Lithium revisited: Is there a baby in the bathwater?

At last year’s Airlie House workshop to develop new ALS/MND Clinical Trial Guidelines the focus was, of course, on MND, but there was also important input and learning from outside the field.

One of the most fascinating presentations was from an oncologist who was explaining how detailed genetic analysis of tumours was leading to an understanding of why some experimental cancer drugs appeared to only work in a small subgroup of patients. The take home message from the cancer field was that there should be more effort made in future MND trials to identify and analyse smaller subgroups of patients, in case a potentially positive effect might be missed.

A new research paper, published in the journal Neurology, raises some intriguing findings from the trials of the drug lithium that were carried out several years ago. Lithium generated a lot of excitement when researchers in Italy reported a positive effect of the drug in the SOD1 mouse model of MND. Almost as an afterthought, their research paper mentioned that they had tested the drug in a small short-term trial in patients and it appeared to have some effect.

This led to a frenzy of activity, with several lithium trials of various size and quality taking place around the world.  The MND Association funded the largest and longest of these trials and when it finally reported that the drug had no benefit, it effectively ended any lingering interest in lithium as a treatment.

A lot has changed in our understanding of MND since then, especially in the genetic factors that are involved in the disease, either as causes or affecting the speed at which MND progresses. This led researchers in The Netherlands to take a look at the genetic makeup of the participants in previous trials, focusing on two of the most common genes linked with MND: C9orf72 (the most common gene known to cause MND) and UNC13A (a gene whose different variants appear to influence disease progression). They found that lithium had no effect on trial participants with the C9orf72 gene mutation, but it did seem to have a beneficial effect in those with a particular variant in the UNC13A gene.

In collaboration with researchers in Italy and the UK, the data from two other lithium trials were analysed and the same effect was seen across all three trials. Combining the data from these trials indicated a statistically significant effect; participants with the UNC13A gene showed an improvement in survival, with a probability of 12-month survival increasing from 40% to 70%, but no effect on survival was found for participants with the C9orf72 mutation.

There is a saying that if you torture your data long enough and hard enough you can get it to tell you anything, so ‘post hoc’ analysis of old data has to be taken with pinch of salt. The researchers are also at pains to point out that the number of people in the trials with this particular UNC13A genetic variant is small and the three trials were carried out in slightly different ways, which throws additional noise into the analysis. It is also not at all clear how lithium should interact with UNC13A, though we do know that lithium is a ‘dirty drug’ that affects a lot of different cellular processes.

But the fact that the results are consistent across all three trials does mean that this effect needs to be followed up in a trial where only people with this UNC13A genetic variant are included. This will almost certainly need to be carried out across several countries to get enough people into the trial. The authors of the paper estimate 1,100 people with MND would need to be screened to have a big-enough sample with this gene variation that only appears in around 15% of MND patients. As UNC13A is not considered a genetic cause of MND but a modifier of disease progression, it is not routinely tested for outside of a research study (there is currently no genetic test for the UNC13A gene variation carried out by the NHS).

These surprising findings show that future MND clinical trials should include at least some analysis of the most common MND genes, but it also shows how vital it is to conduct rigorous trials in the first place. These recent analyses would not have been possible were it not for the collection of detailed clinical information during the trials and the collection of blood samples that could be tested many years later.