Articles of the month Special Edition: Pediatric UTI

A review of the evidence surrounding the diagnosis and management of urinary tract infection in pediatric patients

Pediatric urinary tract infections – not the most exciting topic in emergency medicine, but a topic near and dear to me because it was my first exposure to the value of evidence based medicine. You may not believe it, but as I resident I hated evidence based medicine. It seemed like a lot of work to read dense papers full of opaque statistical tests when there were guidelines that would summarize all that evidence for me. So as a resident, I just religiously followed the guidelines. That meant sending hundreds of children to the ED for cath urines, ordering VCUGs, and prescribing a lot of antibiotics. Unfortunately, once I was exposed to the evidence, this management plan seemed like madness. In fact, I think the practice I learned in residency was probably hurting children. So let’s look at some evidence.

Note: What I am discussing here is the diagnosis and treatment of UTI in well-appearing febrile children without specific urinary symptoms- which represents the majority of pediatric UTIs. Acutely unwell or septic children should be treated as such.

 

Before we jump into the treatment of UTI, a quick note about the accuracy of our UTI gold standard:

Part 1) What are the chances this child has a UTI?

The urine culture came back positive – slam dunk UTI right? Not so fast. Many children have bacteria swimming around their bladders even when they are symptom free.

Kumar CS, Jairam A, Chetan S, Sudesh P, Kapur I, . Asymptomatic bacteriuria in school going children. Indian journal of medical microbiology. 20(1):29-32. 2007. PMID: 17657020 [free full text]

1817 asymptomatic children aged 11-15 were screened for bacteriuria using clean catch mid-stream urines. Anyone with potential urinary symptoms was excluded. 192 children (10.6%) had positive urine cultures despite lacking any symptoms. One might question the generalizability of this data, but I am not sure there is a really good theory that explains to me why would we expect bacteria in the urine of asymptomatic children in one country and not another.

Jha BK, Singh YI. Prevalence of asymptomatic bacteriuria in school going children in Pokhara valley. Kathmandu University medical journal (KUMJ). 5(1):81-4. 2008. PMID: 18603991

502 asymptomatic children aged 5-13 were screened with clean catch mid stream urines. 7 (1.4%) were positive for bacterial pathogens known to cause UTIs.

Roberts KB, Charney E, Sweren RJ. Urinary tract infection in infants with unexplained fever: a collaborative study. The Journal of pediatrics. 103(6):864-7. 1983. PMID: 6644419

This study compared the rate of positive urine culture in children with and without fever. Although the rate of positive culture was higher in children with a fever (4.1%), I think it is worth noting that 0.7% of health, afebrile females ages 2-24 months had 3 consecutive urine cultures that were positive, despite their lack of symptoms.

Ralston S, Hill V, Waters A. Occult serious bacterial infection in infants younger than 60 to 90 days with bronchiolitis: a systematic review. Archives of pediatrics & adolescent medicine. 165(10):951-6. 2011. PMID: 21969396

I frequently hear that kids with obvious clinical bronchiolitis have a high rate of serious bacterial illness (meaning UTI). Indeed, if you culture children with bronchiolitis, you will often find bacteria in the urine. I won’t go through all the individual studies, but this is a review that covers 11 studies. Over the 11 studies, 3.3% of children with clinical bronchiolitis had positive urine cultures. Although the individual studies frequently concluded that we should be culturing all these kids, I think the opposite conclusion (hinted at by this review) is correct. The RSV is obviously the cause of these kids’ fevers – the urines are false positives. This just fits with the previous studies indicating that a percentage of asymptomatic children, or children with non-urinary sources of fever, will also happen to have bacteria in their urine.

Robinson JL, Finlay JC, Lang ME, Bortolussi R, . Urinary tract infections in infants and children: Diagnosis and management. Paediatrics & child health. 19(6):315-25. 2014. PMID: 25332662  [free full text]

I include this reference only so I can copy this quote: “Although positive urine cultures occur with bronchiolitis, it is probable that most positive urine cultures in infants >2 months of age with bronchiolitis are caused by contamination or asymptomatic bacteruria.”

Bottom line: Children will have bacteria in their urine. It frequently will not represent an infection. Even if you happen to get a positive culture for a true pathogen, there is still a reasonable chance that it is a false positive in the absence of urinary symptoms.

So 1-10% of children will have asymptomatic bacteriuria. How does that compare to febrile children?

Hoberman A, Chao HP, Keller DM, Hickey R, Davis HW, Ellis D. Prevalence of urinary tract infection in febrile infants. The Journal of pediatrics. 123(1):17-23. 1993. PMID: 8320616

In this observational study of 945 febrile infants, the authors found that 50 (5.3%) had positive urine cultures that they concluded were UTIs.

Shaikh N, Morone NE, Lopez J. Does this child have a urinary tract infection? JAMA. 298(24):2895-904. 2007. PMID: 18159059

This is a systematic review that includes 12 studies, looking at the rate of positive urine culture in febrile children without an obvious source. In females, the rate of positive culture was 7.5% in those less than 3 months, 7.2% in patients 3-12months, and 2.1% in those 12-24 months. In males, the rates were 8.7% less than 3 months and 2.2% in those 3-12 months of age.

Bottom line: It is hard to know what this means. 1-10% of children always have bacteria in their urine, and here 5-9% of febrile children have bacteria in their urine. There is obviously a large overlap between those numbers. How are you supposed to pick out the kids who have true UTIs from those with asymptomatic bacteriuria with an incidental viral illness? Clearly there will be a large number of false positives.

 

False positives: Is the urine culture like a smoke alarm?

Uses of a fire alarm (False positives).jpg

Part 2) Why do we care about UTI?

Why do the guidelines say that we should be treating pediatric UTI?

Roberts KB. Urinary tract infection: clinical practice guideline for the diagnosis and management of the initial UTI in febrile infants and children 2 to 24 months. Pediatrics. 128(3):595-610. 2011. PMID: 21873693 [free full text]

The AAP guidelines for diagnosis and management of UTI give us these reasons for treating pediatric UTI:

    • UTI can cause acute renal injury
    • Kidney damage increases with delay in diagnosis and recurrent disease
    • Diagnosis allows us to detect obstructive abnormalities or severe reflux
    • UTI can cause hypertension and ESRD later in life

Robinson JL, Finlay JC, Lang ME, Bortolussi R, . Urinary tract infections in infants and children: Diagnosis and management. Paediatrics & child health. 19(6):315-25. 2014. PMID: 25332662  [free full text]

The CPS (Canadian Pediatric Society) guidelines never actually state the reason that treatment of UTI is suggested (unless I missed something) – they just assume that treatment should be done. They do briefly mention renal scarring, but only to mention the risk is really low.

Bottom line: The guidelines discuss a variety of long term renal outcomes as the reason for diagnosis and treatment of pediatric UTI. Of note, preventing sepsis and decreasing mortality are not among the listed reasons for managing UTI.

However, UTIs are “serious bacterial illnesses”, so we probably want to ask:

What is the risk of sepsis from pediatric UTI?

Watson RS, Carcillo JA, Linde-Zwirble WT, Clermont G, Lidicker J, Angus DC. The epidemiology of severe sepsis in children in the United States. American journal of respiratory and critical care medicine. 167(5):695-701. 2003. PMID: 12433670

This is a database study that looked at over 1.5 million pediatric admissions in 7 American states during 1995. They identified a total of 9675 cases that met the criteria for severe sepsis. This translates to a incidence of sepsis of 0.5/1000 children per year. Lesson #1: Severe sepsis is rare in children. This rate is higher the younger you are (1.5/1000 between 1 month and a year, and 3.6/1000 in the first month of life). Only 4% of all pediatric sepsis had a GU source, and half of all children with severe sepsis had previously diagnosed contributing comorbidities. So, lesson #2: in healthy children, sepsis from a urinary source is extremely rare. I would also note that the mortality was lower when sepsis resulted from a urinary source (4% as compared to 10% in the overall cohort.)

Hypothetical math

Let’s consider a hypothetical calculation: If only 0.5/1000 children will have severe sepsis per year, and only 4% of pediatric sepsis is from a GU source, that means that only 1 child out of every 25,000 children will have sepsis caused by a UTI. Compare that to how often you diagnose a child with a “serious bacterial illness” because they happened to have bacteria in their urine. The numbers almost certainly don’t match.

Bottom line: The reason we are supposed to be treating pediatric UTIs is not to prevent sepsis, but rather to prevent renal damage, hypertension, and end stage renal disease later in life.

 

Part 3) Does treating UTI prevent renal damage?

First, we need to ask, how common is renal scarring in the setting of UTI?

Faust WC, Diaz M, Pohl HG. Incidence of post-pyelonephritic renal scarring: a meta-analysis of the dimercapto-succinic acid literature. The Journal of urology. 181(1):290-7; discussion 297-8. 2009. PMID: 19013606

This is a systematic review and meta-analysis looking at the incidence of renal scarring after imaging confirmed pyelonephritis. Overall, if you have pyelonephritis, there is a 41.6% chance that we will see renal scarring on up to 3 month follow up DMSA (99mtechnetium dimercapto-succinic acid) scans.

Hewitt IK, Zucchetta P, Rigon L. Early treatment of acute pyelonephritis in children fails to reduce renal scarring: data from the Italian Renal Infection Study Trials. Pediatrics. 122(3):486-90. 2008. PMID: 18762516

This is a retrospective look at the data from two prospective RCTs. Of the children with initial evidence of renal scarring, 30% still had evidence of scarring on a repeat DMSA scan at one year.

Bottom line: Renal scarring is a real thing, but only about 12% of children with imaging confirmed pyelonephritis will have any evidence of renal scarring after one year (40% will have it initially, only 30% of whom will still demonstrate it one year later).

So we have an imaging based finding that we call renal scarring. Does this radiographic finding have any patient oriented clinical correlates?

When you read reviews and guidelines on this topic that claim that renal scarring matters, there are two commonly cited papers:

Gill DG, Mendes de Costa B, Cameron JS, Joseph MC, Ogg CS, Chantler C. Analysis of 100 children with severe and persistent hypertension. Archives of disease in childhood. 51(12):951-6. 1976. PMID: 1015848 [free full text]

This is a retrospective look at 100 children seen at a single children’s hospital with persistent hypertension. The authors decided that the cause of persistent hypertension was chronic pyelonephritis in 14% of the children. There was no confirmation that pyelonephritis was the etiology, just the opinion of the clinician. Also, this is an extremely select group of sick patients that is unlikely to represent the general pediatric population.

Shaw KN, Gorelick MH. Urinary tract infection in the pediatric patient. Pediatric clinics of North America. 46(6):1111-24, vi. 1999. PMID: 10629676

Although this article is referenced elsewhere, they just make a single statement that “out of 161 children with UTIs, there are GFR changes later in life”. The statement is unreferenced. Perhaps it refers to unpublished research, but it is incredibly difficult to know what to do with such a statement.

Bottom line: The cited literature to support the idea that renal scarring is clinically important is very unimpressive.

 

Those papers are really sort of pathetic, but luckily for us there are other studies to help us answer this question. I’m not sure why authors of pediatric UTI review papers can’t seem to find these studies.

Martinell J, Lidin-Janson G, Jagenburg R, Sivertsson R, Claesson I, Jodal U. Girls prone to urinary infections followed into adulthood. Indices of renal disease. Pediatric nephrology (Berlin, Germany). 10(2):139-42. 1996. PMID: 8703696

This is a prospective observational trial that followed 111 female pediatric patients for a mean of 15 years after their first diagnosis of UTI. 54 of the girls had renal scarring. As compared to controls, there was do difference in blood pressures or renal function among girls with UTIs but no scarring and those with mild renal scarring. There were statistically, but probably clinically insignificant differences noted for those with severe scarring.

Group Mean GFR Mean BP
Control (n=48) 111 114/71
No scarring (n=57) 111 113/74
Mild scarring (n=35) 107 117/75
Severe scarring (n=19) 96 121/78

Wennerström M, Hansson S, Jodal U, Sixt R, Stokland E. Renal function 16 to 26 years after the first urinary tract infection in childhood. Archives of pediatrics & adolescent medicine. 154(4):339-45. 2000. PMID: 10768669

This is a prospective observational trial. There was an original trial that included 1221 pediatric patients with UTI. Of those patients, there were 68 with renal scarring and 57 agreed to long term follow up. There were also 51 patients without scarring that participated in long term follow up. The follow up was between 16 and 26 years. The GFR was the same whether or not you had scarring (99 in both groups). There were some changes in secondary outcomes, such as measuring the GFR of individual kidneys as compared to total GFR, but none appear to be clinically significant.

Wennerström M, Hansson S, Hedner T, Himmelmann A, Jodal U. Ambulatory blood pressure 16-26 years after the first urinary tract infection in childhood. Journal of hypertension. 18(4):485-91. 2000. PMID: 10779101

This is the same cohort of patients as the last study after the same period of follow up, but they decided to share their blood pressure measurements in a second publication. There were no significant difference between the groups.

Bottom line: Renal scarring does not appear to have any clinically important impact on renal function or blood pressure.

If you listen to Jerry Hoffman on this issue, he likes to point out that in the era when children actually had serious bacterial illnesses such as meningitis and epiglottitis (before my time), UTIs were never on the radar. With so many children presumably untreated for their UTIs in the past, where are all the adults dying of end stage renal disease from their renal scarring? Well, these authors help us answer that question:

Salo J, Ikäheimo R, Tapiainen T, Uhari M. Childhood urinary tract infections as a cause of chronic kidney disease. Pediatrics. 128(5):840-7. 2011. PMID: 21987701 [free full text]

This is a report of a systematic review of the literature as well as a retrospective chart review. In the review, they identified 1576 cases of end stage renal disease, and childhood UTI was not identified as the cause of any of them. Of the 308 patients at their single hospital, childhood UTI was a possible (but not definitive) cause of renal failure in only a single patient.

So summing everything up: it appears very unlikely that childhood UTI has any impact on long term, patient oriented outcomes. We do know that there are a percentage of patients with UTI who will develop renal scarring. Forgetting for the moment that renal scarring appears to be clinically irrelevant, we might want to ask ourselves if it is even possible to prevent this imaging finding.

The AAP guidelines make the statement that “clinical and experimental data support the concept that delay in instituting appropriate treatment of acute pyelonephritis increases the risk of renal damage.” For this statement they cite:

Smellie JM, Poulton A, Prescod NP. Retrospective study of children with renal scarring associated with reflux and urinary infection. BMJ (Clinical research ed.). 308(6938):1193-6. 1994. PMID: 8180534 [free full text]

This is was a retrospective chart review of a a group of 52 children who were involved in a study of vesico-ureteric reflux (so a non-representative group of patients) who had renal scars. There were no real chart review methods and no controls. They say that a delay in diagnosis led to the development of scars in 50 of the 52 patients, but seeing as all the children had scars, you obviously can’t make that conclusion using these methods. Also, their definition of a delay was sort of crazy – for example, they argued that 4 children should have been given prophylactic antibiotics based solely on a family history.

So that doesn’t help us much. Luckily, there are rwo studies that actually looked prospectively at children with UTIs that can help us answer the question:

Hewitt IK, Zucchetta P, Rigon L. Early treatment of acute pyelonephritis in children fails to reduce renal scarring: data from the Italian Renal Infection Study Trials. Pediatrics. 122(3):486-90. 2008. PMID: 18762516

This is a secondary analysis of prospective data from two RCTs looking that the antibiotic management of pediatric UTI. There were 287 patients with initial renal scarring who underwent repeat imaging 1 year later. 31% of children still had scarring on repeat scan. When comparing children who had only 1 day of fever (no delay to diagnosis) with those with 5 days of fever (delayed diagnosis) prior to antibiotics, there was no difference in the rate of scarring. Of course, this is not randomized data, so bias from confounders is possible. It is, however, the best data we have.

Doganis D, Siafas K, Mavrikou M. Does early treatment of urinary tract infection prevent renal damage? Pediatrics. 120(4):e922-8. 2007. PMID: 17875650

This is a prospective observational trial of 278 patients comparing the length of time from fever onset to antibiotics being started and the rates of renal scarring. Early antibiotics were associated with less changes on imaging during the acute stage of the illness, there was no difference in renal scarring at one year.

Bottom line: In health appearing children, there doesn’t seem to be a rush to treat UTI.

 

Although that doesn’t help us determine if antibiotics help at all (versus no antibiotics), there at least doesn’t seem to be any reason to rush to diagnose UTI before fever has been present for 5 days.

Part 4) What workup is required after you diagnose a child with a UTI?

This is an area that the guidelines are at least catching up on. During my training, the guidelines suggested that children have ultrasounds and VCUGs. However, the most recent guideline by the AAP no longer recommends VCUG after a single UTI, but they do state that it is required “if renal and bladder ultrasonography reveals hydronephrosis, scarring, or other findings that would suggest either high-grade VUR or obstructive uropathy and in other atypical or complex clinical circumstances. VCUG should also be performed if there is a recurrence of a febrile UTI”.

The guidelines do suggest that children should routinely have ultrasounds done after a UTI, based on level C evidence. Their logic is basically that ultrasound is noninvasive and occasionally finds things so it should be offered, although the do recognize the harms of false positives.

If you were going to order these tests, how often are they positive?

Hoberman A, Charron M, Hickey RW, Baskin M, Kearney DH, Wald ER. Imaging studies after a first febrile urinary tract infection in young children. The New England journal of medicine. 348(3):195-202. 2003. PMID: 12529459 [free full text]

This is a prospective observational trial of 309 children after their first UTI, following the older AAP guidelines. For ultrasounds 88% were normal, leaving 12% abnormal. However of those 12%, not a single one changed the management of the child. For the VCUG, 40% had evidence of vesico-ureteric reflux. However, the authors very astutely point out that this is only a valuable finding if we can do something about it. (Spoiler: we can’t). Looking at DMSA scans, 60% had evidence of pyelonephritis originally, but only 9% had scarring on repeat scan.

Bottom line: If you order tests, you will find things. But the things you find will have no clinical value to the patients.

 

What is the value of the VCUG?

From the previous study, we know that if we perform VCUGs on children after their first UTI, 40% will have evidence of vesico-ureteric reflux.  (As an aside, anything that is present as often as 40% always seems more likely to be a normal variant than pathologic to me.) So kids have reflux. The questions is: can or should we do anything about it?

Nagler EV, Williams G, Hodson EM, Craig JC. Interventions for primary vesicoureteric reflux. The Cochrane database of systematic reviews. 2011. PMID: 21678334

This Cochrane review identified 20 studies including 2324 patients looking at the evidence for treatment options of vesicoureteric reflux. Giving prophylactic antibiotics did not reduce symptomatic UTI (RR 0.68, 95% CI 0.39 to 1.17) or febrile UTI (RR 0.77, 95% CI 0.47 to 1.24). (Also, giving antibiotics every day so you can avoid giving them once in awhile doesn’t really make any sense.) Surgery alone was no better than prophylactic antibiotics. The combination of surgery and daily prophylactic antibiotics decreased the number of febrile UTIs by about 12% over 5 years, but had no effect on renal function. (So you can have 8 children undergo surgery AND take daily antibiotics in order to prevent 1 of them from having to take antibiotics one time for a UTI over 5 years. Think about that for a moment.)

Finnell SM, Carroll AE, Downs SM, . Technical report—Diagnosis and management of an initial UTI in febrile infants and young children. Pediatrics. 128(3):e749-70. 2011. PMID: 21873694 [free full text]

Another meta-analysis looking at the question of prophylactic antibiotics, this one was part of the 2011 AAP guidelines. They conclude that antibiotic prophylaxis is not beneficial in children with vesicoureteric reflux. Prophylaxis did lead to an increase in antibiotic resistance.

Venhola M, Uhari M. Vesicoureteral reflux, a benign condition. Pediatric nephrology (Berlin, Germany). 24(2):223-6. 2009. PMID: 18604562

I have hinted that I doubt the clinical significance of this reflux thing, but I am just a lowly emergency doctor, so I would understand if you didn’t believe me. This is a review and analysis by pediatric nephrologists that comes to the conclusion that VUR is a benign condition.

Bottom line: There is no treatment for VUR that results in clinically important outcomes. (One may wonder if reflux is even really a pathological condition.) Therefore, finding it on a VCUG is utterly useless.

 

What is the value of ultrasound?

Finnell SM, Carroll AE, Downs SM, . Technical report—Diagnosis and management of an initial UTI in febrile infants and young children. Pediatrics. 128(3):e749-70. 2011. PMID: 21873694 [free full text]

Again, this is the review done as part of the 2011 AAP guidelines. They conclude that only 15% of ultrasounds done after a first time UTI will find an abnormality. More importantly, only 10% of those positive findings (1.5% overall) were potentially clinically relevant, but even these didn’t clearly change management. This review also doesn’t tell us if there might have been other clinical features that might have indicated those few true positives, rather than just indiscriminately imaging everyone. That compares to about 24% of the positive ultrasounds (2-3% overall) that will be false positives and lead to unnecessary testing or interventions.

Bottom line: Ultrasound after first UTI is incredibly low yield, and it is not clear that the benefits outweigh the risks.

 

Wrapping it all up

That was a lot of messy evidence. Here are my simple conclusions that guide my practice:

 

  • None of this evidence is high quality – so any conclusions (whether by me or the guidelines) are imperfect.

  • If the child is unwell, culture them as you think is appropriate and treat.

  • In well appearing children, UTI does not seem to cause long term renal problems and the risk of sepsis is incredibly low.

  • In well appearing children, there does not appear to be a difference between starting antibiotics for UTI on the first day of fever or on day number 5.

  • Given the rate of asymptomatic bacteriuria, you should anticipate a high rate of false positive urine cultures.

  • Therefore, it makes sense to wait until at least the fifth day of fever before testing for UTI. This will result in fewer urines being sent and fewer false positives (because many fevers will resolve before day 5), but does not seem to be associated with any harm

  • There does not seem to be any value in routine imaging of children with a first time UTI. Selective imagining based on history and physical makes more sense, but even then it is not clear that imaging changes management.

 

 

Author: Justin Morgenstern

Community emerg doc, FOAM enthusiast, evidence junkie “One special advantage of the skeptical attitude of mind is that a man is never vexed to find that after all he has been in the wrong.” - William Osler

13 thoughts on “Articles of the month Special Edition: Pediatric UTI”

    1. Thanks for the comment.
      First, let me say that the management of pediatric UTI varies a lot around the world, and the post is primarily based on what I have seen as common practice in North America.
      The vast majority of children treated for potential urinary tract infections in North America do not have specific urinary symptoms. They are worked up just because they have a fever. That is why the guidelines talk about renal disease and not managing discomfort – the antibiotics can’t be for the dysuria, because there is none.
      If a child has specific urinary symptoms, I absolutely think they should be treated with antibiotics, with the goal being comfort.
      I have updated the preface to try to emphasize that my conclusions are focused on children without urinary symptoms.

      Like

      1. Sorry, rereading my comment it wasn’t as friendly as I’d intended!

        I’d assumed that a UK population and US population would have pretty similar demographics, but assumptions are dangerous things!

        I completely agree with you; we desperately search for the source of an infection and a positive urine gives us a possible, too easy, win

        Like

  1. So obviously <29 days with fever UA and UC as part of full septic work up. What about 30-90 days of age? What is your practice with this age group. Loved this article as it backs up what Ive been doing in my practice. Aggressively and most often unnecessarily obtaining cathed urine samples in pediatric patients feels a lot like child abuse to me! Thanks for putting this out there!

    Like

    1. There is not really an evidence based answer – I tend to be pretty cautious up until the 2 month vaccines, but seeing as we don’t vaccinate against UTI pathogens, I’m not sure there is any logic in that. I am still not worried about the renal outcomes for these young children, but the chances of sepsis are higher.
      Practically speaking – under 2 months of age I am probably still getting a urine, whereas over 2 months I will generally hold off as long as the child looks well on physical exam.

      Like

  2. Thank you very much for such a nice article. I have a question. What are the indications/guidelines to call a pediatrician on call in a child diagnosed in the ED with a UTI? Thanks

    Like

Leave a Comment

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s