Dr. Fallon: Thank you very much for inviting me. Let’s see, looking for my talk here. Here it is, okay. So I’m going to address the question, is there evidence that repeated antibiotic therapy may be useful? I believe there is such evidence and I hope to present that clearly to you. I want to state that I’m not a member of the Infectious Disease Society of America and I’m not a member of the International Lyme Associated Diseases Society. I am however a member of the American Psychiatric Association. I’m a psychiatrist. And as a psychiatrist and as a research scientist I have specific expertise in studying human behavior and in conducting clinical trials on subjective symptoms associated with human behavior and I hope that expertise will be of help to the committee as I think that’s related to the chronic Lyme symptom story.

Now the particular aspect of the guidelines that I wish to address is the statement that antibiotic therapy has not proven to be useful and is not recommended for patients with chronic subjective symptoms after recommended treatment. I think this statement is partially correct and partially incorrect. And the part of it that’s incorrect is where it says that antibiotic therapy has not proven to be useful. The part that it is correct is that the control trials did end up concluding that antibiotic therapy is not recommended. So what does that mean? That means that there’s a difference between efficacy usefulness versus what’s clinically recommended. Efficacy was definitively demonstrated in the Stony Brook Fatigue Study that was done by Lauren Krupp. And that she found that the percentage of responders on the primary outcome measure of fatigue was over three times greater for patients given one month of IV antibiotic versus one month of IV placebo, and that was at a 0.001 level. So I could stop my talk here because here’s the evidence that repeated antibiotic therapy is helpful and this was for the subjective symptom of fatigue. But she ended up concluding that repeated treatment was not recommended particularly in light of the frequency of serious adverse events four of her patients had serious adverse events out of the fifty five.

So if an antibiotic is effective but carries risks, then what is next? Well you could look for a different antibiotic with a better safety profile or you could find a way of identifying in advance patients most at risk of developing adverse events; a biomarker perhaps or age. We know that people above the age of sixty, for example, have a higher risk of developing thrombi or you can have a careful discussion with the patient on the potential benefits and risks associated with the treatments, that’s what we as physicians do all the time because most of the treatments we give have some kind of risks associated with that.

Just a brief word on study design and clinical trials, the more homogenous the sample, the more likely you’re going to be to show a treatment effect. The more heterogeneous, the less likely you’re going to show a treatment effect, and the two homogenous clinical trials were the Klempner study on fatigue and my study on memory and working memory deficits. Klempner study, he enrolled patients based on reports of functional disability but he did not require a very cut off entry so it was a heterogeneous sample of patients. And in that study, they did not find a treatment difference.

You’ve heard about the study we did at Columbia on persistent Lyme encephalopathy. I’ll just emphasize a few things. It was indeed a smaller sample size than we wanted it. It was thirty seven instead of forty five patients so it’s underpowered in that sense to show a treatment effect. We had highly selected patients so you can’t necessarily generalize to the broader group, the patients with chronic symptoms. Our primary outcome measures were cognitions, specifically memory but we’ve looked at all the different cognitive domains, six of them and secondary measures as well. I wanted to just point out what – this is a summary slide of the results on the cognitive index. So this is the mean cognitive score at baseline at week twelve and in week 24; and what you see, the healthy control group that there’s a gradual improvement in cognition over time. The same thing occurred in the placebo group, and obviously the healthy control group did not get any IV treatment whereas the placebo group did. So the reason the healthy control group got better was because of the practice effect of taking the same neuropsychological test over time, and the same slope was seen in the placebo group. What that means is that there wasn’t much of a placebo effect but rather it was a practice effect of taking the same tests over time. If you add in the drug-treated group, what you see is a much steeper slope of improvement. From baseline to week twelve, and then after week twelve, when they’re off antibiotics, they lose that improvement and they go back down to where the other groups were. So there’s a line here would be same slope as the others, but clearly this is a different slope.

Now some people would suggest that this is actually regression to the mean because they started a little bit lower which they did but it wasn’t a statistically significant difference here between the two groups. Nevertheless, you have to wonder if it was a regression to the mean but the problem is, there is no such thing as regression away from mean, and that’s what happens. So this is more consistent with the treatment effect than regression to the mean effect.

Now the secondary measures of pain and fatigue and physical functioning, what we found, we were advised by our DSMB to address the issue of different baseline severities and to conclude that in our analysis which we

did. And what we found was that patients who had higher levels of baseline severity, an impairment or pain. We’re more likely to show a treatment effect or drug versus placebo than those who had lower levels. So for example at the higher level of pain here, the drug group dropped off dramatically and stayed improved. The placebo group didn’t show much change. In the lower pain levels, you see no difference between drug and placebo. So the way analysis is done is important.

We did however have a number of adverse events in our study that we are all concerned. Seven patients had treatment related adverse events that led to drug discontinuation or hospitalization. The most striking and our greatest concern was associated with the PICC line, two patients did have a thrombus, both were above the age of sixty. One patient had a staph infection. So this highlighted to me and I emphasize it to others that the IV antibiotic treatment is associated with significant risk and at the very least this needs to be discussed very carefully with the patient because it’s serious when it occurs.

So the conclusions were that these highly selected patients had physical impairment comparable to patients with congestive heart failure and we did not recruit people with functional impairments. So this corroborates previous findings. These patients had diffused regions of brain blood flow deficits which are shown here on that picture. The ten weeks of IV ceftriaxone therapy led to short term improvement in cognition that was not sustained. Our primary domain eventually was memory. We did not see any effect on memory. Long term improvement twenty four hour was seen in the more impaired who had pain and physical functioning problems but the risks associated with the IV treatment we’re all concerned.

Now I want to spend the rest of my time talking about the study that was done at Stony Brook by Lauren Krupp and her colleagues. So once again, here’s the evidence that repeated treatment is effective for the subjective symptom of fatigue so that’s – I’ll go back, sorry about that. So this is including all study patients. Sixty four percent versus nineteen percent. If you look just find a subset of those patients were more homogeneous biologically and take only the IgG Western Blot positive patients. The difference is even more striking. Eighty percent of responders in the drug group versus thirteen percent, the placebo group. So a biological sub group had a much higher response rate than the group as a whole.

So there have been a number of comments and critics of the prep study. One was that only one of 3 primary outcome measures showed a treatment effect. These were the three outcome measures. Fatigue severity scale, speed of mental processing and the O sp A antigen. However, the fatigue’s various scale was the only scale that was used for enrollment.

She required a predetermined severity cutoff on fatigue and it was on that scale where the treatment effect was seen. It was a more homogenous sample and that’s where the effect was seen as you would expect.

On a secondary measure which was the visual analog scale of fatigue, there was a trend also supporting the results of fatigue severity scale. In addition, at six months, the mean fatigue visual analog scale scores were better or lower in the ceftriaxone group than in the placebo group. So the secondary measure support the findings of the primary measure.

The fatigue severity scale improvement from her study was also supported by a second study which was my study. Once we were asked by the reviewers of the journal to try to analyze our data in the same way Krupp analyzed her data. We found nearly identical results. Sixty seven percent of those on drug got better, twenty five percent on placebo matching in some ways what was seen earlier. So when you find something from a second study that corroborates the first study that gives greater credence to findings from the first study.

Now what about the second outcome measure speed of processing? Well how impaired were the patients on that? The author state, the observed cognitive deficits were relatively mild. So you can’t expect only mildly impaired patients to show a treatment effect if the sample size is small. Because of insufficient initial impairment on processing speed, the success or failure of repeated antibiotic therapy cannot be determined based on this outcome measure. It wasn’t really adequately tested.

What about the clearance of the Osp A antigen which was there biological measure. Only 9 of the 58 patients actually had the Osp A antigen. There is no way that assess treatment after seeing a sample of 9 patients so therefore, this also is not a good outcome measure so the 3 outcome measures, only 1 was really appropriately designed to be tested. That’s the one we need to rely on to assess clinical outcome and that was the Fatigue Severity Scale.

Second fatigue was that there was an inadvertent unmasking that may have led to an overestimate of the treatment response.

The statement was that significantly more patients in the ceftriaxone group than in the placebo group correctly guessed their treatment assignments. Well, that would be our concern because that would possibly lead to a bias. However, you should also remember that people are more likely to guess correctly if the active treatment is robust, so if you’re really sick, and you get dramatically better, you’re gonna think, wow, I was on the active drug. So you could say it’s a good sign that the ceftriaxone patients are more likely to guess correctly because they felt better.

Conversely, if people on active drug are not more likely to guess correctly, that’s just the treatment. It’s not very effective and the authors in their paper reported of those who believed they were on active therapy, those who guessed they were on active therapy, the ceftriaxone, that people actually got ceftriaxone still had more responders than the placebo group, 83% versus 44%, and then the biological stratification as I mentioned earlier using Western Blot positivity enhancer, they respond or rate the drug or reduce the placebo that would make unmasking unlikely unless you think people who are IgG Western Blot positive are better guessers or at the second study, had nearly identical responder rates. So, for all these reasons, I don’t think unmasking really can discount the results of this trial.

Third was the magnitude of the effect on fatigue was small, that’s another critique and that’s a good question. Was the improvement in fatigue clinically meaningful, 22% improvement compared to baseline in 6 months for the drug group versus 9% for the placebo group. Is this clinically meaningful? So, in inferential statistics, there’s a method that’s used that was elaborated by Cohen to determine whether it’s statistically significant difference is clinically meaningful. An effect size of 0.2 is considered mild effect, 0.5 is medium, and 0.8 is large. Anything above 0.5 is considered clinically meaningful. In Krupp study, using a data from her study, the effect size was 1 for the difference in improvement. So, that would correspond to a large effect size so that’s likely to be a clinically meaningful effect as well as a significant effect. Just to make a [Indiscernible] [0:13:05]with IQ and cognitive testing.

If your IQ improves 1 effect size, if you’re an effect size of 1, that means you’ve gone up 15 points or your IQ went from 115 to 130, we would all be happy to have that kind of increase in our IQ. So, based on large effect size and based on unlikely percentage of improvement among responders that it’s unreasonable to conclude that the magnitude of improvement was too small to be of clinical significance. The last fatigue was that high drop-out rate have misled results, 7 patients were not assessed in 6 months and she did a great analysis in her paper with all the different possible scenarios but the worst case scenario is that all the placebo dropouts would have been responders and all of the drug dropouts would have been non-responders. That’s the worst case scenario and still she found a statistically meaningful difference between the drug and the placebo, so it doesn’t appear that dropouts really skewed the results.

Summarizing my talk here would benefit a repeated antibiotic therapy for fatigue as supported by the primary results and her studies shown here, the biological stratification shown here and the similar findings in the second study.

The conclusion from this review is that repeated antibiotic therapy for patients who are persistent in relapsing fatigue has been shown to be effective but the risks of associated treatment requires a careful discussion with the patient over potential risks and benefits, so the suggested changes to the guidelines, say, lineup here says there was no significant difference between groups and we have improvement in fatigue or pain. That’s what it currently says. I think that’s a little bit, that’s true because it was statistically significant but it’s a little bit misleading because there was a trend towards significance, so I would just put in what the authors wrote in their paper which was the change in the visual analog scale fatigue measure showed a similar pattern to the Fatigue Severity Scale with the trend towards more improvement.

Currently, it says fatigue improved in both groups at the 1 month assessment but improvement was sustained at 6 months only in the ceftriaxone group. That indeed is accurate. I would suggest adding because I think it’s helpful for clinicians to know that there’s a high likelihood of response.

At the primary end point of 6 months, the drug versus placebo responder rate with 64% versus 18% and then to also add the IgG Western Blot positive patients being more likely to respond. So, last slide, some statements were that antibiotic therapy has not proven to be useful. I suggest considering that you might change it to repeated IV antibiotic therapy is efficacious to patients with chronic subjective fatigue but the risk associated with antibiotic therapy requires careful discussion with the patient of the cost benefit ratio. Well, patients without any objective markers of disease benefit of repeated antibiotic therapy for other symptoms such as physical and mental dysfunction and cognitive impairment has not been demonstrated. If an objective marker of cognitive impairment is present; however, they may be short term benefit in cognition and sustained benefit from repeated treatments among the more impaired for pain and physical function. I think that summarizes what exists for the evidence from the studies and I thank you for your attention.

Dr. Baker: Thank you, Dr. Fallon. I’ll start with the first question, but I want to take questions from the other panelists. So if you put yourself in this situation of risk benefit which I mean doctor and treating doctor, going back to your slide where you had the patients with very high impairment, yet some adverse events which were very serious and I turned around and asked you, "Well doctor, you know, I can’t function. My life is on hold. I’ve lost my job," and you had objective findings, would you treat?

Dr. Fallon: I would definitely recommend treatment.

Dr. Baker: And let me ask a different question and this is unfair, but it’s okay to be unfair, right? The different question would be, for how long and would you recommend repeated treatments?

Dr. Fallon: That’s very unfair.

Dr. Baker: At least I announced it ahead of time.

Dr. Fallon: Well, first of all, I’m not an amateur with antibiotics.

Dr. Baker: Yes, I understand, but based on the data . . .

Dr. Fallon: Based on the data, I would say that it makes good sense to get a repeated course of treatment. Generally, if I were recommending clinically to a patient, I probably would say, "If you relapse after a certain length of treatment, you might try a somewhat longer length of treatment, might consider a different type of treatment as well."

For example, in Europe, doxycycline was found to be effective for patients with neurological disease as well. That’s obviously a lot safer than IV antibiotics, so that might be a recommendation.

Dr. Baker: Thank you for answering that unfair question and we’ll go to Dr. Moro first.

Dr. Moro: Thank you for your presentation. My question is related to the clinical trial, the purpose was to use initially about 45 patients, I think, and we hear about the problems with the size of the clinical trial. Did you run into many problems selecting the

Dr. Fallon: Well, it was very hard finding the 37 patients. Our patients, they had to be IgG Western blot positive. They all had to have well-documented Lyme disease in their charts that they brought with them. That’s hard for patients to do. They had to have current subjective memory impairments, so even though they may have memory complaints, they don’t necessarily have memory impairment, they may have verbal fluency impairment but they wouldn’t count for our study. So there were a lot of reasons why people were excluded, plus, if you’re IgG Western blot positive, you’re cognitively impaired, you have very well documented Lyme disease, would you choose to go into a placebo controlled trial where you might get placebo for 6 months?

Most clinical doctors would probably treat that patient because the IgG was Western blot positive, now granted the patients in my study were highly treatment resistant, they had to on average 2 months of IV

antibiotics before 7 months of oral antibiotics. The fact that we saw any treatment effect as remarkable, given how much treatment they previously had, so the difficulty of getting a sample of 45 reflected the nature of the clinical trial. I think, certainly in terms of symptoms, the most common symptoms these people have are the pain and the fatigue. That’s the most common. Cognitive stuff was milder than I thought it would be.

Dr. Baker: Dr. Parsonnet.

Dr. Parsonnet:In the course of preparing for this researching editor or writing up your paper, are you aware of any studies looking at use of antibiotics for chronic fatigue for ideologies or [indiscernible [0:20:07] other than Lyme Disease

Dr. Fallon: No.

Dr. Parsonnet: Based on some sort of non specific or

Dr. Fallon: No, I know that antibiotics are being studied for other purposes. I’m doing a study right now of IV ceftriaxone for refractory psychosis, not because I think it’s due to infection, but because of the [Indiscernible] [0:20:25] effect.

Dr. Parsonnet: Yes.

Dr. Fallon: I don’t know of any chronic fatigue. I know that there are studies. I know that there are some antiviral studies, but I don’t know of any antibacterial studies, it’s not my area of expertise.

Dr. Parsonnet: Right. [Indiscernible] come upon it

Dr. Baker: Dr. Medoff.

Dr. Medoff: So, I wanna separate your comments on insertion, of a new statement into the guidelines from what you were discussing, what you would do with an individual patient, so would you go further than saying that these studies indicate that there was improvement for fatigue specifically the ceftriaxone rather than talking about antibiotics in general.

Dr. Fallon: I mean be specific to

Dr. Medoff: Yes, I think you might if [Indiscernible] modify

Dr. Fallon: I would be specific. I mean, when you think about it, these trials, what have they tested? They’ve tested ceftriaxone and then one trial of doxycycline and that’s it for the controlled trials. So I think it would be

perhaps, a mistake, to make, broad generalizations about antibiotics in general given that the controlled trials really haven’t been done on subjective symptoms

Dr. Medoff: So, you might modify the insertion and state that you’ve just referred to the antibiotic treatment.

Dr. Fallon: Did I

Dr. Medoff: Yes.

Dr. Fallon: Ok. Yes. I would make it more specific. Thank you for clarifying.

Dr. Baker: Other questions: Yes, Dr. Lantos.

Dr. Lantos: What was the P value of the baseline difference between the drug and placebo groups in your trial.

Dr. Fallon: Thank you for asking that. I meant to mention that’s a 0.53 so it fell right above the margin of significance so because it was .053 instead of less than 0.05, there’s a slightly increased chance that it occurred by chance.

Dr. Lantos: Was that the change in baseline characteristics or that the change at the [Indiscernible] [0:22:19] analysis did they mention it?

Dr. Fallon: That was the change every 12 compared to baseline for progressed [Indiscernible] [0:22:25] placebo.

Dr. Lantos: My question is, what is the difference in baseline characteristics between those 2 groups? What was the P value for that? Was it similar to the 0.53.

Dr. Fallon: the 12?

Dr. Lantos: No. At baseline. So, visually, when I look

Dr. Fallon: Yes, there was. I don’t know. I think it was like .25 or something. It wasn’t as significant

Dr. Lantos: Okay.

Dr. Fallon: I have to go back and check but it wasn’t close.

Dr. Baker: Thank you.