Our analysis of UF data delivered by RPM in APD patients treated at home revealed lower net UF during the first cycle of APD session in patients using a dry day regime (APDDD) as compared to those on a wet day regime (APDWD); and some APDDD patients had negative net UF during the first APD exchange. Peritoneal transport modelling indicated that a likely explanation for this phenomenon is that the daytime exchange influenced the hydration state of the peritoneal membrane and that this affected water removal during the initial exchange(s) of the subsequent APD session. Thus, the observed initial lower efficiency of water removal in patients on the APDDD regime appears to be related to a relative increase of peritoneal tissue hydration during the first APD cycles following a dry day.
The acquisition of clinical data relied on the advent of RPM, which was designed to track, on a daily basis, factors such as blood pressure, body weight, dialysis treatment characteristics including ultrafiltration, and patient´s adherence to dialysis prescriptions in patients undergoing APD at home 8, and which offers unique opportunities for studies of net UF of APD exchanges cycle-by-cycle. To the best of our knowledge, this is the first detailed study of UF patterns during APD using this technology.
Lower UF volumes during the initial as compared to subsequent cycles was demonstrated in APDDD but not in APDWD when analysing drain volume ratios (of initial cycle volume to subsequent cycle volumes): in APDDD, but not in APDWD, VdrR1 was lower than VdrR2 and both ratios were significantly lower than 1 (Table 4). Moreover, VdrR was lower in APDDD than in APDWD indicating that the daytime exchange influenced water removal during the subsequent APDDD session. We have not identified any published studies describing this phenomenon.
To further explore why APDDD patients had lower net UF during the first cycles of APD, we performed numerical simulations based on the distributed model 2,7 for a typical APDDD patient with standard prescription (vide supra, and Table 2). These simulations provided results which fit well with clinical data. Thus, while the simulated net UF of the daytime exchange was − 110 mL and 458 mL for the whole night APD session, the corresponding values in the subgroup of APDDD patients was − 188 mL and 440 mL respectively.
Using the applied model, we explored changes in the hydration status of the peritoneal tissue, which is typically not considered in physiological interpretations, and in most mathematical models of the peritoneal transport it is assumed constant 3,7,11,12. However, water reabsorption from the peritoneal cavity (driven by the high intraperitoneal pressure gradient) to the adjacent tissue layers may correspond to more than 70% of total peritoneal fluid absorption 13. The water inflow into the peritoneal tissue induces an increase of interstitial pressure and tissue hydration over their physiological levels. Such an increase, observed also in inflammation, facilitates water and solute transport by changing transport properties of the peritoneal membrane. During the treatment, the interstitial pressure close to the peritoneal cavity increases to equilibrate with the intraperitoneal pressure. Such an increase of interstitial pressure and the corresponding tissue hydration in the layer close to the peritoneal cavity is predicted by the applied distributed model but was also observed experimentally in an animal model of PD and in basic physiology studies 2,6,7,11,14,15.
During a dry day APD regime (APDDD), when the peritoneal cavity is empty or almost empty, intraperitoneal pressure remains lower than the interstitial pressure in the adjacent tissue layers, leading to slow leakage of water accumulated in the tissue into the peritoneal cavity. Consequently, peritoneal tissue hydration decreases to the physiological (lower) state of tissue hydration. At the beginning of APDDD session, the infusion of hypertonic dialysis fluid not only induces ultrafiltration into the peritoneal cavity due to the osmotic force. Besides, a part of the water remains in the tissue and increases its hydration, adapting to the PD condition by equilibrating with the fluid in the peritoneal cavity. In contrast, during a wet day APD regime (APDWD), tissue hydration during the day remains unphysiologically high (as is typical for continuous forms of PD) due to the presence of dialysate throughout the day, resulting in elevated intraperitoneal pressure during daytime. Therefore, any decrease of tissue hydration after a daytime exchange is negligible and there is only a minor (within measurement error) impact on water removal during the subsequent initial APD cycles.
Some strengths and limitations of the study should be noted. Strengths include the availability of detailed data of net UF cycle-by-cycle provided by RPM in two similar groups of patients using the two APD regimes and the long observation period (16 days) in each patient. However, one cannot exclude that at least part of the UF variability between cycles, as observed in our study, might be related also to other factors that are not considered in the applied model, such as changes in the residual peritoneal volume (due to differences in drainage), physical activity of the patients that influences intraperitoneal pressure (such as coughing, patient’s posture during drainage), diuresis, and overall volume status. Therefore, in contrast to solute removal which can be predicted very precisely by available peritoneal transport models, peritoneal ultrafiltration is difficult to predict.
The low number of patients in each of the studied groups is another limitation; further investigations of larger cohorts of patients are warranted to confirm our findings. One problem with the selection of patients for this kind of clinical study is that many patients using cyclers change glucose concentration of dialysis fluid cycle-to-cycle during the night exchanges due to changed prescription, and therefore only a rather low number of patients use the same glucose concentration in all cycles. In our study only patients using dialysis fluid with constant glucose concentration were investigated. Moreover, although both studied groups had statistically similar characteristics, there was a tendency for faster transport status in APDDD than in APDWD group, that might have a slight effect on the comparison of both groups but not the results in each group separately. Furthermore, a difference between the groups is that all but one of the APDWD patients used icodextrin-based dialysis fluid during the long day dwell that significantly increases net UF 16–20 and may have a different impact on peritoneal tissue hydration. Nevertheless, we believe that presented results supported by the mathematical modelling for a typical patient indicate a possible reason for at least part of the observed variability in the water removal between consecutive APD cycles. It should be noted however that there was no difference between the two APD regimes concerning the total obtained volume of net UF from the night and day exchanges and 24 h UF, Table 3.
In summary, our investigation of ultrafiltration cycle-by-cycle during APD assessed by remote monitoring of patients treated at home revealed that compared with a “wet day” APD regime (APDWD), patients on a “dry day” APD regime (APDDD) had lower water removal during the first as compared to subsequent APD cycles, and in some cases even negative UF. According to peritoneal transport modelling, the observed difference in initial UF between the two regimes may reflect local adaptation of the peritoneal tissue to the presence (APDWD) vs. absence (APDDD) of dialysis fluid in the peritoneal cavity during the long daytime dwell. During the dry day (APDDD), peritoneal tissue hydration decreases towards a physiological level of about 18% and then increases during the first APD cycle to about 36% (in tissue close to the peritoneal cavity), a level typical for conditions with continuous presence of dialysis fluid such as in APDWD. We conclude that the increase of peritoneal tissue hydration during the first APD cycle in APDDD patients appears to be a consequence of inflow of water into the peritoneal tissue –thereby reducing the inflow into the peritoneal cavity – resulting in a corresponding decrease of net UF.