First: now you can get email notifications when this blog updates! You won’t receive emails for any other reason, I won’t share your email address with third parties, and you can opt out at any time. Subscribe here
Second: welcome to the new infectious disease fellows starting across the US! You’ve made a great decision. This month I wanted to investigate a question I’ve had for a while about the diagnostic criteria for UTI. If you don’t want to read about that, skip to the bottom of this post for links to this month's ID articles.
The urinary tract infection literature generally defines significant bacteriuria compatible with a UTI as more than 10^5 colony forming units per milliliter (cfu/ml) in a urine culture from an uncatheterized patient, or more than 10^3 cfu/ml in a catheterized patient. The latter definition you can find in the 2009 IDSA CAUTI guideline; the IDSA’s uncomplicated cystitis guidelines actually don’t comment at all on diagnostic criteria, but if you read the UTI literature you’ll see that >10^5 cfu/ml is the most commonly used cutoff. Here was my question: who came up with these quantitative cutoffs for UTI and CAUTI, and on what (if anything) are they based?
To start, I want to be clear that neither urine cultures nor any other test can establish a diagnosis of UTI in the absence of clinical symptoms. Asymptomatic bacteriuria is common and overtreated, and outside of pregnancy, renal transplantation, or impending urologic surgery, treatment of ASB increases adverse events without reducing rates of subsequent UTI, other complications, or mortality.
The traditional teaching has been that normal urine in healthy adults is sterile; you can still find this dogma in Mandell’s, UpToDate, and other respected sources. But it’s not that simple. First, even using culture-based techniques, asymptomatic bacteriuria is detectable in a significant proportion of uncatheterized elderly men and women, and as mentioned above, this is not a pathologic state; it no more represents ‘infection’ than does the gut microbiome. Second, while ‘healthy’ urine may not reliably yield organisms with standard culture techniques, the advent of PCR has lead to the discovery of a normal urinary microbiome containing both bacteria and bacteriophages. In fact, urinary microbiome dysbiosis is associated with a host of genitourinary tract pathologies including recurrent UTI.
Back to the IDSA’s 2009 CAUTI guidelines. Hooton et al propose >10^3 cfu/ml as a compromise cutoff for urine culture positivity that captures most symptomatic CAUTIs while also being feasible for the microbiology lab. The guideline authors point out that as few as 10^2 cfu/ml in cultured urine can be compatible with UTI in a catheterized patient (though most patients with CAUTI will have >10^5 cfu/ml), citing this 1984 NEJM paper by Drs. Randall Stark and Dennis Maki. Stark and Maki noted that if a catheterized patient developed any low-grade of bacteriuria or candiduria, >95% of the time the concentration of organisms would grow to >10^5 cf/ml within 72 hours if not treated. They suggested defining clinically significant bacteriuria in catheterized patients as at least 10^2 cfu/ml because Garibaldi et al had previosuly identified this as the minimum threshold of bacteriuria that was reliably reproducible in catheterized patients (reference here).
The CAUTI guidelines also cite this 1982 NEJM paper by Stamm, in which women with acute dysuria and midstream urine cultures growing as few as 10^2 cfu/ml of coliform bacteria were likely to have the same organisms in the bladder (as assessed by urine culture obtained via suprapubic aspiration). However, a subsequent letter to the editor questions Stamm et al’s methodology, noting that they considered any quantity of coliform growth from the bladder culture proof of UTI even though some of these cultures grew only a few colonies of bacteria and even when the women lacked pyuria (and remember, we now know healthy urine isn’t really sterile). Accordingly, the >10^2 cfu/ml cutoff for urine culture positivity in uncatheterized patients was not widely adopted in subsequent research, though it is mentioned in Mandell’s. Arguably this is a moot point, as in uncatheterized women with acute dysuria we make the decision to treat clinically rather than on the basis of urine culture colony counts.
Anyway, the Stark paper sources the original and more commonly used 10^5 cfu/ml threshold for uncatheterized patients to a Dr. Edward Kass, who demonstrated that this was the threshold of bacteriuria from a mid-stream sample at which an uncatheterized patient’s urine, if left untreated, would repeatedly yield the same organism in similar quantities over time (i.e. indicating it was not a contaminant). The first of Kass’s papers is from 1956 and I couldn’t track it down, but the second has full text available via JAMA IM here. This paper has a beautiful figure 1 showing that while ASB in women spans all concentrations of bacteriuria, patients with pyelonephritis almost universally had bacteriuria of 10^5 cfu/ml or more. So to summarize, here’s my current understanding of how these definitions developed:
Kass 1956-1957: in uncatheterized patients, at 10^5 cfu/ml urine cultures stay positive for the same organism, and nearly all symptomatic patients have this degree of bacteriuria.
Garibaldi 1974: in catheterized patients, urine cultures stay positive for the same organism when as little as 10^2 cfu/ml of bacteria are present.
Stark 1984: in catheterized patients, any amount of bacteriuria, left untreated, will quickly grow to >10^5 cfu/ml.
Hooton 2009: It’s burdensome to have microbiology labs accurately detect less than 10^3 cfu/ml in urine cultures, and most patients with CAUTI have more than 10^5 cfu/ml anyway, so let’s say that in CAUTI the urine culture should grow at least 10^3 cfu/ml.
And that’s your historical sidebar for the month. Here are the articles I have for you:
Antimicrobial agents research this month included drug level monitoring for beta-lactams, isavuconazole, and polymyxin B, the anti-leishmania activity of nelfinavir, and ampicillin-ceftriaxone versus ampicillin-gentamicin for E. faecalis endocarditis.
ID diagnostics research explored the import of different culture techniques for VAP, time-lapse microscopy and MALDI-TOF for rapid identification of antimicrobial resistance, and the diagnostic accuracy of new assays for pneumococcus and syphilis.
General ID topics this month included azithromycin versus doxycycline for murine typhus, chronic HEV infections, ghost vaccines, short-course therapy for community-acquired pneumonia, and the yields of blood cultures and imaging in cellulitis.
HIV and STD research included new IAS guidelines for ART, HIV treatment as prevention, the epidemiology of M. genitalum infection in the southern US, the efficacy of on-demand versus daily PrEP, phase III data on a newly-approved single tablet regimen for HIV, the effect of PD-1 inhibtors on HIV infection, and the importance of ART in treating cervical dysplasia.
Onc and Transplant ID research included topics like T cell infusions for CMV infection, institutional variation in antibiotic use after stem cell transplant, empiric versus pre-emptive strategies for antifungal therapy in persistent febrile neutropenia, C. difficile infection and solid organ graft loss, and donation of organs from patients with HCV infection.
Antibiotic Stewardship and Infection Control topics included probiotics for C. difficile infection, the prevalence of antibiotic overprescription in the US, the safety of bringing stewardship to the ICU, and information control as means to decrease overtreatment of asymptomatic bacteriuria.
As always, if you have questions, comments, or suggestions please drop me a line; the email address is firstname.lastname@example.org and you can find me on twitter at @Cortes_Penfield. Again, if you would like monthly emails notifying you when this site updates, subscribe here