This article is about the unsolved Tylenol and copycat poisonings with cyanide in painkiller capsules. My involvement in the case was as an analytical chemist to develop rapid methods for analyses of the potentially millions of suspect products recalled from shelves of drugstores. The hope was that such chemical testing might give clues to investigators concerning the distribution of tainted products that could lead to an arrest.
It’s been 40 years since seven people died from cyanide poisoning from adulterated extra-strength Tylenol tablets. The victims had purchased bottles of Tylenol, on store shelves, in suburban Chicago. A completely ordinary thing to do. They were tragically unaware of tampering by an unknown attacker who had replaced the contents of the pills with cyanide [1]. It was a shocking incident that completely changed sales and production of pharmaceuticals. Today, we take for granted the safety of drugs and medicines because they’re produced in tamper-evident blister packs or sealed bottles. This was not the case before 1982.
Unfortunately, this incident attracted copycat attacks in the years that followed. It was after such a copycat attack in 1984, this time on a different brand of painkillers in Westchester County, NY, that I became involved. I was at Indiana University, Department of Chemistry, Bloomington IN and I had become friends with my office-mate, now Professor Robert Lodder (at the University of Kentucky, College of Pharmacy). Rob was working on combining Near-infrared Spectroscopy with intelligent algorithms, using statistics and mathematics, for an enhanced interpretation of the data.
Seemingly out-of-the blue, Rob asked me about sampling devices for intact-capsules for near-infrared spectroscopy (NIRS). I was able to get a technical diagram and specifications sketched out and booked into the machine workshop within a day, with a shiny polished aluminium sampling device being returned to me a few days later. The major advantage of this device was that, provided it worked, represented a non-invasive analytical technique. Fortunately, it did work, so analysts didn’t need to open the capsule to find out what was inside. You could do so rapidly so the technology fit the requirements for examining individual capsules from millions recalled from drugstore shelves. Alternately, the analytical technique would apply to a pharmaceutical production line.
The sampling attachment led to a patent: Sample Holders or Reflectors for Intact Capsules and Tablets and Liquid Microcells for Use in Near-infrared Reflectance Spectrophotometers [2]. This device won an R&D100 award as one of the top 100 new products for 1985 by Research and Development Magazine.
With the sampling attachment in hand, I got to work in preparing the samples by placing varying amounts of sodium or potassium cyanide into various brands and types of painkiller. Despite the macabre nature of the study, we often laughed because of the black humor inherent in what we were doing. During a mid-point in this study, we both went out to Pofolks restaurant (Bloomington IN) with our then partners. That evening, we all laughed until our sides were sore. Anyone overhearing our conversion would have considered us to be escaped lunatics, or worse. Fortunately, no-one seemed to eavesdrop that night, or if they did, they didn’t report it to the police. I’m also happy to say that I’m married 36-years to my girlfriend on that night.
I must be one of very few individuals worldwide to have filled medicinal capsules with cyanide for a legitimate scientific purpose. I have never revealed some details, thinking that these details might be important to a police investigation. Filling capsules with cyanide was a messy process since the cyanide salts absorbed water vapour from the atmosphere. You needed to weigh, make NIRS measurements rapidly, and then store the samples in a desiccator. The jell material of the capsule became discolored otherwise. Overall, the work was 6-8 weeks from conception to the first draft of the work. I don’t think I ever worked so fast or hard, or laughed so much. We published the results in Analytical Chemistry [3].
Of all the people I’ve worked with, only Rob Lodder would have called his statistical algorithm the quantile BEAST for bootstrap error adjusted single-sample technique. It was his way of stamping the work with his unique personality. I can attest, because I was there, that he spent barely 30 min. coming up with that BEAST acronym. It was hard to keep up with him. Put simply, the BEAST takes a training set of NIRS data for known untainted capsules of painkiller. A hypercylinder is used to encompass the spectral spread of untainted capsules. If NIRS then measures an unknown capsule and the spectral data is more than a statistically significant (3σ) distance from the hypercylinder, we suspect the unknown capsule of being tainted with a possibly dangerous foreign substance.
Conclusions
Initially, the intent of our research was to develop a rapid method for investigators to work through millions of recalled products in order to help track down the murderer or murderers. For the 1982 Tylenol case, the US-FDA (Food and Drug Administration) tested 2 million capsules for evidence of tampering. Unfortunately, the research and development time from laboratory to scientific paper to field usage made, as it often does, this impracticable. The NIRS technology that we worked on some 40 years ago has become a small gear in a very large suite of platforms for quality control and assurance in pharmaceutical manufacturing that has allowed our modern society to have confidence in the security, safety and efficacy of modern drugs and medicines.
Written with StackEdit.
———
[1] R. A. Vargas, Tylenol murders: daughter tells of toll of unsolved killings, 40 years on, The Guardian, Oct. 02, 2022. Accessed: Oct. 23, 2022. [Online]. Available: https://www.theguardian.com/us-news/2022/oct/02/tylenol-murders-chicago-illinois-unsolved-1982
[2] R. A. Lodder, G. M. Hieftje, and M. Selby, Sample holders or reflectors for intact capsules and tablets and for liquid microcells for use in near-infrared reflectance spectrophotometers, US4882493A, Nov. 21, 1989 Accessed: Oct. 24, 2022. [Online]. Available: https://patents.google.com/patent/US4882493A/en
[3] R. A. Lodder, Mark. Selby, and G. M. Hieftje, Detection of capsule tampering by near-infrared reflectance analysis, Anal. Chem., vol. 59, no. 15, pp. 1921-1930, Aug. 1987, doi: 10.1021/ac00142a008.
Very interesting, Mark!