TY - JOUR
T1 - Gastrointestinal bleeding in patients on long-term dialysis
AU - Trivedi, Hariprasad
AU - Yang, Juliana
AU - Szabo, Aniko
N1 - Publisher Copyright:
© 2014, Italian Society of Nephrology.
PY - 2015/4/1
Y1 - 2015/4/1
N2 - Background: The epidemiology of gastrointestinal bleeding (GIB) in end-stage renal disease (ESRD) has not been adequately characterized. Using United States Renal Data System data we investigated the epidemiology of GIB in hospitalized patients receiving long-term dialysis. Methods: Medicare ESRD patients who began dialysis between 1996 and 2005 were followed from 90 days after starting dialysis to death, transplant, loss of Medicare, or December 31, 2006. GIB events were identified using claims data. Predictors of GIB incidence were analyzed using over-dispersed Poisson regression and Cox regression was used to evaluate the effect on survival. Repeat episodes were modeled using a partially conditional Cox regression model. Results: 406,836 patients were followed for 832,131 person-years, during which 133,967 events were identified. The incidence of GIB was stable through year 2000 but steadily increased thereafter. Chronic gastric ulcer and colonic diverticulosis were the commonest defined causes of upper and lower GIB respectively. Age >49 years, female gender, hypertension as the cause of ESRD, and initiation on hemodialysis was associated with a greater risk of GIB. An episode of GIB conferred a increased hazard of death (hazard ratio 1.9, 95 % CI 1.86–1.93). A previous episode of GIB was associated with greater hazard of another episode (hazard ratio 3.93, 95 % CI 3.82–4.05). Conclusions: In ESRD patients incident to long-term dialysis the incidence of hospital-associated GIB is increasing, is associated with a greater hazard of death, and carries a great hazard of repeat episodes.
AB - Background: The epidemiology of gastrointestinal bleeding (GIB) in end-stage renal disease (ESRD) has not been adequately characterized. Using United States Renal Data System data we investigated the epidemiology of GIB in hospitalized patients receiving long-term dialysis. Methods: Medicare ESRD patients who began dialysis between 1996 and 2005 were followed from 90 days after starting dialysis to death, transplant, loss of Medicare, or December 31, 2006. GIB events were identified using claims data. Predictors of GIB incidence were analyzed using over-dispersed Poisson regression and Cox regression was used to evaluate the effect on survival. Repeat episodes were modeled using a partially conditional Cox regression model. Results: 406,836 patients were followed for 832,131 person-years, during which 133,967 events were identified. The incidence of GIB was stable through year 2000 but steadily increased thereafter. Chronic gastric ulcer and colonic diverticulosis were the commonest defined causes of upper and lower GIB respectively. Age >49 years, female gender, hypertension as the cause of ESRD, and initiation on hemodialysis was associated with a greater risk of GIB. An episode of GIB conferred a increased hazard of death (hazard ratio 1.9, 95 % CI 1.86–1.93). A previous episode of GIB was associated with greater hazard of another episode (hazard ratio 3.93, 95 % CI 3.82–4.05). Conclusions: In ESRD patients incident to long-term dialysis the incidence of hospital-associated GIB is increasing, is associated with a greater hazard of death, and carries a great hazard of repeat episodes.
KW - Dialysis
KW - End-stage renal disease
KW - Gastrointestinal bleeding
KW - Mortality
UR - http://www.scopus.com/inward/record.url?scp=84925935014&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84925935014&partnerID=8YFLogxK
U2 - 10.1007/s40620-014-0132-6
DO - 10.1007/s40620-014-0132-6
M3 - Article
C2 - 25185727
AN - SCOPUS:84925935014
SN - 1121-8428
VL - 28
SP - 235
EP - 243
JO - Journal of Nephrology
JF - Journal of Nephrology
IS - 2
ER -