key: cord-355549-6xnjj5h5
authors: Cécile, Couchoud; Florian, Bayer; Carole, Ayav; Clémence, Béchade; Philippe, Brunet; François, Chantrel; Luc, Frimat; Roula, Galland; Maryvonne, Hourmant; Emmanuelle, Laurain; Thierry, Lobbedez; Lucile, Mercadal; Olivier, Moranne
title: Low incidence of SARS-CoV-2, risk factors of mortality and the course of illness in the French national cohort of dialysis patients.
date: 2020-08-25
journal: Kidney Int
DOI: 10.1016/j.kint.2020.07.042
sha: 
doc_id: 355549
cord_uid: 6xnjj5h5

The aim of this study was to estimate the incidence of COVID-19 disease in the French national population of dialysis patients, their course of illness and to identify the risk factors associated with mortality. Our study included all patients on dialysis recorded in the French REIN Registry in April 2020. Clinical characteristics at last follow-up and the evolution of COVID-19 illness severity over time were recorded for diagnosed cases (either suspicious clinical symptoms, characteristic signs on the chest scan or a positive reverse transcription polymerase chain reaction) for SARS-CoV-2. A total of 1,621 infected patients were reported on the REIN registry from March 16th, 2020 to May 4th, 2020. Of these, 344 died. The prevalence of COVID-19 patients varied from less than 1% to 10% between regions. The probability of being a case was higher in males, patients with diabetes, those in need of assistance for transfer or treated at a self-care unit. Dialysis at home was associated with a lower probability of being infected as was being a smoker, a former smoker, having an active malignancy, or peripheral vascular disease. Mortality in diagnosed cases (21%) was associated with the same causes as in the general population. Higher age, hypoalbuminemia and the presence of an ischemic heart disease were statistically independently associated with a higher risk of death. Being treated at a selfcare unit was associated with a lower risk. Thus, our study showed a relatively low frequency of COVID-19 among dialysis patients contrary to what might have been assumed.

1-REIN registry, Agence de la biomédecine, Saint-Denis La Plaine, France 2-CHRU-Nancy, INSERM, CIC, Epidémiologie Clinique, Nancy, France 3-Nephrology department, Caen University Hospital, France 4-Nephrology department, APHM University Hospital, Marseille, France 5-Nephrology department, GHR Mulhouse Sud-Alsace, France 6-University of Lorraine, CHRU-Nancy, Vandoeuvre, France 7-Calydial, Vienne, France 8-Nephrology department, Nantes University Hospital, France 9-Nephrology department, AP-HP Pitié-Salpêtrière Hospital, Paris, France 10-Nephrology-Dialysis-Apheresis department, Nîmes University Hospital, France Introduction Due to their frequent contact with hospitals and their comorbid condition, dialysis patients are identified as high-risk patients for severe forms of infection from SARS-CoV-2. Guidelines to mitigating risks have been published (1) (2) (3) (4) (5) (6) (7) . However, few studies including case reports or the experience of centres have included sufficient numbers of patients to have a complete overview of their real risk and course of illness (8) (9) (10) (11) (12) (13) (14) (15) (16) . In those studies, case fatality varied from 14% to 31%.

On March 16 th , 2020, the French national End-Stage Kidney Disease REIN Registry began to record all patients on dialysis in France who were diagnosed with COVID-19.

The aim of this first report from the French REIN registry is to describe the population of infected dialysis patients and their course of illness, estimate the incidence and lethality of COVID-19 disease and identify the risk factors associated with the probability of death.

From March 16 th , 2020 to May 4 th , 2020, 1 621 patients were declared as being infected with SARS-CoV-2 on the REIN registry. This represents 3.3% of all 48 669 dialysis patients treated in 1245 dialysis units in metropolitan France and overseas territories.

The clinical and care situation at the first report in the registry was "hospitalized -moderate disease" for 48%, "mild disease treated at home" for 39%, "severe disease in an intensive care unit" for 5%, "death" for 2% and asymptomatic for 2% of cases. The first diagnosis was made in 73% of cases with a PCR on a nasopharyngeal swab, 17% on characteristic signs on the CT scan and 8% on suspicious clinical symptoms. Finally, a positive PCR was available for 1269 patients (79%).

In all, 38% were treated at home. Outpatients were younger (median age 68.7, IQR 56.7-80.4, vs 73.7, IQR 63.7-81.6), more often non-smokers and had less dysrhythmia and incapacity for transfer (Suppl Table1). Their mortality was lower (8.5%) compared to patient who were hospitalised (22.4%).

In all, 9% of patients were admitted to an ICU unit. Those patients were younger than the others (median age 67.2, IQR 58.3-74.5, vs 72.4, IQR 61.3-81.6), less often had cerebrovascular disease, had a higher BMI and were less often treated by hospital-based HD (Suppl Table2). Among the 87 patients for whom information was available, 51% received invasive mechanical ventilation (Suppl figure 1). The mortality of ICU patients was higher (34%) compared to patients who were not admitted to ICUs (15.5%).

The clinical situation at the last report in the registry for patients who were still alive, was "hospitalized -moderate disease" for 11%, "mild disease treated at home" for 16%, "in intensive J o u r n a l P r e -p r o o f care" for 2% and "recovered" for 67% and asymptomatic 4%, with a median follow-up of 19 days (IQR 6-28).

Not all parts of France were affected in the same way. The prevalence of COVID-19 patients varied from less than 1% in the 5 overseas territories and 8 metropolitan regions to over 5% in 3 northeastern regions (especially in Alsace, 10%, one of the first French clusters) and in the Île-de-France, 9%, the most densely-populated region ( Figure 1 ). These variations were not explained by age and were parallel to those of the general population (Figure 1 ). At that time, the percentage of infected persons in the French population was 0.2% and the mortality among confirmed cases was 19% (no systematic screening). The cumulative incidence of new cases after an exponential increase has now stabilized itself ( Figure 2 The clinical characteristics of infected dialysis and control populations are represented in Table 1 .

Compared to the 25 455 selected controls (treated in centres where at least one patient was infected), the probability of being a case was higher in males (OR 1. Among the infected patients, 344 died due to a cause related to SARS-CoV-2 after a median time of 6 days (IQR 3-13). The lethality in diagnosed cases was 21%. In the univariate analysis, higher age, being a former smoker, having a chronic respiratory disease, cardiovascular comorbidities ( e.g. peripheral vascular disease, ischemic heart disease, congestive heart failure or dysrhythmia) and frailty (hypoalbuminemia or inability to walk) were associated with a higher risk of death in SARS-CoV-2 infected dialysis patients. Dialysis in self-care units or out-centres or being a current smoker were associated with a lower risk of death. In fact, most of these clinical characteristics and care J o u r n a l P r e -p r o o f were associated with older age. In the multivariate model, only older age, hypoalbuminemia and the presence of an ischemic heart disease were statistically independently associated with a higher risk of death (Table 2 ). Being treated in a self-care unit was associated with a lower risk of death. Neither chronic respiratory disease, obesity, diabetes nor smoking status were associated with a higher risk of death. The sensitivity analysis including the region of treatment gave similar results.

The trajectory of care is represented in Figure 3 for the 287 deceased patients for whom at least 2 different clinical situations were reported in the registry. For severe cases hospitalized in intensive care units, the median time until death was 7 days (IQR 4-14), whereas the median time for hospitalized patients until death was 5 days (IQR 3-9) and, for patients at home, 6 days (IQR 3-11).

The trajectory of care is represented in Figure 4 for the 799 patients who recovered (clinical situation coded as recovery or asymptomatic). The median time in hospital until recovery was 15 days (IQR 10-21), similar to that for patients who were at home (16 days, IQR 11-21).

So far, more than 1600 dialysis patients have been diagnosed with SARS-CoV-2 infection in France.

Our study shows that the prevalence of SARS-CoV-2 infection in dialysis patients varied throughout the country from 0 to 10%. Mortality in this population of diagnosed cases is high at 21% and is mainly associated with a higher age (13% mortality in patients aged under 75 compared with 30% of patients aged over 75).

The trend of the SARS-CoV-2 epidemic in patients on dialysis shows a parallel development as in the general French population, with North Eastern regions and the Ile-de-France being more affected.

Our global prevalence is 3% of dialysis patients but this reaches 10% in the most affected regions. In the absence of other population-based data, it can only be compared with the 14% of the Haemodialysis Centre in Wuhan at the epicentre of the Chinese epidemic (12) . However, the nonsystematic detection of asymptomatic patients in France may lead to an underestimation of the true dissemination of SARS-CoV-2 in the French dialysis population.

Although the lockdown seemed to have significantly reduced the amount of contact among the general population, dialysis patients have to leave confinement to go to their dialysis units and, consequently, are still in contact with a large number of people. The risk of contamination may occur during transport, at the dialysis unit or during hospitalisation, but also at home with the family or caregivers. Home dialysis was associated with a lower probability of being infected suggesting a protective effect of staying at home. Dialysis centres affected later learned from units contaminated early on in the epidemic's progression and reorganized their patients' circulation and care (14,17).

Indeed, since the beginning of the epidemic, protective strategies have been broadcast by the SFNDT (Société Francophone de Nephrology Dialyse Transplantation) with weekly COVID-19 webinars inviting all French nephrologists to discuss the overall COVID-19 themes and topics available on the SFNDT website (https//www.SFNDT.org˃actualites) . Thanks to this collaboration, the worst may have been avoided. However, we must now remain vigilant and protect our healthcare workers.

The initial incidence of the disease in some dialysis units seemed very high, especially in the initial regions. The incidence in dialysis units is now decreasing, mirroring the decrease in the general population. This can also be associated with the implementation of all the necessary preventive actions prone by the SFNDT, including 1/wearing a mask during transport and for the entire period of care, 2/systematic tracking of patients and screening at the entrance to dialysis units based on fever and symptoms or contact with an infected person and 3/restricting areas for COVID-19 cases (18),.

As in the general population, male gender, diabetes and frailty, but not age, were associated with a higher risk of being infected. A selection bias, due to the fact that these patients may have a more severe form of the disease and are therefore more easily diagnosed, cannot be ruled out. As in the cohort of 627 haemodialysis patients at the Haemodialysis centre in Wuhan, diabetes was associated with a higher risk of infection. This result was still significant when introducing regions in the model to take into account the fact that the epidemic was mainly located in the North East of France where the prevalence of diabetes is higher. Smoking, even after taking comorbidities into account, was associated with a lower risk of infection, as discussed in the general population (19) . Surprisingly, being treated in a self-care unit was associated with a higher risk of being infected. At self-care units, care is provided without supervision by an on-site nephrologist (16) . The presence of a nurse is mandatory and patients are helped with the HD process. All these units collaborate with a hospitalbased dialysis unit. Moreover these units treat younger patients who may have had more contact at risk than elderly. despite the lockdown. These small units, with fewer caregivers on site, could have tarried in implementing protection strategies as proposed by others (1) . Further analyses are required to evaluate the impact of other risk factors, such as living in an institution or in a deprived neighbourhood area associated with overcrowded housing.

International comparison of case fatalities should be made with caution given the case-mix, the various healthcare arrangements and the different dynamics of the epidemic. Our mortality among diagnosed cases, 21% so far, is higher than the 13% reported for the dialysis center in Wuhan (12).

The older age and more frequent comorbidities of French dialysis patients may explain a higher mortality than in China (20, 21) . Furthermore, the non-systematic detection of asymptomatic patients favors more seriously ill patients. In France, case fatality was lower than the 29% reported in J o u r n a l P r e -p r o o f 4 outpatient dialysis facilities in Italy (14), the 30% in a single center in Madrid (15) , or 31% in a single center in New York (16) . Higher mortality in these studies may be explained by a selection bias for more severely ill hospitalized patients. Compared to the general population, the dialysis lethality observed in our cohort was similar to the 20% case fatality rate observed with patients aged over 80 in Italy (20) . It is also similar to the mortality rate for confirmed cases in the French general population, where at least 84% of the people who died had a comorbid condition and 92% were aged 65 or older.

Apart from age, which seems to be the major factor in the general population (22), nutritional status, indirectly assessed by albumin levels and the presence of ischemic heart disease, seem to be the main risk factors. Further in-depth analyses are planned in order to better estimate the excess mortality in dialysis patients at this period, taking into account the underlying mortality risk. Being treated in a self-care units was associated with lower mortality, even after taking into account age and comorbidities. After adjustment, home dialysis mortality did not differ from the mortality rate for hospital-based haemodialysis. However, the small number of patients with home dialysis has not allowed us to make an in-depth analysis so far. Other factors (such as living conditions, delay in alerting and other home-based care) which are not available in our registry, need to be explored.

Although incomplete, the illness trajectory seems to show rapid worsening and a slow healing process. The short lapse of time before death could corroborate the physiopathology with the delay in host inflammatory response phase reported 7 to 10 days after the initial infection (23). This rapid negative development raises the question of reinforced surveillance at home, during dialysis sessions and, why not, preventive hospitalization in a safe environment. Our definition of recovery should be taken with precaution since the definition of recovery is still under debate. Some patients were maintained in hospital under isolation for 15 days.

Very soon after the start of the epidemic, the French-speaking Society of Nephrology, helped by infectious disease specialists made recommendation that for each dialysis patient with fever, a viral syndrome, pulmonary symptoms or diarrhoea, a CT scan should be prescribed as well as a PCR on a nasopharyngeal swab. Contact subjects were also tested in the later period explaining the occurrence of few asymptomatic patients. These recommendations applied to all hospitalized patients and outpatients as well throughout the whole country. However, due to possible variations in diagnosis strategies, Day 1 for each patient may vary from one unit to another. Access to intensive care units was a concern for nephrologists in certain areas. Some tensions could be noted in highly affected regions but, in general, dialysis patients could be transferred to intensive care as required depending on their age and comorbidities.

The strength of this study is its national scale, including the whole population of French dialysis patients. However, these results must be interpreted bearing the following limitations in mind.

Various screening strategies may influence the detection of the disease. This could be the case especially for patients treated at home or asymptomatic patients or sudden death, but mild cases and hospitalized patients can be considered as being exhaustive. Non-systematic screening favours the collection of more severe cases and leads to an overestimation of lethality. The second limitation is the lack of granular data on clinical presentation, laboratory results and treatment and the precise protective strategy implemented in the units. Our study is based on a registry, which gives an exhaustive national overview but with a limited dataset -not on medical records, which could give more detailed data on treatment and clinical presentation but on a limited number of patients with a risk of selection bias. Third, due to the confinement of registry research assistants, the data quality control procedure was limited. Post hoc controls will be taking place to complete the data. Fourth, the total number of patients tested and not considered as COVID positive is unknown. As in the general population, the true lethality of COVID-19 in infected dialysis patients needs to be confirmed by a longer follow-up and deployment of screening methods.

Despite the difficulty to have a "true" estimation, this preliminary report of the French registry shows a relatively low frequency of COVID-19 among dialysis patients contrary to what might have been feared but, as in the general population, the epidemic did not evenly affect the whole territory.

Mortality in diagnosed cases (21%) is, associated with the same causes as in the general population, namely, high age, frailty and comorbid conditions.

The French REIN registry is intended to include all end-stage renal disease (ESRD) patients on renal replacement therapy (RRT) living in France, including overseas territories. Patients with a diagnosis of acute renal failure were excluded, i.e. those who recovered all or some renal function within 45 days or were considered by experts to have acute failure when they died before 45 days. The details of organizational principles and quality control are described elsewhere (24) . The REIN network includes nephrologists, nurses, patients, public health representatives and epidemiologists coordinated within regional and national steering committees. The national coordination center is based at the Agence de la Biomédecine, a public Health Agency that oversees the activity of organ and tissue procurement and transplantation. The 

Clinical characteristics at last follow-up included age, gender, comorbidities, mobility status (walks without help, needs assistance for transfers, or is totally dependent for transfers), body mass index (BMI), tobacco use, haemoglobin and serum albumin, dialysis technique (haemodialysis or peritoneal dialysis) and location (hospital-based, out-centre, self-care unit, home). This study analysed 10 comorbidities: diabetes, congestive heart failure, ischemic heart disease, peripheral vascular disease, aortic aneurysm, cerebrovascular disease, dysrhythmia, active malignancy, cirrhosis, and severe behavioural disorders (defined as including dementia, psychosis, or severe neurosis that may have affected the functional status or adherence to treatment).

The last residence and last dialysis unit before February 15, 2020 were taken into account to avoid misclassification of patients transferred to another dialysis centre due to their infection status. 

The clinical characteristics of patients were expressed as frequencies and percentages for qualitative variables and medians with interquartile ranges for quantitative variables.

The percentage of infected patients in the dialysis units of each region was adjusted on age (indirect standardization) to take into account the underlying age distribution of the dialysed patients. The crude ratio and the standardised ratio are presented on a map, according to the patients' area of residence. To give an overview of the epidemic in France, hospital mortality due to COVID-19 on April 2020, extracted from the platform of the national public health agency, Santé Publique France: https://geodes.santepubliquefrance.fr/#c=indicator was reported.

We also presented the cumulative number of infected patients on a day-to-day graph for the whole country.

To describe the characteristics of infected patients, we compared this population with two control groups. The first one included all the dialysis patients in France who were not infected. The second, to take into account the heterogeneity of the distribution of the epidemic in the country, included only patients treated in the dialysis units where at least one infected patient had been declared. Risk factors associated with being a case in those units were analysed by logistics regression with a stepwise selection of variables. The final model is based on complete data (no imputation). A p-value of <0.05 (two-sided) was considered statistically significant. Results are reported as odds-ratios (ORs) with their 95% confidence interval.

Lethality was estimated from the proportion of deceased patients among the diagnosed cases. To identify the risk factors associated with death in SARS-CoV-2 dialysis patients, a logistics regression with stepwise selection of variables was used. Interactions between age and other factors were explored. A p-value of <0.05 (two-sided) was considered statistically significant. Results are reported as odds-ratios (ORs) with their 95% confidence interval. Sensitivity analyses were made including the region of treatment, either as a fixed effect or with a random intercept.

Finally, when available, the course of illness was represented on a graph to describe the process of care for patients who died and for those who recovered. For each transition between the various care statuses, the number of patients and the median duration before transfer were calculated.

J o u r n a l P r e -p r o o f 

Recommendations for the prevention, mitigation and containment of the emerging

Mitigating Risk of COVID-19 in Dialysis Facilities

COVID-19 and the Inpatient Dialysis Unit: Managing Resources During Contingency Planning Pre-Crisis

Consensus recommendations for the care of children receiving chronic dialysis in association with the COVID-19 epidemic

COVID-19 from the nephrologist's point of view

Pandemic Uncertainty: Considerations for Nephrology Nurses

How we mitigate and contain COVID-19 outbreak in hemodialysis center (HD): lessons and experiences

A Case of Novel Coronavirus Disease 19 in a Chronic Hemodialysis Patient Presenting with Gastroenteritis and Developing Severe Pulmonary Disease

Coronavirus Disease 2019 (COVID-19) Pneumonia in a Hemodialysis Patient

First reported nosocomial outbreak of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) in a pediatric dialysis unit

Peritoneal Dialysis During the Coronavirus 2019 (COVID-19) Pandemic: Acute Inpatient and Maintenance Outpatient Experiences

An Analysis on the Clinical Features of MHD Patients with Coronavirus Disease 2019: A Single Center Study

Deisolation of COVID-19-positive hemodialysis patients in the outpatient setting: a singlecenter experience

Management Of Patients On Dialysis And With Kidney Transplant During SARS-COV-2 (COVID-19) Pandemic In Brescia

COVID-19: clinical course and outcomes of 36 hemodialysis patients in Spain

Presentation and Outcomes of Patients with ESKD and COVID-19

Lessons from the Experience in Wuhan to Reduce Risk of COVID-19

Infection in Patients Undergoing Long-Term Hemodialysis

Hemodialysis with Cohort Isolation to Prevent Secondary Transmission during a COVID-19 Outbreak in Korea

COVID-19 and smoking: A systematic review of the evidence

Spatial Analysis of Case-Mix and Dialysis Modality Associations

Case-Fatality Rate and Characteristics of Patients Dying in Relation to COVID-19 in Italy

OpenSAFELY: factors associated with COVID-19-related hospital death in the linked electronic health records of 17 million adult NHS patients. | medRxiv

Siddiqi HK, Mehra MR. COVID-19 illness in native and immunosuppressed states: A clinical-therapeutic staging proposal

The renal epidemiology and information network (REIN): a new registry for end-stage renal disease in France

COVID-19 Therapeutic Trial Synopsis

We gratefully acknowledge all participants of the REIN registry, nephrologists and research assistants alike, especially at this very particular time. The centres participating in the registry are listed in the REIN annual report: http://www.agence-biomedecine.fr/Le-programme-REIN.We also thank Teresa Sawyers, Medical Writer at Nîmes University Hospital for her help in editing the text.

The authors of this manuscript declare that they have no competing financial interest and no conflict of interests to disclose.

The registry is supported by the Agence de la biomedicine, France.