Statistics from Altmetric.com
Editor,—Pollard et algive an excellent review of the emergency management of meningococcal disease.1 We strongly support the involvement of the paediatric intensive care unit (PICU) at an early stage; however, only a third of all admissions with meningococcal disease are admitted to a PICU. Deciding which children need PICU admission can be a major clinical challenge. It would be inappropriate to transfer all children—they would occupy beds better used for other patients. Although the algorithm in fig 1 of Pollard et al’s article is useful in assessing children who may need PICU admission, it does not cover most children seen in district hospitals who have milder disease. We were disappointed that no mention was made of the role of severity scores, especially the Glasgow meningococcal septicaemia prognostic score (GMSPS)2; which has been validated retrospectively3 and prospectively.4
In a multicentre study of 152 patients in four district general hospitals and one secondary–tertiary centre, Marzouk4compared the performance characteristics for mortality of eight meningococcal scoring systems and laboratory markers of disease severity. The GMSPS performed best, a score ⩾ 8 identifying children at risk of dying from meningococcal disease (odds ratio of 87.0). Prospective validation in 278 children showed a GMSPS of ⩾ 8 to have 100% sensitivity, 75% specificity, a positive predictive value of 29%, and a negative predictive value of 100%.5
The GMSPS is particularly valuable because it can be repeated. This identifies deteriorating disease in patients who initially scored < 8 but then moved into the ⩾ 8 category despite early treatment, as well as patients who were ⩾ 8 but deteriorated further. Characteristics of the score do not seem to change when used by frontline medical staff.5 The GMSPS is useful in identifying severity of disease and predicting mortality. We recommend it to paediatricians to identify children needing PICU admission as an adjunct to Pollard et al’s algorithm.
Dr Pollard et al comment:
We are grateful that Carrol et al have raised the important issue of scoring systems in the evaluation of children with meningococcal disease. Such scores were designed and validated to predict death in cohorts of patients. In contrast, the emphasis of our algorithm is on the emergency management of meningococcal disease and prevention of death in the individual child. However, we recognise that the GMSPS may be useful in alerting the clinician to the important signs of critical illness in children with meningococcal disease and, when used for repeated review of such children, could aid the monitoring of stability or deterioration of the patient. GMSPS is also an important research tool allowing cohorts of patients to be categorised and compared between groups in audit, drug trials, and health planning.
Unfortunately, GMSPS is of only limited value when making decisions about the management of individual patients and could be misleading.1-1 1-2 False reassurance may be provided when the score is low. For example, a child who has severe raised intracranial pressure or compensated shock could score just 5 on the GMSPS but still be critically ill. Carrol et al point out that children with scores > 8 have a 29% chance of dying and suggest that the score may be helpful in deciding which patients require intensive care. However, this is not a reliable use of the GMSPS1-1 and some of the children with initial scores < 8 will also require intensive care and some of these children will deteriorate.
We do not advocate PICU admission for all children with meningococcal disease, as it is clear that most children with the disease do not develop critical illness. Our algorithm encourages clinicians to consider those signs suggestive of severe disease requiring intervention and we only advocate intensive care for those who do not improve after initial management. The algorithm also deals with recognition and initial treatment for children without shock or raised intracranial pressure, those who do not need to go to a PICU, and who should be managed on the general paediatric ward.
GMSPS may help in the evaluation of children with meningococcal disease, particularly where those in the frontline are unfamiliar with the disease, but the prognostic score must not be allowed to distract from urgent resuscitation of those who need it.