DNR orders in SARS-CoV-2 patients: a retrospective validation study in a Swiss COVID-19 Center

INTRODUCTION

During the beginning of 2020, the world witnessed the first COVID-19 pandemic¹. This dramatic situation forced many hospitals to improve their competence and in some countries such as in Switzerland some hospitals became dedicated centers to provide COVID-19 patients’ specific management, following governmental directions ².

In this pandemic scenario, the management of COVID-19 patients at high risk of morbidity and mortality raises new aspects of concerns, due to the risk of futilely investing precious resources such as CPR or ICU-level care³. As a consequence, Health-Care Professionals (HCP) had to find ways to identify those patients to whom the application of resources related to ICU admission resulted useless, especially when faced with limited hospitalization capacity and with the potential risk of shifting resources and time away from patients who could benefit from them. The Do-Not-Resuscitate (DNR) order is applied to those patients for whom further medical treatment is considered to be futile, with a very high probability of death or ventilator dependency despite the use of additional therapies^4,5. A careful management of resources was also applied during the H1N1 influenza pandemic, in 2009, when the UK Resuscitation Council suggested an early identification of DNR patients, in order to avoid futile resource wasting⁶ and thus allowing a reduction in ICU hospitalizations^7,8.

The legal status of DNRs varies from country to country⁵, the Switzerland experience is connected to the role of the Swiss Academy of Medical Sciences (SAMW/ASSM), a bridge builder between science and society, which implemented and applied a national standardized protocol of care based on expert opinion, with the aim to guarantee a better treatment of COVID-19 patients in an emergency situation with limited resources ^9,10. The protocol determined the indications for a correct evaluation and identification of patients suitable for a DNR order. This protocol supported HCP decision at triage and at hospital admission, with the aim to identify patients at high risk of death despite a potential ICU support, thus avoiding futile treatments, ICU overload and additional health costs ^3,9. Nevertheless, until today, the practical consequences of the SAMW/ASSM guidelines are still unknown.

According to other authors, DNR order resulted as an independent risk factor for death in the post-operative period, with DNR patients presenting higher fatality rate¹¹, a more rapid disease progression and a more frequent need of an early Palliative Care (PC) ^12,13, especially to improve the treatment of dyspnea, restlessness, anxiety and delirium ¹⁴. The early identification of DNR patients requiring PC treatment in the short-term during hospital stay is therefore mandatory for a better patients’ management and also from an ethical point of view, this last point acquiring even more importance in the current pandemic scenario.

Aim of this study was to report and analyze the survival rate of DNR COVID-19 patients hospitalized in a single Swiss COVID-19 center, identified and managed as DNR according to the SAMW/ASSM guidelines, to retrospectively validated the effectiveness of the Swiss national standard protocol in an emergency pandemics situation characterized by scarce availability of resources.

METHODS

This was a pilot retrospective validation study, evaluating consecutive DNR patients hospitalized with a diagnosis of COVID-19 during the first two weeks of the first pandemic wave. DNR order was defined according to the first version of SAMW/ASSM national guidelines (Table 1) ⁹, defining patients ineligible for ICU admission when they presented, between others, severe cerebral deficits after a stroke, NYHA class III or IV heart failure, COPD GOLD IV, liver cirrhosis with refractory ascites or encephalopathy at stage > I, stage V chronic kidney disease, a Frailty index higher than 4, age greater than 80 years and/or an estimated survival of less than 24 months ⁹.

RESULTS

During the first two weeks of pandemics, from March 16 to April 1 2020, 242 consecutive patients were admitted to the Hospital; 29 patients were excluded, as they had already been directly admitted as critically ill patients to the ICU. Of 213 patients not-critically ill at hospital admission, 40 (18.8%) were identified by HCP as DNR patients according to SAMW/ASSM national guidelines, thus resulting ineligible for ICU admission (DNR group); 173 (81.2%) patients were instead admitted to the internal medicine Department without a DNR order (control group, Figure 1). Among the DNR group, 27 (67.5%) patients presented a Frailty Index greater than 4, 21 (52.5%) had at least one chronic disease, 7 (17.5%) had a diagnosis of severe circulatory failure and 6 (15%) were affected by oncological active or metastatic disease. Among the DNR group, mean age was 79.9 years (62 – 97, SD 7.9) while among the control group it was 63.2 years (24 – 86, SD 12.1, p = 0.011); male patients were 25 (64%) in the DNR group, and 103 (59.5%) in the control group (Table 2).

• Download figure
• Open in new tab

Figure 1

CLM COVID-19 patients’ stratification during first two weeks of pandemics, according to DNR order identified by SAMW/ASSM guidelines.

Primary Outcome

At 30-days follow-up, 9 (22.5%) patients in the DNR group had died due to cardiovascular arrest (Table 3); mortality rate in the control group was 2.2% with 4 patients deceased (Figure 1 and 2). Concerning the deceased patients in the DNR group, mean age was 82.4 years (72 – 93, SD 6.9), all of them lived at home before the hospitalization, 7 (78%) presented a Frailty index higher than 4, 3 (33%) a pre-existing severe heart disease, 2 (22%) a metastatic oncological disease and 1 (11%) a chronic brain damage (Figure 1). In order to perform a preliminary guidelines’ validation ⁹, a contingency table was structured according to the outcome stratification and DNR group allocation. A Chi-square test comparing mortality in the two groups, revealed a higher mortality rate in the DNR group, p < 0.0001 (Table 4SM); similarly, the Kaplan-Meier distribution reported a significant difference over time (Log-Rank Mantel-Cox 23.104, p < 0.0001, Figure 2).

View this table:

• View inline
• View popup
• Download powerpoint

TABLE 3. Demographic of deceased patients

• Download figure
• Open in new tab

Figure 2

Patients’ survival stratified into DNR/control group, according to standardized national guidelines. As retrospective study, censoring was performed at April 30^th.

Secondary Outcomes

Regarding the deceased patients in the DNR group, average hospital length-of-stay (LOS) was 6.2 days (3 – 11, SD 2.4); for 8 (89%) of them it was possible to meet a family member before the end of life. All 9 patients (100%) received oxygen therapy until death, 4 (44%) of them had a continuous intravenous fluids infusion and 8 (89%) received at least one antibiotics or anti-retroviral therapy during the entire hospitalization (Table 3); all of them (100%) received opioids (89% morphine, 11% fentanyl), 8 (89%) received sedatives as needed, especially midazolam, while no neuroleptics were administered. Treatment implementation resulted in a significant improvement of RASS-PAL, from a median of 2.2 (SD 1.2) to -1.8 (SD 2), p < 0.0001 (Figure 3), and of the modified Borg scale, from a median of 5.7 (SD 3.3) to 4.7 (SD 3.7), p = 0.581 (Figure 1SM).

• Download figure
• Open in new tab

Figure 3

RASS-PAL scale has been used in COVID-19 deceased patients’ agitation assessment. An assessment before and after treatment was performed. After specialistic treatment, RASS-PAL significantly improved from 2.2 to -1.8 (p < 0.0001). Crosses represent the average of the sample.

DISCUSSION

COVID-19 is a disease characterized by important clinical and social impacts, with a mortality rate that is highest in patients over 60 years ¹². In this pandemic setting characterized by limited resources, accurately identifying patients at high risk of short-term death (30 days) represents a key point in both patients’ triage and in-hospital management. Application of the Swiss national standardized DNR protocol allowed the identification of COVID-19 patients with a significantly increased short-term mortality rate in comparison to non-DNR COVID-19 patients ^20,21. This evidence suggests that these guidelines, based on expert opinion and not previously validated, pragmatically identify patients at higher risk of short-term death, thus selecting a population requiring a different specialist approach, rather than intensive care. Unfortunately, DNR definitions change radically between countries, making it difficult to compare these groups of patients. Nevertheless, a national use of shared guidelines made it possible to identify a proportion of patients at high risk of short-term clinical deterioration.

Even though the lack of ICU admission can be considered a bias for the increased mortality in the DNR group, these guidelines are able to identify those patients presenting a more rapid deterioration and requiring a more intense clinical support in the short term. A study assessing the DNR group patients outcomes in the context of ICU admission could not be performed for obvious ethical and health-costs related reasons, but it can be deduced that the higher rate of clinical worsening seen in DNR patients might correlate with a lack of benefit from ICU admission.

The social impact of COVID-19 led the Public Authority to limit family members’ access to hospitalized patients, in order to preserve Public Health ²². In this context, HCP adequate identification of patients at high risk of death allows to predict, and thus to properly organize, a visit permit for patients at the end-of-life. This aspect is essential for Public Health, but also for the long-term social benefits for patients’ relatives, decreasing the risks of post-traumatic-stress-disorder, of unprocessed mourning and of additional psychological diseases²³. Moreover, it may prevent future increases in health costs, due to the better management of these important psychological and ethical aspects.

Palliative treatment played an important role in taking care of COVID-19 DNR patients, both in terms of psychological and physical symptomatic relief ^22,24,25. The effectiveness shown in alleviating dyspnea and agitation, even in this relatively small group of DNR patients, underlines the importance of PC in this scenario. The results we obtained with sedation, besides being significant from a statistical point of view, showed us how the clinical handling of these patients has been satisfactory in guarantying them adequate comfort, even in their last moments of life. In order to achieve this important result, and to adequately adapt therapy and management in relation to the clinical status, we realized that it was mandatory to evaluate patient’s agitation/sedation level through the use of appropriate international scales (RASS-PAL in our case) ^18,26. The assessments carried out so far, in accordance with evidences from other groups ²⁴, highlighted the importance of the PC specialist’s role in the better management of symptoms such as agitation, dyspnea, thirst, fear and solitude, which are unfortunately very frequent in the dying population ²⁵.

Our study is burdened by some limitations. Firstly, it was a retrospective collection of a cohort of patients, with a small sample of DNR patients; however, the significance level of the statistical tests remains high and should induce other groups to carry out future prospective validations in this regard. The group of DNR patients presented a median age greater than the control group; this may reflect the presence of co-morbidities, a key-elements necessary to identify DNR patients according to national guidelines; however, the increased mortality rate in DNR group was at 30 days, therefore the presence of a median age of 80 years alone was not sufficient to explain this increase in short-term mortality. In the analysis of palliative care treatment we decided to evaluate only the administration of opiates and sedatives and their effects, as a more specific information in this context resulted incomplete due to the retrospective design of the study. Finally, we did not investigate the follow-up in terms of hospital discharge, transfer to other facilities or survival after 30 days, mainly because the primary outcome of our study was to investigate the 30-day overall survival in COVID-19 patients, comparing DNR to not DNR groups.

CONCLUSIONS

The high mortality rate in COVID-19 patients enhances how it is relevant to identify the subpopulation at a higher risk of short-term death in a pandemic situation characterized by limited resources; in this scenario, the validation analysis we performed supported the role of the Swiss national standardized protocol in DNR patients’ identification. An early access to specialistic treatment should be mandatory, especially for this subgroup of patients, with the aim to improve the end-of-care management and at the same time to reduce the social impact of this pandemics.