Bacteraemia and occult bloodstream infection — branching viva

Viva format

The examiner takes the candidate through four escalating branches, each anchored to a patient. At each stage the candidate must state the reasoning before the action, name the active guideline or rule, and respect regional boundaries. The examiner probes the why and the corners. [10]

Branch 1 — The well-appearing vaccinated toddler

Examiner: A 14-month-old, fully immunised child, two days of fever to 39.5 °C, no focus, looks well. Would you draw a blood culture and give ceftriaxone? [1]

The model answer reframes the question before answering it. Conjugate vaccines collapsed the prevalence of occult bacteraemia in well-appearing vaccinated young children to well below one per cent, so routine blood culture and expectant empirical ceftriaxone — the 1990s practice — are no longer justified for most such children. The candidate assesses appearance explicitly, examines for a focus, considers a urine test, and reserves blood culture and inflammatory markers for higher-risk features. A specific safety-net with a named owner for any pending test replaces empirical antibiotics. [1] [3] [9]

Examiner probe: "Why did the practice change, and what evidence underpins it?" [1]

The candidate explains that Hib and pneumococcal conjugate vaccines produced large sustained falls in vaccine-serotype invasive pneumococcal disease in young children through both direct and indirect protection, which collapsed occult bacteraemia and reframed the clinical question to risk of invasive bacterial infection. [9]

Branch 2 — The febrile young infant

Examiner: Now a 28-day-old with fever 38.6 °C who looks well, and the RSV panel is positive. Does the positive virus change your plan? [1]

The model answer refuses to be reassured by the well appearance or the positive virus. This 28-day-old sits in the actively managed age band of the American Academy of Pediatrics guideline, between the always-investigate 8-to-21-day group and the PECARN-eligible 29-to-60-day group; she is risk-stratified with inflammatory markers and urinalysis, often with lumbar puncture and empirical therapy. A documented virus lowers but does not abolish bacterial coinfection risk in febrile infants 60 days and younger; appearance, age and markers still drive the decision. An adequate-volume blood culture, urine and bloods are obtained, and the infant is admitted and treated. [2] [3]

Examiner probe: "Which inflammatory marker performs best, and what are its limits?" [2]

The candidate names procalcitonin as the most accurate individual marker for invasive and serious bacterial infection in young febrile infants, ahead of C-reactive protein and white-cell count, and states that no single marker is sensitive or specific enough to rule disease in or out alone — markers work as part of validated rules, not as stand-alone tests. [2]

Branch 3 — The toxic older child

Examiner: Now an 8-year-old who is toxic, tachycardic, mottled and confused. How do you sequence resuscitation, cultures and antibiotics? [12]

The model answer treats this as sepsis first. The Surviving Sepsis Campaign 2026 paediatric first-hour bundle structures the response: recognise and escalate, support airway and breathing, obtain access and draw cultures, give empirical broad-spectrum antibiotics within the first hour, and use aliquot-based fluids with vasoactive support for refractory shock. The candidate draws the blood culture as access is established when this does not delay antibiotics, and does not wait for the work-up to finish before treating. Fluid is given in aliquots with reassessment, not as a target volume. [12]

Examiner probe: "The family gave paracetamol and an oral antibiotic before arrival, then you do a lumbar puncture and the culture is negative. What do you conclude?" [6]

The candidate states that antibiotics sterilise the cerebrospinal fluid within hours, so a negative culture after pretreatment does not exclude bacterial meningitis; decisions rest on the cell count and biochemistry, the blood culture, the clinical course and bacterial PCR where available, and a delayed lumbar puncture is performed once the child is stable. [6]

Branch 4 — The positive culture at 24 hours

Examiner: The well-appearing toddler you sent home has Streptococcus pneumoniae in both bottles at 24 hours. Walk me through your response. [1]

The model answer recalls and reassesses the child, reinterpreting the organism as a true pathogen rather than a contaminant because both bottles grew a recognised pathogen. The child is admitted, examined for a focus including meningism, repeat cultures and inflammatory markers are obtained, and empirical intravenous therapy is started and then narrowed once sensitivities return. The candidate emphasises that a pending culture must always have a named owner so that this call happens within a safe interval, and that the safety-net given at discharge is what made the positive culture actionable rather than dangerous. [1]

Examiner probe: "How would your interpretation differ if the result were coagulase-negative staphylococci in one bottle?" [1]

The candidate states that this is most likely a contaminant, read through the organism, the single-bottle growth and a well child; the response is clinical reassessment rather than reflex antibiotics or line removal, and the culture is repeated only if concern persists. The presence of a central line or immunocompromise would change that interpretation, which is why organism, bottle count and host context are read together. [1] [10]

Closing synthesis

A passing candidate keeps four ideas in tension throughout: conjugate vaccines reframed the question from chasing occult bacteraemia to risk-stratifying for invasive bacterial infection; age is the master variable and the febrile young infant still carries the residual risk; the toxic child is resuscitated first with antibiotics within the hour; and every positive or pending culture is interpreted through the organism, the bottle count and the host, with a named owner and a specific safety-net before any child leaves care. [1] [2] [3] [12]