Posted by Hayleigh McCall · Oct 04, 2018 at 4:57 PM

Consider for November Case Definition Call - Thoughts from Brian Bachaus (MD)

As part of their CSTE case definitions, some conditions (e.g., influenza-associated pediatric death, Hepatitis A) include language that allows for physicians to use clinical judgment in deciding whether or not a "case" (by clinical/laboratory criteria) really is a case and should be counted.

We in Maryland have had concerns regarding the reliability of the Leigonella UATs for some time, and I believe New York City and New York state have also raised concerns. There are reports in the literature of false-positive tests attributed to CKD, rheumatoid factor, etc. -- mostly conditions that lead to excess protein in urine.

We've encountered situations where someone with questionable pneumonia and kidney dysfunction had a positive UAT. The same urine was retested at the state lab, where it was negative, and, additionally, acute and convalescent sera showed no evidence of legionella infection.

I'm curious if other states would be amenable to including an "out" for the legionella case definition, similar to those other conditions listed above.

Comments & Events

Richard Danila

I would not necessarily require clinical criteria to meet the case definition (allowing for an “out” if criteria were not met). The influenza-associated pediatric death requires lab confirmation. Hepatitis A allows for case classification by clinical criteria alone when there is known contact with a hepatitis A case. I can’t think of any nationally notifiable disease that allows for clinical criteria alone in the absence of known contact with another case (for diseases spread person-to-person).

With electronic lab reporting, health departments might indeed receive positive UA test results that in fact may be false positive tests. However, the UA test was undoubtedly ordered by the clinician as part of a clinical work up for pneumonia or severe respiratory disease. It would be hard to think of a situation where a test was ordered for a patient, and tested positive, in the absence of any clinical criteria compatible with Legionnaires’ disease. So in fact was a false positive. You (in Maryland) apparently have seen such cases. On the other hand, surveillance data are dirty, and case counts may include cases that were included on the basis of a positive UA test alone in the absence of some set of clinical criteria. However, to require a thorough review of the clinical symptoms to rule in or rule out cases imposes a resource burden.

Richard N. Danila, Ph.D., M.P.H.
Epidemiology Program Manager, and Deputy State Epidemiologist
Minnesota Department of Health
625 Robert St. N., P.O. Box 64975
St. Paul, MN 55164-0975

(Office) 651-201-5414
(FAX) 651-201-5743
(Work cell) 651-592-0450
(Personal cell) 651-210-3029
(Home) 651-649-0298
richard.danila@state.mn.us<mailto:richard.danila@state.mn.us>

Oct 04, 2018 at 6:20 PM Notified 64 people

Brian Bachaus

Thanks for the feedback, Dr. Danilla.

In my email to Vivian, I was just intending to share with her some of the experiences we’ve had in Maryland and to see if it was worth having the topic as an agenda item for a future call. It’s fine with me that the email was forwarded on to the group, but if I had known that it would be the jumping-off point for the discussion, I would’ve constructed it differently. So, I’ll restate our concerns, and mention how we’ve dealt with similar concerns in other situations.

Every year, the Maryland Department of Health investigates dozens of reports of positive Legionella UATs in individuals who had a healthcare stay during part or all of their exposure window. In making the decision whether or not water restrictions, water testing, etc. are warranted for the facility, we try to determine, among other things, how likely it is that the individual truly has Legionnaires’ disease (based on clinical and laboratory evidence).

For these individuals with substantial healthcare exposures – not ALL positive UATs – the Maryland public health lab often retests the urine specimen that was initially reported as positive by UAT at another lab, and we request acute and convalescent sera for Legionella serology, and sputum for Legionella PCR and culture – though it’s rare that we actually get a lower respiratory specimen. A fairly common scenario for us is that our lab cannot reproduce the positive urinary antigen result, and that the paired sera do not show evidence of Legionella infection (our Legionella serology testing looks for Lp1-6). Obviously, we’ve already had to make a decision regarding water testing and restrictions by the time the convalescent serum has been collected, but we oftentimes still do request it.

Additionally, as mentioned, we evaluate the clinical evidence for individuals with significant healthcare exposure prior to their illness onset. Indeed, it is rare that a UAT is ordered for a patient when there isn’t any clinical criteria compatible with Legionnaires’ disease, given the broad spectrum of illness and the number ways the disease may present itself (as we’ve discussed in the clinical criteria subgroup) – but it does happen. And while some may have clear evidence of respiratory illness and/or pneumonia, it’s not uncommon that we see that an individual’s “current” illness isn’t really a departure from their already-poor respiratory baseline (due to underlying illness) – or that while the radiologist cannot rule out pneumonia or new infiltrate, they also can’t rule out that they’re seeing COPD scarring or CHF w/volume overload or some other, non-infectious process.

Finally, there are reports in the literature of cross-reactions leading to falsely-positive UATs – either from a coinfection, or underlying illness that leads to excess protein in the urine. Unfortunately, the elderly – who, as everyone in this group is already aware, are at greater risk for Legionnaires’ disease – are also more likely to have some of these conditions that are associated with false-positive UATs.

I understand that surveillance data are inherently dirty and that no case definition is perfect – and I’m certainly not suggesting that all health departments use precious resources unnecessarily scrutinizing every single detail of every single case that gets reported. But, in the situations described above, where we’re already collecting the data and seeing an unclear clinical picture paired w/ discordant lab results (especially if the individual has an underlying condition known to be associated w/false-positive UATs), we’d like a case classification that takes into account that additional information. A couple of examples of how this could be addressed include the “out” in the case definitions for influenza-associated pediatric death and the new Hepatitis A case definition, as I mentioned in my email to Vivian.

I know some other states have voiced concerns regarding potentially false-positive UATs, though I don’t know if there’s any interest in modifying the case definition accordingly – which is why I wanted to have the conversation.

Thanks – looking forward to further discussion.

Brian

Oct 05, 2018 at 3:38 PM Notified 63 people