Riley Pediatric Hospital Journal Club

An interactive, virtual journal club for pediatricians

  • Enter your email address to subscribe to this blog and receive notifications of new posts by email.

    Join 26 other followers

  • Archive by month

  • Index

Rule-out sepsis, rule-out breastfeeding?

Posted by ashriner on September 21, 2015

BF rocksReviewer:  Pat Clements, MD

Article:  Mukhopadhyay S, et al.  “Effect of early-onset sepsis evaluations on in-hospital breastfeeding practices among asymptomatic term neonates.”  Hosp Pediatr.  2015 Apr;5(4):203-10. doi: 10.1542/hpeds.2014-0126.

Link:  Click Here

Why this article:  For newborns, I feel the area of suspected neonatal sepsis and chorioamnionitis is full of variations in practice. From variations in what obstetricians label as “chorio” to how long we keep infants in the special care nursery (SCN) while on antibiotics, the literature doesn’t have a lot of definitive information to offer. There are many times in the newborn/NICU setting where breastfeeding takes a backseat to other interventions. This is one of the first studies that looks at the link between early-onset sepsis evaluation and delays in breastfeeding.

Methods:  The study was a secondary observational analysis of data from a previous study looking at epidural analgesia at Brigham and Women’s and Massachusetts General Hospitals. The study group included 820 first-time mothers who intended on breastfeeding after birth. The goal was to look at rates of breastfeeding initiation within the first 2 hours of birth and the use of supplementation in the first 24 hours after birth. Infants were all well-appearing and full term. Infant’s excluded from the study were those with any issues of hypoglycemia or an NPO order. Of the 692 dyads enrolled, 101 infants were separated from their mothers for early-onset sepsis (EOS) evaluation. Criteria for EOS were based on the GBS prevention guidelines from the CDC (including maternal fever >100.4, inadequate GBS treatment, membrane rupture >18 hrs), as well as OB diagnosis of chorioamnionitis. Infants were assessed to be well appearing and had blood cultures and labs drawn, and antibiotics started if necessary. All infants were returned to mom’s room after initial workup initiated, and received antibiotics in the postpartum room if required.

Results:  Of the 101 babies evaluated who were separated from their mother for EOS evaluation after birth, 47% of them had delayed onset of breastfeeding (classified as first feed >2 hrs after birth). Only 13% of infants who were not separated had delays in breastfeeding. The group looked at multiple other variables related to EOS, such as degree of maternal temperature, membrane rupture time etc, and no factors related to the clinical picture itself seemed linked to delayed onset of breastfeeding. The main factor related to delayed onset of breastfeeding was the amount of time child was separated from mother within the first 2 hours after birth. Formula supplementation in the first 24 hrs was also significantly associated with separation for EOS evaluation, with 19% of infants separated ultimately receiving supplementation compared to 10% of infants who were not separated.

Conclusion:  The authors concluded that separation of infants from their mothers in the first 2 hours after birth for EOS evaluation caused delays in breastfeeding and increased the likelihood of receiving formula supplementation. They discussed limitations to the length of the supplementation data being only 24 hrs after birth, as many of the babies in this study were discharged between 24-48 hrs. They did point out, however, other studies which have shown that breastfeeding within 1 hour of birth and avoiding supplementation were strong predictive factors for exclusive breastfeeding at 6 weeks for infants. They also discussed the challenge with generalizing this data to other populations, as each hospital may have their own protocols and criteria for which babies require EOS evaluations.

Discussion:  I felt this study examined some fairly interesting questions. In terms of methods I was a little leery about subjective declaration of “chorio” by OB’s, as there seems to be wide variations in practice (at least at our hospitals). In the end the authors concluded that it wasn’t the EOS evaluation itself that seemed to matter with breastfeeding and supplementation, but rather the act of being separated from mom in general. This made sense to me, and also gave me caution in thinking of all the disruptive activities that take place in the first two hours of birth that might also disrupt breastfeeding (from newborn meds and obtaining measurements to the throng of excited relatives rushing in to visit). I was also interested in how the group kept well-appearing infants in the postpartum room with mom after labs were drawn and IV’s were placed. In our group’s practice I feel even healthy appearing infants may stay in the NICU for 12-24 hours of “observation” while on antibiotics. If this study showed the potential disruptive power of a 1-2 hr separation from mom, it makes me fear what the effects of longer disruptions would be. It would also be interesting to see other data on breastfeeding longer than 24 hrs (i.e do babies that received supplementation in the hospital related to EOS evaluations go on to successfully breastfeed after discharge?) Obviously metrics and predictive tools for neonatal sepsis could be the topic of a whole other journal club. Hopefully this article will remind us to not let breastfeeding be pushed to the back burner!

Posted in Breastfeeding | 1 Comment »

Top 10 articles from PHM 2015

Posted by ashriner on September 2, 2015

Top 10Reviewer:  Emily Webber, MD

What is this?  The Top 10 articles in pediatric hospital medicine presentation is always well attended at the annual PHM conference. Not only is the lecture known for its witticisms, entertainment value and plated lunch, but it also helps paint a picture of the articles that have shaped inpatient pediatric practice of the preceding 12 months.   Two of our own, Dr. Michele Saysana and Dr. Ben Bauer, presented in this venue in 2012 and 2013, respectively.

For the last 3 years, I have worked with the other ‘social media’ faculty to chronicle the links and articles via Twitter, which is convenient for later reflection. What you have in this post is a curated collection of those tweets for your review.

Complex medical research should never be reduced to a mere 140 characters, so the tweet is just the hook to the real article.   They are grouped by theme; I found the hyperbilirubinemia, dehydration and antibiotic overuse articles most relevant to my own practice. If you want to look at some of the other images and discussion from the lecture, search #PHM15 and the feed for @PHMConf. Thank you to everyone who helped attendees attend #PHM15, get send your work in for the #PHM16 submissions process – it’s in Chicago next year!

Enjoy, and please respond to the quiz at the end!


Fevers and antibiotics

  1. Fever w/o source: Biondi et al BCx Time to Positivity in Febrile Infants w Bacteremia , from the PRIS network. …
  2. Antibiotic overuse: Keren et al Comp effectiveness IV vs PO abx in post-d/c tx acute osteomyelitis in children. …
  3. Schroeder/Ralston review article: IV abx duration for common bact infxn in kids …


  1. Bronchiolitis mgmt: Schuh et al Effect of Oximetry on Hospitalization in Bronchiolitis …
  2. The CPG that Shall Not Be Named: Ralston et al. Diagnosis, Management, and Prevention of Bronchiolitis …


  1. Readmissions: Gay et al Rates. impact of pot. preventable readmits @ child hosp …

Venous Thrombus Embolism (VTE)

  1. VTE prophylaxis: Meier et al. Venous thromboembolism in hospitalized adolescents …

Surgery co-management

  1. Another one: Rappaport et al. Pediatric hospitalist comanagement of surgical patients …


  1. Hyperbilirubinemia: Wu et al. Risk for Cerebral Palsy in Infants w/ TS Bilirubin Levels ≥ Exch Tfxn Threshold
  2. Hyperbilirubinemia: Kuzniewicz et al. Incid, Etiol, Outcomes of Nwbrn Hazard Hyperbili …

Complex Care

  1. Medically Complex Children: Bogetz et al Challenges, Sol’ns to Educating Learners About Pediatric Complex Care …
  2. Nolan et al; Diff Percepxn re: QoL, Inpt Tx Goals for Children w/ Severe Disabilities …

Dehydration and IVF

  1. IV Fluids: McNab et al NS vs 0.45NS maint IV fluids for children in hospital @TheLancet …
  2. IV Fluids: Friedman et al isotonic vs hypotonic IV maintenance fluids (in children) …

Honorable mention

Auger et al Summary of STARNet: Seamless Transitions and (Re)admissions Network …


Posted in Uncategorized | Leave a Comment »

Hey, here’s your discharge. So I’ll call you, maybe?

Posted by ashriner on August 5, 2015

Ped phone call

“Oooooooh….so I’m supposed to take that THREE times a day?!”

Reviewer:  Andrew Shriner, MD

Article: Heath J, Dancel R, and Stephens JR.  “Postdischarge phone calls after pediatric hospitalization:  An observational study.”  Hosp Pediatr.  2015 May; 5 (5):241-8.  doi:  10.1542/hpeds.2014-0069. (Link)

Why this article?  Transition of care for children after a hospitalization has been a focus of a significant amount of research in an attempt to reduce issues such as misunderstanding of discharge instructions, difficulty filling prescriptions or giving medications appropriately or readmissions.  While there’s ongoing debate about the preventability of pediatric readmissions, with recent healthcare changes, this focus is unlikely to shift in the near future and there are certainly other ways that patients and families may benefit from efforts to improve the transition of care process.

Methods:  Over a 12-month period, the hospitalists in this study implemented a standardized attending physician phone call to the family of all discharged patients within 72 hours of discharge to check in on any problems with the transition of care.  Charts were reviewed for any difficulties identified, readmissions (14 and 30-days), ED visits, patient satisfaction scores.  Findings were compared to patient’s discharged over the preceding 12 months.

Results:  There were 661 discharges over the study period and 606 discharges in the comparison group from the prior year.  Seventy-eight percent of families (n = 513) were reached via phone call during the study period (up from 50% prior to their intervention).  A problem was identified in nearly 20% of the families who were reached with about half of the issues being related to medication issues.  About 1/3 of the problems identified were found to need a significant action to be taken by the physician (such as contacting another provider or calling in a new prescription).

There were no statistically significant reductions in ED visit rates, readmissions or patient satisfaction scores, but there was a trend towards improvement in all of these outcomes.

Conclusions/Discussion:  This was an interesting study on looking at how to reduce the risk of poor outcomes after inpatient discharges for pediatric patients.  There are some things that made it difficult to determine the impact of this intervention, including the low patient numbers to determine a significant difference on an overall unlikely complication like readmission and the fact that many providers in the group were already performing post-discharge phone calls, so their comparison group was sub-optimal in that respect.

Whenever I read articles like this, I always try to imagine what this would look like if implemented at Riley.  While I think the results noted are encouraging, thinking of the possibility of applying this to our practice seems somewhat overwhelming.  While I’m not sure of our exact number of discharges per year, I would estimate that it would be up to 3-4 times that of the numbers in this study.  The time needed to make all of these phone calls (some of which would undoubtedly require multiple attempts) could quickly become overwhelming on top of the existing demands on our time.  One potential way to get around this issue would be to try to identify groups of patients who are at higher risk for difficult transitions and focus on at-risk groups for follow-up phone calls.  Issues like this are also part of the reason why we aim to communicate with patients’ PCP after a hospitalization and have children follow-up in the office within a few days after discharge – it would be interesting to know what the typical follow-up plans are for patients from the study group as a follow-up phone call may be less helpful if the patient has already been seen their PCP where they theoretically could have these same concerns addressed.

Posted in Discharge planning, readmission to the hospital | Leave a Comment »

Pour Some Sugar on Me: Dextrose Bolus in Pediatric Gastroenteritis

Posted by ashriner on June 23, 2015

Sugar BearReviewers:  Bethany Beard, MD and Angela Dietrich-Kusch, MD


  1. Levy, JA, et al. “Intravenous Dextrose for Children With Gastroenteritis and Dehydration: A Double Blinded Randomized Controlled Trial.” Annals of Emer Med. Vol 61. No3. Mar 2013.  doi: 10.1016/j.annemergmed.2012.08.007 (Link)
  2. Levy JA, Bachur RG, et al. “Intravenous dextrose during outpatient rehydration in pediatric gastroenteritis.” Acad Emerg Med. 2007; 14:324-331.  doi: 10.1197/j.aem.2006.10.098 (Link)

Why this topic?  Imaging you are caring for a 3 year old female with abdominal cramping, diarrhea, and vomiting for the past 72h. She is refusing oral intake and has urinated only once in the past 24hr. On exam, she has dry oral mucosa, sunken dark eyes, and generally appears weak. You’ve learned in medical school and residency that you should only bolus with normal saline. However, you wonder if a bolus of dextrose could make the patient feel better and decrease the chance that mom will bring her back to your emergency department. However, you aren’t confident that dextrose containing fluid bolus is safe and effective when used as a bolus. You perform a literature search to answer your question.

  • P: pediatric patients with gastroenteritis
  • I: dextrose containing IV fluid bolus
  • C: normal saline bolus
  • O: decrease return ED visits/admissions

Literature search: Dr. Beard performed a Pubmed and Ovid search with the search terms ‘dehydration/glucose/or gastroenteritis’, ‘emergency services’, ‘intravenous or infusion’ and limited for children ages 0-18yr, humans, and English language. In total, 21 articles were found to be relatively pertinent to the question and two were selected to answer the specific question.


Article 1: “Intravenous Dextrose for Children With Gastroenteritis and Dehydration: A Double Blinded Randomized Controlled Trial”

Methods: This was a randomized controlled trial of children ages 6mo-6yrs with gastroenteritis at Boston Children’s Hospital. Subjects were randomized to receive a single 20ml/kg bolus of either 5% dextrose in normal saline or normal saline alone. Serum ketone levels and glucose were measured before bolus and again at 1 and 2hr intervals after the fluid bolus. Primary outcome was hospitalization rates. Secondary outcomes were change in serum ketone levels over time.

Results: 188 children were enrolled. The proportion of children hospitalized did not differ between groups (35% in D5 group, 44% in NS group) although there was a trend towards less admission in the dextrose group. Those children who received D5 had a greater reduction in mean serum ketone levels both at 1hr and 2hr. Of those children discharged, 30% of patients in the normal saline group and 11% of patients in the D5 group required unscheduled medical care after discharge home.

Conclusion: The study showed no significant difference in admission rates, however, there was a decrease in ketone levels in the D5 bolus group. Although there was a significant decrease in ketone levels, this did not seem to show any clinical difference in outcomes. Although the study outcome did not specifically study outpatient medical needs, it appears that the D5 group might have experienced less ongoing symptoms and need for medical care (i.e. PCP appointments) once released from the hospital however more research is needed. This may also translate into less repeat ED visits, however, this was not a study outcome in this paper.


Article 2: “Intravenous dextrose during outpatient rehydration in pediatric gastroenteritis”

Methods: This is a case control study of children aged 6mo to 6yr who presented to an urban ED with acute gastroenteritis and dehydration and who received IV rehydration before ED discharge. The study looked at whether the amount of IV dextrose administered at the initial visit was related to return visit admission (RVA).

Results: A total of 168 children diagnosed with gastro and discharged from the ED were studied (56 cases and 112 controls). Cases were defined as children who returned and were admitted within 72hr of initial ED visit and discharge. Patients who had a RVA received significantly less IV dextrose than those who did not have a RVA. Those who received no dextrose had a 3.9 greater odds of having a RVA. The study also found that those who received more dextrose containing IV fluids were 1.9 times less likely to have a RVA. Also, the study found that patients who had been ill for <24hr were more likely to have return visits than those ill for 1-2 days. No other demographic, clinical, or lab factors were associated with greater odds of having a return visit.

Conclusion: Larger amounts of IV dextrose is associated with reduced RVA in children with gastro and dehydration. Those that are sicker for shorter period of time may need to be more highly considered for admission.

Limitations: This was a retrospective study and the degree of symptoms could not be evaluated. It is therefore not known if patients who received more dextrose were sicker. Also, confounding variables influencing the primary outcome measure (return visit in 72hr) could not be controlled. Finally, the patient disposition regarding discharge or admission as well as the original diagnosis of dehydration was not standardized by the physician.


Take Home Points:

Physiologically, it makes sense that patients with dehydration from gastro will have metabolic acidosis with higher ketone levels which could contribute to symptom severity. However, both studies above shown that ketone levels do not contribute to re-admission rates. More research needs to be done to determine if ketone levels actually contribute to symptoms. A bolus with dextrose containing fluids appears to be safe and cheap and seems to show a decrease in re-admission rates as well as outpatient medical care needs. There also needs to be more research in the actual volume and dextrose dose, which has yet to be determined. In patients with very limited oral intake and signs of dehydration I will consider dextrose IV fluid boluses as part of the ED management plan based on the results of these two studies.

Additional Comments from Dr. Dietrich-Kusch:

From the inpatient side of things, I would really love to see bigger study on this. As a hospitalist admitting the children who still fail outpatient management of their gastroenteritis it seems like the use of D5 boluses lead to more labs because nearly always the patients end up with glucosuria and ketonuria on their UA and then people get concerned about DM, which leads to further labs (Hgb A1C, fasting glucose, repeat UA, etc). I personally wonder if there is a smaller amount of glucose that would be just as beneficial, but wouldn’t shoot up the blood sugar so much that they started spilling glucose into their urine.

Posted in diarrhea, fluid/electrolytes | Leave a Comment »

Prednisone vs. Dexamethasone for Treatment of Acute Asthma Exacerbation in Children

Posted by ashriner on May 20, 2015

“Hear me now and believe me later. Dexamethasone is as effective as prednisone for treatment of acute asthma exacerbations in children.” ~ Hans and Franz

Reviewer:  Andrew Shriner, MD

Article:  Meyer JS, Riese J, Biondi E.  “Is dexamethasone an effective alternative to oral prednisone in the treatment of pediatric asthma exacerbations?”  Hosp Pediatr.  2014 May;4(3):172-80. doi: 10.1542/hpeds.2013-0088.

Link:  click here

Why this article?  I’ve started noticing a trend in children being admitted for asthma exacerbations in which they are given either IM or PO dexamethasone as opposed to a dose of oral prednisolone which had been the norm that I was used to.  The reasons that I’d heard were that a 2-day course of dexamethasone had been shown to be at least as effective as a 5-day course of prednisone and that the hope was to improve compliance and side-effects with fewer doses.  I wanted to see for myself what the evidence showed and found an article in Hospital Pediatrics from last year that reviewed some of the recent pediatric literature.

Methods:  Authors performed a Medline search to find studies comparing the use of dexamethasone vs. prednisone in the treatment of asthma exacerbations.  Meta-analysis was performed to assess the rate of revisits to a healthcare provider and symptomatic improvement in the 2 groups after the initial ED visit.

Results:  Six articles were ultimately included in the review and broken down as follows:

  • IM Dexamethasone Studies
    • 3 studies.  Sample sizes of 42, 32 and 143.
    • all studies compared a single dose of IM dexamethasone with doses ranging from 0.3 mg/kg (max 15 mg) to 1.7 mg/kg (max 36 mg) vs. 3-5 days of prednisone with doses of 2 mg/kg/day (but max doses ranging from 50-100 mg/day)
    • Limitations:  lack of blinding, exclusion of patients who were “too sick”, small sample sizes, reliance on subjective reports of symptom improvement
    • No significant differences found between the groups.  Outcomes assessed included:
      • improvement of symptoms, clinical asthma scores through day 5, % returning to baseline by day 5, relapse rates at 28 days, daily albuterol use on day 5, admission rates, unplanned physician visits
  • PO Dexamethasone Studies
    • 3 studies.  Sample sizes of 533, 110, and 89.
    • one study compared one dose of oral dexamethasone to 5 days prednisone.  The other two compared two daily doses of oral dexamethasone to 5 days of prednisone.
      • dexamethasone doses were 0.6 mg/kg in all studies with a max of 16 mg in two and 18 mg in another.
      • prednisone doses were all a total of 2 mg/kg/day, two with a max of 60 mg/day and one with a max of 80 mg/day.
    • No significant differences found between the groups.  Outcomes assessed included:
      • Relapse rates, rates of admission to hospital, mean days to return to baseline, mean patient self-assessment scores at 5 days
  • Meta-analysis results
    • There was no significant difference between the dexamethasone vs. prednisone groups in regards to number of unscheduled physician visits after discharge or the time for symptom improvement back to baseline.
  • Adverse Effects
    • Mixed results with one study showing increased vomiting in the prednisone group while in the ED (but no difference in vomiting rate at home).  Another found no difference in rate of vomiting between groups.
  • Compliance
    • Difficult to tell due to design of some of the studies, but one study found that 17% of children in the study refused > 75% of their prednisone doses and 24% missed between 30-50% of their doses.

It seems like there is mounting evidence that there is likely no significant difference in clinical outcomes when comparing the use of dexamethasone to prednisolone when treating children for asthma exacerbations.  While there was a great deal of variation in the study designs, doses of meds and formulations used; there was never any statistical difference noted between treatments.  What is still unclear to me is what the ideal dexamethasone regimen would be based on these studies.  A dose of 0.6 mg/kg seems to be a good starting point, as we’re used to giving this dose to children with croup and the majority of studies used this as well.  (But, the range of max dosing was anywhere from 16 mg to 36 mg per dose!)  When it comes to dosing, the studies used either one-time dosing or daily x 2 days.  Given all of the variability in the studied regimens, it would be really nice to see a large, high-quality, randomized controlled trial to clarify the ideal treatment regimen and to help ensure that we’re not missing any important differences between treatment with dexamethasone vs. prednisone due to biases from the study design or due to small sample sizes before completely changing our routine management for these patients.

BUT, we’re already seeing some patients who are started on these treatment courses in the ED before they get to us, so we need to at least start thinking about how we are going to handle these situations going forward.  The current dosing schemes that have been in practice at Riley include either a one-time po dose in the ED (usually up to 12 mg max) and for those admitted, sometimes an additional dose 2-3 days later (not the next day as in the protocols in the studies reviewed here).  This makes more sense to me from a pharmacokinetic standpoint, but would likely reduce the positive impact of improved compliance with this regimen if patients needed to go home with a prescription.

IF we’re comfortable saying that 1-2 doses of dexamethasone are equally efficacious as 5-days of prednisolone, I do think there are compelling reasons to switch to dexamethasone, primarily the potential for improving compliance with the prescribed treatment regimen.  The authors also tout advantages of lower cost, fewer adverse effects and better palatability.

I question some of this, though.  While I’m not a connoisseur of medicines, I do have a trusted contact from pharmacy who tells me that liquid dexamethasone is NOT more palatable than liquid prednisolone as one of the studies quoted by the authors suggests.  I can’t find an obvious reason for the discrepancy when looking at the articles they cite, though there was a difference between males and females with males being more likely to prefer dexamethasone to prednisolone (though there was a preference towards dexamethasone regardless of sex).  If we wanted to decide for ourselves which is more palatable, we can make arrangements for a taste testing at one of our next staff meetings if there’s enough interest.

To wrap it up:  I think it’s an interesting shift in practice and one worth looking into, but certainly one with a lot of questions about how best to move forward.  What do you think?

Posted in asthma | Leave a Comment »

Etiology and Incidence of CAP in Hospitalized Children

Posted by ashriner on April 1, 2015

Dr Evil - pneumonia

Or for many children:  “viral lower respiratory tract infections”

Reviewer: Keith Chitty, MD

Article: Jain S et al.  “Community-Acquired Pneumonia Requiring Hospitalization in U.S. Children.” N Engl J Med 2015 Feb 26; 372:835.  DOI: 10.1056/NEJMoa1405870


Why this article? As hospitalists, we take care of patients with pneumonia on a routine basis. We know that both bacteria and viruses can cause pneumonia, but it is difficult to decide which patients not to treat with antibiotics, and which patients to treat empirically for atypicals. This multi-center study provides the newest data regarding the incidence and microbiologic causes of community-acquired pneumonias requiring hospitalization in children.

Methods: This is a prospective, multi-center, population-based, active-surveillance study conducted by the CDC and entitled the Etiology of Pneumonia in the Community (EPIC) study. It took place from 1/1/10-6/30/12 in children’s hospitals in Memphis, Nashville, and Salt Lake City. Children were included in the study if they had evidence of acute infection (fever, chills, hypothermia, leukocytosis or leukopenia), evidence of an acute respiratory illness (new cough, sputum production, chest pain, dyspnea, tachypnea, abnormal lung exam, or respiratory failure), and evidence consistent with pneumonia as assessed by chest radiography within 72 hours before or after admission.

Exclusion criteria were:

  • Recent hospitalization
  • Already enrolled in EPIC study within previous 28 days
  • Resided in an extended-care facility
  • Had an alternative diagnosis of a respiratory disorder
  • Were a newborn who never left the hospital
  • Had a tracheostomy tube, cystic fibrosis or cancer with neutropenia
  • Had received a transplant within the previous 90 days
  • Had active Graft Versus Host Disease or bronchiolitis obliterans
  • Had HIV infection with CD4 count < 200

Laboratory testing included Gram’s stain and bacterial culture from blood samples, pleural-fluid samples, endotracheal aspirates, and bronchoalveolar lavage specimens. PCR for S. pneumoniae and S. pyogenes were performed on whole blood and pleural fluid. Pleural fluid PCR was also done for H. influenzae, S. aureus, S. angiosus and S. mitus. PCR was performed on nasopharyngeal and oropharyngeal swabs for adenovirus, C. pneumoniae, coronaviruses, HMPV, rhinovirus, influenza A and B, M. pneumonia, parainfluenza (1, 2, and 3), and RSV. Serologic testing for adenovirus, HMPV, influenza A and B, parainfluenza, and RSV was performed on available paired acute-phase and convalescent-phase serum specimens.

Results: 2358 children were enrolled in the study and had radiographic evidence of pneumonia. The median age of the children was 2 years. 45% were girls. 40% were white, 33% were black, 19% were Hispanic, and 8% were of another ethnic group. 51% of the children had an underlying condition, with asthma or RAD the most common condition. The median length of stay in the hospital was 3 days.  Overall, 497 children (21%) required ICU care, and 3 (<1%) died. Thirty percent of children 6 months and older had received one or more doses of influenza vaccine for the concurrent season. Eighty-seven percent of children 19 months to 12 years of age had received 3 or more doses of Prevnar. A total of 88% of children received antibiotics during hospitalization.

A nasopharyngeal or oropharyngeal swab was obtained from 96% of the children, blood culture from 90%, whole blood for PCR assays from 87%, paired serum specimens from 44%, pleural fluid from 4%, BAL specimen from 1%, and endotracheal aspirate from 1%. For calculation of proportions of specific pathogens, data were included from only the 2222 (94%) children with radiographic pneumonia who had blood, pleural fluid, endotracheal aspirate, or a BAL specimen available and who also had a nasopharyngeal or oropharyngeal swab or paired serum specimens available.

A pathogen was detected in 81% of children in the study – viruses in 66%, bacteria in 8%, and both bacterial and viral pathogens in 7%. The most commonly detected pathogens were:

  • RSV (28%)
  • Rhinovirus (27%)
  • HMPV (13%)
  • Adenovirus (11%)
  • S. pneumoniae (8%)
  • Parainfluenza (7%)
  • Coronavirus (5%)
  • M. pneumoniae (4%)
  • S. aureus (1%)
  • S. pyogenes (1%)

RSV (37% vs. 8%), adenovirus (15% vs. 3%) and HMPV (15% vs. 8%) were detected more commonly in children younger than 5 years of age than in older children. M. pneumoniae was detected more commonly in children 5 years of age or older than in younger children (19% vs. 3%). Pneumonia peaked in the fall and winter.

A control group was also tested and all pathogens with the exception of rhinovirus were detected in 3% or less of controls. When adjusted for age, rhinovirus was detected in 17% of controls compared with 22% of children with pneumonia.

The annual incidence of hospitalization for pneumonia was 15.7 cases per 10,000 children. Incidence of hospitalization for children younger than 2 years was 62.2 cases per 10,000 children; for 2-4 years of age it was 23.8 cases per 10,000 children; incidence continued to decrease with increasing age.

Discussion: This study reinforced some established knowledge regarding pneumonia; I am also taking away several learning points:

We’ve been taught since medical school that viruses are a common culprit in pneumonia, but I have a new appreciation for that fact after looking at this study. The authors are quick to concede that bacterial culture-based diagnostic tests have limited sensitivity and that bacteremia is detected in a minority of pneumococcal pneumonias. However, this study used state-of-the-art diagnostic testing and only detected bacteria in 15% of cases. Does this mean that we only need to use antibiotics for 15% of our inpatients with pneumonia? Probably not, but especially in the subset of young patients with clinical bronchiolitis and an infiltrate on CXR, I am becoming less excited about treating with antibiotics.

Mycoplasma was detected twice as often as Strep pneumonia overall. However, in children 5 years of age and older Mycoplasma’s detection rate far exceeded that of Streptococcus. I have always considered atypical coverage strongly in older patients, but going forward I think my practice will change to empiric double coverage in that age group.

Lastly, I am interested in the incidence of rhinovirus positive controls. As PCR-based viral testing becomes more commonplace we will see more positive results for rhinovirus/enterovirus. I’ll keep this data in mind; in patients whose hospital courses aren’t typical, these positive results could be a red herring.

Posted in pneumonia | 2 Comments »

Rethinking the rebound bilirubin

Posted by ashriner on March 16, 2015

reboundsReviewer:  Andrew Shriner, MD

Article:  Berkwitt A, Osborn R, Grossman M.  “The utility of inpatient rebound bilirubin levels in infants readmitted after birth hospitalization for hyperbilirubinemia.”  Hosp Pediatr.  2015 Feb;5(2):74-8. doi: 10.1542/hpeds.2014-0074.

Link:  click here

Why this article?  Its been a while since the release of the 2004 AAP guidelines for neonatal hyperbilirubinemia, but one thing that seems to have not been widely adopted is the idea that infants admitted for phototherapy do not need to stay in the hospital after phototherapy is discontinued to check a “rebound” bilirubin level.  I thought this was a nice opportunity for us as a group to stop and ask ourselves a couple of questions:  “Do I routinely order rebound bili levels before discharge after discontinuing phototherapy?”  And, if so, “Why?”

Methods:  This was a retrospective cohort study conducted on infants admitted to 1 tertiary care children’s hospital for hyperbilirubinemia between 1/2007 and 4/2014.  Patients were included in the study if they were > 35 weeks GA and were re-admitted after discharge from newborn nursery for phototherapy to treat indirect hyperbilirubinemia

Exclusion criteria included:

  • history of phototherapy in first 24 hrs of life
  • ICU admission
  • exchange transfusion
  • direct hyperbilirubinemia
  • evidence of sepsis

Children were divided into two groups, those who had repeat bilirubin levels drawn < 12 hours after cessation of phototherapy and those who did not have repeat levels drawn.

Results: 226 infants were included in the final analysis, 130 in the rebound group and 96 in the no-rebound group.  Bilirubin levels were higher in the rebound group at initiation  (19.5 vs. 18.8) as well as at cessation (13.7 vs. 13.0) of phototherapy.  The rebound group also had a significantly shorter duration of phototherapy (15.1 h vs. 17.7 h).  There were 9 re-admissions within 72 hrs from the cohort (5 from the rebound group and 4 from the no-rebound group).  Mean LOS was increased in the rebound group (about 4 hrs longer) as was the mean discharge time.  In the rebound group, there was not a significant difference in mean bilirubin levels at phototherapy cessation and mean rebound bilirubin levels.  Five infants who did have rebound bilirubin levels checked received repeat phototherapy during the same admission – all having had bilirubin levels > 14 at the time of initial phototherapy cessation.  (1 infant had cephalohematoma and 2 had ABO incompatibility)

Infants with bilirubin levels < 14 at the time of phototherapy cessation were significantly less likely to need repeat phototherapy (1.6% vs. 12.4%).

Conclusions: This study was definitely not the most robust in terms of numbers of infants included and is limited by its retrospective design, but I still think there’s a lot that can be taken away from this.  The results were in line with previous studies looking at early rebound bilirubin levels and finding no significant increase after phototherapy cessation.  Re-admissions rates were essentially the same, but LOS was lower and discharge time was earlier in the group not subjected to rebound bilirubin checks prior to discharge.  It was interesting that the group of infants who had rebound levels checked were more likely to have higher bilirubin levels at initiation of phototherapy and at cessation, but received significantly shorter durations of phototherapy (and, by the way, also spent more time in the hospital as well).  I’m inclined to agree with the authors conclusions that our aim should be to follow the AAP recommendations and have a goal of bilirubin levels of 14 or less prior to stopping phototherapy.  Also, this paper is just further evidence that rebound bilirubin levels are probably not going to offer much clinical utility and are not necessary in the majority of patients.  I do still think there may be utility in checking rebound bilirubin levels in a subset of infants who have either been slower to improve on phototherapy than we would expect or have underlying conditions which may put them at higher risk for worsening hyperbilirubinemia after stopping phototherapy, such as significant bruising, cephalohematoma or hemolytic disease.  But, these infants tend to be more the exception than the rule.

Posted in Hyperbilirubinemia | Leave a Comment »

Newborn Weight Loss Nomograms

Posted by ashriner on February 24, 2015

Who doesn't love a good nomogram?

Who doesn’t love a good nomogram?

Reviewer: Melissa Klitzman, MD

Article: Flaherman, VJ., et al. “Early Weight Loss Nomograms for Exclusively Breastfed Newborns.” Pediatrics. 2015. January 1; DOI:10.1542/peds.2014-1532

Link:  click here

Why This Article?: Exclusively breastfed infants lose weight daily while mother’s mature milk is coming in, more so than those receiving any amount of formula. This article studies the weight loss a large cohort of exclusively breastfed infants and presents two nomograms based on delivery modality similar to the Bhutani nomogram for hyperbilirubinemia. This could be a potentially useful tool in identifying those newborns at greater risk for weight loss and therefore earlier intervention.

Methods: This is a retrospective, multicenter cohort study evaluating the daily weight loss of exclusively breastfed newborns greater than or equal to 36 weeks during their initial newborn hospitalization between January 2009 and December 2013. Newborns were in 1 of 14 Kaiser Permanente Northern California hospitals.

Exclusion criteria:

Level II or III NICU care, < 36 week gestation, birth weight <2000g or >5000g, multiple birth, implausible weight loss (>10% in first 24 hours or >15% after first 24 hrs), implausible weight gain (>5%), missing/discrepant data (delivery modality, weight, feeding), and infants were censored if they started formula feeding.

Newborns were included in final analysis if any weight was obtained after 6 hours of age and before exclusive breastfeeding was discontinued. Nomograms stratifying weight percentiles (50th, 75th, 90th and 95th) for weight loss for hours of life were produced based on delivery modality. Vaginal delivery nomogram is between 6-72 hours of life and 6-96 hours of life for the cesarean nomogram.

Additional demographic information was obtained (i.e. maternal race/ethnicity).

Results: In the final analysis, there were 108,907 newborn with 76.6% delivered vaginally and 23.4% delivered by cesarean. There were ~ 1.3 weights recorded per newborn for vaginal deliveries and 2.1 weight per newborn recorded Forty-three percent of the mothers were white non-Hispanic, 24% Hispanic, 24% Asian, 7% black non-Hispanic, and 3% unknown.

The analysis also included examination of the nomograms to see if it was affected by inclusion of late preterm, post term, SGA and LGA. Nomograms were produced by removing those infants and comparing the nomograms (see supplemental Figure 5).

Imputed nomograms were generated (see supplemental figure 6) for those infants censored (removed) from the study due to initiation of formula because of large amounts of weight loss. Weights were included for these newborns up until formula feeding was initiated but then data from an uncensored matched newborn was used. The assumption is that the weight loss history of the matched uncensored newborn approximates the weight loss for the censored newborn.

Discussion: The weight percentile loss nomograms for exclusively breastfed infants will be useful in the initial newborn hospitalization period similar to how the Bhutani nomogram is useful for hyperbilirubinemia. They can help assess the need for more lactation support, supplementation, timing of discharge, and timing of follow-up. There is also a handy-dandy website ( to plug in values similar to bilitool for hyperbilirubinemia.

What I find particularly useful is the stratification of weight loss percentiles based on delivery modality. For instance, a 5% weight loss at 24 hours puts that newborn in the ~75% in the vaginal delivery nomogram, whereas the same % weight loss in an infant born via cesarean would only put that infant in the ~50%. That being said 10% weight loss still seems to be the magic number by 48 hours of life for both vaginal (95%) and cesarean (90%) deliveries.

The imputed nomograms (see supplemental figure 6) overlying the original curves are essentially unchanged for vaginal and cesarean deliveries until ~72 hours of life. Also the nomograms generated excluding late preterm, post term, LGA and SGA (see supplemental Figure 5) are essentially unchanged for both delivery modalities. This lends to reliability of the nomograms.

Approximately 87% of the collected weight data are within the first 36 hours of life for vaginal deliveries and ~90% within the first 60 hours for cesarean. Passed these hours of life the data is less reliable as reflected in the 95% confidence interval nomograms (see supplemental figure 3).

This study also has a smaller white, non-Hispanic infant population and a large non-white infant population compared to the latest US Census data. Therefore the nomogram might be less reliable in population with less racial and ethnic diversities.

While I might not use the nomogram for every infant I see in the well newborn nursery, I have and will continue to use it for infants whose percent weight loss are increased.

Posted in Breastfeeding, neonate | Leave a Comment »

Tamiflu, Tamiflu…What do you do?

Posted by ashriner on February 12, 2015

Would you use it here or there? Would you use it anywhere?

Reviewer: Sean Thompson, MD

Article: Jefferson T, Jones MA, Doshi P, Del Mar CB, Hama R, Thompson MJ, Spencer EA, Onakpoya IJ, Mahtani KR, Nunan D, Howick J, Heneghan CJ. Neuraminidase inhibitors for preventing and treating influenza in adults and children. Cochrane Database of Systematic Reviews 2014, Issue 4. Art. No.: CD008965. DOI: 10.1002/14651858.CD008965.pub4.

Link to Article: Click here 

Why this article?  I thought I would pick something that has had a little controversy associated with it for this month’s article. There is a lot of back-story to this, but briefly, there was significant controversy about oseltamivir and zanamivir that was partly responsible for public outcry for Roche and Glaskow-Smith-Kline (GSK) to release all their previously unpublished data. Prior to this, only select data were released which raised the concern that only positive, industry sponsored studies were being made known. Cochrane reviews prior to 2009 reported that oseltamivir prevented lower respiratory tract complications related to influenza, which led to global stockpiling of neuraminidase inhibitors. There were concerns about the reliability of unpublished data and the author of the Cochrane review was unsuccessful in obtaining data from Roche and so Cochrane reviews prior to 2009 were redacted. After a public campaign that was partly spearheaded by the BMJ, Roche and GSK released previously unpublished data and the Cochrane Collaboration published a new review in April 2014 examining this newly released data.

Objective:  “To describe the potential benefits and harms of NIs for influenza in all age groups by reviewing all clinical study reports of published and unpublished randomized, placebo-controlled trials and regulatory comments.”

Methods:  The authors did a systematic review of all randomized, placebo-controlled trials on adults and children with confirmed or suspected influenza. They focused on newly released clinical study reports from manufacturers, but also did an extensive search for any published RCTs. The data was analyzed for time to first alleviation of symptoms, influenza outcomes, hospitalizations, and adverse events.

 Results:  The authors obtained 107 clinical study reports (160,000 pages). They analyzed 46 stage 2 trials, 20 on oseltamivir (9623 participants) and 26 on zanamivir (14,628 participants). The authors report inadequate reporting for most studies, which places them at high risk of selection bias.

  • Time to first alleviation of symptoms: Oseltamivir reduced symptom duration from 7 days to 6.3 days in adults (95% CI: 8.4 to 25.1 hours, P < 0.0001). In otherwise healthy children, there was a reduction in length of symptoms of 29 hours (95% CI: 12 to 47 hours, P < 0.001). There was no effect found in children with asthma.
  • Hospitalizations: There was no significant effect on hospitalizations in adults with RD = 0.15% (95% CI: -0.78 to 0.91). There was no significant effect in children.
  • Serious Influenza complications: 
    • Pneumonia – There was a significant reduction in pneumonia. The diagnosis was “self-reported, investigator-mediated, and unverified” (RD 1.00%, 95% CI 0.22 – 1.49). Number needed to treat of 100. When the authors looked at radiologically confirmed pneumonia, there was no significant difference. There was no difference in children.
    • Bronchitis, sinusitis, otitis media – There was a significant reduction in bronchitis in adult treatment trials with zanamivir. There were no differences with oseltamivir and no difference with either drug for sinusitis or otitis media in children or adults.
  • Harms: There was increased risk of nausea (RD 3.66%, 95%CI 0.9 – 7.39) and vomiting (RD 4.56%, CI 2.39 – 7.58). The number needed to treat to harm (NNTH) for vomiting was 22. Specifically in children, the main risk was vomiting with a NNTH of 19. Interestingly, there was also a significant lower proportion of children who had a four-fold increase in antibodies after receiving oseltamivir (RR 0.9, 95% CI 0.8 – 1.0).
  • Prophylaxis: Both oseltamivir and zanamivir reduced symptomatic influenza when used as prophylaxis (3.05% for oseltamivir, NNT 33; 1.98% for zanamivir, NNT 51).

Limitations: The study was limited by large amount of variability between trials and what the authors thought was significant amount of bias in the data. There was also a variable degree of participant self-reporting of complications. A further limitation of this type of systematic review is that it only includes RCTs. Though this is considered the highest level of evidence, there may be important evidence that is not included because the Cochrane collaboration only considers RCTs.

Discussion: This review received a lot of media attention and has caused a large amount of debate over the author’s conclusions. The authors conclude that neuraminidase inhibitors have small, and non-specific effects on reducing time to alleviation of symptoms, but there is not good evidence that they reduce any complications of influenza that we really care about it, particularly hospitalization and secondary pneumonia. This was the first time that all of the manufacturer’s data was analyzed and seems to reinforce the view of the Cochrane authors that the benefit of NIs is minimal. It also presents the potential for harm that NIs pose. The Cochrane group and the BMJ question the stockpiling of NIs that has taken place based on this data. Overall, this review seems pretty convincing to me that NIs have been oversold and it causes me to question the CDC’s recommendations. This is where the controversy begins, however. At almost the same time that the Cochrane review was published, a large meta-analysis of observational trials was published in the Lancet, which included data on more than 29,000 patients who were admitted with H1N1 during the 2009-2010 pandemic. Check it out here:

The above meta-analysis reported a decreased mortality attributable to NIs in adults. There was no significant survival benefit found in children.

Critics of the Cochrane review argue that observational studies provide critical information on how patients respond to drugs in the real world. The Cochrane group would argue that observational studies have a high risk of bias. This is true, however most of the data included in the Cochrane review was based on healthy subjects, who in general, do not suffer the serious side effects of influenza. To this point, there are no RCTs that examine the effect of NIs in hospitalized patients with severe disease. Other high-risk groups such as pregnant women and those with significant comorbidities are also not included in RCTs as it would likely be considered unethical at this point to withhold antivirals in these patients.

It can be argued that we have a large amount of data that shows that NIs are effective in reducing duration of illness and mortality in hospitalized patients with severe disease. The Cochrane group and the BMJ have been accused of unfairly casting doubt on NIs because of their conclusions based on RCTs and could jeopardize treatment of patients in a true pandemic when severity of illness and rate of complications are far different from what is reflected in their review.

Your own conclusion on this topic may depend upon how you interpret the data, your own bias on the topic, and what population of patients you treat. As usual, the most convincing data is in adults and the data for benefit in the pediatric world is much weaker. It is hard to argue against the observational data, however, and the potential benefit of NIs in hospitalized patients with severe disease. Those of us who treat outpatients, who are otherwise well, and who have mild disease, will have a harder time arguing that providing oseltamivir will prevent complications and the mild benefit of decreased duration of symptoms needs to be weighed against the increased risk of vomiting. Though the clinical significance is unclear to me, it is also interesting how those treated with NIs had a clinically significant antibody response.


This Cochrane review found that neuraminidase inhibitors decreased the duration of symptoms but did not affect number of hospitalizations or complications from influenza and influenza-like illness. They argue that based on this data, global stockpiling of NIs should be questioned. The results have been strongly debated as observation data on the benefits of NIs have been reported. Despite the results, the review highlights the importance of drug companies and researchers releasing all trial results for analysis. Furthermore, there is a need for further research on neuraminidase inhibitors and effective treatment in the event of a serious pandemic.

Posted in Influenza, pneumonia, Treatment Guidelines | Leave a Comment »

Urine Trouble: How significant is bacteremia in children with a UTI?

Posted by ashriner on January 30, 2015

Urine TroubleReviewer:  Andrew Shriner, MD

Article:  Roman H.K, Chang PW, Schroeder AR.  “Diagnosis and management of bacteremic urinary tract infection in infants.”  Hosp Pediatr.  2015 Jan;5(1):1-8.  doi: 10.1542/hpeds.2014-0051.

Link:  Click here

Why this article?:  UTI is a common diagnosis that we encounter, however, despite helpful clinical practice guidelines, there are still many questions and ongoing debates surrounding appropriate management for these children.  One area of uncertainty is in regards to when blood cultures should be obtained for infants with UTIs and how to appropriately manage them if they are found to be bacteremic.  This article reports on the prevalence of bacteremia in children with UTIs, compares characteristics of children with UTIs and bacteremia to those without bacteremia and also looks at treatment courses and outcomes for children with UTIs and bacteremia.

Methods: This study used a retrospective, cross-sectional design to determine prevalence of bacteremia in infants with UTI and a retrospective double cohort design to compare clinical characteristic and clinical outcomes between infants with UTI (with bacteremia vs. without bacteremia).  Data was gathered from one hospital system in Northern California from 1998-2012 for patients <1 year of age with UTI, though infants in the NICU were excluded given concern for generalizability with the general population. 

Results: A total of 1379 UTIs were identified with blood cultures being obtained in 52% (n=714) of cases.  Fifty-five percent of UTIs were in boys and blood cultures were also more commonly ordered in male infants.  The overall percentage of infants with UTI having a blood culture obtained decreased over the study period by an average of 11%.  Of the blood cultures obtained, 101 (14%) had bacterial growth, 41 of which were due to contaminants.  The 57 remaining infants with bacteremia and UTI represented 4.1% of all infants with UTI and 8% of infants with UTI and had a blood culture obtained.  Prevalence of UTI was highest in infants <1 month of age and decreased with age, but did not differ significantly between males and females.  E. Coli was the most common organism identified. 

Characteristics of infants at presentation with UTI (with and without bacteremia) were very similar with only the following items being significantly different between the two groups: 

  • Bacteremic infants had a higher percentage of band count on their WBC differential 
  • Bacteremic infants more often had urinalysis with bacteria present and > 10 WBCs/hpf. 

Treatment and hospital length of stay (LOS) varied significantly in infants with bacteremic UTI with LOS primarily being driven by duration of parenteral antibiotic course.  Infants treated with 7 days or more of IV antibiotics were younger than those who received shorter courses, but had no other significant differences between the groups.  Treatment varied greatly, even for infants < 2 months of age with 25% receiving ≤ 4 days of IV antibiotics, 46% receiving 5-7 days of IV antibiotics, and 29% receiving > 7 days of IV antibiotics.  Most infants (80%) with bacteremia had fever durations < 24 hrs.  No infants required care in the intensive care unit.  Of infants with bacteremia, 51% also had CSF cultures sent and only one of these returned positive (broth only) and was suspected to be related to a traumatic LP with likely contamination of the CSF culture from the blood present in the sample.  No infants with bacteremia had recurrent UTI or bacteremia with the same bacteria within 30 days of the initial illness, though 2 infants (both with VUR) did develop recurrent UTI with a different organism.  Of infants with UTI and no bacteremia, 4.6% had a recurrent UTI within 30 days, 80% of which were with the same organism.   


 Though the numbers in this study are not large and it is limited to only one hospital system in the US, I think it does offer some valuable information in regards to the prevalence of bacteremia in infants with UTI and the variability with which they are being managed.  The overall prevalence of bacteremic UTIs in this study was about 8%, which is suggestive that this is not an incredibly rare entity.  However, since bacteremia in these infants is hard to identify based on clinical characteristics alone, it is hard to know when blood cultures should be sent to evaluate for this possibility.  Given this uncertainty, it is reassuring to see that while treatment regimens were highly variable, all of the infants in this study had excellent outcomes with low risk for recurrent illness regardless of whether they had long or short courses of IV antibiotics.  Fevers improved quite rapidly in the majority of infants in this study and bacteremia cleared quickly as well in infants who had blood cultures repeated after appropriate treatment was initiated.  While we need to be careful with the conclusions we draw from a retrospective study with relatively low patient numbers, it is encouraging to see data supporting a short course of IV antibiotics which can potentially reduce patient’s LOS without compromising clinical outcomes (especially since my typical practice has been to transition these patients to oral antibiotics after clinical improvement and defervesence!).  Additionally, it raises the question that if bacteremic infants with UTIs can be treated similarly to those without bacteremia, is there any utility to ordering a blood culture in the first place?  And, is there a potential for more harm than benefit  associated with this practice? (i.e. complications that can arise from overdiagnosis of bacteremia, false positives associated with contaminated specimen, and harm associated with prolonged hospitalizations and long courses of IV antibiotics).  There are, however, some studies indicating that infants who have UTIs with bacteremia are at higher risk of having high-grade reflux or urinary obstruction and the UK guidelines actually recommend VCUG for these infants.  Knowing this would make me more likely to order renal bladder ultrasounds while these children are admitted, but it seems that if it was normal, VCUG could still be avoided.

Posted in bacteremia, urinary tract infection | Leave a Comment »


Get every new post delivered to your Inbox.

Join 26 other followers