

part of multiorgan failures and shows poor prognosis.

Mortality of AKI ranges from 23 to 64 % [8, 11–15].

In this study, we will first describe demographical and

clinical features of the patients who had RRT during their

ICU stay. Then, we will describe and compare demo-

graphical and clinical characteristics of the CD and AKI

patients, ending with a comparison of the mortality rate

between the two groups.

Materials and methods

The study population consisted of all patients who had any

type of RRT after admission to the medical ICU of Mar-

mara University Hospital between 2000 and 2007. The

university hospital had 280 beds where the medical ICU

occupied 8 beds among a total of 21 adult ICU beds

(medical, anesthesia and coronary ICU). Nearly 12,000

patients/year were admitted to the hospital where around

260 patients/year admitted to the medical ICU during the

study period. The university hospital has an actively

working dialysis unit with 20 beds and works as a reference

hospital for health-related problems of dialysis-dependent

patients. We reviewed the past records of the patients

through the computer database and medical notes of the

ICU and dialysis unit of the hospital. The patients were

admitted to the ICU from the emergency service or other

wards of the hospital except ten patients from the operation

room. Of the ten patients, nine were on CD and eight had

the operation of renal transplantation. In the event of

multiple admissions, only the initial ICU admission was

accepted. Exclusion criterion was being less than 18 years

old. The study was approved by Ethics Committee of

Marmara University (09.2013.0253).

Data elements included demographics, admission diag-

nosis, comorbidity, Acute Physiologic and Chronic Health

Evaluation (APACHE) II score, need of mechanical venti-

lation (MV), presence of organ failure, development of

hypotension during RRT (systolic blood pressure

\90 mmHg after two successive measurements were con-

sidered as hypotension), type of dialysis, length of ICU stay

(LOS), ICU mortality and cause of mortality. The APACHE

II score was calculated by using the worst values within the

first 24 h of admission to the ICU. MV included noninvasive

MV (NIMV) and invasive MV (IMV). Individual organ

failures were defined as the sequential organ failure assess-

ment score C2 points for each system [12].

A diagnosis of infection included admissions for pri-

marily sepsis-related diagnosis. These were lung,

gastrointestinal, urinary tract, central nervous system, soft

tissue, and catheter-related infections. A cardiac diagnosis

encompassed cardiogenic shock, heart failure, acute myo-

cardial infarction, hypervolemia, cardiac arrest, and rhythm

problems. A hepatic diagnosis included admission with

chronic/active hepatic failure. A diagnosis of hemorrhage

described any bleeding. A diagnosis of epilepsy included

status epilepticus and uncontrolled seizures. Vasculitis

comprised patients who had immune suppressive therapy

or plasmapheresis due to vasculitis-related organ failures.

Uremia included those patients who were admitted to the

ICU with acute deterioration of renal function due to any

cause and CD patients who required urgent RRT due to any

reason. COPD exacerbation described patients with COPD

who had MV support because of respiratory failure.

The baseline serum creatinine levels of the patients were

retrieved from the hospital computer database or medical

files 3 months before the ICU admission; otherwise, the

lowest serum creatinine levels during the admission were

considered. Patients’ previous renal functions were cate-

gorized into three groups: patients who had a documented

history of chronic kidney disease (CKD) or whose glo-

merular filtration rates were \60 ml/min, as calculated by

MDRD using baseline serum creatinine levels, were

accepted as patients with CKD stages 2–5 [16]. Develop-

ment of AKI in those patients was named as acute-on-

CKD. Patients being chronically dialysis-dependent prior

to ICU admission were put into CKD stage 5D [16].

Patients who had normal baseline serum creatinine levels

were accepted as patients with normal kidney function

(NKF). Comorbidities were extrapolated from patients’

medical notes.

All patients requiring dialysis in this study were evalu-

ated and managed by the nephrology department of the

hospital. RRT was either intermittent hemodialysis (IHD),

or continuous venovenous hemodialysis (CVVHD), or

peritoneal dialysis (PD). IHD was employed for hemody-

namically stable patients, while CVVHD was used pri-

marily for patients who were hemodynamically unstable.

Some patients received two modalities (IHD-CVVHD or

PD-IHD). In this case, the last modality used was accepted.

Patients who were on PD before ICU admission had PD

unless contraindicated due to peritonitis, catheter-related

problems or other technical problems. Types of dialysis,

number of dialysis sessions and dialysis prescriptions were

arranged based on outpatient therapy and current status of

the patients.

Statistical analysis

Normally distributed continuous variables were analyzed

by the Student’s t or ANOVA tests and results were

expressed as means with standard deviations (SD).

Clin Exp Nephrol

123

Continuous variables without normal distribution (age and

LOS) were reported as medians with interquartile ranges

and compared using the Mann–Whitney U or Kruskal–

Wallis tests. Categorical variables were expressed as pro-

portions and compared with the Chi-square test. Multi-

variate analysis was performed using logistic regression

with mortality as the dependent outcome variable. The

variables yielding p values\0.15 by univariate analysis or

considered clinically important were entered in a backward

multivariate logistic regression, with goodness of fit

determined by the method of Hosmer–Lemeshow. When

the degree of correlation between the variables entered in

the multiple regression equation is high, the regression

model can produce incorrect variance estimates and may

give a misleading interpretation of the model. To address

the problem of multicolinearity, a correlation matrix was

used to asses the overlap between the data. A high spear-

men correlation (r [ 0.5) indicates that the data highly

correlate with each other. One way of mitigating the

harmful effects of multicolinearity is to delete the offend-

ing variables from the regression model [17]. The receiver

operating characteristic (ROC) curve analysis was used to

determine the performance of APACHE II probability of

death. p \ 0.05 was considered as statistically significant.

Results

During the 7-year period, 222 patients (M/F, 113/109;

median age 66 (53–73), range 19–91) had extracorporeal

dialysis support. The baseline characteristics of the patients

are listed in Table 1. The predominant reasons for admis-

sion were infection, cardiac problems and bleeding. The

mean APPACHE II score was 27.6 ± 8.3 and 45.5 % of

the patients were on CD. The rates of NIMV and IMV were

30.4 and 61.3 %, respectively. Organ system failure data

were obtained for 112 patients where 59.8 % of them had

non-renal organ system failures. The data of LOS was

reached in 61 patients and the median LOS was 6 (2–19.5)

days. Thirteen patients (5.8 %) had IHD at first, but then

switched to CVVHD due to hemodynamic instability and

all of them except one died. One septic patient had PD

initially then, followed by IHD because of technical

problems. Renal-transplant patients had RRT because of

allograft nephropathy. The data of blood pressure during

dialysis session was obtained for 45 patients where 44 % of

them experienced hypotension. Overall observed ICU

mortality was 58.1 %. The mean APACHE II score and

rates of CVVHD, IMV, non-renal organ system failure,

AKI, history of chronic liver disease and hypotension

during RRT were significantly high in nonsurvivors

(Table 2). All surgical patients survived. Mortality

increased with the number of organ system failures,

Table 1 Characteristics of 222 patients who had RRT in the ICU

Sex, n (% men) 113 (50.9)

Age, median (25–75 percentiles) 66 (53-73)

Primary diagnosis, n (%)

Infection 94 (42.3)

Cardiac disease 44 (19.8)

Bleeding 22 (9.9)

Uremia 16 (7.1)

Postoperative follow-up 10 (4.5)

Epilepsy 9 (4.1)

Hepatic failure 6 (2.7)

Othersa 21 (9.6)

Comorbidity, n (%)

CVD 113 (50.9)

DM 68 (30.6)

Cancer 40 (18.0)

COPD 19 (8.6)

Liver disease 19 (8.6)

APACHE II, mean (SD) 27.6 (8.3)

IMV, n (%) 136 (61.3)

NIMV, n (%) 67 (30.2)

Type of dialysis, n (%)

IHD 156 (70.3)

CVVHD 53 (23.9)

PD 13 (5.8)

Dialysis indication, n (%)

Regular dialysis 74 (33.3)

Sepsis-related AKI 59 (26.6)

Uremiab 59 (26.6)

Hypervolemia 25 (11.2)

Hyperkalemia 4 (1.8)

Intoxicationc 1 (0.5)

Mortality, n (%) 129 (58.1)

Cause of mortality, n (%)

Sepsis 106 (82.2)

Cardiac disease 13 (10.1)

Liver failure 6 (4.7)

Bleeding 4 (3)

APACHE acute physiologic and chronic health evaluation, CVD

cardiovascular disease, COPD chronic obstructive pulmonary disease,

DM diabetes mellitus, CVVHD continuous venovenous haemodialy-

sis, ICU intensive care unit, IHD intermittent haemodialysis, IMV

invasive mechanical ventilation, NIMV noninvasive mechanical

ventilation, n number, PD peritoneal dialysis, RRT renal replacement

treatment, SD standard deviationa Others include vasculitis (7), COPD exacerbation (7), pancreatitis

(4), intoxication (2) and HELLP syndrome (1)b Uremia describes indication of RRT due to worsening of kidney

function for any reason except for sepsis for AKI patients and defines

urgent dialysis due to any cause except for hyperkalemia for CD

patientsc One CD patient with the diagnosis of intoxication had normal

regular dialysis during ICU stay

Clin Exp Nephrol

123

reaching 88 % after C3 organ system failures (Fig. 1).

Patients with organ failures had high APACHE II scores

(organ failure?, 30.4 ± 8.6; no organ failure, 24.9 ± 8.2;

p = 0.005).

Among 121 patients who had RRT first time in the ICU,

six patients had extracorporeal dialysis support because of

heart failure-related volume overload (5) and intoxication

(1). The remaining 115 (51.8 %) patients were accepted as

AKI. When we compared AKI patients with CD patients,

the rates of respiratory failure, IMV, cancer and mortality

were found to be significantly high in AKI patients

(Table 3). All AKI patients who were admitted to the ICU

with the diagnosis of liver failure except one died. When

the patients were analyzed according to past renal function,

Table 2 Demographic and

clinical characteristics of

survivors and nonsurvivors

Bold values indicate statistically

significant results

AKI acute kidney injury,

APACHE acute physiologic and

chronic health evaluation, CVD

cardiovascular disease, CD

chronic dialysis, COPD chronic

obstructive pulmonary disease,

DM diabetes mellitus, CVVHD

continuous venovenous

haemodialysis, IHD intermittent

haemodialysis, IMV invasive

mechanical ventilation, LOS

length of ICU stay, NIMV

noninvasive mechanical

ventilation, N/A not applicable,

n number, PD peritoneal

dialysis, SD standard deviationa Others include vasculitis (7),

COPD exacerbation (7),

pancreatitis (4), intoxication (2)

and HELLP syndrome (1)b LOS covers 61 patientsc Hypotension during dialysis

session encloses 45 patientsd Non-renal organ failure

covers 112 patientse Uremia describes indication

of RRT due to worsening of

kidney function for any reason

except for sepsis for AKI

patients and defines urgent

dialysis due to any cause except

for hyperkalemia for CD

patientsf One CD patient with the

diagnosis of intoxication had

normal regular dialysis during

ICU stay

Parameters Alive (n = 93) Dead (n = 129) p

Sex, % men 55.9 47.3 0.20

Age, median (25–75 percentiles) 66 (52–73) 66 (56–73) 0.47

APACHE II, mean (SD) 24.9 (7.1) 29.6 (8.7) 0.001

Primary diagnosis (%)

Infection 24.7 55 N/A

Cardiac disease 28 14

Bleeding 14 7

Uremia 9.5 5.3

Postoperative follow-up 10.8 0

Epilepsy 4.3 3.9

Hepatic failure 0 4.7

Othersa 8.7 10.1

Comorbidity (%)

CVD 53.8 48.8 0.46

DM 34.4 27.9 0.30

Cancer 14 20.9 0.18

COPD 4.3 11.6 0.054

Liver disease 1.1 14 <0.001

IMV (%) 24.7 87.6 <0.001

NIMV (%) 28 31.8 0.54

LOS, median (25–75 percentiles)b 5 (3–14) 9 (2–21) 0.56

Hypotension, n (%)c 6 (20.7) 14 (87.5) <0.001

Type of renal disease (%)

AKI 32.2 67.8 0.010

CD 49.5 50.5

Non-renal organ failure, (%)d

Respiratory 31.5 89.1 <0.001

Cardiovascular 20 86.4 <0.001

Haematologic 5.8 49.2 <0.001

Hepatic 5.8 27.4 0.002

Neurologic 3.8 35.5 <0.001

Type of dialysis (%)

IHD 88.2 57.3 <0.001

CVVHD 3.2 38.8

PD 8.6 3.9

Dialysis indication (%)

Regular dialysis 35.5 31.8 N/A

Sepsis-related AKI 4.3 42.6

Uremiae 33.4 21.7

Hypervolemia 22.5 3.1

Hyperkalemia 3.2 0.8

Intoxicationf 1.1 0

Clin Exp Nephrol

123

AKI patients with previous NKF had higher rates of

chronic liver disease, respiratory failure, IMV and mor-

tality than acute-on-CKD and CD patients (Table 4).

The following variables were entered into the backward

logistic regression model: comorbidities, IMV, age, sex,

APACHE II and past kidney function. Hypotension during

RRT and non-renal organ system failure were not put into

the system due to limited number of the patients. No cor-

relation was encountered between variables. The indepen-

dent predictors of overall ICU mortality were IMV [odds

ratio (OR), 14.8; 95 % CI, 5.47–40.05; p B 0.001] and

presence of NKF before ICU admission (OR, 2.8; 95 % CI,

1.04–7.37; p = 0.041; Hosmer–Lemeshow goodness of fit

v2: 10.59, p = 0.226) (Table 5). After analyzing variables

according to AKI and CD, IMV emerged to be a risk factor

of mortality for both AKI (OR, 18.7; 95 % CI, 4.48–77.72;

p B 0.001) and CD patients (OR, 8.14; 95 % CI,

2.01–33.04; p = 0.003; Hosmer–Lemeshow goodness of

fit v2:9.05, p = 0.249), but APACHE II score was impor-

tant only for CD patients (OR, 1.13; 95 % CI, 1.02–1.26;

p = 0.024; Hosmer–Lemeshow goodness of fit v2:10.71;

p = 0.218). The areas under the ROC curves for APACHE

II score were 0.78 (95 % CI, 0.55–0.89; p = 0.004) for CD

and 0.52 (95 % CI, 0.39–0.66; p = 0.69) for AKI patients

(Figs. 2, 3).

Discussion

Although APACHE II scores were similar between AKI

and CD patients, more AKI patients died than did CD

patients. The result is similar to other studies which showed

higher mortality for AKI patients than CD patients [1, 7, 8,

18–20]. AKI develops in critically ill patients usually in the

context of multiple organ dysfunctions; it shows the

severity of the disease [1, 15, 21]. The rate of MV is high

among AKI patients and studies reported that respiratory

failure requiring MV is a key risk for mortality in these

patients as demonstrated also in our study [11, 18, 19].

Another reason for having high mortality in our AKI

patients was the inclusion of patients with medical condi-

tions. Mortality rate of patients with AKI in medical

diagnosis was reported higher than surgical diagnosis [14,

19, 20]. These patients are more likely to have severe

diseases and need other organ supports like MV and

vasopressors.

The mortality rate of our dialysis-dependent ESRD

patients was higher than other studies [1, 3–8, 22]. This

could be due to admission of severely ill patients, consid-

ering high rate of IMV and mean APACHE II score in our

patients. The studies that reported high mortality in ESRD

patients had rates of MV higher than 50 % [9, 18]. Addi-

tionally, MV was displayed as a mortality predictor for

ESRD patients [1, 18, 19, 23]. We, as well, displayed that

IMV was associated with an eightfold risk of mortality in

CD patients. The other reason for having high mortality

could be due to inclusion of more patients with medical

reasons. Admission for a medical as opposed to a surgical

reason is associated with a worse outcome among ESRD

patients [4, 9, 10, 20]. It was reported that ICU mortality

rate was 61 % for CD patients admitted for medical rea-

sons in comparison to CD patients admitted for surgical

reasons which had 19 % of mortality [9]. In our study, 9 %

of CD patients were admitted to the ICU after elective

surgeries and none of them died.

In this study, the mortality rate of AKI patients with

NKF before ICU admission was higher than acute-on-CKD

and CD patients. Although APACHE II scores were not

different between these three groups, those with past NFK

had high respiratory failure and IMV rates which actually

show the severity of the disease and impact of respiratory

failure on mortality mentioned above. Any critical illness

severe enough to result in AKI requiring RRT in those

patients with earlier NKF would be associated with par-

ticularly high mortality reported also by Walcher et al.

[18].

In the present study, more than 50 % of AKI patients

had the diagnosis of infection and sepsis was found as a

main cause for mortality. Sepsis, a highly prevalent syn-

drome that prompts admission to ICU, is a leading pre-

cipitant of AKI. Between 50 and 80 % of AKI is

considered associated with sepsis and presence of sepsis is

correlated with the severity of AKI and mortality as also

seen in our study [7, 11, 15, 18]. Infection is also reported

as the principal admission diagnosis and mortality reason

in ESRD patients admitted to ICU [6, 7, 24]. This can be

because of high susceptibility of ESRD patients to infec-

tion due in part to the acquired immune deficiency state of

Fig. 1 The relationship between the number of non-renal organ

system failures and mortality among all patients who had RRT in the

ICU

Clin Exp Nephrol

123

uremia, advanced age, comorbid conditions such as DM,

and frequent and repetitive exposure to potential infectious

risk factors during the normal course of dialysis therapy

[25]. In this study, primary admission and mortality diag-

noses were, as well, infection in CD patients.

When we compared patients according to comorbidities,

cancer was found to be more common in AKI patients.

AKI, frequently seen in critically ill cancer patients,

reported to be between 32 and 54 % [26, 27], and 32 % of

them had RRT during ICU stay [27]. Although mortality in

Table 3 Differences between

AKI and CD patients

Bold values indicate statistically

significant results

AKI acute kidney injury,

APACHE acute physiologic and

chronic health evaluation, CVD

cardiovascular disease, CD

chronic dialysis, COPD chronic

obstructive pulmonary disease,

DM diabetes mellitus, CVVHD

continuous venovenous

haemodialysis, IHD intermittent

haemodialysis, IMV invasive

mechanical ventilation, LOS

length of ICU stay, NIMV

noninvasive mechanical

ventilation, N/A not applicable,

n number, PD peritoneal

dialysis, SD standard deviationa Others include vasculitis (7),

COPD exacerbation (7),

pancreatitis (4), intoxication (1)

and HELLP syndrome (1)b LOS covers 59 patientsc Hypotension during dialysis

session encloses 42 patientsd Non-renal organ system

failure covers 106 patientse Uremia describes indication

of RRT due to worsening of

kidney function for any reason

except for sepsis for AKI

patients and defines urgent

dialysis due to any cause except

for hyperkalemia for CD

patients

Parameters AKI (n = 115) CD (n = 101) p

Sex, % men 53 49.5 0.60

Age, median (25–75 percentiles) 67 (55–74) 65 (53–72) 0.09

APACHE II, mean (SD) 27.8 (7.9) 27.5 (9) 0.50

Admission diagnosis (%)

Infection 54.8 30.7 N/A

Cardiac disease 9.6 27.6

Bleeding 7.8 12.9

Uremia 12.1 2

Postoperative follow-up 0.9 8.9

Epilepsy 0 8.9

Hepatic failure 5.2 0

Othersa 9.6 9

Comorbidity (%)

CVD 48.7 51.5 0.68

DM 31.3 29.7 0.79

Cancer 27.8 7.9 <0.001

COPD 9.6 6.9 0.48

Liver disease 11.3 5.9 0.16

NIMV (%) 32.2 25.7 0.29

IMV (%) 73 51.5 0.002

LOS, median (25–75 percentiles)b 8 (2–21) 4.5 (2–20) 0.65

Hypotension, n (%)c 11 (50) 9 (45) 0.74

Non-renal organ failure, (%)d

Respiratory 55.4 76.5 0.001

Cardiovascular 55.4 61.7 0.49

Haematologic 33.9 26.4 0.39

Hepatic 23.2 13.2 0.17

Neurological 22.2 21.4 0.92

Type of dialysis (%)

IHD 67 72.2 N/A

CVVHD 33 14.9

PD 0 12.9

Dialysis indication (%)

Regular dialysis 0 73.2 N/A

Sepsis-related AKI 50.4 0

Uremiae 40.1 13.9

Hypervolemia 7.8 10.9

Hyperkalemia 1.7 2

Mortality (%) 67.8 50.5 0.010

Cause of mortality (%)

Sepsis 89.8 70.6 N/A

Cardiac disease 1.3 23.5

Liver failure 7.6 0

Bleeding 1.3 5.9

Clin Exp Nephrol

123

cancer patients with AKI requiring RRT was reported

between 64 and 78 % [27, 28], the presence of cancer was

not able to be shown as a prognostic factor [28], except for

uncontrolled cancers [27]. Likewise, we failed to show

cancer as an important mortality predictor in the multi-

variate analysis although it was significant in the univariate

Table 4 Comparison of the

patients according to past

kidney function

Bold values indicate statistically

significant results

APACHE acute physiologic and

chronic health evaluation, CVD

cardiovascular disease, CKD

chronic kidney disease, COPD

chronic obstructive pulmonary

disease, DM diabetes mellitus,

D dialysis, IMV invasive

mechanical ventilation, LOS

length of ICU stay, NIMV

noninvasive mechanical

ventilation, NKF normal kidney

function, N/A not applicable,

n number, SD standard

deviationa LOS covers 59 patientsb Hypotension during dialysis

session includes 42 patientsc Non-renal organ system

failure encloses 106 patients

Parameters NKF

(n = 70)

CKD stages 2–5

(n = 45)

CKD stage 5D

(n = 101)

p

Sex, % men 52.9 53.1 49.5 0.87

Age, median (25–75

percentiles)

65.5 (52–73) 70 (64.5–75) 65 (53–72) 0.048

APACHE II,mean (SD) 27.9 (8.5) 27.7 (6.9) 27.5 (9) 0.95

Comorbidity (%)

CVD 41.4 60 51.5 0.14

DM 22.9 44.4 29.7 0.048

Cancer 25.7 31.1 7.9 0.001

COPD 7.1 13.3 6.9 0.39

Liver disease 15.7 4.4 5.9 0.044

NIMV (%) 32.9 31.1 25.7 0.57

IMV (%) 77.1 66.7 51.5 0.002

LOS, median (25–75

percentiles)a10 (2–22) 6 (4–14) 4.5 (2–20) 0.583

Hypotension, n (%)b 8 (50) 3 (50) 9 (45) 0.94

Non-renal organ failure (%)c

Respiratory 84.3 64.4 55.4 <0.001

Cardiovascular 55.6 55 61.7 0.788

Haematologic 36.1 29.4 26.4 0.61

Hepatic 28.2 11.8 13.2 0.13

Neurological 24.3 17.6 21.4 0.85

Mortality (%) 78.6 51.1 50.5 <0.001

Cause of mortality (%)

Sepsis 87.2 95.7 70.6 N/A

Cardiac disease 1.8 0 23.5

Liver failure 9.2 4.3 0

Bleeding 1.8 0 5.9

Table 5 Results of multivariate models

Variables B p Odds ratio (95 % CI)

All patients

Liver disease 2.212 0.069 9.13 (0.84–98.81)

COPD 1.945 0.108 6.99 (0.65–75.10)

IMV 2.695 \0.001 14.81 (5.47–40.05)

NKF 1.019 0.041 2.77 (1.04–7.37)

APACHE II 0.064 0.053 1.06 (0.99–1.13)

AKI patients

IMV 2.927 \0.001 18.66 (4.48–77.72)

CD patients

IMV 2.097 0.003 8.14 (2.01–33.04)

APACHE II 0.125 0.024 1.13 (1.02–1.26)

AKI acute kidney injury, APACHE acute physiologic and chronic

health evaluation, CVD cardiovascular disease, CD chronic dialysis,

COPD chronic obstructive pulmonary disease, CI confidence interval,

IMV invasive mechanical ventilation, NKF normal kidney function

Fig. 2 The ROC curve for sensitivity and specificity of APACHE II

score in ESRD patients [0.78 (95 % CI 0.55–0.89); p = 0.004]

Clin Exp Nephrol

123

analysis. In this study, all AKI patients who were admitted

to the ICU with liver failure except one died. This was

because of having severely ill patients. The development of

AKI in these patients is a poor prognostic criterion with

reported mortality rates of between 52 and 91 % [29, 30].

Another reason for having high mortality could be through

failures in providing a liver-transplant for the patients due

to the shortage of organs and less transplant units. Mor-

tality was demonstrated to be 98 % in liver failure patients

who had CVVHD but no liver transplantation [31].

The primary limitation of this study is its retrospective

design conducted at a single medical center and its small

sample size. As such, the results might lack wider appli-

cability. Second, we did not examine AKI patients who did

not require RRT and patients without AKI. Therefore, we

can not generalize our findings to all ICU patients. Third,

information available from the early ICU database was not

sufficient to allow full collection of data on variables of

organ failure systems, LOS and hypotension during RRT.

This could be a reason for failing to show a difference

between AKI and CD patients along these variables.

Fourth, due to insufficient database, we could not study

hospital mortality, hospital LOS and renal survival for AKI

patients at the time of hospital discharge. Thus, the out-

comes for the patients after ICU discharge are unknown.

In conclusion, we have shown that overall mortality is

high among patients requiring RRT in the ICU. Although

disease severity score is the same between AKI and CD

patients, the rates of IMV and mortality are significantly

high in AKI patients. AKI patients with NKF before ICU

admission have an increased risk of dying than do patients

with acute-on-CKD and CD patients. IMV is an important

prognostic predictor for both AKI and CD patients,

whereas APACHE II is displayed as important only for CD

patients.

Acknowledgments We thank Dr. Pınar Ay, M.D., Assoc. Prof. of

Health Care and Statistic at Marmara University, School of Medicine,

for his help with the statistical analysis. T. Akbas designed the study,

gathered the raw data of the study and contributed to the writing of

the paper. Sait Karakurt helped the study design, reviewed the paper

and contributed to the writing of the manuscript. Serhan Tuglular

designed the study, reviewed the paper and contributed to the writing

of the manuscript. All authors read, edited and ultimately approved

the final manuscript.

Conflict of interest All authors have declared that no conflict of

interest exists.

References

1. Strijack B, Mojica J, Sood M, et al. Outcomes of chronic dialysis

patients admitted to the intensive care unit. J Am Soc Nephrol.

2009;20:2441–7.

2. Metnitz PGH, Moreno RP, Almeida E, et al. SAPS 3—from

evaluation of the patient to evaluation of the intensive care unit.

Part 1: objectives, methods and cohort description. Intensive Care

Med. 2005;31:1336–44.

3. Uchino S, Morimatsu H, Bellomo R, Silvester W, Cole L. End-

stage renal failure patients requiring renal replacement therapy in

the intensive care unit: incidence, clinical features, and outcome.

Blood Purif. 2003;21:170–5.

4. Hutchison CA, Crowe AV, Stevens PE, Harrison DA, Lipkin

GW. Case mix, outcome and activity for patients admitted to

intensive care units requiring chronic renal dialysis: a secondary

analysis of the ICNARC Case Mix Programme Database. Crit

Care. 2007;11:R50.

5. Dara SI, Bekele A, Bajwa AA, Albright RC. Outcome of patients

with end-stage renal disease admitted to the intensive care unit.

Mayo Clin Proc. 2004;79:1385–90.

6. Manhes G, Heng AE, Aublet-Cuvelier B, Gazuy N, Deteix P,

Souweine B. Clinical features and outcome of chronic dialysis

patients admitted to an intensive care unit. Nephrol Dial Trans-

plant. 2005;20:1127–33.

7. Rocha E, Soares M, Valente C, et al. Outcomes of critically ill

patients with acute kidney injury and end-stage renal disease

requiring renal replacement therapy: a case-control study.

Nephrol Dial Transplant. 2009;24:1924–30.

8. Clermont G, Acker CG, Angus DC, Sirio CA, Pinsky MR,

Johnson JP. Renal failure in the ICU: comparison of the impact of

acute renal failure and end-stage renal disease on ICU outcomes.

Kidney Int. 2002;62:986–96.

9. Chapman RJ, Templeton M, Ashworth S, Broomhead R, McLean

A, Brett SJ. Long-term survival of chronic dialysis patients fol-

lowing survival from an episode of multiple-organ failure. Crit

Care. 2009;13:R65.

10. Allegretti AS, Steele DJ, David-Kasdan JA, Bajwa E, Niles JL,

Bhan I. Continuous renal replacement therapy outcomes in acute

kidney injury and end stage renal disease: a cohort study. Crit

Care. 2013;17:R109.

11. Uchino S, Kellum JA, Bellomo R, et al. Acute renal failure in

critically ill patients: a multinational, multicenter study. JAMA.

2005;294:813–8.

12. De Mendonca A, Vincent JL, Suter PM, et al. Acute renal failure

in the ICU: risk factors and outcome evaluated by the SOFA

score. Intensive Care Med. 2000;26:915–21.

Fig. 3 The ROC curve for sensitivity and specificity of APACHE II

score in AKI patients [0.52 (95 % CI, 0.39–0.66); p = 0.69]

Clin Exp Nephrol

123

13. Chertow GM, Levy EM, Hammermeister KE, Grover F, Daley J.

Independent association between acute renal failure and mortality

following cardiac surgery. Am J Med. 1998;104:343–8.

14. Abosaif NY, Tolba YA, Heap M, Russell J, El Nahas AM. The

outcome of acute renal failure in the intensive care unit according

to RIFLE: model application, sensitivity, and predictability. Am J

Kidney Dis. 2005;46:1038–48.

15. Avasthi G, Sandhu JS, Mohindra K. Acute renal failure in

medical and surgical intensive care units—a one year prospective

study. Ren Fail. 2003;25:105–13.

16. Levey AS, Eckardt KU, Tsukamoto Y, et al. Definition and

classification of chronic kidney disease: a position statement from

kidney disease: improving global outcomes (KDIGO). Kidney

Int. 2005;67:2089–100.

17. Van Steen K, Curran D, Kramer J, et al. Multicollinearity in

prognostic factor analyses using the EORTC QLQ-C30: identi-

fication and impact on model selection. Statist Med. 2002;21:

3865–84.

18. Walcher A, Faubel S, Keniston A, Dennen P. In critically ill

patients requiring CRRT, AKI is associated with increased

respiratory failure and death versus ESRD. Ren Fail. 2011;33:

935–42.

19. Ostermann M, Chang R. Renal failure in the intensive care unit:

acute kidney injury compared to end-stage renal failure. Crit

Care. 2008;12:432–3.

20. Bagshaw SM, Mortis G, Doing CJ, Godinez-Luna T, Fick GH,

Laupland KB. One-year mortality in critically ill patients by

severity of kidney dysfunction: a population-based assessment.

Am J Kidney Dis. 2006;48:402–9.

21. Steinvall I, Bak Z, Sjoberg F. Acute kidney injury is common,

parallels organ dysfunction or failure, and carries appreciable

mortality in patients with major burns: a prospective exploratory

cohort study. Crit Care. 2008;12:R124.

22. Senthuran S, Bandeshe H, Ranganathan D, Boots R. Outcomes

for dialysis patients with end-stage renal failure admitted to an

intensive care unit or high dependency unit. Med J Aust.

2008;188:292–5.

23. Juneja D, Prabhu MV, Gopal PB, Mohan S, Spridhar G, Nayak

KS. Outcome of patients with end stage renal disease admitted to

an intensive care unit in India. Ren Fail. 2010;32:69–73.

24. Sood MM, Miller L, Komenda P, et al. Long-term outcomes of

end-stage renal disease patients admitted to the ICU. Nephrol

Dial Transplant. 2011;26:2965–70.

25. Arulkumaran N, Annear NMP, Singer M. Patients with end-stage

renal disease admitted to the intensive care unit: systematic

review. Br J Anaesth. 2013;110:13–20.

26. Liborio AB, Abreu KL, Silva GB Jr, et al. Predicting hospital

mortality in critically ill cancer patients according to acute kidney

injury severity. Oncology. 2011;80:160–6.

27. Soares M, Salluh JI, Carvalho MS, Darmon M, Rocco JR, Spector

N. Prognosis of critically ill patients with cancer and acute renal

dysfunction. J Clin Oncol. 2006;24:4003–10.

28. Maccariello E, Valente C, Nogueira L, et al. Outcomes of cancer

and non-cancer patients with kidney injury and need of renal

replacement therapy admitted to general intensive care units.

Nephrol Dial Transplant. 2011;26:537–43.

29. De Carvalho JR, Villela-Nogueira CA, Luiz RR, et al. Acute

kidney injury network criteria as a predictor of hospital mortality

in cirrhotic patients with ascites. J Clin Gastroenterol.

2012;46:e21–6.

30. Cholongitas E, Senzolo M, Patch D, Shaw S, O’Beirne J, Bur-

roughs AK. Cirrhotics admitted to intensive care unit: the impact

of acute renal failure on mortality. Eur J Gastroenterol Hepatol.

2009;21:744–50.

31. Fraley DS, Burr R, Bernardini J, Angus D, Kramer DJ, Johnson

JP. Impact of acute renal failure on mortality in end-stage liver

disease with or without transplantation. Kidney Int. 1998;54:

518–24.

Clin Exp Nephrol

123