Information available on pediatric patients requiring continuous renal replacement therapy (CRRT) is scant and partially deriving from the adult critically ill patients: in many cases pediatric CRRT (pCRRT) is prescribed according to local expertise and without the utilization of any specific recommendation. The prospective pCRRT (ppCRRT) registry was created in order to improve and increase observational information on pCRRT practice of several US centers and it is currently the only source, worlwide, which gathered the experience derived from hundreds of treated children [1]. The ppCRRT registry was founded in 2001 and comprises 13 pediatric centers in the USA. So far, the ppCRRT enrolled 344 patients with age ranging from 1 day to 25 years and weights ranged from 1.3 to 160 kg; 11 different primary diagnoses were described in the registry the most represented of which are sepsis, stem cell transplantation, cardiac disease, liver disease, and malignancies [1]. The overall survival of these patients was 58 % and it appeared to be significantly reduced in patients with multiple organ dysfunction syndrome and fluid overload, weighing less than 10 kg or receiving stem cell transplantation [1]. Provided this unique set of information (and other related to technical aspects such as position and size of dialysis catheters, the association with filter lifespan or anticoagulation) the ppCRRT registry left several questions unanswered: optimal timing of pCRRT has still to be clarified (although it is now clear that it should be considered before fluid is accumulated in children), optimal dosing of CRRT prescription is currently unknown (even if, after adult experience and the possibility to easily achieve highest doses in such small patients, this may not be an issue significantly affecting outcomes) and, above all, long-term outcomes of acute kidney injury (AKI) children undergoing CRRT are still to be explored (it is possible that patients surviving a CRRT treatment suffer a reduced renal reserve with a significant risk of becoming chronic renal failure patients).

Although the analysis of timing of pCRRT may appear difficult due to the lack of prospective evidence, the lack of absolute indications for dialysis inception and the lack of a common definition of “timing”, as a common sense rule “timely” CRRT start has been advocated by several authors [2]. It seems particularly important that the pCRRT start occurs before the fluid overload threshold of 10–20 % has been reached in an AKI pediatric patient [3]. A recent small retrospective observation of pCRRT patients undergoing extracorporeal membrane oxygenation (ECMO) showed that not only mortality is affected by the level of fluid overload achieved at the time of CRRT start, but also that it seems hard to “force” fluid removal in order to improve patient outcomes [4]. In light of this, a post hoc analysis of the RENAL trial showed that apart from what level of fluid overload is reached at CRRT start (in adult patients), mortality was affected by the level of fluid balance reached in the first 48 h: critically ill adult patients tended to survive more if a negative level of fluid balance was rapidly achieved [5]. In a large retrospective 10-year cohort analysis, a thorough analysis of timing of pCRRT initiation was recently attempted [6]. The authors defined timing as the time from intensive care unit (ICU) admission to CRRT initiation. They found that late initiators (>5 days) had higher mortality than early initiators (≤5 days) with a hazard ratio of 1.56 (95 % confidence interval: 1.02–2.37) with an increase of 5 % in mortality for every day of delay in CRRT initiation after adjusting for significant confounders. On multivariable regression, independent predictors of mortality were also fluid overload, indication for CRRT initiation (simultaneous presence of renal dysfunction and fluid overload), severity of illness at ICU admission, and active oncologic diagnosis. Interestingly, according to pediatric-modified risk, injury, failure, loss, and end stage (pRIFLE) classification, the authors seemed to realize that many severe cases at ICU admission were more likely to be treated earlier (hence, most severe pRIFLE classes were more common in the survivor group); on the other side, if a patient slightly but significantly progressed to higher AKI stages during the course of ICU admission, the treatment was delayed and a correlation between pRIFLE progression and time to CRRT start was found. It is possible that this subgroup of patients may receive the most benefit from an earlier CRRT start. Pediatric CRRT has to be started before 5 days from ICU admission if a patient has AKI in order to be effective in terms of mortality [6].

Currently, the two dialysis modalities most frequently used in infants are peritoneal dialysis (PD) and CRRT. PD is currently the RRT treatment of choice in neonates, unless specific contraindications are present (i.e. peritonitis, abdominal masses, or bleeding) [7]. PD uses peritoneum as a semipermeable membrane to achieve solute diffusion and plasma water ultrafiltration: dialysate is infused through an abdominal catheter and after a period of so-called “dwell time”, waste solution is drained from the abdomen. Typically, in pediatric cardiac surgery neonates, in order to avoid excessive intra-abdominal pressure rise during dialysis solution infusion and to prevent hemodynamic instability, a “low flow” prescription of 10 ml/kg dialysate is recommended [8]. Dwell times may vary from 10 to 30 min according to the need for higher to lower solute clearances. Water ultrafiltration may be regulated by dialysate tonicity (provided by glucose concentration, 1.36–2.5 %). PD is a simple and safe technique, which does not require dedicated technology or a steep learning curve and may be administered by ICU nurses without specific nephrologic expertise. Nonetheless, it must be acknowledged that PD is limited by a relative lack of efficiency especially in water removal which may be particularly relevant in severely overloaded patients. Furthermore, the most important side effects of PD in neonates may be interstitial fluid accumulation in case of suboptimal dialysate drainage, hyperglycemia, and higher risk of peritoneal infection. PD is finally contra-indicated in patients with recent abdominal surgery or abdominal bleeding [7]. However, PD in the post-operative phase may be applied proactively and in a very early phase of oliguria or fluid retention. PD has been recently showed to be associated with improved survival in a large cohort of post-cardiac surgery neonates if started in the first 24 post-operative hours when compared with patients who received PD after the second post-operative day [9].

Extracorporeal dialysis in children can be managed with a variety of modalities, including intermittent hemodialysis, and CRRT, delivered as hemofiltration, hemodialysis or hemodiafiltration [10]. The choice of dialysis modality to be used is influenced by several factors, including the goals of dialysis, the unique advantages and disadvantages of each modality, and institutional resources. Intermittent dialysis may not be well tolerated in infants because of its rapid rate of solute clearance and in particular in hemodynamically unstable pediatric critically ill patients. These children are generally treated by CRRT that seems to better provide both fluid and solute re-equilibration and pro-inflammatory mediators removal. Circuits with reduced priming volume together with monitors providing an extremely accurate fluid balance are still not commercially available [10]. Post-heart surgery patients, in particular, are a peculiar and interesting model of acute water accumulation and inflammation: they receive ultrafiltration soon after cardiopulmonary bypass weaning in order to remove water and inflammatory mediators before the harmful effects of inflammation and fluid overload become clinically relevant [10]. Several studies showed a statistical difference in the percentage of fluid overload of children with severe renal dysfunction requiring RRT. At the time of dialysis initiation, survivors tend to have less fluid overload than non-survivors, especially in the setting of MODS [11]. For this reason, in children now priority indication is given to the correction of water overload. In fact, differently from the adult patients where dialysis dose may play a key role, an adequate water content in small children is the main independent predictor of outcome.