Indian J Pediatr DOI 10.1007/s12098-014-1357-z
ORIGINAL ARTICLE
The Role and Reliability of Rapid Bedside Diagnostic Test in Early Diagnosis and Treatment of Bacterial Meningitis Arun Kumar & Pradeep Kumar Debata & Amitabh Ranjan & Rajani Gaind
Received: 26 June 2013 / Accepted: 13 January 2014 # Dr. K C Chaudhuri Foundation 2014
Abstract Objective To evaluate the role and reliability of rapid bedside diagnostic test in early diagnosis and treatment of bacterial meningitis in children using reagent strips. Methods This prospective, single blinded study was conducted in the Department of Pediatrics of VMMC & Safdarjung Hospital, New Delhi in collaboration with the Department of Microbiology of VMMC & Safdarjung Hospital, New Delhi, over a period of 15 mo (August 2009 to Nov 2010). Seventy-five children aged 3 mo to 12 y admitted in the pediatric ward with suspected diagnosis of acute meningitis were included. All enroled patients underwent lumbar puncture. CSF samples were taken and divided in 2 parts for laboratory evaluation and rapid strip analysis. The sensitivity, specificity, positive predictive value and the negative predictive values of the reagent strips for the diagnosis of bacterial meningitis were calculated. Accuracy of the reagent strips was established using kappa statistics. Latex agglutination for antigen detection and microbiological culture were also done. Results Highly significant association was observed between CSF examination in routine laboratory method and dipstick method. The number of laboratory values that correlated were- for cells 71(94.63 %), for protein 68 (90.67 %), for glucose 68(90.67 %) out of total 75 cases. The sensitivity and specificity of reagent strip in diagnosing acute bacterial meningitis were 96.7 % and 97.8 % respectively. The positive A. Kumar : P. K. Debata (*) : A. Ranjan Department of Pediatrics, Vardhman Mahavir Medical College and Safdarjang Hospital, New Delhi 110029, India e-mail: [email protected] R. Gaind Department of Microbiology, Vardhman Mahavir Medical College and Safdarjang Hospital, New Delhi, India
predictive and negative predictive values of reagent strip in diagnosing acute bacterial meningitis were 96.7 % and 97.8 % respectively. Staphylococcus aureus was found to be the most common organism isolated (50 %). Conclusions Thus reagent strip analysis is a very rapid, reliable and effective method for diagnosis of acute bacterial meningitis in children. Staphylococcus aureus was the most common organism isolated.
Keywords Acute bacterial meningitis . Lumbar puncture . CSF examination . Reagent strip analysis . Rapid diagnosis . Culture positivity
Introduction Acute meningitis is one of the most serious infections in children. Accurate early diagnosis and treatment of the entity are the keys in the management of these cases to reduce the morbidity and mortality [1,2]. Meningitis can be lifethreatening because of the inflammation’s proximity to the brain and spinal cord; therefore the condition is classified as a medical emergency [3,4]. Diagnosis of meningitis traditionally depends on cytological, biochemical and bacteriological tests of CSF. Conventionally to get the biochemistry report and cell count takes about 6–8 h, gram stain takes 8–24 h and culture takes about 72 h in most of the centers. On the other hand, many centers like District hospital and Community health centers which are the backbone of our health care system do not have laboratory facilities particularly for analysis of CSF. So, for the developing countries, it is very essential to have a diagnostic method which is simple to be carried out at the levels of primary health centre and the first referral units (FRU).
Indian J Pediatr
Material and Methods This prospective, single blinded study was conducted in the Department of Pediatrics of VMMC & Safdarjung Hospital, New Delhi in collaboration with the Department of Microbiology of VMMC & Safdarjung Hospital, New Delhi over a period of 15 mo (August 2009 to Nov 2010) after obtaining clearance from hospital ethical committee. Seventy-five children from 3 mo to 12 y with a suspected diagnosis of acute meningitis based on clinical findings admitted in one of the pediatric ward were included in the study. Children with age less than 3 mo, conditions where lumbar puncture (LP) is contraindicated, shunt related infections and the patients who did not give consent for lumbar puncture or those who had already received treatment from other hospitals were excluded from the study. After informed consent, all the enroled patients underwent lumbar puncture following thorough clinical evaluation and funduscopic examination. CSF samples were subjected to routine diagnostic tests in the laboratory immediately and reagent strip analysis for glucose, protein and leukocyte examination. The diagnosis of meningitis was made according to the results of the CSF analysis (Table 1). The reagent strip analysis was carried out immediately after obtaining the sample by the investigator and an independent observer and the second part of the sample were sent to the laboratory for analysis. The technicians performing the analysis in the laboratory was unaware of the results obtained. Reagent strip Combur-10 (Boehringer Mannheim) is a 10-patch strip used for semi-quantitative determination of specific gravity, protein, glucose, leukocytes, nitrites, pH, hemoglobin, ketones, bilirubin, and urobilinogen in urine by visual reading. Combur 10 was preferred in the present study because it can also detect very low (less than 2.8 mmol) values of glucose. The reagent strip was directly dipped in to the CSF or else the CSF was withdrawn into a pipette and 1–2 drops were poured on the glucose, protein and leukocyte patches. The CSF was washed off after 60 s and the color change was read against the standard provided. The presence of leukocytes was graded as negative, 10–25, 25–75, 75–500 and >500/microL; protein as 30, 100 and 500 g/L; and glucose as less than 2.8, 2.8, 5.5 mmol/L. Laboratory values of CSF glucose, protein and leukocytes were compared against the corresponding value of the reagent strip. Because of the skewed deviation, natural log values were used; direct correlation was not possible because the strip data was categorical and the laboratory data was continuous. The authors took the midpoint values between the dipstick categories as cut-points to determine the ranges for each category. The number of laboratory values, falling within these, was taken and agreement was assessed by the kappa
Table 1 Interpretation of CSF samples Gram stain and/or Antigen detection and/or CSF culture
Glucose < 2.8 mmol/L Protein > 500 mg/dL Any number of cells
Or Positive blood culture with meningitis, CSF polymorphs, nuclear cytosis
Glucose < 2.8 mmol/L Protein > 500 mg/dL Cell count ≥ 500 cell/μL
statistics. Other variables studied included distribution of age, sex, symptoms and signs, latex agglutination (L A) test, grams stain and culture.
Results Among 75 suspected cases of meningitis, 30 cases (40.00 %) were positive and 45 cases (60.00 %) were negative by laboratory criteria. Among laboratory positive cases, 29 cases (96.66 %) were also positive by reagent strip, 1 case was falsely negative by reagent strip and 1 case was falsely positive by reagent strip. Reagent strip showed a positive correlation with laboratory test which has a high statistical significance (k value 0.944, p value 0.000). The number of laboratory values that correlated were- for cells 71(94.63 %), for protein 68 (90.67 %), for glucose 68(90.67 %) out of total 75 cases. Reagent strip showed a positive correlation with laboratory test which has a high statistical significance (kappa 0.94, 0.819, 0.819; p
ORIGINAL ARTICLE
The Role and Reliability of Rapid Bedside Diagnostic Test in Early Diagnosis and Treatment of Bacterial Meningitis Arun Kumar & Pradeep Kumar Debata & Amitabh Ranjan & Rajani Gaind
Received: 26 June 2013 / Accepted: 13 January 2014 # Dr. K C Chaudhuri Foundation 2014
Abstract Objective To evaluate the role and reliability of rapid bedside diagnostic test in early diagnosis and treatment of bacterial meningitis in children using reagent strips. Methods This prospective, single blinded study was conducted in the Department of Pediatrics of VMMC & Safdarjung Hospital, New Delhi in collaboration with the Department of Microbiology of VMMC & Safdarjung Hospital, New Delhi, over a period of 15 mo (August 2009 to Nov 2010). Seventy-five children aged 3 mo to 12 y admitted in the pediatric ward with suspected diagnosis of acute meningitis were included. All enroled patients underwent lumbar puncture. CSF samples were taken and divided in 2 parts for laboratory evaluation and rapid strip analysis. The sensitivity, specificity, positive predictive value and the negative predictive values of the reagent strips for the diagnosis of bacterial meningitis were calculated. Accuracy of the reagent strips was established using kappa statistics. Latex agglutination for antigen detection and microbiological culture were also done. Results Highly significant association was observed between CSF examination in routine laboratory method and dipstick method. The number of laboratory values that correlated were- for cells 71(94.63 %), for protein 68 (90.67 %), for glucose 68(90.67 %) out of total 75 cases. The sensitivity and specificity of reagent strip in diagnosing acute bacterial meningitis were 96.7 % and 97.8 % respectively. The positive A. Kumar : P. K. Debata (*) : A. Ranjan Department of Pediatrics, Vardhman Mahavir Medical College and Safdarjang Hospital, New Delhi 110029, India e-mail: [email protected] R. Gaind Department of Microbiology, Vardhman Mahavir Medical College and Safdarjang Hospital, New Delhi, India
predictive and negative predictive values of reagent strip in diagnosing acute bacterial meningitis were 96.7 % and 97.8 % respectively. Staphylococcus aureus was found to be the most common organism isolated (50 %). Conclusions Thus reagent strip analysis is a very rapid, reliable and effective method for diagnosis of acute bacterial meningitis in children. Staphylococcus aureus was the most common organism isolated.
Keywords Acute bacterial meningitis . Lumbar puncture . CSF examination . Reagent strip analysis . Rapid diagnosis . Culture positivity
Introduction Acute meningitis is one of the most serious infections in children. Accurate early diagnosis and treatment of the entity are the keys in the management of these cases to reduce the morbidity and mortality [1,2]. Meningitis can be lifethreatening because of the inflammation’s proximity to the brain and spinal cord; therefore the condition is classified as a medical emergency [3,4]. Diagnosis of meningitis traditionally depends on cytological, biochemical and bacteriological tests of CSF. Conventionally to get the biochemistry report and cell count takes about 6–8 h, gram stain takes 8–24 h and culture takes about 72 h in most of the centers. On the other hand, many centers like District hospital and Community health centers which are the backbone of our health care system do not have laboratory facilities particularly for analysis of CSF. So, for the developing countries, it is very essential to have a diagnostic method which is simple to be carried out at the levels of primary health centre and the first referral units (FRU).
Indian J Pediatr
Material and Methods This prospective, single blinded study was conducted in the Department of Pediatrics of VMMC & Safdarjung Hospital, New Delhi in collaboration with the Department of Microbiology of VMMC & Safdarjung Hospital, New Delhi over a period of 15 mo (August 2009 to Nov 2010) after obtaining clearance from hospital ethical committee. Seventy-five children from 3 mo to 12 y with a suspected diagnosis of acute meningitis based on clinical findings admitted in one of the pediatric ward were included in the study. Children with age less than 3 mo, conditions where lumbar puncture (LP) is contraindicated, shunt related infections and the patients who did not give consent for lumbar puncture or those who had already received treatment from other hospitals were excluded from the study. After informed consent, all the enroled patients underwent lumbar puncture following thorough clinical evaluation and funduscopic examination. CSF samples were subjected to routine diagnostic tests in the laboratory immediately and reagent strip analysis for glucose, protein and leukocyte examination. The diagnosis of meningitis was made according to the results of the CSF analysis (Table 1). The reagent strip analysis was carried out immediately after obtaining the sample by the investigator and an independent observer and the second part of the sample were sent to the laboratory for analysis. The technicians performing the analysis in the laboratory was unaware of the results obtained. Reagent strip Combur-10 (Boehringer Mannheim) is a 10-patch strip used for semi-quantitative determination of specific gravity, protein, glucose, leukocytes, nitrites, pH, hemoglobin, ketones, bilirubin, and urobilinogen in urine by visual reading. Combur 10 was preferred in the present study because it can also detect very low (less than 2.8 mmol) values of glucose. The reagent strip was directly dipped in to the CSF or else the CSF was withdrawn into a pipette and 1–2 drops were poured on the glucose, protein and leukocyte patches. The CSF was washed off after 60 s and the color change was read against the standard provided. The presence of leukocytes was graded as negative, 10–25, 25–75, 75–500 and >500/microL; protein as 30, 100 and 500 g/L; and glucose as less than 2.8, 2.8, 5.5 mmol/L. Laboratory values of CSF glucose, protein and leukocytes were compared against the corresponding value of the reagent strip. Because of the skewed deviation, natural log values were used; direct correlation was not possible because the strip data was categorical and the laboratory data was continuous. The authors took the midpoint values between the dipstick categories as cut-points to determine the ranges for each category. The number of laboratory values, falling within these, was taken and agreement was assessed by the kappa
Table 1 Interpretation of CSF samples Gram stain and/or Antigen detection and/or CSF culture
Glucose < 2.8 mmol/L Protein > 500 mg/dL Any number of cells
Or Positive blood culture with meningitis, CSF polymorphs, nuclear cytosis
Glucose < 2.8 mmol/L Protein > 500 mg/dL Cell count ≥ 500 cell/μL
statistics. Other variables studied included distribution of age, sex, symptoms and signs, latex agglutination (L A) test, grams stain and culture.
Results Among 75 suspected cases of meningitis, 30 cases (40.00 %) were positive and 45 cases (60.00 %) were negative by laboratory criteria. Among laboratory positive cases, 29 cases (96.66 %) were also positive by reagent strip, 1 case was falsely negative by reagent strip and 1 case was falsely positive by reagent strip. Reagent strip showed a positive correlation with laboratory test which has a high statistical significance (k value 0.944, p value 0.000). The number of laboratory values that correlated were- for cells 71(94.63 %), for protein 68 (90.67 %), for glucose 68(90.67 %) out of total 75 cases. Reagent strip showed a positive correlation with laboratory test which has a high statistical significance (kappa 0.94, 0.819, 0.819; p
