Discussion:
The present study is the first and the largest to assess the outcomes of
ESRD with LGIB. We analyzed 2187954 admissions from the NIS database to
examine outcomes of ESRD among the patients with LGIB. To compare
outcomes of ESRD among patients with LGIB versus ESRD patients, we used
a propensity-matched model. This model allowed us to account for
confounding factors and reduce the effect of selection bias. The
in-hospital mortality was higher in ESRD patients with LGIB group. This
group has higher odds of requiring mechanical ventilation and
vasopressor for the shock. Advanced age, acute coronary syndrome,
anemia, acute respiratory failure, mechanical ventilation, and sepsis
were identified as predictors of mortality in this group. Mean LOS and
mean total charges were also higher in this group.
ESRD is a growing public health problem with number tripled between 1990
and 2010. ESRD patients have a higher tendency to develop LGIB. As the
primary etiological factors (diabetes and hypertension) for ESRD are
increasing, there will be an increasing rate of ESRD, which might as
well increase the case of GIB related to ESRD. [6 20] ESRD patients
are likely to bleed more due to a variety of causes, which include
heparin used during dialysis, platelet dysfunction, and medications.
[7 11 12] It is essential to understand that LGIB in ESRD patients
could be due to a variety of causes. [21] The most common cause of
bleeding is angiodysplasia. The other causes are diverticulosis,
hemorrhoids, stercoral ulcer, and ischemic colitis. Due to many
pathophysiology and locations of bleeding, ESRD patients who develop
acute LGIB require a thorough evaluation.
In 2016, there were 35.7 million hospital stays in the United States,
and the cost of these stays totaled over $417 billion, with a mean cost
per stay of $11,700 [22]. Hospitals and health systems across the
nation are under pressure to avoid patient harm and to reduce the length
of stay (LOS) [23]. A longer LOS also increases the likelihood of a
hospital-acquired condition (HAC), which harms patients and contributes
to an even costlier and longer stay [24 25]. The mean LOS is higher
among ESRD patients with LGIB compared to ESRD patients. Our findings
are consistent with previous studies which have shown that the patients
with CKD have an average hospital length of stay (LOS) that is 2 to 3
days longer compared to the general population [26 27]. Increased
LOS results in increased cost of a stay in the hospital. Our results
show that the mean cost of hospitalization in ESRD patients who develop
LGIB is 19000 $ more compared with ESRD patients. Early identification
and intervention would be prudent to reduce the length and cost of the
stay.
Many studies have shown increased mortality for ESRD patients with GIB.
The all-cause mortality of ESRD patients hospitalized for UGIB is
significantly high, with the first-month mortality of 11 to 14%, and
the one-year mortality could go up to 27%. [28 29] There is limited
data available for ESRD patients hospitalized for LGIB. A study by
Trivedi et at; showed an episode of hospital-associated gastrointestinal
bleeding in long-term dialysis patients could increase the hazard of
death by two-fold [5]. Our findings show that ESRD patients who get
hospitalized for LGIB are likely to have more mortality (OR 1.3, 95% CI
1.2-1.3; P=0.00) compared with ESRD patients. ESRD patients have many
comorbid conditions and higher mortality due to cardiovascular disease.
Cardiovascular events may increase after gastrointestinal bleeding
causing hemodynamic insult or decline in hemoglobin [5]. GIB would
cause hypovolemia. Hypovolemia from any cause increases mortality in the
patients hospitalized for GIB. [30] To reduce mortality in ESRD
patients with LGIB, early volume resuscitation, and adequate blood
transfusion might be necessary. Also, in the patient with a history of
coronary artery disease, the optimization of risk factors and medicines
is of very high value to reduce mortality.
ESRD patients who are hospitalized for LGIB have higher odds of
requiring mechanical ventilation (OR 1.4, 95% CI 6.4-16.4; P=0.00) and
vasopressor for shock (OR 1.2, 95% CI 4.9-5.4; P=0.002). These could be
reasoned as these patients have many comorbidities, and they are likely
to get sicker quickly. [30] We hypothesized that patients with ERSD
and LGIB have a longer LOS and likely to develop more complications.
Complications like cardio-respiratory failure and sepsis would require
them to need more ventilator and vasopressor support. Early
identification of shock, volume status, and sepsis would be prudent to
improve outcomes in ESRD patients and LGIB.
In this study, we have identified a few predictors for increased
mortality in ESRD patients who get hospitalized for LGIB. These
predictors are advanced age (OR 1.02 CI 1.02-1.03 P=0.00), anemia (OR
1.04 CI 1.59-1.91 P=0.006), acute coronary syndrome (OR 1.8 CI 1.6-2.1,
P=0.00), acute respiratory failure (OR 1.29 CI 2.0-2.6, P=0.00),
mechanical ventilation (OR 1.9, CI 3.5-4.4, P=0.00), and sepsis (OR 1.5,
CI 4.1-5.08, P=0.00). ESRD patients have a high number of comorbid
conditions. One study for LGIB has shown that in-hospital mortality
increased > 7 % with comorbid score ≥ 2 vs. 1.7% for
those with no comorbidities [30]. We are going to emphasize the
importance of recognizing that ESRD patients who get hospitalized for
LGIB need close attention as they have a higher likelihood of becoming
sicker and dying while hospitalized. These predictors could help in the
future to create a risk score to identify high-risk patients earlier
during hospitalization.
The primary limitation of the study is the cross-sectional design of the
data, which does not allow longitudinal follow-up of patients, limiting
us from incorporating a temporal effect into our analysis. Second, the
NIS contains hospitalization records and not individual patients
implying that adjusting for multiple hospitalizations was impossible.
Despite these limitations, our study diminishes bias similarly to a
randomized trial by including a wide array of covariates in both our
propensity matching algorithm and our final models. We were able to
provide the results of a large cohort composed of patients from numerous
institutions, all geographic regions, economic and demographic groups in
the USA. Therefore, our results could be generalized to the USA
population.
In summary, ESRD patients who get hospitalized for LGIB are sicker than
we believe. They have a higher chance of requiring critical care support
leading to increased length of stay and cost for care. These patients
have a higher tendency to bleed, and they bleed more due to a variety of
insults, and due to a few different causes. Clinicians are compelled to
have a broader knowledge of this clinical entity as the management of
these patients involves a variety of tests and treatments. In the
future, developing a risk score to identify the high-risk patients would
be of help.
Table 1. Demographicscomparing ESRD with LGIB
vs. ESRD only.