Can we afford to lose more antimicrobial drugs?

Every known drug can cause unpredictable reactions. Moreover, a significant number of drug-induced serious adverse events are detected only after their release on the market. In the last 20 years of antimicrobial pharmacotherapy development, we have been helpless witnesses of the downfall of many drugs that displayed very good clinical efficacy but were plagued with unforeseen serious adverse events. The events were all identified after the release of these drugs for clinical use.

Some of the problems discovered were impossible to predict. The so-called temafloxacin syndrome is an example of such an unexpected serious adverse event. This promising oral broad-spectrum fluoroquinolone was withdrawn from the market 4 months after its approval owing to an unusually high frequency of serious hemolysis reports, sometimes associated with organ system dysfunction [1].

Another drug that has received its share of attention over the last 5 years is linezolid. Despite a somewhat uneventful development, there has been an increasing number of warning signs in the medical literature regarding myelosuppression [2], mitochondrial toxicity that might lead to lactic acidosis and peripheral/optical neuropathy [3,4], and serotonin syndrome (SS) [5,6] associated with linezolid use. SS reflects a hyperstimulation of the serotonin receptor 5-HT_1A. The use of high doses of selective serotonin reuptake inhibitors (SSRIs) or combination of SSRIs (or other serotonergic agents, such as bupropion or fentanyl) with drugs that exhibit monoamine oxidase inhibition activity (such as moclobemide or linezolid) will typically induce SS. It is noteworthy that no cases of SS had been reported before the launch of linezolid in the USA in 1999. Since then, at least 30 spontaneous case reports of SS with a possible causal relationship with linezolid use were published or reported to regulatory agencies [3]. Most of the cases reported were mild in severity. The likelihood of encountering SS associated with linezolid will probably be even more significant now, considering the widespread use of SSRIs in clinical practice. Patients prescribed higher doses of SSRIs or other serotonergic drugs, or those receiving more than one drug with serotonergic activity, are probably at higher risk of displaying this kind of reaction, although risk factors for the development of SS in patients treated with linezolid have not yet been clearly defined [7].

Even the so-called ‘old’ drugs, such as ceftriaxone, can still give us some concerns. Even though ceftriaxone has been on the market for more than 20 years, with an enviable safety record, there have recently been a number of deaths reported in neonates and newborns who were treated with ceftriaxone and received parenteral calcium concomitantly [101]. Crystalline material was found in the renal and pulmonary vasculature at autopsy in some of these patients. Incompatibility between ceftriaxone and calcium was, if not well known by practitioners, at least well documented [8]. This physical–chemical interaction appears to have played a significant role in the deaths reported. As a result, the US FDA recommended that: “Generally, fatalities have been associated with simultaneous administration of ceftriaxone and calcium-containing products. However, administration of the two products at different times and via different infusion lines has also been fatal. Therefore, ceftriaxone should not be mixed with calcium-containing products and not administered in the same or different infusion lines or sites in any patient within 48 h of each other” [101]. Further to this recommendation, some pediatric healthcare facilities have already withdrawn ceftriaxone from their drug formulary as a precaution, since all deaths related to this adverse reaction have been observed in neonates or newborns. However, solely considering its potential for worsening hyperbilirubinemia in neonates, the relevance of using cetriaxone in this patient population was already highly debatable from a safety point of view [9]. As for the adult population, the course of action following this warning has given rise to controversy [9]. This reaction has never been reported in adults before, and the amount of calcium administered to adults is usually lower proportionally to what is used in children. Furthermore, will pharmacy systems be able to avoid the administration of calcium in the 48 h following the last dose of ceftriaxone? Should ceftriaxone be banished from critical care or surgical units where the use of calcium is more common and why is this problem arising after more than 20 years of use? All these questions are left unanswered for now.

Unfortunately, we often do not know beforehand which factors might trigger an adverse reaction, whether being of an immune, metabolic (e.g., inhibition or induction of cytochrome P450 isoenzyme activity), physical–chemical or another nature. Clinical trials required by regulatory agencies during the drug development process are designed to evaluate the efficacy and safety of a treatment with a population exhibiting very limited comorbidities and, therefore, limiting the impact of confounding factors. Hence, patient populations in Phase I–III studies are often quite different from the patients consulting physicians on a daily basis. Potential drug interactions are thus more common after the introduction of a drug for clinical use. Moreover, as the population grows older, we also witness increases in comorbidities that can influence drug response.

In this regard, postmarketing surveillance is of the utmost importance in bringing to light potential serious adverse events. Nevertheless, one of the problems we face is the scarcity of well-organized postmarketing surveillance systems for antimicrobial agents. The need for postmarketing studies has been pointed out for more than 20 years, with the hope that computerized systems would help gather pharmacovigilance data [10]. However, most of the information concerning drug safety, and particularly serious adverse events, still arises from spontaneous reporting to regulatory agencies or sometimes through cases reported in the medical literature. The causal relationship between the observed reaction and the drug is often poorly documented and it sometimes takes years to link, with some degree of certainty, an adverse reaction to a particular drug [11,12]. The recent example of gatifloxacin-associated dysglycemic events [13,14] illustrates the time gap between the observation of an adverse event associated with the use of a particular drug and the confirmation of the problem being caused by this drug. For 3 years after the first publication reporting three cases of gatifloxacin-associated dysglycemic events, we either suspected gatifloxacin to be the culprit or simply wondered whether it was just an amplification phenomenon. Many clinicians considered the possibility that it could be a class effect for all fluoroquinolones, since glycemic abnormalities had been reported with other quinolones as well, albeit very infrequently. Even after gatifloxacin withdrawal by Bristol-Myers Squibb in 2006, we still do not know exactly to what extent this reaction might occur in patients treated with other fluroquinolones.

Recent advances in postmarketing surveillance strategies indicate that some web- and healthcare-based strategies might prove useful in the near future [15,16]. Some government-supported initiatives, such as the Research on Adverse Drug Events and Reports (RADAR) program, also exist and can improve the quality of documentation of spontaneous adverse reaction reports [17]. However, these programs depend largely on public funding and their long-term survival is, unfortunately, always uncertain. The pharmaceutical industry has made a commitment to regulatory agencies, such as the FDA, to conduct postmarketing studies. Recently, some criticism was raised following evidence that this commitment does not seem to translate very often into tangible results [18]. Some will also question the impartiality of industry-driven postmarketing studies and often consider them biased. Therefore, we should exploit the networking potential with widespread access to internet and develop networks of surveillance involving practicing physicians, pharmacists, nurses and regulatory agencies, as well as the pharmaceutical industry. These communication systems could contribute to raise awareness in clinicians, trigger research initiatives to investigate potential mechanisms involved in the observed problem, and identify risk factors.

We must not forget that many treated patients receive a complex pharmacotherapy associated with multiple comorbidities. Therefore, they could display unpredictable reactions to antimicrobial agents. We should always expect the unexpected, especially when a new drug reaches the market. This may also prove useful for drugs already in use for some time, as recent adverse events reported with ceftriaxone point out. As new antimicrobial agents in clinical development are very few, it is our responsibility to make sure that the drugs we have in our armamentarium are used carefully to avoid a premature and unfortunate withdrawal and to ensure first and foremost that we are providing safe antimicrobial therapy to our patients.

Financial & competing interests disclosure

L Bergeron served on advisory panels for Pfizer, Bristol-Myers Squibb, Wyeth and Oryx Pharmaceuticals. The author has no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

No writing assistance was utilized in the production of this manuscript.