Slides
Transcript
So I thought I was going to be following Wouter and of course thought that’s a little unfair. Everybody else gets to debate each other, but I have to debate Wouter. And then I found out that I get to go first, and so he will have the final word. But we will try to give something to think about anyhow.
So what we’re really talking about in this session is how to deal with the patients who are conventional imaging negative, negative brain MRI, negative conventional CT myelogram, and what to offer them and what the options are. So that’s kind of what we think of as conventional imaging negative.
I want to point out that I think this is actually most patients. Hosoya and colleagues put together a consortium of Japanese hospitals, 11 of them, prospectively enrolled 100 patients with orthostatic headaches that they suspected based on clinical symptoms had CSF leaks. When they did the MRIs, what they found was that 69 out of the 89 were negative. And when they did a variety of spinal imaging techniques, they found that ultimately 84 of the 100 were negative. And so overall about 80% of prospectively enrolled patients with orthostatic headache will be conventional imaging negative.
And it can be dangerous to tell those patients that “You do not have a leak and you should not have any treatment oriented towards a leak.” Part of what’s developed since Hosoya’s study is of course looking for CSF-venous fistulas. When do you do lateral decubitus imaging?
But I’m actually even more interested in what do you do with those patients who are leak negative or imaging negative no matter what you do, because it’s still the vast majority of patients. Some of those who are conventional imaging negative, maybe 15% of them will wind up with CSF-venous fistulas on imaging, but the rest still will be imaging negative.
And if you have high blood pressure and the doctor suggests something and it does not work, the patient just lives with that. You have systolic pressure of 140, I’m going to give you enalapril, your blood pressure still 140. Most of those patients, they don’t go out and look for an alternative treatment. They just live with it if their doctor says that’s what’s going to happen.
But if you have pain and the doctor tells you there’s nothing to be done here, you don’t go home and just live with the pain. Pain activates motivational circuits, emotional circuits in the brain that drive you to make a plan to do something. You will do something when you have pain. If you have really bad osteoarthritis of your knee and the knee surgeon tells you “You are too young to have a knee replacement.” — you don’t just live with that. You go out, maybe you start getting injections in your knee, maybe you start on chronic opioids, you go look for another surgeon, you do something.
And the patients with orthostatic headaches, they’re also going to do something when you tell them, “Oh, this isn’t a leak. There’s nothing to do. You don’t need to do a myelogram to look for a CSF-venous fistula.” They’re going to do something.
This patient came to me and was interested in being evaluated for a leak, was referred by an outside neurologist. She was a 45-year-old woman whose pain started at age 42, complained of bilateral occipital pain and also pain into the tongue, all of which were better when flat. She was treated in the pain clinic at Stanford for both glossopharyngeal neuralgia and occipital neuralgia. Also noted comorbid tinnitus, nausea, and fatigue. One of the fellows in the pain clinic gave her a 48-hour flat test, which we’ll talk a little bit more about, and she reported quote that her “symptoms were 75% better.” Notice how non-compelling that is to just have that as a simple sentence in the note.
She was evaluated for a leak. Her MRI was negative, her CT myelogram was negative, but they noted that her total protein when they did the myelogram was elevated at 71, suggesting that there might be a diagnosis other than occipital neuralgia and other than a primary headache disorder. And we did what we do at Stanford with such patients. I presented her case to a group of headache neurologists and neuroradiologists and we voted. I was the only person who thought we should offer this patient a patch, and so we did not. If at Stanford we have a case like this, if no one else thinks this person should be patched, they don’t get a patch.
And the feeling of the group was this protein was high enough that the patient deserved a neuroimmune evaluation. And so she was referred to that group and they did their work-up, decided there was nothing neuroimmune going on. They did not send the patient back to me, and that patient went out and did something else because that’s what patients with pain will do. So she went and saw an outside pain doctor who repeated her occipital nerve blocks, and when she got two days of relief, they decided to put in an occipital nerve stimulator. And this is the MRI after the occipital nerve stimulator misadventure. You can see there the residua of a needle track into the cerebellum leading to a hemorrhage. Leaving her neurologically in quite bad shape.
And so the point is there is a risk to the things that we do and there is a risk to the things that we do not do. And while this is the only lead I’ve seen put into the brain, I have seen plenty of people when they were told that they were not leaking despite their orthostatic symptoms go on and pursue craniocervical instability in the setting of imaging that appears normal to me and the radiologists who I have evaluate it. I’ve seen them go on and have their heads fused to their neck. I’ve seen people have of course posterior craniotomies for ostensibly Chiari. People will do dangerous things for their orthostatic headache and so we should be thinking about that when we decide what we might do and the risks, benefits, and alternatives for these patients.
I wanted to offer her a patch and what kind of data, what kind of outcomes might we have seen had I offered her a patch? We started to follow that between 2016 and 2018 using validated metrics of outcomes of various symptoms focused on the core symptoms of CSF leak: head pain, neck pain, tinnitus, dizziness, fatigue, neurocognitive deficits, and a PROMIS measure of global health. The PROMIS measure of global health is divided into a physical and a mental subscale. The important thing about the PROMIS metrics is a score of 50 is the US mean and 10 points on a PROMIS scale is a standard deviation in the population. So when you go up 10 points on a PROMIS scale, you’ve moved up a full standard deviation in terms of your global physical health according to the US population.
And between August of 2016 and November of 2018, we had 85 patients who completed those measures before receiving a patch when they were imaging negative after MRI and CT myelogram. Oh, excuse me. 85 patients who presented with symptoms of a CSF leak, prospectively gathered, who underwent MRI and CT myelogram, completed measures beforehand, subsequently underwent at least one patch, and completed at least one set of measures afterwards.
Among those 85 patients you can see that in fact the duration of symptoms in this cohort is an average of eight years. So these are very different than the patients in the meta-analysis that is shown in JAMA Neurology by D’Antona and colleagues where the average duration of symptoms in this oft-cited paper is 31 days. So this is just a very different set of people. Most of the patients I see with chronic orthostatic headaches are not in their first month of symptoms. They have chronic symptoms and those symptoms are generally quite severe.
You can see that 83% reported orthostatic head pain and their global physical health at baseline was a 32, which is essentially two standard deviations below the mean. Their global physical health is at the fourth percentile, similar to the global physical health that you see in patients with lung cancer or AIDS. And they’d been living that way for an average of eight years. Their Headache Impact Test-6 score, their HIT-6, is in the severe range and their PROMIS fatigue is a standard deviation and a half above the US norm, with their fatigue being in the 95th percentile comparable to cohorts published with advanced multiple sclerosis and with chronic fatigue syndrome. Really identical.
After doing MRIs and CT myelograms, conventional CT myelograms of the 85, roughly 20% (16) met ICHD-3 criteria, either having positive imaging or a low opening pressure, but 81% would have failed to meet ICHD-3 criteria, having essentially normal imaging and/or a normal opening pressure. So we have again 20% conforming to ICHD-3 criteria and 80% not. They underwent an average of 3.6 epidural patches, reflecting my obstinacy and persistence.
Following those patches, when we looked at durable outcomes, changes in those metrics from pre-patch to the last assessment, an average of 377 days (over a year after their last patch), what you see is that their PROMIS physical health is significantly improved, highly statistically significant. Their mental health beneath that also highly significantly improved. Their HIT-6 is improved. Their Neck Disability Index has improved. Their fatigue is improved. Nausea and vomiting have a non-significant tendency to improve. And so they really improve across a broad range of measures.
Is this clinically meaningful? This paper reported a threshold of something like 2.8 points for what is a clinically meaningful change in the PROMIS global physical health for an individual. We looked at what percentage of our patients actually had not statistically significant improvements, but clinically meaningful improvements.
Of the 16 people who met ICHD-3 criteria, 13 had clinically meaningful response over a year after having been patched despite eight years of treatment. I’m kind of surprised by that when I think about it. Why should someone with an imaging proven leak for eight years actually get that much better in a sustained way from a pretty limited number of patches? I think all of those defects are epithelialized. I don’t think any of them are really being sealed, but nonetheless, this is compelling data that they really do benefit even after years of symptoms.
But what about the 81% of patients who did not conform to the ICHD-3 criteria, all of whom are imaging negative and had normal opening pressure? They also underwent an average of 3.6 epidural patches. And they had a 54% clinically meaningful response rate. Of the 69, 37 went on to have a durable clinically meaningful response despite being incredibly sick for a long time.
So if we look at the clinically meaningful response rate among those who met ICHD-3 criteria, the response rate is about 81%, and in those who did not meet criteria, the response rate is still 54%, which almost meets statistical significance. I’m sure if we had more patients, we would see that you really do get a higher response rate when somebody’s ICHD-3 positive. And so if you want to maximize your success rate, only patch the people who are ICHD-3 conformant. But if you want to help the highest number of people who can meaningfully benefit, don’t do that. 73% of the clinically meaningful responders would never have received the treatment that helped if patching had been contingent on positive imaging or a low opening pressure.
So what can help us identify patients who are most likely to have a clinically meaningful response among patients not conforming to ICHD-3 criteria? We looked at the measures that we evaluated before we patched these patients, and I want to point out a couple of things.
Number one, their primary report of whether their head pain was postural did not significantly predict whether they were going to benefit a year after treatment. And similarly, how much time they could be upright before their symptoms started had no predictive value at all in whether they were going to be better a year later. Nor did migrainous symptoms, photophobia, phonophobia, nausea, vomiting, stigmata of primary headaches, things that would cause many a headache neurologist to say these patients are having migrainous symptoms — we should not offer them patching. Those had no negative predictive value. They did not predict photophobia, did not predict a lower likelihood of benefiting.
And in contrast, how good their head pain felt after being flat for either an hour, or how low their head pain got after being flat all night, were strongly predictive of the likelihood of having a benefit over a year after their last patch — a clinically meaningful one.
So, let’s look at that raw data. This is what it looks like. It’s a lot, but I’m going to try and break it down for you. What we have here on the x-axis is the pre-patch headache severity after laying flat an hour. In other words, how low did their head pain get after they stayed flat for an hour? On the y-axis, we have their improvement in that PROMIS global physical health over a year after their last patch. This is the line of no change. The graph is skewed to have that line lower in the graph because we have many more people improving and improving by greater amounts than we had people declining over the time, which is reassuring.
This is the line above which, if you’re one of those green dots, over a year after treatment — you are still at least a full standard deviation better than the norm of global health for the population than you were at the beginning. And that’s the line of declining by a similar amount. You’ll notice we don’t have any dots really below that. And so, in fact, if we go back to that, the number of dots above the upper green line outweigh the number of dots below the lower green line by a ratio of 13 to 1 — favoring treatment.
And if we look at the people whose pain severity went down to a zero or a one after being flat for an hour, what you see is almost everybody improved and stayed improved for over a year. And roughly half of those people in fact improved by more than a full standard deviation, not just this clinically meaningful improvement as described by that paper. Half of the patients in this group had improvements of more than a standard deviation over a year after being treated. And in fact, really when you look at people whose head pain went down to a zero, one, or a two, that holds true.
So for these people, I think this is really clearly clinically meaningful. It’s not just that they met an arbitrary threshold in a paper. A full standard deviation is like having the US mean income average is about $100,000 a year. The standard deviation of that is $30,000 to $40,000. Do you think it makes a big difference to you if your income goes from $100,000 to $140,000 or if your SAT scores go from 1,200 to 1,400? It’s a big change. And that’s what the change is when you go from the mean to one standard deviation above the mean.
So in conclusion, it’s hard to explain the statistically significant relationship between relief when flat, two separate measures, and the subsequent clinically meaningful improvement among ICHD-3 negative patients with a placebo response or regression to the mean or the Hawthorne effect. And among patients with chronic disabling symptoms, these weren’t people who were just not doing well. Disabling symptoms suggesting CSF leak and whom imaging does not confirm a leak. 54% of those will show clinically meaningful improvement at long-term follow-up after an average of 3.6 epidural patches and two factors specifically associated with CSF leak. More complete resolution of head pain upon being flat for an hour or overnight significantly predicted the likelihood of being a clinically meaningful responder. And most of the patients who experienced clinically meaningful improvement came from the ICHD-3 negative group by a ratio of 3:1. So think about what your goals are. A high rate of response versus helping the most people.
I want to add one little coda to that. What if patching fails in these people who are ICHD-3 negative? I had a patient not included in that cohort but with who I shared with Wouter. So I’m going to bring it up. A 23-year-old woman with a history of pediatric epilepsy, factor V Leiden, on hormonal contraceptives, who in October of 2018 had a grand mal seizure on an airplane from SFO to Boston flying off to college and within 3 days started having terrible orthostatic headaches. Bad neck pain, essentially became bedridden. Had to drop out of college, return home. It was assumed she had a concussion because of the headache and difficulty concentrating that was persistent, but noted that she always felt better in the morning. She was referred to me for evaluation. Her MRI, full spine and brain, and conventional CT myelogram were all totally normal.
We did give her the 48-hour flat test where we asked them to rate the symptoms of a CSF leak from 0 to 10 after being upright all day, then spend the next 48 hours essentially laying flat, only getting up to go to the bathroom. These were her scores. Her head pain went from a seven to a two. Her neck pain went from a four to a zero. All of her other symptoms went to a zero when she stayed flat. And this was really the reason why I started giving this test to people. Not because I think time will tell does it have some predictive validity, but what it is is it’s a much more compelling way of measuring orthostatic symptoms than that resident saying she felt 75% better when flat. If it matters, you should measure it and you should write it down. And so that was the purpose of that.
Here’s her brain MRI. She’s 21. I’m not sure that her upgoing pituitary border really is meaningful in a 21-year-old and there’s plenty of people with a suprasellar distance among healthy controls that are 2.38. There’s no venous distension, no venous distension sign really. So I patched her. She had a complete recovery for roughly a year, went back to college, resumed her studies, and then her leak came back. So I patched her again, complete relief, but this time for only eight weeks, but she had that nice year. So I patched her again, again gets better like six weeks. So I think, okay, now time has gone on. I know about CSF-venous fistulas now. We do lateral decubitus myelograms, negative. But there’s little hazy funny nerve roots. So I try putting fibrin glue on those. No relief.
And then I think maybe she’s leaking from that myelogram we did originally. Maybe she just developed a chronic post-puncture from me. Maybe I’m the problem. Following the meeting in Hawaii, at Stanford, we redid our MR protocol. And so now we’re getting these 3D T2-weighted MRIs on everyone. And we got that sequence on her which we had not previously had. And when I looked at it, I was wondering if this was anything. I don’t know. Is that an artifact? It’s not. To my surprise, I thought maybe that’s a bleb. She never had a myelogram at L1-2, which is where that thing is. And I’d asked her before, “Have you ever had a needle in your back?” “No, I’ve never had a needle in my back.” So, let’s take a look at this. That’s what it looks like blown up a little bit. And if we scroll back and forth through it on the sagittal, it’s going to look like this. All right. So, I look at that. I’m not sure it’s real or not. Is it a vein? Is it a real bleb? Does anyone — Jürgen — do you want to opine? Is it a real bleb? Am I getting faked out? I don’t know either. All right. This is, by the way, what it looks like on the axial.
So I go back to her and I say, “Are you sure you never had a needle in your back before the CT myelo?” “No.” “Ask your mom.” “Mom says I didn’t.” I said, “I think it’s a bleb, but I’ve never seen that in somebody who didn’t have a needle in her back. Go back into your hospital records into the lab section and see if you’ve ever had any labs that have CSF.” So it turns out in her pediatric epilepsy when she had her first seizure at age one, at age one she had a lumbar puncture to evaluate her for that first seizure. So I looked at that. I patched it. I thought maybe I was seeing some little contrast enter the epidurals or intrathecal at that spot. I’m not really sure. So I sent her to Wouter and said, “Wouter, would you consider exploring that area?” He did. He found no bleb. That was 18 months ago and she’s been thick since and graduated college.
So why I favor treatment? The risks of not treating should not be assumed to be trivial. Many patients improve and some improve dramatically. Those having dramatic improvements more than a full standard deviation outweigh those declining by the same amount by a ratio of 13 to 1. And this is something we don’t write about a lot. The response to treatment, the fact that she got better for a year from a patch — response to treatment is informative and it changes the way we deal with people and it leads to additional therapies. And that’s an issue that I think the field has to address. And with that, if you want to see the paper from the publication that I went through to actually look at the data and see where I’m misleading you, you can find it there. And if I’ve misled you, it’s my mistake.
Controversies in SIH: Dr. Wouter Schievink on Treatment with Confirmation of a Leak
So this is the clinical dilemma, right? Just like what Dr. Carroll talked about, there’s a patient who has positional headaches. I usually ask them, how do you feel in the morning when you wake up after a good night’s sleep, and they say, “Yeah, that’s when I feel my best.” And they have a normal brain MRI and they have normal, you know, what we call conventional imaging. So what do you do with a patient like that, right?
In our practice, the vast majority of people are either referred by a headache physician or they’re what we call self-referred. And that does make a difference, but there are different things that you can do for a patient like that — provided they’ve gone through some other type of headache evaluation and possible treatment. So you can tell them to drink more caffeinated beverages. Most patients who have a CSF leak do report some benefit of increasing their caffeine intake. And in the US we have these caffeine pills that are called NoDoz. So you can try that. You can try an abdominal binder. We give that to patients for free in my office. At the end it’s kind of like in the old-fashioned way, you don’t really know what’s going on and you just write them a prescription for some antibiotics. So it’s a little bit like that. But I do also think that there’s a possible predictive value to that. Or you can just go on Amazon and order an abdominal binder. It’s about $17-$18 or so.
Or you can do what some patients and doctors call a blind blood patch. This is Dr. Maya doing a blood patch. Obviously, it’s not blind, right? It’s kind of a stupid description. It’s just non-targeted. So you can do that. And this is from Dr. Carroll’s paper, and you can see that there’s a real significant difference in the headache impact. What is the T stand for? Test. Yeah, but still, right, if you go, and it’s a pretty big arrow, but it’s still right. So people are clinically significantly better, but it’s not necessarily that they’re cured. But what I think is really important from Dr. Carroll’s paper, as we can see at the bottom there, is that there were no really adverse effects of him doing three, four blood patches on these patients — the majority of them not having any objective evidence for a CSF leak. So that’s really important.
Then I think it was maybe Dr. Amrhein who mentioned this paper earlier. It’s a patient population that was treated by Dr. Louy, who’s one of the anesthesiologists that we’ve been working with for the last 25 years. And it’s actually, even though it was published I think in 2020, it’s really old data. It’s between 2001 and 2010 when we first started treating a good number of patients with CSF leaks.
If you read the paper, you know we kind of get away from the paper that we use ICHD-2 criteria to include patients with that. But then when you read that, he says that the headaches are consistent with the definition of a CSF leak. But also he says if the clinical symptoms were not classic, the diagnosis was presumed if the patient had a confirmatory physical exam in which increased intra-abdominal pressure resulted in improvement or resolution of their headache or symptoms. This was accomplished with the patient standing or sitting, and the examiner, meaning Charles Louy, wrapping his arms around the patient’s abdomen and progressively squeezing.
So that’s, you know, and look, if you ask me, would you rather have this 48-hour Stanford flat test or what we somewhat jokingly called the Louy Manoeuvre? I don’t really know what my answer would be. But I’m just saying that paper, the vast majority of those patients at least had no objective—they might have had a leak, but there was no objective evidence for a CSF leak.
So, as you probably know, when we started writing these ICHD-3 criteria, we kind of went overboard a little bit and we made it a purely radiographic diagnosis basically. While ICHD-2 criteria was really mostly a clinical diagnosis with some radiographic confirmation, and one of them was that the headache would resolve within 3 days of epidural blood patching. Obviously that’s not something that happens all the time. So we kind of kicked that out for ICHD-3.
Also, Dr. Carroll had no complications. We did have some complications. There was one patient I remember really well who woke up and couldn’t move his legs after the blood patch, and he called the OR. I was going to take him to the OR. The patient first went to the MRI, and when he got out of the MRI he had regained full strength. But you know that was kind of a scary situation as you can imagine. Also, it was a young man who had no objective evidence for a leak. And then we had another patient who for many, many years suffered with cauda equina syndrome, was incontinent of bladder and bowel. But some of these patients had really high-volume blood patches, I think in this series up to 124 milliliters of blood. So that’s a lot, a lot of blood.
But there are some risks. So I think the risk of doing a blood patch in somebody who has an obvious leak is really, really low. But doing it in somebody who does not have a leak, the risk is quite a bit higher. And you also put yourself at some risk even by just ordering a blood patch.
So this is a patient who had an epidural blood patch. And what happens occasionally is that the blood is injected subdurally. Some anesthesiologists feel that that’s a better way of treating CSF leaks. But that’s a virtual space, right, that’s not a real space. So it’s really difficult to do that on purpose. But sometimes it happens, and then you can get what’s called arachnoiditis and it can be really a horrible chronic disease. It’s untreatable. There’s no cure for that. And a doctor could even be sued just by ordering a blood patch like that.
This is another example of arachnoiditis. This was somebody who just after two regular water-soluble myelograms developed this really bad arachnoiditis. The nerve roots look horrible, they’re all clumped together. Interestingly this patient had no symptoms of it though.
And this is somewhat similar to what Dr. Carroll showed. This was a patient who had very orthostatic headaches with a lot of neck pain. So the local anesthesiologist thought oh it must be in the neck, right, even though all the scans were normal. Patient had a regular interlaminar cervical blood patch and as you can see, the needle went into the spinal cord and this patient woke up hemiplegic after the blood patch. So there’s some risk, and it totally makes sense, right.
So on this MRI that you can see there on the left side, right? Outside of the dura, there’s this large collection of CSF that has leaked out, right? So there’s a giant epidural space, right? We say, “Oh, you can drive a truck through it.” And you can imagine, that’s pretty easily accessible. You can put your blood there. You can probably inject a good 40, 50, 60 milliliters of blood. But if you don’t have any extra dural CSF, then it becomes a little bit more risky, and how much more risky in the real world, we don’t really know how that is.
So what about targeted treatment? It’s another option. This is a patient who had, I think I showed that earlier, multiple cysts, and I think she was referred for surgery and didn’t really want to do that because she had so many cysts. And I asked Dr. Maya if he would pick two or three targets and generally go by, not necessarily the largest cyst, but just the most irregular cyst and what’s more easily accessible, and that resolved her CSF leak.
Sometimes I do surgical exploration on patients who have a history of a ventral leak where the leak is gone on imaging. As you can see here, this is somebody where we found this piece of bone that had migrated intradurally, but it had sealed around the piece of bone. And so after we took the piece of bone out, you can see that there’s the hole in the dura. Here around it, you can see that was the partially self-healed portion of the dural defect. And then these are two other patients who had sealing of the defect, but it just wasn’t completely sealed. Or maybe it was, because certainly not everybody who has surgery like that feels better after surgery.
Or you can go look for a CSF-venous fistula, right, because those patients don’t have an epidural CSF collection. But there’s a risk of doing a search. You can get a post-dural puncture headache, because still the neuroradiologists in this room have not figured out how to diagnose this non-invasively. There’s ionizing radiation. It also stigmatizes the patient, right? So sometimes people come to us and we do all this testing and we say it’s normal, and then they go back to their doctors and they say, well, I mean, look, if you went to a specialist center, they must really think you have a CSF leak because they did all this expensive invasive imaging. It cost a lot. It disrupts the patient’s life, as you can imagine. And for the hospital, a lot of resources. It’s reimbursed not nearly as well as just reading a scan. It costs money, and also especially when you start in the CSF leak program and you have all these patients combined where all the imaging is normal, that doesn’t really create a lot of enthusiasm of taking care of these patients.
We’ve been looking: are there some ways to predict the presence of a leak? We’re running short on time, but one of the things we found is that if you don’t have any cysts along the spine, it’s really unlikely that you have a fistula. The other thing that we found out is that if your optic nerve sheath is really diminished, if there’s not enough CSF around the optic nerve, that really increases the likelihood of finding a fistula.
Now, and this is something I learned from Dr. Carroll, you shouldn’t tell a patient, “Well, you don’t have a leak.” You can say, “Well, we didn’t find a leak.” But not finding a leak doesn’t necessarily mean that the patient does not have a leak. And if you fix these fistulas in people with normal brain MRIs, you can see that this optic nerve sheath becomes totally normal after treatment.
So what are some of the other predictors? This is also from Dr. Carroll. I’m kind of disappointed you didn’t say that, because he usually says, “Oh, this patient had a really good story for a leak.” And then the response to blood patching, I think, is important. So this I do this in jest, this is like the batting average for referring headache doctors. And that just means if you send at least three patients, we have this thing in our office, how often do these patients actually have a leak? And the batting average, if you’re really, really good, it’s more than 20%. That’s called the Mendoza line in American baseball. Again, if it’s less than 20%, then you know, and baseball is really weird. If you miss hitting the ball eight out of ten times, you’ll be fired. But if you miss it seven out of ten times, you’ll get a $20 million a year contract. So I find it difficult to believe.
And then what about the response to patching? I think that is important. Dr. Kranz’s group and our group, we do things a little bit different. We find fistulas in about 10% of people with normal brain MRIs, and then he published a study where the yield was zero in his practice. And so Dr. Maya thought, well, you know, that proves that DSM are better than dynamic CT myelograms. Now he didn’t really say that. But we did think about that, but that’s not the case, right? If you have a patient with an abnormal brain MRI, we found it like three out of four times. Dr. Kranz’s group found it about three out of four times. But the main difference between these two studies was that we kind of pre-screened patients and we basically, at the time of this study, would only offer this invasive test if somebody had had a positive response to a blood patch. So I think that’s important as well.
So in conclusion, if we have a patient with some cysts, the optic nerve sheath is diminished, but otherwise the brain MRI is normal, and the headache referring doctor has a good batting average and there’s a good response to blood patch, we actually are able to identify fistula about 50% of the time. But if that’s all the opposite, then it’s essentially zero. Thank you.