Background and Aims
Gastrointestinal (GI) symptoms are present in 20% of patients with SARS-CoV-2 coronavirus infection (COVID-19). We studied the association of GI symptoms (in patients with COVID-19) with adverse outcomes and factors associated with poor outcomes in these patients.
Methods
The study cohort included 100,902 patients from the Cerner Real-World Data COVID-19 Database of hospital encounters and emergency department visits with COVID-19 infection from December 1, 2019, to November 30, 2020. Multivariate analysis was used to study the effect of GI symptoms on adverse outcomes and the factors associated with mortality, acute respiratory distress syndrome (ARDS), sepsis, and ventilator requirement or oxygen dependence in patients with COVID-19 and GI symptoms.
Results
Patients with COVID-19 and GI symptoms were significantly more likely to have ARDS (odds ratio [OR] 1.20, 95% confidence interval [CI] 1.11, 1.29), sepsis (OR 1.19, 95% CI 1.14, 1.24), acute kidney injury (OR 1.30, 95% CI 1.24, 1.36), venous thromboembolism (OR 1.36, 95% CI 1.22, 1.52), or GI bleed (OR 1.62, 95% CI 1.47, 1.79) and less likely to experience cardiomyopathy (OR 0.87, 95% CI 0.77, 0.99) or death (OR 0.71, 95% CI 0.67, 0.75). Among those with GI symptoms, older age, higher Charlson comorbidity index scores, and use of proton pump inhibitors/H2 receptor antagonists were associated with higher mortality, ARDS, sepsis, and ventilator or oxygen requirement.
Conclusion
Patients with COVID-19 who have GI symptoms have overall worse in-hospital complications but less cardiomyopathy and mortality. Older age, higher comorbidity scores, and the use of proton pump inhibitors and H2 receptor antagonists are associated with poor outcomes in these patients.
Keywords
Abbreviations used in this paper:
AKI (Acute kidney injury), ARDS (acute respiratory distress syndrome), CCI (Charlson comorbidity index), CI (confidence interval), COVID-19 (SARS-CoV-2 coronavirus infection), CRWD (Cerner Real-World Data), ER (emergency department), GERD (gastroesophageal reflux disease), GI (gastrointestinal), H2RA (H2 receptor antagonist), PPI (proton pump inhibitors), OR (odds ratio), VTE (venous thromboembolism)Introduction
The SARS-CoV-2 coronavirus (COVID-19) pandemic continues to afflict the global community, with 570,159,393 cases and 6,384,239 deaths reported globally as of July 24, 2022.
1
Initially recognized as a primary respiratory disorder, subsequent expansive research has identified COVID-19 as a multisystemic condition.John Hopkins University [database on the Internet].
https://coronavirus.jhu.edu/map.html
Date accessed: July 25, 2022
2
Although our understanding of the disease has grown exponentially since its identification, the relationships between specific systemic involvement and outcomes remain elusive.Gastrointestinal (GI) symptoms, including diarrhea, nausea, vomiting, anorexia, and abdominal pain, affect approximately 20%–60% of patients with COVID-19.
3
,4
There is also a growing interest in the role of proton pump inhibitors (PPIs) and H2 receptor antagonists (H2RAs) in COVID-19–related outcomes with studies reporting divergent results regarding the impact of PPI/H2RA use on adverse outcomes.5
, 6
, 7
The association of GI symptoms with adverse outcomes, however, has been controversial in previous studies.8
, 9
, 10
Knowledge of specific factors driving outcomes in this subset of patients with COVID-19 would be beneficial in triage, monitoring, risk stratification, and treatment decisions. We investigated the effect of demographics, comorbidities including GI diseases, and specific medication use (PPI/H2RA) on adverse outcomes in patients with COVID-19 with GI symptoms.Methods
Study Design and Population
Data from Cerner Real-World Data (CRWD) COVID-19 database was used for our study. CRWD is extracted from the electronic medical records (EMRs) of hospitals in which Cerner has a data use agreement. Encounters may include pharmacy, clinical and microbiology laboratory, admission, and billing information from affiliated patient care locations. All admissions, medication orders and dispensing, laboratory orders, and specimens are date and time stamped, providing a temporal relationship between treatment patterns and clinical information. Cerner Corporation has established Health Insurance Portability and Accountability Act–compliant operating policies to establish deidentification for CRWD. Access is granted under a data use agreement to organizations and individuals after approval of a submitted research proposal. Patients with at least one encounter (emergency department [ER] or inpatient) with a diagnosis code associated with COVID-19 exposure or infection or an emergency department or inpatient encounter with a positive result for a COVID-19 laboratory test were eligible for inclusion in CRWD COVID-19 database. Only diagnoses and laboratory results from encounters with a service date of December 1, 2019, forward were considered. Longitudinal data were included from all encounters for each patient (labeled by a unique identification number) from January 2015 (to account for variables of interest in medical history) until November 2020.
The data set included demographics, including patient age, race, gender, ethnicity, and hospital characteristics. Mortality was defined using the “deceased” indicator provided by the database. In addition to capturing in-hospital mortality, this indicator allowed administrative updates based on available knowledge of any patient’s death outside the hospital. The International Classification of Diseases, Ninth Revision, Clinical Modification and the International Classification of Diseases, Tenth Revision, Clinical Modification, diagnosis codes recorded during each visit were also separately available. First, laboratory test results for COVID-19 of all patients were examined to filter out those patient records without a confirmed COVID-19 test or only exposure. The included patient data sets were then merged with visits categorized as COVID-19 by Cerner. Cerner identifies each patient with a unique identification number, which allowed us to track multiple visits of a patient. Patients were then categorized as those with and without GI symptoms. GI symptoms were defined as stated in Table A1. Charlson comorbidity index (CCI) was calculated using the Quan et al. coding algorithm.
11
Binary variables were created for the presence or absence of disease and complications in each patient (Table A1). The study was exempt from review under the minimal risk protocol from Providence St Joseph Hospital Institution Review Board, Seattle, WA, USA.Statistical Analysis
Bivariate analysis was used to compare patient demographics, hospital characteristics, comorbidities, concomitant GI conditions, and medication use (Table 1). In each case, proportions in the 2 groups (GI symptoms vs without GI symptoms) were compared using Z-scores. Bivariate and multivariate analyses were used to compute the associations between GI symptoms and complications. These multivariate models were adjusted for race, gender, age (stratified), CCI (stratified), gastroesophageal reflux disease (GERD), other GI conditions (composite of GI malignancy, inflammatory bowel disease, celiac disease), and PPI and H2RA use (Table 2). Statistical significance was defined as a P value of <.05. The outcomes of interest were acute respiratory distress syndrome (ARDS), sepsis, acute kidney injury (AKI), venous thromboembolism (VTE), cardiomyopathy, mortality, and use of invasive mechanical ventilation or oxygen dependence. In addition, the GI-specific complications studied were GI bleed, intestinal ischemia, pancreatitis, acute liver injury, and intestinal pseudoobstruction. Subsequently, multivariate logistic regression models were used to estimate factors associated with ARDS, sepsis, ventilator requirement/oxygen dependence, and mortality in the subset of patients with GI symptoms (Table 3). Statistical analysis was performed using Python and SAS programming languages.
Table 1Baseline Characteristics of Patients With COVID-19 With and Without Gastrointestinal Symptoms Presenting to the Hospital
| Variable | Inpatients/emergency room | GI symptoms | No GI symptoms | P value |
|---|---|---|---|---|
| N = 100,902 | N = 19,915 (19.7%) | N = 80,987 (80.3%) | ||
| n (%) | n (%) | n (%) | ||
| Age (y) a To protect patient identity in smaller populations based on age groups, Cerner Real-World Data codes individuals aged <18 years as age 17 years and more than 90 years of age as age 90. The mean age was thus calculated by removing all patients of age <18 and >89 years, with an age range of 18–89 years. | ||||
| Mean (SD) | 52.39 (19.32) | |||
| Median (IQR) | 52 (67–36) | |||
| ≤17 y | 8238 (8.2%) | 1366 (6.9%) | 6872 (8.5%) | <.0001 |
| 18–34 | 20,822 (20.6%) | 3455 (17.3%) | 17,367 (21.4%) | <.0001 |
| 35–49 | 20,661 (20.5%) | 3946 (19.8%) | 16,715 (20.6%) | .0099 |
| 50–64 | 23,821 (23.6%) | 5010 (25.2%) | 18,811 (23.2%) | <.0001 |
| 65–79 | 18,316 (18.2%) | 4155 (20.9%) | 14,161 (17.5%) | <.0001 |
| ≥80 | 9044 (9%) | 1983 (10%) | 7061 (8.7%) | <.0001 |
| Gender | ||||
| Men | 49,331 (48.9%) | 8978 (45.1%) | 40,353 (49.8%) | <.0001 |
| Women | 51,172 (50.7%) | 10,845 (54.5%) | 40,327 (49.8%) | <.0001 |
| Unknown/other | 399 (0.4%) | 92 (0.5%) | 307 (0.4%) | .095 |
| Race | ||||
| White | 58,655 (58.1%) | 11,999 (60.3%) | 46,656 (57.6%) | <.0001 |
| Black | 18,538 (18.4%) | 3855 (19.4%) | 14,683 (18.1%) | <.0001 |
| Other | 23,709 (23.5%) | 4061 (20.4%) | 19,648 (24.3%) | <.0001 |
| Ethnicity | ||||
| Hispanic | 43,266 (42.9%) | 8321 (41.8%) | 34,945 (43.1%) | .0005 |
| Non-Hispanic | 53,343 (52.9%) | 10,774 (54.1%) | 42,569 (52.6%) | .0001 |
| Other or unknown | 4293 (4.3%) | 820 (4.1%) | 3473 (4.3%) | .28 |
| Hospital region | ||||
| Midwest | 8309 (8.2%) | 1504 (7.6%) | 6805 (8.4%) | .0001 |
| Northeast | 20,827 (20.6%) | 3764 (18.9%) | 17,063 (21.1%) | <.0001 |
| South | 40,448 (40.1%) | 9305 (46.7%) | 31,143 (38.5%) | <.0001 |
| West | 31,318 (31%) | 5342 (26.8%) | 25,976 (32.1%) | <.0001 |
| Hospital beds, no | ||||
| <300 | 4600 (4.6%) | 993 (5%) | 3607 (4.5%) | .0012 |
| 300–499 | 7983 (7.9%) | 1594 (8%) | 6389 (7.9%) | .6 |
| 500–999 | 20,972 (20.8%) | 3622 (18.2%) | 17,350 (21.4%) | <.0001 |
| >1000 | 67,347 (66.7%) | 13,706 (68.8%) | 53,641 (66.2%) | <.0001 |
| Charlson comorbidity index | ||||
| CCI, mean (SD) | 1.217 (2.02) | 1.58 (2.3) | 1.13 (1.94) | <.0001 |
| 0 | 55,848 (55.3%) | 9479 (47.6%) | 46,369 (57.3%) | <.0001 |
| 1–4 | 36,072 (35.7%) | 7929 (39.8%) | 28,143 (34.8%) | <.0001 |
| ≥5 | 8982 (8.9%) | 2507 (12.6%) | 6475 (8%) | <.0001 |
| Charlson comorbidities (components of CCI) | ||||
| Myocardial infarction | 4060 (4%) | 908 (4.6%) | 3152 (3.9%) | |
| Congestive heart failure | 8269 (8.2%) | 1863 (9.4%) | 6406 (7.9%) | |
| Peripheral vascular disease | 2995 (3%) | 796 (4%) | 2199 (2.7%) | |
| Cerebrovascular disease | 3240 (3.2%) | 1028 (5.2%) | 2212 (2.7%) | |
| Dementia | 5597 (5.5%) | 1392 (7%) | 4205 (5.2%) | |
| Chronic pulmonary disease | 13,865 (13.7%) | 2953 (14.8%) | 10,912 (13.5%) | |
| Rheumatological disease | 1458 (1.4%) | 425 (2.1%) | 1033 (1.3%) | |
| Peptic ulcer disease | 475 (0.5%) | 185 (0.9%) | 290 (0.4%) | |
| Mild liver disease | 3517 (3.5%) | 1191 (6%) | 2326 (2.9%) | |
| Diabetes | 25,522 (25.3%) | 5955 (29.9%) | 19,567 (24.2%) | |
| DM with no complications | 23,411 (23.2%) | 5483 (27.5%) | 17,928 (22.1%) | |
| DM with chronic complications | 7842 (7.8%) | 2131 (10.7%) | 5711 (7.1%) | |
| Hemiplegia/paraplegia | 787 (0.8%) | 273 (1.4%) | 514 (0.6%) | |
| Kidney disease | 10,345 (10.3%) | 2645 (13.3%) | 7700 (9.5%) | |
| Any malignancy, including leukemia or lymphoma | 2389 (2.4%) | 694 (3.5%) | 1695 (2.1%) | |
| Moderate/severe liver disease | 711 (0.7%) | 225 (1.1%) | 486 (0.6%) | |
| Metastatic solid tumor | 605 (0.6%) | 205 (1%) | 400 (0.5%) | |
| AIDS/HIV | 428 (0.4%) | 109 (0.5%) | 319 (0.4%) | |
| Hypertension | 31,258 (31%) | 7592 (38.1%) | 23,666 (29.2%) | |
| Hyperlipidemia | 17,549 (17.4%) | 4335 (21.8%) | 13,214 (16.3%) | |
| Concomitant GI conditions | ||||
| Acute liver failure | 444 (0.4%) | 144 (0.7%) | 300 (0.4%) | <.0001 |
| GERD | 7541 (7.5%) | 2296 (11.5%) | 5245 (6.5%) | <.0001 |
| GI malignancy, ALL | 376 (0.4%) | 132 (0.7%) | 244 (0.3%) | <.0001 |
| GI malignancy, esophagus/stomach/colon/liver | 265 (0.3%) | 87 (0.4%) | 178 (0.2%) | <.0001 |
| IBD (UC + Crohn's) | 316 (0.3%) | 137 (0.7%) | 179 (0.2%) | <.0001 |
| Celiac disease | 36 (0.04%) | 11 (0.1%) | 25 (0.03%) | .1 |
| Medications | ||||
| PPI (1 mo history) | 15,741 (15.6%) | 4566 (22.9%) | 11,175 (13.8%) | <.0001 |
| H2RA (1 mo history) | 11,335 (11.2%) | 3281 (16.5%) | 8054 (9.9%) | <.0001 |
CCI, Charlson comorbidity index; DM, diabetes mellitus; GERD, gastroesophageal reflux disease; GI, gastrointestinal; H2RA, H2 receptor antagonists; IBD, inflammatory bowel disease; IQR, interquartile range; PPI, proton pump inhibitors; SD, standard deviation; UC, ulcerative colitis.
a To protect patient identity in smaller populations based on age groups, Cerner Real-World Data codes individuals aged <18 years as age 17 years and more than 90 years of age as age 90. The mean age was thus calculated by removing all patients of age <18 and >89 years, with an age range of 18–89 years.
Table 2Adjusted Associations of GI Symptoms With Outcomes and Complications in Patients With COVID-19 Presenting to the Hospital
| Outcomes and complications | Total | GI symptoms | No GI symptoms | P value | Adjusted OR (95% confidence intervals) | P value for adjusted OR |
|---|---|---|---|---|---|---|
| n (%) | n (%) | n (%) | ||||
| Total | 100,902 | 19,915 | 80,987 | |||
| Outcomes | ||||||
| ARDS | 4135 (4.1%) | 1137 (5.7%) | 2998 (3.7%) | <.0001 | 1.20 (1.11, 1.29) | <.0001 |
| Sepsis | 16,145 (16%) | 4127 (20.7%) | 12,018 (14.8%) | <.0001 | 1.19 (1.14, 1.24) | <.0001 |
| Acute kidney injury | 14,782 (14.6%) | 3999 (20.1%) | 10,783 (13.3%) | <.0001 | 1.30 (1.24, 1.36) | <.0001 |
| Venous thromboembolism | 1559 (1.5%) | 467 (2.3%) | 1092 (1.3%) | <.0001 | 1.36 (1.22, 1.52) | <.0001 |
| Cardiomyopathy | 8475 (8.4%) | 1993 (10%) | 6482 (8%) | <.0001 | 0.87 (0.77, 0.99) | .027 |
| Mortality | 8574 (8.5%) | 1666 (8.4%) | 6908 (8.5%) | .46 | 0.71 (0.67, 0.75) | <.0001 |
| Mechanical ventilation/oxygen dependence | 3177 (3.1%) | 817 (4.1%) | 2360 (2.9%) | <.0001 | 1.05 (0.96, 1.14) | .2729 |
| GI complications | ||||||
| GI bleed | 2000 (2%) | 706 (3.5%) | 1294 (1.6%) | <.0001 | 1.62 (1.47, 1.79) | <.0001 |
| Intestinal ischemia | 37 (0.037%) | 16 (0.1%) | 21 (0%) | .0003 | 1.75 (0.88, 3.50) | .1122 |
| Pancreatitis | 554 (0.55%) | 294 (1.5%) | 260 (0.3%) | <.0001 | 0.84 (0.67, 1.05) | .1266 |
| Acute liver injury | 444 (0.44%) | 144 (0.7%) | 300 (0.4%) | <.0001 | 1.13 (0.90, 1.42) | .2777 |
| Pseudoobstruction | 13 (0.01%) | 7 (0.035%) | 6 (0.007%) | .002 | 0.32 (0.04, 2.47) | .2745 |
ARDS, Acute respiratory distress syndrome; GI, gastrointestinal; OR, odds ratio.
a Adjusted for race, gender, age (stratified), Charlson comorbidity index (stratified), gastroesophageal reflux disease, other GI conditions (composite of GI malignancy, Inflammatory bowel disease, celiac disease), and proton pump inhibitor and H2 receptor antagonist use.
b Given only 13 patients, not adjusted for other GI conditions and gender to get reliable estimates.
Table 3Factors Associated With Adverse Outcomes in Patients With COVID-19 and Gastrointestinal Symptoms Presenting to the Hospital
| Variable | Mortality | ARDS | Sepsis | Ventilator/oxygen dependence | ||||
|---|---|---|---|---|---|---|---|---|
| OR (95% CI) | P value | OR (95% CI) | P value | OR (95% CI) | P value | OR (95% CI) | P value | |
| Black race vs White | 1.02 (0.88, 1.18) | .8032 | 1.10 (0.92, 1.32) | .2838 | 1.03 (0.96, 1.14) | .6063 | 0.87 (0.71, 1.07) | .1913 |
| Other race vs White | 0.98 (0.84, 1.15) | .8283 | 1.90 (1.63, 2.21) | <.0001 | 1.15 (1.04, 1.27) | .0056 | 0.97 (0.80, 1.17) | .7388 |
| Men vs women | 1.47 (1.32, 1.64) | <.0001 | 1.47 (1.30, 1.68) | <.0001 | 1.44 (1.33, 1.55) | <.0001 | 1.14 (0.99, 1.32) | .0727 |
| Unknown gender vs men | 1.27 (0.58, 2.80) | .5465 | 1.02 (0.36, 2.89) | .9674 | 1.43 (0.85, 2.41) | .1749 | 0.63 (0.15, 2.61) | .5202 |
| Bed size 300–499 vs <300 | 1.33 (0.9, 1.96) | .1533 | 0.84 (0.56, 1.26) | .3973 | 1.50 (1.14, 1.96) | .0037 | 1.05 (0.70, 1.55) | .8265 |
| Bed size 500–999 vs <300 | 1.73 (1.22, 2.47) | .0024 | 1.50 (1.06, 2.13) | .0223 | 2.33 (1.82, 2.97) | <.0001 | 0.69 (0.48, 0.998) | .049 |
| Bed size >1000 vs <300 | 1.24 (0.88, 1.74) | .2157 | 1.07 (0.77, 1.47) | .7008 | 2.26 (1.80, 2.84) | <.0001 | 0.59 (0.42, 0.82) | .0015 |
| Region Northeast vs South | 1.25 (1.08, 1.46) | .0036 | 1.05 (0.87, 1.27) | .6171 | 1.46 (1.31, 1.62) | <.0001 | 1.81 (1.47, 2.21) | <.0001 |
| Region Midwest vs South | 0.72 (0.57, 0.90) | .0044 | 1.12 (0.87, 1.45) | .3706 | 0.45 (0.38, 0.55) | <.0001 | 0.92 (0.68, 1.25) | .578 |
| Region West vs South | 1.20 (1.04, 1.38) | .0145 | 2.13 (1.81, 2.50) | <.0001 | 1.78 (1.62, 1.95) | <.0001 | 1.62 (1.34, 1.96) | <.0001 |
| Age group 18–34 vs ≤17 | 2.46 (0.86, 7.02) | .0924 | 5.27 (1.90, 14.64) | .0014 | 0.86 (0.68, 1.09) | .2196 | 0.75 (0.37, 1.52) | .4239 |
| Age group 35–49 vs ≤17 | 3.82 (1.39, 10.49) | .0094 | 9.42 (3.46, 25.65) | <.0001 | 1.14 (0.91, 1.43) | .2677 | 1.74 (0.94, 3.24) | .0779 |
| Age group 50–64 vs ≤17 | 9.53 (3.52, 25.74) | <.0001 | 15.18 (5.61, 41.09) | <.0001 | 1.51 (1.21, 1.88) | .0003 | 2.33 (1.27, 4.25) | .0062 |
| Age group 65–79 vs ≤17 | 19.46 (7.21, 52.52) | <.0001 | 15.39 (5.67, 41.74) | <.0001 | 1.81 (1.45, 2.27) | <.0001 | 3.22 (1.76, 5.89) | .0001 |
| Age group ≥80 vs ≤17 | 37.21 (13.76, 100.63) | <.0001 | 6.32 (2.28, 17.48) | .0004 | 1.67 (1.32, 2.12) | <.0001 | 1.96 (1.05, 3.66) | .0351 |
| CCI score 1–4 vs 0 | 3.88 (3.20, 4.71) | <.0001 | 2.32 (1.95, 2.76) | <.0001 | 2.63 (2.39, 2.89) | <.0001 | 3.84 (3.00, 4.90) | <.0001 |
| CCI score ≥5 vs 0 | 8.23 (6.70, 10.12) | <.0001 | 2.79 (2.27, 3.43) | <.0001 | 5.02 (4.44, 5.67) | <.0001 | 7.07 (5.41, 9.25) | <.0001 |
| GI malignancy, all | 1.45 (0.73, 2.86) | .2906 | 1.65 (0.80, 3.39) | .1717 | 1.02 (0.60, 1.73) | .9495 | 0.76 (0.27, 2.14) | .6055 |
| Acute liver failure | 1.68 (0.83, 3.41) | .1505 | 0.83 (0.32, 2.17) | .7002 | 0.72 (0.40, 1.30) | .2692 | 1.30 (0.51, 3.32) | .5901 |
| IBD (UC and Crohn's) | 1.01 (0.52, 1.93) | .9977 | 1.00 (0.47, 2.15) | .9972 | 1.96 (1.29, 2.97) | .0015 | 1.43 (0.70, 2.94) | .3251 |
| GERD | 0.85 (0.73, 0.98) | .0264 | 0.71 (0.59, 0.85) | .0003 | 0.91 (0.82, 1.02) | .0983 | 1.19 (0.99, 1.43) | .0639 |
| Celiac disease | 7.44 (1.44, 38.46) | .0166 | 1.05 (0.21, 5.32) | .9526 | 2.47 (0.28, 21.55) | .4128 | ||
| PPI | 1.48 (1.32, 1.66) | <.0001 | 2.19 (1.91, 2.50) | <.0001 | 1.88 (1.73, 2.05) | <.0001 | 1.73 (1.48, 2.02) | <.0001 |
| H2RA | 1.78 (1.57, 2.02) | <.0001 | 3.75 (3.29, 4.28) | <.0001 | 2.50 (2.28, 2.73) | <.0001 | 1.97 (1.68, 2.30) | <.0001 |
| Percent concordant | 84.7 | 81 | 77.5 | 79.3 | ||||
P values <.05 are designated in bold.
ARDS, acute respiratory distress syndrome; CCI, Charlson comorbidity index; CI, confidence interval; GERD, gastroesophageal reflux disease; GI, gastrointestinal; H2RA, H2 receptor antagonists; IBD, inflammatory bowel disease; OR, odds ratio; PPI, proton pump inhibitors; UC, ulcerative colitis.
a Celiac disease removed from multivariate analysis for mortality as it was generating unreliable estimates.
Results
Baseline Characteristics
Of 100,902 patients who presented with a confirmed diagnosis of COVID-19 to the hospital or ER, 20% had GI symptoms. The mean age was 52 years (range: 18–89 years). A higher proportion of patients with GI symptoms were aged ≥50 years (vs without GI symptoms; Table 1). Of those who had GI symptoms, 54.5% were women. Overall, 58.1% of the study population identified as White. Patients with GI symptoms were more commonly of the White race or Black race. Among patients with GI symptoms, a higher 46.7% (vs 38.5% without GI symptoms) came from the Southern region. Overall, 66.7% of all patients studied were evaluated at large hospitals with more than 1000 bed size. Patients with GI symptoms were more likely to be evaluated at large hospitals at 68.8% compared with 66.2% without GI symptoms. The mean CCI was higher in patients with GI symptoms, 1.58 compared with 1.13. The CCI score was 1–4 in 39.8% (vs 34.8%) and ≥5 in 12.6% (vs 8%) of patients with GI symptoms (compared with those without GI symptoms). All component comorbidities of the CCI as well as acute liver failure, GERD, GI malignancy, and inflammatory bowel diseases were more common in patients with GI symptoms (Table 1). The use of PPI (22.9% vs 13.8%, P < .0001) and H2RA (16.5% vs 9.9%, P < .0001) was also more common in patients with GI symptoms.
Outcomes
Impact of the Presence of GI Symptoms on Complications of COVID-19
Patients with COVID-19 and GI symptoms were more likely to have ARDS (odds ratio [OR] 1.20, 95% confidence interval [CI] 1.11, 1.29), sepsis (OR 1.19, 95% CI 1.14, 1.24), AKI (OR 1.30, 95% CI 1.24, 1.36), and VTE (OR 1.36, 95% CI 1.22, 1.52) and less likely to have cardiomyopathy (OR 0.87, 95% CI 0.77, 0.99) and mortality (OR 0.71, 95% CI 0.67, 0.75; Table 2). The mean hospital length of stay was 6.8 days for all patients.
Presence of GI Symptoms in COVID-19 and Risk of GI Complications
GI bleed was the most common GI complication (2%) and was more likely in patients with GI symptoms (OR 1.62, 95% CI 1.47, 1.79). Intestinal ischemia, pancreatitis, acute liver injury, and intestinal pseudoobstruction were not associated with GI symptoms (Table 2).
Factors Associated With Adverse Outcomes in Patients With COVID-19 and GI Symptoms: Mortality, ARDS, Sepsis, and Invasive Mechanical Ventilation or Oxygen Dependence
In a multivariate analysis of patients with GI symptoms in COVID-19 (Table 3), men had a higher risk of mortality (OR 1.47, 95% CI 1.32, 1.64), ARDS (OR 1.47, 95% CI 1.30, 1.68), and sepsis (OR 1.44, 95% CI 1.33, 1.55). Black race compared with White was not associated with adverse outcomes. A hospital bed size of 500–999 compared with <300 was associated with higher mortality (OR 1.73, 95% CI 1.22, 2.47), ARDS (OR 1.50, 95% CI 1.06, 2.13), and sepsis (OR 2.33, 95% CI 1.82, 2.97). West region compared with the south was associated with higher mortality (OR 1.20, 95% CI 1.04, 1.38), ARDS (OR 2.13, 95% CI 1.81, 2.50), sepsis (OR 1.78, 95% CI 1.62, 1.95), and ventilator requirement or oxygen dependence (OR 1.62, 95% CI 1.34, 1.96). Northeast region was associated with a higher mortality (OR 1.25, 95% CI 1.08, 1.46), sepsis (OR 1.46, 95% CI 1.31, 1.62), and ventilator or oxygen dependence (OR 1.81, 95% CI 1.47, 2.21), whereas Midwest region was associated with lower mortality (OR 0.72, 95% CI 0.57, 0.90) and sepsis (OR 0.45, 95% CI 0.38, 0.55).
A higher age, particularly age 50 years and above (compared with age 17 years or less) as well as an increasing CCI score (compared with a CCI score of 0) was associated with a higher risk of mortality, ARDS, sepsis, and ventilator or oxygen dependence (Table 3).
Impact of PPI and H2RA in Patients With GI Symptoms in COVID-19
PPI use was associated with a higher risk of mortality (OR 1.48, 95% CI 1.32, 1.66), ARDS (OR 2.19, 95% CI 1.91, 2.50), sepsis (OR 1.88, 95% CI 1.73, 2.05), and ventilator or oxygen dependence (OR 1.73, 95% CI 1.48, 2.02). Similarly, the use of H2RA was associated with a higher risk of mortality (OR 1.78, 95% CI 1.57, 2.02), ARDS (OR 3.75, 95% CI 3.29, 4.28), sepsis (OR 2.50, 95% CI 2.28, 2.73), and ventilator or oxygen dependence (OR 1.97, 95% CI 1.68, 2.30).
Discussion
In this study of 100,902 patients with COVID-19 presenting to the hospital or ER, 1 in 5 patients had GI symptoms. Patients with GI symptoms were older and more likely to be women. Two-thirds of the patients studied were evaluated at large hospitals. This is likely reflective of the higher usage of Cerner EMR in larger hospitals compared with smaller hospitals. Patients with GI symptoms had a higher comorbidity burden. In addition, concomitant GI conditions and medication use of PPI or H2RA were more common in patients with GI symptoms. This demonstrates that patients with GI symptoms in COVID-19 represent an overwhelmingly sicker patient population. In a previous small study of 150 hospitalized patients with COVID-19, the presence of GI symptoms was not associated with a difference in comorbidities.
9
Conversely, a retrospective study of 2804 patients hospitalized in New York with COVID-19 found a greater comorbidity burden and younger age among patients with GI symptoms.12
Association of GI Symptoms With Adverse Outcomes in COVID-19
Our study found that the presence of GI symptoms in patients with COVID-19 was associated with a higher risk of ARDS, sepsis, AKI, and VTE. Conversely, GI symptoms were associated with a lower risk of mortality and cardiomyopathy. GI symptoms were not associated with ventilator or oxygen requirements. Few of these associations have been studied before. A previous study of 2804 patients found GI symptoms to be associated with lower mortality and less need for mechanical ventilation.
12
A meta-analysis of 12,797 patients did not find a difference between the mortality in patients with GI symptoms and overall mortality in COVID-19.3
A separate smaller study of 150 patients did not find an association between GI symptoms and mortality or mechanical ventilation in COVID-19.9
Furthermore, a meta-analysis of 186 articles demonstrated that patients with COVID-19 and GI symptoms had an increased risk of ARDS but not mortality,13
whereas another meta-analysis of 125 articles found that patients with COVID-19 presenting with GI symptoms had a higher risk of mortality, ARDS, AKI, and cardiac injury.14
Our findings emphasize the need to consider GI symptoms in the determination of prognosis in COVID-19.Association of GI Symptoms With GI Complications in COVID-19
Among GI complications, GI bleed was more common in those presenting with GI symptoms, but intestinal ischemia, pancreatitis, acute liver injury, or pseudoobstruction was not. Although the causality of the association between COVID-19 and pancreatitis has been questioned,
15
the association between COVID-19 and liver injury has been previously demonstrated.16
The proposed mechanism of liver injury in COVID-19 is a combination of direct viral invasion as well as inflammatory, hypoxic, and drug-induced effects.17
Impact of Demographic Factors and Comorbidities on Outcomes in Patients With GI Symptoms in COVID-19
Previous large studies have shown older age to be associated with a higher risk of mortality and mechanical ventilation in COVID-19.
18
,19
In our study of patients with COVID-19 and GI symptoms, the odds of mortality were approximately 4-fold, 10-fold, 20-fold, and 37-fold in age groups 35–49, 50–64, 65–79, and ≥80, respectively. Increasing age also had higher odds of ARDS, sepsis, and ventilator requirement or oxygen dependence, dropping respectively to 6.3-, 1.6-, and 1.9-fold in the age group ≥80 years. A previous study of hospitalized patients with risk factors for ARDS postulated that the anatomical, physiological, and immunological differences in those more than 80 years of age conferred protection against ARDS; in addition, more conservative goals of care led to less aggressive testing for ARDS along with less use of mechanical ventilation.20
Male sex has been associated with mechanical ventilation in previous studies on COVID-19, whereas its association with mortality has been inconsistent.
18
,19
In our study of COVID-19 with GI symptoms, men had higher odds of mortality, ARDS, and sepsis but not ventilator requirement or oxygen dependence. The results of previous studies exploring the relationship between COVID-19 mortality and race have been incongruent. Black race or Hispanic ethnicity was not found to be associated with mortality in a previous study of 10,131 veterans with COVID-19.19
A University of Michigan study of 5698 patients with COVID-19 also did not find mortality differences by race.21
A study using the same CRWD COVID-19 database of patients with COVID-19 found increased mortality in non-Hispanic Black or African American patients.22
In our study of patients with COVID-19 with GI symptoms, Black race was not associated with adverse outcomes. This is likely because, in addition to studying a subset of patients with GI symptoms, we also adjusted for different variables, including the use of PPI/H2RA. Overall, being in the West and Northeast regions predicted worse outcomes, and the Midwest region had better outcomes compared with the Southern region in our study. This is consistent with earlier reports of worse mortality outcomes in COVID-19 in the Northeast region 18
and the relationship of mortality with COVID-19 disease burden.19
We found that an increase in CCI score was associated with increasing mortality, ARDS, sepsis, as well as ventilator requirement/oxygen dependence in patients with GI symptoms and COVID-19. Previous studies have found CCI score to be associated with mortality and poor outcomes in COVID-19.
23
,24
A previous meta-analysis found that patients with COVID-19 and severe disease had a higher prevalence of abnormal liver function.25
We did not find an independent association between acute liver injury and severe outcomes, including respiratory failure and death in patients with COVID-19 who had GI symptoms.Impact of GERD, PPI, and H2RA in Patients With GI Symptoms in COVID-19
In patients with GI symptoms in COVID-19, GERD appeared to be protective against ARDS and mortality but had no association with ventilator requirement or oxygen dependence or sepsis. Hypochlorhydria has been hypothesized to be conducive to COVID-19 GI infection.
26
Perhaps, a conversely increased acidic environment associated with GERD suppresses COVID-19 viral load at the gastrointestinal point of entry, favoring a milder disease course.In a nationwide Korean study of 132,316 patients, the use of PPI was associated with an increased risk of a composite outcome of critical care requirements, mechanical ventilation, and mortality.
5
The effect of H2RA on COVID-19 outcomes has been controversial in previous studies.27
,28
A recent study of 295 patients with COVID-19 elucidated that PPI use before hospitalization was associated with mortality and ARDS.29
The presence of GI symptoms in our study cohort likely marks the presence of SARS-CoV-2 in the GI tract in these patients. A possible explanation is that the suppression of gastric acid favors increased viral load and enterocyte invasion by SARS-CoV-2 and a possible greater cytokine storm as previously hypothesized.5
In addition, a recent study found PPI to be associated with GI symptoms and discussed a possible role of PPI-related dysbiosis in the clinical course of COVID-19.4
These hypotheses in part could explain our study findings that PPI and H2RA were associated with a higher risk of ARDS, mortality, sepsis, as well as ventilator requirement or oxygen dependence in patients with COVID-19 and GI symptoms. We were unable to confirm the role of PPI and H2RA in the causation of these adverse outcomes, however, and future studies are needed that account for the indications for the use of PPI and H2RA, whether such use is a modifiable factor or unavoidable use due to medical necessity, and for the cause of death.Strengths and Limitations
To our knowledge, this is the largest study to date evaluating the characteristics and outcomes in patients with COVID-19 and GI symptoms and the first study to explore the detailed predictors of specific adverse outcomes, including ARDS, sepsis, mortality, and ventilator requirement or oxygen dependence in this subset of patients. Our study is not without limitations, however. The CRWD COVID-19 database includes COVID-19 patients with an ER or inpatient encounter and omits those treated at home. This confers a bias toward patients with a more aggressive disease course. Moreover, a comparison group without COVID-19 was not available using the CRWD COVID-19 database.
Another limitation is the potential for coding errors and variations in coding practices among the hospitals captured by the database, as well as limitations in the reconciliation of such variations using coding-based queries retrospectively. There is also the possibility of unmeasured individual patient-level confounders not captured by International Classification of Diseases codes or the CRWD COVID-19 database. We were unable to ascertain if the differences in an individual patient and/or family values, preferences, goals of care, and decisions such as palliative care or hospice affected the ventilator requirements and/or mortality. Finally, we could not account for the differences in treatment protocols by individual hospitals and clinicians, as well as the temporal changes in treatment protocols.
Conclusion
Patients with COVID-19 and GI symptoms have a higher comorbidity burden and overall worse in-hospital complications but lower mortality and cardiomyopathy than patients without GI symptoms. A higher age and comorbidity burden, as well as the use of PPI and H2RA, are associated with poor outcomes in patients with COVID-19 that have GI symptoms. Factors associated with adverse outcomes specific to this subset of patients could guide further disease management, prognostication, and resource utilization decisions.
Authors' Contributions
Nikita Patil contributed to conceptualization, methodology, and writing, reviewing, and editing the article. Pankush Kalgotra contributed to data curation, formal analysis, software, and reviewing and editing the article. Suneha Sundaram contributed to conceptualization, methodology, and reviewing and editing the article. Stephanie Melquist contributed to methodology and writing, reviewing, and editing the article. Sravanthi Parasa contributed to conceptualization, methodology, supervision, and reviewing and editing the article. Madhav Desai contributed to conceptualization, methodology, and reviewing and editing the article. Prateek Sharma contributed to conceptualization, methodology, supervision, and reviewing and editing the article.
Supplementary Materials
References
- John Hopkins University [database on the Internet].https://coronavirus.jhu.edu/map.htmlDate accessed: July 25, 2022
- Covid19, beyond just the lungs: a review of multisystemic involvement by Covid19.Pathol Res Pract. 2021; 224: 153384
- Prevalence and mortality of COVID-19 patients with gastrointestinal symptoms: a systematic review and meta-analysis.Mayo Clin Proc. 2020; 95: 1632-1648
- Prevalence of gastrointestinal symptoms in severe acute respiratory syndrome coronavirus 2 infection: results of the prospective controlled multinational GI-COVID-19 study.Am J Gastroenterol. 2022; 117: 147-157
- Severe clinical outcomes of COVID-19 associated with proton pump inhibitors: a nationwide cohort study with propensity score matching.Gut. 2021; 70: 76-84
- Increased risk of COVID-19 among users of proton pump inhibitors.Am J Gastroenterol. 2020; 115: 1707-1715
- Proton pump inhibitor use is not strongly associated with SARS-CoV-2 related outcomes: a nationwide study and meta-analysis.Clin Gastroenterol Hepatol. 2021; 19: 1845-1854.e6
- Gastrointestinal symptoms predict the outcomes from COVID-19 infection.J Clin Gastroenterol. 2021; 56: e145-e148
- Gastrointestinal symptoms and outcomes in hospitalized coronavirus disease 2019 patients.Dig Dis. 2020; 38: 373-379
- Diarrhea is associated with increased severity of disease in COVID-19: systemic review and metaanalysis.SN Compr Clin Med. 2021; 3: 28-35
- Coding algorithms for defining comorbidities in ICD-9-CM and ICD-10 administrative data.Med Care. 2005; 43: 1130-1139
- Disease course and outcomes of COVID-19 among hospitalized patients with gastrointestinal manifestations.Clin Gastroenterol Hepatol. 2021; 19: 1402-1409.e1
- Meta-analysis of outcomes of patients with COVID-19 infection with versus without gastrointestinal symptoms.Baylor Univ Med Cent Proc. 2020; 33: 366-369
- Gastroenterology manifestations and COVID-19 outcomes: a meta-analysis of 25,252 cohorts among the first and second waves.J Med Virol. 2021; 93: 2740-2768
- COVID-19 and acute pancreatitis: examining the causality.Nat Rev Gastroenterol Hepatol. 2021; 18: 3-4
- Hepatic complications of COVID-19 and its treatment.J Med Virol. 2020; 92: 1818-1824
- Pathophysiological mechanisms of liver injury in COVID-19.Liver Int. 2021; 41: 20-32
- Patient characteristics and outcomes of 11,721 patients with COVID19 hospitalized across the United States.Clin Infect Dis. 2020; 72: e558-e565
- Risk factors for hospitalization, mechanical ventilation, or death among 10 131 US veterans with SARS-CoV-2 infection.JAMA Netw Open. 2020; 3: e2022310
- Older adult patients are at lower risk of ARDS compared to younger patients at risk: secondary analysis of a multicenter cohort study.J Intensive Care Med. 2020; 35: 42-47
- Characteristics associated with racial/ethnic disparities in COVID-19 outcomes in an academic health care system.JAMA Netw Open. 2020; 3: e2025197
- Racial disparities in COVID-19 outcomes exist despite comparable Elixhauser comorbidity indices between Blacks, Hispanics, Native Americans, and Whites.Sci Rep. 2021; 11: 8738
- Charlson comorbidity index and a composite of poor outcomes in COVID-19 patients: a systematic review and meta-analysis.Diabetes Metab Syndr. 2020; 14: 2103-2109
- Older age and comorbidity are independent mortality predictors in a large cohort of 1305 COVID-19 patients in Michigan, United States.J Intern Med. 2020; 288: 469-476
- The prevalence of gastrointestinal symptoms, abnormal liver function, digestive system disease and liver disease in COVID-19 infection: a systematic review and meta-analysis.J Clin Gastroenterol. 2021; 55: 67-76
- Direct COVID-19 infection of enterocytes: the role of hypochlorhydria.Am J Infect Control. 2021; 49: 385-386
- Comparative effectiveness of famotidine in hospitalized COVID-19 patients.Am J Gastroenterol. 2021; 116: 692-699
- Famotidine use is associated with improved clinical outcomes in hospitalized COVID-19 patients: a propensity score matched retrospective cohort study.Gastroenterology. 2020; 159: 1129-1131.e3
- Pre-hospitalization proton pump inhibitor use and clinical outcomes in COVID-19.Eur J Gastroenterol Hepatol. 2022; 34: 137-141
Article info
Publication history
Published online: August 26, 2022
Accepted:
August 12,
2022
Received:
May 12,
2022
Footnotes
∗Author names in bold designate shared co-first authorship.
Conflicts of Interest: These authors disclose the following: P.S. is a consultant for Medtronic, Olympus, Boston Scientific, Fujifilm, and Lumendi; and received grant support from Ironwood, Erbe, Docbot, Cosmo pharmaceuticals, and CDx labs. S.P. is a member of Medical Advisory Board, Fujifilm. M.D. contributed to grant support from Intercept Pharma. The remaining authors disclose no conflicts.
Funding: The authors report no funding.
Ethical Statement: The corresponding author, on behalf of all authors, jointly and severally, certifies that their institution has approved the protocol for any investigation involving humans or animals and that all experimentation was conducted in conformity with ethical and humane principles of research. The study was exempt from review under the minimal risk protocol from Providence St Joseph Hospital Institution Review Board, Seattle, WA, USA.
Data Transparency Statement: The authors are unable to make the data available to other researchers, but data are available at Cerner.
Identification
Copyright
© 2022 The Authors. Published by Elsevier Inc. on behalf of the AGA Institute.
User license
Creative Commons Attribution – NonCommercial – NoDerivs (CC BY-NC-ND 4.0) | How you can reuse 
Elsevier's open access license policy
Creative Commons Attribution – NonCommercial – NoDerivs (CC BY-NC-ND 4.0)
Permitted
For non-commercial purposes:
- Read, print & download
- Redistribute or republish the final article
- Text & data mine
- Translate the article (private use only, not for distribution)
- Reuse portions or extracts from the article in other works
Not Permitted
- Sell or re-use for commercial purposes
- Distribute translations or adaptations of the article
Elsevier's open access license policy
