Time for paediatrics to screen out sepsis “screening”

Despite the success of vaccinations against organisms such as Haemophilus influenza type B and Neisseria meningitidis,1 mortality and morbidity from these invasive infections remains high. One in five deaths in children were associated with infection pre-pandemic.2 This includes neonates who are particularly vulnerable due to immature immune systems and, in whom worldwide, infection rates have increased over the last 20 years.3 The sequelae of infection can cause lasting morbidity, through limb amputation, chronic respiratory disease and organ impairment, as well as death. As such, an aggressive approach is often taken to managing infection in children, particularly in the pre-school period, including a low threshold for numerous investigations and empirical antibiotics.

The practice of investigating febrile children has become so common, its own terminology, the “sepsis screen” has developed within the paediatric vernacular. The origin of the “sepsis screen” (which incorporates blood inflammatory markers, blood culture, urine and cerebrospinal fluid for microscopy, culture and sensitivity) is not clear. The earliest it appears in the literature is in regards to neonates deemed at high risk of sepsis.4

In the last decade, another shorthand description, “sepsis screening” arose from the process whereby patients were quickly identified as being at risk of sepsis. Sepsis screening is a concept in both adult and paediatric practice whereby physiological measurements are recorded and if certain thresholds are met the patient is deemed in need of an urgent clinical review. The development of sepsis red flags, notably from the National Institute for Health and Care Excellence,5 provided the thresholds for healthcare providers to use. The delivery of sepsis screening was given national priority by the creation of financial targets for their implementation.6

However, in our view, both “sepsis screen” and “sepsis screening” are inaccurate and unhelpful terms. Neither are, and never could be, a test for “sepsis.” Paediatric sepsis is inherently difficult to define, however, the most recent sepsis definition for adults is, “a life-threatening organ dysfunction caused by a dysregulated host response to infection.”7 The critical part of sepsis is that the life-threatening nature of the condition is due to the body’s response to the infection, not the infection itself. A bloodstream infection is a serious condition; however, it is not synonymous with sepsis. Although it is common to have sepsis and a bloodstream infection, it is also possible to have sepsis without a bloodstream infection in children (indeed the majority of cases of sepsis may have no positive culture8), and vice versa. The combination of a blood, urine, and cerebrospinal fluid culture cannot diagnose sepsis, merely confirm or deny the presence of a bacterial infection at these sites.

In addition, “screening,” by definition, requires the patient to be asymptomatic. However, many patients receiving a “sepsis screen” are symptomatic e.g. children presenting with a fever and coryza. This renders the term “screening” inaccurate and problematic for a number of reasons. It leads to confusion and misconception about the important differences between a blood stream infection and sepsis by both health professionals and the families of children. Many families are now familiar with the term sepsis following widespread awareness campaigns and tragic cases where the severity of a child’s clinical condition has not been appropriately recognised. This awareness is useful for families to be able to overcome an often challenging hierarchical gradient if a professional dismisses their concerns,9 but can cause inappropriate reassurance if the screening ‘test’ is determined to be negative or unnecessary anxiety when poorly specific vital signs highlight sepsis is present when it isn’t.10

The term sepsis screen triggers a bundle of observations or investigations to be carried out, often without due thought into their purpose. The outcomes of a sepsis screen may result in harm due to the risk of false positive tests from blood cultures11 and urine cultures12 resulting in longer patient stay, further investigations which are often painful for the child13 and risk of iatrogenic harm.14

Campbell and colleagues15 have highlighted when considering whether to perform any test one needs to:

1. Decide what diagnosis they are investigating

2. Determine the pre-test probability of the condition in question

3. Decide whether they are looking to rule in or rule out a diagnosis

4. Decide what they will do if the test is either positive or negative, and importantly

5. Consider whether performing the test could harm the patient.

Outside of the neonatal period where the decision to treat for invasive infections can be undertaken on the basis of risk factors alone, the application of Campbell’s rules would suggest “Sepsis screening” is not useful for diagnosis of sepsis, which is a clinical syndrome. In order to help us detect those febrile children who may have, or be showing early signs of sepsis, we need to develop a system of identifying children who are at higher risk of sepsis. This could be by either earlier identifying validated risk factors (e.g. very young age and comorbidities) such as suggested by the Kaiser Permanente calculator,16 or undertaking interventions in children showing overt signs of sepsis, or who are at high enough risk that further investigations or interventions are required. It is notable that the current NICE approach has not been validated to be effective in children17 and sepsis risk thresholds based on presenting criteria have yet to be established.

Ultimately for children with potential sepsis there is no screening involved, as all patients come to us because of signs or symptoms. We must then make clinical judgements based on the individual patient scenario, and perform targeted testing which delivers high value care. This approach is supported by the recently published Academic of Medical Royal Colleges sepsis guidance18 which has moved away from screening to a clinical decision support framework with the aim to balance risk (based on a composite of clinical features) with response (urgency of clinical review and intervention). This recognition was also incorporated into the new Surviving Sepsis guidelines for children, which have expanded the window to initiate antimicrobials to within three hours in the absence of shock to allow for more judicious use of investigations and treatment.19

In our view, it is time to remove the term, “sepsis screen,” from the paediatric lexicon. Instead we need to focus on identifying patients with infection at high risk of sepsis, and to provide focused, patient centred management based on individual risk profiles.