- Train Derailment results in sudden release of PCB near couples home
- Spouse of wife dies of cancer related to PCB
- Test results show PCBs in her body fat
- Wife has subcutaneous body fat surgically removed
- Subsequent review determines validating data errors
- Needless surgery due to erroneous data
A rail-road ran through the mid-western property of a couple I will refer to as “husband” and “wife.” A derailment on their property was probably much less of a surprise to them than news the rail company was hauling environmentally toxic cargo. Some of the cargo containers were ejected and at least one broke open, releasing PCBs (poly chlorinated biphenyls) into the soil which was uncomfortably close to their home and living area.
A cleanup was conducted but sometime later the husband became afflicted with a type of cancer believed to be connected to PCB poisoning. Indeed PCBs are long lived in the environment and if there were significant residues in the soil surrounding their home, they could easily have taken a chronic dose of the PCBs in one way or another. The rail and chemical owners took responsibility and assumed all expenses associated with his illness but sadly he died from the cancer and, naturally, they suspected the wife would also be carrying toxic residues of PCBs in her body fat. So they contracted a local laboratory to biopsy her body fat and test it for PCBs. The results of the tests were incredible! The report said she had absolutely huge residues of PCBs in her body fat.
PCBs are fat soluble and if one is exposed to PCBs much of it will partition into the fat. Over time some may leak out and disrupt cellular processes leading to cancer or maybe some other symptom or disease. It is also known that if, for some reason, she lost weight the residues of PCBs could be displaced into the blood and reach acute toxic thresholds in short order. Something had to be done. So she was subjected to surgical fat removal. Approximately 90% of her subcutaneous body fat was surgically removed. This indeed is a very invasive and painful procedure!
Several years later the physicians recommended she undergo another biopsy and PCB analysis to verify the operation was successful. A biopsy was taken and submitted to the same laboratory which had done the first analysis. The results startled everyone! She still had massive residues of PCBs in her body and it was recommended she have another defatting surgery because of it. Fortunately, this aroused some suspicion. The rail and chemical officials requested all pertinent raw and processed data of the analysis from the laboratory and sent it to the validation company I worked for at the time. I was assigned to the project.
It did not take long before I could see remarkable deficiencies in the how the lab work was done. I was struck by the data interpretation, the statistical processing and the nearly non-existance of any kind quality assurance or quality control. First, they didn’t follow any named analysis procedure approved by either government regulatory agencies or by inter-corporate standards. The fat tissue was subjected to some “cleanup” procedures but it never would be clean enough to be successfully measured using their gas chromatograph with a flame ionization detector. This type of detector reacts to almost everything. In order for it to work properly the PCBs would have to be almost perfectly separated from all the lipid residues of the fat. This would be nearly impossible to achieve. The actual graphics of the raw data coming out of the instrument were, in non-technical terms, slimed! There was nothing there that looked like PCBs. They also gave no graphics of actual PCB standards and since there are many types of PCBs there should have been three or five examples showing they could actually see them and tell one from the other.
To measure residue levels an analyst requires a “standard curve” generated from known levels of the actual chemical running from low to high concentrations. It should have at least five points of residue levels. It should also be in a straight line so a mathematical formula can be used to measure the sample (“unknown”) residue level(s). But this laboratory used only three points and it was not linear at all. The standard curve looked better suited for triangles.
When validating data one would expect other matters of quality control and assurance which should be standard. If a sample had nothing in it at all, the instrument should indeed see nothing. We call these “Blanks.” They show the system is clean and operating properly. Ideally blanks should be run before and after each sample series showing the system remained clean throughout the analysis time frame. There was none supplied by this lab. Also, if there was something there, the instrument should see it. This is done by running a known standard of the chemical being tested or a portion of the actual sample can be spiked and run as a “Matrix Spike” showing that if there was something in the sample the instrument could show it. None was supplied by this lab. There should also be at least some data showing their instruments were capable of reaching acceptable limits of detection, but again, no such supporting evidence was given. Ideally there should be instrument logs telling when samples (including blanks, spikes etc.) were run and in what order. Usually modern autoinjection systems will automatically generate such logs. But once again, no such paperwork was included.
Missing these quality assurance and control items did not hinder our final judgment. It was clear the lab was utterly inept and this incompetence was likely longstanding as several years had passed between the first and second analysis. This is what we reported. We were not a part of the followup decisions by our clients but expected another fat sample to be taken and submitted to a properly run and maintained laboratory. If they did this, it is almost certain the final results were very different. In the end, our data validation team could only shake our heads in astonishment how this woman was once and likely would have once again been subjected to a terribly invasive surgical procedure because of totally erroneous data.
Most “modern” labs have their own quality teams so this sort of incompetence is largely self correcting. However, some labs are under contract requiring their work to be validated by an unaffiliated third party. There are some labs who will occasionally cut corners and submit their reports simply ignoring the fact their quality goals have not been thoroughly met. And, I am sad to say, there are labs who will falsify their quality assurance data. A falsification can be hard to determine but if full documentation is supplied, it can be found. Ultimately, an unaffiliated review of data can be worthwhile. Indeed it can save money in the long run.
About the Author:
Analytical Chemist Toxicologist Biologist Expert Witness No. 719 has experience that stretches from the physiology of poison mode-of-action to consulting for environmental risk assessment and litigation. He has conducted analytical method development, data validation for RCRA/CERCLA Laboratories, environmental fate projects under FIFRA and laboratory management.