The most frequent contributing factor to acute kidney injury (AKI) is sepsis and multi-organ dysfunction syndrome (MODS) [1]. The high mortality rate of patients with AKI reaches 60% despite considerable improvement of renal replacement therapy (RRT). This fostered the interest on the impact of AKI on distant organ function. One interesting interaction is the one between lung and kidney. AKI induces increased lung vascular permeability, cellular inflammation, and dysregulated salt and water channels resulting in respiratory failure [2]. Also in patients with non-severe pneumonia AKI is associated with higher immune response and an increased risk of death [3]. In contrast to several studies investigating the effect of AKI on survival of patients undergoing extracorporeal membrane oxygenation [4, 5], there is only scarce data on the effect of AKI on the survival of patients undergoing pumpless interventional lung assist (iLA membrane ventilator; Novalung, Talheim, Germany) treatment. This technique is increasingly used by intensivists in patients with life-threatening respiratory failure or acute respiratory distress syndrome (ARDS) suffering from persistent hypercapnia. The 28 day survival of patients undergoing iLA therapy ranges from 41% in a retrospective analysis [6] to 51% in a prospective cohort study in patients with ARDS from multiple etiologies [7]. However, those studies disregarded patients with AKI on renal replacement therapy (RRT). Only two studies provide short term survival data on patients undergoing iLA membrane ventilator therapy also suffering from AKI. In a retrospective analysis by Liebold et al. 39% of the iLA patients required intermittent haemodialysis [8]. In a study by Floerchinger et al. 42% of the patients were treated by continuous veno-venous hemofiltration (CVVH) [9]. So far it is unknown, what impact extended dialysis (ED), an increasingly used RRT method in the ICU [10, 11], has on long term survival of patients undergoing iLA treatment. Moreover, it is not known whether the improvement in respiratory acidosis has any impact on renal function in these patients. Hence, the aim of our study was to investigate the effect of AKI on long term mortality of patients treated by iLA. Furthermore the potential impact of iLA treatment on renal function was investigated in this single centre retrospective study.

This retrospective cohort study included all adult patients that underwent iLA treatment from Januar 1st 2005 to December 31st 2009 at a tertiary care university hospital. The Institutional Review Board of the Medical School Hannover waived the need for approval and informed consent for this analysis. A total of 61 patients were identified in the central documentation system of the hospital. From all of those patients charts as well as laboratory data were reviewed. RRT dependence was defined as AKIN stage 3. All patients requiring RRT in addition to iLA treatment received ED using the GENIUS system. Details of the system are summarized elsewhere [11].

Estimated glomerular filtration rate (eGFR) was calculated using the CKD-EPI formula [12].

We identified 61 patients undergoing iLA treatment within our observation period of five years. Patient characteristics are presented in Table 1. The underlying diseases are summarized in Table 2. More than one third (n = 21, 34%) of all patients required ED. Exactly one third of these patients had already been RRT dependent before (n = 7; 33%) iLA insertion. Half of the patients became RRT dependent on the day of iLA insertion (n = 10; 48%). Within two days after iLA placement another five patients (24%) became RRT dependent. Hence, within two days of iLA insertion 71% (n = 15) of patients ever to become RRT dependent were on ED (Figure 1). This corresponds to roughly 25% of all patients undergoing iLA treatment. Patients undergoing ED treatment were not different in respect to gender, age, weight, iLA support duration, use of nephrotoxic antiobiotics, PaO₂, paCO₂, Horowitz index, MAP, P_max and PEEP at start of iLA support (Table 1) and underlying disease to patients not requiring RRT. Also time from ICU admission to iLA insertion was not different in the two groups. However, patients undergoing ED had a significant lower pH, a higher SOFA and APACHE II score as well as a significant higher iLA flow 2 h after implantation than patients not requiring RRT (p < 0.05) (Table 1). The 28 day survival of the whole cohort was 23%. There was a clear difference in survival between patients requiring ED and those not requiring RRT. While patients undergoing ED exhibited a 19% 28 day survival, the survival rate at day 28 in patients without need for dialysis was 40% (p = 0.001) (Figure 2). This difference became even more marked after one year. While patients undergoing ED had a 5% one year survival, patients without the need for RRT had a 25% one year survival (p < 0.001). Patients in whom iLA was placed as a bridge to lung transplantation and that were successfully transplanted showed a markedly higher 28 day survival rate of 78%, while all of the patients listed for lung transplantation without subsequently receiving a graft died before day 25 (Figure 3). In contrast to the whole cohort, the necessity of RRT had no effect on survival in the transplant candidates.

	all	iLA and ED	iLA - no ED	p
patients	61	21	40
male/female	30 | 31	9 | 12	21 | 19
age (years)	40 (28 to 52)	36 (28 to 52)	41 (28 to 49)	0.942
iLA treatment (days)	6 (3 to 11)	5 (2 to 9)	7 (5 to 13)	0.191
iLA-flow after 2 h (l/min)	1.1 (1.0 to 1.4)	1.4 (1.3 to 1.6)	1.1 (1.0 to 1.1)	0.009
days on ICU until iLA	7 (3 to 15)	8 (5 to 16)	7 (3 to 12)	0.880
paO2 (mmHg)	82 (70 to 98)	77 (70 to 96)	82 (70 to 100)	0.97
after 24 h iLA	68 (65 to 84)	71 (62 to 90)	78 (68 to 83)	0.900
paCO2 (mmHg)	100 (81 to 115)	98 (81 to 114)	100 (79 to 115)	0.880
after 24 h iLA	52 (43 to 61)	44 (53 to 59)	56 (45 to 61)	0.840
pH	7.15 (7.08 to 7.26)	7.10 (7.04 to 7.15)	7.23 (7.13 to 7.27)	0.007
after 24 h iLA	7.38 (7.31 to 7.46)	7.34 (7.26 to 7.35)	7.41 (7.33 to 7.46)	0.013
Horowitz index (PaO2/FiO2)	96 (75 to 137)	98 (77 to 114)	94 (67 to 144)	0.280
MAP (mmHg)	80 (75 to 86)	80 (70 to 85)	80 (78 to 87)	0.249
Pmax (cmH2O)	38 (34 to 41)	38 (35 to 40)	39 (32 to 42)	0.747
PEEP (cmH2O)	8 (5 to 10)	9 (5 to 13)	8 (5 to 10)	0.370
Creatinine* (μmol/l)	53 (40 to 91)	93 (49 to 174)	50 (37 to 73)	0.005
after 24 h iLA			52 (34 to 78)
eGFR* (ml/min/1,73 m2)	120 (79 to 138)	66 (33 to 138)	124 (100 to 140)	0.003
SOFA-Score	11 (10 to 12)	14 (11 to 17)	11 (10 to 11)	< 0.0001
APACHE II - Score	32 (28 to 34)	32 (31 to 38)	31 (28 to 33)	0.023
Vancomycin use day 1-3 of iLA	32%	40%	29%
Aminoglycoside use day 1-3 of iLA	26%	13%	32%
28 day survival	33%	19%	40%
one year survival	18%	5%	25%

Data presented as median (interquartile ranges). *Creatinine and eGFR only from patients without ED dependence before iLA insertion.

underlying disease	n	28 day survival	one year survival
cystic fibrosis	15	27%	27%
pneumonia	10	40%	20%
status post lung Tx	8	25%	13%
pulmonary fibrosis	7	43%	0%
COPD	7	29%	0%
lung cancer	2	50%	0%
pulmonary hypertension	2	100%	50%
leukemia	2	50%	50%
polytrauma	2	50%	50%
others	6	17%	17%

Not surprisingly, the median eGFR of patients with no current or future need of ED on admission to the ICU was 124 (100 to 140) [ml/min/1.73 m²]. Patients which became RRT dependent had a significant lower eGFR (66 (33 to 138) [ml/min/1.73 m²]). Despite the improvement in pH and pCO₂ the insertion of iLA had no effect on renal function in patients that did not become RRT dependent (Table 1).

The pertinent findings of our study were that 1) the combination of iLA treatment and ED is associated with an increased short and long term mortality 2) most of the patients with this combination either were dialysis dependent before the start of iLA treatment or within 48 hours after start of iLA therapy 3) the bridging to successful lung transplantation was associated with the best 28 day survival 4) decreased eGFR at the time on the day of iLA institution was predictive of future dialysis dependence.

The combination of iLA treatment with acute kidney injury requiring renal replacement therapy dramatically increases mortality. Most of the patients that become dialysis dependent do so within 48 hours of iLA insertion. These patients have already a decreased eGFR at the time of ICU admission. The best outcome in our cohort could be obtained in patients that were successfully bridged to transplantation.