Urinalysis Voodoo

I must have missed the class on urinalysis interpretation in medical school.  When I hit the ward I didn’t have a clue.  Those first few weeks I nodded along knowingly when the residents talked confidently about how the patient did/did not have a UTI, initially being too embarrassed about not knowing something so simple to ask what, exactly, on the urinalysis led them to their conclusion.  A few rotations in I noted what the obstetricians considered a UTI the medicine men sometimes (but not always) called a contaminated specimen.  In pediatrics I became even more confused as interpretations seemed to shift depending on whether the urine was obtained via bag, catheter or clean-catch as well as the mood, time of day and temperament of the child/pediatrician.  After finally working up the courage to ask some residents to explain themselves, I found that these differences made for a myriad of inconsistencies and initially concluded that it was all just “a bunch of voodoo!”

In emergency medicine we often use a urinalysis +/- microscopy to determine whether we should treat and/or culture the patients that we suspect have a urinary infection for any number of reasons.  As a resident, I amalgamated the black arts that I observed into my own brand of Urinalysis Voodoo (TM) that went something like this:

-if nitrites + –> UTI
-if leukocyte esterase (LE) super high –> UTI
-if pregnant and urine not pristine –> UTI
-if not pregnant and LE was only a bit up but it walked and talked like a UTI –> UTI
-if nurse tells me that the urine smelled really bad –> lean towards UTI

When the lab added microscopy the voodoo continued:
-if bacteria seen –> UTI
-if >5 WBC seen –> UTI

I became quite comfortable with my Urinalysis Voodoo (TM) and, realizing that investigating it further would be unlikely to get me an interview with Scott Weingart, I abandoned attempts to learn about it further and instead went to work on a device that would allow me to intubate myself with only brutaine, a laryngoscope and a mirror.

Until now.

A quick search of the literature led me to a crazy amount of pee research (henceforth known as peesearch, hehe).  Fortunately, before I delved too deeply into it (errr…), I found a review article:

Simerville, J.A., Maxted, W.C. & Pahira, J.J. (2005). Urinalysis: A Comprehensive Review. American Family Physician, 71(6), 1153-1162.

Pearls from the review article:
-cleansing of the women’s genitalia prior to providing a sample does not seem to affect contamination rates
-the proper descriptive word for the smell of urine is “urinoid” while UTI urine is “pungent”
-LE (produced by neutrophils in the urine) can read falsely negative with oxidizing antibiotics such as cephalexin, nitrofurantoin, tetracycline and gentamicin as well as vitamin C
-Nitrites (converted from nitrate by bacteria in the urine) can read falsely positive if the dipstick was exposed to air (ie 1 week of exposure to air results in 33% false positive)
-Normal WBC per HPF (high power field) for men is <2 and women is <5
-As suspected, most of the studies examined found Nitrites and bacteria to be highly specific but not sensitive while LE and WBC >5 were found to be quite sensitive but not specific.

While I found this information interesting and somewhat helpful, some of it was unreferenced (ie “whiff test” efficacy, normal amounts of WBC per HPF) and the studies reviewed had quite variable results overall.

My search for an objective approach to UTI’s then led me to the UK where the NHS recently studied the issue extensively and even developed a decision rule with LIKELIHOOD RATIOS!  (I ❤ LR’s). Reference:

Little, P., Turner, S., Rumsby, K., Warner, G., Moore, M., Lowes, J.A., Smith, H., Hawke, C., Turner, D., Leydon, G.M., Arscott, A. & Mullee, M. (2009). Dipsticks and diagnostic algorithms in urinary tract infection: development and validation, randomised trial, economic analysis, observational cohort and qualitative study. Health Technology Assessment, 13(19).

Unfortunately, they aptly concluded that their score added little to diagnosis (LR’s only became high enough for practical use if there was a positive nitrite) and had too poor of a NPV to rule it out.  Notably, in their 451 participant validation study they did provide an estimate of the strength of each variable by calculating OR’s for nitrites (5.56), 1+ or more LE (4.49), and blood (2.12).

As for the “whiff test,” I reviewed multiple articles searching for even one that would validate the perception that our nostrils could effectively differentiate between “urinoid” and “pungent” odors.  Unfortunately, none of them showed results that would substantially contribute to diagnosis or workup for UTI leaving my nose dripping in disappointment.

In conclusion, Urinalysis Voodoo (TM) doesn’t seem to be all that out of touch.  Unfortunately, while urinalysis can be helpful, we do not have a great time-sensitive test to rule in or out UTI’s and need to rely on our experience, our history/physical exam, and consider the additional wrinkles that the patient’s status as a febrile child, a female or a pregnant female bring to the situation.  While this certainly won’t be practice changing (that wouldn’t be boring!), I hope that anyone who reads this review will leave with a greater understanding of why we do what we do.

That’s it for this week and my first ever post!  As promised, it was nothing exciting.  In the future I will consider writing about equally unintriguing topics such as IV fluid selection in moderate dehydration, constipation, and any other unsexy topics the crickets or my first and only commenter purdye (http://manuetcorde.wordpress.com/) request.

Feedback/comments are welcome and appreciated – thanks for reading!

Brent Thoma @boringem