Objective To examine the association between residual renal function at initiation of dialysis and prognosis in maintenance dialysis patients.Methods Incident patients with end-stage renal diseases initiating dialysis between 1 January 2005 and 30 September 2009,followed up to 31 March 2010 were enrolled in this study.Residual renal function was evaluated using eGFR estimated by the abbreviated MDRD equation.Patients were classified into four groups according to eGFR of ≥10.5,8 to ＜10.5,6 to ＜8,＜6 ml·min-1·(1.73 m2)-1.The outcome was all-cause and cardiocerebral vascular mortality.Results (1) A total of 562 patients were included.The median eGFR at initiation of dialysis was 5.60 (2.26-12.62) ml·min-1·(1.73 m2)-1.The median follow-up time was 17 (0-58) months from initiation of dialysis and 141 patients died within this period.The median survival time was 45.48 (43.05-47.90) months.With eGFR declined,Scr,BUN,serum uric acid,serum prealbumin,phosphorus,calcium and phosphate product,iPTH,mean arterial pressure (MAP) at initiation of dialysis increased (P＜0.05),and hemoglobin,proportion of male,proportion of diabetes comorbidity,proportion of the Charlson comorbidity index ≥5 decreased (P＜0.05).Though there was no significant difference among the four groups,the proportion of left ventricular hypertrophy comorbidity increased when eGFR declined.(2) There was no significant difference of all-cause mortality among four groups using Kaplan-Meire survival curve.Cox regression model indicated no significant difference of all-cause mortality in levels of eGFR (HR=1.012,95％CI 0.961-1.065,P=0.654).Without patients died in the first 3 months,the multivariate Cox regression model indicated eGFR at initiation of dialysis was the protective factor to 1 year survival (HR=0.791,95％CI 0.669-0.935,P＜0.01).(3) The multivariate Cox regression model indicated the risk of overall and 1 year cardiocerebral vascular death decreased with eGFR at initiation of dialysis increased (HR=0.868,95％CI 0.777-0.971,P＜0.05; HR=0.937,95％CI 0.851-0.992,P＜0.05,respectively).(4) The multivariate Cox regression model indicated eGFR at initiation of dialysis was benefit to survival of patients treated by peritoneal dialysis,with all-cause death risk decreased by 10％ when eGFR increased by 1 ml·min-1·(1.73 m2)-1 (HR=0.90,95％CI 0.81-0.99,P＜0.05).In hemodialysis patients,Kaplan-Meire survival curve was significantly different among the four groups (Log-rank test,P=0.047); the survival of the group of 8 to ＜10.5 ml·min-1·(1.73 m2)-1 was lower as compared to the groups of 6 to ＜8 (Log-rank test,P=0.033) and ＜6 ml·min-1(1.73 m2)-1 (Log-rank test,P=0.005); but the multivariate Cox regression model indicated no relationship between survival and eGFR.In the subgroup of chronic glomerulonephritis as primary renal disease,the eGFR at initiation of dialysis was the benefit factor,with all-cause death risk decreased by 16.6％ (HR=0.834,95％CI 0.736-0.946,P＜0.01) and cardiocerebral vascular death risk decreased by 18.2％ (HR=0.818,95％CI 0.669-0.999,P＜0.05) when eGFR increased by 1 ml ·min-1 ·(1.73 m2)-1.In the subgroup of chronic glomerulonephritis treated by peritoneal dialysis,the all-cause death risk decreased by 32.1％ with eGFR increased by 1 ml·min 1·(1.73 m2)-1 (HR=0.679,95％CI 0.535-0.862,P＜0.01).Conclusions Early initiation of dialysis may not be associated with improved overall survival,but may reduce cardiocerebral vascular and 1 year all-cause mortality,improve the survival of chronic glomerulonephritis patients and peritoneal dialysis patients.

Objective To assess the risk factors of intradialytic-hypotension (IDH) among maintaining hemodialysis (MHD) patients and to explore the relation between NT-proBNP and IDH,thus to provide clinical evidence for the prevention and treatment of IDH.Methods A total of 202 MHD patients during March 2009 to May 2009 in our dialysis center were enrolled in the study.Intradialytic blood pressure (BP) was measured during a 3-month period.IDH was defined as an event characterized by a sudden drop in systolic BP more than 20 mm Hg or in mean artery pressure (MAP) more than l0 mm Hg.Logistic regression analysis was used to assess the risk factors of IDH.ROC curve was used to evaluate the diagnostic efficacy of serum NT-proBNP.Results The incidence of IDH was 42.1％.One hundred and seventeen patients with no-IDH (＜1/10 hypotensive events per 3 months) were served as controls.Fifty-five patients with o-IDH (≥ 1/ 10 but ≤1/3 hypotensive events per 3 months) and 30 patients with f-IDH (＞1/3 hypotensive events per 3 months) were identified among 202 patients.Multivariate regression analysis showed that age,gender,ultrafiltration rate,serum NT-proBNP,serum albumin,aortic root dimension (AoRD) were associated with IDH among MHD patients.Serum NT-proBNP was positively correlated with IDH.The area under the ROC curve (AUC) of NT-proBNP was 0.76 (95％ CI 0.69 to 0.83,P＜0.01).The cut-off value of serum NT-proBNP for IDH was 1746.5 ng/L,with a sensitivity of 88.61％ and a specificity of 51.10％.Furthermore,the AUC of NT-proBNP for f-IDH was 0.65 (95％ CI 0.53 to 0.763,P＜0.01).The cut-off value of serum NT-proBNP for f-IDH was 8208.0 ng/L,with a sensitivity of 33.33％ and a specificity of 91.30％.Conclusions Elderly,female,high ultrafiltration rate,high level of serum NT-proBNP,hypoalbuminemia,shorter AoRD are independent risk factors of IDH among MHD patients.Serum NT-proBNP can be used as a predictor of IDH.

Objective To explore the ways to decrease the postoperative complications of pancreaticoduodenectomy. Methods Thirty-four patients who underwent pancreaticoduodenectomy between January 1998 and December 2004 were reviewed retrospectively. A duct-to-mucosa pancreaticojejunostomy was performed mostly for patients, and an end-to-end pancreaticojejunal invagination for 5-patients with a soft pancreas and a small pancreatic duct. The end-to-side hepaticojejunostomy and the Roux-en-Y reconstruction of gastrointestinal continuity were performed for all patients. Results The hospital mortality was zero. The postoperative complications occurred in the form of wound infection was 4(12%), delayed gastric emptying was 1(3%), pneumonia was 1(3%), intra-abdominal collections was 1(3%) and pancreaticojejunostomy leak was 1(3%). In 1 patient with pancreaticojejunostomy leak, the closure was achieved with the conservative treatment. Intra-abdominal bleeding, intra-abdominal abscess and other anastomotic leakage were not seen in any of patients. The median follow-up was 21 months (ranging from 6 months to 5 years), none of patients had clinical evidence of steatorrhea, bile reflux gastric disease, anastomotic ulcer, retrograde cholangitis and dumping syndrome, there was no new case of diabetes. Conclusions Proper method of reconstruction produces encouraging results in decreasing the complications after pancreaticoduodenectomy.