Abstract: Objective: To explore the predictive value of combining fecal microbiota 16sRNA high-throughput sequencing with serum albumin (Alb) for the fast peritoneal solute transport rate (PSTR) in PD patients.
Methodology: In this center’s retrospective patient cohort (January 2004-January 2024), patients were divided into non-fast PSTR group and fast PSTR group based on the ratio of dialysate creatinine to serum creatinine from the first peritoneal equilibration test (PET) performed 30 days post-catheter insertion (D/Pcr=0.65). The ability of Alb to predict fast PSTR was analyzed using multivariate Logistic regression and receiver operating characteristic (ROC) analysis, and the external validation of Alb’s predictive capability was conducted using the PSTR empirical formula proposed by David, which includes parameters such as gender, race, Alb, and sodium. On the other hand, from May to October 2024, 50 newly admitted PD patients were recruited, and the first PET was performed 30 days post-catheter insertion. Concurrently, fecal samples were collected, frozen at -80℃, and subsequently subjected to high-throughput sequencing of fecal 16S rRNA. Patients were classified into low transport group (D/Pcr ≤0.49), low average transport group (D/Pcr 0.50-0.64), and fast PSTR group (D/Pcr ≥0.65) based on the D/Pcr ratio. The differences in intestinal microbiota levels among the three groups were compared, and significant differential bacterial genera were identified using linear discriminant analysis, analyzing the discriminative value of differential genera combined with Alb for fast PSTR.
Results: A retrospective cohort study included 1008 patients with PD, among which the fast PSTR group accounted for 26.9%. Logistic regression analysis indicated that the variables independently associated with fast PSTR included systolic blood pressure, uric acid, Alb. Compared to other variables, the level of Alb had the largest area under the ROC curve for determining fast PSTR. The predictive ability of Alb was not statistically significantly different from that of the David formula [AUC 95% CI 0.636 (0.593-0.680) vs 0.629 (0.589-0.670), P=0.480]. The fecal microbiota study included 50 newly admitted PD patients, revealing statistically significant differences in Alb levels among the low transport group, low average transport group and fast PSTR group (P=0.012). However, there were no statistically significant differences in the α and β diversity of intestinal microbiota among the three groups. The composition of the microbiota showed no statistically significant differences at the phylum, class, order, and family levels, but significant differences were observed at the genus and species levels. Linear discriminant analysis revealed that the relative abundance of the genus Ruminococcus in the low transport group was significantly higher than that in the fast PSTR group (P=0.018). Correlation analysis indicated a negative correlation between the relative abundance of Ruminococcus and D/Pcr levels (r=-0.351, P=0.013). The DeLong test demonstrated that the relative abundance of Ruminococcus significantly improved the predictive value of Alb for fast PSTR determination [AUC 95% CI 0.871 (0.733-0.979) vs 0.725 (0.554-0.895), P=0.04]; the predictive capability of fecal bacteria combined with Alb was also significantly superior to that of the David formula [AUC 95% CI 0.668 (0.475-0.860), P<0.001].
Conclusion: Serum Alb levels are independent predictors of peritoneal solute transport function in patients with PD. The relative abundance of Ruminococcus species derived from fecal 16S rRNA high-throughput sequencing, when combined with Alb levels, can significantly enhance the predictive value of the latter in determining the fast PSTR.

Key words: peritoneal dialysis, fast peritoneal solute transport rate, fecal microbiota, high-throughput sequencing