When researchers in the Diflunisal Trial Consortium gathered for a meeting last March, they had no idea what was in store. John Berk of Boston University reviewed their attempt to treat the rare disease familial amyloid polyneuropathy (FAP) with the repurposed anti-inflammatory drug. He took the researchers through the fine details of the statistical analysis before offering up the first result. It showed the treatment had halted the fatal illness, while preserving quality of life. “The room went completely silent,” Berk recalled. “No one had any idea that diflunisal was going to have that kind of effect.” The trial results, published December 25 in the Journal of the American Medical Association, not only offer a new treatment for FAP but also show that an artificial chaperone can stabilize an amyloid-forming protein and keep it from building aggregates.

FAP arises from mutations in transthyretin, a thyroid hormone transporter. This progressive disease causes pain, weakness, and loss of skin sensations, such as the ability to detect heat. Estimated to affect fewer than 10,000 people worldwide, FAP typically starts between ages 30 and 60 and kills within 10 to 15 years. Because the liver produces most of the body’s transthyretin, liver transplant has been a standard FAP treatment since 1990. However, few donors are available, and transplants are costly, risky, and require recipients to take immunosuppressants for the rest of their lives. Moreover, Berk said, doctors have found that in many people with FAP, a disease-linked thickening of the heart muscle continues despite a transplant (Liepnieks et al., 2010Olofsson et al., 2002). “Liver transplant is less and less seen as a panacea,” he said. “Alternative treatments are needed.”

Study co-author Jeffery Kelly of The Scripps Research Institute in La Jolla, California, has been pursuing an alternative for decades. His idea: stabilize transthyretin in its native tetramer form, preventing it from dissembling into monomers that can misfold into amyloid. His group designed hundreds of new compounds to fit the thyroid hormone pocket and stabilize transthyretin, choosing a drug called tafamidis as the top candidate and developing it through FoldRx Pharmaceuticals, Inc., a company Kelly co-founded. At the same time, they looked for known drugs that might fit the cavity. Among the pharmacopeia of medicines already approved for use in the United States, they found diflunisal, an early-generation, nonsteroidal anti-inflammatory drug (NSAID). Diflunisal stabilized transthyretin and prevented amyloid aggregation in vitro. Though it has fallen out of regular use in favor of newer NSAIDS, diflunisal remains available by prescription. Kelly and colleagues pursued trials of both drugs. Results of a Phase 2/3 trial of tafamidis (Coelho et al., 2012) showed that patients on the drug retained some muscle strength, sensation, and reflexes in their legs, leading to approval in the European Union but not the United States (see Aug 2011 news story).

Berk, senior author Peter Dyck of the Mayo Clinic in Rochester, Minnesota, and colleagues followed up on early, positive results from a small Phase 1 double-blind trial of diflusinal (Sekijima et al., 2006) with a larger trial. They randomized 130 people with FAP to receive either placebo or 250 milligrams of the NSAID—a dose at which the drug weakly blocks inflammation—twice daily for two years. They measured neuropathy by a 270-point scale called the Neuropathy Impairment Score plus seven. It allots points for a variety of clinical signs, including muscle weakness, reduced reflexes, and sensory loss. Higher scores indicate worse symptoms and a change of two points indicates a significant difference. A gain of two points could indicate muscle strength fell by 25 percent and another symptom, such as motion sensation, dropped by 50 percent. Berk noted that doctors typically use the scale to measure progression of diabetes, and in that population it takes two years to see a noticeable change in the score. Progression was much more rapid in the FAP trial. In the placebo group, the scores rose an average of 25 points over the two-year study. For those taking diflunisal, the average uptick was 8.7 points. 

The researchers also examined quality of life using a generic 36-part questionnaire called the SF-36, which rates physical and mental health on a 100-point scale. Higher scores indicate healthier, happier people, though the subjective nature of the scale makes it difficult to rate an individual’s score as good or bad. Study participants had baseline scores of about 35 out of 100 for physical factors and 47 out of 100 for mental factors. Placebo recipients saw their physical scores go down by an average of 4.9 and mental scores decline by 1.1 points over the study period; in people on diflunisal, scores increased an average of 1.5 and 3.7 points, respectively.

Both Kelly and Berk said they were surprised at the performance of the generic NSAID. They noted that compared with tafamidis, diflunisal binds transthyretin poorly. However, diflunisal easily enters the bloodstream when taken orally, making plenty of the drug available to bind all the transthyretin, they said. Because other transporters carry the vast majority of thyroid hormones in the body, the drug should have no effect on hormone dynamics. Transthyretin also shuttles vitamin A, but grabs it via a different binding site, suggesting that function should remain unaffected.

Researchers have not compared diflunisal and tafamidis head-to-head in matching trials. “At the end of the day, I think both are effective drugs,” Kelly said. “It is really hard to say which of these compounds will turn out to be better.” Diflunisal has the disadvantage of being an anti-inflammatory, meaning it could irritate the stomach, alter blood flow in the kidney, and cause fluid to build up in the heart, Berk noted. The diflunisal researchers avoided enrolling people with heart failure or kidney problems. On the other hand, generic diflunisal should be cheaper than tafamidis, which remains under patent. “I think both drugs will prove useful,” Berk said. In time, they may be joined or supplanted by similar-acting, more specific chaperones, said Isabella Graef at Stanford University School of Medicine in Palo Alto, California, who has developed molecules that are more selective for transthyretin than diflunisal or tafamidis (Penchala et al., 2013).

Berk said the researchers plan to request that the Food and Drug Administration formally approve FAP as an indication that can be treated with diflunisal. However, given the drug’s current availability, he pointed out, “there is very little to prevent or even to discourage a primary care physician or rheumatologist from prescribing this stuff for [FAP].” In fact, he suspects many people in the study’s placebo arm who dropped out of the trial believed they were getting the dummy pills, and wanted to obtain the active drug off-label. The researchers also want to follow their participants long-term in an open-label extension study.

As for tafamidis, Kelly said Pfizer, which bought FoldRx in 2010, plans to submit additional data for FDA approval. He suspects that the Phase 2/3 trial—at only 18 months—was too short to yield results as good as diflunisal’s. Having now followed up people on tafamidis for six to seven years, he said the treatment continues to stabilize disease.

Could a similar approach prevent amyloid formation in other neurodegenerative conditions, such as Alzheimer’s or Parkinson’s? In principle, yes, Graef said, though in practice it will be more complicated than with transthyretin. Unlike transthyretin, which has an obvious pocket for an artificial chaperone to fill, Aβ and tau form pretty loose structures. Gregory Petsko of Brandeis University, Waltham, Massachusetts, added that the drug would have to cross the blood-brain barrier, an additional challenge not required to treat FAP. Petsko, and researchers led by Dennis Selkoe at Brigham and Women's Hospital, Boston, have reported that α-synuclein also forms a native tetramer (see Aug 2011 news story), though Petsko told Alzforum he has had no luck finding a chaperone for α-synuclein to treat PD and other scientists have debated how robustly this finding replicates in other labs.

Should scientists come up with a chaperone to try for AD or PD, Kelly has some advice. “You have to be patient,” he said. “Even when you stop new protein aggregation, you still need a long trial to see a positive result.”—Amber Dance


  1. This study is commendable as a trial of a "repurposed" NSAID that had otherwise lost its clinical relevance. The trial showed both formal efficacy and quality-of-life benefits in patients with a rare but lethal inherited syndrome that leads to multi-organ amyloid deposition. In a particular, fatal form, the disease involves the peripheral nerves. It is important to point out, however, that there are many sources of amyloid, and that the type involved in this disease comes from mis-metabolism of transthyretin, a thyroid hormone carrier protein.

    The paper's discussion section contains the statement "this study provides proof of concept that kinetically stabilizing an amyloidogenic precursor protein (transthyretin) translates to successfully modifying amyloid-related neurological disease progression." While this statement is undoubtedly correct, it would be unwise to interpret these findings as having implications for a role of diflunisal or other NSAIDs in the treatment of Alzheimer's disease (AD) dementia. It is true that impressive epidemiological evidence suggests users of conventional "dual-inhibitor" NSAIDs have reduced frequency of AD dementia, at least among "young-old" individuals (Szekely and Zandi, 2010). Those findings suggest that administration of NSAIDs to individuals in the early pre-symptomatic stages of AD might retard disease progression at that point. One should remember, however, that this notion has never been proven in clinical trials, and that all trials attempting to show a therapeutic effect of NSAIDs (including one cyclo-oxygenase 2 inhibitor) as treatments of symptomatic AD or its prodrome MCI have failed. The largest of these, the ADCS treatment trial with naproxen and rofecoxib, actually showed a trend toward negative (not just null) results in the rofecoxib-treated group (Aisen et al., 2003). Another large trial of rofecoxib in MCI patients showed negative results on the primary outcome of "conversion" to dementia (Thal et al., 2005). Even the Alzheimer's Disease Anti-inflammatory Prevention Trial (ADAPT)—which was intended to examine NSAID treatments for prevention rather than treatment of symptoms—showed a negative treatment effect over its first several years in participants treated with naproxen or celecoxib (Lyketsos et al., 2007). Adverse effects were especially noteworthy in ADAPT participants who (it was later discovered) had entered the trial with MCI (Breitner et al., 2011). At least one large observational study also found increased incidence of AD dementia in prior heavy users of NSAIDs (Breitner et al., 2009). The latter finding was derived from a particularly elderly cohort who might, accordingly, have had a relatively high proportion of individuals with significant pre-symptomatic or early symptomatic Alzheimerization in their brains. If anything, these results suggest that NSAIDs have adverse treatment effects on persons with symptomatic or late pre-symptomatic AD.

    Especially because transthyretrin amyloid is different and the target neural structures are distinct from AD pathology, I consider that this new JAMA paper does not counter that suggestion in any way.


    . Non-steroidal anti-inflammatory drugs and Alzheimer's disease: the epidemiological evidence. CNS Neurol Disord Drug Targets. 2010 Apr;9(2):132-9. PubMed.

    . Effects of rofecoxib or naproxen vs placebo on Alzheimer disease progression: a randomized controlled trial. JAMA. 2003 Jun 4;289(21):2819-26. PubMed.

    . A randomized, double-blind, study of rofecoxib in patients with mild cognitive impairment. Neuropsychopharmacology. 2005 Jun;30(6):1204-15. PubMed.

    . Naproxen and celecoxib do not prevent AD in early results from a randomized controlled trial. Neurology. 2007 May 22;68(21):1800-8. Epub 2007 Apr 25 PubMed.

    . Extended results of the Alzheimer's disease anti-inflammatory prevention trial. Alzheimers Dement. 2011 Jul;7(4):402-11. PubMed.

    . Risk of dementia and AD with prior exposure to NSAIDs in an elderly community-based cohort. Neurology. 2009 Jun 2;72(22):1899-905. PubMed.

    View all comments by John Breitner

Make a Comment

To make a comment you must login or register.


News Citations

  1. Amyloid-Blocking Drug Poised for Approval for Rare Disease
  2. An α-Synuclein Twist—Native Protein a Helical Tetramer

Paper Citations

  1. . Progression of transthyretin amyloid neuropathy after liver transplantation. Neurology. 2010 Jul 27;75(4):324-7. PubMed.
  2. . Progression of cardiomyopathy after liver transplantation in patients with familial amyloidotic polyneuropathy, Portuguese type. Transplantation. 2002 Mar 15;73(5):745-51. PubMed.
  3. . Tafamidis for transthyretin familial amyloid polyneuropathy: a randomized, controlled trial. Neurology. 2012 Aug 21;79(8):785-92. Epub 2012 Jul 25 PubMed.
  4. . Orally administered diflunisal stabilizes transthyretin against dissociation required for amyloidogenesis. Amyloid. 2006 Dec;13(4):236-49. PubMed.
  5. . AG10 inhibits amyloidogenesis and cellular toxicity of the familial amyloid cardiomyopathy-associated V122I transthyretin. Proc Natl Acad Sci U S A. 2013 Jun 11;110(24):9992-7. Epub 2013 May 28 PubMed.

External Citations

  1. trial
  2. trial
  3. study

Further Reading


  1. . Trans-suppression of misfolding in an amyloid disease. Science. 2001 Sep 28;293(5539):2459-62. PubMed.
  2. . Tafamidis for transthyretin amyloidosis. Drugs Today (Barc). 2012 May;48(5):331-7. PubMed.
  3. . Mechanisms of transthyretin inhibition of β-amyloid aggregation in vitro. J Neurosci. 2013 Dec 11;33(50):19423-33. PubMed.
  4. . Transthyretin stabilization by iododiflunisal promotes amyloid-β peptide clearance, decreases its deposition, and ameliorates cognitive deficits in an Alzheimer's disease mouse model. J Alzheimers Dis. 2014;39(2):357-70. PubMed.
  5. . Genetic stabilization of transthyretin, cerebrovascular disease, and life expectancy. Arterioscler Thromb Vasc Biol. 2013 Jun;33(6):1441-7. Epub 2013 Apr 11 PubMed.

Primary Papers

  1. . Repurposing diflunisal for familial amyloid polyneuropathy: a randomized clinical trial. JAMA. 2013 Dec 25;310(24):2658-67. PubMed.