The Duration Of Antibiotic Therapy And Criteria Of Hospital Discharge For Pediatric Bacterial Meningitis

Objective1. To understand the usage and duration of antibiotic therapy in pediatric inpatientswith bacterial meningitis in our hospital.2. To investigate the times for children with bacterial meningitis to recover thenormal level of cerebrospinal fluid in our hospital.3. To explore the relationship between the cerebrospinal fluid level of children withbacterial meningitis at discharge and their prognosis.4. To explore the duration of antibiotics therapy and the criteria for discharge forchildren with bacterial meningitis.Subjects and Methods1. SubjectsWe selected pediatric patients with bacterial meningitis who were admitted to theneurological rehabilitation department of Guangzhou Women and Children’s MedicalCenter from January2005to June2013. These patients all gestational aged from37to42weeks, with normal birth history, normal intelligence and growth, and normalweight before illness, excluding those with other serious neurological, psychiatric orcongenital metabolic disease, endocrine system disease, immunodeficiency disease,congenital brain malformations, head trauma, or infections other intracranialinfectious meningitis. They all accepted the entire course of antibiotic treatment andthe times for their follow-up were all≥6months from their discharge, and their WBC of cerebrospinal fluid was＜20*106/L when they were discharged. A total of221cases were collected and we telephoned their parents to bring their children back tothe hospital for follow-up examination and evaluation. Finally a total of99childrenwere successfully followed up.2. Methods2.1Data collection: We collected medical records of pediatric bacterialmeningitis cases that met the inclusion criterion, and summarized their generalinformation (name, sex, age, gestational age, birth history, family history, pre-illnesshealth status and nutritional status) and hospitalization information (clinicalmanifestation, clinical general situation, complications, clinical diagnosis, laboratoryexamination results, pathogenic examination results and other related examination\).We estimated their prognosis score according to their situation at the time ofadmission. Scoring method: Coma required intubation3points, seizures2points,hypothermia2points, shock1point,<12months1point, CSF WBC>1000*106/L1point, Hb<110g/L1point, still fever three days after admission0.5point, CSF Glu＜1.1mmol/L0.5point, in total≥4.5points, the mortality and unfavorable outcome willbe more[9].2.2Follow-up: We emailed the follow-up questionnaire and called the parents tobring their children back to the hospital for follow-up. Through the questionnaire weacquire the information on the parents’ education levels and the children’s recovery after leaving the hospital. Then we conducted the physical examinations for the children and tested them with Gasell Developmental Scale (＜3years and10monthsold) or Wechsler Intelligence Scale(≥3years and10months old) and parents of thechildren who were older than2years old filled in the Child Behaviour Checklist.According to the last examination results of the head MRI, BAEP and EEG of thechildren and the compliance of their parents, the children were selectively re-checkedwith those three examinations. Children who failed to return to the hospital wereexcluded.2.3Data arrangement: We checked the medical records and divided the cases into4groups according to results of the last cerebrospinal fluid(CSF) examination when discharged:(1) group GP: including children whose CSF WBC count was＜20*106/L, neither CSF protein nor glucose concentration were normal whendischarged;(2) group G: including children whose CSF WBC count was＜20*106/L,CSF protein concentration was normal but glucose concentration was abnormal whendischarged;(3) group P: including children whose CSF WBC count was＜20*106/L,CSF glucose concentration was normal but protein concentration was abnormal whendischarged;(4) group N: including children whose CSF WBC count was＜20*106/L,and both CSF protein and glucose concentration were normal when discharged. Based on the prognostic scores and follow‐up results of each group, we evaluated whether the children developed sequelae or not and the severity of sequlae (The grading of sequelae severity was based on the Classification of Diseases of the Global Burden of Disease‐Disease Control Priorities Project) in each group, and explored the relationship between CSF results at discharge and the prognosis.2.4Statistical analysis: The database was set up by using Microsoft Excel andanalyzed by using Statistical Package of Social Science version16.0(SPSS16.0). Thesignificance level of statistical analysis was α=0.05. We used the one way ANOVAto compare the prognosis score of these4groups, while the follow-up prognosisresults were compared by using Kruskal-Wallis test. Finally we used ordinal logisticregression to analyze the relationship among grouping, prognosis score and thefollow-up prognosis. We used Chi-square test or the Fisher’s exact probabilities tocompare other qualitative variables. Quantitative variables data were changed intoqualitative variables according to clinical practice as previously reported in otherliteratures. We used the median survival time to calculate the time for recovery ofCSF protein and glucose concentration.Results1. The use of antibioticsThis study included99pediatric cases who all accepted antibiotic treatmentwhen suspected of bacterial meningitis, and among them21cases accepted single antibiotic therapy during the acute phase of hospitalization,77cases accepted thecombination therapy of two antibiotics, and1case accepted the combination therapyof three antibiotics. The range of duration of antibiotics treatment for the99cases wasfrom7to145days, with an average of59.10±28.28days.2. Prognostic scoreAccording to the CSF level at discharge, the99cases were divided into fourgroups: Group N of20cases, group GP of39cases, group G of22cases, and group Pof18cases. Prognostic scores were estimated for the four groups according to theircondition at admission, and cases with the scores≥4.5were found to havesignificantly higher mortality and more adverse clinical outcomes[9].10cases (25.6%)in group GP had scores≥4.5, while29cases (74.4%)＜4.5;6cases (27.3%) ingroup G hand scores≥4.5, and16cases (72.7%)＜4.5points;6cases (33.3%) ingroup P had scores≥4.5, and12cases (66.7%)＜4.5;5cases (25.0%) in group C hadscore≥4.5, and15cases (75.0%)＜4.5. The difference for the percentage of≥4.5among the4groups had no statistical significance with P=0.932.3. The time for recovery of cerebrospinal fluid to normal levelThe CSF WBC count for all the99cases were found to recover to＜20*106/Lwith a time range of4to108days and an average of34days;41cases CSF glucoseconcentration were not tracked to restore normal, its recovery normal time the shortestis1day, the longest is by the388th day had not been restored to normal, median is90days;23cases CSF protein were not tracked to return to normal, the shortest time is5days, the longest is by the388th day not yet recovery normal, median89days.4. The follow-up prognosisThe99children who were follow-up with no obvious neurological sequelae werein77cases (77.78%), mild sequelae were in8cases (8.08%), severe sequelae were in14cases (14.14%).5. The relationship between CSF results at discharge and prognosisFor the99cases, all of their CSF WBC counts were＜20*106/L when they were discharged, with an average of2.58±0.48mmol/L for their CSF glucoseconcentration and0.58±0.43g/L for CSF protein concentration.77cases had nosequelae with an average of2.61±0.45mmol/L for CSF glucose concentration and0.52±0.22g/L for CSF protein concentration at discharge.8cases had mild sequelaewhose glucose average was2.38±0.40mmol/L and protein on average0.55±0.16g/L at discharge.17cases had severe sequelae whose glucose average was2.56±0.63mmol/L and protein average0.88±0.98g/L at discharge. The difference ofCSF glucose concentration among no sequelae group, mild sequelae group and severesequelae group had no statistical significance with P=0.427. The difference of CSFprotein concentration among the three groups had statistical significance with P=0.013, and there was a significant difference between no sequelae group and severesequelae group for pairwise comparison with P=0.003, the ROC curve threshold was＜0.87g/L.Among4groups, group GP had28cases (71.8%) with no sequelae,4cases(10.3%) with mild sequelae, and7cases(17.9%) with severe sequelae; group G had18cases(81.8%) with no sequelae,2cases (9.1%) with mild sequelae, and2cases(9.1%) with serious sequelae; group P had15cases (75.0%) with no sequelae,only1(5.0%) case with mild sequelae, and4cases(20.0%) with severe sequelae;group N had16cases (76.2%) with no sequelae,1case (4.8%) with mild sequelae,and4cases(19.0%) with severe sequelae. The differences in percentage of follow-upoutcomes of no sequelae, mild sequelae and severe sequelae for the four groups hadno statistical significance with P=0.721.Conclusions1. These studies of99cases bacterial meningitis in children in our hospital weredischarged after treatment without a recurrence.8.08%of them had mild sequelae,14.14%of them had severe sequelae.2. Our study suggested that stopped antibiotic treatment to discharge when the CSFprotein concentration was＞0.87g/L, does influence the prognosis, the higherCSF protein the worse prognosis. 3. Based on this study, we recommend that the clinical criteria for stopping antibiotictreatment in children with bacterial meningitis should be: absence of fever formore than one week after therapy, disappearance of all acute clinical symptoms,CSF WBC count of＜20*106/L, CSF glucose concentration should be＞1.8mmol/L and CSF protein concentration should be＜0.87g/L. One week afterthe discontinuation of antibiotics, CSF examination should be repeated. Patientscan be discharged only when they have been clinically stable for2weeks. Afterdischarge, patients should be monitored for any neurological sequelae duringfollow-up and performed a CSF examination every two weeks until their CSFbecomes normal.