Ultragenyx filed UX111 for accelerated approval, took a complete response letter in July 2025, went away, and came back to win STANDARD FULL approval on 17 September 2026, two days before its target date. Full approval is a higher bar than the accelerated pathway originally sought, on a 23.5 point higher Bayley-III cognitive score over natural history, p<0.0001. It is the first ever treatment for its indication. It is not the exception: of the four refusals on this board whose FDA letter is published, three were approved afterwards, and the one that was not is the one whose letter says the registrational trial failed its primary endpoint. A resolved refusal is evidence of execution and an unresolved one is evidence of a problem, and this record tells them apart because every event is typed and dated.

On 17 September this record read hereditary angioedema as a crowded indication that was not really a contest: deucrictibant treated an attack on demand, lonvo-z was a one time prophylactic, so the collision was a split. Pharvaris then reported CHAPTER-3, which puts deucrictibant XR into prophylaxis, and the reading inverted. The indication now carries 4 drugs (Deucrictibant IR, Deucrictibant XR, Icatibant (Firazyr), Lonvo-z) and Pharvaris is competing against Intellia with its own second asset. Nothing was corrected here, because nothing was wrong: the world moved and the record moved with it, one week apart, both states dated. That is what the as-of on every claim is for.

Regulatory events are typed, so adverse history is a class and not a judgment call. 7 drugs carry at least one adverse event between them: Apitegromab (Isembyld) (Complete response letter); Deramiocel (Complete response letter, PDUFA extended, indication narrowed, Advisory committee voted 9 to 3 against); Fayuvi (rebisufligene etisparvovec-hopf) (Complete response letter); Lonvo-z (FDA request for preclinical fetal data); Molgramostim (Refuse to File, insufficient CMC, PDUFA extended 3 months, major amendment); Pixclara (floretyrosine F 18) (Complete response letter); Zipalertinib (REZILIENT3 reported 3 deaths from septic shock). Building the scorer forced a taxonomy fault into the open. One label, adcom_declined, was carrying both “the FDA is not planning an advisory committee”, which is favourable and removes panel risk, and “the advisory committee voted 9 to 3 against”, which is the opposite. Scoring those alike would have been wrong in both directions at once. The taxonomy now separates them, and a trial safety signal is distinguished from a request by the regulator, because the actor is different.

The query asks for a person who shipped an approved drug where that drug is the regulatory precedent for a drug now pending. It returns 4 rows from 2 people. Berndt Modig and Jochen Knolle both connect to icatibant, and icatibant is precedent for both Pharvaris assets. Knolle is named as an inventor on icatibant and ran R&D at Jerini; Modig was its chief financial officer through the 2005 listing and the 2008 sale to Shire. THE SCOPE MATTERS AND WE STATE IT: this record holds 280 named executives, and the other 278 (Adam Scott Levy, Adrienne S Brown, Agosti Marcello Pignagnoli, Akshay Vaishnaw, Alexander Hardy, Alexander M Gold, Ali Behbahani, Ali J. Satvat, Amrit Ray, Anat Hakim, Andrea Ellis, Andrew Lo, Andrew R Allen, Andrew R Robbins, Andrew Stewart Plump, Anna Nijdam, Anne Erickson, Anne Lesage, Anne-Marie Martin, Anthony A. Florence Jr., Anthony Bergmann, Anthony Rosenberg, Arlene Morris, Arturo Molina MD, Asif Ali, Asuka Miyabashira, Athena Countouriotis, Awny Samaan Botros Farajallah, Barbara W. Bodem, Berndt Modig, Betty D Larson, Birgit C Schultes, Braden C. Parker, Brian Foard, Brian Goff, Brian Mueller, Brian Taylor Slingsby, Bruce D Broussard, Bruce Trapnell, MD, Bryan Giraudo, Carmen Chang, Caroline Litchfield, Caryn Parlavecchio, Charles Greg Guyer, Charles J Homcy, Chirfi Guindo, Christian Behrenbruch, Christine E Seidman, Christophe Pierre Weber, Christopher W. Cain, Cole Pinnow, Corazon (Corsee) D. Sanders, Corrine Savill, Craig McDonald, MD, Cristin Hubbard, Dalton E. Smart III, Daniel G Welch, Daniel Skovronsky, Daphne Karydas, Darren Charlton Smith, Darren Joel Patti, David A Ramsay, David A Ricks, David B Musket, David D. Meek, David E I Pyott, David Hallal, David L Lowrance, David Lebwohl, David Malcom Rodman, David Michael Williams, David N Gill, David Norman Cade, David P. Ryan, David R. Maraldo, David W. Nassif, Dean Y Li, Deborah Dunsire, Dennis Karl Huang, Derek DiRocco, Dinesh V PH D Patel, Diogo Rau, Donald A Zakrowski, Douglas A. Dachille, Douglas M Baker JR, Earl M Collier JR, Edgardo Hernandez, Edward J Dulac III, Edward T Mathers, Elaine J Heron, Elaine Mary Shannon, Eliana Clark, Elisabeth Bjork, Elizabeth M Anderson, Emiko Higashi, Emil D Kakkis, Eric Aguiar, Eric Crombez, Eric Dozier, Eric Warren, Erik Harris, Erin Burkhart, Es-Johansson An van, Evan Kearns, Forest Baskett, Frank Litvack, Frank Mccormick, Frank Verwiel, Fred E Cohen, Fred Hassan, Gabriele Ricci, Gabrielle Sulzberger, Gary Clark Biddle, Gaurav K Gupta, George Eric Davis, George W Dunbar JR, Georgia Keresty, Giles Richard Platford, Glenn Sblendorio, Gregory R Friberg, Hannah Valantine, Hao-Yuan Chuang, Harold E Selick, Hendrik Peter Scholl, Henry J Fuchs, Howard Horn, Ian T Clark, Ilya Yuffa, Inge G Thulin, Ita Lu, J Erik Fyrwald, Jacquelyn L Sumer, Jad Adaimi, James Basta, James C Momtazee, James J. Ferguson III, Jann Elizabeth Skinner, Jean Luc Butel, Jeffrey Alan Jones, Jeffrey S. Flier, Jennifer E. Cook, Jennifer Marie Smoter, Jennifer Michaelson, Jennifer Zachary, Jesse Goodman, Jessica Sachs, Jing L. Marantz, Johannes Gerardus Christiaan Petrus Schikan, Johannes Jacobus Oosthuizen, John Edward Robinson, John L. Green, John Leonard, John M Leonard, John Maraganore, John Michael Longo, John Richard Pinion, Jon Congleton, Jon Congleton, Joseph S McCracken, Joshua Reed, Juan R Luciano, Julie So-Young Kim, Junlin Ho, Karah Herdman Parschauer, Karen Jean Ferrante, Karen Krasney, Karimah Es Sabar, Kathleen R. McCabe, Kathy J Warden, Katie Peng, Keith Woods, Kenneth L. Custer, Kimberly A. Reed, Koichiro Kimura, Koji Hatsukawa, Kristina Burow, Lara Haddad RPh, RAC, Lauren Rusckowski Duprey, Lewis T Williams, Linda Marban, Lucas Montarce, Mani Sharma, Margaret Kayo, PhD, MBA, Maria Rivas, Maricel Apuli, Marie Elizabeth McDonald, Mark Alexander Nelson, Mark Awadalla, Mark J Enyedy, Mary Ellen Coe, Mary Kay Fenton, Mary Thistle, Masami Iijima, Matt Pauls, Matthew K Fust, Matthew Pauls, Matthew Ros, Maykin Ho, Melissa Seymour, Michael Gilman, Michael Narachi, Michael P Dube, Michael T Maurer, Michel Bruno Orsinger, Mike Kelliher, Miki Tsusaka, Milano Furuta, Mittie Doyle, Mo Qatanani, Mohamad Makhzoumi, Muna Bhanji, Naarden Jacob Van, Nadim Ahmed, Nadim Ahmed, Natalie Anne Marie Furney, Nathan L. Mata, Neda Aghajani Memar, Neil Kumar, Nevan C Elam, Nicola Deidre Petal Greenway, Pamela J. Craig, Patricia F Russo, Patrik Jonsson, Paul Edward Walker, Paul Gisbert Auwaerter, Paul Rothman, Paulus Stoffels, Peng Lu, Peter Evan Harwin, Philip J Gotwals, Rachael Easton, MD, Ph.D., Ralph Alvarez, Ramy Riad Ahmed Ibrahim, Randal W. Scott, Rashmi Virendra Ramchandani, Raymond Dennis Pratt, Rhonda Janice Pacheco, Richard A Meier, Richard Brudnick, Richard J Hawkins, Richard R. DeLuca, Rick Yang, Ricky Sun, Risa J Lavizzo-mourey, Robert Glassman, Robert J. Hombach, Robert M Davis, Robert M Plenge, Robert Matthew Lutz, Ronald J Daniels, Sanat Chattopadhyay, Sarah A O'Dowd, Scott D Sandell, Shehnaaz Suliman, Srinivas Akkaraju, Stephen Mayo, Stephen W Webster, Steven Gillis, Surendralal Lanca Karsanbhai, Teresa Marie Bitetti, Theodore Alan Huizenga, Thomas H Glocer, Thomas Trimarchi, Timothy P Walbert, Todd Shegog, Vikas Sinha, Viviane Monges, Wan-Shan Chen, Willard H Dere, William D. Waddill, William J Chase, William R Jellison, Wim Souverijns, Xiao-Hui Chen, Yasmine Wasfi, Yoshiaki Fujimori, Yu-Hsin Lin, Zosia Piotrowska, MD) carry an employment edge and no prior approval that this search found. That is a searched and found nothing result across 280 people, not a survey of every executive in the sector, and it is not evidence that nobody else in biotech closes a loop like this.

Every claim in this record carries the address it came from and the words it was read from: 467 of 467, with no exceptions. The dates were the weak point, because they sat on the drug as properties rather than as claims, and properties carried no provenance at all. Re-sourcing them found that the most important number on each row was the least evidenced thing in the file. 6 were re-read against the sponsor's own release this week. 8 are carried forward from 17 September with the reason recorded per date: 0 sit behind walls that refuse automated access, proved as a refusal rather than assumed from a failed fetch, and the rest were read in full and the date is genuinely not on the page this record cites. None of the 8 was deleted. Method, so the check can be repeated: KXCO fetched each sponsor release on 2026-10-05, extracted the sentence carrying the date, and stored that sentence against the drug; where no such sentence could be obtained the date kept its 17 September provenance and the reason was recorded. A page that prints every date as though all were equally checked is the normal thing, and it is the thing this room refuses to be.

Baxdrostat carried a target of 18 May 2026. Nothing in this record supports that value: the AstraZeneca release it cites carries no date at all, and the FDA approval history gives 15 May 2026 as the decision. The unsourced target was removed rather than corrected to the sourced one, because we never knew 18 May was the goal date and replacing it would have invented a second fact to cover the first. The same test then caught a worse case: 6 drugs held a target that was simply a copy of their approval date, which made them look like decisions that landed exactly on target. A difference between a number and a copy of itself is not a measurement. All 6 were cleared and one genuine measurement survives, Fayuvi at 2 days early.

Both cohorts were searched again under one rule: phase 3, the drug named, the sponsor agreeing, no access or extension programmes. The approved cohort went from 0.12 registrational trials per drug to 4.25 and the pending cohort from 0.83 to 3.50. The gap closed and the sign flipped, approved +3.38 against pending +3.08. The formula now passed. What it produced was a board topped by CagriSema on 22 registered phase 3 studies and Orforglipron on 18, because Novo Nordisk and Eli Lilly can afford to run them. The trials term counts studies, so a ranking built on it measures which sponsor is largest. The formula could be made to pass its test and passing was worse than failing, so it is withdrawn rather than repaired. Its components are kept and shown per drug because each is a counted fact. The board is ordered by regulatory standing and implied upside, from dated events rather than from a model. Correction, 30 September 2026: three of those CagriSema studies, RENEW 1, 2 and 4, test cagrilintide alone and were removed from the record. CagriSema now stands on 19 and the pending cohort on 3.25 trials per drug. The reasoning is unchanged.

This room carried an execution score, trials plus twice precedent plus twice closed loop minus twice adverse events, and published it failing its own backtest. The test compared the mean score of the approved cohort against the pending one and the approved names came out lower. We wrote that up as the formula being unable to credit back a refusal a company had overcome. That explanation was wrong. Traced on 23 September, the gap was 1.45 points and the adverse term accounted for 0.08 of it. The rest was us. ONE of the 8 approved drugs carried a registrational trial in this record and NINE of the 12 pending ones did, because the record was built as a screen of live filings, so the pending names were researched and the approved ones were entered as outcomes. A test between a researched cohort and an unresearched one measures which cohort was researched.