Clinical Research in the Lay Press: Irresponsible Journalism Raises a Huge Dose of Doubt

In an era in which science and medicine make front-page news in the lay press, it is critical that the complex workings of clinical investigation be portrayed accurately to the public. The appetite for news of medical “breakthroughs” seems insatiable at times. In this setting, sensational articles about medicine, physicians, and pharmaceutical companies can easily find an attentive audience that may be unable to distinguish truth from sensationalism. To provide a case study for how inaccurate and dangerous the mainstream press can be if articles are not carefully written, as well as to correct inaccuracies and defend honesty in research, we offer our counterpoint to a recent article [1] that questions the various systems of checks and balances that govern the conduct of clinical trials and implicitly accuses one of our infectious diseases colleagues of unethical conduct in 2 clinical trials. In “A Times Investigation: Drug Trials with a Dose of Doubt” [1], an article written by reporter David Willman and published in the Los Angeles Times, Dr. Thomas Walsh, the Chief of the Immunocompromised Host Section, Pediatric Oncology Branch, National Cancer Institute, National Institutes of Health (NIH), was implicitly accused of unethical conduct in 2 randomized, controlled trials of empirical antifungal therapy for persistent neutropenic fever. Walsh was the lead author of both of these multicenter trials, both of which were published in the New England Journal of Medicine [2, 3]. The first study compared conventional deoxycholate amphotericin B (D-AmB) with liposomal amphotericin B (L-AmB), and the second compared L-AmB with caspofungin. Specifically, Walsh, the principal investigator, was implicitly accused of deliberately underdosing the standard therapy drug to favor the investigational agent of the pharmaceutical sponsor. We use the term “implicitly accused,” because Willman is careful to not make explicit accusations of wrongdoing. However, when all of the misleading statements, nonsequiturs, loaded and pejorative descriptions, and selected quotations are strung together, Willman creates the image of a respected NIH scientist and pharmaceutical companies colluding to rig trials to win US Food and Drug Administration (FDA) approval for the favored drug (table 1). Willman implies that the biased trial design jeopardized patient safety because patients with life-threatening infections were treated with inadequate doses of antifungals. Implicit in these accusations is the contention that experienced physicians at multiple levels of oversight—including physicians at the NIH and FDA, site investigators and members of institutional review boards at dozens of health care centers, and the members of the New England Journal of Medicine editorial boards—either actively colluded in this conspiracy or had such poor knowledge about empirical antifungal therapy that they did not realize that a conspiracy had occurred. Although the title “Drug Trials with a Dose of Doubt” is an attention grabber that sells newspapers, an accurate representation of the facts will show that these trials were conducted with sterling integrity. Statements made in the Willman article [1] and our rebuttal. Willman opens his “investigation” with the implicit accusation that Walsh had extraordinary influence over the FDA's approval of caspofungin. Caspofungin was initially approved as therapy for invasive aspergillosis in patients with refractory infection or intolerance to standard therapy. Willman correctly notes that the database involved a limited number of patients. However, this drug, representing a new class, was free of serious toxicity and had encouraging results in treating this life-threatening infection [5]. The FDA weighed the limited but supportive database, the acute, unmet need for effective therapeutics against aspergillosis, and the recommendations of its 12 independent advisory board members, and arrived at the very reasonable decision to approve caspofungin. Yet, Willman gives the false impression that Walsh had a singular influence over the FDA's decision. “Merck summoned to the microphone one of its announced consultants, a man whose government job was nearby, at the NIH. Dr. Thomas J. Walsh assured the committee that Merck's data describing the patients was `extremely robust and very, very rigorous.'.... The advisory committee voted unanimously to endorse the drug... Sixteen days later, the FDA approved it” [1]. On the basis of Willman's remarks, an intelligent lay person may question whether the FDA actually looked at the data or whether Walsh's remarks at the microphone were all the assurances that the FDA needed. The absurdity of this scenario is stunning, and yet Willman supports this allegation with a series of non sequiturs. Because Walsh and the FDA physicians are federal employees, Willman implies that Walsh must be in a unique position to influence FDA decisions and to single-handedly attain FDA approval of major drugs. The NIH and the FDA are entirely separate federal administrations with distinct missions and oversight. What is the evidence for Walsh's purported excessive influence over the FDA? None exists. Unfortunately, Willman was just getting started. At the heart of the Willman “exposé” is the contention (although it is never explicitly stated) that the 2 clinical trials were rigged to increase the likelihood of FDA approval of the investigational drugs. Willman's evidence consists of selective quotations from letters to the editor that raised concerns about the dose of standard drug used in the trials. Willman's use of these selective quotations is antithetical to the rigorous, unbiased science that he claims to defend. Letters to the editor serve as a forum for debating points in published material and are typically critical. Some letters make cogent arguments. Others do not. More importantly, letters to the editor are typically authored by one or a few physicians, do not reflect broad consensus, and, unlike the article that they critique, are not subject to rigorous peer review. The web of deceit that is implicitly alleged in this conspiracy theory is intricate but centers on 2 issues: selective enrollment of patients and picking dosages of antifungals to bias outcome. The first layer of purported deceit implicitly alleged by Willman's article [1] involved the selection of the type of trial that would have the highest likelihood of securing FDA approval. Willman states that, “for makers of new antifungal drugs, less burdensome clinical study standards could make it easier to get the products approved... for instance some companies wanted to enroll cancer patients with suspected but unproven fungal infections,” thereby implying that studies of empirical therapy are not scientifically valid and are only designed for cherry-picked patients. We disagree. The rationale for empirical therapy is to treat a potential occult invasive fungal infection before it becomes clinically overt. This concept is based on the central tenet that early treatment of invasive fungal infections improves outcome and that lower doses, when used early, may benefit patients—particularly those patients with infections that are difficult to diagnose (such as invasive fungal infections). Before the development of empirical therapy, unsuspected infections with Candida species, Aspergillus species, and other fungi were frequently diagnosed at autopsy [6, 7]. Empirical antifungal therapy, as conducted in the 2 trials at issue, has been studied in >3000 patients and has been endorsed in authoritative guidelines from infectious diseases and hematology professional societies in North America and Europe [8, 9]. At the time that the 2 trials in question were designed, empirical antifungal therapy was a bedrock principle and standard-of-care for patients at risk for suspected fungal infections and was, therefore, a legitimate subject of clinical investigation. The recent availability of safe antifungal agents and improved diagnostic tools has opened a scientific debate about the current role of empirical therapy, with some investigators advocating different approaches [10]. These and other differences of opinion represent the scientific debate that is part and parcel of any scientific field. The second and more serious of Willman's implicit accusations is that Walsh deliberately chose to administer lower, less-effective dosages of comparator drugs in the 2 trials. In trial 1, D-AmB (0.6 mg/kg per day) was compared with L-AmB (3 mg/kg per day). In trial 2, L-AmB (3 mg/kg per day) was compared with caspofungin (70 mg administered once, followed by a regimen of 50 mg per day). Willman [1] implicitly alleges that D-AmB (the control drug) was underdosed in trial 1. He further suggests that, in trial 2, it was L-AmB (now the control drug), that was underdosed to favor caspofungin. That the dose of L-AmB was the same in both trials makes these allegations self-contradictory and logically untenable. In fact, the dosages of all drugs used in both trials were appropriate on the basis of substantial published data. Consensus supporting the 0.6 mg/kg per day D-AmB dosage in trial 1 [2] and the 3 mg/kg per day L-AmB dosage in trial 2 [3] is based on the following data. First, D-AmB was administered at a dosage of 0.5–0.6 mg/kg per day in prior studies of empirical therapy that established the safety of this approach and suggested a protective benefit [6, 11]. Second, no evidence of superior outcomes associated with higher dosages of D-AmB or L-AmB has ever been published. In a randomized, controlled trial of empirical therapy that compared D-AmB 1 mg/kg per day versus L-AmB 1 mg/kg per day versus L-AmB 3 mg/kg per day, D-AmB recipients had a response rate comparable to that of patients treated with L-AmB, but they experienced greater nephrotoxicity (23%) than did patients receiving L-AmB 1 mg/kg per day (0%) or L-AmB 3 mg/kg per day (3%) (P =.01) [12]. These results were similar to those for trial 1 (D-AmB vs. L-AmB) [2], which showed comparable efficacy but higher nephrotoxicity in the D-AmB 0.6 mg/kg per day group than in the L-AmB 3 mg/kg per day group. In another study of empirical antifungal therapy, L-AmB administered at a dosage of 3 mg/kg per day and 5 mg/kg per day had similar efficacy and toxicity [13]. Finally, a recent trial of primary therapy for invasive aspergillosis (the AmBiLoad study [4]) showed that L-AmB (3 mg/kg per day, which is the same dosage that was administered in the empirical trials) was equally effective but less toxic than a 10 mg/kg per day regimen of L-AmB. Yet, Willman [1] creates the misleading impression that patients enrolled in these empirical trials may have died of breakthrough aspergillosis because of inadequate dosing of D-AmB or L-AmB. The results discussed above, particularly the findings of the AmBiLoad study, dispel the false notion that administering higher dosages of drug is equated with improved efficacy, and they unequivocally give further validation of the dosages used in the Walsh trials in question [2, 3]. Third, there is evidence of increasing dose-dependent toxicity with D-AmB. Even a superficial review of the literature would find multiple reports of high rates of dose-limiting nephrotoxicity associated with D-AmB use [4, 12, 14–18]. D-AmB-related nephrotoxicity has been shown to be an independent risk factor for mortality [14]. Yet, Willman [1] chose to ignore these well-documented, substantial patient safety concerns in his discussion of the D-AmB versus L-AmB empirical trial. The majority of patients who receive empirical antifungal therapy do not have an occult fungal infection. Therefore, this approach necessarily entails treating many individuals to benefit a minority of patients. It is, therefore, of key importance that the regimen be safe. Willman's contention that patients were put at risk by unethical trial design flies in the face of his article's [1] total disregard of the inherent toxicities in AmB-based antifungal therapy mentioned above [4, 12, 14–16, 18]. Furthermore, the dosages of D-AmB and L-AmB in trial 1 [2] were agreed upon by all 32 investigators and by senior members and statisticians of the Mycoses Study Group and were approved by participating health care center institutional review boards, the National Institute of Allergy and Infectious Disease protocol review committee, and the FDA. Patient safety measures were stringent and relied on baseline evaluation to exclude invasive fungal infection, included monitoring for breakthrough invasive fungal infection during therapy, allowed for dosage modification (with dosage to be increased if invasive fungal infection was suspected and decreased in response to toxicity), and included prospective data review by an independent data safety monitoring board. D-AmB recipients had significantly more frequent dose reductions because of toxicity than did L-AmB recipients. This finding the theory of underdosing of D-AmB. the in the number of associated with therapy, in Willman's article [1], was, in fact, not in this of patients Willman [1] further suggests a disregard for patient safety by that the L-AmB dosage in the L-AmB versus caspofungin trial [3] could not be increased a patient had 5 days of the dosage of investigational or comparator drugs and to The is that the and health were by on treatment the However, Willman that “A patient could be from the study and treated and that was to the for increasing the and approved by all institutional review boards, and the [1]. That is, the study design investigators to as and to on the of patient if a patient from the because patient safety was at the heart of the The accusation that the L-AmB dosage was in trial 2 [3] not in trial 1, which used the same dosage is with evidence that the 3 mg/kg per day L-AmB dosage is for empirical therapy. The implicit accusation a of knowledge of oversight. On the basis of the results of trial 1 [2], L-AmB was approved by the FDA as empirical therapy for neutropenic at a dosage of 3 mg/kg per The FDA the use of control when new such as caspofungin. Therefore, the use of L-AmB at 3 mg/kg per day in trial 2 was and for approval. Although Walsh was an in to dosage selection a by a group of investigators who had to approve the study design and of approval were to the a major in the of the of drug approval has been the of checks and These same by and institutional review an part of our approach to clinical research, for all of the drug development as it this the standard for drug Willman [1] in that dozens of investigators and actively with Walsh to patients patient or were that a conspiracy was this is the of of medical in the and is or In he not explicitly Willman [1], in implicitly the NIH with the pharmaceutical and for the of dozens of senior Mycoses Study Group members, and FDA patient enrollment in unethical trials with the institutional review boards at the study the editorial board at the New England Journal of Medicine of unethical and the FDA and its advisory board drugs on the basis of all of these systems of independent to have for the conspiracy Walsh and the pharmaceutical to be We Willman's of Walsh's is no for an or to design a trial with a control that is not standard of [1]. We one We it is to conduct a study dozens of health care centers if the control not to a of is to be standard of Willman [1] further Walsh on the of his to In Willman the major of principal investigators and the data review committee with other for study and as well as for results to the It is entirely appropriate for a principal and data review committee to provide to both the pharmaceutical and the FDA, particularly when he to Walsh, an with scientific and on scientific advisory a federal not Walsh from such he is as to do as any other with similar is, a very of individuals who can such to both the pharmaceutical and the FDA. is critical to the scientific the pharmaceutical and the scientific in this time of decreased federal In fact, all drugs are to the pharmaceutical independent some whose is and the FDA. The federal and pharmaceutical of the Mycoses Study Group is but one of that had to major in antifungal therapy. a of This not with both of an It must give some of the of the evidence supporting to give some to and to to it to to a review of any article for the of the evidence and the of the article's The Los Angeles Times has its on all The of Willman's implicit allegations and the made by investigators and by Walsh's before the of Willman's article [1] have Willman's to carefully the of his to do question the of the Los Angeles Times as a serious the with for this at and the of this give the a dose of about this The of Willman's accusations against dosages of antifungals were to bias the 2 empirical trials in favor of the investigational been in the another is the of during the conduct of the that to the inaccurate article [1]. Willman his of Walsh in with accusations Walsh, of Willman's accusations in a response in This response did not but to another in which more response was to Willman and his in that further the of his was and Willman's article [1] was published in of Willman's are in in his by such as the dosage and of therapy for patients with suspected published study has established that a higher dose of an antifungal drug is more effective in treating suspected infection and that some studies have suggested that lower dosing may provide similar dosages were not by Walsh but with of other and, were and approved by the FDA and institutional review boards of participating [1]. would that these statements, in and of would Walsh of with to the However, by the Willman on less authoritative such as and selective quotations from the of medical to the implicit allegations of on a respected and government on the and to the lay that individuals enroll in clinical trials at may be to more for this Walsh as a colleagues of Dr. Walsh, we are that his is To his patients and Dr. Walsh is a whose to the care of and is central to his professional To his colleagues Dr. Walsh is an who has the of antifungal therapy. This is the of Dr. The of articles such as Willman's [1] is that members of the lay do not medical the Los Angeles Times is is and is as an authoritative news there is to that the will on the basis of Willman's article [1], that the of drug development in the and is and that they from participating in clinical trials. We question whether Willman and the Los Angeles Times the that patients as a of Willman's clinical who the of the we and by the lay we have the to defend our colleagues when professional has been and to in clinical and its systems of independent when they have been of has from and is a for and and is a of the for and has from is a for and and is a of the for and has from and is a for and and is a of the for and is a in has from and and is a for and has from and is a of the for and has from and has from and and is a for and has from and is a for and and is a of the for and has from and is a for and and is a of the for and has from and is a for and is a of the for and has from and is a for and and is a of the for and is a for and is a of the for and and is a in is a of the for and has from and and is a for and has from and is a for and and is a of the for and has from and and is a of the for and has from and has from the Institute of is a of the for has from and is a for and and is a of the for and has from and has from and is a for and is a of the for and has from and and is a of the for and has from and has from and is a for and is a of the for the Institute of Allergy and Infectious Institutes of Health has from and is a for and and is a of the for and has from and and is a of the for and has from and is a for and and is a of the for and has from and is a for and and is a of the for has from and has from and is a for and is a of the for has from and is a for and and is a of the for and has from is a for and and is a of the for and has from is a for and and is a of the for and has from and is a for and and is a of the for and is associated with has from and and is a of the for and is a for and and is a of the for and has from and has from and is a for and and is a of the for and has from and has from and has from and has from and is a for and and is a of the for and is a for and a of the for has from and is a for and and is a of the for and has from and and is a of the for and is a of the for has from and is a for and and is a of the for and has from and is a for and and is a of the for and has from and is a for and is a of the for and and is a in has from and is a for and and is a of the for and has from has from and is a for and and is a of the for and has from and and is a of the for and has from and and is a for and has from is a for and and is a of the for and has from and and is a of the for and has from and is a for and is a of the for and and is a in is a for and is a of the for and has from and has from and and is a of the for and has from and has from and is a for and is a of the for is a of the for is a for and a in and has from and is a for and and is a of the for and has from and has from and and is a for and other no

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.