Abstract
Background
Patients with acute hepatitis A virus (HAV) infection are at risk of developing acute kidney injury (AKI) which may result in increased healthcare resource utilization and worse clinical outcomes. We investigated the impact of AKI on healthcare utilization and clinical outcomes in patients hospitalized with acute HAV infection utilizing a large database.
Methods
We queried the National Inpatient Sample (NIS) 2007–2014 to identify acute HAV infection-related hospitalizations with and without AKI. Primary outcomes were prevalence of AKI and its predictors with secondary outcomes included the mean length of stay (LOS), hospitalization cost and mortality in both groups.
Results
Out of 68 364 acute HAV infection-related hospitalizations, 47 620 met our study criteria and 7458 (15.7%) had concurrent AKI. HAV patients with AKI were older (62.5 vs. 53.7 years; P value <0.001). A higher mean LOS (10.03 vs. 5.6 days; P value <0.001) and mean total hospitalization cost ($27 171.35 vs. $12 790.26; P value <0.001) were observed in HAV patients with the AKI group. A total of 1032 patients (13.8%) in the AKI group died during the same hospitalization as compared to 681 patients (1.5%) in the non-AKI group, P value <0.001. AKI in HAV was also found to be an independent predictor of mortality [adjusted odds ratio (aOR), 3.28; 95% confidence interval, 2.23–4.84; P value <0.001) after adjusting for the confounding factors.
Conclusion
We found that 15.67% of patients hospitalized with acute HAV had AKI which contributed to increased healthcare utilization and higher mortality which is preventable.
Keywords: Charlson Comorbidity Index, Healthcare Cost and Utilization Project, hepatitis A, HCUP, hospitalization cost, national inpatient sample, predicters of acute kidney injury
Introduction
Acute hepatitis A virus (HAV) infection is a communicable yet preventable infectious disease that is more prevalent in under-developed countries. According to the WHO, almost 1.4 million cases are reported each year worldwide [1]. In the USA, there have been recent outbreaks of acute HAV infections, and more than 34 347 cases with 21 104 (61%) hospitalizations and 333 deaths have been reported since 2016 [2]. In another study, the incidence of HAV infections has been reported with an increase of 294% between 2016 and 2018 [3]. Most of the time, it is a self-limiting disease, and clinical presentations vary from mild to severe disease with progression to the fulminant hepatic failure in less than 1% of cases [4]. Acute kidney injury (AKI) is one of the most common extrahepatic manifestations of acute HAV infection resulting from oral intake intolerance resulting into prerenal azotemia or due to the development of immune complex-mediated glomerulonephritis. Although AKI is more prone to develop in fulminant HAV cases, few small studies have reported an increased incidence of AKI in nonfulminant cases with an overall incidence ranging between 5 and 7% [5]. Existing studies have shown that patients with HAV infection who develop AKI are at an increased risk of higher morbidity and mortality with major economic implications. However, the majority of these studies are limited by small sample size and lack of generalizability; furthermore, the prevalence of AKI with its impact on in-hospital mortality, length of stay (LOS) and costs in patients hospitalized with HAV infection at a large scale is not well described. We aimed to estimate the prevalence of AKI, its effect on healthcare utilization and clinical outcomes in patients hospitalized with acute HAV infection through this study utilizing the largest US national inpatient database.
Methods
Data source
We queried the National Inpatient Sample (NIS) database from 2007 to 2014 to identify HAV-related hospitalizations, with and without concomitant AKI. The description of NIS is described elsewhere [6]. Briefly, it is a large publicly available all-payer nationwide database, an initiative of the Healthcare Cost and Utilization Project (HCUP), that contains information on more than 7 million (unweighted) and 35 million (weighted) hospitalizations each year. This database provides national estimates of in-hospital mortality and resource utilization associated with various conditions and provides data for the policymakers and healthcare organizations to improve the quality and outcomes related to the in-hospital care. The structure of the NIS database follows the complex survey design methodology, and each year data contained a 20-percent stratified sample from all the acute care hospitals from the states participating in the HCUP. In 2012, the sampling method was changed to a 20-percent stratified sample from all the discharges from US community hospitals, excluding rehabilitation and long-term acute care hospitals. The stratification is based on the US census region, division, hospital control, location, teaching status and hospital bed size based on the American Hospital Association criteria. The unit of observation in the database is a single hospitalization identified through a principal discharge diagnosis with up to 29 secondary discharge diagnoses. These discharge diagnoses are based on the International Classification of Diseases, Ninth Revision or Tenth Revision, Clinical Modification Codes (ICD-9-CM) system depending upon the year.
Cohort selection and study variables
Each hospitalization was considered to be HAV infection-related if it included the ICD-9 codes 070.0 (hepatitis A with hepatic coma) and 070.1 (viral hepatitis A without mention of hepatic coma) as a primary or secondary discharge diagnosis code. Similarly, AKI was defined by the ICD codes 584.9, 584.6, 584.7, 584.5 and 584.8. We divided the patients into two groups, that is, (1) HAV infection with AKI and (2) HAV infection without AKI.
We excluded patients with age less than 18 years, elective admissions and cases with preexisting end-stage renal disease to correctly identify the need for hemodialysis in patients with HAV who develop or present with AKI. Each observation in the NIS contains patient, treatment and hospital-level information, including age, gender, race, median income by zip code, number and type of procedures, LOS, hospitalization charges, hospital region, hospital teaching status, hospital size based on the number of beds available and its geographical location. We utilized the Charlson Comorbidity Index (CCI) to assess the burden of comorbidities of patients in both groups. The CCI score predicts 1-year survival in patients based on the presence or absence of multiple comorbidities in studies utilizing the administrative databases [7].
Primary and secondary outcomes
The primary outcomes of our study were the prevalence and predictors of AKI in patients hospitalized with acute HAV infection. Secondary outcomes included estimation of healthcare utilization including LOS, inflation-adjusted total hospitalization costs, requirement of mechanical ventilation, need for the hemodialysis and all-cause in-hospital mortality.
Statistical analysis
We used STATA, version 16.0 (StataCorp, College Station, Texas, USA), to conduct all statistical analyses. Descriptive statistics for patients and hospital characteristics were analyzed using Pearson’s chi-square test for categorical and Student’s t-test for the continuous variables. Categorical variables are reported as proportions or percentages, whereas continuous variables are described as means with standard deviations (SD). We treated the missing data with case-wise deletion, and only less than 5% of the data for LOS, hospitalization cost and mortality were missing. We chose an a-priori alpha value of less than 0.05 as a level of significance. We used bivariate linear and logistic regression analysis followed by multivariate regression models to adjust for the confounding factors to assess the effect of AKI on secondary outcomes. We selected the variables to construct the multivariate models, which showed a significance level of 0.2 in the bivariate analysis. The total hospitalization cost variable is not available in the NIS as a default; hence we used the HCUP cost-to-charge ratio Supplementary files, Supplemental digital content 1, http://links.lww.com/EJGH/A706 to calculate the hospital cost [8]. We adjusted the hospitalization cost to the year 2018 for inflation utilizing the Medical Expenditure Panel Survey factors [9].
Results
Baseline characteristics
A total of 68 364 HAV-related adult hospitalizations were identified between 2007 and 2014; however, a total of 47 620 met our study inclusion criteria and of these, 7458 (15.7%) had a concomitant diagnosis of AKI as shown in Fig. 1. Patients in the AKI cohort were significantly older with the mean age of 62.5 vs. 53.7 years in non-AKI patients (P value <0.001), with approximately 88% older than 45 years. The majority of patients were males (59.7% in AKI and 52.3% in the non-AKI group; P value <0.001) and majority belonged to white race (62.1 % in AKI and 59.3% in the non-AKI group; P value <0.001). More than half of the patients in the AKI group (54.5%) had Medicare as a primary payer compared to the non-AKI group (36.7%), P value <0.001. There were no significant differences in median incomes of patients based on their zip codes in two groups. Patients in the AKI group had a higher comorbidity index, with 56.1% having CCI >3 compared to 25.4% in the non-AKI group, P value <0.001. In terms of individual comorbidities, the proportions of hypertension, diabetes mellitus, chronic kidney disease (CKD) and liver cirrhosis were significantly higher in the AKI group relative to the non-AKI group. Almost twice as many patients in the AKI group required a skilled nursing facility on discharge compared to the non-AKI group (22.5 vs. 11.6%; P value <0.001). Rest of the hospital characteristics and geographical regions were similar in the two groups (Table 1).
Fig. 1.
Patient selection flow diagram.AKI, acute kidney injury; HAV, hepatitis A virus; *, nonadult patients, elective admissions, end-stage renal disease patients and those with missing data related to outcomes.
Table 1.
Baseline patient and hospital characteristics for hospitalized cases of acute hepatitis A virus infection with and without acute kidney injury
| Variable | HAV without AKI N (%) | HAV with AKI N (%) | P value |
|---|---|---|---|
| No. of patients | 40161 (84.34%) | 7458 (15.66%) | |
| Patient Characteristics | |||
| Male | 21004 (52.3%) | 4452 (59.7%) | <0.001 |
| Age (years) | |||
| Mean age | 53.74 | 62.52 | <0.001 |
| 18–34 | 6586 (16.4%) | 410 (5.5%) | |
| 35–44 | 5301 (13.2%) | 492 (6.6%) | |
| 45–64 | 17912 (44.6%) | 3229 (43.3%) | |
| ≥65 | 10362 (25.8%) | 3319 (44.5%) | |
| Race | <0.001 | ||
| White | 23815 (59.3%) | 4631 (62.1%) | |
| Black | 5502 (13.7%) | 1313 (17.6%) | |
| Hispanic | 7028 (17.5%) | 962 (12.9%) | |
| Other | 3815 (9.5%) | 552 (7.4%) | |
| Median income based on the zip code | 0.95 | ||
| $1–$38 999 | 12972 (32.3%) | 2439 (32.7%) | |
| $39 000–$47 999 | 10241 (25.5%) | 1939 (26%) | |
| $48 000–$62 999 | 9277 (23.1%) | 1693 (22.7%) | |
| >$63 000 | 7671 (19.1%) | 1380 (18.5%) | |
| Insurance provider | <0.001 | ||
| Medicare | 14739 (36.7%) | 4065 (54.5%) | |
| Medicaid | 9719 (24.2%) | 1387 (18.6%) | |
| Private | 11165 (27.8%) | 1402 (18.8%) | |
| Uninsured | 4538 (11.3%) | 604 (8.1%) | |
| Charlson comorbidity index | <0.001 | ||
| 0 | 16105 (40.1%) | 1193 (16%) | |
| 1 | 8956 (22.3%) | 1029 (13.8%) | |
| 2 | 4900 (12.2%) | 1052 (14.1%) | |
| 3 or more | 10201 (25.4%) | 4184 (56.1%) | |
| Comorbidities | |||
| Hypertension | 16386 (40.8%) | 4236 (56.8%) | <0.001 |
| Diabetes mellitus | 7992 (19.9%) | 2334 (31.3%) | <0.001 |
| Chronic kidney disease | <0.001 | ||
| Stage 3 | 683 (1.7%) | 731 (9.8%) | |
| Stage 4 | 120 (0.3%) | 321 (4.3%) | |
| Stage 5 | 40 (0.1%) | 60 (0.8%) | |
| Liver cirrhosis | 3896 (9.7%) | 1178 (15.8%) | <0.001 |
| Complications | |||
| Electrolyte imbalance | 12450 (31%) | 5280 (70.8%) | <0.001 |
| Acute liver failure | 683 (1.7%) | 701 (9.4%) | <0.001 |
| Hemodialysis | 40 (0.1%) | 619 (8.3%) | <0.001 |
| Mechanical ventilation | 1125 (2.8%) | 1260 (16.9%) | <0.001 |
| Discharge disposition | <0.001 | ||
| Home | 29197 (72.7%) | 3416 (45.8%) | |
| Home with home health | 3574 (8.9%) | 895 (12%) | |
| Skilled nursing facility | 4659 (11.6%) | 1686 (22.6%) | |
| Against medical advice | 924 (2.3%) | 97 (1.3%) | |
| Hospital characteristics | |||
| Teaching status | 0.28 | ||
| Nonteaching hospital | 17229 (42.9%) | 3334 (44.7%) | |
| Teaching hospital | 22932 (57.1%) | 4124 (55.3%) | |
| Hospital location | 0.02 | ||
| Rural hospital | 3655 (9.1%) | 500 (6.7%) | |
| Urban hospital | 36506 (90.9%) | 6958 (93.3%) | |
| Hospital size | <0.001 | ||
| Small | 4739 (11.8%) | 679 (9.1%) | |
| Medium | 10321 (25.7%) | 1805 (24.2%) | |
| Large | 25101 (62.5%) | 4982 (66.8%) | |
| Hospital region | <0.001 | ||
| Northeast | 9317 (23.2%) | 1350 (18.1%) | |
| Midwest | 6707 (16.7%) | 1320 (17.7%) | |
| South | 16225 (40.4%) | 3229 (43.3%) | |
| West | 7912 (19.7%) | 1559 (20.9%) |
AKI, acute kidney injury; HAV, hepatitis A virus.
Prevalence and predictors of acute kidney injury
We observed that prevalence of AKI in hospitalized patients with acute HAV was 15.7% (7 458/47 620). A total of 619 (8.3%) patients who developed AKI required hemodialysis during their hospitalizations. Multivariate logistic regression analysis showed that advanced age [adjusted odds ratio (aOR), 1.02; 95% confidence interval (CI), 1.01–1.04; P value <0.001], African American race (aOR, 1.45; 95% CI, 1.16–1.8; P value <0.001), CCI ≥2 and higher (aOR, 1.73; 95% CI, 1.32–2.26; P value <0.001), CKD stages 3 and 4 (aOR, 3.64; 95% CI, 2.56–5.17; P value <0.001), liver cirrhosis (aOR, 4.8; 95% CI, 1.12–20.63; P value 0.04), fluid and electrolyte imbalance (aOR, 4.08; 95% CI, 3.49–4.78; P value <0.001), acute liver failure (aOR, 4.12; 95% CI, 2.9–5.86; P value <0.001) and the requirement for mechanical ventilation (aOR, 3.69; 95% CI, 2.81–4.85; P value <0.001) were independent predictors for the development of AKI as shown in Table 2.
Table 2.
Bivariate and multivariate logistic regression showing predictors for acute kidney injury in patients hospitalized with acute hepatitis A virus infection.
| Bivariate logistic regression | Multivariate logistic regression | |||||
|---|---|---|---|---|---|---|
|
|
|
|||||
| Odds ratio | 95% confidence interval | P value | Odds ratio | 95% confidence interval | P value | |
| Female | 0.72 | (0.64–0.82) | <0.001 | 0.76 | (0.65–0.9) | <0.001 |
| Race | ||||||
| White | Reference | |||||
| Black | 1.25 | (1.05–1.48) | 0.01 | 1.45 | (1.16–1.8) | <0.001 |
| Hispanic | 0.69 | (0.56–0.84) | <0.001 | 0.82 | (0.64–1.05) | 0.12 |
| Other | 0.77 | (0.6–1) | 0.05 | 0.92 | (0.67–1.27) | 0.62 |
| Age (mean) | 1.03 | (1.03–1.04) | <0.001 | 1.02 | (1.01–1.04) | <0.001 |
| 18–34 | Reference | |||||
| 35–44 | 1.47 | (1.07–2.03) | 0.02 | 0.83 | (0.54–1.29) | 0.41 |
| 45–64 | 2.81 | (2.17–3.64) | <0.001 | 0.95 | (0.58–1.55) | 0.83 |
| ≥65 | 5.14 | (3.96–6.67) | <0.001 | 0.92 | (0.45–1.88) | 0.82 |
| Charlson comorbidity index | ||||||
| 0 | Reference | |||||
| 1 | 1.59 | (1.29–1.97) | <0.001 | 1.24 | (0.95–1.62) | 0.11 |
| 2 | 3.08 | (2.5–3.78) | <0.001 | 1.73 | (1.32–2.26) | <0.001 |
| 3 or more | 5.44 | (4.61–6.42) | <0.001 | 2.74 | (2.17–3.44) | <0.001 |
| Median income based on the zip code | ||||||
| $1–$38 999 | Reference | |||||
| $39 000–$47 999 | 1.02 | (0.87–1.21) | 0.78 | – | – | – |
| $48 000–$62 999 | 0.99 | (0.82–1.18) | 0.89 | – | – | – |
| >$63 000 | 1 | (0.83–1.21) | 0.97 | – | – | – |
| Insurance provider | ||||||
| Medicare | Reference | |||||
| Medicaid | 0.52 | (0.44–0.61) | <0.001 | 0.98 | (0.77–1.26) | 0.9 |
| Private | 0.47 | (0.4–0.55) | <0.001 | 0.97 | (0.77–1.22) | 0.78 |
| Uninsured | 0.48 | (0.38–0.59) | <0.001 | 1.3 | (0.95–1.76) | 0.1 |
| Comorbidities | ||||||
| Diabetes mellitus | 1.82 | (1.59–2.07) | <0.001 | 1.08 | (0.9–1.29) | 0.42 |
| Hypertension | 1.88 | (1.67–2.13) | <0.001 | 1.18 | (1–1.41) | 0.06 |
| Chronic kidney disease | ||||||
| Stage 2 | Reference | |||||
| Stage 3 | 6.19 | (4.7–8.15) | <0.001 | 1.69 | (0.77–3.71) | 0.19 |
| Stage 4 | 15.57 | (9.13–26.55) | <0.001 | 3.64 | (2.56–5.17) | <0.001 |
| Stage 5 | 9.29 | (3.72–23.19) | <0.001 | 5.99 | (2.72–13.18) | <0.001 |
| Liver Cirrhosis | 1.71 | (1.45–2.03) | <0.001 | 4.8 | (1.12–20.63) | 0.04 |
| Complications | ||||||
| Acute liver failure | 6.4 | (4.97–8.26) | <0.001 | 4.12 | (2.9–5.86) | <0.001 |
| Mechanical ventilation | 7.05 | (5.77–8.61) | <0.001 | 3.69 | (2.81–4.85) | <0.001 |
| Electrolyte imbalance | 5.42 | (4.76–6.18) | <0.001 | 4.08 | (3.49–4.78) | <0.001 |
| Hospital location | ||||||
| Rural | Reference | |||||
| Urban | 1.27 | (0.97–1.65) | 0.08 | 1.43 | (1.01–2.03) | 0.05 |
| Hospital teaching status | ||||||
| Nonteaching | Reference | |||||
| Teaching | 0.89 | (0.78–1.01) | 0.07 | 0.96 | (0.81–1.14) | 0.63 |
| Hospital size | ||||||
| Small | Reference | |||||
| Medium | 1.21 | (0.96–1.53) | 0.11 | 1.29 | (0.95–1.74) | 0.11 |
| Large | 1.36 | (1.1–1.67) | <0.001 | 1.47 | (1.11–1.94) | 0.01 |
| Hospital region | ||||||
| Northeast | Reference | |||||
| Midwest | 1.39 | (1.13–1.7) | <0.001 | 1.01 | (0.77–1.34) | 0.94 |
| South | 1.43 | (1.2–1.7) | <0.001 | 1.03 | (0.83–1.27) | 0.79 |
| West | 1.38 | (1.13–1.69) | <0.001 | 1.23 | (0.96–1.59) | 0.1 |
Resource utilization
AKI contributed to an increase in the mean LOS (10.03 vs. 5.63 days; P value <0.001) with increased hospitalization costs ($27 171.35 vs. $12 790.09; P value <0.001) as shown in Table 3. On adjusted analysis, AKI was found to be an independent predictor of increased LOS (β: 2.12; 95% CI, 1.39–2.85; P value <0.001) and hospitalization cost (β: 5763.39; 95% CI, 3790.5–7736.28; P value <0.001) as depicted in Table 4. Moreover, patients in the AKI group had more electrolyte and fluid imbalance (70.80 vs. 31%; P value <0.001) and required more hemodialysis (8.3 vs. 0.1%; P value <0.001). Similarly, patients in the AKI group needed mechanical ventilation (16.9 vs. 2.8%; P value <0.001) when compared to the patients without AKI.
Table 3.
Secondary outcomes and the impact of acute kidney injury on the length of stay, hospitalization cost and all-cause inpatient mortality, in patients hospitalized with acute hepatitis A virus infection.
| HAV without AKI | HAV with AKI | Crude odds ratio/coefficient (95% confidence interval) | Adjusted odds ratio/coefficient (95% confidence interval) | P value | |
|---|---|---|---|---|---|
| Length of Stay | 5.63 days | 10.03 days | 4.40 (3.78–5.03) | 2.12 (1.38–2.84) | <0.001 |
| Mean hospitalization cost | $12 790.09 | $27 171.35 | 14 381.26 (12284.28–16478.24) | 5763.39 (3790.50–7736.28) | <0.001 |
| Mortality | 611 | 1023 | 10.27 (7.98–13.23) | 3.28 (2.23–4.84) | <0.001 |
| Hemodialysis | 40 (0.1%) | 619 (8.3%) | – | – | <0.001 |
| Mechanical Ventilation | 1125 (2.8%) | 1260 (16.9%) | – | – | <0.001 |
AKI, acute kidney injury; HAV, hepatitis A virus.
Table 4.
Multivariate regression analysis to study the impact of acute kidney injury on secondary outcomes including length of stay, hospitalization cost and all cause in-hospital mortality
| Length of stay | Hospitalization cost | Mortality | |||||||
|---|---|---|---|---|---|---|---|---|---|
|
|
|
|
|||||||
| Coefficient | 95% confidence interval | P value | Coefficient | 95% confidence interval | P value | Odds ratio | 95% confidence interval | P value | |
| Acute kidney injury | 2.12 | (1.39–2.85) | <0.001 | 5763.39 | (3790.5–7736.28) | <0.001 | 3.29 | (2.23–4.85) | <0.001 |
| Female | −0.34 | (−0.66 to −0.02) | 0.04 | −1210.75 | (−2040.45 to −381.05) | <0.001 | 0.77 | (0.54–1.1) | 0.15 |
| Race | |||||||||
| White | Reference | ||||||||
| Black | 0.09 | (−0.45 to 0.62) | 0.76 | – | – | 0.56 | (0.34–0.95) | 0.03 | |
| Hispanic | 0.28 | (−0.29 to 0.85) | 0.34 | – | – | 1.09 | (0.67–1.77) | 0.73 | |
| Other | 0.13 | (−0.44 to 0.7) | 0.65 | – | – | 0.8 | (0.36–1.78) | 0.59 | |
| Age (mean) | 0.03 | (0.01–0.06) | 0.01 | 72.54 | (−2.8 to 147.88) | 0.06 | 1.07 | (1.04–1.1) | <0.001 |
| 18–34 | Reference | ||||||||
| 35–44 | 0.1 | (−0.6 to 0.8) | 0.78 | 904.73 | (−783.09 to 2592.54) | 0.29 | 1.05 | (0.35–3.2) | 0.93 |
| 45–64 | −0.18 | (−1.05 to 0.7) | 0.69 | 606.22 | (−1548.12 to 2760.55) | 0.58 | 0.68 | (0.2–2.3) | 0.53 |
| ≥65 | −0.84 | (−2.28 to 0.61) | 0.26 | −415.99 | (−4539.12 to 3707.15) | 0.84 | 0.38 | (0.07–1.95) | 0.25 |
| Charlson comorbidity index | |||||||||
| 0 | Reference | ||||||||
| 1 | −0.34 | (−0.76 to 0.09) | 0.12 | 937.05 | (−70.14 to 1944.24) | 0.07 | 1.11 | (0.57–2.17) | 0.76 |
| 2 | −0.07 | (−0.56 to 0.43) | 0.79 | 5610.02 | (3777.32–7442.72) | <0.001 | 1.56 | (0.8–3.05) | 0.19 |
| 3 or more | −0.12 | (−0.62 to 0.38) | 0.64 | 4135.22 | (2722.1–5548.34) | <0.001 | 3.43 | (1.9–6.18) | <0.001 |
| Median income based on the zip code | |||||||||
| $1–$38 999 | Reference | ||||||||
| $39 000–$47 999 | −0.34 | (−0.76 to 0.09) | 0.12 | 1124.52 | (144.59–2104.44) | 0.03 | 0.87 | (0.55–1.38) | 0.56 |
| $48 000–$62 999 | −0.07 | (−0.56 to 0.43) | 0.79 | 2337.87 | (1021.11–3654.62) | <0.001 | 0.84 | (0.51–1.39) | 0.51 |
| >$63 000 | −0.12 | (−0.62 to 0.38) | 0.64 | 2845.68 | (1560.35–4131) | <0.001 | 0.71 | (0.41–1.24) | 0.23 |
| Insurance provider | |||||||||
| Medicare | Reference | ||||||||
| Medicaid | 0.55 | (−0.04 to 1.14) | 0.07 | 962.68 | (−426.43 to 2351.8) | 0.17 | 1.44 | (0.84–2.46) | 0.18 |
| Private | −0.61 | (−1.09 to −0.13) | 0.01 | 233.46 | (−1208.45 to 1675.37) | 0.75 | 1.72 | (1.05–2.8) | 0.03 |
| Uninsured | 0.39 | (−0.27 to 1.05) | 0.25 | 717.32 | (−967.24 to 2401.88) | 0.4 | 2.51 | (1.39–4.54) | <0.001 |
| Comorbidities | |||||||||
| Diabetes mellitus | – | – | – | – | – | – | – | – | – |
| Hypertension | – | – | – | – | – | – | 0.55 | (0.38–0.8) | <0.001 |
| Chronic kidney disease | |||||||||
| Stage 2 | Reference | Reference | |||||||
| Stage 3 | −2.36 | (−3.14 to −1.57) | <0.001 | −6143.27 | (−8345.9 to −3940.64) | <0.001 | 0.51 | (0.23–1.13) | 0.1 |
| Stage 4 | −1.11 | (−2.73 to 0.5) | 0.18 | −5020.16 | (−9414.02 to −626.3) | 0.03 | 0.37 | (0.09–1.45) | 0.15 |
| Stage 5 | −0.63 | (−2.4 to 1.14) | 0.49 | −4761.41 | (−9536.85 to 14.04) | 0.05 | 2.12 | (0.31–14.49) | 0.44 |
| Liver Cirrhosis | −0.7 | (−1.29 to −0.11) | 0.02 | −1712.31 | (−3324.09 to −100.52) | 0.04 | 1.31 | (0.86–2) | 0.22 |
| Complications | |||||||||
| Acute liver failure | 0.59 | (−0.74 to 1.91) | 0.39 | 3683.99 | (−257.06 to 7625.04) | 0.07 | 1.5 | (0.83–2.72) | 0.18 |
| Mechanical ventilation | 7.27 | (5.76–8.78) | <0.001 | 34713.52 | (29759.79–39667.25) | <0.001 | 22.62 | (15.28–33.5) | <0.001 |
| Hemodialysis | 3.84 | (1.78–5.9) | <0.001 | 14699.01 | (7360.13–22037.89) | <0.001 | 2.31 | (1.16–4.57) | 0.02 |
| Electrolyte imbalance | 1.54 | (1.19–1.9) | <0.001 | 2683.94 | (1785.99–3581.88) | <0.001 | 1.42 | (0.99–2.04) | 0.06 |
| Hospital location | |||||||||
| Rural | Reference | ||||||||
| Urban | 0.78 | (0.27–1.3) | <0.001 | 3717.26 | (2706.74–4727.78) | <0.001 | – | – | – |
| Hospital teaching status | |||||||||
| Nonteaching | Reference | ||||||||
| Teaching | 0.86 | (0.49–1.23) | <0.001 | −207.45 | (−1444.72 to 1029.82) | 0.74 | – | – | – |
| Hospital size | |||||||||
| Small | Reference | ||||||||
| Medium | −0.31 | (−0.88 to 0.25) | 0.28 | −779.11 | (−2116.11 to 557.9) | 0.25 | – | – | – |
| Large | 0.54 | (0.01–1.08) | 0.05 | 1716.45 | (385.51–3047.38) | 0.01 | – | – | – |
| Hospital region | |||||||||
| Northeast | Reference | ||||||||
| Midwest | −1.08 | (−1.65 to −0.51) | <0.001 | −2311.61 | (−3853.82 to −769.4) | <0.001 | 0.6 | (0.3–1.21) | 0.15 |
| South | −0.3 | (−0.82 to 0.21) | 0.25 | −1250.56 | (−2693.77 to 192.64) | 0.09 | 1 | (0.63–1.59) | 0.99 |
| West | −0.86 | (−1.44 to −0.27) | <0.001 | 1927.68 | (125.21–3730.15) | 0.04 | 0.83 | (0.49–1.42) | 0.5 |
All-cause in-hospital mortality
Increased all-cause in-hospital mortality was also observed among the HAV-infected patients with AKI [1023/7 458; 13.8% as compared to HAV infection without AKI (681/40 161; 1.5%), P value <0.001]. We found that AKI was an independent predictor of mortality (aOR, 3.29; 95% CI, 2.23–4.85; P value <0.001) after adjusting for confounders such as age, gender, race, comorbidities, insurance status, hospital location, hospital teaching status, hospital bed size and geographic region.
Discussion
In our study, we used the largest national inpatient database to assess the overall prevalence of AKI in patients admitted with acute HAV infection and we found that about 15.7% patients presented with or developed AKI during their course of hospitalization. Patients with advanced age, black race, higher comorbidity burden, especially with CKD and those who develop complications such as fluid and electrolyte imbalance, acute liver failure or require mechanical ventilation were at a higher risk of having AKI. These patients also stayed longer in the hospital, that is, twice as compared to patients without AKI which eventually increased the hospitalization cost substantially. Similarly, patients hospitalized with HAV who had concomitant AKI are more likely to die during the hospitalization with AKI as an independent predictor of mortality after adjusted for confounding factors.
Our study is an extension of the findings published in the previous studies from outside the USA. The prevalence of AKI in hospitalized patients with HAV in small single center retrospective studies is reported between 3 and 7.2% [5,10–12]. In a retrospective study performed by Choi et al., [13] authors have reported an incidence of 7.6% in patients with acute HAV infection in both fulminant and nonfulminant cases. However, after excluding the fulminant cases of acute HAV, the incidence of AKI remained as high as 5.2%. Some of the risk factors for the development of AKI reported in these studies are advanced age, diabetes, heavy alcohol intake and fulminant course of hepatitis A [10,11,13]. However, we found that 15.7% of the patients admitted with acute HAV infection had a concomitant discharge diagnosis of AKI. In our analysis, advanced age, black race, CCI ≥2 and more, preexisting CKD, fluid, and electrolyte imbalance, acute liver failure, and the requirement for mechanical ventilation are identified as independent predictors for the development of AKI. The findings in our study are more generalizable due to large sample size, which represent patients from different hospitals varying in size, academic status, rural and urban location as well as geographical region.
The precise mechanism of AKI in patients with HAV remains unknown. Prerenal causes, hyperbilirubinemia, pigment nephropathy and immune-complex-mediated nephritis have all been proposed for the pathophysiology in development of AKI in patients with acute hepatitis infection [14]. In a study of 26 patients with HAV and AKI, Kim et al. [10] found that only three patients had urinary sodium of <20 meq/L suggesting the majority of the cohort had AKI secondary to nonprerenal causes. Pigment nephropathy and cholemic nephrosis related to direct tubular dysfunction and acute tubular necrosis have also been well described in renal biopsies [15]. Human hepatocyte growth factor (hHGF) is a paracellular growth factor and has been implicated as one of the pathogenic mechanisms in fulminant liver failure due to acute viral hepatitis [16]. Although high levels of hHGF were found in nonsurvivors as compared to survivors, others have speculated that its ability to help in tissue regeneration and angiogenesis might help with recovery from liver and kidney failure [17].
In general, AKI is also associated with increased healthcare utilization in patients hospitalized with various conditions [18]. In a large retrospective study performed by Bedford et al. [19] consisting of 66 829 hospitalized patients admitted for various reasons, the patients who developed AKI had 2.2 times higher LOS after adjusting for confounding factors. We found that HAV patients who developed AKI had a longer hospital stay (10.03 days in the AKI group vs. 5.63 days in the non-AKI group), which is also directly responsible for a higher hospitalization cost. A recent report on the hospital cost per HAV hospitalization reported an average cost of $16 232 (SD $602; 95% CI, $15 052–$17 411) [20]. However, we observed that the hospital cost is almost twice in patients who develop AKI as compared to the non-AKI group. The increase in LOS and higher hospitalization cost in the AKI group can be attributable to dialysis-related procedure expenses and nephrology consultation; however, Collister et al., [18] reported that even the mildest forms of AKI contributed 1.2–1.3 times higher cost of care in patients who developed AKI than those without AKI.
The overall mortality rate for HAV-infected patients have been reported less than 1% (range 0.3–1%) [21]; however, the mortality in patients who develop AKI is estimated to be significantly higher in previous studies. In a study reported by Choi et al., [13] the authors reported in-hospital mortality of 16.7% for patients in the AKI group, slightly higher than our study findings (13.8%) [13]. Lin et al. [22] reported in their prospective study including patients with fulminant HAV infection, day one serum creatinine level of >2.0 mg/dl was found to be 54% sensitive and 88% specific for predicting nonsurvival, defined as death or need for liver transplantation. We found AKI to be an independent predictor of mortality (adj OR, 3.29; 95% CI, 2.23–4.85; P <0.001) after adjusting for confounders such as age, gender, race, comorbidities, insurance status, hospital location, teaching, bed size and geographic region. Although it is not possible to establish a causal link between AKI and mortality through this epidemiologic study but higher mortality in patients with AKI can be explained by the presence of concomitant multiple organ system failures, especially in critically ill patients.
We acknowledge several limitations to our study findings; majority of them are related to the database itself. First, we utilized the NIS database, which is subjective to coding errors due to its administrative nature and dependence on the ICD-9 coding system [23]. However, the ICD-9 codes we used in our study are very well-validated and used in several studies based on the NIS database [24,25]. Second, the lack of data availability on medications, laboratory and radiologic data has significantly limited our ability to determine the baseline assessment of kidney function, the severity of AKI, and dynamicity in the serum creatinine level and individual patient level management strategies. Nevertheless, we used CCI, which is a validated tool to determine the comorbidity burden of patients as an important prognostic factor. Third, Agency for Healthcare Research and Quality recommends against using the NIS to derive state-level estimates owing to the small representation of hospitals in the NIS from many states [26]. State-level data, if obtainable, could have been helpful to public health policymakers to help focus public health efforts in states affected more with HAV infection. However, we included the geographical regions to study the variations in the outcomes across the USA. Fifth, as noted by Hofmeister et al., [20] cost estimates utilizing an administrative database can significantly overestimate hospitalization costs as patients who were incidentally diagnosed with HAV infection but were admitted primarily for other more costly conditions might have skewed the cost analysis. We limited our analysis to nonelective hospitalizations which excludes admissions planned for procedures, hence minimizing the probability of overestimation in the cost analysis. Also, our data only represent the costs and outcomes of HAV-infected patients admitted to hospitals and do not capture visits to emergency rooms, primary care physician offices and urgent care; hence, it may not represent the true incidence and resource utilization of HAV infections in the USA.
To our knowledge, despite of all the limitations, this is the first study to report the impact of AKI in the hospitalized patients with acute HAV infection in the USA on a large scale. Our report suggests that AKI is an independent predictor of increased length of hospital stay, higher total hospitalization costs and inpatient mortality. Our study highlights that AKI in acute HAV patients affects the healthcare utilization, morbidity and mortality. Therefore, it is very important to identify patients who are at risk for developing AKI and timely manage them with adequate fluid resuscitation to prevent AKI which ultimately will help to reduce healthcare utilization, morbidity and mortality.
Supplementary Material
Fig. 2.
Primary and secondary outcomes in patients with acute hepatitis A with and without acute kidney injury (AKI); (a) length of stay (b) hospitalization cost (c) inpatient mortality (d) mechanical ventilation. HAV, hepatitis A virus.
Acknowledgements
The authors would like to thank Dr. Karla D Passalacqua, PhD (Senior Medical Writer) from Henry Ford Health System to review the manuscript and provided the suggestions to improve the quality of the manuscript.
The authors received no financial support or grants for the research, authorship, and publication of this article. The study is performed utilizing a de-identified national database.
Study concept and design: A.K., S.S., A.S., E.M. and K.M. Data analysis: A.K., K.G. and A.M. Drafting of the manuscript: A.K., K.S., A.M., M.M.B., K.N., K.G. and A.S. All authors interpreted the data, and critically revised the manuscript for important intellectual content and approved the final version of the manuscript and give consent to publish the study.
Footnotes
Conflicts of interest
There are no conflicts of interest.
Supplemental Digital Content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s website, www.eurojgh.com
References
- 1.Hepatitis A. 2015. https://www.who.int/immunization/diseases/hepatitisA/en/. [Accessed 7 September 2020]
- 2.Widespread person-to-person outbreaks of hepatitis A across the United States. https://www.cdc.gov/hepatitis/outbreaks/2017March-HepatitisA.htm. [Accessed 7 September 2020]
- 3.Foster MA, Hofmeister MG, Kupronis BA, Lin Y, Xia GL, Yin S, et al. Increase in Hepatitis A Virus Infections - United States, 2013–2018. MMWR Morb Mortal Wkly Rep 2019; 68:413–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Kemmer NM, Miskovsky EP. Hepatitis A. Infect Dis Clin North Am 2000; 14:605–615. [DOI] [PubMed] [Google Scholar]
- 5.Kim HW, Yu MH, Lee JH, Chang JW, Yang WS, Kim SB, et al. Experiences with acute kidney injury complicating non-fulminant hepatitis A. Nephrology (Carlton) 2008; 13:451–458. [DOI] [PubMed] [Google Scholar]
- 6.INTRODUCTION TO THE HCUP NATIONAL INPATIENT SAMPLE (NIS) 2017. Updated May 12, 2019. https://www.hcup-us.ahrq.gov/db/nation/nis/NISIntroduction2017.pdf. [Accessed 15 September 2020]
- 7.Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 1987; 40:373–383. [DOI] [PubMed] [Google Scholar]
- 8.Cost-to-Charge Ratio Files. 2020. https://www.hcup-us.ahrq.gov/db/ccr/costtocharge.jsp. [Accessed 15 September 2020]
- 9.Medical Expenditure Panel Survey (MEPS). Agency for Healthcare Research and Quality. Rockville, MD. https://www.ahrq.gov/cpi/about/otherwebsites/meps.ahrq.gov/index.html. [Accessed 9 September 2020] [Google Scholar]
- 10.Kim SH, Yoon HE, Kim YK, Kim JY, Choi BS, Choi YJ, et al. Acute hepatitis A-associated acute renal failure in adults. Nephron Clin Pract 2008; 109:c127–c132. [DOI] [PubMed] [Google Scholar]
- 11.Jung YJ, Kim W, Jeong JB, Kim BG, Lee KL, Oh KH, et al. Clinical features of acute renal failure associated with hepatitis A virus infection. J Viral Hepat 2010; 17:611–617. [DOI] [PubMed] [Google Scholar]
- 12.Kwon SY, Park SH, Yeon JE, Jeong SH, Kwon OS, Lee JW, et al. Clinical characteristics and outcomes of acute hepatitis a in Korea: a nationwide multicenter study. J Korean Med Sci 2014; 29:248–253. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Choi HK, Song YG, Han SH, Ku NS, Jeong SJ, Baek JH, et al. Clinical features and outcomes of acute kidney injury among patients with acute hepatitis A. J Clin Virol 2011; 52:192–197. [DOI] [PubMed] [Google Scholar]
- 14.Lin CC, Chang CH, Lee SH, Chiang SS, Yang AH. Acute renal failure in non-fulminant hepatitis A. Nephrol Dial Transplant 1996; 11:2061–2066. [DOI] [PubMed] [Google Scholar]
- 15.Betjes MG, Bajema I. The pathology of jaundice-related renal insufficiency: cholemic nephrosis revisited. J Nephrol 2006; 19:229–233. [PubMed] [Google Scholar]
- 16.Tsubouchi H, Kawakami S, Hirono S, Miyazaki H, Kimoto M, Arima T, et al. Prediction of outcome in fulminant hepatic failure by serum human hepatocyte growth factor. Lancet 1992; 340:307. [DOI] [PubMed] [Google Scholar]
- 17.Oe S, Shibata M, Miyagawa K, Honma Y, Hiura M, Abe S, Harada M. Hepatitis A complicated with acute renal failure and high hepatocyte growth factor: a case report. World J Gastroenterol 2015; 21:9671–9674. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Collister D, Pannu N, Ye F, James M, Hemmelgarn B, Chui B, et al. ; Alberta Kidney Disease Network. Health care costs associated with AKI. Clin J Am Soc Nephrol 2017; 12:1733–1743. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Bedford M, Stevens PE, Wheeler TW, Farmer CK. What is the real impact of acute kidney injury? BMC Nephrol 2014; 15:95. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Hofmeister MG, Yin S, Aslam MV, Teshale EH, Spradling PR. Hepatitis A hospitalization costs, United States, 2017. Emerg Infect Dis 2020; 26:1040–1041. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Ly KN, Klevens RM. Trends in disease and complications of hepatitis A virus infection in the United States, 1999–2011: a new concern for adults. J Infect Dis 2015; 212:176–182. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Taylor RM, Davern T, Munoz S, Han SH, McGuire B, Larson AM, et al. ; US Acute Liver Failure Study Group. Fulminant hepatitis A virus infection in the United States: incidence, prognosis, and outcomes. Hepatology 2006; 44:1589–1597. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Khera R, Krumholz HM. With great power comes great responsibility: big data research from the National inpatient sample. Circ Cardiovasc Qual Outcomes 2017; 10:e003846. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Cheungpasitporn W, Thongprayoon C, Ungprasert P, Wijarnpreecha K, Mao MA, Aeddula NR, et al. Hepatitis A hospitalizations among kidney transplant recipients in the United States: nationwide inpatient sample 2005–2014. Eur J Gastroenterol Hepatol 2020; 32:650–655. [DOI] [PubMed] [Google Scholar]
- 25.Pavkov ME, Harding JL, Burrows NR. Trends in Hospitalizations for Acute Kidney Injury - United States, 2000–2014. MMWR Morb Mortal Wkly Rep 2018; 67:289–293. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Why the NIS should not be used to make State-level estimates. Updated May 1, 2016. https://www.hcup-us.ahrq.gov/db/nation/nis/nis_statelevelestimates.jsp. [Accessed 20 September 2020]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.