In this study, we demonstrated that a multi-modal approach using PAL-treatment in patients with ALI achieves negative cumulative fluid balance without compromising organ function. Furthermore, compared to a matched control group, we found improved oxygenation, EVLWI and IAP reduction, fewer days on mechanical ventilation, shorter ICU stay, and reduced 28-day mortality in the PAL group.
The idea behind PAL-treatment aiming for negative fluid balance in a setting of capillary leak is based on the recently rediscovered concept of the ebb and flow phase [2, 13, 28, 29]. The ebb phase represents a distributive shock characterized by increased capillary permeability and albumin leak [2, 28, 30]. Excess interstitial fluid leads to organ dysfunction , including ALI, secondary IAH , and associated acute kidney injury . Shock reversal and subsequent hemodynamic and renal recovery sets in the transition to the flow phase resulting in mobilization of excess extravascular (lung) water . Previously, a neutral to negative cumulative balance [1, 7, 10, 13, 33, 34] and reduction of EVLWI were shown to correlate with improved survival [15, 16, 18, 19].
PAL-treatment intends to initiate the flow phase, limiting capillary leak and promoting interstitial fluid removal while ensuring organ perfusion at the same time. Therefore, it is a specific form of restrictive fluid management, combining open lung ventilation, small volume resuscitation with hyperoncotic albumin, and aggressive fluid removal.
Open lung ventilation strategy in ALI signifies application of high levels of PEEP  and is correlated with decreases in EVLWI . Both the percentage of potentially recruitable lung and EVLWI are related to outcome [15, 16, 18, 19, 35]. The open lung strategy in ALI is associated with increased alveolar fluid clearance and reduced EVLWI [37–39]. In this study, PAL-treatment was initiated in patients with low oxygenation index and high EVLWI, potentially indicating a higher proportion of recruitable lung. Within the concept of the polycompartment syndrome, we set PEEP level (cmH2O) equal to IAP (mmHg) in order to counteract IAP  (Figure 3).
Induction of the flow phase with PAL-treatment implies vascular refilling from the interstitium and subsequent removal of fluids from the body producing a net negative fluid balance. In this context, addition of small volume resuscitation with hyperoncotic albumin to a fluid removal regimen resulted in a greater negative fluid balance while maintaining better hemodynamic stability . Moreover, restoration of colloid osmotic pressure in absence of elevated hydrostatic pressure may prevent further interstitial edema formation [41, 42]. Accordingly, hypoproteinemia is highly predictive of positive fluid balance and development of ALI in patients with sepsis . In view of PAL-treatment as a therapy for capillary leak, beneficial effects of albumin on the microcirculation may be of particular interest, attenuating capillary permeability and pulmonary inflammation [44–47].
PAL-treatment proved to be effective to achieve negative fluid balance. After 1 week, PAL-treated patients had a net negative cumulative fluid balance while control patients added up to a positive cumulative fluid balance, similar to other cohorts [12, 20]. In contrast to previous studies, negative fluid balance was pursued as a specific goal. One week of PAL-treatment did not significantly worsen cardiovascular function. Yet, more patients required vasopressor therapy and administered doses were higher. Furthermore, although renal SOFA score on each day was similar in both groups, the PAL group had higher average creatinine for the observed week. In line with previous reports, successful restrictive fluid strategy with PAL-treatment led to improved oxygenation and shortened duration of MV [12, 21]. Moreover, PAL-treated patients had a significant greater reduction of EVLWI. This observation possibly reflects improved healing of lung injury, better shock reversal with transition to the flow phase [15, 16, 18, 19]. We found PAL-treated patients to have a significant reduction of CLI as a result of restoration of serum albumin.
Overzealous fluid therapy in a setting of capillary leak is an important risk factor for IAH, associated with organ failure and increased mortality . Therefore, a fluid strategy aimed at negative fluid balance and avoiding crystalloid over-resuscitation may play an important role in preventing and even treating IAH . In this context, our observations demonstrated a significant increase of IAP in controls, whereas IAP dropped in PAL-treated patients.
Our study has several important limitations. First, the use of historical controls may raise difficulties to ensure that obtained differences in outcome are related to the studied treatment. Indeed, we cannot deny that continuous evolving standard care has led to better outcomes in patients with acute lung injury . In particular, there are indices that a more protective ventilation (not only higher PEEP as per protocol) was applied in the PAL group since at baseline controls had slightly higher tidal volumes and lower PEEP levels. However, as patients were selected for PAL-treatment based on low oxygenation indices and high EVLWI, they may have had a higher percentage of potentially recruitable lung, requiring higher PEEP levels .
Second, the large difference in mortality between the two groups has to be placed in context. Expected mortality in the control group and PAL group was 47% and 48%, computing a standardized mortality ratio of 1.04 and 1.70, respectively. Apart from the presumed better standard care over time, a selection bias may have been introduced by including only mechanically ventilated patients with thermodilution catheter monitoring. Thus, we selected a specific case mix of severely ill ALI patients prone to exhibit fluid retention, in which attention to fluid balance may be expected to have great potential benefit. In this regard, we note a considerable high cumulative fluid balance after 1 week in controls (8,027 mL), albeit similar to earlier reports [12, 20]. Third, since this was an open trial in which fluid therapy decisions were made by the treating physician, the lack of a strict protocol to guide fluid therapy may have introduced bias. Fourth, our database did not supply detailed information on amounts of fluids administrated in the first 6 h. The fluid balance on the day before enrolment was almost 1 L higher in controls, possibly indicating a more aggressive initial volume replacement. Exact data on the type of fluid used and the rate of hourly ultrafiltration in patients with CRRT were not recorded either.