Case def proposal outline_08152018.docx
(20.6 KB)
This is the first draft proposal of changes for the new SFR case definition. I have included comments within the document to explain the format and changes made. If you still have questions, please feel free to comment below or email me (wwd7@cdc.gov).
Please make comments/edits by Friday, 8/31, so we can discuss during our next RMSF Working Group call.
Thanks,
Kristen
Please make comments/edits by Friday, 8/31, so we can discuss during our next RMSF Working Group call.
Thanks,
Kristen
I have read over your proposed SFR Case Definition and I have added comments about how I would like the new Case Definition to be. In particular the proposed case definition does nothing to address the issue of patients who have been tested by acute and convalescent serology and have been positive on both tests without a four-fold increase in titer. I have attached the proposed case definition with my comments.
From my understanding of human immune response to pathogenic SFR agents, a patient's titer should rise from undetectable, or low levels (in the event that a person has had recent past exposures to a non-pathogenic SFR agent) at least four-fold within about 2 weeks of illness onset. In the past, I have been told by persons at the CDC Rickettsial Diseases Branch that people exposed to highly pathogenic SFR agents such as RMSF typically develop a strong immune response to such an exposure, i.e., at least a 1:1024 or 1:2048 titer on their convalescent sample. We see a high percentage of cases where person are tested at 1:64, or 1:128 on both their acute and their convalescent samples, a few that have a 1:512 on both the acute and convalescent samples, and very few that have a four fold rise in titer. Over the past 10 years in Virginia, less than 1% of patients (only 17 patients out of ~2,650 SFR cases counted) had a four-fold rise in titer from their acute to convalescent serum sample.
Our tick surveillance efforts across Virginia in 2012 determined that about 55% of lone star ticks carry Rickettsia amblyommii, an SFR agent that has not been proven to be pathogenic (less than 1% carried R. parkeri and none carried R. rickettsii. Tick studies in other states have had similar results. And as lone star ticks are, by far, the most common cause of tick bites in Virginia, many people are exposed to them more than once a year, e.g., are bitten in the springtime and early summer by nymph stage lone star ticks, and then bitten again in the late summer by one or more larval stage lone star ticks. Although American dog ticks occur in Virginia, they generally constitute less than 2% of what is collected on tick drags, and over many years I have heard of very few people being bitten by this species.
I would think that as health departments, our case surveillance would primarily be interested in disease surveillance, so spending a lot of epidemiology time investigating and reporting patients who have not had a four fold rise in titer from their acute to their convalescent sera, seems pointless. No doubt that many of these patients are ill, but testing has shown that it was not an SFR agent that caused their illness, so why count them as SFR cases?
Furthermore, our annual reporting of the many SFR cases we have counted may also mislead the medical community and public into believing that RMSF is rampant in Virginia, and this might cause them to be less likely to consider other much more appropriate diagnoses such as ehrlichiosis. It appears that some doctors do test acute patients for both ehrlichiosis and RMSF, but when this acute sample only comes up RMSF positive, they may tend to focus on RMSF and discount an ehrlichiosis diagnosis. They often do not see the patient again to draw a convalescent serum sample and there seems to be little understanding about the time it takes for patients to develop a detectable immune response to rickettsial agents.
David N. Gaines, Ph.D.
State Public Health Entomologist
Virginia Dept. of Health
Division of Environmental Epidemiology
Richmond, VA
804-864-8112 – Office
804-864-8131 - FAX
804-221-1028 – Cell
Thanks for posting. And thank you, David for including some thoughtful discussion! I'm attaching GA's comments and questions to this. To summarize:
We're looking forward to the next RMSF working group call!
Shawna Stuck
Vectorborne Disease Epidemiologist
Georgia Department of Public Health
I too have been frustrated with the prevalence of patients that only have acute titer positives on their RMSF testing. However, I can understand why a patient, after being treated and feeling better, would want to avoid returning to the doctor to pay another co-pay fee, and get stuck with a needle again to draw a convalescent sample.
I reviewed Virginia's SFR case data over three years (2014-2016) and saw that approximately 37% of the cases we were counting as SFR each year only had a convalescent sample tested. Furthermore, in my analysis, I erred on the side of caution, so this 37% only represented those patients with SFR positive samples that were drawn within the first 5 days of illness onset (i.e., obviously much too early in the course of illness to represent an immune response to whatever illness brought the patient to the doctor in the first place).
However, as we do not have a convalescent sample collected or tested on any of these patients, we cannot immediately conclude that these patients did not have an SFR illness. And as it was probably some compatible illness symptoms that led these patients to go to the doctor and be tested for RMSF in the first place, having only an acute serum sample test to go by does not give us enough information to either reject, or accept the case. Therefore, I proposed that such cases be counted as "Suspect" SFR cases (i.e., we would note them as "Suspect" in our VEDSS entries, but they would not contribute to the total annual case count).
Therefore, unlike those patients that had acute and convalescent sera drawn and tested without a four-fold rise in titer, we cannot just reject them as "Not a Case" and move on to the next investigation. That is because some of these single acute positive titers may have been drawn after a period when they would begin to register an immune response to whatever has infected them, so it would be necessary for the epidemiologist to contact the provider to determine what the onset date was on these cases to see if the single titer was drawn before, or after an immune response would have become detectable. If the single titer was drawn after when an immune response would become detectable then I think that we could not count them as just "Suspect" SFR cases and we would have no choice but to count these cases as a "Probable" SFR cases. The only question at this point would be how many days after illness onset do we begin to count a single titer as evidence of a "Probable infection". As it takes a person from 7 to 10 days to develop an immune response to an SFR exposure, should it be >=7 days, or should we err on the side of caution and make it >= 6 days?
David N. Gaines, Ph.D.
Kristen just added me and I have included my comments over Shawna's here for Arizona. Our RMSF and SFGR epidemiological picture is a little different, though some challenges in surveillance are likely the same as other states. I am looking forward to working with this group on proposed changes to this case definition.
I will add some expanded comments shortly.
Hayley
Hayley D. Belisle-Yaglom, MS, MPH
David made the comment about a "suspect" category and we actually do use that in Arizona. Personally, I don't like it because the criteria are very non-specific. But I have inserted it below for your viewing.
Suspect
• A case with laboratory evidence of past or present infection but no clinical information available (e.g. a laboratory report).
• A case that meets the clinical criteria with a negative acute specimen results and missing convalescent testing.
I always wanted to mention that Arizona receives ALL positive AND negative rickettisal serology results. This was put into place before I started 4 years ago, but helps us get an understanding of potential burden. Keep in mind again that our epidemiology is drastically different than in the rest of the U.S. and we have endemic hotspots for the disease versus widespread transmission...mostly due to our tick vector.
Some additional comments in response to David and Shawna.
In Arizona, we also see a high percentage of cases where people test 1:64 or 1:128 on their acute/ acute and convalescent samples. The antibody rise even more confirmed cases is rarely as dramatic as we immunologically think it should be.
We do not do regular tick surveillance work, however, we do know that ticks are carrying non- or low- pathogenic rickettsia and therefore cause these non-specific antibody responses in patients being tested for RMSF.
Shawna asked:
b. Getting convalescent titers on patients is incredibly challenging for reasons already mentioned. This is always going to be a flaw and we miss A LOT of cases.
I agree that the goal of updating this case definition is to help states conduct surveillance and report cases in a more efficient and epidemiologically sound way. But one consideration I keep in mind is that spotted fever rickettsiosis is not necessary the same across the U.S. So what I think would work best for Arizona is likely not going to work for everyone else.
Hayley
I very much like your idea of improving physician education about the utility of PCR testing or other specific assays and about the proper sample collection methods needed for such testing. It is apparent in Virginia that very few of the many hundreds of SFR cases we count each year are real. Furthermore, tick surveillance and tick testing by various groups in Virginia over the past decade has shown that the only pathogenic agent of spotted fever rickettsiosis we have ever detected among thousands of tested ticks was R. parkeri. None of the tested ticks of any species has ever tested positive for R. rickettsii in VA, and at least 6,000 Dermacentor variabilis have been tested in that time. Among lone star ticks, which are the cause of >95% of tick bites in areas of the state where SFR cases are reported, R. parkeri was seen in <1% of tested ticks, but the agents of ehrlichiosis were quite common [i.e. about 5%]. Furthermore although Dermacentor variabilis accounts for up to about 2% of ticks collected in some areas of Virginia, we rarely hear about bites by this species and I have not been bitten by one since the early 1970s , although the number of lone star tick bites I have had since then have been way too many to count.
Since last spring, we have put on three educational presentations to ER Staff and doctors and told them that they should focus on ehrlichiosis by collecting whole blood from suspected rickettsiosis patients and submitting it for multiplex PCR for ehrlichia/anapalsma. It appears that most patients suffering from a rickettsial illness will see their doctors within the first five days of illness, so with this strategy, doctors will detect ehrlichiosis cases at a time when serology is not useful (i.e., it takes from 7 to 10 days for a patient to develop a detectable immune response after onset of a rickettsial illness, but ehrlichia pathogens can be detected by PCR or microscopy in a patient's whole blood from the first day of illness).
If the doctors forget about RMSF and focus primarily on ehrlichiosis, test their patients by multiplex PCR, and treat their patients for ehrlichiosis with doxycycline, if by chance one of these patients is actually a real case of SFR illness, the doxycycline treatment should solve that problem. With this strategy we hope to correctly identify a lot more ehrlichiosis cases and will probably not miss many real SFR cases as they appear to be rare. We have also educated ER doctors to collect swab samples, or scabs from patients with eschars, and punch biopsies from patients with macular, papular or petechial rashes, or whole blood samples from patients with a petechial rash. All of these SFR samples would be submitted for testing at the CDC. It is too bad that the commercial labs are not offering such PCR testing of swab, biopsy or blood samples for SFR, because most docs appear to be very much in the habit of using commercial labs and some of them even have Lab Corp personnel that draw the blood samples in their offices, which means the samples must be sent to Lab Corp..
Hopefully, our educational strategy will improve our annual case counts for ehrlichiosis, and reduce the number of useless SFR cases being identified, at least in the few jurisdictions where I have, thus far, conducted educational programs for ER staff.
A comment on the eschar, punch biopsies--our experience is that it has been a hit or miss. With Rickettsia parkeri detection in AZ, we have done messaging about eschar biopsies and have had success if public health gets to the table quickly. Sometimes it is a missed opportunity or the eschar is not always very visible depending on area of tick bite. For the bunch biopsies of rash in the case of RMSF, we have had some providers not want to pursue because they view it as "an invasive procedure".
Abelardo
https://doodle.com/poll/aimcdbizzacqk2nq
On another matter, and I may have missed this but I did not see it in the case definition proposal, what is the current goal of SFR surveillance? I know we are discussing a lot of technical details, but what more, specifically, do we want to learn?
Here are the goals from the previous CSTE position statements on RMSF/SFR
07-id-05: The improved case definition will provide the tools to help establish the burden of disease due to this pathogen. This will provide a greater understanding of this disease among physicians, nurses, and public health professionals and allow for appropriate prevention messages to the public.
03-id-08: Identify cases and track incidence of RMSF in the US. Further define the clinical spectrum of disease.
09-id-16: To provide information on the temporal, geographic, and demographic occurrence of Spotted Fever Rickettsiosis (including Rocky Mountain Spotted Fever) to facilitate its prevention and control.
The most recent goal was to facilitate prevention and control, so I suppose we have met the previous goal of defining the clinical spectrum of disease? If identifying cases is too difficult using traditional laboratory methods maybe we should look at alternate data streams to obtain information on the temporal, geographic and demographic occurrence. Data from commercial labs, insurance claims data, more robust tick surveillance to identify vector/pathogen occurrence. I think some states have already done this and published on it...Tennessee?
My two cents worth.
Carl
I closed the Doodle poll and it looks like the date and time that worked for the majority of folks is Friday, October 26th at 9:30 am - 10:30 am (central time).
This is the WebEx call-in information for the meeting:
Audio connection: 415-655-0003
Access Code: 647 661 040
Meeting password: 3JUFqdGB
Meeting link: https://tngov.webex.com/tngov/j.php?MTID=m07df2c5a2163da38624eac820f548669
I have also tried to attach an Outlook calendar invite but I'm not sure if you're able to see it. :)
Looking forward to next week!
Kind regards,
Kelly Orejuela
I do not see that the goals of any of the past CSTE position statements are being served by these Case Definitions. As several studies have shown that a significantly large proportion of the healthy U.S. population would test positive on serological testing for RMSF. we need to change the Case Definition so that it no longer picks up all of this misleading chatter. Furthermore the people who most often get tested for RMSF are probably also the people who most often have tick encounters, so it is likely that a much higher proportion of these people would test false positive for RMSF. Counting cases based on IgG positive IFA results on an acute serum sample makes no sense. At best these should be called "Suspect Cases". Counting cases on the basis of IgG positive acute and convalescent samples in which there was not a four-fold increase in titer also makes no sense. I would love someone to explain the rational for either of these actions.or why the current Case Definition is serving a useful purpose.
What would help us most is a campaign to educate the healthcare community about SFR and ehrlichia disgnostics (e.g., the multiplex PCR), as well as their entomological prevalence, and encourage them to "think ehrlichiosis" when they see a potential SFR case. The entomological evidence suggests that the agent of RMSF and other pathogenic SFR agents such as R. parkeri are either rare or uncommon in the tick species that most commonly bite people, whereas the agents of ehrlichiosis are relatively common. If more doctors considered ehrlichiosis in their differential, and used the right diagnostic assays, they would probably be right a lot more times than they would be wrong. Right now, with the rampant diagnosis of RMSF based on unreliable serological assays, they are wrong much more often than they are right, and the laboratory results they are getting only reinforce their misconception..
BTW,.someone has got to tell Lab Corp to add Ehrlichia ewingii to their multiplex PCR assay. About 4% of the Virginians that test positive for ehrlichia on the Mayo Clinic multiplex assay test positive for E. ewingii. That means that by not offering testing for E. ewingii in their multiplex assay, Lab Corp is causing a misdiagnosis (a negative test result) in a significant number of ehrlichiosis patients that they test, particularly if the ordering physician did not also do acute and convalescent serology. Furthermore, this year Virginia counted its first known fatality due to an E. ewingii infection (this in a healthy middle aged person).
David N. Gaines, Ph.D.
State Public Health Entomologist
Virginia Dept. of Health
Division of Environmental Epidemiology
Richmond, VA
804-864-8112 – Office
804-864-8131 - FAX
804-221-1028 – Cell