Zantac and Cancer, Round 2: Admissibility of Scientific Evidence Gets Bungled

Related articles

In December, a federal judge dismissed 50,000 Zantac cases because the scientific evidence establishing cancer causation didn’t pass legal muster. In March, a California state judge reached the opposite conclusion. What happens next?

More than 100,000 Zantac cases remain active in state courts around the country. The claim is not that Zantac (ranitidine)an antacid, per se, is dangerous. Rather, it is that, over time and intense heat, the drug could degrade to a compound called N-Nitrosodimethylamine (NDMA), a probable and ubiquitous human carcinogen at sufficient doses.

Florida’s Judge Robin Rosenberg virtually eviscerated the federal cases based on the inability of the plaintiffs to establish causal proof between Zantac and five alleged cancers under prevailing standards for evidentiary admissibility governed by the Daubert case. The remaining cases are pending in California, New York, Delaware, and Illinois, where Daubert doesn’t apply.

The strongest (albeit still weak) evidence associates five cancers with ranitidine breakdown products:  esophageal, bladder, stomach, liver, and pancreatic. Some state cases claim also include other cancers.

The first state case scheduled for trial is Goetz v. GlaxoSmithKline, to be tried in California in July.  In March, the judge ruled on the admissibility of testimony of the plaintiff’s experts, a new cast from those proffered in Florida, essentially allowing all to testify on causation, at least in part, stopping short of ruling on whether the evidence establishes causation.  The case, apparently the best in the plaintiff’s basket of cases, is called a bellwether trial.

“A bellwether trial is essentially a test case in a mass tort litigation that sets a precedent for the hundreds or thousands of cases making the same or similar allegations. The first bellwether trial will be a good indicator for out-of-court settlement resolutions. If the plaintiffs win a jury verdict, it generally increases the likelihood that cases will end in Zantac settlement agreements.”

The issues before the court include:

  • whether ranitidine degrades to NDMA outside the body
  •  whether ranitidine creates NDMA inside the body
  • whether NDMA causes cancer (general causation)
  • whether Goetz consumed meaningful doses of NDMA
  • whether the NDMA from ranitidine caused Goetz to develop bladder cancer (specific causation)

Rules of Evidence

To determine whether the experts will present legally admissible scientific evidence, California requires judges to conduct a Sargon hearing (an analogue of the Daubert hearing). Judge Evelio M. Grillo, who conducted the Sargon hearing in the Goetz case, believes the standards for admissibility under Sargon require the trial court acts to exclude expert opinion testimony that is:

  • based on materials of a type on which an expert may not reasonably rely
  • based on reasons unsupported by the material on which the expert relies
  • material that does not actually supports the expert's reasoning
  • is speculative

By comparison, Daubert also requires that judges ensure that the evidence is

  • Reliable
  • Relevant
  • Fits the facts of the case

General rules of evidence in civil cases require that experts’ testimony reach a “more probably than not” standard (also known as the 51% probability) and that they are testifying as an “expert” according to the practice of scientific experts of their professional community.

With this background, we highlight a few examples of how Judge Grillo perverts the laws of evidence to allow questionable and speculative testimony to come before the jury.

The strongest case?  The Goetz case fails, lacking evidentiary relevance, among other reasons.

Confounding Mr. Goetz’s case that Zantac caused his bladder cancer is that there are no existing Zantac pills taken by Mr. Goetz to assess the dose or the amount of breakdown product he ingested. All that is available are four generic tablets – expired for over a year.  From this evidence, the plaintiff must establish the amount of NDMA that might realistically be consumed from Zantac. Holding that all generics are the same, Judge Grillo allows the experts to testify both on the breakdown characteristics of the samples and the dose the plaintiff allegedly ingested.

A Litigation Expert” – an arbitrary judicial construction failing the Kumho Tire Test [1]

Rather than holding the experts to the customary standards of their professional, scientific community, the judge creates a new category of “litigation expert” to espouse as would any such person, irrespective of whether it comports to the consensus of the scientific community: Judge Grillo holds:

“A litigation expert is providing a litigation opinion for litigation and not for professional purposes, so the litigation opinion does not need to be "of a type that reasonably may be relied upon” for professional purposes. A litigation expert’s litigation opinion only needs the level of certainty that the plaintiff must present to prevail at trial.”

Drawing support not from case law but from a potential expert, the court provided a unique litigation standard: Claiming that because the conventional courtroom standard, requiring “within a reasonable degree of medical probability” is different from and less certain than the standards in academia or medical practice. it was admissible in a lawsuit

Statistical Significance – Fails the Reliability Standard

Perhaps this new construct manifests most clearly in accepting evidence that is not statistically significant and where epidemiological studies do not present a consistent risk level above the conventional “more probable than not” standard. In civil cases, the customary rule of law is that the expert must proffer testimony that is at least “51% certain” (the quantitative equivalent of “more probable than not”) – anything less would be speculative.

As I have written, two types of causation are required to prove a case

  • general causation: can a substance cause a disease
  • specific causation: did the substance cause the condition in this specific plaintiff.

 Yet, even courts allowing testimony where the results are less than “more probable than not” require the evidence presented to be reliable. [2]

The aim ... is not to admit only persuasive expert opinion; it is to exclude only ‘clearly invalid and unreliable’ expert opinion.”

Statistical significance, specifically a p-value of 0.05, is regarded as the level at which reliability is conventionally accepted. (In essence, one can expect similar results 95% of the time, and are assured that the results are not random).

In an epidemiological study, a greater relative risk value (over a baseline level of 1) indicates the degree of association between the substance under investigation with the disease. A relative risk of 1.1 signifies a 10% increase in risk in those exposed compared to the background level. A relative risk over 1.5 would be the bare minimum level to indicate a more probable than not causal association [3]

Another measure of reliability is the confidence interval - the range of results expected if the study is repeated. Only studies with confidence intervals exceeding 1 are statistically significant.

“Plaintiff argues that even if no single study meets the traditional statistical standard of “statistical significance,” that the aggregate effect of multiple studies that show some correlation can demonstrate general causation even if no single study meets the standard of “statistical significance.”

Surprisingly, Judge Grillo opined that he could conclude that risk ratios consistently above 1.0 [without specifying how much over one] but less than 2 suggest general causation and that confidence intervals below 1.0 do not preclude the use of the data.

Dose, Specific Causation, and Rank Speculation

While the court acknowledges the importance of dose in assessing specific causation, echoing Paracelsus, that the “dose makes the poison,” the court allowed testing done on the expired generic tablets to stand in for unexpired Zantac pills taken by the plaintiff.

One expert did conclude that “no consistent signals emerged across studies, and studies … [and] found no association between ranitidine and overall or specific cancer risk. …

I will only conclude that it is possible for NDMA formation from ranitidine to also result from reaction in the gastrointestinal tract following oral administration of ranitidine formulations…. I have provided the hypothesized mechanisms for NDMA breakdown in the human body and the potential toxic byproducts that can occur.”

Rather than denying the admissibility of these statements as rank speculation, as would most courts, under either Frye or Daubert Standard, Judge Grillo states:

“Because Benet limited his opinion to “potential” or “possibility” level of certainty, the Sargon analysis of that opinion concerns whether his reasoning supports an opinion at that level of certainty.”

Judge Grillo admitted it, even though the words “possibility” indicate speculation and a hypothesized mechanism can be held to be sound science, in contravention to the established laws of scientific acceptability

The defendant allows the Judge, is always free to cross-examine on these points, which we can expect them to do vigorously. Whether the jury will accept the speculative nature of the plaintiff’s experts remains to be seen.

[1] Kumho Tire v Carmichael extends the holding on the scope of scientific and technical experts to procedures and expertise as practiced within their profession.

[2]  Some judges allow plaintiffs to cobble together a case based on pieces of evidence below the more probable than not standard, if on the aggregate all the evidence proves the causation element by the 51% degree of certainty.

[3] The nine Bradford Hill factors which specify certain pre-conditions in inferring a causal association must also be shown.