Why repurposed drugs failed against COVID-19: evidence, policy and the Bulgarian case

When SARS-CoV-2 emerged in late 2019, no proven therapeutic or prophylactic agent existed, and clinical practice was driven by urgent necessity rather than evidence-based medicine. The dominant strategic response was the rapid repurposing of existing drugs with well-documented safety profiles: the antimalarials hydroxychloroquine and chloroquine, the macrolide azithromycin, the tetracycline doxycycline, and the antiparasitic ivermectin. A new systematic review with network meta-analysis, using a data lock point of 31 December 2022 and drawing on randomized controlled trials from major platforms including RECOVERY, SOLIDARITY, TOGETHER and PRINCIPLE, now provides a definitive synthesis of how these candidates performed—and why their failure was not a failure of evidence generation but of evidence translation.

The pharmacological rationale for hydroxychloroquine and chloroquine rested on their chemistry as lipophilic weak bases. With pKa values of approximately 8.3 for hydroxychloroquine and 10.2 for chloroquine, the compounds diffuse uncharged across membranes at neutral pH but become protonated and trapped inside acidic vesicles such as endosomes and lysosomes. This ‘lysosomotropic’ effect neutralizes vesicular pH and disrupts a broad repertoire of pH-dependent processes. Their extraordinarily large volume of distribution, exceeding 5,000 liters, and terminal half-life of 40 to 50 days produce high concentrations in lung parenchyma, the primary site of SARS-CoV-2 replication—a compelling pharmacological premise on paper.

Two mechanistic hypotheses supported clinical testing. As immunomodulators, the drugs’ alkalinization of endosomes impairs activation of Toll-like receptors 7 and 9, attenuating MyD88-dependent signaling, reducing type I interferon and pro-inflammatory cytokine production, and inhibiting NLRP3 inflammasome assembly—processes implicated in severe COVID-19’s cytokine release syndrome. As antivirals, chloroquine had been shown in SARS-CoV-1 studies to inhibit terminal glycosylation of the ACE2 receptor, and endosomal alkalinization was posited to block cathepsin L–dependent viral entry. The pivotal moment came with an in vitro study reporting potent inhibition of SARS-CoV-2 replication in Vero E6 cells, with EC50 values of 1.13 micromolar for chloroquine and 0.72 micromolar for hydroxychloroquine, concentrations thought to be clinically achievable. Amplified by political endorsements and media coverage, and reinforced by a small, nonrandomized French trial of 36 patients, this evidence transformed hydroxychloroquine into a perceived miracle cure, driving emergency authorizations and bulk government purchases.

Ivermectin’s trajectory followed a strikingly similar arc. The macrocyclic lactone, whose 2015 Nobel Prize–recognized antiparasitic mechanism involves glutamate-gated chloride channels largely absent in vertebrates, had prepandemic in vitro activity against dengue and West Nile viruses through inhibition of the nuclear import protein importin α/β1. A landmark 2020 study reported an approximately 5,000-fold reduction in SARS-CoV-2 RNA in Vero/hSLAM cells 48 hours after treatment, spawning hypotheses that ivermectin could block nuclear import of viral proteins such as NSP1, inhibit the RNA-dependent RNA polymerase, suppress NF-κB and MAPK signaling, or block spike–ACE2 binding. But a fundamental pharmacokinetic–pharmacodynamic flaw was evident from the outset: the in vitro inhibitory concentrations of roughly 2 to 5 micromolar were 50 to 100 times higher than the peak plasma concentrations of about 40 to 100 nanomolar achieved with approved antiparasitic dosing of 150 to 200 micrograms per kilogram. A systematic evaluation found that compounds showing clinical benefit against COVID-19 exhibited unbound Cmax/EC50 ratios of at least 6.8 in human respiratory cell models—a threshold ivermectin could not approach without doses risking neurotoxicity.

Azithromycin and doxycycline were deployed on different grounds. Azithromycin, a 50S ribosomal subunit inhibitor with established immunomodulatory effects in chronic airway disease, suppresses IL-8 production, attenuates neutrophil activation, shifts macrophages toward a pro-resolving M2 phenotype, enhances efferocytosis, and reduces NLRP3 inflammasome activity. Its justification in COVID-19 was threefold: coverage against secondary bacterial pneumonia, immunomodulation of the hyperinflammatory phase, and—least substantiated—putative direct antiviral activity via endosomal pH alteration. Doxycycline, lacking significant in vitro antiviral activity against SARS-CoV-2, was proposed as a disease-modifying agent through inhibition of matrix metalloproteinases, zinc-dependent enzymes whose dysregulation contributes to alveolar–capillary barrier breakdown in severe disease. Elevated MMP-8 and MMP-9 levels correlate with COVID-19 severity, and a murine model showed that doxycycline-mediated MMP inhibition reduced tissue pathology, but the zinc-chelation hypothesis of direct viral inhibition remained speculative.

The randomized trial evidence dismantled each rationale. A meta-analysis of 16 randomized controlled trials encompassing 15,000 patients found no significant effect of hydroxychloroquine on 28-day all-cause mortality (pooled risk ratio 1.08, 95% confidence interval 0.97–1.19; I² = 0.0%), with the RECOVERY trial contributing 76% of the analytical weight. Clinical improvement at day 28 across eight trials (N = 6,449) showed a pooled risk ratio of 0.98 (95% CI 0.95–1.01), and progression to WHO Clinical Progression Score ≥7 showed no benefit (RR 0.93, 95% CI 0.71–1.22). Viral clearance was not accelerated (day 7 RR 1.02, 95% CI 0.82–1.27; I² = 70.8%), directly refuting the foundational hypothesis. Adverse events were borderline significantly increased (RR 1.13, 95% CI 1.00–1.27), corresponding to a number needed to harm of 20, and serious adverse events showed a consistent trend toward harm (RR 1.10, 95% CI 0.92–1.31). Chloroquine, evaluated in the WHO SOLIDARITY trial of 11,330 hospitalized patients, showed no mortality benefit (rate ratio 1.19, 95% CI 0.89–1.59), with a trend raising concern about cardiac harm given its known QTc-prolonging risk.

Ivermectin fared no better in rigorous trials. The TOGETHER trial, a double-blind placebo-controlled study of 1,358 high-risk outpatients, found no significant benefit for hospitalization or death (relative risk 0.90, 95% CI 0.70–1.16), corroborated by ACTIV-6 (hazard ratio 1.02, 95% CI 0.92–1.13 for sustained recovery). Meta-analyses showed no mortality benefit in outpatients (RR 0.97, 95% CI 0.73–1.29) or hospitalized patients (hazard ratio 0.84, 95% CI 0.65–1.08), no acceleration of viral clearance, and a trend toward increased serious adverse events (RR 1.30, 95% CI 0.74–2.29). Azithromycin showed no benefit for hospitalization or death in outpatients (RR 0.83, 95% CI 0.54–1.29) or 28-day mortality in hospitalized patients (RR 0.97, 95% CI 0.89–1.06 across 7,874 patients), and one trial reported a 149% relative increase in adverse events, mostly gastrointestinal. Doxycycline’s PRINCIPLE trial of 1,471 outpatients found no reduction in hospitalization or death (RR 0.91, 95% CI 0.59–1.42), and a small open-label inpatient trial (N = 387) showed no mortality benefit (RR 0.96, 95% CI 0.53–1.75).

The study’s distinctive contribution lies in analyzing why definitive evidence failed to translate into policy, using Bulgaria as a critical case: the highest COVID-19 mortality rate in the European Union, structural healthcare vulnerabilities, five health ministers in 21 months, and the lowest institutional trust in the EU, with prepandemic trust in government at 21% and in the health system at 35%. The analysis identifies three mutually reinforcing dimensions of failure. Structural vulnerability—chronic underfunding and infrastructural depletion—left the system without adaptive capacity under surge conditions. Governance fragmentation eroded institutional memory and decision-making authority, producing policy paralysis and de facto authority of unofficial protocols. Social legitimacy deficits meant official recommendations were systematically undermined by alternative narratives, in contrast to Nordic countries with high trust and coherent governance. International living guidelines from the WHO, NIH, and IDSA converged on recommendations against these drugs by late 2021, demonstrating that evidence-synthesis infrastructure worked—but guideline infrastructure proved insufficient where institutional carriers of evidence lacked legitimacy.

A further legacy is antimicrobial resistance. With bacterial coinfection rates among COVID-19 patients consistently below 8%, yet 75% of Bulgarian patients treated with antibiotics, the tenfold overprescription constituted an uncontrolled experiment in evolutionary selection pressure. Bulgarian data document a 27% surge in antimicrobial consumption, carbapenem resistance in Klebsiella pneumoniae rising from 15% to 40.6%, and an estimated 8% of all deaths linked to antimicrobial resistance. The authors argue that stewardship programs must be hardwired into emergency plans with protected personnel and funding, that pre-positioned platform trial networks should be maintained as always-on infrastructure, and that trust-building through transparency and accountability must be cultivated long before crises. The limitations are acknowledged: some included trials were open-label with risk-of-bias concerns, the temporal scope ends in December 2022 before the Omicron-dominated and vaccinated era, and the Bulgarian case may not generalize fully to other health system contexts. The central lesson stands, however: the failure of repurposed drugs during COVID-19 stemmed not from insufficient evidence, which arrived within months, but from systemic breakdowns in structure, governance, and legitimacy that determined whether that evidence could shape practice.

Subject of Research: Drug repurposing for COVID-19 and the translation of randomized trial evidence into national policy

Article Title: Strategies for drug repurposing in COVID-19

Article References: Kokudeva, M., Velikova, T., & Momekov, G. (2026). Strategies for drug repurposing in COVID-19. Pharmacia, 73, Article e197010. https://doi.org/10.3897/pharmacia.73.e197010

Image Credits: AI Generated

DOI: 10.3897/pharmacia.73.e197010

Keywords: drug repurposing, COVID-19, SARS-CoV-2, hydroxychloroquine, ivermectin, azithromycin, doxycycline, randomized controlled trials, meta-analysis, antimicrobial resistance, health policy, Bulgaria