Prevalence and Outcomes of COVID-19 Infection among Hemodialysis Patients in Nephrology-accredited Training Institutions in the Philippines

1. INTRODUCTION

COVID-19 had a significant impact on the hemodialysis (HD) population. It has led to outbreaks, as seen in two dialysis centers in Lombardy, Italy [1]. Other aspects of management among these patients were greatly affected. Due to infection control policies, most elective surgeries were advised to be postponed, leading to delays in the creation of arteriovenous fistula [2]. Due to concerns of immune suppression from medications, there was also a delay in performing kidney transplants. In the United Kingdom, about 1,670 kidney transplant opportunities were lost, which led to 6,317 active patients on the kidney-alone waiting list instead of 4,649 individuals [3].

Studies documenting the incidence of COVID-19 in dialysis patients are limited. In the case of the hemodialysis population, articles published in the past were case reports [4] or case series [5]. But with the steady rise of cases, more studies were presented. In one study done at the dialysis center of Renmin Hospital in Wuhan, China, where the first outbreak of the disease was seen [6], forty-two of 230 maintenance hemodialysis patients (18.26%) were diagnosed with COVID-19 infection during the study period. The duration of the study was done during the start of the outbreak (January 14, 2020) up to the control of the epidemic on March 12 that same year. Fifteen HD patients (6.52%), including 10 COVID-19-diagnosed died.

Another study documented COVID-19 cases in a dialysis center at a University Hospital in Madrid, Spain [7]. Among the 90 patients in their unit, 37 (41.1%) were diagnosed with COVID-19 infection, 36 of whom were confirmed by a positive PCR test. About 60% of patients were symptomatic, with 16 of the cases needing hospital admission and 6 resulting in death. From this data, they have modified their policies and organized a dialysis area for COVID-19 cases with protocols for infection control and management.

A study done in an in-patient dialysis center in a hospital in Paris, France, saw forty-four COVID-19 patients who were on dialysis. The majority were hemodialysis patients, and only five patients were on peritoneal dialysis (PD). Three of these patients were switched to hemodialysis during their course in the hospital [8].

In a multicenter study involving four large medical centers in Wuhan, China, among eight hundred eighteen patients on PD, eight were found to be positive for COVID-19. They have similar symptoms, radiographic changes, and lab findings to the other COVID-19 cases in general. Six recovered, and two of these patients died [9].

Regarding prevalence studies, there are two large studies. In a retrospective study by a national dialysis provider in the United States, 438 of 7948 (5.5%) maintenance dialysis patients developed COVID-19. Factors associated with developing COVID-19 infection were male sex, African American race, in-center dialysis (vs home dialysis), treatment at an urban clinic, residence in a congregate setting, and greater comorbidity. Of these maintenance dialysis patients with COVID-19, 109 (24.9%) died. Older age, heart disease, and markers of frailty were associated with mortality [10]. Meanwhile, in Germany, a study that involved 238 dialysis centers, the prevalence was 14% and the mortality for COVID-19 dialysis patients was 30% [11].

The European Renal Association - European Dialysis and Transplant Association has developed ERACODA (ERA-EDTA COVID-19 Database for patients on dialysis or living with a kidney transplant) that also collects individual data of dialysis and kidney transplant patients who were infected with COVID-19 [12]. Upon examination of its data collection form, certain details were not gathered, such as medications given, specific respiratory modality if required, and adjustment of dialysis dose. Our database aims to collect local, comprehensive data to understand the epidemiology, clinical course, and outcomes of COVID-19 infection among our hemodialysis and peritoneal dialysis patients. This will guide us in making recommendations as to the diagnosis and management of dialysis patients to improve survival for dialysis patients. It will lead to better outcomes and avoid unnecessary burdens to family members of dialysis patients, and minimize transmission of infection to other people. The study’s main objective is to describe the prevalence, clinical profile, and clinical outcomes (i.e., hospital stay, mortality) of hemodialysis patients infected with COVID-19.

2. METHODOLOGY

3. RESULTS

An amount of 785 unique hemodialysis patients had COVID infection during the study period. Considering the hemodialysis population of regions covered by these institutions, the prevalence of COVID is 5.1% among hemodialysis patients. Among institutions, the Medical City reported the greatest number of cases (112), followed by Southern Philippines Medical Center, Chong Hua Hospital, St. Luke’s Medical Center- Quezon City, and Veterans Memorial Medical Center (Table 1).

In the study population, 171 patients died while admitted, accounting for 22% of mortality. There is no significant difference between the survivors and non-survivors based on demographic data (Table 2). In terms of co-morbidities, more than three-quarters of the subjects were hypertensive, with a significantly higher proportion among non-survivors. Almost three-fifths of the subjects were diabetic, and almost a third with dyslipidemia, but there was no statistical difference between the two groups. In terms of maintenance medications, more deceased patients also used diuretics and insulin, while more survivors had a higher proportion of taking calcium carbonate and sevelamer. A higher percentage of survivors also had supplementary epoetin, with most of them using epoetin alfa. The average epoetin dose among users was not statistically significant.

Also, in Table 1, results show a large proportion of patients in survivor and mortality groups have hemodialysis sessions three times a week. Most of them use a high-flux dialyzer, but there was no statistical difference among the groups. Among the initial symptoms, cough is the most common symptom, experienced by two-thirds of the subjects, followed by dyspnea and fever. On analysis, only dyspnea was significantly higher among mortalities than survivors.

Based on the initial laboratory determinations (Table 3), those who died surprisingly had a higher baseline hemoglobin/ hematocrit and neutrophils, with lower eosinophils, lymphocyte, and platelet counts in Complete Blood Count (CBC). In terms of electrolytes, non-survivors had a higher baseline serum potassium and lower serum chloride and calcium. Arterial blood analysis showed lower pH, pO2, PaO2/FiO2 ratio, and serum bicarbonate values among mortalities. As expected, those who died also had higher baseline inflammatory markers. The rest of the laboratory parameters were not statistically significant.

In terms of clinical course (Table 4), there was a significantly higher proportion of patients in the mortality group who developed cytokine storm, sepsis, and septic shock. Overall, a quarter of all patients needed ICU admission, with most of the patients in the mortality group. However, ICU days between survivors and non-survivors were not statistically significant. The rest of the patients were admitted to the COVID area and home quarantine. 43% of patients received remdesivir, with a significantly higher proportion in the mortality group. Similar trends were noted in dexamethasone, tocilizumab, and hydrocortisone. 43% of patients also received anti-coagulation for prophylaxis, but still with a significantly higher proportion in the mortality group. There is also a higher mortality in those given anti-coagulation for extracorporeal treatment.

Almost 80% did not require inotropes, and those who required such intervention also had a significantly higher proportion in the mortality group. There is also less mortality among those who do not require or need minimal supplementary oxygen. In terms of hemodialysis prescription, more than two-thirds of patients did not require modification. For those who shifted to SLED, CVVHDF, or increased frequency, the percentage was significantly higher for the mortality. Almost 13% of patients also had hemoperfusion, and while the proportion favored the survivor group, a significantly higher percentage of patients in the mortality group underwent the procedure. A small percentage (2.92%) had convalescent plasma therapy. In terms of mechanical ventilatory days and length of hospital stay, there is no statistical difference between the two groups.

A summary of all significant factors is enumerated in Table 5 using univariate analysis. On multivariate analysis (Table 6), only five factors were found to be significant independent of other factors. These are the presence of cytokine storm, sepsis, use of heparin for extracorporeal therapy, use of tocilizumab, and elevated D-dimer values.

4. DISCUSSION

In our study, the prevalence of COVID among hemodialysis patients in PSN-accredited training institutions is 5.1%. Variations exist depending on the country, area, or center/s covered. The prevalence in the study is slightly higher than the initial study in Wuhan, China, of 2.2% [15]. However, this was lower in another study in a hemodialysis center in Wuhan, which was 11% [15]. The highest prevalence of COVID among hemodialysis patients was in two centers in the United Kingdom, with a combined 22.2% prevalence of asymptomatic patients using serologic screening [16].

Locally, there are two major studies published before this. The first one was a retrospective, observational study of 68 COVID-positive hemodialysis patients in the University of the Philippines-Philippine General Hospital (PGH) [17]. Prevalence was not determined, and the study period lasted for 4 months during the start of the pandemic. The second was performed in the National Kidney and Transplant Institute (NKTI), involving not just hemodialysis patients but also peritoneal dialysis patients and those initiated on renal replacement therapy. 68% of the subjects were hemodialysis patients infected with COVID [18].

This study was performed in more centers than in the previous studies mentioned. Results showed that the mean age of subjects is more than 60 years old, the majority were males, with diabetes as the primary cause of CKD, on hemodialysis for an average of 2 years, and hypertension as the most common and significant co-morbidity, followed by diabetes and dyslipidemia. The patients in the PGH were younger average-wise and hypertension was the most common cause of CKD. Study characteristics in this study are almost similar to the NKTI study, albeit a younger age in the latter and lesser duration of dialysis (a mean of 1 year). Cough was the most common symptom in this study, followed by dyspnea and fever. However, dyspnea was the only significantly higher percentage among those who died compared to survivors. The studies from PGH and NKTI have dyspnea as its most common symptom.

In terms of laboratories, non-survivors surprisingly have higher hemoglobin, hematocrit, and neutrophil values, with lower lymphocyte and platelet counts. This is different from the PGH study, in which the only significant hematologic parameter was a higher WBC among mortalities. In terms of electrolytes, the mortalities also had significantly higher serum potassium levels despite no recorded hyperkalemia (Serum K >6 mmol/l) and lower calcium and chloride. Hypochloremia and hypocalcemia were found to be associated with poor outcomes in COVID-19 hospitalized patients in other studies [19, 20]. This is likely since these imbalances can worsen the effects of the virus on the body. For example, hypocalcemia can weaken the heart and make it more difficult for the body to fight off infection. Hypochloremia can also worsen respiratory symptoms, as chloride is important for the function of the lungs [20].

Among the biomarkers, non-survivors also had significantly higher LDH and procalcitonin levels. These trends were also noted in the same studies. High LDH levels indicate increased cell damage and damage to muscles [21]. Procalcitonin levels are also significantly higher among mortalities. This protein, produced primarily by the thyroid gland and white blood cells, is also an inflammatory marker. It can indicate direct damage to the lungs because of the infection [21]. Unique to this study is that other markers were also elevated in mortalities. D-dimer is significantly higher, which is a protein fragment produced when fibrin clots dissolve. It is elevated in conditions that cause blood clots, like pulmonary embolism and deep venous thrombosis. COVID-19 is a pro-thrombotic disease, and this could mean that thrombi formation and possibly a contributory cause of death among these patients despite the use of anti-coagulation. C-Reactive Protein, Ferritin, and ESR were also elevated, indicative of a high inflammatory state of COVID patients with unfavorable outcomes.

There are still variations in terms of the management of these patients. Since the two local studies were performed early in the pandemic, Azithromycin and Hydroxychloroquine were the most common medications given in the PGH study, while Tocilizumab and Hydroxychloroquine were given in the NKTI study. For this study, almost half of the patients were given Dexamethasone, followed by Remdesivir. A significantly higher proportion though, was noted among those who died, which reflects that most of these patients have severe presentations that warranted the use of these medications. There was a small percentage of patients given Tocilizumab and other steroids like Hydrocortisone. This study also determined the use of inotropes among patients. While most patients did not require inotropes, those who required at least one most likely have an unfavorable outcome.

An advantage of this study was that it classified patients with unfavorable complications. About a third of all patients developed cytokine storm and sepsis, while more than 18% developed septic shock. A significantly higher proportion of these patients were in the mortality group, consistent with the premise that disease severity portends poorer outcomes. Also, 17% of patients required mechanical ventilation, as in the PGH study, and this is a significant factor in mortality.

Another unique outcome of the study was it determine if there were modifications with the hemodialysis prescription or if other extracorporeal therapies were performed. The majority of the patients did not require any modification to their usual treatment. For those with increased frequency, who used SLED and even CRRT, a significantly higher percentage of patients were in the mortality group, reflecting the severe effect of the infection. Hemoperfusion was performed in 13% of patients, slightly more than in the PGH and NKTI studies, and more patients who underwent convalescent plasma therapy.

In terms of multivariate analysis, five factors were found to be independently associated with mortality. Presence of sepsis and cytokine storm is correlated with mortality among COVID-19 dialysis patients [22, 23]. Cytokine storm refers to an excessive and uncontrolled release of pro-inflammatory cytokines that leads to severe inflammation, organ damage, and poor outcomes. It also leads to sepsis, causing organ dysfunction. Hence, therapies targeting pathways of minimizing cytokine production were investigated in these patients [23]. One of these is Tocilizumab, which is a humanized monoclonal antibody that binds to the interleukin-6 (IL-6) receptor. IL-6 is a cytokine that plays a role in inflammation and immune responses. Tocilizumab blocks the binding of IL-6 to its receptor, which prevents IL-6 from activating its downstream signaling pathways. This leads to a reduction in inflammation and immune responses [24]. However, there are still conflicting results as to its efficacy. In a study conducted in nine hospitals in Brazil, a randomized controlled trial was performed by giving Tocilizumab to some patients. Adverse events and need for mechanical ventilation were not different from standard care alone, and there was even a higher percentage of death in the treatment group, albeit not statistically significant [25]. On the other hand, a study by Eskazan and colleagues showed that mortality was lower among patients given tocilizumab [26]. Further investigations must be performed to establish its efficacy and also look at other factors such as timing of administration and patient sub-groups.

Use of an anti-coagulant for extracorporeal therapy was another independent factor. This could be related to the fifth factor, which is elevated D-dimer. Studies have highlighted that anti-coagulant use in COVID-19 patients is due to the prothrombotic state induced by the virus, causing endothelial dysfunction [27]. As mentioned, an elevated D-dimer is related to thrombotic episodes. Use of this drug has been associated with improved survival in critically ill COVID-19 patients by either preventing or lysing any blood clots present [28]. However, it is possible that timing of administration, heavy thrombi load, and interaction with other medications can affect the efficacy of anti-coagulants among the mortalities. These independent factors were different from the PGH study, where patients who needed mechanical ventilators, elevated procalcitonin, and a low PaO2/ FiO2 ratio were noted.

The initial plan of this study was also to analyze COVID infections among peritoneal dialysis patients. However, we only collected four patients, which is far below the minimum sample size. A possible recommendation is to continue data collection until the required number of subjects is reached. However, the data for hemodialysis is one of the largest numbers of subjects involved. The only two published studies with larger samples to date were the ERACODA study, which was an initiative of the European Renal Association-European Dialysis Transplant Association with 1141 subjects [29], and the study investigating dialysis attributes and strategies with COVID cases in London, with about 990 confirmed cases [30].

It could be possible that sub-group analysis can be performed, and the investigators will investigate this possibility. Another recommendation would be randomized controlled trials on different management aspects, from various drugs to the use of hemoperfusion or convalescent plasma. Since it can be assumed that the management of these patients is better due to the monitoring of both Nephrologists, Nephrologists-in-Training, and other Specialists, looking at clinical characteristics and outcomes of COVID-19 hemodialysis patients in hospitals without Nephrology Fellowship Training can be investigated.

CONCLUSION

This study has demonstrated one of the more susceptible groups of individuals that was affected by COVID-19, affecting more than 5% of hemodialysis patients. Mostly, it has similar clinical characteristics, management, and outcomes to other studies of this kind. Hemodialysis patients with COVID-19 infection presenting with sepsis, cytokine storm, and thrombotic episodes should be considered as high risk for unfavorable outcomes, and timely interventions must be done to minimize death in these individuals.

REFERENCES