Comparison of Two Intravenous Fluid Management Strategies in COVID-19 Patients in Hubei, China: A Retrospective Multicenter Study

Optimal uid management in patients with COVID-19 has not been reported. This retrospective, multicenter study investigated the impact of intravenous infusion volume in the early stage of COVID-19 on clinical outcomes. Methods 127 patients from two tertiary hospitals were separated into the “conservative” and “liberal” groups based on average daily intravenous infusion volume within the rst seven days after admission. Basic information, demographic and epidemiological characteristics, laboratory ndings, treatments, and outcome measures were retrieved from medical records. The disease progression and prognosis were analyzed and compared.


Abstract Background
Optimal uid management in patients with COVID-19 has not been reported. This retrospective, multicenter study investigated the impact of intravenous infusion volume in the early stage of COVID-19 on clinical outcomes. Methods 127 patients from two tertiary hospitals were separated into the "conservative" and "liberal" groups based on average daily intravenous infusion volume within the rst seven days after admission. Basic information, demographic and epidemiological characteristics, laboratory ndings, treatments, and outcome measures were retrieved from medical records. The disease progression and prognosis were analyzed and compared.

Results
The average daily intravenous infusion volume within 7 days was 500 (150-700) ml/day in the conservative-strategy group (n = 87), and 1100 (1000-1288) ml/day in the liberal-strategy group (n = 40) (p < 0.001). There were no statistical differences in median age, male-to-female ratio, epidemiology, laboratory ndings on admission, comorbidities, and average daily urine output within the seven days (p > 0.05). The nal K + in the liberal group was slightly higher than that at admission, and the nal hematocrit level in the conservative group had a signi cant difference than that at admission (p < 0.05).

Conclusions
Although there appeared to be no signi cant difference in the duration of hospitalization between using conservative and liberal uid management strategies, the former was associated with lower rates of disease progression, mechanical ventilation and in-hospital mortality without increased nonpulmonaryorgan dysfunction. These results support the importance of implementing conservative intravenous uid infusion in the early stage of COVID-19.

Background
Since its outbreak in Wuhan, COVID-19 has spread globally, causing over 11 million infections and 537,000 deaths as of July 6, 2020. 1,2 The World Health Organization (WHO) had declared COVID-19 a pandemic in March, 2020. 3 Despite the gradual decrease in cases in some countries, the pandemic has not yet been brought under control in many other regions of the world.
Studies on its clinical presentation, laboratory and pathological ndings along with its clinical outcome have shown that a large proportion of patients died of SARS-CoV-2 associated acute lung injury (ALI) or acute respiratory distress syndrome (ARDS). 4,5 Yet, the optimal uid management for ALI/ARDS has not been determined. 6-9 Most practitioners believed that appropriate restriction of uid input is bene cial to a good prognosis of ARDS, and previous studies have supported that a conservative uid-management strategy for patients with ALI or ARDS may alleviate lung edema, improve lung function, shorten Intensive Care Unit (ICU) stay, decrease duration of mechanical ventilation, and improve prognosis. [10][11][12] However, some scholars believed that possible risks of this approach are the decrease of cardiac output and deterioration of nonpulmonary-organ function. [13][14][15][16] The incidence of advanced ARDS in patients with COVID-19 pneumonia has been reported to be over 20%. 17 Nevertheless, the impact of uid restriction in the early stages of COVID-19 on clinical outcomes is unclear. 18 Current evidence is insu cient to support the use of either a liberal or conservative intravenous uid strategy in the early stages. In this study, we enrolled patients with con rmed COVID-19 from two hospitals in the epicenter province of Hubei (with Wuhan as its capital) and investigated the impact of the two strategies on the outcomes in COVID-19 patients.

Study design and participants
We conducted a retrospective, cross-sectional, multicenter study focusing on intravenous uid management in patients diagnosed with COVID-19 from January 1 to April 1, 2020, recruited from the A liated Hospital of Wuhan University of Science and Technology and Huanggang Central Hospital, two tertiary hospitals in Hubei, China. This study was approved by the ethics committees of the two hospitals.
A total of 236 patients with unexplained pneumonia on general wards in the two hospitals were randomly selected as the initial study population. They would then be further checked with the following inclusion and exclusion criteria.
Based on the clinical diagnostic standards in the New Diagnosis and Treatment Scheme for Novel Coronavirus Infected Pneumonia (Trial Edition 7) 19 issued by the National Health Commission of the People's Republic of China, the inclusion criteria were as follows: 1, their COVID-19 infection was con rmed by SARS-CoV-2 positive nding by real-time polymerase chain reaction (RT-PCR) according to WHO interim guidance using nasal and/or throat swabs. 2, their COVID-19 pneumonia was further con rmed by typical chest computerized tomography (CT) ndings.
Those patients who met the following criteria were excluded from the study: 1, without complete laboratory data, including blood counts, blood biochemistry and coagulation parameters. 2, without complete physician's orders sheet. 3, with fewer than seven days' in-hospital stay. 4, having developed shock within seven days since hospitalization. 5, with severe cardiovascular diseases on admission, such as acute myocardial infarction or congestive heart failure. 6, with severe renal impairment, such as acute kidney injury. 7, respiratory failure on admission or requiring mechanical ventilation. And/or 8, with respiratory infections caused by other viruses, including in uenza A virus, in uenza B virus and respiratory syncytial virus.

Study Procedures
Demographic characteristics, epidemiological history, clinical manifestation, daily intravenous uid administration and urine output within seven days after admission, admission assessment (curb-65 score), laboratory data, treatment programs and outcome measures were extracted from quali ed patients' medical records. Clinical outcomes were followed up to May 1, 2020. Patients were divided into two groups according to the average daily intravenous infusion volume within seven days after admission. Those with an average daily intravenous infusion volume of < 1000 ml were assigned to the conservative-strategy group whereas those with the volume of ≥ 1000 ml to the liberal-strategy group. All data were separately obtained by two authors, and when necessary, missing data were re-collected through communications with attending physicians and other medical personnel. Through comparative analyses of the laboratory data and outcomes of the two groups, the impact of intravenous infusion volume on progression and prognosis was observed. Progression of the disease was assessed in terms of an changed curb-65 score.

Outcomes
In this study, we calculated average daily intravenous uid administration and urine output of seven days after admission, and collected the epidemiological data (i.e., clear or unclear contact history and illness onset to admission), admission assessment (curb-65 Score), coexisting illnesses, laboratory data, medical treatments and clinical outcomes. Our main observed outcomes included duration of hospitalization, organ functional changes, the rate of disease progression and the necessity for mechanical ventilation. The secondary outcome was in-hospital mortality.

Statistical Analysis
For continuous variables, such as age and average daily urine volume within seven days, mean (±SD) was used for normally distributed data and median with interquartile range for non-normally distributed data, followed by t-test and non-parametric tests when appropriate. Categorical variables such as chronic medical illness and in-hospital mortality were expressed as number (%) and compared using the χ 2 test.
For laboratory results, we also assessed whether the measurements were outside the normal range. All analyses were performed using SPSS version 23.0 (IBM Corporation, Armonk, NY, USA). A two-tailed value of p < 0.05 was considered statistically signi cant.

Enrollment and Exclusions
A total of 236 patients with unexplained pneumonia were initially enrolled during screening. Figure 1 shows the reasons for subsequent exclusion of 109 patients. First excluded were 46 patients for having no RT-PCR test or negative test results. Sixty-three patients were excluded for other reasons such as lacking laboratory data or complete physician's orders sheet and duration of hospitalization fewer than seven days. Some of the patients had more than one reason for exclusion. Of 127 patients eventually included, 87 patients received conservative uid administration while 40 liberal uid administration.

Baseline characteristics
The baseline characteristics of the 127 patients with COVID-19 are provided in Table 1. The two groups were similar with respect to demographic characteristics, epidemiological history, time from onset of illness to admission, disease severity on admission, organ dysfunction, coexisting illnesses and treatment programs (all p > 0.05). There were no signi cantly statistical differences in the results of laboratory data on admission between the two groups (all p > 0.05).

Main outcome measures
The average daily intravenous infusion volume over seven days was 500 (150-700) ml/day in the conservative group, as compared to 1100 (1000-1288) ml/day in the liberal group (p < 0.001). There was no signi cant difference between the two groups in the average daily urinary output during 7 days after admission (1514.93 ± 137.79 ml/day vs. 1511.23 ± 243.28 ml/day, p = 0.751) ( Table 2).
In the nal laboratory examination after seven days of uid management and before discharge, the lymphocyte count and albumin in the conservative group were higher than those in the liberal group (all p < 0.05). In addition, we found that urea nitrogen value and ESR were lower in the conservative group (all p < 0.05). No signi cant differences were found in hematocrit, bilirubin, aminotransferases, creatinine, CRP, NT-pro BNP, ESR and other laboratory results between the two groups (all p > 0.05) ( Table 2). In the intragroup comparison of laboratory tests, the mean nal K + in the liberal group was higher than that at admission, and the mean nal hematocrit level in the conservative group was lower than that at admission (all p < 0.05), but within the normal range. In addition, the hematocrit, Na + in the liberal group and K + , Na + in the conservative group showed no difference (all p > 0.05).
The duration of hospitalization was 22.41±11.99 days in the conservative strategy group and 25.28±12.08 days in the liberal group (p = 0.120). However, when compared to the liberal group, the conservative group was found to have statistically lower rates of disease progression (9.3% vs 37.5%, p < 0.001), mechanical ventilation (2.3% vs 27.5%, p < 0.001) and in-hospital mortality (2.3% vs 15.0%, p = 0.012) ( Table 2).

Discussion
While knowledge regarding epidemiologic features and diagnostic tools of COVID-19 is rapidly evolving, uncertainties surrounding various aspects of optimal management strategies persist. Front-line multidisciplinary experts are deeply concerned about the irregular use of antimicrobials and adjuvant drugs in the international guidelines, as this has resulted in an excessive amount of intravenous uids in some patients in the short term. 20 Whether uid management should be introduced early in COVID-19 patients to slow down disease progression has not been studied. Relevant previous research into uid management has mostly focused on patients with ARDS or in other critical conditions. In a study in 416 patients with COVID-19, Shi et al 17  To the best of our knowledge, this study has been the rst to focus on the relationship between early intravenous uid volume and prognosis of COVID-19. It was conducted by examining the medical records of patients during the initial outbreak of COVID-19 in two hospitals in Hubei Province from January to April 2020. In this retrospective study, we chose the daily intravenous infusion volume as a quantitative measure. Although we did not detect a signi cant difference between the conservative and liberal strategies of intravenous uid management on the duration of hospitalization, the conservative strategy was signi cantly associated with lower rates of disease progression, mechanical ventilation and hospital mortality. Therefore, our data appear to suggest that a conservative intravenous uid strategy leads to fewer complications and better prognosis than a liberal intravenous uid strategy when managing COVID-19 patients. We speculate that the worse prognosis in the liberal group might be caused by a side effect of over intravenous uid infusion, resulting in pulmonary interstitial edema.
No signi cant differences were found in the WBC, hemoglobin, hematocrit levels and other laboratory results between the two groups. The available data are not su cient to distinguish among these or offer any other potential explanations.
However, during the period studied, most patients' food and drink oral intake was based on their personal appetite and not quanti ed. The patients' body weight was not monitored, either. These may affect the accuracy in assessing body water retention and need to be addressed in future study.
Although the mean blood urea nitrogen level was slightly lower in the conservative group than that in the liberal group, there were no signi cant differences in the creatinine level, cardiac and renal function between the two groups. In addition, the conservative strategy was associated with a slightly higher lymphocyte count and albumin level, which might be related to mild dehydration in the patients. The intragroup comparison of laboratory tests showed that there was no dehydration in the conservative strategy and both groups had anhydrous electrolyte imbalance.
In summary, our initial study suggests that conservative intravenous uid management may result in a favorable prognosis of COVID-19 within the rst seven days after patients' admission. Therefore, clinical treatment may need considering the total amount of daily intravenous infusion in such patients. This should be investigated in prospective and large-scale studies.

Conclusions
Although there appeared to be no signi cant difference in the duration of hospitalization between using conservative and liberal uid management strategies, the former was associated with lower rates of disease progression, mechanical ventilation and in-hospital mortality without increased nonpulmonaryorgan dysfunction. These results support the importance of implementing conservative intravenous uid infusion in the early stage of COVID-19. Availability of data and materials: The datasets used and analyzed during the current study are available from the corresponding author on reasonable request.

Competing interests:
The authors declare that they have no competing interests.

Funding:
This study was funded by Beijing Municipal Natural Science Foundation General Program (CN, 7192197).