Occam’s Razor

Arete-Zoe, December 14, 2016

Occam’s (or Ockham’s) razor is a principle attributed to early medieval logician William of Ockham (1287–1347). He used the principle to justify many of his conclusions including his statement “God’s existence” that “cannot be deduced by reason alone”. The original principle as formulated by Ockham says:

“Pluralitas non est ponenda sine neccesitate (Entities should not be multiplied unnecessarily)”

Ockham’s razor 

Many scientists have adopted Occam’s Razor. German rationalist philosopher, Gottfried Wilhelm Leibniz (1646-1716) reinvented the same principle in his “Identity of Indiscernibles”.

Natural philosopher Isaac Newton (1642-1727) in his Principia: the mathematical principles of natural philosophy stated the rule:

” We are to admit no more causes of natural things than such as are both true and sufficient to explain their appearances.” (Newton’s Principia)

Bertrand Russell offered the following form of Occam’s Razor:

“Whenever possible, substitute constructions out of known entities for inferences to unknown entities”

Ernst Mach’s version of Occam’s razor as defined in his Principle of Economy, stated:

“Scientists must use the simplest means of arriving at their results and exclude everything not perceived by the senses.”

Stephen Hawking writes in A Brief History of Time:

“We could still imagine that there is a set of laws that determines events completely for some supernatural being, who could observe the present state of the universe without disturbing it. However, such models of the universe are not of much interest to us mortals. It seems better to employ the principle known as Occam’s razor and cut out all the features of the theory that cannot be observed.”

Did we forget these principles in clinical research?

Recent analysis by PAREXEL evaluated data on 38 Phase III trials from the period between 2012 and 2015 that failed to meet primary or secondary efficacy endpoints. These 38 trials summarily enrolled nearly 150,000 patients. (Phase III Trial failures: Costly, but preventable)

Darapladib (SB480848): Phase I (12 studies, enrollment 283); Phase II (3 studies, enrollment 264); phase III (3 studies, enrollment 28,918); observational (1 study, 28855 subjects). GSK blockbuster hopeful darapladib had little effect on strokes, one of the key measures on that endpoint, and the drug failed to demonstrate greater efficacy than placebo [FierceBiotech, 2014].

Evacetrapib (LY2484595): Phase I (20 studies, enrollment 953); Phase II (2 studies, enrollment 580); Phase III (4 studies, enrollment 12,604). According to the American College of Cardiology, evacetrapib is third late failure in a class of cholesteryl ester transfer protein (CETP) inhibitors. These drugs are designed to disrupt conversion of HDL cholesterol into LDL cholesterol. Another CETP inhibitor, torcetrapib, was abandoned after a phase 3 clinical trial because of increased risk of cardiovascular events and death. Dalcetrapib, was stopped when a phase 2 clinical trial which failed to meet efficacy goals [AAC, 2016].

Losmapimod (GW856553): Phase I (3 studies, enrollment 92); Phase II (5 studies, enrollment 908); Phase III (1 study, enrollment 3,503). GSK’s decision to terminate losmapimod development for COPD came in January 2016, following a long and hard look at Phase II data. Termination of the losmapimod heart attack program came after the drug failed to demonstrate efficacy in a Phase III trial. Losmapimod is a selective inhibitor of p38 mitogen-activated protein kinases, which are mediators of acute inflammation. According to CT.gov, the drug was tested for glomerulosclerosis, COPD, acute coronary syndrome, and major depression. [Biospace, 2016]

Otamixaban (XRP0673): Phase I (2 studies, enrollment 73); Phase II (2 studies, enrollment 4,188); Phase III (1 study, enrollment 13,220). In Phase I, otamixaban was studied on patients with renal and hepatic impairment, respectively. The Phase III study failed to show superiority of otamixaban over a combination of heparin and eptifibatide (standard of care) in moderate- to high-risk patients. Moreover, otamixaban doubled the risk of major or minor bleeding and thrombotic procedural complications. [PharmaTimes, 2013].

Serelaxin (RLX030): Phase I (1 study, 36 patients); Phase II (11 studies, enrollment 810); Phase III (3 studies, enrollment 10,788). The FDA granted breakthrough status to serelaxin in June 2013. A year later, the drug was rejected by both EMA and the FDA. The project continues, however, and the latest study RELAX-REPEAT is now again recruiting patients with chronic heart failure. [Medscape, 2016]

Aclerastid (DSC127): Phase I (0/0); Phase II (1 study, enrollment 78); Phase III (3 studies, enrollment 1,255). In November 2015, an assessment of aclerastide found the drug did not meet its efficacy endpoints of healing diabetic foot ulcers. Consequently, Derma Sciences halted the clinical trial and is terminated the aclerastide program, and its CEO Edward Quilty resigned. [BusinessWire, 2015]

Aleglitazar: Phase I (8 studies, enrollment 267); Phase II (6 studies, enrollment 899); Phase III (8 studies, enrollment 9,838). Roche has announced the halting of a phase 3 trial of aleglitazar due to failure to meet efficacy endpoints and safety concerns, specifically PPAR class-related toxicity such as bone fractures, heart failure, and gastrointestinal bleeding. Aleglitazar turns on peroxisome proliferator-activated receptor (PPARs), receptors that are known to be problematic. Dr. Bernard Charbonnel in a piece on PPARs noted that more than 50 such drugs had failed clinical trials due to safety concerns. After the Phase III failure of another three PPARs, aleglitazar, muraglitazar and tesaglitazar the whole class of drugs appears to be doomed. [FierceBiotech, Medscape, 2013]

Basal insulin Peglispro (LY2605541): Phase I (14 studies, enrollment 480); Phase II (2 studies, enrollment 269); Phase III (10 studies, enrollment 6087). In December 2015, Eli Lilly announced that it will cease development of basal insulin peglispro, a potential treatment for type 1 and type 2 diabetes [Lilly press release, 2015]

“While we are encouraged by the efficacy data we observed for BIL, we know that moving forward would have required a significant amount of time and investment with no assurance that we would find conclusive answers,” said Enrique Conterno, president, Lilly Diabetes.

Saxagliptin: Phase I (18 studies, enrollment 641); Phase II (5 studies, enrollment 784); Phase II/III (1 study, enrollment 450), Phase III (38 studies, enrollment 18,216) Phase IV (37 studies, enrollment 26229). SAVOR was a large randomized, double-blind, placebo-controlled Phase 4 study in patients with T2DM at high risk of CV disease. The primary objective of the study was to evaluate the CV safety of saxagliptin. SAVOR, showed an association between saxagliptin treatment and an increased risk for hHF; however, a mechanism to account for the hHF finding has yet to be determined [AstraZeneca Advisory Committee Briefing Document, 2015]. FDA safety warning followed in February 2014.

Alisertib (MLN8237): Phase I (32 studies, enrollment 1234); Phase I-II (7 studies, enrollment 605); Phase II (17 studies, enrollment 1440); Phase III (1 studies, enrollment 271). Japanese drug-maker Takeda tested Alisertib for the treatment of a variety of cancers, including peripheral T-cell lymphoma, acute myeloid leukemia, small-cell lung carcinoma, solid tumors, and others. The program for lymphoma was halted after the drug failed to show any benefit in progression-free survival. Takeda acquired the program through the acquisition of Milennium Pharmaceuticals in 2008. [PharmaTimes, 2015]

Cabozantinib Phase I (19 studies, enrollment 950); Phase I/II (3 studies, enrollment 184); Phase II (37 studies, enrollment 2840); Phase III (5 studies, enrollment 2887), Phase IV (2 studies, enrollment 1188); Phase not stated (5 studies, enrollment 121); Observational (1 studies, enrollment 50). In 2014, following phase III failure of its drug against advanced prostate cancer, Exelixis had to ax about 70% of its workforce. In March 2016, French drug-maker Ipsen bought the drug with a hope that it may treat kidney and liver cancers. The drug is already approved for severe thyroid cancer (FierceBiotech).

Dacomitinib (PF00299804): Phase I (18 studies, enrollment 591); Phase I/II (4 studies, enrollment 593); Phase II (16 studies, enrollment 992); Phase III (3 studies, enrollment 2,037); Phase not stated (1 studies, enrollment 6). In January 2014, Pfizer announced that dacomitinib failed in two Phase III studies. Dacomitinib irreversibly inhibits enzyme pan-HER kinase in patients with advanced non-small cell lung cancer. Dacomitinib did not improve progression-free survival when compared with Roche’s Tarceva (erlotinib) and placebo, respectively. Research for the treatment of non-small cell lung cancer with specific mutations continues. [PharmaTimes, 2014, Cancer Therapy Advisor, 2016]

Enzastaurin (DB102, formerly LY317615 HCl): Phase I (9 studies, enrollment 280); Phase I/II (2 studies, enrollment 137); Phase II (35 studies, enrollment 2,595); Phase III (2 studies, enrollment 1,106). In 2013 Eli Lilly stopped on development of enzastaurin, which failed to improve disease-free survival in patients with diffuse large B-cell lymphoma. Enzastaurin is a serine/threonine kinase inhibitor of the PKC beta and AKT pathways and has been studied for a range of solid tumors and hematological cancers. [Fierce Biotech, 2016]

Etirinotecan pegol (NKTR-102): Phase I (2 studies, enrollment 57); Phase II (9 studies, enrollment 377); Phase III (2 studies, enrollment 1,202). The results of the BEACON trial failed to demonstrate improvement in survival and progressio-free survival when compared to active control in patients with advanced breast cancer. [PMLive, 2015]

Ganetespib + docetaxel (STA-9090): Phase I (13 studies, enrollment 337); Phase I/II (6 studies, enrollment 434); Phase II (15 studies, enrollment 2,353); Phase II/III (1 studies, enrollment 385); Phase III (3 studies, enrollment 2,296). In October 2015, Synta Announced Termination of Ganetespib Phase 3 GALAXY-2 Trial in Lung Cancer because of futility. [Business Wire, 2015]

Iniparib (BSI-201):

Lapatinib + Trastuzumab: Phase 0 (1 study, enrollment 40); Phase I (19 studies, enrollment 1,564); Phase I/II (11 studies, enrollment 468); Phase II (51 studies, enrollment 3,888); Phase II/III (1 study, enrollment 697); Phase III (15 studies, enrollment 13,919); Observational (7 studies, enrollment 715). The results from the phase III ALTTO trial show no additional benefit for adding lapatinib (Tykerb) to trastuzumab (Herceptin) in the adjuvant treatment of HER2-positive breast cancer in terms of progression-free survival and overall survival. [The ASCO Post, 2014]

MAGE-A3: Phase 0 (1 studies, enrollment 25); Phase I (11 studies, enrollment 353); Phase I/II (9 studies, enrollment 400); Phase II (10 studies, enrollment 664); Phase II/III (1 studies, enrollment 4); Phase III (2 studies, enrollment 3,629); Observational (1 studies, enrollment 983). GlaxoSmithKline invested high hopes and lot of fortune in the cancer vaccine MAGE-A3. The vaccine failed to meet efficacy endpoints for non-small cell lung cancer and then again for melanoma. GSK continues to search for a possible sub-population of cancer patients who might benefit from the therapy. [FierceBiotech, 2014]

Motesanib (AMG 706): Phase I (19 studies, enrollment 781); Phase I/II (2 studies, enrollment 51); Phase II (17 studies, enrollment 1,617); Phase III (3 studies, enrollment 1,650); Observational (1 study, enrollment 50). Japanese pharmaceutical giant Takeda acquired the drug from AMGEN and tested in a series of trials for non-small cell carcinoma. Motesanib failed to meet efficacy endpoints and the trial was terminated in 2011. [In the Pipeline, 2015]

Onartuzumab (MetMab): Phase I (4 studies, enrollment 59); Phase I/II (1 studies, enrollment 0); Phase II (7 studies, enrollment 1,141); Phase III (5 studies, enrollment 1,610). In March 2014, Roche oncology R&D arm at Genentech halted a Phase III study comparing MetMab with Tarceva. The drug failed to block metastasis in non-small cell lung cancer and the trial was halted for futility. [Fierce Biotech, 2014]

Ramucirumab: Phase I (20 studies, enrollment 1,292); Phase I/II (2 studies, enrollment 204); Phase II (34 studies, enrollment 3,406); Phase II/III (1 study, enrollment 908); Phase III (12 studies, enrollment 7,785); Phase IV (1 study, enrollment 1,000); Expanded Access (2 studies, enrollment 0). In September 2013, Eli Lilly announced that its Phase III study for ramucirumab failed to meet its efficacy endpoints on progression-free survival in patients with metastatic breast cancer. Phase III study for ramucirumab in gastric cancer hit its primary and secondary endpoints. The announcement was followed by immediate drop in shares by more than 5%. [FierceBiotech, 2013]

Selumetinib + Dacarbazine: Phase I (1 studies, enrollment 211); Phase II (3 studies, enrollment 783); Phase III (1 studies, enrollment 152). AstraZeneca acknowledged that its non-small-cell lung cancer (NSCLC) candidate selumetinib missed its primary endpoint in a Phase III trial assessing the drug with docetaxel. The combination failed to improve progression-free survival and overall survival compared with placebo. [GEN, 2016]

Trebananib + paclitaxel (AMG 386): Phase I (11 studies, enrollment 453); Phase I/II (2 studies, enrollment 122); Phase II (15 studies, enrollment 3,208); Phase III (3 studies, enrollment 2,157); Expanded Access (1 study, enrollment 20). Following disappointing results, trebananib was terminated for ovarian cancer. This is a comment by Sean E. Harper, M.D., executive vice president of R&D at Amgen: [Amgen, 2014]

“While the overall survival results of the TRINOVA-1 study are disappointing, this study is the first of three Phase 3 trials designed to evaluate the safety and efficacy of trebananib in patients with ovarian cancer. We continue to explore the potential of trebananib’s novel anti-tumor mechanism of action in other cancer settings”

Xilonix: Phase III (2 studies, enrollment 1,076). The Phase III trial for oncology drug Xilonix had to be halted not because of safety or efficacy issues but because of a series of embarrassing errors and mishaps that made the study impossible to evaluate. The drug was tested for colorectal cancer. [FierceBiotech, 2015]

Vintafolide Phase I (3 studies, enrollment 72); Phase II (6 studies, enrollment 458); Phase III (1 studies, enrollment 640). Merck & Co. and Endocyte announced that the Data Safety Monitoring Board for the PROCEED trial has recommended that the Phase III study be halted as vintafolide did not meet efficacy endpoints and failed to improve progression-free survival in patients with platinum-resistant ovarian cancer. [FirstWord Pharma, 2014]

Vosaroxin +cytarabine (Qinprezo): Phase I/II (1 studies, enrollment 115); Phase II (1 studies, enrollment 61); Phase III (1 studies, enrollment 675); Expanded Access (1 studies, enrollment 17). Sunesis Pharmaceuticals announced disappointing results from its Phase III trial VALOR trial. The drug failed to improve overall survival in patients with acute myeloid leukemia. [Clinical Trials Arena, 2015]

Fluticason furoate + vilanterol (Breo/Relvar): Phase I (8 studies, enrollment 258); Phase II (7 studies, enrollment 877); Phase III (36 studies, enrollment 29,361); Phase IV (4 studies, enrollment 332). GlaxoSmithKline tested their new drug Breo to demonstrate improved survival compared to placebo. The SUMMIT trial missed efficacy endpoints and the program was halted. [PharmaForum, 2016]

Bitopertin (RG1678): Phase I (2 studies, enrollment 71); Phase II (2 studies, enrollment 400); Phase III (4 studies, enrollment 2,450). Roche’s Phase III failed to meet its primary endpoint, and the company was forced to halt the trial for futility. Bitopertin is a glycine transporter-1 inhibitor designed for the treatment of schizophrenia. Adding bitopertin to antipsychotics failed to improve persistent negative symptoms of schizophrenia as measured by PANSS (Positive and Negative Symptoms Scale) scores when compared to placebo. [PsychCentral, 2015]

Edivoxetine (LY2216684): Phase I (16 studies, enrollment 362); Phase II (2 studies, enrollment 697); Phase II/III (3 studies, enrollment 1,101); Phase III (6 studies, enrollment 6,197). Eli Lilly’s antidepressant candidate edivoxetine used as an add-on to a SSRI failed to meet its primary endpoint in three Phase III trials when compared to a SSRI combined with placebo. [GEN, 2013]

Pomaglumetad methionil: Phase I (2 studies, enrollment 160); Phase III (2 studies, enrollment 432); Expanded Access (1 study, enrollment 16). In summer 2012, Eli Lilly wrote of another schizophrenia drug, pomaglumetad methionil. In the first two pivotal Phase III studie the drug did not show superiority to placebo. [GEN, 2013]

Lessons Learned and Potentials for Improvement in CNS Drug Development: ISCTM Section on Designing the Right Series of Experiments (2015)

Solanezumab: Phase I (1 studies, enrollment 130); Phase II (3 studies, enrollment 113); Phase II/III (1 studies, enrollment 210); Phase III (6 studies, enrollment 3,190). Drug designed to treat Alzheimer’s disease, through targeting the amyloid plaque, failed a huge Phase 3 clinical trial. The announcement caused Eli Lilly’s shares to drop 10.5 percent. [Bloomberg Gadfly, 2016]

Apremilast: Phase I (8 studies, enrollment 398); Phase I/II (2 studies, enrollment 13); Phase II (25 studies, enrollment 2,047); Phase II/III (1 study, enrollment 15); Phase III (10 studies, enrollment 4,446); Phase IV (6 studies, enrollment 418); Observational (8 studies, enrollment 6875); not stated (3 studies, enrollment 45). In summer 2014, Celgene announced that apremilast failed a Phase 3 study in improving symptoms of patients with ankylosing spondylitis. [Xconomy, 2014]

Tabalumab:  Phase I (4 studies, enrollment 122); Phase II (8 studies, enrollment 866); Phase III (9 studies, enrollment 7,858). Lilly’s two Phase III trials for tabalumab – ILLUMINATE I and II- did not meet primary endpoint and failed to measure up compared to existing therapies. The drug was designed to treat autoimmune diseases such as rheumatoid arthritis and lupus erythematosus. [FierceBiotech, 2014]

Vercirnon (Traficet-EN): Phase I (2 studies, enrollment 54); Phase II (3 studies, enrollment 596); Phase III (4 studies, enrollment 1,492). Data from the SHEILD-1 study failed to meet the primary and secondary endpoint for secondary treatment of patients with Crohn’s Disease. [BioWorld]

Innovation always involves risk of failure. It is an art to see what the data show, and what they don’t, and which projections are the result of our wishful thinking or unsubstantiated assumptions. It may be just my impression that 14th century logician William of Ockham whispers in my ear that entities shall not be multiplied unnecessarily.