Why Pharmacovigilance Can Detect “Resistance as an Adverse Event”
The idea that antimicrobial resistance could appear in a pharmacovigilance database might initially seem counterintuitive. Spontaneous-reporting systems were designed to capture adverse drug reactions like unexpected physiological effects, organ toxicity, severe allergic responses, and not microbiological outcomes. Yet in recent years, large-scale datasets such as EudraVigilance have begun to reveal a pattern: when an antibiotic repeatedly “fails” in real-world clinical settings, the event label “drug ineffective” becomes statistically disproportionate. This does not mean resistance is recorded directly, but it does mean that therapeutic failure, often driven by resistant organisms, appears frequently enough to become an identifiable pharmacovigilance signal.
Antibiotic failure is rarely random. When a drug consistently performs below expectations, clinicians become more likely to report ineffectiveness, especially if the patient experiences complications, prolonged illness, or requires hospitalization. Over time, these reports accumulate, and disproportionality analysis can distinguish sporadic failure from a meaningful trend. A strong signal for “drug ineffective” therefore becomes a soft but valuable proxy for the possibility of resistance in the population or for clinical misuse, but when the signal favors one antibiotic over another, resistance becomes a more plausible explanation.
In the case of amoxicillin versus amoxicillin/clavulanate (Amoxicillin and amoxicillin/clavulanate (Augmentin): a clinical guide for patients), both drugs are prescribed for nearly identical infections, to similar patient groups, in overlapping settings. If one accumulates significantly more “drug ineffective” reports, random variability becomes an unlikely explanation.
Co-amoxiclav is, mechanistically, more susceptible to resistance-driven failure than amoxicillin alone. The broader spectrum exerts heavier ecological pressure on commensal and pathogenic microbes, encouraging beta-lactamase proliferation and selection of organisms less responsive to the combination. Clavulanate itself neutralizes beta-lactamase activity but also exerts selective pressure that may paradoxically expand populations of resistant bacteria when exposure is widespread. This larger ecological footprint increases the chances that therapeutic failure will manifest in real-world prescribing.
Pharmacovigilance thus serves as an unexpected but increasingly important window into antimicrobial resistance. While laboratory testing and epidemiological studies remain the gold standards, spontaneous reporting captures failures across diverse regions, prescriber habits, comorbidities, and dosing patterns, offering a population-level signal that resistance is becoming clinically visible. The EudraVigilance findings reflect precisely this dynamic: therapeutic failure accumulating not as isolated events but as a reproducible pattern strong enough to be treated as a pharmacovigilance signal in its own right.
What the EudraVigilance Analysis Actually Shows: Hepatotoxicity and “Drug Ineffective” Signals
The EudraVigilance review of 18,229 individual case safety reports between 2020 and 2024 offers one of the clearest comparative snapshots of real-world outcomes for amoxicillin versus amoxicillin/clavulanate. The numerical distribution alone is revealing: nearly 11,000 reports involved the combination therapy, despite amoxicillin on its own being widely prescribed across Europe. That imbalance reflects not only usage patterns but the clinical consequences of adding clavulanate, a drug known to increase the frequency of adverse reactions, particularly within the hepatobiliary system.
The analysis confirmed what has been documented for more than a generation: co-amoxiclav is substantially more hepatotoxic than amoxicillin alone. The signal was dominated by cholestatic and mixed-pattern liver injury, a toxicity profile strongly associated with clavulanate metabolism. Although most cases resolved, the severity and recurrence risk highlight why hepatobiliary disorders remain the most frequent reason for hospitalization related to this combination.
The more surprising finding lies not in the toxicity profile but in the frequency of reports labeled “drug ineffective.” Disproportionality analysis revealed that co-amoxiclav produces a stronger ineffectiveness signal than amoxicillin, suggesting that, in real-world usage, the combination fails more often than its narrower counterpart. This counters the intuitive belief that a broader-spectrum agent should offer greater therapeutic success. The authors interpret the signal as a population-level indicator of antimicrobial resistance, specifically, that the selective pressure created by routine co-amoxiclav use may be driving resistance patterns that reduce its clinical utility. While “drug ineffective” can have multiple origins, including viral infections or dosing errors, these explanations would be expected to appear equally across both drugs. The fact that the signal was stronger for co-amoxiclav, despite similar clinical indications, supports the interpretation that resistance contributes to the discrepancy.
Methodologically, the analysis applied standard pharmacovigilance tools to minimize artifact. Disproportionality was assessed using reporting odds ratios and information component values, providing robustness across both frequentist and Bayesian frameworks. The consistency of the hepatotoxicity and ineffectiveness signals across methods further strengthens the findings. EudraVigilance cannot establish causality, but it effectively identifies patterns unlikely to arise from chance alone. Here, both toxicity and therapeutic failure point toward a clinical message: the addition of clavulanate changes not just safety, but real-world efficacy in ways invisible in controlled trials yet visible at the continental level through pharmacovigilance data.
Should Co-Amoxiclav Really Be a First-Line Therapy? Rethinking Its Routine Use in ENT and Respiratory Infections
The EudraVigilance findings arrive at a time when prescribing habits in many European countries continue to favor co-amoxiclav as a reflexive first-line choice for common ENT and lower-respiratory infections. Sinusitis, acute otitis media, pharyngitis, bronchitis, dental infections are the conditions for which the combination is routinely given, often as a “stronger” alternative that seems safer to prescribers and reassuring to patients. Yet this practice stands in tension with long-standing stewardship principles and with most evidence-based guidelines, which consistently recommend narrower agents when clinical features do not suggest beta-lactamase producing organisms. The new pharmacovigilance data now adds a compelling new dimension: not only does co-amoxiclav carry more toxicity, it may also be failing more often.
Part of the problem is therapeutic overkill. The addition of clavulanate broadens coverage far beyond what is necessary for most uncomplicated ENT infections, which are frequently viral or caused by pathogens that remain highly susceptible to amoxicillin. When a broad-spectrum combination is used inappropriately, it increases selective pressure on commensal flora and pathogens alike, accelerating the proliferation of beta-lactamase producing organisms. Over time, this ecological disruption erodes the drug’s effectiveness, a phenomenon now mirrored in the “drug ineffective” signal emerging from EudraVigilance. What appears at first to be a theoretical stewardship concern has manifested as a measurable pattern of therapeutic failure across Europe.
In clinical practice, the tendency to “play it safe” by choosing the broader agent often reflects diagnostic uncertainty or time pressure rather than microbiological rationale. Yet this pattern paradoxically raises the risk of both adverse events and resistance. The hepatotoxic potential of co-amoxiclav has long been known, but the combination of increased toxicity and reduced effectiveness makes the risk benefit balance less comfortable than many clinicians assume. Moreover, many ENT and respiratory infections resolve without antibiotic therapy altogether, and guidelines increasingly emphasize watchful waiting, delayed prescriptions, or narrow-spectrum therapy for precisely this reason.
Reconsidering co-amoxiclav as a first-line choice does not mean abandoning it. Rather, it requires targeted, evidence-based selection: reserving the combination for cases with a clear indication or strong suspicion of beta-lactamase positive organisms. The EudraVigilance findings shift the conversation from “is co-amoxiclav safe?” to “is it clinically justified?” For many routine and mild infections, the answer may be no, both for the individual patient and for the broader microbial landscape on which future treatment success depends.
Practical Clinical Scenarios: When Pure Amoxicillin Is the Better Choice and When Co-Amoxiclav Is Justified
In daily practice, choosing between amoxicillin and co-amoxiclav often hinges less on microbiology than on habit. The EudraVigilance findings offer a timely reminder that the decision carries meaningful clinical consequences. Amoxicillin alone remains an excellent first-line therapy for many common infections, particularly when the likelihood of beta-lactamase producing organisms is low. Streptococcal pharyngitis, most uncomplicated cases of otitis media, early-community pneumonia without significant risk factors, and mild sinusitis respond reliably to amoxicillin, which achieves high tissue penetration and has a far cleaner hepatic safety profile. In these settings, adding clavulanate provides little advantage while exposing patients to increased toxicity and greater ecological pressure.
There are, however, situations where co-amoxiclav remains entirely appropriate. Human and animal bites, diabetic foot infections, persistent or severe sinusitis, and dental infections with suspected anaerobic involvement all warrant broader coverage. Likewise, patients with documented or strongly suspected beta-lactamase producing pathogens benefit from the clavulanate component, which can restore activity where amoxicillin monotherapy would fail. The key is that these indications represent specific clinical scenarios not the majority of routine ENT or respiratory complaints.
Applying the EudraVigilance insights means sharpening antibiotic selection rather than restricting access. Amoxicillin should be the default when its spectrum matches the likely pathogen. Co-amoxiclav should be reserved for situations where its added breadth is clinically meaningful. This distinction helps preserve the effectiveness of both agents while reducing preventable toxicity and slowing resistance trends already visible in pharmacovigilance data.
References
- Reyes, M., Rosales, A., Álvarez, A., Caccia, R., & Hernández, E. (2025). EudraVigilance analysis of amoxicillin and amoxicillin/clavulanic acid: Hepatoxicity and drug-ineffective signals as indicators of antimicrobial resistance. Molecules, 30(18), 3825. https://www.mdpi.com/1420-3049/30/18/3825
