Introduction

The World Health Organization (WHO) declared the coronavirus-19 disease (COVID-19) pandemic in March 2020. This was initially identified in a group of patients with respiratory compromise in December 2019 in Wuhan. The infection spread rapidly throughout Asia, Europe, and other continents; in Peru, the first case was identified on March 6th, 2020, and to date, several cases and high mortality rates have been reported, ranking as the fifth most affected country in Latin America [1,2,3,4,5,6,7].

Infection is more frequent in older adults and males, with comorbidities such as diabetes mellitus and other immunosuppressive conditions. The clinical presentation is characterized by persistent fever, lymphopenia, high ferritin, pulmonary involvement in more than 50% of cases, and other findings associated with increased interleukins [8,9,10,11,12].

Chronic kidney disease shows a progressive increase in frequency associated with the need for replacement therapy in any of its modalities [2]. The population with chronic kidney disease in all stages has multiple comorbidities associated with increased mortality during COVID-19 disease. The immune response of patients with chronic kidney disease is diminished, with less chance of presenting cytokine storm, which explains why some case series presents low percentages of severe manifestations. However, thay are at increased risk of severe respiratory infections due to associated comorbidities, nutritional status and hospital admissions [12,13,14,15].

Studies of hemodialysis patients with COVID-19 show different results. Some report mild disease with non-severe pneumonia, associated with reduced immune response and decreased cytokine storm [15,16]. In contrast, another study reported a 28% mortality with severe clinical manifestations [17]. There is consensus that: older age, diabetes, hypertension, cardiovascular disease, lung disease, a less efficient immune system, outpatient hemodialysis in overcrowded dialysis centers, together with the number of years on dialysis and increased inflammatory markers, lead to increased prevalence and mortality [18,19,20,21].

In Latin America, there is little information on the burden of COVID-19 in hospitalized patients with stage-5 chronic kidney disease on dialysis so this data would be important to optimize care strategies and clinical outcomes.

The study's main objective is to evaluate the factors related to the mortality of patients with stage-5 chronic kidney disease on dialysis with COVID-19 disease in a public hospital in Lima, Perú.

Methods

Design and population

This is a retrospective case series study made at the Hospital Nacional Arzobispo Loayza in Lima, Perú, between April and December 2020. This public hospital is a national referral center and belongs to the Peruvian Ministry of Health (MINSA). The study population was selected by non-probabilistic sampling, including the entire universe of patients hospitalized with stage-5 chronic kidney disease on renal replacement therapy and COVID-19 disease during the study period.

Inclusion criteria were hospitalized patients older than 18 years; having stage-5 chronic kidney disease on hemodialysis or chronic peritoneal dialysis; renal transplant patients with functioning kidneys; patients who started chronic hemodialysis or peritoneal dialysis during hospitalization; clinical diagnosis of COVID-19 disease made by their treating physician and/or diagnosis of COVID-19 in the discharge summary, including antigen testing or reverse transcriptase polymerase chain reaction (RT-PCR) test on a nasopharyngeal swab. Exclusion criteria were patients diagnosed with acute kidney injury, an episode of acute kidney injury on chronic kidney disease, and incomplete data in clinical history records. All were new cases of COVID-19; 107 patients diagnosed with stage-5 chronic kidney disease and COVID-19 disease were registered; 105 met the inclusion criteria. The primary data source was clinical records, which were recorded retrospectively, assigning sequential numbering.

Variables

The study included demographic, clinical, and laboratory variables; the dependent variable was vital status at discharge. Clinical and laboratory data were considered within the first 24 hours of hospital admission and processed in the laboratory. Patients did not receive any vaccine against COVID-19 since, during that period, it was not yet available in our country.

The variables were: age categorized ≤ 60 years and > 60 years, sex, body mass index (underweight, normal, overweight, and obese), hypertensive and non-hypertensive etiology of chronic kidney disease, smoking history, cause of hospitalization, dialysis admission modality (new: patient with stage-5 chronic kidney disease starting dialysis during hospitalization, and chronic: patient with stage-5 chronic kidney disease, receiving regular maintenance dialysis prior to hospitalization), the severity of COVID-19 disease: mild, moderate and severe [22], inflammatory parameters (hemoglobin, C-reactive protein, ferritin, D-dimer, Lactate dehydrogenase, lymphocyte, and leukocyte counts), blood gas values (pH, bicarbonate, partial oxygen pressure and arterial oxygen pressure / inspired oxygen fraction), urea, creatinine, hospital stay, renal replacement therapy modality: chronic hemodialysis or chronic peritoneal dialysis.

Additionally, laboratory variables were classified according to reference value: lymphocytes (> 1000 and ≤ 1000 cells per microliter) [16], leukocytes (≤ 11 000 and > 11 000 cells per milliliter) [16], platelets (>150 000 and ≤ 150 000 cells per microliter) [16], D - dimer (≤ 1.5 and > 1. 5 milligrams per liter) [14], ferritin (≤ 1000 and > 1000 milligrams per deciliter), C-reactive protein (≤ 10 and > 10 milligrams per deciliter) [14], lactate dehydrogenase (≤ 400 and > 400 international units per liter) [14], bicarbonate > 20 and ≤ 20 millimoles per liter) [13], potassium (≤ 5.5 and > 5.5 mmol/L) [13], alanine aminotransferase (< 40 and ≥ 40 U/L) [13], aspartate aminotransferase (< 40 and ≥ 40 U/L) [13], and arterial oxygen pressure/inspired oxygen fraction (> 150 and ≤ 150 millimeters of mercury).

Ethics

The Institutional Ethics and Research Committee of the Cayetano Heredia Peruvian University approved the study under the number 039-01-21. Since it is a retrospective study, it is exempt from informed consent.

Statistical analysis

Descriptive statistics: We used means and standard deviation to describe the numerical variables with normal distribution. Normality was assessed with the Shapiro-Wilks test. Categorical variables were described in proportions (%). Frequency tables are presented, including clinical, demographic, laboratory characteristics, and mortality.

Inferential statistics: Being an exploratory analysis, we used relative risk to assess the association between categorical variables and outcome (relative risk and 95% confidence intervals). For multivariate analysis, logistic regression models were used. In addition, the variables sex and age were also included in the model because they were considered potential confounders. Data were analyzed with Stata software version 17. For the analysis, p < 0.05 was considered statistically significant.

Results

In the study period, 105 patients diagnosed with stage-5 chronic kidney disease and COVID-19 disease were evaluated. Mean age was 58.6 years (standard deviation: 14.3), 57 patients (54.3%) were male, and the most frequent etiology of chronic kidney disease were arterial hypertension and diabetic nephropathy (Table 1).

Of the patients who started hemodialysis during hospitalization, 40 (38.1%) did so for the first time. The indications for dialysis were: pulmonary congestion, present in 46 (43.8%); hyperkalemia in 40 (38.1%); metabolic acidosis in 12 (11.4%) and uremic encephalopathy in 5 (4.8%); 104 (99.04%) received hemodialysis and only one (0. 96%) peritoneal dialysis. The initial vascular access used was temporary central venous catheter 51 (48.5%) followed by tunneled catheter 34 (32.4%) and 20 (19.1%) arteriovenous fistulas; 46 (43.8%) required oxygen mask and 39 (37.1%) nasal cannula, the mean body mass index was 24.3 kilogram per square meter (SD 3.87).

In the bivariate analysis, patients with moderate to severe COVID-19 disease had 1.10 (95% CI: 1.09 to 4.21) times the risk of dying than those with a mild presentation. Those who were overweight and obese had 1.51 (1.11 to 1.92) times the risk of dying than those with normal body mass index. Similarly, patients with lymphocytes ≤ 1000 cells per microliter, leukocytes > 11 000 cells per microliter, D-dimer >1.5 milligrams per liter, ferritin >1000 milligrams per deciliter, C-reactive protein >10 milligrams per deciliter, lactate dehydrogenase >400 international units per liter, bicarbonate < 20 millimoles per liter and arterial oxygen pressure/ fraction of inspired oxygen ≤ 150 millimeters of mercury, were at higher risk than those with normal values (tables 2 and 3).

In the multivariate analysis using logistic regression, the following outcomes were associated with mortality: severe clinical condition of COVID-19, C-protein reactive mg/dL > 10, and PaO₂/FiO₂ < 150 mmHg, had 1.48 (95% CI 2.24 to 7.77), 9.72 (1.41 to 8.58) and 10.23 ( 5.87 to 36.06) times the risk of dying compared to those who did not have this condition, adjusted for age and sex. (Table 4)

The mortality rate was 20%, and there was no difference between patients in the chronic hemodialysis program and those who started hemodialysis during hospitalization.

Discussion

In this study, an exploratory analysis was carried out to evaluate possible factors related to mortality in patients with stage-5 chronic kidney disease on dialysis with COVID-19, taking into account that during the pandemic, patients with stage-5 chronic kidney disease on renal replacement therapy were a vulnerable population, presenting with mild, moderate and severe infection. Among them, there was a higher need for hospitalization, intensive care unit admission, and mortality [4,23].

In the present study, most patients were male, and hypertension (97.1%) and diabetes mellitus (23.8%) were the most frequent comorbidities, similar to the general population [9,10,13]. The main cause of hospitalization was respiratory failure (76.2%), requiring oxygen support with oxygen masks, similar to other series [15,23,24,25,26]. Cai et al. reported that six (15.5%) of patients with chronic kidney disease were admitted to the intensive care unit [26], differing from the present study. No patient was admitted to critical care units, due to high demand and limited capacity to care for this type of patients [15,23,24,25,26]. Concerning body mass index, a direct association between obesity and increased risk of COVID-19 disease is known. In this study, the prevalence of overweight was 30.48% and obesity 7.1%, lower than that reported in other studies and unrelated to mortality. This is explained by the fact that in our population on chronic dialysis, obesity is less prevalent than in the general population [15,18,27,28,29].

The indications for dialysis were pulmonary congestion, hyperkalemia, and metabolic acidosis [15,23,24,25,26,27,28], due to irregular outpatient hemodialysis (shortened dialysis time, irregular dialysis attendance due to fear of infection, or lack of money).

A short-term mortality rate in patients with COVID-19 disease and on chronic dialysis has been documented to range between 20 and 30% [29,30,31]. A Peruvian study reported mortality rates of 16.3% in hospitalized patients [32]; in our investigation, mortality was 20%, similar to what has already been reported.

Sosa et al. [29] reported a decrease in survival during hospitalization for COVID-19 disease related to the disease severity and inflammatory markers. In our study, survival was 65.3% and 56% at days 20 and 30 of hospitalization, respectively, lower than the general population [31,33].

There is evidence that the factors associated with increased mortality are age, male gender, comorbidities such as hypertension, diabetes mellitus, obesity, inflammatory markers, and respiratory involvement [26,27,28,29,30,31,32,33]. In the present investigation, and a spanish study, no relationship was found between age, gender, and comorbidities [34].

Regarding laboratory markers, there are reports of the relationship between increased leukocytes, C-protein reactive, lactate dehydrogenase, ferritin, D-dimer, and decreased lymphocytes with mortality, related to an exaggerated inflammatory response in patients with COVID-19. Such findings are congruent with ours [31,32,33,34]. It is described that oxygen saturation < 90% persistent despite oxygen administration, oxygen partial pressure < 68 millimeters of mercury, and decreased arterial oxygen pressure/inspired oxygen fraction are associated with increased mortality in the context of severe respiratory compromise [35]. In this series, most patients were observed to require oxygen support and showed a significantly low arterial oxygen pressure/inspired oxygen fraction.

In our study, increased inflammatory markers and decreased oxygenation status indicators were related to increased mortality. Variables associated with lower survival were C-reactive protein values >10 milligrams per deciliter and arterial oxygen pressure/inspired oxygen fraction values ≤ 150 millimeters of mercury, similar to that described by Wang et al. [35], who mentioned that arterial oxygen pressures/inspired oxygen fraction < 136 millimeters of mercury were characteristic of patients with higher mortality. On the other hand, Turgutalp et al. [36] noted that an increase of more than ten times the normal value of C-reactive protein was associated with lower survival, the latter in patients hospitalized on maintenance hemodialysis. In addition, severe COVID-19 disease was associated with increased mortality in a study evaluating more than 300 patients undergoing chronic hemodialysis, as in the present report [37].

The study's main limitations are that it is a retrospective study and that the main source of data was hospital medical records, which may implicitly compromise the validity and reliability of the recorded data. Two different groups of patients are described: chronic and new, which could constitute two different populations, with different hospitalization admission conditions and mortality risk, despite the fact that the groups were comparable in all the variables evaluated, except for creatinine levels. This potential bias source could be handled with a larger sample size and a longer follow-up time. Data such as time on dialysis prior to hospital admission (of patients on chronic dialysis) and the number of hemodialysis sessions were unavailable.

The results obtained in the study provide insight into the presentation and clinical outcome of COVID-19 disease in patients with chronic kidney disease on dialysis. This information could be used to reduce the morbidity and mortality of these patients, even more so knowing that in Latin America, there is a progressive increase in patients with chronic kidney disease whose approach and maintenance are still deficient, generating high costs for the healthcare system. Additional studies should be carried out in our region, including a larger number of patients, distributed according to the dialysis modality, and evaluating the evolution of survivors after COVID-19 disease.

Conclusion

In patients with stage-5 chronic kidney disease hospitalized for COVID-19, severe COVID-19 disease, C-protein reactive levels > 10 milligrams per deciliter, arterial oxygen pressure/inspired oxygen fraction ≤ 150 millimeters of mercury and severe COVID-19 disease were the main factors associated with mortality.