Multivariable predictive models for adverse outcome of invasive meningococcal disease in children

Hematologic Dysfunction Criteria in Critically Ill Children: The PODIUM Consensus Conference

CONTEXT

Studies of organ dysfunction in children are limited by a lack of consensus around organ dysfunction criteria.

OBJECTIVES

To derive evidence-informed, consensus-based criteria for hematologic dysfunction in critically ill children.

DATA SOURCES

Data sources included PubMed and Embase from January 1992 to January 2020.

STUDY SELECTION

Studies were included if they evaluated assessment/scoring tools to screen for hematologic dysfunction and assessed outcomes of mortality, functional status, organ-specific outcomes, or other patient-centered outcomes. Studies of adults or premature infants, animal studies, reviews/commentaries, small case series, and non-English language studies with inability to determine eligibility were excluded.

DATA EXTRACTION

Data were abstracted from each eligible study into a standard data extraction form along with risk of bias assessment.

RESULTS

Twenty-nine studies were included. The systematic review supports the following criteria for hematologic dysfunction: thrombocytopenia (platelet count <100000 cells/µL in patients without hematologic or oncologic diagnosis, platelet count <30000 cells/µL in patients with hematologic or oncologic diagnoses, or platelet count decreased ≥50% from baseline; or leukocyte count <3000 cells/µL; or hemoglobin concentration between 5 and 7 g/dL (nonsevere) or <5 g/dL (severe).

LIMITATIONS

Most studies evaluated pre-specified thresholds of cytopenias. No studies addressed associations between the etiology or progression of cytopenias overtime with outcomes, and no studies evaluated cellular function.

CONCLUSIONS

Hematologic dysfunction, as defined by cytopenia, is a risk factor for poor outcome in critically ill children, although specific threshold values associated with increased mortality are poorly defined by the current literature.

[Invasive meningococcal disease in Navarra in the era of a meningococcal C vaccine]

INTRODUCTION

Systematic childhood vaccination against meningococcus C has had a considerable impact on meningococcal invasive disease (MID). The aim of this study is to perform an analysis on the epidemiology, the clinical features, and the factors associated with a worse prognosis of MID, in the era of a meningococcal C vaccine.

MATERIAL AND METHODS

The study included confirmed cases of MID in children less than 15 years of age in Navarra, Spain, between 2008 and 2014. The risk of death or permanent sequelae was evaluated according to the presence of clinical features and analytical parameters at diagnosis.

RESULTS

The average annual incidence was 7.9 cases per 100,000 children, with the highest attack rate in children < 1 year. Of 53 cases analysed, 87% were due to meningococcus B. Fever (100%), rash (91%), and elevation of procalcitonin (94%) were the most frequent findings at diagnosis. Some sign of shock was observed in 70% upon arrival at the hospital. The case-fatality rate was 3.8% and 10 % survived with permanent sequelae. Glasgow coma scale < 15 (odds ratio [OR]= 9.2), seizure (OR=8.3), sepsis without meningitis (OR=9.1), thrombocytopenia (OR=30.5), and disseminated intravascular coagulation (OR= 10.9) showed a greater association with a worse prognosis.

CONCLUSION

The MID continues to be a significant cause of morbidity and mortality in children. Therefore, new advances are needed in the prevention, early diagnosis, and detection of the factors associated with poor prognosis.

Invasive pneumococcal disease in children prior to implementation of the conjugate vaccine in the Zurich region, Switzerland

OBJECTIVE

To describe symptoms, disease manifestations and outcome of invasive pneumococcal disease in children prior to implementation of the pneumococcal vaccine.

PATIENTS AND METHODS

Analysis of children younger than 16 years of age with invasive pneumococcal disease (IPD; n = 119). Children with culture-confirmed IPD, without underlying illness at risk for invasive disease, were included.

RESULTS

IPD in 90 children (age: median 2, mean 3.2 years) included 15 with meningitis, 16 with septicaemia, 14 with bacteraemia, 24 with pneumonia and 21 with skin, bone and joint infections. Symptoms of IPD most often described were fever and gastrointestinal symptoms (abdominal pain, vomiting, or diarrhoea), and coughing. More than 90% of children with pneumonia were coughing. Most importantly, clinical signs significantly predictive for severe IPD included tachycardia for sepsis, tachypnea for pneumonia, and meningeal signs for meningitis. Leukocyte, neutrophil and platelet counts were lower and C-reactive protein concentrations were higher on admission in children with complicated than in children with uncomplicated IPD but, due to wide overlap of these numbers, the difference was not of prognostic help to predict clinical course and outcome. Overall, 40% of children with IPD manifested complications and IPD showed a mortality rate of 6.6%.

CONCLUSIONS

IPD is a serious disease with a high complication rate and mortality. The clinical signs tachycardia, tachypnea, and meningism were highly predictive for severe IPD. The initial clinical presentation and laboratory evaluation were mostly unpredictable with respect to complications and outcome in contrast to the clinical signs.

[Foudroyant meningococcal sepsis in emergency medicine. Medical treatment and post-exposure prophylaxis]

A paediatric emergency case with a rescue helicopter, which was reported as unexpected cardiac arrest of a 4-year-old child will be described. Based on this case, the symptoms and therapy of meningococcal sepsis and Waterhouse-Friderichsen-Syndrome under emergency conditions will be discussed as well as aspects of post-exposure prophylaxis.

Severity of meningococcal infections is related to anthropometrical parameters

BACKGROUND

Invasive meningococcal infections remain an important cause of death in children. In addition, malnutrition has been classically associated with increased severity of infectious diseases. However, in our experience lethal meningococcaemia in clinically malnourished children is extremely rare. Our purpose was to determine whether there is an association between nutritional status and outcome in children with invasive meningococcal infection.

METHODS

We carried out an observational study and prospectively determined anthropometrical parameters in 127 children aged 1 month to 4 years with invasive meningococcal infection seen in our inpatient facilities from August 1999 to May 2004. Severity and survival were the clinical end points analysed.

RESULTS

Children with severe disease had higher weight for age (1.02 vs -0.19) and height for age (1.12 vs -0.58) z scores than those with non-severe disease. Non-survivors had higher weight for age (0.90 vs -0.16) and height for age (0.73 vs -0.57) z scores than survivors. Clinical and biological variables usually accepted as predictors of high mortality or severity in patients with meningococcal infection were not significantly associated with weight for age and height for age z scores.

CONCLUSION

In the present prospective series of children with invasive meningococcal disease, severity and death were linked to anthropometrical parameters and seemed to be associated with a very good nutritional status, which confirmed our previous uncontrolled observations.

Invasive meningococcal disease in children in Greece: comparison of serogroup A disease with disease caused by other serogroups

Although invasive meningococcal disease caused by serogroup A is not prevalent in developed countries, a considerable number of cases were recently recorded in Greece. In this study, serogroup A meningococcal disease was compared prospectively with meningococcal disease caused by other serogroups, using similar settings of testing and management during a 5-year period between 1999 and 2003. The Neisseria meningitidis serogroup was determined in 262 cases. Serogroup B predominated, accounting for 158 (60%) of the cases. Serogroup A was second most frequent (19%), followed by serogroups W135 (11%), C (8%), and Y (2%). No cases due to serogroup C were recorded during the last year of the study. Patients with serogroup A disease were older and had a milder course compared to patients infected with serogroups B or C. Toxic appearance, purpura, thrombocytopenia, abnormal coagulation tests, and the need for admission to the intensive care unit, fluid resuscitation, inotropic drugs, and mechanical ventilation were less common. Although morbidity and mortality were lower in these patients, the differences were not significant. Serogroup B is predominant in our area, and the introduction of an effective vaccine against it is a priority. Serogroup A has emerged as the second most common serogroup, but the illness associated with it is milder.

Genomic polymorphic profiles in an Irish population with meningococcaemia: is it possible to predict severity and outcome of disease?

Patients with meningococcal disease have increased plasma levels of proinflammatory cytokines IL-6, IL-1beta, and TNF-alpha, with higher levels associated with fatal outcome. This study investigated whether polymorphisms in genes encoding these cytokines, and in those encoding anti-inflammatory IL-10 and IL-1Ra, are associated with the outcome in patients with meningococcal disease. Seven polymorphisms were genotyped in 183 meningococcal disease patients and 389 controls. The IL-6 -174 G/G and IL-10 -1082 A/A genotypes were more frequent in nonsurvivors compared with survivors (P=0.023 IL-6, 0.25 IL-10), and in patients with severe disease compared to those with mild disease (P=0.037 IL-6, 0.0078 IL-10). An association was also found between meningococcal disease and the IL-1RN VNTR polymorphism, but no association was observed with the LTA +252, TNF -308, IL-10 -592, or IL-1B +3953 polymorphisms. We conclude that genetic variability in the IL-6, IL-10, and IL-1RN genes is associated with a poor outcome in meningococcal disease.

Cardiac rescue with enoximone in volume and catecholamine refractory septic shock

In December 2000 and February 2001, two children with suspected meningococcal disease were admitted to our pediatric intensive unit. Their Glasgow Meningococcal Septicaemia Prognostic score was 12 points. General treatment including antibiotics, steroids in case of meningitis, and fluid replacement, was performed. Despite appropriate volume replacement, intubation and ventilation, noradrenaline and adrenaline continuous infusions < or =1.0 microg/kg/min, and additional bolus infusions, cardiac output deteriorated within minutes in both children. Calcium and bicarbonate were given without sustained effect. Echocardiography demonstrated no pericardial effusion and shortening fraction was <10%. External cardiac massage had to be performed immediately in one case for electromechanical uncoupling. Both patients received a bolus of enoximone 2 mg/kg and 5 mg/kg body weight, respectively, followed by a continuous infusion of 20-23 microg/kg/min. Thereafter, both children had an adequate blood pressure and their shortening fraction increased to >30%. Within minutes, the catecholamine infusion could be reduced in both patients. The children completely recovered from their life-threatening situations. In patients with severe prolonged catecholamine and volume refractory endotoxin shock in Waterhouse-Friderichsen syndrome, even with electromechanical uncoupling and complete myocardial arrest, enoximone can immediately restore myocardial contractility.

Comparison of prediction models for adverse outcome in pediatric meningococcal disease using artificial neural network and logistic regression analyses

The objective of this study was to compare artificial neural network (ANN) and multivariable logistic regression analyses for prediction modeling of adverse outcome in pediatric meningococcal disease. We analyzed a previously constructed database of children younger than 20 years of age with meningococcal disease at four pediatric referral hospitals from 1985-1996. Patients were randomly divided into derivation and validation datasets. Adverse outcome was defined as death or limb amputation. ANN and multivariable logistic regression models were developed using the derivation set, and were tested on the validation set. Eight variables associated with adverse outcome in previous studies of meningococcal disease were considered in both the ANN and logistic regression analyses. Accuracies of these models were then compared. There were 381 patients with meningococcal disease in the database, of whom 50 had adverse outcomes. When applied to the validation data set, the sensitivities for both the ANN and logistic regressions models were 75% and the specificities were both 91%. There were no significant differences in any of the performance parameters between the two models. ANN analysis is an effective tool for developing prediction models for adverse outcome of meningococcal disease in children, and has similar accuracy as logistic regression modeling. With larger, more complete databases, and with advanced ANN algorithms, this technology may become increasingly useful for real-time prediction of patient outcome.

Extremity pain and refusal to walk in children with invasive meningococcal disease

OBJECTIVE

Early recognition of invasive meningococcal disease in children may be difficult. Extremity pain and refusal to walk (extremity symptoms) are uncommonly mentioned as clinical findings in children who present with this disease. We sought to determine 1) the frequency of extremity symptoms as part of the clinical presentation in children with invasive meningococcal disease and 2) whether these symptoms help identify children with otherwise unsuspected meningococcal disease.

METHODS

We reviewed the medical records of patients who were younger than 20 years and had invasive meningococcal disease from 1985 to 1996 at 3 pediatric referral centers. Children with extremity symptoms were identified and described. We compared clinical and laboratory findings and frequency of adverse outcomes between these children and those with invasive meningococcal disease without extremity symptoms.

RESULTS

We identified 274 children with invasive meningococcal disease, 45 (16%) of whom had either history or physical examination evidence of extremity pain (31) or refusal to walk (14) as part of their clinical presentations. Five of the 45 patients had arthritis at the time of presentation. Patients with extremity symptoms at presentation were significantly older (77.9 +/- 62.2 vs 44.0 +/- 56.9 months), had lower temperatures (38.8 +/- 1.2 degrees C vs 39.2 +/- 1.2 degrees C), and had higher band counts (28.2 +/- 15.2% vs 18.1 +/- 12.4%) than did patients without extremity symptoms. There were no significant differences, however, between groups with regard to rash, white blood cell counts, coagulation parameters, prevalence of meningitis, or adverse outcomes. Seventy-three (27%) of the 274 patients had unsuspected disease, and 5 (7%) of these had extremity symptoms at the time of diagnosis.

CONCLUSIONS

Sixteen percent of children with invasive meningococcal disease have extremity symptoms at the time of diagnosis. These symptoms may help to identify some patients with otherwise unsuspected invasive meningococcal disease.

Neurodevelopmental outcome in meningococcal disease: a case-control study

AIMS

To determine long term neurodevelopmental outcome following the spectrum of meningococcal infection.

METHODS

Between 1988 and 1990, 152 cases of meningococcal disease were recruited; 139 survived. Between 1998 and 1999, 115 survivors (83%) were evaluated, together with 115 sex and age matched controls. Standard measures of neurological function, coordination, cognition, behaviour, and hearing were used to assess neurodevelopmental status.

RESULTS

One case has spastic quadriplegia. Gross neurological examination was normal in all other cases and all controls. Five cases and no controls have significant hearing loss. Cases performed at a lower level than controls on measures of coordination, cognition, and behaviour. Four cases and no controls had major impairments. The adjusted odds ratios for moderate and minor impairments were 3.6 (95% CI 1.3 to 10.3) and 1.6 (95% CI 0.8 to 3.4) respectively.

CONCLUSION

The majority of survivors from this cohort do not have gross neurological deficits. However, when objective measures of motor function, cognitive ability, and behaviour were applied significant detriments were found in meningococcal survivors.

Early severe neutropenia and thrombocytopenia identifies the highest risk cases of severe meningococcal disease

OBJECTIVE: To determine the performance of established predictors of mortality in pediatric acute meningococcal disease (MD) in a contemporary population and to develop a simple predictive score that will not vary with observer. DESIGN: Prospective study for development set and mixed retrospective and prospective study for validation set. Setting and PATIENTS: A total of 227 patients with clinical meningococcal disease who were referred to three multidisciplinary pediatric intensive care units from 1993 to 1999. Early deaths before transfer to pediatric intensive care unit and deaths from cerebral herniation were included in the analysis. MEASUREMENTS AND MAIN RESULTS: The product of platelet and neutrophil counts at presentation (PN product) predicts mortality from meningococcal disease better than either count alone and at least as well as established severity scores. The Glasgow Meningococcal Septicaemia Prognostic Score and Malley scores performed poorly in these populations. The positive predictive value (PPV) for a Glasgow meningococcal septicemia prognostic score of >/=8/15 was 17.5% (16 of 91; 95% CI = 9%-25%), significantly lower than published estimates of 30%-74%, (p <.01). The PPV for death (or amputation) with a Malley score of 3/3 was 50% (12 of 24; 29%-71%), significantly lower than the published value of 100% (p <.001). The PN product appears to be a useful predictor. For a PN product of <40, PPV = 82% (9 of 11), specificity = 99% (195 of 197), and sensitivity = 73% (23 of 30). The performance of this score was greatest in younger children <5 yrs of age in whom clinical cerebral herniation was not seen as a cause of death (0 of 21 deaths at <5 yrs of age; 4 of 9 deaths at >/=5 yrs of age). CONCLUSION: Established scores significantly overestimate the occurrence of adverse outcomes in meningococcal disease. This may reflect improved resuscitation and outcome or variability in the application of these scores. The PN product achieves similar prediction to the scores currently in use and is independent of the observer. Factors that reflect the extent of the inflammatory response rather than the care before presentation are becoming increasingly important.

Can generic scores (Pediatric Risk of Mortality and Pediatric Index of Mortality) replace specific scores in predicting the outcome of presumed meningococcal septic shock in children?

OBJECTIVE

To compare, in children with septic shock and purpura, the accuracy in predicting death of two specific scores (the MenOPP bedside clinical [MOC] score of Gedde Dahl and the score of Groupe Francophone de Réanimation Pédiatrique [GFRP]), the C-reactive protein (CRP) level, and the two pediatric generic scores (the Pediatric Risk of Mortality [PRISM] and Pediatric Index of Mortality [PIM] scores).

DESIGN

Prospective, population-based study with analysis of previous comparative studies.

SETTING

A 14-bed pediatric intensive care unit in a university hospital.

PATIENTS

All children admitted consecutively to the pediatric intensive care unit with septic shock and purpura (n = 58, with 16 deaths [27.6%]) from January 1993 to May 2000.

INTERVENTIONS

None.

MEASUREMENTS AND MAIN RESULTS

The MOC and GFRP scores and the CRP level were prospectively determined at admission. The PRISM score was prospectively calculated within 24 hrs of admission or at the time of death, and the PIM score was calculated retrospectively between 1993 and 1997 and then prospectively from admission data. The nonparametric estimate of the area under the receiver operating characteristic curves (AUC) was calculated from the raw data using the Wilcoxon-Mann-Whitney two-sample statistic, and the standard error of the AUCs was calculated with DeLong's method. All the scores had an AUC >0.80, the PRISM probability of death having the best one (0.96 +/- 0.02). The PRISM value, which is easier to calculate, had an AUC of 0.95 +/- 0.02. The PRISM score performed significantly better than the PIM score (AUC, 0.83 +/- 0.06; p <.01) and the CRP level (AUC, 0.80 +/- 0.06; p <.01); however, there was no significant difference between the MOC (AUC, 0.91 +/- 0.04) and GFRP scores (AUC, 0.87 +/- 0.05). Analyzing literature and calculating AUCs from original data of previous studies, we observed that the superiority of the PRISM score had never been demonstrated in meningococcal diseases.

CONCLUSIONS

The PRISM score performed better than the PIM score, and was not surpassed by specific scores. Thus, we propose its use for outcome prediction in children with septic shock and purpura. However, if the PRISM score is to be used as inclusion criterion for clinical trials, it should be evaluated within a few hours after admission.

Meningococcal disease among children who live in a large metropolitan area, 1981-1996

Neisseria meningitidis is an important cause of serious bacterial infections in children. We undertook a study to identify meningococcal infections of the blood, cerebrospinal fluid, or both of children in a defined geographic area to describe the burden of disease and the spectrum of illness. We reviewed the medical records of all children aged <18 years who had meningococcal infections at the 4 pediatric referral hospitals in Boston, Massachusetts, from 1981 through 1996. We identified 231 patients with meningococcal disease; of these 231 patients, 194 (84%) had overt disease and 37 (16%) had unsuspected disease. Clinical manifestations included meningitis in 150 patients, hypotension in 26, and purpura in 17. Sixteen patients (7%) died. Although meningococcal disease is devastating to a small number of children, we found that the burden of pediatric disease that it caused at the 4 pediatric referral centers in this geographic region was limited; that patients with overt meningococcal disease are most likely to have meningitis; and that individual practitioners are unlikely to encounter a patient with unsuspected meningococcal disease.

Recognition, treatment and complications of meningococcal disease

Meningococcal disease remains a major cause of death in young children. A decrease in mortality requires recognition and treatment of the disease at a number of stages in the illness. Life-threatening meningococcal disease usually presents as septicaemia rather than meningitis. The cardinal feature of meningococcal septicaemia is the purpuric rash. Many parents recognise the rash and seek medical advice because of it. When primary care physicians recognise the rash, the administration of parenteral penicillin may decrease mortality. However, antibacterials are not given promptly if there is no rash or if the disease presents in an atypical form. In hospital, antibacterial therapy with a third-generation cephalosporin should be given. Disease severity needs to be assessed by a valid method, such as the Glasgow Meningococcal Septicaemia Prognostic Score (GMSPS). This can identify those patients who need intensive care and/or might benefit from new therapies. The 2 life-threatening complications are septic shock and meningoencephalitis with raised intracranial pressure. Despite numerous case reports of success with potential new treatments, none has been proven safe and/or effective by controlled trials. Although it is tempting to focus on new treatments, the early recognition of severe meningococcal disease by parents, primary care physicians and junior hospital doctors is equally, if not more, important as a potential means of decreasing mortality.

Clinical and hematologic features do not reliably identify children with unsuspected meningococcal disease

OBJECTIVE

To determine the frequency of unsuspected meningococcal disease (UMD) in young febrile children with meningococcal infections and evaluate whether clinical and laboratory parameters commonly used in the evaluation of fever can help identify children with UMD.

METHODS

We reviewed the records of children with meningococcal disease from 1985 to 1996 at four referral centers. Children who were evaluated as outpatients and then discharged to home, from whom Neisseria meningitidis was isolated from blood or cerebrospinal fluid cultures obtained during these outpatient visits, were considered to have UMD. We compared clinical and laboratory parameters between these children and 6414 febrile outpatients 3 to 36 months old with negative blood cultures enrolled in a separate study of occult bacteremia.

RESULTS

We identified 381 children with meningococcal disease, of whom 45 (12%) had UMD. Of the 45 with UMD, 37 (82%) were 3 to 36 months old. Compared with the 6414 culture-negative patients, these 37 patients with UMD were significantly younger (8.9 +/- 5.4 vs 14.2 +/- 8.1 months) and had significantly higher band counts (14.3 +/- 11.1 vs 7.3 +/- 7.5%). There were no significant differences, however, in temperature, white blood cell counts, and absolute neutrophil counts. Multivariate analysis identified young age and the band count as independent predictors of UMD.

CONCLUSIONS

Children ultimately diagnosed with meningococcal disease have commonly been evaluated as outpatients and discharged to home before diagnosis. Of the hematologic parameters frequently used in the evaluation of fever, only the band count differs significantly between young febrile children with UMD and those with negative cultures. Because UMD is uncommon in young febrile pediatric outpatients, however, the predictive value of the band count is low. Thus, the complete blood count is not routinely helpful for the diagnosis of UMD.

Cerebrospinal fluid pleocytosis and prognosis in invasive meningococcal disease in children

BACKGROUND

The absence of cerebrospinal fluid (CSF) pleocytosis in invasive meningococcal disease (IMD) has been associated with an increased risk of death. It is unknown whether patients who lack a cellular response to central nervous system (CNS) infection are at the same risk of adverse outcome as patients who lack CNS infection.

OBJECTIVES

To determine the frequency of presentation and outcome of three groups of children with IMD: Group 1, children with CSF pleocytosis; Group 2, children without CSF pleocytosis and with negative CSF cultures (bacteremia alone); and Group 3, children without CSF pleocytosis but with positive CSF cultures (CNS infection without CSF pleocytosis).

METHODS

We reviewed the medical records of children with IMD at four pediatric referral hospitals between 1985 and 1996. Clinical and laboratory indices and severe adverse outcomes (defined as death or limb loss) were compared in the three groups. Multivariable logistic regression analysis was performed to determine whether CNS infection without CSF pleocytosis was independently associated with adverse outcome in IMD.

RESULTS

Three hundred seventy-seven children with IMD were identified. Eighty-six patients were excluded because their CSF analysis either was not done or was unevaluable; of these patients 22 (25.6%) had an adverse outcome. Of the 291 evaluable patients 204 (70.1%) had CSF pleocytosis, 52 (17.9%) had bacteremia alone and 35 (12.0%) had CNS infection without CSF pleocytosis. Patients with CNS infection without CSF pleocytosis had significantly lower white blood cell and platelet counts and more coagulopathy than patients with bacteremia alone (P < or = 0.05) or patients with CSF pleocytosis (P < or = 0.01). The frequency of adverse outcome was 40% for patients with CNS infection without CSF pleocytosis compared with 9.6% for patients with bacteremia alone (P = 0.001) and 3.4% for patients with CSF pleocytosis (P < 0.001). CNS infection without CSF pleocytosis was independently associated with adverse outcome by multivariable logistic regression analysis (P = 0.003).

CONCLUSIONS

Approximately 30% of all children with IMD present without CSF pleocytosis. Of these patients those with CNS infection without pleocytosis are at higher risk of adverse outcome than either patients with CSF pleocytosis or patients with bacteremia alone.