Anti‐inflammatory agents in atherosclerosis—and a need for reform: Extraordinary claims require extraordinary evidence

Abstract

Since 1858, human atherosclerotic plaques have been shown to contain immune cells. But are these cells suitable therapeutic targets? Dozens of clinical trials of anti‐inflammatory agents other than colchicine have been performed in patients with clinically evident atherosclerosis. None of these trials led any regulatory body in any jurisdiction to allow a cardiovascular indication for any of these agents. This discouraging work provides a background to evaluate new data on colchicine. In 2024–2025, new clinical trials of colchicine in patients with atherosclerosis showed benefit, no benefit, or harm. In 2025–2026, over a dozen meta‐analyses so far have appeared, drawing on ∼34 trials, but do not agree with each other. One of these meta‐analyses, Xie et al.’s in the Journal of Internal Medicine, provides a compelling graphical display of the pre‐specified primary outcomes of six long‐term clinical trials as they were published over time. The pattern of early positive trials, then newer negative (null) trials, suggests regression to the truth. Jeon and Cho et al.’s population‐wide study in the Journal of Internal Medicine of patients with Type 2 diabetes and gout found no benefit from colchicine over nonsteroidal anti‐inflammatory drugs on major adverse cardiovascular events. This information suggests several areas for reform of research on inflammation in atherosclerosis. Fully informed consent should require that clinical trials of anti‐inflammatory agents in atherosclerotic arterial disease must disclose to trial participants the failures of this approach in over 50 clinical trials to date, spanning decades. Additionally, the field might reconsider its commitment to the Big Idea that inflammation must be a definitive therapeutic target in atherosclerosis. The track record so far for inflammation inhibition in atherosclerosis is extensive and disappointing—a fact that merits wider discussion. Therapeutic successes from targeting cholesterol‐rich, apolipoprotein‐B‐containing lipoproteins—the proven causative agents of this disease—provide extraordinary evidence. Current data for colchicine and other anti‐inflammatory therapies do not.

“Extraordinary claims require extraordinary evidence.” The astronomer and science communicator Dr. Carl Sagan popularized the concept, most famously in the way he evaluated reported sightings of extraterrestrials [1]. The assertion has a long intellectual history that includes the mathematicians Thomas Bayes and Pierre‐Simon Laplace. What it takes to convince a person depends on what that person knows already [2, 3].

Is the claim that colchicine helps patients with advanced atherosclerosis extraordinary? And if so, is the evidence so far regarding this claim extraordinarily strong? Two articles [4, 5] in the December 2025 issue of the Journal of Internal Medicine advance this discussion, particularly regarding the second question.

An extraordinary claim?

To begin with, any claim of therapeutic benefit is extraordinary in the sense that patients’ lives and well‐being are at stake. The burden of proof—the onus—rests entirely on the person, commercial concern, or regulatory body that is making claims of benefit. The rest of us are not responsible for verifying, or disproving, every assertion out there in the ongoing blizzard of health claims.

Some skepticism is always warranted, but especially when money, professional recognition, prominent intellectual commitments to a particular approach, and individual or institutional prestige are involved. These items constitute financial [6, 7, 8] and nonfinancial [2, 9, 10] conflicts of interest (COIs).

Regarding probabilities, colchicine is a first‐in‐class agent to achieve regulatory approval for the treatment of patients with advanced atherosclerotic arterial disease. Despite over a century and a half of research into inflammation and atherosclerosis, colchicine remains the sole anti‐inflammatory medication with an indication for this disease [11]. These facts are extraordinary.

For comparison, a steady effort—starting most convincingly with the large, long‐term Scandinavian Simvastatin Survival Study (4S) [12, 13]—gradually persuaded our field that an entire class of medications—statins—substantially lowers the rates of atherosclerotic cardiovascular events and, in 4S, improves patient survival. Statins—and then the non‐statins ezetimibe, PSCK9 inhibitors, and bempedoic acid—were shown to reduce event rates directly in relation to their ability to lower plasma concentrations of cholesterol‐rich apolipoprotein‐B (apoB)‐containing lipoproteins, primarily low‐density lipoprotein (LDL) [13, 14, 15, 16, 17, 18, 19, 20], now proven as the causative agents of this disease [21, 22, 23, 24]. Importantly, long before 4S, the new pharmacologic approaches to LDL lowering had already accumulated a wealth of supportive basic, epidemiologic, clinical, and genetic data [21, 22, 23, 24, 25]. In other words, the huge body of successful clinical trials of these LDL‐lowering agents provides us with extraordinary evidence to support a claim that is not so extraordinary because it fits with extensive prior knowledge. Practitioners in the field of preventive cardiology became convinced that these numerous different medications to lower plasma LDL levels provide cardiovascular benefit.

What is known about anti‐inflammatory compounds in the treatment of atherosclerotic arterial disease? There are at least 168 years of histologic and other basic‐science studies on the immune system and atherosclerosis [6, 11, 26]. Those studies motivated—and have now been largely superseded by—human clinical data, the most important of which came from an extensive body of prospective randomized clinical trials of anti‐inflammatory interventions in patients with this disease. Table 1 summarizes the most advanced of these trials—meaning Phase 3 when available—of immunosuppressives other than colchicine in patients with atherosclerotic arterial disease [11, 27, 28, 29, 30, 31]. In Table 1, there are 23 such clinical trials. The actual number is higher because the list in Table 1 (a) omits Phase 1 and 2 trials whenever a Phase 3 trial was available; (b) omits studies beginning in the 1960s showing a lack of benefit, or even harm, from two key classes of anti‐inflammatory medications—corticosteroids and nonsteroidal anti‐inflammatory drugs (NSAIDs)—in patients with atherosclerotic arterial disease [32, 33, 34, 35, 36]; and (c) might not be comprehensive despite best efforts. Such a large number of clinical trials in the same area casts doubt on the validity of the commonly used p‐value threshold of 0.05. At that threshold, we should expect that approximately one trial in 20 would falsely appear positive [2, 37].

Table 1.

Prospective randomized clinical trials of anti‐inflammatory agents other than colchicine in patients with atherosclerotic arterial disease.

Agent	Inflammatory targets	Overview of clinical trials	Outcomes (journals and years published)	Comments
Pexelizumab	Complement factor C5	Three Phase 3 RCTs: one in STEMI‐PCI, two in CABG	All negative (JAMA 2004, JAMA 2007, J Thorac Cardiovasc Surg 2011)
Veliflapon (aka DG‐031)	5‐Lipoxygenase–activating protein (FLAP)	Prospective randomized Phase 2 biomarker trial in MI patients who carry at‐risk variants in the FLAP gene or in the leukotriene A4 hydrolase gene	Positive, but based on biomarkers only (JAMA 2005)	Clinical development of this compound has been terminated [30]
Succinobucol (aka AGI‐1067)	‐TNFα‐inducible expression of VCAM1, MCP1, and E‐selectin in endothelial cells [38]; ‐ Vcam1 and Mcp1 mRNA levels in mice [38]; ‐Circulating biomarkers of inflammation such as myeloperoxidase, white blood cell count, and fibrinogen in humans [39]	Phase 3 RCT in patients with recent ACS	No effect on the primary endpoint of cardiovascular events (Lancet 2008 [39])	The name given to this trial was highly positive‐sounding: “Aggressive Reduction of Inflammation Stops Events (ARISE)” [39]. See the main text for literature on potential ethical issues with such trial names
Atreleuton (aka VIA‐2291)	Arachidonate 5‐lipoxygenase (5‐LOX)	Two Phase 2 trials and one post hoc analysis, together spanning a decade (2010–2020); imaging and biomarker endpoints only	‐ Circ Cardiovasc Imaging 2010 (met 1° endpoint of a reduction in whole‐blood stimulated leukotriene LTB4 at trough drug level in post‐ACS patients; favorable preliminary data from the coronary CTA substudy); ‐ Atherosclerosis 2015 (no significant difference in inflammation in index artery by FDG‐PET); ‐ J Cardiovasc Comput Tomogr 2020 (post hoc analysis of the 2010 CTA substudy, now to assess epicardial fat volume, which decreased)	Clinical development of this compound has been terminated [30]
Setileuton (aka MK‐0633)	Arachidonate 5‐lipoxygenase (5‐LOX)	Phase 3 trial in myocardial infarction suspended (2006)	Information available in review articles	Clinical development of this compound has been terminated [30]
Inclacumab	P‐selectin, a key mediator of “thrombo‐inflammatory reactions”	Two Phase 2 RCTs: one in NSTEMI, the other in CABG	‐The NSTEMI trial met its pre‐specified 1° endpoint (JACC 2013), but inclacumab was not pursued in a Phase 3 CV trial, ‐The CABG trial was negative (JACC 2016)	No regulatory approval for a CV indication
Varespladib	Secretory phospholipase A2 (sPLA2) This inhibitor, varespladib, had “favorable effects on lipid and inflammatory markers”	Phase 3 RCT in ACS	‐NS trend to harm for the composite 1° endpoint ‐A highly significantly greater rate of MIs ‐Terminated early by the DSMB for “futility and possible harm” (JAMA 2014)	Well‐known authors
Darapladib	Lipoprotein‐associated phospholipase A2 (Lp‐PLA2): “In atherosclerotic plaques, lipoprotein‐associated phospholipase A2 increases the production of proinflammatory and proapoptotic mediators”	Two RCTs: one in stable CAD and the other in ACS (STEMI or NSTEMI)	Both negative (NEJM 2014, JAMA 2014)	‐Well‐known authors ‐The name given to the trial published in the NEJM was highly positive‐sounding: “Stabilization of Atherosclerotic Plaque by Initiation of Darapladib Therapy (STABILITY).” See the main text for literature on potential ethical issues with positive‐sounding trial names
Salsalate (a prodrug dimer of salicylates)	“[M]ultiple potential pathways for the anti‐inflammatory effects of salicylates have been identified” in part through NF‐κB inhibition.	RCT in overweight and obese statin‐using patients with clinically evident stable CAD. The pre‐specified 1° endpoint was progression of non‐calcified plaque volume by CTA.	Negative study, but successful, specific anti‐inflammatory effects: “decreased total white blood cell, lymphocyte, monocyte, and neutrophil counts and increased adiponectin levels without change in C‐reactive protein levels” (JAMA Cardiol 2016)	“Urinary albumin levels increased, with tinnitus and atrial arrhythmias more common, in the salsalate group compared with the placebo group”
Vitamin D	Number and function of Treg cells	Prospective randomized trial of 3000 IU of vitamin D vs. no supplementation, in patients aged 30–80 years, with medically treated CAD. Pre‐specified 1° outcome was the increase in Treg cells.	The vitamin D trial “was terminated owing to slow recruitment of participants.” [40]	Treg cells are a high‐profile target [41].
Low‐dose IL2	Number and function of Treg cells	Prospective, randomized, double‐blinded, placebo‐controlled, Phase 2 IVORY trial of low‐dose IL2 in patients presenting with ACS and serum CRP levels >2 mg/L	‐The IVORY trial met its pre‐specified 1° outcome of lower arterial inflammation by FDG‐PET/CT in the IL2‐treated group compared with placebo‐treated controls. Arterial inflammation was 7.7% lower, p = 0.015 (Nat Med 2026 [42]). ‐IL2 increased Treg cell levels compared with placebo (2° outcome, p < 0.0001). ‐CRP levels fell in both treatment groups to a statistically indistinguishable extent (p = 0.6–0.9) ‐There was a possible early signal for fewer MACE after 2 years of IL2 treatment (zero IL2‐treated patients with events vs. three patients with events in the placebo group) [42]. The ongoing IVORY‐FINALE observational study is following patients who completed the IVORY trial, to assess their future MACE.	‐Treg cells are a high‐profile target [41] ‐Regarding the trial's inclusion criteria, see the main text for literature on the considerable proportion of patients who have low CRP levels yet persistent atherosclerotic events. In the IVORY trial, “Most [ACS] patients failed screening due to hsCRP levels <2 mg L−1, accounting for 65% of screen failures.” These individuals comprised 22.6% of the initial group of 106 ACS patients who underwent screening [42]
*Plozalizumab (aka MLN1202)	Binds CCR2 to block its interaction with the chemokine CCL2 (aka MCP1) to thereby inhibit monocyte trafficking into atherosclerotic plaques	One Phase 2 trial; biomarker endpoint only	Met 1° endpoint of a significant reduction in serum CRP levels, p = 0.027. ∼30%–40% of patients with a key SNP in the MCP1 promoter achieved CRP levels ≤3 mg/L on plozalizumab (Am J Cardiol 2011)	‐No regulatory approval for a CV indication ‐See the main text for literature on the considerable proportion of patients who have low CRP levels yet persistent atherosclerotic events
*Anakinra	IL‐1 receptor antagonist (IL1ra)	Phase 2 RCT in NSTEMI‐ACS	Met its 1° endpoint of reduction in the AUC for CRP over the first 7 days (Eur Heart J 2015)	‐“MACE at Day 30 and 3 months was similar but at 1 year there was a significant excess of events in the IL‐1ra group” ‐Small study (n = 182)
*Losmapimod	p38 MAPK: ‐“Activation of p38 MAPK leads to amplification of the inflammatory cascade through enhanced production of multiple cytokines, including tumor necrosis factor, interleukins 1 and 6, metalloproteinases, and cyclooxygenase 2” ‐This inhibitor, losmapimod, lowered CRP levels by 57% in hypercholesterolemic patients (Circ 2011). ‐The Phase 2 trial of losmapimod in NSTEMI found that this agent lowered CRP levels early (72 h) and BNP levels later (12 weeks; Lancet 2014).	Phase 3 RCT in AMI (NSTEMI or STEMI)	Negative study (JAMA 2016)
*Canakinumab	Interleukin‐1ß (IL1ß)	Phase 3 RCT in post‐ACS patients with CRP>2 mg/L (CANTOS)	Negative at the lowest and highest doses; statistically significant benefit only at the middle dose—and just barely (NEJM 2017)	‐No subsequent regulatory approval for a CV indication for this medication despite applications by Novartis to the FDA‐USA (rejected) and to the EMA (withdrawn right after the FDA‐USA's rejection); ‐No dose–response (in contrast with the dose–response for lung cancer); ‐Greater rates of fatal infections or sepsis; ‐Remarkably expensive medication; ‐Extensive funding
*Canakinumab	Interleukin‐1ß (IL1ß)	Patients with symptomatic peripheral artery disease	No benefit on the 1° endpoint of plaque progression by MRI of the superficial femoral artery (SFA) (Vasc Med 2019)	Again, no subsequent regulatory approval for a CV indication for this medication in any jurisdiction
*Methotrexate	Prior work by the same first author stated that “very‐low‐dose‐methotrexate (VLDM, 10 mg weekly) [is] a proven anti‐inflammatory regimen that reduces TNFα, IL‐6, and CRP levels” [43]	Large‐scale RCT	Negative study: “Among patients with stable atherosclerosis, low‐dose methotrexate did not reduce levels of interleukin‐1β, interleukin‐6, or C‐reactive protein and did not result in fewer cardiovascular events than placebo.” (NEJM 2019)	‐The name given to this trial, “Cardiovascular Inflammation Reduction Trial (CIRT)” overpromised the efficacy of methotrexate, which failed here to reduce inflammatory markers. See the main text for literature on potential ethical issues with positive‐sounding trial names ‐“Methotrexate was associated with elevations in liver‐enzyme levels, reductions in leukocyte counts and hematocrit levels, and a higher incidence of non–basal‐cell skin cancers than placebo.” ‐Generous public funding from the NIH‐USA
*Allopurinol	Inhibits “xanthine oxidase‐mediated vascular oxidative stress” thereby leading to “improvements in markers of oxidative stress and inflammatory response” including a drop in serum CRP levels in a prior study that was cited in the Introduction of the final report of the cardiovascular clinical trial	Large, randomized clinical trial in patients aged 60 years or older with ischemic heart disease but no history of gout	No benefit on the 1° composite three‐point cardiovascular outcome (Lancet 2022)	Generous public funding (UK National Institute for Health and Care Research)

Note: Here, the focus is the most advanced of these trials, meaning Phase 3 when available. The clinical trials of agents chosen to lower serum CRP levels, the key read‐out of the inflammasome‐IL1ß‐IL6‐CRP pathway, which is also targeted by colchicine, are grouped at the bottom of the table, and each of them is indicated with an asterisk (*). For further details, including additional clinical trials that were precursors to the ones listed here, as well as primary citations and proposed mechanisms of action, please see [11, 27, 28, 29, 30, 31]. Not included here are studies beginning in the 1960s showing a lack of benefit, or even harm, from two key classes of anti‐inflammatory medications, corticosteroids and NSAIDs, in patients with atherosclerotic arterial disease [32, 33, 34, 35, 36].

Abbreviations: ACS, acute coronary syndrome; aka, also known as; AMI, acute myocardial infarction; AUC, area under the curve; BNP, B‐type natriuretic peptide; CABG, coronary artery bypass graft; CAD, coronary artery disease; CC, C–C (Cys–Cys) motif; CCL2, CC‐chemokine ligand‐2 (aka MCP1); CCR2, CC chemokine receptor‐2; CRP, C‐reactive protein; CT, computed tomography; CTA, computed tomography angiography; CV, cardiovascular; DSMB, data and safety monitoring board; EMA, European Medicines Agency; FDA‐USA, United States Food and Drug Administration; FDG‐PET, ¹⁸fluorodeoxyglucose positron emission tomography; IL, interleukin; MACE, major adverse cardiovascular event; MAPK, mitogen‐activated protein kinase; MCP1, monocyte chemoattractant protein‐1 (aka CCL2); MI, myocardial infarction; MRI, magnetic resonance imaging; “negative” in common parlance means “null”; NIH‐USA, National Institutes of Health (USA); NS, non‐significant; NSAIDs, nonsteroidal anti‐inflammatory drugs; NSTEMI, non‐ST‐segment elevation MI; PCI, percutaneous coronary intervention; RCT, randomized controlled trial; SNP, single‐nucleotide polymorphism; STEMI, ST‐segment elevation MI; TNF, tumor necrosis factor; T_reg cells, regulatory T cells; VCAM1, vascular cell adhesion molecule‐1.

Strikingly, not a single agent listed in Table 1 achieved regulatory approval for use in patients with atherosclerosis. Over a decade ago, when two of these trials—CIRT of methotrexate and CANTOS of canakinumab—still required initial or ongoing funding, the principal author of both trials justified them by promising that the trials would “enroll more than 25,000 patients worldwide and provide a fundamental test of the inflammatory hypothesis of atherothrombosis” thereby “Closing the loop on inflammation and atherothrombosis” [44]. The CIRT and CANTOS trials would bring us “to a translational endgame testing whether inflammation inhibition will or will not have a major role in cardiovascular clinical practice.” [45]. “These two trials are the end game of testing these hypotheses.” [46].

The CIRT trial was negative (null)—no cardiovascular benefit from methotrexate (Table 1 and [47]). The CANTOS trial was negative at the highest and lowest doses of canakinumab, weakly statistically positive for cardiovascular benefit at only the middle dose (using a p‐value threshold of 0.05 shared among the three doses), with no evident dose–response (Table 1 and [48]). In contrast, the exploratory endpoint of lung cancer in CANTOS showed a highly statistically significant benefit, with a dose–response. Administration of canakinumab in CANTOS produced serious adverse side effects that included fatal infections and sepsis [48].

Using results from the CANTOS trial, Novartis—the developer of canakinumab—submitted applications to the United States Food and Drug Administration (FDA‐USA) and to the European Medicines Agency for a cardiovascular indication for their drug. As publicly predicted based on the weaknesses of the CANTOS results right after their release [49], the FDA‐USA ultimately rejected the Novartis application, and just a few weeks later, Novartis withdrew its European application [50]. Thus, the CIRT trial and the CANTOS trial failed to bring any immunosuppressive agent to any cardiovascular clinic in any jurisdiction (Table 1).

Yet no one speaks of an “endgame” for inflammation inhibition in atherosclerosis anymore. Intellectual commitment to this approach—namely, that inflammation must be a definitive therapeutic target in this disease—became paradoxically stronger after CIRT and especially CANTOS. References [51, 52, 53, 54] provide examples from opinion leaders in this part of our field. The newest of these references just appeared as the cover story in a prominent popular‐science journal, with a nearly identical title and highly similar content to what the same journal had published in another cover story a quarter‐century ago [54, 55].

Several of the negative trials in Table 1 had been given positive‐sounding names, hence potentially coercive to desperately ill patients and their physicians. For dozens of years, articles in the cardiovascular literature have criticized this ethically dubious practice [56, 57, 58]. As just one example, a negative trial published in 2008 in a prestigious high‐impact journal bore the name “Aggressive Reduction of Inflammation Stops Events (ARISE)” (Table 1 and [39]). It seems unintentionally reminiscent of a fictional satirical trial name from 2002: “A Surefire Cure for Cancer.” [56].

The repetitive failures summarized in Table 1 have now exceeded the unsuccessful history of antioxidant supplements in patients with atherosclerotic arterial disease. The field completed one dozen or so prospective randomized trials of antioxidants in atherosclerosis before becoming exhausted with the once wildly popular oxidative modification hypothesis. Decades of highly funded work on that hypothesis ended in ashes: no approved antioxidant medications in atherosclerosis and no approved diagnostic tests [24, 59, 60, 61, 62]. By comparison, Table 1—without colchicine, corticosteroids, or NSAIDs—lists just under two dozen failed clinical trials of anti‐inflammatory agents—approximately twice as many—and still counting.

The two theories are closely related: Many leading researchers have asserted that rancid lipids are a key factor in provoking inflammation in the atherosclerotic arterial wall [28, 38, 39, 52, 55, 63, 64, 65, 66, 67, 68]. Recently renewed interest in lipoprotein(a) (Lp[a])—a particularly atherogenic variant of LDL [69] that contains peroxidized lipids [70, 71]—seems unlikely to resurrect the oxidized LDL hypothesis. Lp(a) contains key biologically active lipids that are not oxidized—such as diacylglycerols and lysophosphatidic acid—that may be responsible for most of its unusual effects [67].

Many of the immunosuppressives listed in Table 1 caused harm. Much of the harm could be anticipated, such as increased fatal and serious non‐fatal infections (canakinumab); increased urinary albumin and tinnitus (salsalate); and elevations in liver‐enzyme levels and reductions in blood‐cell counts (methotrexate). But some harm was a surprise, such as increased atrial arrhythmias (salsalate), increased heart attacks (varespladib), and increased major adverse cardiovascular events (MACE, anakinra).

Many of the agents in Table 1 targeted different components of the immune system and so perhaps should not be lumped together. But many were chosen to inhibit a single well‐recognized pathway—namely, from the NLRP3 inflammasome to interleukin (IL)1ß to IL6 to C‐reactive protein (CRP) [31, 72]. It is the same pathway that colchicine is said to inhibit in this disease [31, 72, 73, 74, 75]. Clinical trials of agents chosen to lower serum CRP levels, a key read‐out of this pathway, are presented together at the end of Table 1 and indicated with asterisks.

The clinical problem that these trials sought to address is “residual risk” of MACE. Residual risk means the persistence of atherosclerotic heart attacks, strokes, and new‐onset symptomatic peripheral artery disease even after optimal management of conventional modifiable cardiovascular risk factors. But is residual risk really so hard to understand? One should keep in mind that the typical clinical picture for residual risk of MACE occurs in adults who, unfortunately, had been allowed to develop advanced arterial disease long before the initiation of optimal management. “[I]magine if we had a policy that patients with pneumococcal pneumonia could receive penicillin, but only after they are already on a ventilator. Many would die despite antibiotics and even prednisone, and it would not be a mystery.” [11].

In contrast, the underlying hypothesis behind the trials in Table 1 is that residual risk of MACE in patients with advanced atherosclerosis must arise primarily from residual inflammatory risk. But abundant clinical data contradict this hypothesis [11, 76]. For example, a classic study from 2011 showed that nearly half of the patients presenting with a first heart attack had low or very low serum levels of CRP [77]. Subsequent work documented substantial event rates in patients enrolled with baseline CRP levels <1 mg/dL who were then placed on optimal medical therapy [78]. These findings imply reasons for residual risk other than persistent sterile inflammation through that famous pathway [11, 76, 77, 79].

If we seek to eradicate the clinical burden of atherosclerosis, then these data imply that targeting inflammation—or any other factor—in end‐stage arteries is highly unlikely to achieve this goal [11, 76]. Despite extensive efforts, the approach remains a dead end. The fact that we focus our interventions so late in the course of atherosclerotic arterial disease may be a better explanation for residual risk of MACE. Residual risk may become a tractable problem on the road to eradication if we can screen and intervene earlier [11, 24, 76, 80]—an assertion that deserves testing [24, 76, 81].

Thus, in advance of the recent cardiovascular outcome trials of colchicine, inflammation—and specifically the inflammasome‐IL1ß‐IL6‐CRP pathway that was already targeted by several agents in Table 1—had consistently failed to provide any new medication with regulatory approval to treat patients with atherosclerosis. Nor does the pathway explain the considerable proportion of patients who have low CRP levels yet persistent atherosclerotic events [11, 42, 76, 77, 78, 79]. Although practitioners had hoped for an effective new medication, the assertion that colchicine would provide clinical benefit in advanced atherosclerotic arterial disease seemed at that point to be an extraordinary claim.

And the evidence?

Prospective randomized controlled clinical trials of colchicine on clinical outcomes in patients with atherosclerotic arterial disease

Table S1 summarizes major randomized clinical trials of colchicine in patients with atherosclerosis. Many of these trials are also discussed in the next section, below, in the context of “new meta‐analyses” published in 2025–2026.

Two prospective randomized clinical trials, COLCOT, reported in 2019 [82], and LoDoCo2, reported in 2020 [83], formed the basis for the FDA‐USA to grant a cardiovascular indication for colchicine in 2023. Both trials had studied the effect of colchicine versus placebo on cardiovascular outcomes in patients with clinically evident atherosclerotic disease (secondary prevention) during approximately 2 years of follow‐up. Both trials had met their pre‐specified primary outcomes, with significant p‐values (0.02 and <0.001, Table S1) and significant confidence intervals (CI).

But the story was not entirely straight‐forward. Neither trial had used circulating CRP levels as an inclusion criterion, and so the regulatory approval was broad for advanced arterial disease, with no CRP requirement. Because of trial design, it is not known whether colchicine had acted via anti‐inflammatory mechanisms or, say, its well‐known anti‐platelet effects [84]. In fact, in a small substudy of the COLCOT trial, the effect of colchicine on serum CRP levels was indistinguishable from the results with placebo (Table S1 and [82, 85]). Regarding consistency, colchicine in the LoDoCo2 trial produced only a tiny treatment effect in the Dutch patients but a robust effect in the Australian patients (Table S1 and [86]). As discussed by Jeon and Cho et al. [5], COLCOT found a greater cardiovascular benefit for patients with Type 2 diabetes mellitus (T2DM), which is often described as an “inflammatory” condition, whereas LoDoCo2 did not. Of note, an additional cardiovascular trial of colchicine—the Australian COPS, which was published 12 days after LoDoCo2 in 2020—failed to meet its pre‐specified primary composite endpoint (Table S1 and [87]). A key component of that primary composite endpoint—all‐cause mortality—was significantly worse in the colchicine arm [87]. Nevertheless, the two positive trials—COLCOT and LoDoCo2—were sufficient for the FDA‐USA in 2023.

But the next year, 2024, turned into colchicine's annus horribilis [11, 88]. Two new randomized controlled clinical trials of colchicine in patients with recent stroke—CHANCE‐3 and CONVINCE—failed to show benefit (Table S1 and [89, 90]). More troubling, the largest clinical trial to date of colchicine in myocardial infarction (MI), CLEAR SYNERGY—which had a similar design to the earlier COLCOT trial—failed to show any benefit in the primary outcome, nor across any component of the primary outcome, during 3 years of follow‐up (Table S1 and [91]). In the CLEAR SYNERGY trial, colchicine successfully lowered serum CRP levels, indicating an anti‐inflammatory effect [91], but “The only signal from this trial was an increased risk of diarrhea,” said the first author, Dr. Sanjit Jolly, in an interview just after public presentation of the results [92]. In the discussion section of their full report, Jolly et al. argued, “In our trial, 649 first primary‐outcome events occurred; an analysis of previous trials showed that those with more than 600 outcome events rarely produced spurious results disproven by subsequent trials.” [91].

Two major objections have been raised against the CLEAR SYNERGY trial [4, 75, 93, 94]. The first is that much of the trial took place during the COVID‐19 pandemic. In his compelling analysis of the dal‐GenE trial, which also took place during the pandemic and was also negative, Professor Allan Sniderman wrote, “Randomized controlled trials are the most credible tests of hypotheses because both groups should be affected randomly—and therefore similarly—by any event, known or unknown, that could alter the outcome, positively or negatively. Unknown events include COVID‐19.” [95]. Jolly et al.’s report of the CLEAR SYNERGY trial addressed this issue as well: “The results seemed to be similar during the different phases of the Covid‐19 pandemic.” [91]. Dr. Jolly subsequently expanded on this point: “The event rates aren't any different. We extended the follow‐up, and we extended the recruitment period, so we [continued] after COVID‐19 and there's still no treatment effect.” [96]. Additionally, and in contrast, many other prospective randomized clinical trials performed during the COVID‐19 era were positive and produced important new information to improve patient care—for example, the steady stream of successful cardiovascular outcome trials of GLP‐1 receptor agonists and SGLT2 inhibitors [97].

The other major objection to the CLEAR SYNERGY trial was the degree of CRP lowering. After 3 months, circulating CRP values were 4.27 ± 0.19 mg/L in the placebo group versus 2.98 ± 0.19 mg/L in the colchicine group, that is, significantly lower with colchicine, yet the adjusted mean was still above 2 mg/L (difference = −1.28 mg/L; 95% CI, −1.81 to −0.75, not crossing zero). But objections to the lack of any detectable CRP lowering by colchicine in the substudy of the COLCOT trial have been infrequent in my experience; Reference [85] provides a rare example. What does the degree of CRP lowering in the CLEAR SYNERGY trial mean? Nonadherence? Unlikely, because rates of diarrhea, a known side‐effect of colchicine, were statistically and medically significantly higher in the colchicine arm, as noted above [91]. But even if one accepts the argument that serum CRP levels must be pushed below 2 mg/L for a beneficial cardiovascular effect, the fact remains that colchicine comprehensively failed in the largest long‐term prospective randomized double‐blinded placebo‐controlled trial to date in MI patients.

The year 2025 was not much better for this medication. The prospective randomized placebo‐controlled COCOMO‐ACS trial found no significant benefit from 17.8 months of colchicine on the pre‐specified primary outcome of fibrous cap thickness by repeat optical coherence tomography of lipid‐rich non‐culprit coronary plaques in patients with acute non‐ST‐segment elevation MI (Table S1 and [98]). This result does not confirm prior reports that colchicine improves plaque composition by imaging in patients with acute coronary syndrome (ACS) (Table S1 and [11, 99, 100]).

More troubling is the prospective randomized Co‐STAR trial of colchicine versus placebo, given to patients before and after transcatheter aortic valve replacement. Results of the trial appeared online on July 15, 2025 [101], after the Journal of Internal Medicine had received and reviewed [4, 5]. The Co‐STAR trial was stopped early by pre‐specified protocol, owing to a statistically significant excess of new strokes in the colchicine group: five new strokes with colchicine (8.3%, i.e., one in 12 of the 60 patients in the colchicine arm of the study) versus zero new strokes with placebo (p = 0.022, Table S1). The authors correctly noted, “due to a play of chance, patients in the colchicine group more often had a history of prior stroke at baseline compared to patients in the placebo group (5 [8.3%] versus 0).” [101]. But the report did not specify whether the new strokes after administration of colchicine occurred in those specific patients. Nor did it state whether the new strokes were ischemic or hemorrhagic. Nevertheless, a history of prior events should be kept in mind when designing possible future clinical trials of colchicine in atherosclerotic arterial disease, a point covered in more detail in the “Learning from meta‐analyses” section.

An explosion of new meta‐analyses since 2024—plus one new population‐wide study in 2025

In 2025 and the beginning of 2026, 14 new meta‐analyses of prospective randomized clinical trials of colchicine in patients with advanced atherosclerotic arterial disease were published, most of which were stimulated by the negative trials in 2024 and the ensuing clinical confusion [4, 74, 93, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112]. At least seven of these meta‐analyses appeared in just the first 4 months of 2026 [103, 107, 108, 109, 110, 111, 112].

During this period, however, only two new cardiovascular trials of colchicine appeared: COCOMO‐ACS [98] and Co‐STAR [101], mentioned above and in Table S1. Nonetheless, based on established literature trends [113], it would not be surprising if more meta‐analyses of the published cardiovascular trials of colchicine in patients with atherosclerosis are on the way.

For simplicity, I have chosen a selection of these new meta‐analyses to highlight—namely, the first since 2024 [102], the two meta‐analyses by principal authors of recent major colchicine trials [74, 93], Xie et al.’s from the Journal of Internal Medicine [4], and the largest, which comprises 20 clinical trials of colchicine in patients with atherosclerotic arterial disease [103]. Table S1 was constructed by listing all clinical trials that were included in at least one of these five new meta‐analyses or, in the case of Co‐STAR, published separately. There are 23 such trials in Table S1. It should be noted that the literature contains additional clinical trials of colchicine in patients with atherosclerosis that were not included in these five meta‐analyses (e.g., four in [114] as noted in the “Comments” column in Table S1 and an additional seven in the first table of [108]). Taken together, this information yields a tentative total so far of 34 prospective cardiovascular trials of colchicine in patients with atherosclerosis.

The first of the meta‐analyses in 2025 was by Shaikh et al., analyzing nine clinical trials of colchicine in patients with ACS (Table S1 and [102]). The meta‐analysis included shorter and longer term trials, as well as trials that used surrogate outcomes such as imaging or serum CRP levels. Nevertheless, there was sufficient information on the study's stated primary outcomes of MACE or recurrent ACS for the authors to draw a conclusion: no benefit from colchicine [102]. Of the five secondary outcomes, only re‐hospitalizations (better with colchicine) and gastrointestinal side‐effects (worse with colchicine) achieved statistical significance [102].

d'Entremont et al., a collaborative group that includes several authors from the CLEAR SYNERGY trial, also published a meta‐analysis in 2025 of nine clinical trials of colchicine [74]. Their meta‐analysis had been planned before the results of the CLEAR SYNERGY trial were known [96]. d'Entremont et al.’s focus was patients with clinically evident atherosclerosis, not necessarily ACS, who had been treated with colchicine versus no colchicine, followed for at least 90 days, and for whom data on efficacy or safety were available. These features, including the length of follow‐up, were consistent with previous meta‐analyses in this area [74]. As shown in Table S1, d'Entremont et al. studied a different but overlapping set of nine trials from Shaikh et al.—and the two groups reached somewhat different conclusions. d'Entremont et al. inferred that colchicine had produced a modest 12% reduction in their primary outcome of a composite of cardiovascular death, MI, or stroke [74]. The only component of their primary outcome to reach significance was MI, with a 16% reduction attributed to colchicine, p = 0.02.

Two additional meta‐analyses from 2025, Samuel et al.’s in the European Heart Journal [93] and Xie et al.’s in the Journal of Internal Medicine [4], used nearly identical criteria for selecting clinical trials to include—namely, those trials that had enrolled patients with clinically evident atherosclerosis, then randomized the patients to colchicine versus placebo, followed them for 12 or more months, and compared the incidence of cardiovascular events, not surrogate outcomes. Consequently, as shown in Table S1, Samuel et al. and Xie et al. summarized exactly the same set of six randomized controlled clinical trials: LoDoCo (reported in 2013) [115], COLCOT (2019) [82], LoDoCo2 (2020) [83], COPS (2020) [87], CONVINCE (2024) [90], and CLEAR SYNERGY (ePub in 2024) [91]. To emphasize, both of these meta‐analyses included the recent negative trials CONVINCE in stroke [90] and CLEAR SYNERGY in MI patients [91]. Both excluded the large negative stroke trial CHANCE‐3 owing to its short duration—90 days of follow‐up—even though a high percentage of recurrent strokes occur in that narrow window (Table S1 and [11, 89]). Perhaps because of timing, the published meta‐analyses in 2025 and early 2026 did not include the increased rate of new strokes in the colchicine arm of the recent Co‐STAR trial [101], but that trial was likely too small to substantially affect the overall pooled analyses.

Both Samuel et al. and Xie et al. came to the same conclusion—namely, that colchicine had significantly reduced the rates of MI, stroke, and other key cardiovascular endpoints, provided that the medication had been administered for at least 12 months [4, 93]. The estimated reductions in their primary outcome of MACE by random effects models were 25%–26%. The commentary published in the same issue as Samuel et al.’s article noted that those authors’ results “are prone to exaggeration because (i) their meta‐analysis used a random‐effects model that gives undue weight to smaller trials which happen to have larger effects sizes; and (ii) the CHANCE‐3 trial with its neutral findings was excluded.” [116]. Similar criticisms can be applied to Xie et al.’s meta‐analysis, although those authors also included a fixed‐effect model that gave a somewhat lower 19% estimate for reduction in MACE based on the six trials [4].

Importantly, Xie et al.’s [4] first figure on the effects of colchicine on cardiovascular events (adapted here as Fig. 1) shows in explicit graphical form that the first three of the six included trials were positive—LoDoCo, COLCOT, and LoDoCo2. But their figure also shows that the most recent three of the six trials were negative—COPS [87], CONVINCE [90], and CLEAR SYNERGY [91].

Fig. 1.

Forest plot of the effects of colchicine on the pre‐specified primary outcomes of the six randomized controlled clinical trials (RCTs) of colchicine in patients with advanced atherosclerotic arterial disease that formed the basis for the new meta‐analyses by Samuel et al. [93] and by Xie et al. [4]. Data from the six trials are presented in their order of publication. Total events are emphasized. Adapted with permission from the first figure of Xie et al. [4]. The results of those authors’ fixed‐effect model and their random‐effects model are indicated at the bottom of the figure, as in the original. The figure shows that the first three of the six included trials were positive, whereas the most recent three of the six were negative (null). A semitransparent red arrow has been added to indicate this temporal trend, which raises the disturbing possibility of regression to the mean, also known as regression to the truth. 95%‐CI, 95% confidence interval; RR, risk ratio.

For perspective, would practitioners prescribe statins or any other LDL‐lowering medications if half of the cardiovascular outcome trials of those agents were negative (null)? Here is an answer: When just one ill‐conceived imaging trial by the manufacturers of ezetimibe turned out negative, major opinion leaders in academic medicine and in the lay press publicly abandoned the causative role of LDL in atherosclerosis in favor of the idea that statins decrease cardiovascular events via alleged anti‐inflammatory actions that ezetimibe, a non‐statin, would not share [117, 118]. Later clinical trials repeatedly and definitively repudiated this idea, as explained in more detail in the “Learning from meta‐analyses” section. Statins, ezetimibe, PCSK9 inhibitors, and bempedoic acid decrease cardiovascular events by lowering plasma concentrations of LDL and apoB, with no detectable signal for any other imaginable action of statins [13, 14, 15, 16, 17, 18, 19, 20].

Moreover, the temporal pattern displayed in Xie et al.’s [4] crucial first figure raises the disturbing possibility of regression to the mean, also known as regression to the truth [37, 119, 120, 121, 122]. The alleged benefits of colchicine in atherosclerotic arterial disease have become smaller and less significant as more trials accumulate, especially when those trials have higher numbers of events [2], as in CLEAR SYNERGY [91]. Figure 1 here contains the addition of a superimposed arrow to indicate this temporal trend. Whether this pattern continues in ongoing and possible future trials will be of considerable interest.

Of note, Samuel et al.’s [93] similar display in their first figure, meaning their top panel on MACE, presents the negative COPS trial as if it had been significantly positive. Samuel et al.’s panel shows an exploratory endpoint from COPS instead of the pre‐specified primary outcome. Nor does that panel include the CONVINCE trial—nor does it present the trials in chronological order. Samuel et al.’s display makes it look as if there were three positive trials of colchicine regarding MACE and just one neutral trial, CLEAR SYNERGY. The impression it gives of CLEAR SYNERGY, as another long‐term advocate of inflammation as a therapeutic target in atherosclerosis [118] put it, is that “you have one trial that's kind of the odd man out, and I don't buy it” [94]. But it is not the case.

The largest of these recent meta‐analyses, by Laudani et al. [103], separated clinical trials of colchicine in patients with chronic coronary syndrome (CCS) from ACS (see their first table and first figure). Other authors have suggested that colchicine may provide more benefit in chronic CAD than in the acute setting [96, 103, 116], although “the acute trials were essentially chronic [disease] after the first year” (Dr. Jolly quoted in [96]). Laudani et al.’s first figure, “Forest plot for MACE,” however, could justify the opposite conclusion regarding colchicine in chronic versus acute CAD. Their meta‐analysis of colchicine trials in patients with CCS yielded a relative risk of MACE of 0.62, with a 95% CI of 0.36–1.07, crossing 1.00, and a corresponding non‐significant p‐value of 0.087. Their analysis of colchicine trials in patients with ACS gave a relative risk of MACE of 0.77 (95% CI, 0.62–0.95) with a significant p‐value of 0.016. Bundling all those trials together produced a relative risk of MACE of 0.71 (95% CI, 0.58–0.87, p = 0.001), although “Quality of evidence was deemed moderate for MACE in the pooled and both CCS and ACS cohorts” [103]. Laudani et al. concluded, “the use of colchicine in patients with CAD reduces MACE without significantly increasing SAEs [serious adverse events] compared to control, although increasing gastrointestinal adverse events” [103].

Do these meta‐analyses mean that we should ignore the three [87, 90, 91], now six [89, 98, 101], new prospective randomized clinical trials showing no cardiovascular benefit, or even cardiovascular harm, from colchicine? That is, in effect, what several well‐known inflammation researchers stated shortly after the first public presentation of the results of the CLEAR SYNERGY trial—and apparently before key meta‐analyses from 2025 had been completed [94]. Here is what the lead author of the COLCOT trial [82], who was also the senior author of the recent meta‐analysis by Samuel et al. [93], said,

“The results of CLEAR SYNERGY have absolutely not changed my practice” [94].

Again, it appears to be the problem of prominent intellectual commitments to a particular approach [9, 10], also known as “strongly held views that are not easily relinquished” [10]. In the part of our field devoted to inflammation in atherosclerosis, similar difficulties arise from “a vested interest in complexity.” Many high‐profile careers depend on the immune system to constantly provide the next cytokine, cell type, or “inflammatory” patient cohort to study.

Wiersma et al. [9] and Bauer et al. [10] summarized literature that identified prominent intellectual commitments and ensuing prestige as nonfinancial COIs. Ioannidis emphasized COIs as a particular problem with many meta‐analyses [113]. In this circumstance, a principal author of one of the recent meta‐analyses of colchicine in advanced atherosclerosis appears to have had his conclusion firmly in mind, and said so publicly, before performing and publishing the study.

Dr. Sanjit Jolly, the first author of the CLEAR SYNERGY trial, responded:

“There's a subgroup of people whose life is built on inflammation who outright dismissed [CLEAR SYNERGY] … As a scientist, if you don't believe the data, then you're biased. And when data come out that conflict with what you believe in, you have two choices. You can look at the data objectively or you can dismiss them. And I think outright dismissing the data is the wrong thing to do” [94].

Evidence regarding possible effects of colchicine in primary prevention of cardiovascular events is poor [123], and the information we have on the effects of colchicine on cardiovascular outcomes in patients with T2DM is inconsistent, as noted above and in [5]. To fill these gaps, Jeon and Cho et al. [5], also in the Journal of Internal Medicine, performed a real‐world population‐wide assessment of T2DM patients without clinically evident atherosclerosis. Jeon and Cho et al. used the Republic of Korea's National Health Insurance Service database to select T2DM patients who also had gout—common in T2DM—and these patients’ gout was treated with one of two first‐line therapies: colchicine versus NSAIDs. As noted above and in Table 1, however, use of NSAIDs has been linked to increased cardiovascular event rates in other circumstances [33, 34, 35, 36], and so it may not have been an entirely neutral comparator.

After extensive propensity‐score matching of more than 50 covariates, Jeon and Cho et al.’s results were spectacularly and consistently negative. Use of colchicine was associated with no detectable benefit whatsoever over NSAIDs for the primary outcome of four‐point MACE, nor across any MACE component [5]. It would be of further interest to examine MACE rates in T2DM patients with gout who are taking colchicine or NSAIDs versus no‐colchicine/no‐NSAIDs, as had been done for gout patients without diabetes [36]. Jeon and Cho et al. also reviewed previously published observational studies of colchicine use and cardiovascular events, which—like the randomized clinical trials—have given conflicting results, showing both decreased and increased [36] cardiovascular event rates in association with this medication [5].

For comparison, another recent real‐world study of patients with T2DM from the same Korean National Health Insurance Service database found significantly lower rates of MACE in association with the intensity and especially the duration of therapy with statins [124]. These two population‐wide studies of statins [124] and colchicine [5], key medications of interest, have the advantage of collecting real‐world data. Prospective randomized clinical trials have been criticized for their narrow inclusion criteria, which can potentially limit broad applicability [37], and for highly labor‐intensive follow‐up that can be unrealistic in ordinary practice. But population‐wide studies have a disadvantage from lack of randomization: Confounders cannot always be reliably controlled. Still, the careful analyses and unambiguous findings of Jeon and Cho et al. do not help the case for colchicine, particularly when the comparator was NSAIDs, and the results can be contrasted with positive data from T2DM patients in the same database who take statins, a class of medications that provide clinical benefit in atherosclerotic arterial disease solely through their ability to lower plasma concentrations of LDL and apoB [13, 14, 19].

Next steps

Learning from meta‐analyses

What are the best uses of meta‐analyses? There is a growing literature critical of the methodology and proliferation of these studies. One author estimated that there may now be more meta‐analyses than actual clinical trials and that perhaps only 3% of all meta‐analyses are well‐conducted and clinically useful [113].

Here are three possibilities for high value from meta‐analyses. The first is in unusual circumstances involving a number of human studies, each underpowered but all pointing in the same general direction. Rigorously combining them may be the only way to generate findings with sufficient statistical power [125, 126]. But the situation with colchicine and atherosclerosis is different: The disease is by no means unusual, and the trials do not all point in the same direction.

The second potential high‐value use of meta‐analyses is in a common disease when there are many well‐powered trials all pointing in the same direction. Meta‐analyses can provide mechanistic insights. In atherosclerosis, so‐called pleiotropic effects of statins—meaning effects allegedly independent from LDL or apoB lowering—attracted a great deal of attention over the years [118, 127, 128]. But the Cholesterol Treatment Trialists’ Collaboration, among others, clearly identified the single factor that accounts for differential benefits of different statins at different doses on subsequent atherosclerotic events—namely, the degree and duration of plasma LDL or apoB lowering [13, 14]. The same has held true of ezetimibe, PCSK9 inhibitors, and bempedoic acid, which are not statins yet produce identical benefits as statins when matched for LDL or apoB lowering [15, 16, 17, 18, 19, 20]. These comparative analyses of the outcomes from various clinical trials, within and across several classes of medications, have provided powerful evidence against any significant role for pleiotropic effects of statins in the cardiovascular benefits that these medications provide to patients with atherosclerotic arterial disease.

The third potential use for meta‐analyses is when clinical trials do not point in the same direction. That is now the case with colchicine.

Can a dozen‐plus new meta‐analyses uncover a convincing clinical result from the current set of conflicting clinical trials? So far, practitioners have voted no. Prescription rates for colchicine by cardiologists remain low [129], owing at least in part to “divergent trial data” [75]—a problem that these meta‐analyses cannot change. Moreover, it does not seem compelling to combine clinical trials that had come to disparate conclusions, particularly when they studied disparate patient groups [75, 86, 116] and arrived in a troubling sequence of positive then neutral trials (Fig. 1).

Along related lines, the new meta‐analyses in 2025–2026 do not even agree with each other. Even when an overall effect was deduced, examples of underwhelming p‐values [74, 116] cannot be “considered a robust result for these types of analyses.” [96] Thus, the new meta‐analyses do not appear to constitute extraordinary evidence at a level that should direct practitioners to rely on colchicine in the clinical management of patients with atherosclerotic arterial disease.

In fact, based on the newly published randomized controlled clinical trials of colchicine showing no benefit [89, 90, 91, 98] or even harm [101], a re‐examination by the FDA‐USA and other regulatory agencies of their cardiovascular indications for this medication would be justified and helpful.

What these many new meta‐analyses can offer is guidance for the design of possible future clinical trials, with the hope of providing clarity at some point. Here are some features to consider: Have a minimal serum CRP level for enrollment to be able to test if that is where this medication might act in atherosclerotic arterial disease; ensure balance in prior MACE between the treatment arms (cf., [101]); verify that CRP levels drop below 2 mg/L in the treatment group during the trial (cf., [82, 85, 91]); assess medication adherence; assess platelet function [84, 130]; aim for more than 600 outcome events to minimize the possibility of the trial being overturned in the future [91, 131]; designate a treatment period of at least 12 months [4, 93] even if it means abandoning the goal of reducing early events such as recurrent strokes (cf., [89]); and ask the authors of prior trials and of these meta‐analyses for their ideas.

Here are related points from the commentary that was published simultaneously with the negative CONVINCE trial: “On the basis of the results from CONVINCE, implementing strategies that overcome challenges imposed by the relatively high rates of non‐adherence [to colchicine], dropouts, and crossovers, will be pivotal for the potential success of ongoing and future colchicine stroke prevention trials.” [132]. References [31, 75, 96, 132] provide lists of ongoing clinical trials of colchicine in patients with atherosclerotic arterial disease. Results of those trials should be followed closely.

A need to reform the part of our field devoted to inflammation in atherosclerosis

One more issue deserves discussion: the need to reform this part of our field. Three areas of concern come to mind regarding clinical and basic investigation of inflammation and atherosclerosis. First, trial names must be scrupulously neutral‐sounding, as noted in the initial section on “An extraordinary claim?”, in Table 1, and in decades of prior cardiovascular literature [56, 57, 58]. This point is particularly important for outcome trials of anti‐inflammatory compounds in patients with atherosclerotic arterial disease, given that so many of these trials to date have failed to achieve a cardiovascular indication for their experimental agents.

Second, given this lengthy disappointing history, institutional review boards should mandate that ongoing and future clinical trials of anti‐inflammatory agents in atherosclerotic arterial disease must disclose to prospective trial participants the extensive failures of this approach in scores of clinical trials so far (Table 1, references therein on anti‐inflammatory agents other than colchicine, and Table S1 and [87, 89, 90, 91, 98, 101] on colchicine). Anything less is not fully informed consent. It bears emphasizing that human clinical trials demand time, inconvenience, and risk from the patients who generously volunteer to participate.

Third, contributors to this part of the field—public funding agencies, regulatory bodies, pharmaceutical companies, private foundations, and individual scientists—might want to reconsider their intellectual commitment to the Big Idea that inflammation must be a definitive therapeutic target in atherosclerosis. How many dozen clinical trials that failed to bring a new therapy to patients in need, over how many decades, and at what personal and financial cost will it take for us to move past a concept that has produced so little for patient care? Table 1, Table S1, and this text list a total of at least 57 human clinical trials of anti‐inflammatory agents in patients with atherosclerotic arterial disease, spanning more than the past third of a century, without clear benefit and with no signs of stopping. Do we need a 58th trial? A 98th?

The field would benefit from an explicit expiration date, or shelf life, after which special scrutiny would be applied to any further support—financial and nonfinancial—for an entrenched Big Idea that has underperformed, or not performed, clinically [133, 134]. To protect patients, clinical trials are routinely stopped by protocol for futility or harm. Perhaps funding agencies could consider similar provisions for entire fields of long‐standing, but under‐performing, clinical and basic investigation.

These issues have become especially acute, given the therapeutic successes from targeting cholesterol‐rich, apoB‐containing lipoproteins, the proven causative agents of this disease. The field might make more progress to lessen or even eradicate the clinical burden of atherosclerosis, a lifelong disease, if we prioritized earlier screening of patients and earlier management of these causative agents, as well as proven exacerbators such as hypertension and smoking [11, 76, 80, 135, 136, 137, 138, 139, 140]—instead of focusing on exotic means of immunosuppression in end‐state arteries [11].

Conclusions

Long before Dr. Carl Sagan, as a fascinating scientific episode known as the N‐ray affair was drawing to a close, the man at its center, Professor René Blondlot, summarized how researchers convince each other: “Let each one form his personal opinion … either from his own experiments or from those of others in whom he has confidence.” [141].

For now, Dr. Sanjit Jolly, the first author of the CLEAR SYNERGY trial [91] and the senior author of one of 2025–2026's many meta‐analyses [74], provides helpful guidance. Dr. Jolly had placed his own father on colchicine after an MI but then immediately took him off as soon as CLEAR SYNERGY was unblinded [11, 92]. At this point, it is difficult to recommend starting or continuing a patient or family member on that medication—or on any other anti‐inflammatory therapies—for atherosclerotic arterial disease. Ongoing and newly designed clinical trials might change that conclusion, but the track record so far for inflammation inhibition in atherosclerosis is extensive and disappointing—a fact that merits wider discussion.

Author contributions

Kevin Jon Williams: Conceptualization; writing—original draft; writing—review and editing; data curation; visualization.

Conflict of interest statement

Dr. Williams declares no financial, and no nonfinancial, conflicts of interest related to the topics of this commentary. Dr. Williams was the first and sole corresponding author of the original formal presentation of the response‐to‐retention model of the pathogenesis of atherosclerosis [21]. The response‐to‐retention model includes downstream immune and non‐immune reactions, but not as the central causative factor in this disease.

Supporting information

Table S1: Prospective randomized clinical trials of colchicine versus no‐colchicine in patients with atherosclerotic arterial disease that were included in at least one of the five selected meta‐analyses from 2025–2026 [4, 74, 93, 102, 103] or published separately.