Patients and extracorporeal circuits
A total of 395 episodes (Fig. 1) in 131 patients, accounting for 16,244.1 h of effective treatment time, were included in the study. Over the course of our study, 96 cases (24.3%) were electively ended (i.e., the circuit had been used for 72 h). Clotting of the filter or air-trap chamber occurred in 299 cases (75.7%). The average lifespan of the extracorporeal circuit was 41.1 ± 24.8 h. For anticoagulation, RAC was the primary choice (48.6%), followed by no anticoagulation (31.1%), and LMWH (20.3%). In the cluster of modality, the proportion of CVVHDF was 81.3%, and CVVH was 18.7%. The average prescribed dose of CRRT was 31.3 ± 3.2 ml/kg/h. The femoral vein accounted for 93.2% of vascular access. Right side femoral vein access was used in 61.6% of the cases, and the left side was used in 38.4% of the cases. Alternative vascular accesses included the jugular vein (6.8%), of which 88.2% of the cases were accessed on the right side. The details are reported in Table 1.
Dynamic pressure changes during CRRT with different extracorporeal circuit failures (ECF)
For further analysis, we defined three types of extracorporeal circuit failures  according to circuit lifespan, including early (≤ 12 h), intermediate (> 12 h, ≤ 24 h) and late (> 24 h). A total of 134 circuits (33.9%) experience early-intermediate failure, and 261 circuits (66.1%) experienced late failure. The mean changes in the AOP, PFP, EP, RIP and TMP data were completely distinct in the different groups. The Dynamic mean pressure curve graphs are shown in Fig. 2.
The negative value of AOP was smallest in the early group (-62.87 ± 2.31 mmHg), which was 23.5 and 4.87 mmHg lower than that in the late and intermediate groups, respectively. The overall changes in the PFP were also varied among the different types of ECF: the mean value in the early, intermediate and late groups were 133.43 ± 21.95 mmHg, 150.47 ± 28.09 mmHg and 104.92 ± 3.89 mmHg, respectively. About EPs, intermediate group had the smallest value of mean extracorporeal circuit data, followed by the late and early groups. In data of RIPs, the lowest and highest mean values were 46.38 ± 1.11 mmHg and 61.22 ± 7.74 mmHg in the late group and intermediate group, respectively. In cure graph of TMP, the line in the early and intermediate groups increased rapidly, with mean data of 98.12 ± 34.48 mmHg and 120.15 ± 38.891 mmHg, respectively. Moreover, the variability of late groups was statistically smaller than that compared to the other groups (P༜0.05) in all totally different extracorporeal circuit pressure cluster (AOP, PFP, EP, RIP, TMP). The detail variability data are shown in Table 2.
Access outflow dysfunction (AOD) events under different anticoagulants
A total of 143 circuits experienced at least one AOD episode, and no significant difference was found (41.0 ± 25.7 vs. 41.3 ± 23.6 h, P = 0.91) in the lifespan of the circuits in which no AOD event occurred. However, the circuits without moderate-severe AOD events were significantly prolonged compared to those with moderate-severe AOD events during CRRT (43.0 ± 24.4 vs. 28.6 ± 24.2 h, P = 0.003) (Fig. 3).
In our study, RCA was associated with longer circuits survival (31.3 ± 20.0 h vs. 23.9 ± 19.1 h vs. 54.6 ± 22.2 h, P < .0.05). Moreover, different anticoagulation strategies had distinct effects on moderate-severe AOD events in the circuit lifespan. When no anticoagulation was used, the lifespan of circuits without moderate-severe AOD events was significantly prolonged (17.6 ± 11.2 h vs. 35.1 ± 17.1 h, P = 0.001). The same effect existed when RCA was used (40.3 ± 22.2 h vs. 55.9 ± 21.7 h, P = 0.016). However, the effect of moderate-severe AOD events on circuit survival disappeared with the use of LMWH (24.4 ± 15.5 h vs. 24.9 ± 16.3 h, P = 0.96; Fig. 4).
Analysis of risk factors of extracorporeal circuit survival
Comparison between the early-intermediate and late groups revealed that circuits in the chronic group had a lower occurrence of moderate-severe AOD episodes (22.4% vs. 8.0%, P < 0.001), lower platelet count (102.67 ± 90.11 vs. 133.46 ± 84.86 *109/l, P = 0.011) and higher use of the CVVHDF modality (90.4% vs. 63.4%, P < 0.001). However, mild AOD events, hemoglobin, PT, INR, APTT and vascular access (P > 0.05) were not significantly different between these two groups (Table 3). Variables associated with a shorter lifespan of the extracorporeal circuit are shown in Table 4. According to the Cox regression model, moderate-severe AOD events (HR 1.893, 95CI% 1.300 to 2.756, P = 0.001) were risk factors for circuit survival during CRRT. RCA (HR 0.391, 95CI% 0.293 to 0.521, P < 0.001) and CVVHDF (HR 0.546, 95CI% 0.376 to 0.793, P = 0.001) were independently associated with a longer lifespan of the extracorporeal circuit.
Solute removal efficiency and dynamic pressure changes
The removal efficiency of medium-macro molecular solutes (β2-microglobulin) was significantly lower than that of BUN and creatinine at different time during CRRT. All efficiencies of tested solutes removal (BUN, creatinine and β2-microglobulin) dropped gradually with operation time prolonged (Fig. 5). The details of solute removal efficiency in different anticoagulation modalities were presented in Supplementary Appendix File. According to the precise TMP data which was matched with sample collection time,two groups were formed: TMP < 150 mmHg and TMP ≥ 150 mmHg. The solute removal efficiency in the lower TMP group showed greater clearance ability than that in the higher TMP group. Moreover, this phenomenon significantly occurred between the TMP < 150 mmHg and TMP ≥ 150 mmHg group for BUN (0.92 ± 0.10 vs. 0.83 ± 0.16, P = 0.001), creatinine (0.77 ± 0.20 vs. 0.63 ± 0.23, P = 0.007), and β2-microglobulin (0.46 ± 0.11vs. 0.29 ± 0.08, P < 0.001) at 24 h (Fig. 6).