Below the x-axis, the number at risk for the individual time points with number of deaths (in parentheses) are reported. ICU indicates intensive care unit; OHCA, out-of-hospital cardiac arrest.
eMethods 1. Search Strategy
eMethods 2. Aggregate-Data Meta-analysis
eFigure 1. PRISMA Flow Chart of Relevant Study Identification
eFigure 2. Forest Plot Summarizing Overall Survival Rates After Hospital Discharge Stratified by Inclusion in Kaplan-Meier (KM) Based Meta-Analysis
eFigure 3. Forest Plot Summarizing Risk Ratio for Overall Mortality After Hospital Discharge of Patients With Shockable vs Non-Shockable Initial Cardiac-Arrest Rhythm Stratified by Inclusion in Kaplan-Meier (KM) Based Meta-Analysis
eFigure 4. Forest Plot Summarizing Survival Rates of Survivors to Hospital Discharge at 3-Years of Follow-up
eFigure 5. Forest Plot Summarizing Survival Rates of Survivors to Hospital Discharge at 5-Years of Follow-up
eFigure 6. Forest Plot Summarizing Survival Rates of Survivors to Hospital Discharge a 10-Years of Follow-up
eFigure 7. Kaplan-Meier Survival Curve of Survivors to Hospital Discharge/30 Days After Exclusion of the Study From Andrew et al (Sensitivity Analysis)
eFigure 8. Forest Plot Summarizing Overall Survival Rates After Hospital Discharge Stratified by Follow-up Category After Exclusion of the Study by Andrew et al (Sensitivity Analysis).
eFigure 9. Forest Plot Summarizing Overall Survival Rates After Hospital Discharge Stratified by Follow-up Category After Exclusion of the Study by Sanghavi et al (Sensitivity Analysis)
eTable 1. Detailed Summary of Included Studies
eTable 2. Risk of Bias Assessment
eResults. Supplemental Results
Customize your JAMA Network experience by selecting one or more topics from the list below.
Amacher SA, Bohren C, Blatter R, et al. Long-term Survival After Out-of-Hospital Cardiac Arrest: A Systematic Review and Meta-analysis. JAMA Cardiol. 2022;7(6):633–643. doi:10.1001/jamacardio.2022.0795
What is the long-term survival rate of patients who survive the initial hospital stay for an out-of-hospital cardiac arrest?
In this meta-analysis of 21 studies (11 800 patients) using a Kaplan-Meier–based approach and 33 studies (16 933 patients) using a classic aggregate data approach, the 10-year survival rates in patients with out-of-hospital cardiac arrest who survived the initial hospital stay were between 62% and 64%.
These findings suggest that additional research is needed to improve the long-term survival of patients with out-of-hospital cardiac arrest.
Data on long-term survival beyond 12 months after out-of-hospital cardiac arrest (OHCA) of a presumed cardiac cause are scarce.
To investigate the long-term survival of adult patients after surviving the initial hospital stay for an OHCA.
A systematic search of the EMBASE and MEDLINE databases was performed from database inception to March 25, 2021.
Clinical studies reporting long-term survival after OHCA were selected based on predefined inclusion and exclusion criteria according to a preregistered study protocol.
Data Extraction and Synthesis
Patient data were reconstructed from Kaplan-Meier curves using an iterative algorithm and then pooled to generate survival curves. As a separate analysis, an aggregate data meta-analysis was performed.
Main Outcomes and Measures
The primary outcome was long-term survival (>12 months) after OHCA for patients surviving to hospital discharge or 30 days after OHCA.
The search identified 15 347 reports, of which 21 studies (11 800 patients) were included in the Kaplan-Meier–based meta-analysis and 33 studies (16 933 patients) in an aggregate data meta-analysis. In the Kaplan-Meier–based analysis, the median survival time for patients surviving to hospital discharge was 5.0 years (IQR, 2.3-7.9 years). The estimated survival rates were 82.8% (95% CI, 81.9%-83.7%) at 3 years, 77.0% (95% CI, 75.9%-78.0%) at 5 years, 63.9% (95% CI, 62.3%-65.4%) at 10 years, and 57.5% (95% CI, 54.8%-60.1%) at 15 years. Compared with patients with a nonshockable initial rhythm, patients with a shockable rhythm had a lower risk of long-term mortality (hazard ratio, 0.30; 95% CI, 0.23-0.39; P < .001). Different analyses, including an aggregate data meta-analysis, confirmed these results.
Conclusions and Relevance
In this comprehensive systematic review and meta-analysis, long-term survival after 10 years in patients surviving the initial hospital stay after OHCA was between 62% and 64%. Additional research is needed to understand and improve the long-term survival in this vulnerable patient population.
Out-of-hospital cardiac arrest (OHCA) remains a leading cause of death, despite important advances in prehospital and in-hospital care during the past few decades.1-3 Nearly 90% of patients with OHCA in the US do not survive until discharge from the index hospital stay.4 For patients admitted to the hospital after OHCA, the survival rate to hospital discharge is approximately 50% to 60%.1,5 Most patients experience an OHCA because of underlying medical conditions and only a few after trauma or surgery.3,6 In Western countries, several key factors that directly influence early survival have been identified and mainly lie within the initial rescue process,7,8 including early high-quality cardiopulmonary resuscitation,9,10 the initial rhythm of cardiac arrest,11 and early defibrillation.10,12,13
In addition to efforts to improve initial survival after OHCA, the interest in better understanding and improving longer-term outcomes in this population is increasing, as outlined in the recently updated European postresuscitation care guidelines.6 Based on a recent analysis, the mean global survival rate of patients with OHCA is 7.7% at 1 year after hospital discharge with variations worldwide.1,14 Of importance, previous studies15-19 have concluded that patients who survive the initial OHCA often also have a favorable neurologic outcome. Several prognostic factors for long-term survival have been identified, including younger age,20 receipt of bystander cardiopulmonary resuscitation,21 absence of diastolic dysfunction,22 functional status at hospital discharge,23-25 initial cardiac arrest rhythm,11 and post–cardiac arrest care at a tertiary referral hospital.26
Still, although there is considerable research examining 1-year survival rates,1,9,14 larger-scale studies of survival beyond 1 year are lacking. The purpose of this systematic review and meta-analysis was to assess pooled reconstructed patient data (RPD) of long-term survival after OHCA beyond 1 year and present data on potential factors associated with long-term survival after OHCA.
Data collection and reporting for this systematic review and meta-analysis followed the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) reporting guideline and the Meta-analysis of Observational Studies in Epidemiology (MOOSE) reporting guideline.27,28 To avoid reporting bias and duplication, we preregistered the study protocol on PROSPERO29 (CRD42021243689).
For the purpose of this study, long-term survival was defined as survival after OHCA for more than 12 months. Prospective, retrospective, and interventional clinical studies reporting long-term OHCA survival rates were eligible. For the primary analysis, the focus of the current study was long-term survival after hospital discharge or 30 days after OHCA. A secondary analysis focusing on long-term survival after hospital admission was also performed.
A systematic search of the online databases EMBASE (via Elsevier) and MEDLINE (via Ovid) was conducted from their inception to March 25, 2021. The search strategy was developed in collaboration with a medical information specialist (C.A.-H.) experienced in systematic reviews and meta-analysis. The final search strategy adapted for MEDLINE and EMBASE is available in eMethods 1 in the Supplement. To identify additional records, all references of eligible studies were screened (backward citation tracking), and a similar articles search was conducted for all included records on July 2, 2021.
After an initial common screening calibration phase, 2 reviewers (S.A.A. and C. Bohren) performed the title and abstract screening. The same reviewers performed the full-text screening independently, and disagreements were discussed until consensus was found.
Studies were included if they met the following inclusion criteria: adult patients (≥18 years of age) with OHCA, survival to hospital admission (including intensive care unit [ICU] admission) or until hospital discharge or until 30 days after OHCA, and reporting of survival beyond 12 months. Studies were excluded if they had unseparable mixed populations of patients with OHCA and patients with in-hospital cardiac arrest, if the cardiac arrest was mainly of a noncardiac origin (eg, trauma, drowning, electrocution, suicide, or hypothermia), and if the language of the publication was other than English, French, German, or Spanish. Animal studies, reviews, congress abstracts, editorials, letters, comments, case reports, and case series were also not considered. Furthermore, because of significant changes in resuscitation guidelines in November 2005, studies were excluded if most of the study period was before January 1, 2006.
In addition, only studies that had Kaplan-Meier (KM) survival curves with reported numbers at risk were included for the KM-based meta-analysis. For the aggregate data meta-analysis, all studies that reported the proportion of survivors were included.
For an unbiased meta-analysis of survival data, it is imperative to obtain individual patient data from the original studies, allowing additional subgroup analysis, which would otherwise not be possible.30 However, reported data details are frequently insufficient in studies that report survival data because of aggregated summary statistics, leading to critical methodologic issues for a meta-analysis.31 Because access to original individual patient data proves difficult owing to organizational, logistic, and privacy reasons,32 the criterion standard for a meta-analysis of survival data is computational reconstruction of patient data from published KM curves.32 Therefore, to obtain RPD from the included original studies, an iterative approach was used, which was initially developed by Guyot et al33 and refined by Wei and Royston34 and Liu et al.32 This approach has been described in detail previously and its high accuracy and reproducibility demonstrated repeatedly.32-34 For this study, to import the quality data coordinates (time and survival probability) of the published KM curves, a digital software program (DigitizeIt, version 2.5.9; DigitizeIt Services) was used. Combined with the numbers at risk reported for several time points, the patient data of each included study were reconstructed using the web application Shiny, version 126.96.36.199,35 which was developed by Liu et al32 and integrates the iterative algorithm into a user-friendly web application. The authors checked the accuracy of the RPD by comparing the survival probability at each read-in point with the corresponding reconstructed survival probability and the estimated number of patients at risk compared with reported values in the original data.32 In addition, using the RPD, KM curves of each study cohort were computed, and 2 independent reviewers (S.A.A. and C. Bohren) assessed them for accuracy through visual inspection. In case of uncertainties concerning data in the original publication, the authors were contacted for clarification.
For further analysis, the RPD from the individual studies were pooled. The RPD from studies reporting survival from hospital discharge or 30 days after OHCA were pooled to obtain the primary and secondary outcomes. For long-term survival after hospital admission or ICU admission, the RPD from studies that reported survival from hospital admission or ICU admission were pooled. For the stratified analysis by initial cardiac arrest rhythm, the RPD were pooled for each arm, and new hazard ratios were calculated.
An aggregate data meta-analysis of all eligible studies was also performed for all studies that reported long-term outcomes (>12 months) of patients with OHCA surviving to hospital discharge or 30 days after OHCA. The detailed methods can be found in eMethods 2 in the Supplement. In brief, a meta-analysis of survival proportions (overall survival and survival at 3, 5, and 10 years of follow-up) of eligible studies was performed using a random-effects model.
The risk of bias was assessed by the ROBINS-I (Risk of Bias in Non-randomized Studies of Interventions) tool36 as far as applicable. The ROBINS-I tool assesses 7 domains (confounding, selection of participants into the study, classification of interventions, deviations from intended interventions, missing data, measurement of outcomes, and selection of the reported result) by which bias might be introduced into nonrandomized studies. Because the included studies did not test a specific intervention, the domains “classification of interventions” and “deviation from intended intervention” were rated as not applicable. Two reviewers (S.A.A. and C. Bohren) independently assessed the risk of bias for each included study using a standardized form. Disagreements were discussed until consensus was found.
All statistical analyses were performed using Stata MP, version 15.1 (StataCorp LLC). A 2-sided P < .05 was considered to be statistically significant. In the KM-based meta-analysis, mean (SD) survival times, median (IQR) survival times, and percentage of survival at different time points with 95% CIs were calculated. In the KM-based meta-analysis, differences in survival between groups were assessed by the log-rank test for difference and a Cox proportional hazards regression model.
A total of 15 342 records were identified through database searches and 5 records through other sources. After removing duplicates (n = 5296), the remaining 10 051 records were screened on titles and abstracts. Of the resulting 854 full-text articles, 38 studies8,11,15,16,20-23,37-66 were included in the final qualitative and quantitative analysis. A total of 21 studies with 11 800 patients11,16,23,37-54 were used for KM-based meta-analysis, whereas 33 studies with 16 933 patients8,11,15,16,20-23,37,40-43,45-51,54-66 were used for aggregate data meta-analysis. eFigure 1 in the Supplement outlines the study selection process.
The characteristics of the included studies8,11,15,16,20-23,37-66 are given in Table 1, and a detailed summary of the included studies can be found in eTable 1 in the Supplement. The studies were conducted on 4 continents: Europe (n = 20), North America (n = 8), Asia (n = 6), and Australia (n = 4). All studies had an observational design. Sample sizes ranged from 3552 to 344937 for KM-based meta-analysis and from 1662 to 344937 for aggregate data meta-analysis. Most studies were conducted after January 1, 2006, with a range of publication years from 2011 to 2021. None of the studies included patients with COVID-19.
Risk of bias was assessed for all included 38 studies8,11,15,16,20-23,37-66 (eTable 2 in the Supplement). Thirteen studies11,15,16,21,37,43,53,55,57,58,61,62,66 had a low risk of bias, and 21 studies8,22,23,39-42,44-52,54,59,60,64,65 had a moderate risk of bias from reporting outcomes of a selected subgroup of patients with OHCA. Four studies20,38,56,63 were identified with a serious risk of bias because they included only a highly selective subgroup of patients with OHCA: the studies by Sanghavi et al63 and Chan et al20 included only patients 65 years or older, the study by Conte et al38 included only patients with idiopathic ventricular fibrillation patients, and the study by Herman et al56 included only apparently unexplained cardiac arrests. No study had a critical risk of bias.
Studies used for the KM-based meta-analysis had different starting points regarding long-term follow-up. For the main analysis, the initial in-hospital mortality was excluded, and 10 studies23,37,38,41,43,44,46,51,52,54 starting at hospital discharge and 4 studies11,39,42,45 starting 30 days after OHCA were pooled. For our secondary analysis, 1 study40 starting at ICU admission, 3 studies47,50,53 starting at hospital admission, and 2 studies48,49 starting at either hospital or ICU admission were included.
Overall, 14 studies11,23,37-39,41-46,51,52,54 with a total of 8175 patients reported long-term overall survival of patients surviving to hospital discharge or 30 days after OHCA with a maximum follow-up of 15 years. During a total observation time of 44 515 person-years, 1992 deaths occurred. The mean (SD) survival time was 5.4 (3.8) years, and the median survival time was 5.0 years (IQR, 2.3-7.9 years). The survival rates were 82.8% (95% CI, 81.9%-83.7%) at 3 years, 77.0% (95% CI, 75.9%-78.0%) at 5 years, 63.9% (95% CI, 62.3%-65.4%) at 10 years, and 57.5% (95% CI, 54.8%-60.1%) at 15 years (Figure 1A and Table 2).
The results of 2 studies11,16 with a total of 1111 patients were available for a stratified analysis regarding long-term survival in patients with and without shockable initial rhythm. Compared with patients with a nonshockable initial rhythm, patients with a shockable initial rhythm had a significantly better long-term survival (HR, 0.30; 95% CI, 0.23-0.39; P < .001) (Figure 2 and Table 2).
In total, 6 studies40,47-50,53 with 3385 patients investigated the long-term survival rate after hospital admission or ICU admission with a maximum follow-up time of 11 years. During a total observation time of 5753 person-years, 2150 deaths occurred. The mean (SD) survival was 1.7 (2.5) years, and the median survival 0.1 years (IQR, 0.0-3.0 years), with a 5-year survival rate of 32.6% (95% CI, 30.8%-34.4%) and a 10-year survival rate of 28.2% (95% CI, 25.7%-30.7%) (Figure 1B and Table 2).
The aggregate data meta-analysis included a total of 33 reports8,11,15,16,20-23,37,40-43,45-51,54-66 on outcomes from 30 cohorts of OHCA survivors. The results of 30 studies8,15,20-22,37,40-43,45-51,54-66 with a total of 15 599 patients were eligible for an analysis of overall survival after hospital discharge (Figure 3). The results were stratified according to follow-up duration (>1-3 years, >3-5 years, or >5 years). The results showed evidence of high overall heterogeneity (I2 = 99%, P < .001) and significant between-group heterogeneity (survival rates of 70% in the >1- to 3-year category, 77% in the >3- to 5-year category, and 64% in the >5-year category, P = .04), with proportions of survivors ranging from 0.22 to 0.92 in studies in the category of >1 to 3 years, 0.49 to 0.99 in studies in the category of more than 3 to 5 years, and 0.50 to 0.76 in studies in the category of greater than 5 years. The additional results of the aggregate data meta-analysis can be found in the eResults and eFigures 2 to 6 in the Supplement.
A post hoc secondary analysis was performed for the aggregate data analysis excluding the 2 largest studies37,63 and for the KM-based meta-analysis excluding the study by Andrew et al.37 Results were similar to the overall analysis (eFigures 7-9 in the Supplement).
This systematic review and meta-analysis of long-term survival beyond 12 months after OHCA of presumed cardiac origin brings together observational evidence from 38 studies8,11,15,16,20-23,37-67 with a total of 28 733 patients. In KM-based meta-analysis, patients who initially survived the hospital stay after OHCA had survival rates of 82.8% at 3 years, 77.0% at 5 years, 63.9% at 10 years, and 57.5% at 15 years of follow-up (Figure 1). The stratified analysis by initial cardiac arrest rhythm suggests that an initial shockable rhythm is associated with a strong survival benefit during at least 8 years of follow-up (Figure 2). Finally, the long-term survival in patients admitted to the hospital was considerably lower, with a 32.6% survival at 5-year follow-up and 28.2% survival at 10-year follow-up, mainly because of high mortality during the index hospitalization (Figure 1). The KM-based meta-analysis of survival after hospital discharge was also validated in an aggregate data meta-analysis showing similar survival rates and a risk ratio for long-term mortality favoring a shockable initial cardiac arrest rhythm, thus indicating the reliability of the KM-based approach (eFigures 3-6 in the Supplement).
Based on observational data, patients with OHCA have significantly lower long-term survival when compared with age- and sex-matched cohorts or an unselected overall population, especially during the first year after OHCA.37,57 However, the validity of this comparison is limited because patients with OHCA usually carry a high burden of prearrest comorbidities.68,69 Populations with similar prearrest comorbidities would be more suitable for comparison of long-term survival and include general ICU patients, patients with acute myocardial infarction, or patients with stroke.70-73 Indeed, a large Australian general ICU cohort (n = 19 921), excluding cardiac surgical patients, reported a survival rate of 54.7% at 15 years of follow-up,72 which is comparable to the results of this meta-analysis. Of interest, long-term survival is lower in patients after coronary artery bypass grafting, with 51% survival after 10 years and 27% survival after 15 years.74 Nevertheless, arguably the most suitable comparators are patients with acute myocardial infarction due to similar comorbidities and overlapping origins.75,76 Studies75,77,78 based on older data sets reported 10-year survival rates of discharged patients with ST-elevation myocardial infarction (STEMI) of approximately 50%. A newer study79 reports a 10-year survival of 76.2%. Hence, the 10-year survival of patients with OHCA found in this meta-analysis of 63.9% is slightly inferior to the survival of discharged patients with STEMI.
The initial rhythm of cardiac arrest is a well-documented factor associated with short-term survival after OHCA.9,80 In a large meta-analysis9 of 82 854 patients, the pooled odds ratios for survival to hospital discharge of patients with a shockable initial rhythm of cardiac arrest ranged from 2.91 (95% CI, 1.10-7.66) in the studies with the lowest survival rates to 20.62 (95% CI, 12.61-33.72) in the studies with the highest survival rates when compared with nonshockable rhythms. However, data on the association of the initial rhythm and long-term survival are scarce. Majewski et al11 found a significant and persistent positive association of an initial shockable rhythm and long-term survival for up to 8 years after hospital discharge. These results are now consolidated by the increased statistical power of this meta-analysis, which suggests a positive association with survival for 8 years or more.
The presented data have several implications for clinical practice and research. First, although survival to hospital discharge of patients with OHCA remains low, with in-hospital mortality of approximately 50% to 60%,1,5 the presented data for the long-term survival of patients surviving the index hospitalization are comparable to those for general ICU patients, including patients with STEMI.70-73,81 The long-term survival of patients with OHCA is substantially higher compared with that of patients after coronary artery bypass grafting.74 The similar long-term survival of discharged patients with OHCA is an important finding but has to be interpreted with caution because of the possibility of severe brain injury in OHCA survivors.6 Still, data suggest that patients with unfavorable neurologic outcomes represent only a small proportion of long-term survivors, as 2 large, long-term cohorts with a total of 1858 patients have shown that patients with an unfavorable neurologic outcome (cerebral performance category 3-4) account for only approximately 15% of patients with OHCA discharged alive.23,25 Of the latter, only approximately 33% survive to 5 years of follow-up.23,25 The good long-term survival as shown in this meta-analysis, combined with the probability of a good neurologic outcome, contributes to the evidence used in the decision-making process and discussion with patients and their relatives in the critical care setting and rehabilitation as well as the discussion in do not attempt resuscitation orders.
Second, this meta-analysis found an initial shockable cardiac arrest rhythm to be associated with improved long-term survival compared with a nonshockable cardiac arrest rhythm. This finding reinforces the importance of using the initial cardiac arrest rhythm in prognostic models for long-term survival and supports the notion of early defibrillation in the survival chain and the need for widened availability and accessibility of defibrillators.
Third, although an increasing body of evidence supports the KM-based meta-analysis, the meta-analysis might be considered inaccurate because of its reconstructive nature.32-34 The data presented in this report, which used classic aggregate data as well as a KM-based meta-analysis, again support the accuracy and applicability of the KM-based approach. This finding should encourage future research in the field of survival data to apply this innovative approach.
This study has several strengths. First, pooling long-term data on survivorship after cardiac arrest adds a novel perspective of epidemiologic data. Second, this study focused on data collected after 2005, accounting for this up-to-date report regarding treatment protocol of long-term survival and associated factors in this specific group of patients. Third, the applied KM-based approach allows for an accurate meta-analysis of survival data, which is usually prone to bias because of insufficient detail of reported aggregated summary statistics.32 Fourth, with pooled data from 28 733 patients, this study has adequate power to support the conclusions. Fifth, the internal validity of the present systematic review is supported by an overall low to moderate risk of bias in most of the included studies.
This systematic review and meta-analysis also has several limitations. First, 1 study37 provided approximately one-third of patients for the KM analysis of long-term survival after hospital discharge, which limits the external validity and generalizability of the results. However, an additional sensitivity analysis yielded similar results (eFigure 7 in the Supplement). Second, important confounders could not be adjusted for. Thus, there is a risk of confounding by age, resuscitation circumstances, and other vital parameters, such as worse survival attributable to early withdrawal of life-sustaining therapy. Third, studies that reported outcomes 1 year or less after OHCA were excluded, which may cause selection bias. Fourth, this study is not able to provide survival estimates of a matched control group without OHCA, which would improve the interpretation of data. Fifth, because this study is unable to provide information on neurologic status at hospital discharge and functional capacity during long-term follow-up, the analysis provides little direction to improve post-OHCA care.
In this comprehensive systematic review and meta-analysis, long-term survival to 10 years in patients surviving the initial hospital stay after OHCA was 62% to 64%. Additional research is needed to understand and improve long-term survival in this vulnerable patient population.
Accepted for Publication: March 1, 2022.
Published Online: May 4, 2022. doi:10.1001/jamacardio.2022.0795
Corresponding Author: Sabina Hunziker, MD, MPH, Intensive Care, University Hospital Basel, Petersgraben 4, CH-4031 Basel, Switzerland (firstname.lastname@example.org).
Author Contributions: Dr Amacher and Ms Bohren had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Dr Amacher and Ms Bohren contributed equally to this work.
Concept and design: Amacher, Bohren, Becker, Hunziker.
Acquisition, analysis, or interpretation of data: Amacher, Bohren, Blatter, Beck, Müller, Loretz, Gross, Tisljar, Sutter, Appenzeller-Herzog, Marsch, Hunziker.
Drafting of the manuscript: Amacher, Bohren.
Critical revision of the manuscript for important intellectual content: Amacher, Blatter, Becker, Beck, Müller, Loretz, Gross, Tisljar, Sutter, Appenzeller-Herzog, Marsch, Hunziker.
Statistical analysis: Amacher, Bohren, Blatter, Hunziker.
Obtained funding: Becker, Hunziker.
Administrative, technical, or material support: Amacher, Bohren, Beck, Müller, Loretz, Gross, Sutter, Hunziker.
Supervision: Tisljar, Marsch, Hunziker.
Design of search strategy: Appenzeller-Herzog.
Conflict of Interest Disclosures: Dr Sutter reported holding stock in Novartis, Roche, Alcon, and Johnson & Johnson and receiving grants from UCB Pharma outside the submitted work. No other disclosures were reported.
Funding/Support: Dr Hunziker and her research team were supported by grants 10001C_192850/1 and 10531C_182422 from SNF and the Swiss Society of General Internal Medicine during the conduct of the study.
Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.