NIH CLINICAL CENTER GRAND ROUNDS
Episode 2010-13
Time: 53:08
Recorded April 7, 2010
Pitfalls and Promise for Individualized Pain Medicine
Hyungsuk Kim, DDS, PhD, Staff Scientist, Pain Research Unit, Symptoms
Management Branch, Intramural Research Division, NINR
Symptom Distress Mechanisms in Digestive Disorders
Wendy A. Henderson, CRNP, MSN, PhD
Assistant Clinical Investigator
Biobehavioral Unit, Symptoms Management Branch, Intramural Research
Division, NINR
ANNOUNCER: Discussing Outstanding Science of the Past, Present and Future - this is NIH Clinical Center Grand Rounds.
(Music establishes, goes under VO)
ANNOUNCER: Greetings and welcome to NIH Clinical Center Grand Rounds, recorded March 31st, 2010. Today, we'll hear from two researchers from the Symptoms Management Branch at the National Institute of Nursing Research will present.
Dr. Hyungsuk Kim, staff scientist, Pain Research Unit, will address "Pitfalls and Promise for Individualized Pain Medicine." Then, Dr. Wendy A. Henderson, assistant clinical investigator, Biobehavioral Unit, will present "Symptom Distress Mechanisms in Digestive Disorders."
We take you to the Lipsett Ampitheater at the NIH Clinical Center in Bethesda, Maryland for today's presentation.
KIM: Thank you very much for the introduction. The title of my talk today is pitfalls and promises for individualized pain medicine. Before I start, I'd like to announce I do not have any conflicts of interest for my presentation.
These are the main purposes of my presentation. First of all, I'd like to explain the importance of individualized pain medicine. Pain is very important to us. Pain can provide very useful information to evaluate the conditions or diseases of the patient and it is considered the fifth vital sign. Pain is one of the main reasons for visiting doctors and also one of the main reasons for patients to avoid seeing doctors.
As these numbers show we are spending a lot of our resources because of the pain. Apparently, we are not taking care of pain very well because of the most widely used analgesic drugs, because of that, 110,000 hospitalizations are caused and because of the adverse reactions, nearly 20,000 deaths happen annually in the United States. Because of acetaminophen the second most widely used analgesic drug, we have a lot of liver failures in the western world. Not only those kinds of things happen but also many different diseases such as fibromyalgia or CRPS, those kind of diseases do not respond well to the treatment strategies including acetominophen. Additionally, people respond very differently to the analgesic drugs, for example, after the surgery, the total consumption ranges from two milligrams to 345 milligrams of patient populations with a mean dosage of 65.2 milligrams. So if we give the mean value of analgesic drugs of the opioid, 65.2 milligrams, to the patient who requires additional amount of medication, that kind of patient does not get any kind of analgesic affect.
This kind of variability in pain and analgesic responses in humans are very well-known phenomenon. For example, last week, "USA Today" says, we are born with different vulnerabilities to pain and that all these painful experiences shape us in ways that can profoundly affect our responses. And one of us needs a single pain pill after surgery, while another has chronic pain for years after leaving hospitals. Of course, we cannot believe everything in " USA Today," so this is the scientific evidence.
This graph shows the response to the heat painful stimuli which is 49 degrees celsius. Every 735 normal healthy subjects was exposed to the 49 degrees celsius, but their responses ranges from zero, which is no pain, to the 100, which is worst pain imaginable. As you can see from the graph all the people are almost evenly distributed from 0-10. So this graph clearly shows that people are very, very different. They are differently responding to the same painful stimuli. So you agree that individualized pain medicine is very important.
I'd like to read you the pitfalls in genetics for pain and analgesia research. Pain is defined as an unpleasant sensory and emotional experience which is primarily associated with tissue damage or described in terms of tissue damage, or both. Pain is a complex trait. It is regulated by genetics and environmental influences and thousands of genes are involved in pain perception and transmission, pain recognition. It is a very, very complex trait. Probably one of the most complex traits among the common diseases or common conditions. Therefore, the effect of each genes is thought of, it's very difficult to detect that kind of single gene affect.
When we talk about the genetics, whenever we do the genetic association study. The first test that should be tried to come from whether that specific trait is heritable or not. If the trait is not heritable, there is no point in trying to find the responsible genetic loci for that trait. So, heritable estimate is the best way to confirm if it's heritable or not and calculate it like this. Because monozygotic twins are considered as 100% identical in the genome, dizygotic twins are considered average sharing 50% of their genome. So calculation of correlation and correlation of the monozygotic twins minus correlation of the dizygotic twins times two is calculated as heritability estimate and expressed as a percentage. Heritability estimates may be one of the most misunderstood concepts in genetics. It represents the portion of the trait variation caused by genes, not the portion of the trait itself caused by the genes.
Probably an example could show that. For example, if the father's height is 180 centimeters, 6 foot, this is my height. And mother's height is 152 centimeters, average height of a Korean man -- I came from Korea. Average height is 177 centimeters and average height of women is 160 centimeters. And the heritability estimate by percentage in 65% in men and 70% in women. When I calculated the expected height of my son, the calculation is like this, from my height minus average of man and my wife's height minus average of women divided by 2 times heritability estimates. Then the average or expected difference of the height of my son is 3.9 centimeters. So the genetically expected height of my son is this value. And the same calculation of height for my daughter but different heritability. The expected height of my daughter is this number.
Heritability estimates have some limitations. This value applies to the specific population. So 65% of heritability estimates is only applied to the man and Korean population. If number can be different if it is calculated from Chinese population or U.S. population.
The other limitation of a heritability estimate is the basic assumption of this calculation is no interactions of the genes with environment, which is clearly not true. So it has a lot of limitations but the true meaning of heritability estimate calculation is if it is high, it is more likely to succeed to detect the action of individual genes. The height is, as everybody agrees, height is heritable. And height is reported as 50-80%. And that's a lot of pain. Pain ranges from 40-60% in this heritability estimate. Acute pain is 25-35. Back and neck pain: menstrual pain and some pain like temporal mandibular joint pain is zero.
From this, heritability estimates of pain, we can get lessons. First, the pain is mild to moderately heritable. And as we can see from the different numbers from the population, it looks like different types of pain have different genetic backgrounds. Otherwise they seem to be similar. That kind of fact that different pain came from the different genetic background can also be confirmed to be the animal study which was published in 2003 by these two investigators and this is from an experiment using different techniques and this shows this is stomach pain and this is mechanical pain. All those pains have different clusters based on genetic background. Therefore, with the genetic association studies, if the phenotype is combined pain like additional heat pain and cold pain, you have to be very, very careful to get that result.
And another pitfall of the genetic association studies in pain and analgesia is population stratification. This is the actual data, subjects are asked to put their hands in ice water and when it is unbearably painful, they can withdrawal their hand from the ice water. This is the result we have when I click this sample into the different ethnic populations, Americans and African Americans, both showed no association. The reason why is because African American population are reported that they tend to report their pain more frequently. The African Americans take the majority part of this population. So even though in African Americans, this t/t allele showed higher. This type of population stratification can affect the results of genetic association studies. Therefore, whenever you create the result, you have to make sure that the population is not mixed.
And the other problem as to the pitfalls of the genetic research in pain is the measurement. Visual analogue scale is the most widely used measurement of the pain. And it is anchored one side no pain, and this is anchored worst pain imaginable and this is 100 millimeter ruler. But the problem is, this anchor, worst pain imaginable is different from people to people. For example, the worst pain imaginable for a woman who gave birth probably would be very, very different from the worst pain imaginable for the teenaged girl, or the worst pain for the experience for a soldier with a gunshot wound would be different from a normal civilian who has not have any broken arms or something like that. So this is the problem. It's like comparing the small and giant ants. It looks like much bigger and more scary. But everybody will understand that smaller elephant is much bigger than the giant ant.
Objectivity is the issue. And I believe the next speaker will discuss details about this issue. The other problem is the pitfall of nongenetic factors. Pain is influenced by a lot of nongenetic factors. As everybody knows, men are not supposed to complain about the pain. This is the same test with the male data. We can see here that males put their hands into the ice water as long as possible and almost 50% of subjects could keep their hand in the ice water up to three minutes. Meanwhile, females, they are not like that. A lot of females withdraw their hands pretty quickly and only 25% of women keep their hand in the ice water. So of course it's not all because of the gender expectations. But it's clear the expectations have something to do with the pain response. This maybe presents the women are smarter than the men.
And the other nongenetic factors that exist, for example, culture. In some cultures, we are taught that we are not supposed to express ourselves very easily. It's different from western culture. And secondary gain is another. For example, an accident, then the patient in the middle of the lawsuit or that kind of situation, they usually do not get well unless that kind of situation is solved. And publication bias is another issue we have to think about. The authors do not want to publish the negative results. Because of that, most of the results are not published. So the genetic positive association tends to be exaggerated in the results. And also epigenetic influence on the pain too. However, it's not well characterized yet.
So far, we have reviewed the pitfalls of the genetic research in pain and I'd now like to share the status of the individualized pain medicine.
First example maybe is the liver enzyme. And one is involving metabolism of multiple drugs and metabolism from codeine to morphine and 7-10% of Caucasians are poor metabolizers. It doesn't function very well. So they do not get pain relief from codeine. This is the morphine to codeine ratio and poor metabolizers have very low value compared to the ultrametabolizers. This type of thing can be also confirmed with the clinical data that poor metabolizers requires more tramadol which is codeine compared to the extensive metabolizers. And because of this finding, cytochrome chips have been developed and available since 2005. But I heard they very rarely use this in clinics.
The other example is an allele that showed less amount of morphine consumption compared to other allele groups and even shows some joint affects. So that plus AA Allele group shows the consumption working for the cancer pain treatment.
Other examples for the end-stage drug use is this one. One SNP located in the promotor region of the COX 2 gene shows different gene expressions. GG homozygote group shows high expression of COX 2 genes compared to these heterozygotes and cc homozygote group. This result also is consistent with the clinical data that GG group showed higher expression of COX induced COX 2 genes. They responded well to the COX 2 selective inhibitors while the other groups showed better response to the nonselective COX inhibitors.
So far we have reviewed the genome association studies. But we preferred to expand it to the genome-wide association studies because we don't know the exact number of genes in human genome yet. And there are a lot of striking variances and we have more than thirteen million SNPS identified. And for the known genes, 31% of the function is known when the genome predicts in 2001 and year 2008 it goes 34%. Therefore genome-wide association studies is the next direction we have to go and this is what we got from the genome-wide association study in all the models and the phenotype of this graph is analgesic onset.
And this is developing a sequence. And when we extend the region, we find that in the high equilibrium with this gene and when genotype this gene, we could fine four SNPS are completely linked to this equilibrium with that SNP from the GWAS result. And this minor zygote people showed more than 3 times longer analgesic onset time than compared to heterozygous and major homozygous group.
From now, we have to do the replication studies to confirm the previous findings. And as I said before, the whole genome scale with appropriate studies should be performed. And function of genomic studies. The association studies can only provide the statistical significance but cannot provide the mechanism. The functional studies are needed and we also need new analysis too. Not only for the phenotype measurement, but also the new technologies including mass generation sequencing machine, we will have a lot of variables. Currently from one million SNPS data to one billion of the entire human sequencing data for each individual. So we need new analysis too.
We also need to educate people. As you can see from the example the USA Today, too much can be found on the genetics publicly probably because of the CSI TV show. But meanwhile, too less attention can be found in positions as I showed. Amplitude can be used to determine the analgesic drug selections for the patients but physicians never use that. Usually physicians prescribe analgesics when patients complain of pain and they adjust a dosage based on the response of the pain, not using the genetic association study.
So we are still very far from the realization of individualized pain medicine. And it may not be feasible in our generation but if we achieve our effort in this area, our next generation or our next generation, can get the benefit from the genome-based prescription system as we get the benefits from the GPS association when finding directions.
And I'd like to thank all those people and especially for the patients and their families for preparing this presentation. Thank you very much.
[APPLAUSE]
We have a few minutes of questions from the audience.
AUDIENCE STATEMENT: Thank you very much for the very interesting presentation. You pointed out a problem with the poor metabolizers. I just wanted to comment and of course you're aware of this, ultrarapid metabolizers can generate a lot of morphine from codeine and there are reports in overdoses of morphine in lactating babies whose mother have taken codeine and happen to be ultrarapid metabolizers. There are population whereas there is a high prevalence of ultrarapid metabolizers among Ethiopians. So I just wanted to add that commentary.
KIM: And also some drugs should stay there. If it metabolized too fast, the patient is an ultrametabolizer ,then the effectiveness of the drug will go down rapidly.
HENDERSON: So the title of my talk is symptom distress mechanisms and digestive disorders. I wanted to take a moment to thank you for being here today and thanking the clinical center for this wonderful opportunity for our institute. I have no disclosures that are relevant.
However I'd like to note that the gastrointestinal pain pointer here forward called the GI pain pointer is a research assessment tool and not approved as a medical device. The objectives of my talk are to describe current and experimental outcome assessments Of digestive disease symptoms and to further describe the utility of a recently developed and validated instrument, called the GI pain pointer, for realtime subjective assessments of digestive symptoms.
Key facts: 20% of the United States population reports digestive symptoms of one sort or another. Abdominal pain is the most common GI diagnosis. It is one of the top 10 reasons for outpatient visits and affects predominantly women. Digestive disorders symptoms, for those who are not clinicians, are many and vary. Here are a few. Abdominal pain, constipation, diarrhea, nausea, vomiting and so forth. Symptoms can overlap in that if you have a patient presenting with abdominal pain, that same patient may have fatigue, joint pain, sleep disturbance, or other distressive type symptoms.
More key facts include that abdominal pain symptoms can be mild and intermittent to severe and continuous. Patients present in many different ways. It can disrupt their daily life. There are current treatment options including dietary modification, pharmacologic interventions and biobehavioral interventions.
However, as Dr. Kim indicated, pain treatment is still lacking in some respects. Lucky for us, the health care reform act of 2010 has pain provisions for increasing pain research to encourage the research. The pain consortium in consultation with the division of program coordination and planning and strategic initiatives will make recommendations based on the reform act for pain research initiatives that will be paid for by the common fund. The secretary through the institute of medicine will hold a conference and establish an interagency pain research coordinating committee. That's the new news. Now from the news that we are aware of from the past 10 years forward, our joint commission on accreditation of health care organizations. The standards mandate specific standards on pain assessments. If the patient is ready to appropriate assessment to the nature and intensity of their pain, and regular assessments that are documented in the medical records including education to patients and family members, which are reported at regular intervals and note patient related outcomes or follow-up care.
As previously indicated, pain is typically a subjective phenomenon in that it is how the patient reports it. So if a patient tells you they're in pain, they're in pain. This is how they state it. Therefore it is subjective. Typically from a clinical perspective, it's rated on the 0-10 scale, 0 being no pain and 10 being the worst pain ever. The location of abdominal pain specifically as a digestive disorder symptom is generally recorded at the discretion of the health care provider. Of note, most studies show that clinicians underestimate the pain intensity in their patients.
So now let's think about pain as more of an objective phenomenon. This is a way that we as a clinician or a family member could observe a patient is in pain. Some of the physiologic correlates of this include heart rate, blood pressure, skin temperature, or heart rate variability.
I'd like to focus the rest of this talk on abdominal pain of unknown origins. As Dr. Kim spoke of pain, generally being regarded to tissue damage, this is pain that is not associated with an organic cause. It doesn't have associated tissue damage. These symptoms of abdominal pain typically develop after acute illness. The etiology is unknown, however there are a few theories. There is a possibility genetic link running in families. Multiple subtypes and here are a few listed. Three main theories involved in the pathophysiology of pain of unknown origin will be discussed in the rest of this presentation, the first of which is being visceral hypersensitivity. In this, patients express pain differently than others. So as Dr. Kim previously spoke, there may be a genetic component to this. What we were looking at was more of a mechanistic standpoint from a GI perspective. In these early studies, we used technology for those who aren't familiar with that, we would present a polyethylene balloon seen here at the right, which would be deflated to mimic a meal. So these are patients that typically complain of abdominal pain after eating. They will have pain, early satiaty, other feelings of reflux and so forth. We use the balloon as it has no other nutritional affects to mimic the meal. Polyethylene bag introduced in the mouth down the esophagus and into the stomach. And in this particular protocol, would be infated at varying intervals and the patient would be asked to tell us what they feel and how much pain they feel. At that time, what was available technology wise was a response paddle which would have six oneses on it. 0 being they felt nothing and up to 5 being severe pain with the distension of the balloon. This data would be recorded at 11 intervals in a computer program.
What we found in this group of children of note there were not controls, per se that were asymptomatic as it wasn't approved by the IRB but we had predominantly females. About a little more than half had upper abdominal pain and the others had organic disease. We found symptom presentation was primarily abdominal pain noted in blue here are the nonulcerative dispepsia. That means no sign of organic cause. And in the red are those with organic disease although we didn't know it at the time. The person doing the research study didn't know that. Here are other symptoms, nausea, vomiting, et cetera. What we found clinically, there was a clear difference -- let's look here at the 2 groups. So the red being the organic disease group where there was evidence for why their pain would be present and the blue group was the nonulcerrative group. The nonulcerrative pepcid group sensed the pressure inflated at much lower pressures and complained ever pain sooner than those with organic disease. This was not statistically different when calculating it based on this 5.0-5 scale.
Next I'd like to talk about the work that we have done regarding gut inflammation as a potential pathophysiology of abdominal pain of unknown origin. In this schematic done in our lab, we described some of the recent literature in that they found mass cell tryptase excretion could contribute to abdominal pain when there was no histologic or tissue injury. Park and colleagues, found that mass cells, il6 and 5h2 called serotonin, could contribute to increased colonic permeability and lead to increased abdominal pain.
So here at the NIH, we did a retrospective analysis looking at pediatric biopsies of patients who all had painful complaints of abdominal pain. In a sample of approximately 48 patients, approximately 60% and 40% were noninflammatory verses inflammatory cause. Again, slightly more females than males. Age range 5 to 17 years of age. We looked at their biopsies and stained them with the help of the national cancer institute and we stained them blue and looked at the mass cells in these colonic biopsies of those with inflamed phenotypes and those with noninflammatory phenotypes. We found that correlation between mass cells and 5ht with self reported pain. Interestingly, we found in those with the noninflamed biopsies, an increase in mass cells compared to the inflamed biopsies. As one would expect, we found elevated il6 in the inflamed biopsies given the clearer gross inflammation present.
Moving on to brain-gut interactions. This is the model we are now testing in our prospective natural history study here at the clinical center at NIH. This study includes assessment of sympathetic nervous system as well as immune and inflammatory function and hypopituitary adrenal access. We knew from our prior work, there is an unmet need from a clinical research perspective in terms of proper characterization of these pain phenotypes there. There is also a need for a realtime better assessment of both subjective or the patient reported or how they state their pain, and objective pain assessments.
As Dr. Kim showed before, in a clinical research environment, we used something more of a visual analogue scale or the patient will mark with a line what their pain is and then it's measured. This gives a continuous score as opposed to the 0-10 which is used in the clinical environment, which forces the patient to pick a single number. So this gives you a little bit more of a better description of their pain from a subjective standpoint. It's typically pen and paper, some have been in electronic media. There is no objective component.
So with the help of our colleagues, we developed here at NIH in the clinical center, the Gastrointestinal pain pointer. Also called the GI pain pointer. What this device does it is assesses abdominal pain in children and adults noting location as well as the intensity, quality and physiological parameters in realtime. It combines the subjective scores with objective measures. Now I'll show you what the GI pain pointer does. Through a graphical interface, the clinician or the research clinician enters in the participant's id, that means it's a male or female participant. And they pick the domain for where they will save the data file. And they pick whether a reading prompt is needed for that particular participant. Then the tool is handed to the participant for them to select their gender and for them to select their body type. So here we will select a male gender. We also have normal weight and overweight physiques. So we'll pick the overweight physique. This body is able to be rotated if there is pain on the front and also on the back. And then the participant clicks on the pencil and then points to where it hurts. So as they normally would for us, they are able to do so on their own. They circle it and can move the pain around. They can make it a little bit larger or smaller if it's not quite where their pain is located. Then we asked them to dial up their pain level in terms of how severe it is. Of note, there is no number here. It's blinded. So therefore they can't say I picked 57 last time, I have a 57 amount of pain this time so it's good for repeated measures. Then here you'll see the pain descriptors and they pick whether it's a sensory neurostabbing type of pain or whether it's some other type of pain that is more of an exhausting or bloating or general dull pain. This data is captured. Meanwhile at the same time, the clinician or an electronic capture for heart rate and blood pressure captures that data. It can be entered in by the clinician or it could be synced datawise from that electronic device. So we capture blood pressure and heart rate. This information, all of it including the subjective and objective components are exported to a data file. And then the device will prompt you for the time to the next reading. So it will have an alert system as to when the next time to assess their pain is.
So, we presented this to the patients so that they could use this tool. What we needed to do was validate this tool compared to the research standard here at NIH. So we validated the GI pain pointer in combination with the McGill pain questionnaire which is a self-reported tool. We need to validate the accuracy of sensitivity. We used statistical methods here and I'll move quickly because we are limited on time. And then the methods including using both instruments and 43 patients who are then given an intestinal permeability test solution which could induce GI discomfort and then the patients were monitored for 5 hours following ingestion of solution. The McGill and GI pain pointer were administered simultaneously at 11 time points based on gastric emptying. In 30 minutes prior to in adjusting solution and at the time of injection and 35 minutes and so forth.
Here are sample demographics. We found that the pain scales were significantly correlated at a p of less than .007. One particular patient shows pain reports on the McGill and on the GI pointer. You can see they are highly correlated at .9903. We also found within subject correlation across time to be highly correlated at.7521. And again this is one patient's data showing the blue is the GI pain pointer and the red is the McGill. You can see how they closely follow each other over time. The results are the GI pain pointer discriminated between the persons with and without abdominal pain appropriately and between sample and coefficients were acceptable for the validity testing.
The conclusions, not only was the device found to be valid, it also allowed clinicians to have a more integrated resource for pain assessments that include location, intensity and the physiologic parameters in realtime.
The implications for clinical practice are it could be amended for other uses and other electronic medias such as PDA and it also could be downloaded for data capture in electronic medical records. The protocol could be amended for different time points depending as well as potentially utility and nonverbal patients such as a patient at MICU or someone who is unable to speak to a stroke. It could be translated into other languages as well.
It's clear there are further needs from a clinical and translational perspective. There is a need as you saw from our prior studies to better characterize patients pain and pain reports and to be able to put a subjective component to those pain reports as well. And also to see the change in those over time for patient related outcomes.
We further need to look into a genetic understanding of the patterns and pathways of abdominal pain specifically of unknown origin and identify those etiologies and further to target specific treatments.
I'd like to thank our collaborators, those in my lab for helping with this presentation and this data collection. As well as other NIH collaborators who helped with developing this tool. We also had some extramural collaborators as well as the engineers and psychological software, our statistical support and also for the patient recruitment. Many thanks to those individuals who took the time to participate in our study as well as the medical and nursing staff here at NIH for their support.
Here is my e-mail if you have other questions and also the web address for the brain-gut natural history protocol. And I welcome questions at this time.
[APPLAUSE]
QUESTION: It's nice to see this all at one point. I'm wondering have you thought about or considered broadening this into other areas? This is very specific for GI pain which of course is difficult to assess. But I'm wondering if you thought about that feasibility of doing that?
HENDERSON: We have. We actually had some interest from other folks to ask, could we amend this for a different population, for example neuropathic pain, could you shade on the body where it hurts and describe it in that perspective? Other individuals have asked, could it be used in ICU settings or even in theatre for a military perspective for folks who are unable to speak to you at the time to be able to capture that. The other interest we had are different medication studies to be able to document the change in their pain and how well that particular medication may improve their symptoms.
STATEMENT: Sounds like you have a lot of work cut out for you.
HENDERSON: Thank you very much for your time.
[APPLAUSE]
ANNOUNCER: You've been listening to NIH Clinical Center Grand Rounds recorded April 7, 2010. On today's presentation, we heard from two researchers from the Symptoms Management Branch at the National Institute of Nursing Research. Dr. Hyungsuk Kim, staff scientist, Pain Research Unit, addressed "Pitfalls and Promise for Individualized Pain Medicine." Then, Dr. Wendy A. Henderson, assistant clinical investigator, Biobehavioral Unit,presented "Symptom Distress Mechanisms in Digestive Disorders." You can see a closed-captioned videocast of this lecture by logging onto http://videocast.nih.gov -- click the "Past Events" link -- or by clicking the "View Videocast" link on the podcast homepage at www.cc.nih.gov/podcast. The NIH CLINICAL CENTER GRAND ROUNDS podcast is a presentation of the NIH Clinical Center, Office of Communications, Patient Recruitment and Public Liaison. For more information about clinical research going on every day at the NIH Clinical Center, log on to http://clinicalcenter.nih.gov. From America's Clinical Research Hospital, this has been NIH CLINICAL CENTER GRAND ROUNDS. In Bethesda, Maryland, I'm Bill Schmalfeldt at the National Institutes of Health, an agency of the United States Department of Health and Human Services.
