Gastrointestinal bleeding is associated with higher in-hospital mortality, longer length of stay and higher cost in patients with in-hospital cardiac arrest

Background In-hospital cardiac arrest (IHCA) carries a high mortality and significant morbidity in survivors. Gastrointestinal bleeding (GIB) can complicate cardiac arrests. We aim to study the association between GIB and the in-hospital outcomes of patients with IHCA. Methods and results The National Inpatient Sample 2016–2018 databases were used. IHCA were identified using ICD-10-PCS code for cardiopulmonary resuscitation. Other diagnoses including GIB were identified using ICD-10-CM codes. Multivariate logistic regression was used to study the effect of GIB on in-hospital mortality. Gamma regression with log link was used to determine the effect of GIB on length of stay and cost of admission. In patients with IHCA, GIB as a secondary diagnosis is associated with an increased in hospital mortality (unadjusted 74.2% vs 68.3%, adjusted OR 1.17, 95% confidence interval [CI] 1.09–1.25, p < 0.001), longer length of stay (unadjusted median 16 vs 10 days, IQR 9–27 vs 5–17 days, exponentiated coefficient 1.45, 95% CI 1.36–1.54, p < 0.001 for survivors; unadjusted median 4 vs 3 days, IQR 1–10 vs 1–7 days, exponentiated coefficient 1.27, 95% CI 1.22–1.34, p < 0.001 for patients who died in hospital), and higher cost for hospital stay (unadjusted median $226065 vs $151459, IQR $117551–434003 vs $76197–287846, exponentiated coefficient 1.40, 95% CI 1.32–1.49, p < 0.001 for survivors; unadjusted median $87996 vs $77056, IQR $42566–186677 vs $34066–149009, exponentiated coefficient 1.26, 95% CI 1.20–1.32, p < 0.001 for patients who died in hospital) adjusted for baseline characteristics and other comorbidities. Conclusions In patients with IHCA, GIB as a secondary diagnosis is associated with a higher in-hospital mortality, longer length of stay and higher cost for the admission.


Introduction
Over 290,000 adults in the United States suffer in-hospital cardiac arrest (IHCA) each year which carries a high mortality and can lead to significant morbidity in survivors. 1 Outcomes of in-hospital cardiac arrest depend on many factors including the cause of the cardiac arrest 2 and arrest rhythm. 3 While studies have found a higher Charlson Comorbidity Index (CCI) or the presence of certain comorbidities such as diabetes and renal disease is associated with a higher mortality in out-of-hospital cardiac arrests, 4,5 few studies have looked at the effects of comorbidities on the outcomes of IHCAs. Here we study the association between the presence of a secondary diagnosis gastrointestinal bleeding (GIB) and the outcomes of IHCA using the National Inpatient Sample (NIS).
The NIS is developed by the Agency for Healthcare Research and Quality (AHRQ). It is the largest publicly available all-payer inpatient

Methods
The National Inpatient Sample 2016À2018 databases were used. We identified IHCA using ICD-10-PCS code for cardiopulmonary resuscitation. GIB and severe sepsis were identified using ICD-10-CM codes. A shockable arrest rhythm was identified with the presence of ICD-10-CM codes for ventricular tachycardia or ventricular fibrillation. Other comorbidities were identified, and Charlson Comorbidity Index and Elixhauser Comorbidity Index were calculated using coding algorithms developed by Quan et al. 6 with codes from University of Manitoba. 7 Interventions such as targeted temperature management and mechanical ventilation were identified using ICD-10-PCS codes. The ICD-10-CM and ICD-10-PCS codes used are listed in Supplementary Table 1. Patients with a primary diagnosis of GIB were excluded.
Data processing and statistical analysis were done using SAS 9.4 (SAS Institute Inc., Cary, NC, USA) and RStudio 1.3.1093 (RStudio, PBC, Boston, MA, USA). Multivariate logistic regression was used to study the effect of GI bleeding on in-hospital mortality, controlling for age (in decades), gender, race, primary expected payer, shockable arrest rhythm, targeted temperature management, other acute conditions or comorbidities including severe sepsis and individual groups of CCI except for peptic ulcer disease. Generalized linear regression using a gamma distribution with log-link was used to determine the effect of GIB on length of stay and the cost of the index admission controlling for the above factors stratified by if death occurred during hospital stay. Cost was adjusted for inflation to 2021 US dollars. P value of <0.01 were considered significant. To study the effect of targeted temperature management on GIB in patients with IHCA, logistic regression was used adjusting for age (in decades), gender, race, primary expected payer, shockable arrest rhythm, comorbidities including severe sepsis and individual groups of Charlson Comorbidity Index including peptic ulcer disease.
Results 67351 adult patients with IHCA were identified in the databases. 1059 patients have a primary diagnosis of GIB and was thus excluded. 48 patients with missing mortality data were excluded. 66244 adult patients with IHCA were included in the analysis (Fig. 1). Among them, 5204 (7.86%) patients have a second diagnosis of GIB. The baseline characteristics of the IHCA patients with and without cardiac arrest were listed in Table 1. Unadjusted for other factors, IHCA patients with a secondary diagnosis of GIB more likely to be of non-Caucasian race (42.2% vs 38.3%, p < 0.001) and less likely to have a shockable arrest rhythm (25.3% vs 28.2%, p < 0.001). Similar percentage of patients in both groups underwent targeted temperature management (1.9% vs 1.8%, p = 0.937). Patients with a secondary diagnosis of GIB have a higher Charlson Comorbidity Index (CCI) when peptic ulcer disease (PUD) is included in the CCI (mean AE SD 2.70 AE 1.57 vs 2.53 AE 1.59, p < 0.001). The difference in CCI is smaller when PUD is excluded from the CCI (2.59 AE 1.53 vs 2.52 AE 1.58, p = 0.004).
Among different age groups (age 18-39, age 40-64, and age 65 and above), GIB as a secondary diagnosis is associated with an increased in-hospital mortality in all age groups, with the highest OR in patients between age 18 and 39 (Fig. 3).
In patients with IHCA, a secondary diagnosis of GIB is associated with a longer length of say both in patients who survived the hospital stay (unadjusted median 16 vs 10 days, IQR 9À27 vs 5À17 days, exponentiated coefficient 1.45, 95% CI 1.36À1.54, p < 0.001) and in patient who died in hospital (unadjusted median 4 vs 3 days, IQR 1À10 vs 1À7 days, exponentiated coefficient 1.27, 95% CI 1.22À1.34, p < 0.001), adjusted for age (in decades), sex, race, arrest rhythm, comorbidities including individual groups of CCI excluding PUD, severe sepsis, and if patient underwent targeted temperature management (Fig. 4). A secondary diagnosis of GIB is also associated

Discussion
IHCA carries a high mortality but survival to discharge had improved over the past 2 decades. 8 The effects of comorbidities on the outcomes and healthcare resource utilization of IHCA have not been well studied. Our study showed in patients with IHCA, GIB as a secondary diagnosis is associated with an increased in-hospital mortality, longer length of stay and higher cost of hospital stay.
The study has important limitations. Coding errors or omissions may be present in this administrative database and information such as medication use, alcohol use or previous history of GIB is not available from this data base. Other factors that will affect outcomes include the direct cause of the cardiac arrest and time to initiation of CPR are not available. It is also impossible to ascertain from the database whether GIB occurred before or after IHCA. GIB, especially esophageal variceal bleeding can also be the cause of ICHA and is associated with worse outcomes. 2 The effects of this on our result should be limited because patients with a primary diagnosis of GIB were excluded from our analysis and after this exclusion no patient carries a diagnosis of esophageal varices with bleeding. Most importantly, GIB may also be a maker of patients who were sicker or more complicated in ways that were not captured by comorbidity indices and other factors we included in the analysis. Despite these limitations, the increased mortality and healthcare resource utilization associated with a secondary diagnosis of GIB in this large real-world cohort is notable.
GIB is an important complication of hospitalization 9 especially in critically ill patients. 10 As a complication, it is associated with worse outcomes in may acute illnesses such as sepsis and stroke. 11À13 Here we showed its association with worse outcomes in patients with IHCA. Patient with IHCA may be particularly at high risk for GIB given they are often critically ill and require invasive mechanical ventilation (70.1% in our sample). Our study failed to show any association between targeted temperature management and GIB in IHCA patients. This is consistent with previous studies. 14,15 The association between GIB and increased mortality holds true in all age groups, but the OR was highest in the age 18À40 group. This is possibly related to the different disease spectrum and direct causes of IHCA in this age group.
In summary, GIB is an important comorbidity for patients with IHCA and is associated with increased mortality, longer length of stay and high hospital charge. It is important for clinicians to be vigilant about GIB when taking care of patients of IHCA. Prevention strategies such as acid suppression with a proton pumps inhibitor or histamine-2 receptor antagonist and/or early institution of enteral nutrition should be considered for appropriate patients at high risk for GIB. 10 Funding National Inpatient Sample Databases were purchased using educational stipends from Nuvance Health Danbury Hospital.

Declaration of interests
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.