Clinical Pharmacology and the Best Use of Public Research Investment

In our May editorial, editor-in-chief Piet van der Graaf reviewed predicting success in industrial drug development1 and pointed to Hampson’s2 and Lalonde’s3 thoughtful explorations of optimized decision making in this domain. Those involved in generating new therapeutics understand the path involves (i) understanding the pathophysiology and the biology of the system under study, and (ii) then expanding this understanding by modulation of drug targets. There are many steps between discovery to drug approval, and even more to later use optimization in practice. Our successes in new targeted therapeutics are built largely upon prior public investment in basic biomedical research (the discovery phase). The striking roll out of coronavirus disease 2019 (COVID-19) mRNA vaccines is but one recent example.4 The discovery-to-development track that leads to new patent protected (and expensive) medications is the domain of the pharmaceutical industry. The early phase development involves relatively modest costs and few patients. The large expenditures in dollars, patient numbers, and opportunity costs occur in phase III. This has led to the concept of kill fast, with early identification of compounds with low potential for success. Another path to new therapeutics involves the repurposing of already approved drugs, applied to new indications. This activity takes place primarily in academic medical centers and is sponsored by governmental agencies or foundations. Perhaps the most successful example is the use of humble aspirin,5 which remains the bedrock of secondary prevention of ischemic arterial events and for prevention of pre-eclampsia. In this issue of Clinical Pharmacology and Therapeutics, we examine the use of N-acetylcysteine to prevent iodinated radiographic contrast-induced renal dysfunction (Figure 1). Sandilands6 describes a group of four soundly designed and well-conducted experimental medicine studies investigating the mechanistic basis of this hypothesis. They measured pharmacokinetics of oral and i.v. N-acetylcysteine in healthy volunteers and those with renal insufficiency and paired these parameters with biomarkers of renal blood flow, renal injury, and plasma antioxidants. Oral N-acetylcysteine generated only trace systemic concentrations and no measurable physiologic effect. Intravenous N-acetylcysteine was noted to increase renal perfusion in healthy volunteers, but not in those with renal dysfunction – the only group that would benefit from the treatment. There were no appreciable changes in biomarkers of antioxidant effect, the other mechanism by which N-acetylcysteine was proposed to protect kidneys in the face of contrast exposure. In the kill fast paradigm, their studies would have led to the termination of a research program examining N-acetylcysteine for the prevention of contrast-induced nephropathy. There would be no phase III trial for N-acetylcysteine. However, this is not how things proceeded in real life. Instead, the actual story began in 2000 with a small efficacy trial in which a post-contrast rise in creatinine was seen in 9 of 42 patients with placebo compared to 1 in 41 receiving oral N-acetylcysteine.7 This publication kicked off a series of clinical trials, society recommendations, and enthusiastic use in clinical practice. The Sandilands study reported this month was first registered in ClinicalTrials.gov in 2007, with the filing noting a 2006 consensus working group conclusion that “N-acetylcysteine is not consistently effective in reducing the risk for contrast-induced nephropathy.”8 In 2008, Stenstrom9 pointed out the extremely low bioavailability of oral N-acetylcysteine (a property known for years)10 and warned against “premature phase III trials” before there was an adequate exploration drug exposure, dose, and response. Fast forward to the primary completion of the Sandilands studies listed as January 2015. Publication was not until 2022. The delay in publication is not ideal, although it is unlikely a more rapid communication would have made a difference in the route the N-acetylcysteine story took. The potential use of oral N-acetylcysteine was dimmed by the lack of efficacy in the 2010 ACT trial (2,308 patients) and settled by the null 4,993 patient PRESERVE trial (described in 2013,11 terminated early for futility, and published fully in 201812). In the Sandilands trial 4 (renal dysfunction cohort administered standard of care i.v. contrast), there were 6 cases of patients who had elevated creatinine described as contrast-induced nephropathy. None had elevations in acute kidney injury biomarkers. There is a temporal relationship to arterial i.v. contrast, although angiography is the procedure that involves more than just i.v. contrast administration. Regardless of the causal link in this case, epidemiologic studies indicate that potential risk with venous administration for computed tomography is much weaker. Indeed, the American College of Radiology Manual on Contrast Media preferred description was changed in 2015 from contrast-induced kidney injury to contrast-associated nephropathy, as varied evidence streams brought into question a direct linkage. This further supports the premise that the epidemiology and biology of the disease – acute kidney injury (AKI) around the time of contrast rather than due to contrast – was not fully explored before a series of phase III trials were embarked upon. Have we been here before? Small clinical trials in 1996 (n = 200)13 and 1999 (n = 112)14 reported large effects from the peri-operative use of beta-adrenergic blockage to improve outcomes in non-cardiac surgery. These lit a 15-year cycle of enthusiastic endorsement of the approach in society guidelines, which was softened over time by the weight of a series of null or negative randomized trials eventually enrolling 19,191 patients.15 Sandilands posits that funding agencies have learned from prior experiences such as these and suggests this is less likely to happen today. This is likely true, although repeating patterns show up, such as the recent burst of excitement around the use of vitamin C in critically ill patients based upon large effect studies containing tens of patients. Subsequent clinical trials published in 202016, 17 and 202118 showing no effect have dampened enthusiasm. Use of vitamin C in COVID has similarly failed to demonstrate efficacy.19 Whereas the order of investigation of vitamin C in sepsis was still backward – phase III before phase I – the positive outcome in this example is that a consensus conclusion has been reached with hundreds rather than thousands of patients. The experience of COVID, in which there was a rush to try anything, has also demonstrated that we also will likely continue to see small, underpowered, and locally funded studies going directly into attempts of efficacy demonstration. The argument, it is suspected, is that during a public health emergency, it is “worth trying anything.” A search of COVID in clinicaltrials.gov supports this guess. However, underpowered COVID trials without the demonstration of biologic plausibility, or basic principles of pharmaceutics and optimized dose, is not just a distraction but eats resources that could be used to a more systematic testing of potential treatments. We cannot prevent poorly designed studies from being conducted, but we should raise the bar for taking these as a demonstration of efficacy. What lessons can be taken from this study to the larger enterprise of drug repurposing? The largest is that, in the past, there has been public investment in later phase clinical trials that perhaps has been not well spent as it could have. The tools of clinical pharmacology can optimize the use of public research funding. There is also an ethical aspect, as we strive to minimize the number of patients in clinical trials that are unlikely to be successful. The demonstration of a sequential approach to drug development starting with exposure/response characterization and proof of concept has proved itself over the past 50 years with wildly successful improvements in human health due to new drugs. The experience with N-acetylcysteine is the counter example of the dangers of hoping for a hunch to be true, and luck to be on one’s side. In addition, there are also the uncounted resources wasted in actual practice of the use of N-acetylcysteine for years in patients with no benefit. Large studies are not conducted without the sponsorship of large organizations. Funding agencies should require phase III studies to have proof of concept and pharmacokinetic plausibility demonstrated before moving on to exposure in large patient populations. This may involve a more purposeful attempt to fill the gap of funding experimental medicine and phase I trials for repurposed drugs. This can be done on the programmatic level, but also in the selection of reviewers across the spectrum of discovery, development, and utilization of drugs. It is likely that this would also end up being a cost-effective investment. Improved and expanded training of investigators into clinical trial theory at academic institutions likely is of benefit, as is participation of academic researchers in industrial clinical trials. Investigators who understand modern drug development are less likely to imagine a world in which we move quickly to phase III. Improvements in healthcare digitization and availability of medical records for investigation can also provide guidance. These tools will assist in providing evidentiary weight for drugs already in use and to characterize the natural history of disease and response to treatments. Real-world evidence will not provide definitive answers. However, it is another source of information to assess if a hypothesis has strong enough evidence warranting extensive efficacy testing in patient trials. Clinical pharmacology has a key role to play in the attempts at repurposing. A prescient article early in the pandemic outlined principles of optimizing therapeutics against COVID-19.20 The concepts remain remarkably accurate 2 years on, which speaks to the power of drug development science our discipline endorses. Last, journals have a role as peer review gatekeepers. The editors of CPT turned away provocative COVID-related papers that subsequently were cited often but had unsupported scientific assertions that did not meet the standards of the journal. Most articles will eventually find some home but keeping standards high and examining a totality of evidence avoids providing a high impact platform for studies that may not warrant it. In summary, Sandilands provides us with a beautiful demonstration of a well-run academic series of early phase studies. These are not easy or cheap, but much cheaper than a series of negative phase III efficacy trials. The use of N-acetylcysteine in practice and clinical guidelines is settled, and the publication will not change this. However, this paper does point to how such questions can be asked and answered in an academic setting. This hopefully will reduce the number of future articles that suggests the lazy ending line of “future research must be done” to assess the efficacy of N-acetylcysteine. More importantly, it can provide funding agencies a case study when thinking about funding priorities. The author declared no competing interests for this work.