Clinical trial research on COVID-19 in Germany – a systematic analysis

Background: In 2020, the COVID-19 pandemic led to an unprecedented volume of almost 3,000 clinical trials registered worldwide. We aimed to describe the COVID-19 clinical trial research agenda in Germany during the first year of the pandemic. Methods: We identified randomized clinical trials assessing interventions to treat or prevent COVID-19 that were registered in 2020 and recruited or planned to recruit participants in Germany. We requested recruitment information from trial investigators as of April 2021. Results: In 2020, 65 trials were completely (n=27) or partially (n=38) conducted in Germany. Most trials investigated interventions to treat COVID-19 (86.2%; 56/65), in hospitalized patients (67.7%; 44/65), with industry funding (53.8%; 35/65). Few trials were completed (21.5%; 14/65). Overall, 187,179 participants were planned to be recruited (20,696 in Germany), with a median number of 106 German participants per trial (IQR 40 to 345). From the planned German participants, 13.4% were recruited (median 15 per trial (IQR 0 to 44). Conclusions: The overall German contribution to the worldwide COVID-19 clinical trial research agenda was modest. Few trials delivered urgently needed evidence. Most trials did not meet recruitment goals. Evaluation and international comparison of the challenges for conducting clinical trials in Germany is needed.


Introduction
The COVID-19 pandemic has led to an unprecedented volume and speed in the international clinical research agenda.Almost 700 clinical trials were registered worldwide in the first 100 days, which planned to recruit almost 400,000 participants to assess interventions to treat or prevent COVID-19. 1 However, this research agenda has been tainted by a multitude of small trials with 50 from the 700 trials aimed to include more than 1,000 participants. 1Nevertheless, there were remarkable efforts to conduct large adaptive pragmatic trials directly informing therapeutic decisions in COVID-19 patients: the UK Randomised Evaluation of COVid-19 thERapY (RECOVERY) trial recruited 10,000 patients in two months, 2 the established Randomised, Embedded, Multifactorial Adaptive Platform trial for Community-Acquired Pneumonia (REMAP-CAP) was rapidly adapted to COVID-19, 3 and the COVID-19 Solidarity Trial for COVID-19 Treatments (SOLIDARITY) sponsored by the World Health Organization (WHO), was conducted in six months. 4In the first 100 days, China, the US, Spain and France, had the largest share of the initiated trials, with many planned as international collaborations. 1Germany contributed to large-scale international studies such as REMAP-CAP 5 and WHO-SOLIDARITY, 6 and the German Federal Ministry for Education and Research (BMBF) provided 1.6 billion € for COVID-19-related research. 7e aim of this study was to describe the German contribution to the worldwide COVID-19 clinical trial agenda, including all randomized clinical trials (RCTs) with German participants, that were registered during the first year of the pandemic.

Data sources
We searched the COVID-evidence database to identify RCTs in international trial registries (ClinicalTrials.gov;WHO International Clinical Trials Registry Platform).Additionally, we manually searched the German Clinical Trials Register (DRKS).For all eligible registered trials, we searched corresponding publication results and preprints in the Living OVerview of Evidence platform for COVID-19 (LÁOVE), Cochrane COVID-19 Study Register, MEDLINE/PubMed, and Google Scholar using the trial registry number (last search 1 April 2021, see extended data 8 ).

Eligibility criteria
We included all planned, ongoing, or completed RCTs assessing interventions to treat or prevent COVID-19, registered in 2020, that recruited or planned to recruit at least one participant in Germany.

Selection of trials and data extraction
One author (JH) conducted the searches and screened the trials for eligibility and a second author (PJ or LGH) was consulted in cases where eligibility was unclear.Four authors (JH, PJ, AR, MB) extracted data on trial characteristics (status, duration, design, population, intervention and control, recruiting countries, actual/planned trial size, funding) and published results.
Trials with industrial/commercial sponsors were classified as industry-funded, those with public/non-commercial sponsors as public-funded, and trials with public/non-commercial sponsors with collaborators from industry as publicly funded with industry contribution.Trials were also classified as completely (i.e., national trials) or partially (i.e., international) conducted in Germany.Details on the extraction process are provided as extended data. 8

Recruitment information
We identified contact details of corresponding investigators for all eligible trials using registry information and web search.For each trial, we asked investigators via email about the current trial status, inclusion date of the first patient, number of patients recruited in Germany, and any published results (prior to April 2021, see extended data 8 ).For trials completely conducted in Germany with public funding, we further asked (i) if they were aware of other trials with public funding that we have not identified and (ii) to confirm the extracted information on their trial (see extended data 8 ).We received replies from 60% (39 out of 65) of all eligible trials.This complemented the data contained in registries or publications and provided full information on actual and target sample size in Germany for 34 trials (52.3%).

COVID-evidence update
COVID-evidence is a living database that is continuously being updated. 9,10Registry entries retrieved weekly from ClinicalTrials.gov and the WHO International Clinical Trials Registry Platform are automatically pre-screened for eligibility using basic filters to identify RCTs.Unclear and identified RCT entries are then manually screened to verify eligibility.
In August 2021, the RCT filter of COVID-evidence was updated.The data presented herein are based on the COVIDevidence processes (including RCT filter and automatic pre-screening) as of April 2021.For transparency and exhaustiveness, we updated our search in the COVID-evidence database in August 2021 with the use of the most recent processes and post-hoc identified trials.

Data analysis
We report medians with interquartile ranges (IQRs) if not stated otherwise.For all analyses, we used R (version 4.1).

Results
In 2020, 65 RCTs were planned to investigate treatments or preventive interventions for COVID-19 with participants from Germany (Table 1 and extended data 8 ).They aimed to include a median of 300 participants per trial (IQR, 174 to 830) internationally, including 106 in Germany (IQR, 40 to 345, Table 2).
Our results also indicate that 17 trials (26.2%) had published results; 15 were partially conducted in Germany, 15 were explored COVID-19 therapies and 12 were industry-funded.From those 17 trials with published results, 12 trials had results published as peer-reviewed articles or preprints, and five trials had published results exclusively as press release or in the registry.Results were reported by 11 trials, while six stated interim results (Table 3).
All trials assessing interventions to prevent COVID-19 were vaccine trials (13.8%; 9 from 65).They were considerably larger than therapy trials with a total planned sample size of 153,639 participants (median 2,520 per trial [IQR, 1,200 to 34,000]); seven of the nine vaccine trials (77.7%) planned to include over 1,000 participants.From the 187,179 planned participants in the 65 trials, 82.1% were healthy participants planned to be recruited in vaccine trials.
None of the trials investigated non-pharmaceutical interventions to prevent the pandemic spread, such as social distancing or behavioral interventions.
Hospitalized patients were recruited in 44 (67.7%), outpatients in 14 (21.5%), and both inpatients and outpatients in seven (10.8%) trials.No trial was conducted in nursing homes, kindergarten, childcare, or schools (Table 1 and extended data 8 ).Adolescents (12 years and older) were included in five trials partially conducted in Germany, however, trials that included children below the age of 12 did not exist.
Most trials were double blinded (n = 36, 55.4%) and used a two-arm parallel group design (n = 54, 83.1%).From the 11 trials that had an adaptive design (17.0%), five re-estimated the target sample size.
From the 38 trials partially conducted in Germany, nine (23.7%) planned to recruit over 1,000 participants compared with five from 27 (18.5%) of the trials that were completely conducted in Germany.Notes: * The number of trials corresponds to trials with full recruitment information; ** For the 13 trials that provided recruitment information and announced a completion date by April 2021 (timepoint of the recruitment status assessment), the median proportion of target sample size was 13.6% (IQR, 0.6 to 24.0%, range 0 to 113%).   he convalescent plasma intervention arm reached a pre-specified statistical threshold for futility among patients who are critically ill with COVID-19.Although recruitment continued for moderate patients, following the publication of the RECOVERY trial results showing no stat.significant benefit on mortality outcomes the convalescent plasma intervention arm was closed to recruitment. 34,35esults reported in multiple peer reviewed publications, [30][31][32] preprints 33 and press releases 34,35 ; industry-funded.

Actual recruitment in Germany
The median number of planned participants from Germany was 106, while the median recruited was actually 15 participants per trial (IQR, 0 to 44), corresponding to a median proportion of 13.4% of the Germany-specific target sample size recruited per trial (IQR, 0 to 29.2%, range 0 to 113%).This proportion was almost identical (13.6%) in the 13 trials that announced a completion date by April 2021 and provided recruitment information (Table 2).Two of these 13 (15.4%)reached their target sample size (i.e., at least 99%; Figure 1).
From the 65 trials, 11 (16.9%) have not yet been recruiting or, although they had planned to, will never recruit participants in Germany.
The trials with recruitment information from the investigator survey were more often completely conducted in Germany (74.1% vs. 36.8%partially conducted in Germany), more often publicly funded (80% vs. 34.3%industry-funded), and more often registered in the first half of 2020 (64.3% vs. 31.8%second half of 2020; see extended data 8 ).

Trials identified post-hoc
The search in COVID-evidence in August 2021 identified three additional small trials aiming to include 64 to 130 patients.Two were registered late December 2021, and two have been completed, however none of the results are available.For details see extended data. 8  Notes: All trials aimed at treating COVID-19 and had a two-arm parallel group design with no treatment/usual care as control if not stated otherwise.In four trials (CVC for COVID-19, COVID-PREVENT, NIC-002, HERO-19), there was information on commercial co-funding since a private company was named as a collaborator.Two trials (ECMO-VID, NCT04377334) by a principal investigator from a university hospital had an unclear funding type with no information on external commercial or public funding; + Withings ScanWatch is a smart watch for recording SpO2, ECG, and heart rate; First named trial center is lead/sponsor (exception CAPSID, sponsor: German Red Cross DRK Blood Donor Service Baden-Wurttemberg-Hesse).

Discussion
This systematic analysis of the clinical trial research agenda on COVID-19 in Germany showed that from the almost 3,000 trials registered worldwide in 2020, only 65 trials planned to include participants in Germany.
Approximately 20,696 participants from Germany were planned to be included in COVID-19 clinical trials, of which 10,613 patients would be treated for this disease.The typical COVID-19 trial with a German contribution aimed to include 106 persons in Germany, however achieved to recruit a fraction (13.4%, or 15 persons per trial).This estimate did not change when we considered trials which planned to be completed at the timepoint of our recruitment assessment in April 2021.While precise information was not available for all trials, it can be estimated that under 3,000 individuals have been included in COVID-19 RCTs in Germany (13.4% of the 20,696 German participants planned in 55 of 65 trials).In trials for COVID-19 treatments, the estimated total participants were approximately 1,500 (13.5% of 10,613 participants in 48 of 56 treatment trials).This is a small fraction of the approximately 155,000 COVID-19 hospitalizations reported during 2020 in Germany. 11These data indicate that in Germany about 1 out of 100 hospitalized patients participated in a trial to investigate potential COVID-19 therapies, while in the United Kingdom, 1 out of 6 hospitalized patients for COVID-19 took part in the RECOVERY trial. 12 was also unexpected that despite the very prominent and successful position of Germany in research and development of highly effective vaccines, more German participants have not contributed to the overall vaccine research agenda.The darker part of the bars illustrates the proportion of the target sample size in Germany that was actually recruited within each trial with the corresponding proportion (percentages reported next to the bar).The dashed vertical line corresponds to the timepoint of the recruitment status assessment, April 2021.Trials can be declared as "completed" with or without reaching their target sample size."Terminated" trials end without reaching their intended goals and started recruiting but are not considered "complete"."Withdrawn" trials will never start or recruit patients.* Trials that were partially conducted in Germany: the proportion of target sample size focus only on German participants and does not reflect the international recruitment accrual; $ End date was not available and was arbitrarily set at 1 June 2022.
There was a considerable public investment in clinical research for COVID-19.The German Federal Ministry of Education and Research (BMBF) spent 1.6 billion € for research projects related to COVID-19, a large part being allocated to vaccine development. 7The RECOVERY trial reported the cost per patient at £250, corresponding to a total of approximately 1 million Euro for a trial with 3,000 persons. 13As illustrated in Figure 1, the planned number of participants in Germany has been reached in three of the completed, and none of the early terminated trials.Moreover, 2 out of the 13 trials that announced a completion date by April 2021 reached their target sample size.The proportions of already recruited participants in the ongoing trials agree with the overall interpretation, and do not exclude delayed recruitment, which is further corroborated by the six not yet recruiting trials.Decisions for early termination may very well indicate reasonable and well-founded strategic decisions to avoid wasting research resources, however it may also reflect recruitment difficulties or other challenges.An overestimation of eligible participants or prejudiced views by recruiters and participants on trial interventions are common reasons for low recruitment. 14However, the massive impact of the COVID-19 pandemic on health care systems and clinical research institutions has challenged trial recruitment as well. 15,16Most countries do not have the long-established and effective clinical trial infrastructure and academic environments such as the UK, where the feasibility of rapid setup and implementation of massive clinical trials has been impressively demonstrated.The RECOVERY trial was planned in two days, included more than 10,000 patients with the first patient enrolled after nine days, and discovered the first mortality-lowering treatment for COVID-19 in two months. 2The highly pragmatic and embedded-in-usual-care study design were major drivers of this unprecedented successful clinical research.
As the reasons that might have led to successful or unsuccessful recruitment in the German trials were not assessed in this study, further research is needed to understand the recruitment challenges, and how they can be mitigated to facilitate recruitment success for clinical trials in Germany in situations where decision-relevant evidence is urgently needed.A careful comparison of the German research environment for clinical trial research with international circumstances seems warranted.
The focus of the trial research in Germany was similar to the worldwide trial landscape with a strong weight on exploring treatment with drugs and biologicals. 1Our results indicate that there were no RCTs registered in 2020 assessing strategies to control the pandemic spread, social distancing, or behavioral interventions although those have been strictly applied in Germany.There was also no German contribution with RCTs conducted in nursing homes, kindergarten, childcare, or schools although the role of these settings was under considerable discussion.Conversely, there are international examples of RCTs aiming to determine the best strategies for preventing virus transmission.8][19] Another US randomized trial included almost 20 million people to assess how digital information may affect local case rates. 20Different testing strategies to ensure safe gatherings at mass events have also been investigated in Spain, 21 France 22 and Norway. 23In case of facing future pandemics with similar challenges, such research would be highly important to develop evidence-based public health interventions and to improve policymaking in Germany.
This study has several limitations.First, our sample depended on accurate trial registrations and reports in the biomedical literature.We cannot exclude those trials conducted in Germany that have not been identified, for example because they were not registered, the registry data gave no indication of a German contribution, or the registry entry was delayed.Nevertheless, searching the Germany-specific registry DRKS did not yield any additional trials.The use of the most recent version of COVID-evidence in August 2021 resulted in just one more trial not previously captured by the filter, and two additional studies registered very late (28 and 30 December 2020) not previously captured due to time lags in registries.While these three studies are not included in the recruitment analysis, they have similar characteristics to the other trials and would not impact our overall interpretation.Second, although we obtained information for more than half of all included trials in an investigator request, we still had incomplete data on the actual recruitment in Germany with imbalances related to some trial characteristics.For example, we had recruitment information for 34.3% of the industryfunded trials compared to 80% of the publicly-funded trials.However, it is unlikely that the other trials have substantially higher recruitment and completion rates, and that missing recruitment information significantly affected our results.Even a five-fold higher recruitment rate in these remaining trials would not change the overall picture: which is that most trials have not been successfully recruited.Finally, the estimation of a target sample size might not be applicable for the few trials with an adaptive design, since sample size (and other methodological trial characteristics) may change over time.Therefore, our analysis related to adaptive designs should be seen contextualized within the trial progress.However, the median reported target sample size for trials with and without adaptive designs was similar.

Conclusions
The overall German contribution to the worldwide clinical trial research agenda for COVID-19 has been relatively modest.While few excellent examples of successful individual trial recruitment exist, most trials were not able to meet their goals and did not deliver the much needed evidence.A close evaluation and international comparison of the challenges and barriers for conducting clinical trials in Germany is urgently needed.
Sometimes there is a number after the decimal point and sometimes not.Please be consistent.

Discussion:
You could give more information on the funding structure and trial infrastructure in Germany to enhance the understanding of the results for the reader.The benefits of publishing with F1000Research: Your article is published within days, with no editorial bias • You can publish traditional articles, null/negative results, case reports, data notes and more • The peer review process is transparent and collaborative • Your article is indexed in PubMed after passing peer review • Dedicated customer support at every stage • For pre-submission enquiries, contact research@f1000.com

100 2020 -
Abbreviations: ECMO = extracorporeal membrane oxygenation; H = hospital (non-universitary); n = number; n.r.= not reported; SOFA = sequential organ failure assessment; U = university hospital, university medical center, or medical school.Notes: All trials aimed at treating COVID-19 and had a two-arm parallel group design with no treatment/usual care as control if not stated otherwise.In four trials (CVC for COVID-19, COVID-PREVENT, NIC-002, HERO-19), there was information on commercial co-funding since a private company was named as a collaborator.Two trials (ECMO-VID, NCT04377334) by a principal investigator from a university hospital had an unclear funding type with no information on external commercial or public funding; + Withings ScanWatch is a smart watch for recording SpO2, ECG, and heart rate; First named trial center is lead/sponsor

Figure 1 .
Figure1.Timeline, status, and actual recruitment for all 34 trials with recruitment information.Notes: The figure illustrates the length of the trials with their start and end date of conduct, sorted by start date and colored by status (based on trial registration details or investigators' request).The darker part of the bars illustrates the proportion of the target sample size in Germany that was actually recruited within each trial with the corresponding proportion (percentages reported next to the bar).The dashed vertical line corresponds to the timepoint of the recruitment status assessment, April 2021.Trials can be declared as "completed" with or without reaching their target sample size."Terminated" trials end without reaching their intended goals and started recruiting but are not considered "complete"."Withdrawn" trials will never start or recruit patients.* Trials that were partially conducted in Germany: the proportion of target sample size focus only on German participants and does not reflect the international recruitment accrual; $ End date was not available and was arbitrarily set at 1 June 2022.

○
Do authors plan to re-assess the data in 2023?(all trials should have finished recruiting then).○Youmention some valid implications for research.Do you also have recommendations or implications for practice?E.g., trialists, funding bodies.Or implications for (meta-)researchers how?○Is the work clearly and accurately presented and does it cite the current literature?YesIs the study design appropriate and is the work technically sound?YesAre sufficient details of methods and analysis provided to allow replication by others?YesIf applicable, is the statistical analysis and its interpretation appropriate?YesAre all the source data underlying the results available to ensure full reproducibility?YesAre the conclusions drawn adequately supported by the results?YesCompeting Interests: No competing interests were disclosed.Reviewer Expertise: Meta-research, meta-analysis, GRADEI confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard.PartlyCompeting Interests: I am employed by the UK NIHR as National Specialty Lead for the Comprehensive Research Network Respiratory disorders group (NIHR-CRN).I have received funding from a variety of pharmaceutical companies as grant support unrelated to this work.Additionally my funding institution receives fees for site activities for clinical trials I participate in as a Principal / Coordinating Investigator Reviewer Expertise: Research background Respiratory medicine, Clinical Trials, COVID 19, I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard.

Table 1 .
Characteristics of all 65 randomized COVID-19 trials planned to be conducted in Germany.

Table 2 .
Number of planned and actually recruited participants overall and from Germany in randomized COVID-19 trials.

Table 3 .
Trial results summary and recruitment status of all 17 trials with published results.Remdesivir for 5 days or 10 days vs standard care in patients with moderate COVID-19 on clinical status at day 11, assessed on a 7-point ordinal scale.Patients randomized to 5 days remdesivir had a stat.significant benefit (odds ratio, 1.65, 95% CI [1.09; 2.48], p = 0.02) but not those randomized to 10 days remdesivir (p=0.18).Results are based on 596 randomized patients.The trial was then extended to include additional patients.
31mdesivir versus placebo in hospitalized patients with COVID-19 with evidence of lower respiratory tract infection on the time to recovery, defined by either discharge from the hospital or hospitalization for infection-control purposes only assessed on an 8-point ordinal scale.Remdesivir had a stat.significantbenefitcompared with placebo (rate ratio for recovery, 1.29, 95% CI [1.12; 1.49], p < 0.001).Results reported in a peer reviewed publication26and posted on Clinicaltrialsregister.eu (EUCTR2020-001052-18-DE); industry-funded.A phase 2/3 trial on anti-C5a antibody IFX-1 (vilobelimab) versus standard care in hospitalized patients with severe COVID-19 pneumonia on percentage change in PaO 2 /FiO 2 in the supine position from baseline to day 5. Results focus on the phase 2 which randomized 30 patients in the Netherlands.No stat.significantdifference for the percentage change in PaO 2 /FiO 2 (difference -24%, 95% CI [À58; 9], p = 0.15), but preliminary findings on mortality at 28-days promising enough to continue with phase 3 and mortality as primary endpoint.Results reported in a peer reviewed publication27; industry-funded.30Treatmentwitha7-day fixed-dose course of hydrocortisone or shockdependent dosing of hydrocortisone suggested a benefit assessed on 384 patients randomized.The intervention arms were terminated early following the press release from the RECOVERY trial showing a stat.significantbenefit on mortality outcomes.31Bothtocilizumab and sarilumab (interleukin-6 receptor antagonists) met the predefined criteria for efficacy assessed on

Table 4 .
Characteristics of all 22 randomized COVID-19 trials planned to be conducted in Germany that received public funding.