Hunting for Invasive Bacterial Illness in Infants with a Positive UA
/Mahajan P, VanBuren JM, Tzimenatos L, et al. Serious Bacterial Infections in Young Febrile Infants With Positive Urinalysis Results. Pediatrics 2022;150(4). 10.1542/peds.2021-055633
Question
What is the prevalence of bacteremia and/or bacterial meningitis (“invasive bacterial illness”, ISI) in febrile infants ≤60 days of age with a positive urinalysis (UA) result?
Study Design
This study was a secondary analysis of a dataset collected prospectively to examine noncritical febrile infants ≤60 days old who had at least a blood culture collected (original study: doi.org/10.1001/jama.2016.9207).
Thoughts: The original dataset was collected as a convenience sample, ie probably collected when research staff were available for enrollment, which is likely during business hours.
Population
7180 noncritical febrile infants ≤60 days old presenting to a PECARN ED (list of centers: https://pecarn.org/research-nodes/) between March 2011 and April 2019. Fever meant a temperature ≥38C, either in the ED, from a referring facility, or by history. Infants had to have had a UA performed and results from CSF obtained at that ED visit were available.
Exclusion criteria
History of prematurity (<37 weeks)
Significant comorbid conditions
Antibiotic use in the preceding 48hr
Those with critical illness (requiring endotracheal intubation or vasoactive medications)
Thoughts: Note that “tactile fever” measurements at home were likely not included in this study, as there was a specific temperature number requirements. This may have excluded There was also no indication of how a temperature had to be taken, and we all know the variability between temporal, ear, axillary, oral, and rectal temperatures in infants. However, this may be more likely to miss “low grade” fevers as externally measured temps are more often lower than internal ones.
Methods & Definitions
A positive UA was defined by any of the following:
Positive nitrites
Any leukocyte esterase
>5 WBC/hpf
Comparisons for a “positive UA” were done both on the individual components of the UA as listed above and for the aggregate. A positive UTI was defined by the presence of a known urinary pathogen where growth varied by method of collection (≥10,000 CFU/ml for a cath specimen or ≥1,000 CFU/ml for a suprapubic aspirate). This definition of a positive UA and UTI were based on a prior PECARN study (doi.org/10.1542/peds.2017-3068). Definition of bacteremia and bacterial meningitis were defined as growth of a known pathogen. Keep in mind that, unlike in most of adult medicine, the common practice at some pediatric EDs is to collect only a single blood culture, so there was no discussion in this paper about single specimen growth of possible contaminants like staph epidermidis.
The authors also described their results in two groups: ≤28 days old and those 29-60 days old. They also did a subgroup analysis on those infants aged 22-28 days which was likely due to the recent AAP guidelines (doi.org/10.1542/peds.2021-052228 although the results here don’t support treating this group any differently in regards to this question).
In assessing possible predictor variables for IBI the authors included the following:
Age
Qualifying temperature
Yale Observation score (doi.org/10.1542/peds.2017-0695)
WBC count
Absolute neutrophil count
Procalcitonin (one model done without procal and one with)
They repeated the same model for factors associated with IBI in febrile infants with UTIs. Their last outcome was determining the prevalence of IBI in febrile infants with a positive UA who were deemed low risk by the PECARN febrile infant rule.
Results
1090 of the 7180 eligible subjects (15.2%) had a positive UA. Of those with a positive UA, 541 (49.6%) ended up having a UTI. Reassuringly, of those with a negative UA only 0.8% had a UTI.
One of the authors’ biggest conclusions comes from the group of infants aged 29-60 days with a positive UA: of those 697 patients, there were no cases of bacterial meningitis. In other words, in a febrile, noncritically ill, not otherwise baseline high risk (comorbidities or prematurity), infant aged 29 dasy and older with a positive UA, you can safely skip the LP. Supplemental table 4 has some interesting results regarding the pathogens responsible for each type of infection and if this paper at all interests you (or if you’re just wondering about antibiotic coverage) I’d suggest looking it over. One notable thing is that, in this cohort, the pathogens commonly found in the urine aren’t the same as those commonly found in the CSF.
When looking at the multivariate model for those with a positive UA that did include procalcitonin, only younger age and serum procalcitonin were independently associated with IBI. In the model that didn’t include procalcitonin, younger age, higher temperature, and higher ANC were independent predictors. Happy news for those of us using the PECARN febrile infant rule already, in those patients who were classified as low risk (by ANC & procalcitonin) with a positive UA, there were no cases of IBI for any age.
Thoughts: Figure 1 has some interesting takeaways. Of those patients with a negative UA result (6090), 0.6% ended up having bacterial meningitis (34). In comparison, for a positive UA results 0.4% of those patients had bacterial meningitis (4). This is different than the pattern of bacteremia which was seen more often (5.8%, 63 patients) in those with a positive UA than in those with a negative UA (1.1%, 69 patients). The authors point this out by mentioning that the higher rate of IBI in patients with a positive UA was driven by those with bacteremia. I would tend to agree with their unsaid conclusion that a negative UA in these infants is less reassuring against bacterial meningitis and I would be getting the LP.
So why did I choose to discuss this article? The LP is a time suck in the ER. However, they’re really technically easy to do on an infant and knowing if an infant has meningitis versus a different infection that requires a much shorter course of less intense antibiotics in a developing gut biome is a pretty important result. This study had a pretty big data set when you’re thinking about cultures from urine, blood, and CSF and the authors looked at the data in a whole bunch of different, interesting, and useful ways. I like their conclusions and I consider it to be either practice affirming or practice changing.
Now, if only someone could come up with a decision rule to determine what degree of osteoarthritis will make getting the LP on the altered granny impossible, I could really sleep better at night.
Authorship
Written by, Audio Podcast by Marlena Wosiski-Kuhn, MD, PGY-3, University of Cincinnati Department of Emergency Medicine
Editing, Posting, Peer Review by Jeffery Hill, MD MEd, Associate Professor, University of Cincinnati Department of Emergency Medicine
Cite As
Wosiski-Kuhn, M., Hill, J. Hunting for Invasive Bacterial Illness in Infants with a Positive UA. TamingtheSRU. www.tamingthesru.com/blog/journal-club/hunting-for-invasive-bacterial-illness-in-infants-with-a-positive-ua. 7/3/2023
