+
MODEL
Pediatrics and Neonatology xxx (xxxx) xxx
Available online at www.sciencedirect.com
ScienceDirect journal homepage: http://www.pediatr-neonatol.com
Original Article
Diagnostic performance of the Sputum Gram Stain in predicting sputum culture results for critically ill pediatric patients with pneumonia Wun-Yan Huang a, Ming-Sheng Lee b,c, Liang-Mei Lin d, Yi-Chen Liu d,* a Department of Pediatric Emergency Medicine, China Medical University Children’s Hospital, China Medical University, Taichung, Taiwan b Department of Pediatric Pulmonology and Critical Care, Changhua Christian Children Hospital, Changhua, Taiwan c Graduate Program in Nutrition, Chung Shan Medical University, Taichung, Taiwan d Respiratory Therapy Section for Children, Changhua Christian Children Hospital, Changhua, Taiwan
Received May 16, 2019; received in revised form Jan 21, 2020; accepted Mar 31, 2020
Available online - - -
Key words gram stain; pediatric; sputum culture
Background: The sputum Gram stain is an inexpensive, rapid, and convenient laboratory method that predicts the bacterial pathogens in patients with pneumonia. This study aimed to evaluate the diagnostic performance of this method in predicting sputum culture results for critically ill pediatric patients. Methods: From June 2008 to June 2018, patients with pneumonia with an endotracheal or a tracheostomy tube in place in the Pediatric Intensive Care Unit at Changhua Christian Hospital were enrolled retrospectively. Sputum was collected from each patient via the artificial airway for Gram stain and culture evaluations of bacterial pathogens. Mixed culture results were excluded. A successful prediction was defined as a match of the sputum Gram stain and culture results. Results: A total of 622 records were reviewed, of which 542 were analyzed. Haemophilus influenzae, Pseudomonas aeruginosa, and Streptococcus pneumoniae were the three most common pathogens found. The overall prediction success rate of the sputum Gram stain was 59.23%. The sensitivity of the method in predicting gram-negative bacilli (GNB), gramnegative cocci (GNC), and gram-positive cocci (GPC) was 0.45, 0.67, and 0.61, respectively. Its specificity in predicting GNB, GNC, and GPC was 0.87, 0.98, and 0.87, respectively. Its positive likelihood ratio in predicting GNB, GNC, and GPC was 3.46, 33.50, and 4.69, respectively. The highest prediction success rate among all pathogens was for GNC.
* Corresponding author. No. 135, Nanxiao St., Changhua City, Changhua County, 500, Taiwan. E-mail address:
[email protected] (Y.-C. Liu). https://doi.org/10.1016/j.pedneo.2020.03.014 1875-9572/Copyright ª 2020, Taiwan Pediatric Association. Published by Elsevier Taiwan LLC. This is an open access article under the CC BYNC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Please cite this article as: Huang W-Y et al., Diagnostic performance of the Sputum Gram Stain in predicting sputum culture results for critically ill pediatric patients with pneumonia, Pediatrics and Neonatology, https://doi.org/10.1016/j.pedneo.2020.03.014
+
MODEL
2
W.-Y. Huang et al Conclusion: The sputum Gram stain had high specificity and relatively low sensitivity in predicting the bacterial pathogens in critically ill pediatric patients. Its high specificity in predicting sputum culture results means that clinicians can confidently use sputum Gram stain results to guide their antibiotic choice for treatment. Copyright ª 2020, Taiwan Pediatric Association. Published by Elsevier Taiwan LLC. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/ by-nc-nd/4.0/).
1. Introduction The sputum Gram stain and sputum culture are routine tests for patients with suspected pneumonia.1 Bacteria are broadly divided into two groups depending on whether the cells are round (cocci) or rod shaped (bacilli). The Gram stain is used to distinguish and classify bacterial species into gram-positive and gramnegative groups by detecting the peptidoglycan that is present in the bacterial cell wall. In this procedure, bacteria that stain purple are known to be gram positive and those that stain red are considered gram negative. Because either of the bacterial shapes may be gram positive or gram negative, a mixture of gram-positive and gram-negative bacteria can occur in the same microscope field. On the basis of these features of shape and color in the Gram stain method, any bacterium can be classified into one of four groups: gram-negative bacilli (GNB), gram-negative cocci (GNC), gram-positive bacilli (GPB), and gram-positive cocci (GPC). A sputum culture is another diagnostic test for the pathogen that causes pneumonia. The bacterial type is identified using a microscope or chemical tests. The sputum culture test remains the most common method used by clinicians to make the specific diagnosis of bacterial pneumonia and to guide antibiotic use for treatment.2 White blood cells (WBCs) and C-reactive protein (CRP) are indirect indicators of an acute bacterial infection. The higher the WBCs or CRP in the blood, the greater is the chance of there being a bacterial infection. In the vast majority of patients with severe pneumonia, significant increases in WBCs and CRP can be found in blood draws, and sputum cultures generally have a higher chance of developing bacteria. There are some reasons why the Gram stain has garnered so much acceptance. The procedure has low risk and is fast and inexpensive.3 On the basis of the rapid smear result, the clinician can choose adequate antibiotics for treatment as soon as possible. However, Gram stain results are sometimes not compatible with sputum culture results, and research on this issue with regard to children is still warranted.4,5 The aim of our study was therefore to analyze the accuracy of sputum Gram staining in predicting sputum culture results for critically ill pediatric patients. Moreover, we also analyzed whether WBC and CRP levels could be used to predict positive results of sputum cultures in this patient population.
2. Methods 2.1. Patients, and calculations of the diagnostic performance of the sputum gram stain This research was a retrospective study of the Pediatric Intensive Care Unit (PICU) at Changhua Christian Children’s Hospital (Taiwan), which is a tertiary care center that receives referrals from central Taiwan. The PICU has 15 beds and admits over 500 patients annually. Between June 2008 and June 2018, the medical records of all pediatric patients who were suspected of having pneumonia on the basis of clinical symptoms and chest X-ray findings during admission, and who had an endotracheal or a tracheostomy tube placement, were included in this study. Sputum samples were collected from the patients via the artificial airway for Gram stain and culture evaluations of the bacterial pathogen. Each Gram stain was performed on a purulent portion of each sputum specimen. Sputum samples with less than 5 squamous epithelial cells and more than 25 polymorphonuclear cells per low-power field were enrolled. The frequency of occurrence in each bacterial morphology in oil immersion field (OIF) was recorded as 1þ (rare, <1/OIF), 2þ (few, 1e5/OIF), 3þ (moderate, 6e30/ OIF), or 4þ (heavy, >30/OIF).6 If the Gram stain indicated only one type of bacterium, the result was reported regardless of the number. If the Gram stain indicated two or more types of bacteria, the type with the highest number was designated the predominant morphotype, but the result was recorded only if the number was more than 3. Gram stain results that did not indicate either of these cases were classified as mixed stains. The specimens were cultured on plates containing 5% sheep blood, chocolate, and MacConkey agar, and then inoculated with 5% carbon dioxide at 35e37 C for a minimum of 48 h. Samples with mixed culture results were excluded and only the ones with pure growth were included. The clinical characteristics including age, gender, outcome, disease distribution, and intubation days of patients with positive and negative sputum cultures were compared. Analysis of bacterial pathogens of sputum cultures was done. A positive match of the sputum Gram stain result with the sputum culture result was considered a successful prediction. For example, if we wanted to calculate the diagnostic performance of the sputum Gram stain in predicting GNB, we defined true positives as smears
Please cite this article as: Huang W-Y et al., Diagnostic performance of the Sputum Gram Stain in predicting sputum culture results for critically ill pediatric patients with pneumonia, Pediatrics and Neonatology, https://doi.org/10.1016/j.pedneo.2020.03.014
+
MODEL
Sputum Gram Stain in Predicting Culture Results showing GNB and cultures showing GNB; false negatives as smears showing non-GNB and cultures showing GNB; false positives as smears showing GNB and cultures showing nonGNB; and true negatives as smears showing non-GNB and cultures showing non-GNB. The above description is shown in Supplementary Fig. 1 for easier visualization. Then, the following diagnostic parameters were calculated for Gram stain predicting GNB, GNC, GPB and GPC: sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), positive likelihood ratio (LRþ), and negative likelihood ratio (LR). In addition, 403 eligible patients also received laboratory examinations for WBC and CRP levels when collecting sputum samples. Using these results, the diagnostic performance of a certain WBC or CRP value in predicting bacterial sputum culture could be calculated; namely, the sensitivity, specificity, PPV, NPV, LRþ, and LR. This study was approved by the institutional review board of Changhua Christian Hospital (Approval No. 181209). The need for informed consent was waived.
2.2. Data analysis The acquired data were analyzed using SPSS Version 10 software (SPSS, Inc., Chicago, IL, USA). Comparisons of the patient characteristics were performed using analysis of variance for continuous variables and the chi-square test for categorical variables. The sensitivity, specificity, PPV, NPV, LRþ, and LR of the sputum Gram stain in the different bacterial classifications were calculated. A receiver operating characteristic (ROC) curve was created to evaluate the possibility of using WBC and CRP levels to predict bacterial sputum cultures. Possibility levels of less than 0.05 were taken as significant. To establish this, the best clinical cut-off point was chosen and the area under the curve (AUC) and 95% confidence interval (CI) were calculated.
3. Results Of the records of 622 patients initially included in this study, those of 80 patients were excluded owing to mixed cultures. Therefore, the records for the remaining 542 patients were selected for this study. Of these patients, 122 had positive identification of bacteria in the sputum cultures and the other 420 had negative findings. Comparisons of age, gender, outcome, disease distribution and intubation days between patients with positive and negative sputum cultures are shown in Table 1. The percentages of males to females and death to survival showed no significant p values (0.106 and 0.320, respectively). Only three items, including hematology, sepsis, and trauma in disease distribution, showed significant p values (0.034, 0.035, and 0.003, respectively). The mean age of positive and negative sputum cultures is 5.32, and 5.17 years, respectively; the mean of the intubation days of positive and negative sputum cultures is 9.82 and 9.75 days, respectively. The pathogens identified in the 122 sputum cultures are listed in Table 2. The results revealed that Haemophilus influenzae, Pseudomonas aeruginosa, Streptococcus pneumoniae, Staphylococcus aureus, and Moraxella catarrhalis were the
3 Table 1 The clinical characteristics of patients with positive and negative sputum cultures. Variable
Gender Malea Femalea Outcome Deatha Survivala Disease distribution Cardiovascular diseasea Neurologya Hematologya Sepsisa Endocrinologya Traumaa Chronic lung diseasea Gastroenterologya Nephrologya Othersa >1 diagnosisa Age, year-oldb Intubation daysb a b
Positive sputum culture (n Z 122)
Negative sputum culture (n Z 420)
P value
78 (63.93) 44 (36.07)
234 (55.71) 186 (44.29)
0.106
19 (15.57) 103 (84.43)
51 (12.14) 369 (87.86)
0.320
12 (9.84)
61 (14.52)
0.182
10 (8.2) 0 (0) 38 (31.15) 0 (0) 17 (13.93) 36 (29.51)
48 (11.43) 15 (3.57) 92 (21.9) 5 (1.19) 24 (5.71) 126 (30)
0.309 0.034 0.035 0.226 0.003 0.917
0 (0) 0 (0) 9 (7.38) 0 (0) 5.32 5.52 9.82 25.69
11 (2.62) 2 (0.48) 26 (6.19) 10 (2.38) 5.17 6.13 9.75 11.87
0.071 0.445 0.639 0.085 0.818 0.966
n (%). Mean SD.
top five pathogens in the bacterial sputum cultures, with isolation rates of 23.77%, 21.31%, 16.39%, 10.66%, and 9.84%, respectively. The Gram stain successfully predicted the culture results for 321 patient samples only, where 62 cultures grew bacteria that corresponded with the Gram stain findings, and 259 cultures had no growth and thus corresponded with the
Table 2 cultures.
Classification of pathogens in 542 sputum
Organism
n (%) in all samples
n (%) in bacterial samples
No growth Haemophilus influenzae Pseudomonas aeruginosa Streptococcus pneumoniae Staphylococcus aureus Moraxella catarrhalis Acinetobacter baumannii Klebsiella pneumoniae Staphylococcus capitis MRSA Escherichia coli Others
420 (77.49) 29 (5.35) 26 (4.79) 20 (3.69) 13 (2.4) 12 (2.21) 5 (0.92) 4 (0.74) 3 (0.55) 2 (0.37) 1 (0.18) 7 (1.29)
29 (23.77) 26 (21.31) 20 (16.39) 13 (10.66) 12 (9.84) 5 (4.09) 4 (3.28) 3 (2.46) 2 (1.64) 1 (0.82) 7 (5.74)
MRSA Z methicillin-resistant Staphylococcus aureus.
Please cite this article as: Huang W-Y et al., Diagnostic performance of the Sputum Gram Stain in predicting sputum culture results for critically ill pediatric patients with pneumonia, Pediatrics and Neonatology, https://doi.org/10.1016/j.pedneo.2020.03.014
+
MODEL
4
W.-Y. Huang et al
negative Gram stain results. Thus, the overall prediction success rate of the sputum Gram stain was 59.23% for this pediatric population. Based on the original case numbers divided into subgroups, including stain þ/culture þ, stain þ/culture , stain /culture þ and stain /culture e, the predictive accuracy of the sputum Gram stain for GNB, GNC, GPB, and GPC is listed Table 3. Here, “stain “ includes both mixed stain and negative findings. The sensitivities for GNB, GNC, and GPC were 0.45, 0.67, and 0.61, respectively. The specificities and NPV for GNB, GNC, GPB, and GPC were all greater than 0.87 and 0.92, respectively. In terms of the sensitivity, specificity, PPV, and LRþ, the results for GNC were the best. Inaccuracy of Gram stain is of two major forms: false positive stain (stain þ/culture ) and false negative stain (stain /culture þ). Table 4 shows the rates of false positive smears in inaccuracy of Gram stain of GNB, GNC, GPB and GPC were 11.07%, 1.48%, 5.17% and 11.99%, respectively. In contrast, the rates of false negative smears of GNB, GNC and GPC were 7.01%, 0.74% and 2.77%, respectively. The dominant morphotypes of false positive stains are GPC and GNB, and GNB is the predominant morphotype of false negative stains. Table 5 lists some of the data on the accuracy of WBC and CRP levels in predicting a positive bacterial sputum culture. We chose WBC values of 10,000, 15,000, 20,000, 25,000, and 30,000 cells/mm3 and CRP values of 5, 10, 15, 20, 25, and 30 mg/dL for predicting positive bacterial cultures. ROC curves are shown in Supplementary Fig. 2. With regard to the accuracy of WBC levels in predicting a positive bacterial sputum culture, the AUC value and p-value were 0.585 and 0.011, respectively. The best clinical cut-off point in the ROC curve was the WBC value of 20,000 cells/mm3. For CRP, the AUC value and p-value in the ROC curve were 0.538 and 0.258, respectively.
4. Discussion Our study of a PICU population suspected of having pneumonia demonstrated the accuracy of the sputum Gram stain in predicting GNB, GNC and GPC of cultures, showing that it had high specificity and NPV. In terms of the accuracy of WBC levels in predicting a positive bacterial sputum culture, the best clinical cut-off point in the ROC curve was the WBC value of 20,000 cells/mm3.
Table 3
Classifications of inaccuracy of Gram stain.
Type of inaccuracy of Gram stain
n (%) in 542 samples
false positive stain stain þ/culture GNB GNC GPB GPC false negative stain stain /culture þ GNB GNC GPC
60 (11.07) 8 (1.48) 28 (5.17) 65 (11.99)
38 (7.01) 4 (0.74) 15 (2.77)
GNB Z gram-negative bacilli; GNC Z gram-negative cocci; GPB Z gram-positive bacilli; GPC Z gram-positive cocci.
Table 1 shows clinical characteristics of patients with positive and negative sputum cultures. Sepsis and trauma patients tend to have positive sputum cultures, while hematology patients tend to have negative sputum cultures. Bacterial pneumonia is known to be a common cause of sepsis in pediatric patients. Patients with major trauma often need long-term ventilator support and might develop ventilator-associated pneumonia (VAP) during ICU admission. Arumugam et al.7 also reported that trauma patients with head injury and chest injury have high risk of VAP. In contrast, hematology patients might have higher risk of opportunistic infection. One example is Pneumocystis jirovecii or fungus infection, which was hard to culture in the medium.8,9 In a previous pediatric study regarding washed sputum cultures, Cao et al.10 showed that the isolation rates of H. influenzae, Str. pneumoniae, M. catarrhalis, and S. aureus were 22.4%, 9.4%, 4.9%, and 0.4%, respectively. Those authors also reported that H. influenzae was the most common bacterium in cultures of sputum from children with lower respiratory tract infections.10 Our study obtained the same result and similar clinical data about the prevalence rate of H. influenzae. However, our study found P. aeruginosa to have the second highest prevalence rate. This may be due to the fact that some patients in our PICU had cerebral palsy, making it easy for P. aeruginosa to colonize their lungs. Two other research studies also found that P. aeruginosa was the common pathogen of
Predictive accuracy of the sputum Gram stain for GNB, GNC, GPB, and GPC.
Organism
GNB GNC GPB GPC
Table 4
n (%) in 542 sputum samples Stain þ Culture þ
Stain þ Culture
Stain Culture þ
Stain Culture
31 (5.72) 8 (1.48) 0 (0) 23 (4.24)
60 (11.07) 8 (1.48) 28 (5.17) 65 (11.99)
38 (7.01) 4 (0.74) 0 (0) 15 (2.77)
413 522 514 439
(76.2) (96.31) (94.83) (81)
Sn
Sp
PPV
NPV
LRþ
LR
P value
0.45 0.67 L 0.61
0.87 0.98 0.95 0.87
0.34 0.5 0 0.26
0.92 0.99 1.00 0.97
3.46 33.50 L 4.69
0.63 0.34 L 0.45
<0.0001 <0.0001 <0.0001 <0.0001
Sn Z sensitivity; Sp Z specificity; PPV Z positive predictive value; NPV Z negative predictive value; LRþ Z positive likelihood ratio; LR Z negative likelihood ratio; GNB Z gram-negative bacilli; GNC Z gram-negative cocci; GPB Z gram-positive bacilli; GPC Z grampositive cocci.
Please cite this article as: Huang W-Y et al., Diagnostic performance of the Sputum Gram Stain in predicting sputum culture results for critically ill pediatric patients with pneumonia, Pediatrics and Neonatology, https://doi.org/10.1016/j.pedneo.2020.03.014
+
MODEL
Sputum Gram Stain in Predicting Culture Results Table 5
5
Predictive accuracy of WBC and CRP for positive bacterial sputum culture.
Parameter 3
WBC(cells/mm )
CRP (mg/dL)
Cut-off values
Sn
Sp
PPV
NPV
LRþ
LR
AUC (95% CI)
P value
>10,000 >15,000 >20,000a >25,000 >30,000 >5 >10 >15 >20 >25 >30
0.69 0.43 0.23 0.12 0.07 0.29 0.22 0.14 0.09 0.07 0.06
0.40 0.69 0.84 0.94 0.96 0.72 0.84 0.90 0.92 0.97 0.99
0.26 0.29 0.30 0.37 0.35 0.25 0.31 0.33 0.28 0.44 0.60
0.81 0.80 0.78 0.78 0.77 0.76 0.77 0.76 0.76 0.76 0.76
1.15 1.39 1.40 1.94 1.77 1.05 1.41 1.48 1.20 2.39 4.61
0.77 0.83 0.92 0.94 0.97 0.98 0.92 0.95 0.98 0.96 0.95
0.585 (0.520e0.649)
0.011
0.538 (0.471e0.604)
0.258
Sn Z sensitivity; Sp Z specificity; PPV Z positive predictive value; NPV Z negative predictive value; LRþ Z positive likelihood ratio; LR Z negative likelihood ratio; AUC Z area under the curve; CI Z confidence interval; WBC Z White blood cell; CRP Z C-reactive protein. a Best clinical cut-off point.
pneumonia in critically ill pediatric patients with cerebral palsy.11,12 We calculated the predictive accuracy of the sputum Gram stain in predicting corresponding culture results with GNB, GNC, GPB, and GPC. Obviously, in our study, if the sputum Gram stain showed GNC positivity, the diagnostic performance was the best in all groups. In our study, cultures with GNC were all of M. catarrhalis and there were no cultures of Neisseria species. Sensitivity, specificity, LRþ, and LR were not affected by the prevalence, while PPV and NPV were affected. An LR þ value of greater than 10 indicates a large and conclusive probability. In our study, the LR þ value of the sputum Gram stain with GNC positivity in predicting a culture reached 33.50, and the main component of GNC in our study is M. catarrhalis. This result suggests that if the Gram stain tested positive for GNC, the culture would very likely be of M. catarrhalis in PICU. In our pediatric study, the values of sensitivity and specificity of sputum Gram stain predicting cultures, range from 0.45 to 0.67 and 0.87e0.98, respectively (Table 3).Two adult studies reported values of sensitivity and specificity of sputum Gram stain predicting cultures, ranging from 0.57 to 0.82 and 0.97e1.00, respectively.13,14 Another adult systematic review showed values of sensitivity and specificity ranging from 0.69 to 0.76 and 0.91e0.97, respectively.15 It is interesting that the diagnostic performance of Gram stain in adult studies was somewhat better than our pediatric study. In other words, the inaccuracy of the sputum. Gram stain was higher in the pediatric group. Gram stain error is a common cause of inaccuracy in sputum Gram stain, which includes specimen errors and technical errors.4,5,16e18 Gram-negative bacteria have a very thin layer of peptidoglycan, whereas gram-positive bacteria have a thick one. Specimen errors are caused by the fact that some gram-positive bacteria can become gram-negative at a certain age (hours to days), or that gram-negative bacteria can become gram-positive in older cultures.19 Technical errors occur when performing Gram stain. For example, a thick smear can decrease specificity, making a gram-negative organism appear as a gram-positive
one because of the under-decolorized smear. In contrast, a thin smear can decrease sensitivity, making a gram-positive organism appear as a gram-negative one because of the over-decolorized smear. Yunusa et al.20 also reported the above-mentioned causes of Gram stain errors. However, the proportion of Gram stain errors may be very close between pediatric and adult groups, as long as it is performed soon after sampling and follows standard procedures in our hospital. There might be some other reasons that cause inaccuracy of Gram stain in our pediatric group, for example, incidence of lung aspiration, antibiotic usage, etc. We believed that the use of uncuffed endotracheal tube and lower tension of the esophageal sphincter contribute to higher risk of lung aspiration in pediatric critical patients. If lung aspiration causes pneumonia in intubated patients, sputum Gram stain may show positive results but negative culture findings, which increases the risk of false positives. As we know, the predominant anaerobic isolates of aspiration pneumonia were GPC.21 The high incidence of GPC results in our false positive group also supports this notion. In addition, failure to predict GNB pathogens was the main cause of either false positive or false negative results in our study, and the most common GNB pathogen in our work was P. aeruginosa. Some of our patients with P. aeruginosa infection had cerebral palsy with need for longterm respiratory care, who frequently receive antibiotic treatment. We suspect frequent antibiotic usage may affect results of Gram stain. In recent years, some research studies have shown that biomarkers, such as WBCs and CRP, are effective in selecting bacterial pneumonia.22e26 We studied the correlation between WBC levels and positive bacterial sputum cultures. Although the WBC value was meaningful to a positive bacterial culture, the AUC value was not high enough. The higher the WBC value was, the higher the possibility was of obtaining a positive bacterial culture. The best clinical cut-off point in the ROC curve was the WBC value of 20,000 cells/mm3, with a sensitivity of 0.23 and a specificity of 0.84. This suggests that the sputum culture
Please cite this article as: Huang W-Y et al., Diagnostic performance of the Sputum Gram Stain in predicting sputum culture results for critically ill pediatric patients with pneumonia, Pediatrics and Neonatology, https://doi.org/10.1016/j.pedneo.2020.03.014
+
MODEL
6 test should be carried out for children with lower respiratory tract infection and a WBC value of greater than 20,000 cells/mm3. Although the correlation between CRP levels and positive bacterial sputum cultures was not good enough, the specificity in such prediction at greater than 0.9 occurred when the CRP value was greater than 15 mg/ dL. Therefore, we concluded that both WBC and CRP values can be used to predict positive bacterial sputum cultures, especially the WBC value. This study had several limitations that have to be pointed out. First, the correlation between sputum culture results and accurate diagnosis of bacterial pneumonia was uncertain. Positive sputum culture is not a definitive indicator of bacterial pneumonia. Second, absence of previous antibiotic usage data may have influenced the diagnostic performance of the Gram stain. Third, we did not analyze the association between mixed stain and mixed culture results, which was very complicated. Large-scale prospective research may be helpful in fixing these issues. In conclusion, given its high specificity and relatively low sensitivity, the sputum Gram stain can be used to predict sputum culture results in critically ill pediatric patients. The NPV for each pathogen was high, especially in detecting GNC. Our results help clinicians understand the benefits and pitfalls when using sputum Gram stain results to predict bacterial pathogens of pneumonia in PICU.
Declaration of Competing Interest The authors declare no conflicts of interest in relation to this work.
Acknowledgements None.
References 1. Murdoch DR, O’Brien KL, Scott JA, Karron RA, Bhat N, Driscoll AJ, et al. Breathing new life into pneumonia diagnostics. J Clin Microbiol 2009;47:3405e8. 2. Torres A, Lee N, Cilloniz C, Vila J, Van der Eerden M. Laboratory diagnosis of pneumonia in the molecular age. Eur Respir J 2016;48:1764e78. 3. Nagata K, Mino H, Yoshida S. Usefulness and limit of Gram staining smear examination. Rinsho Byori 2010;58:490e7 [Article in Japanese]. 4. Samuel LP, Balada-Llasat JM, Harrington A, Cavagnolo R. Multicenter assessment of Gram stain error rates. J Clin Microbiol 2016;54:1442e7. 5. Rand KH, Tillan M. Errors in interpretation of Gram stains from positive blood cultures. Am J Clin Pathol 2006;126:686e90. 6. Deirdre LC. Staining procedures, chapter 3.2. Clinical microbiology procedures handbook. 3rd ed.vol. 1. Washington DC: ASM Press; 2007. p. 1e16. 7. Arumugam SK, Mudali I, Strandvik G, El-Menyar A, Al-Hassani A, Al-Thani H. Risk factors for ventilator-associated pneumonia in trauma patients: a descriptive analysis. World J Emerg Med 2018;9:203e10. 8. White PL, Backx M, Barnes RA. Diagnosis and management of Pneumocystis jirovecii infection. Expert Rev Anti Infect Ther 2017;15:435e47.
W.-Y. Huang et al 9. Cooley L, Dendle C, Wolf J, Teh BW, Chen SC, Boutlis C, et al. Consensus guidelines for diagnosis, prophylaxis and management of Pneumocystis jirovecii pneumonia in patients with haematological and solid malignancies, 2014. Intern Med J 2014;44:1350e63. 10. Cao LD, Ishiwada N, Takeda N, Nigo Y, Aizawa J, Kuroki H, et al. Value of washed sputum Gram stain smear and culture for management of lower respiratory tract infections in children. J Infect Chemother 2004;10:31e6. 11. Thorburn K, Jardine M, Taylor N, Reilly N, Sarginson RE, van Saene HK. Antibiotic-resistant bacteria and infection in children with cerebral palsy requiring mechanical ventilation. Pediatr Crit Care Med 2009;10:222e6. 12. Morar P, Singh V, Makura Z, Jones AS, Baines PB, Selby A, et al. Oropharyngeal carriage and lower airway colonisation/infection in 45 tracheotomised children. Thorax 2002; 57:1015e20. 13. Roso ´n B, Carratala ` J, Verdaguer R, Dorca J, Manresa F, Gudiol F. Prospective study of the usefulness of sputum Gram stain in the initial approach to community-acquired pneumonia requiring hospitalization. Clin Infect Dis 2000;31:869e74. 14. Miyashita N, Shimizu H, Ouchi K, Kawasaki K, Kawai Y, Obase Y, et al. Assessment of the usefulness of sputum Gram stain and culture for diagnosis of community-acquired pneumonia requiring hospitalization. Med Sci Monit 2008;14:CR171e6. 15. Ogawa H, Kitsios GD, Iwata M, Terasawa T. Sputum gram stain for bacterial pathogen diagnosis in community-acquired pneumonia: a systematic review and Bayesian meta-analysis of diagnostic accuracy and yield. Clin Infect Dis 2019. pii: ciz876, [Epub ahead of print]. 16. Flournoy DJ. Interpreting the sputum Gram stain report. Lab Med 1998;29:763e8. 17. Becerra SC, Roy DC, Sanchez CJ, Christy RJ, Burmeister DM. An optimized staining technique for the detection of Gram positive and Gram negative bacteria within tissue. BMC Res Notes 2016;9:216. 18. Johnson MJ, Thatcher E, Cox ME. Techniques for controlling variability in gram staining of obligate anaerobes. J Clin Microbiol 1995;33:755e8. 19. Magee CM, Rodeheaver G, Edgerton MT, Edlich RF. A more reliable gram staining technic for diagnosis of surgical infections. Am J Surg 1975;130:341e6. 20. Thairu Y, Nasir IA, Usman Y. Laboratory perspective of Gram staining and its significance in investigations of infectious diseases. Sub-Saharan Afr J Med 2014;1:168e74. 21. Brook I, Finegold SM. Bacteriology of aspiration pneumonia in children. Pediatrics 1980;65:1115e20. 22. Zhydkov A, Christ-Crain M, Thomann R, Hoess C, Henzen C, Werner Z, et al. Utility of procalcitonin, C-reactive protein and white blood cells alone and in combination for the prediction of clinical outcomes in community-acquired pneumonia. Clin Chem Lab Med 2015;53:559e66. 23. Chalmers JD, Singanayagam A, Hill AT. C-reactive protein is an independent predictor of severity in community-acquired pneumonia. Am J Med 2008;121:219e25. 24. Principi N, Esposito S. Biomarkers in pediatric communityacquired pneumonia. Int J Mol Sci 2017;18. pii: E447. 25. Blasi F, Stolz D, Piffer F. Biomarkers in lower respiratory tract infections. Pulm Pharmacol Ther 2010;23:501e7. 26. Bauer S, Lamy O. Role of C-reactive protein in the diagnosis, prognosis and follow-up of community-acquired pneumonia. Rev Med Suisse 2010;6:2068e70. 2072-3, [Article in French].
Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi.org/10.1016/j.pedneo.2020.03.014.
Please cite this article as: Huang W-Y et al., Diagnostic performance of the Sputum Gram Stain in predicting sputum culture results for critically ill pediatric patients with pneumonia, Pediatrics and Neonatology, https://doi.org/10.1016/j.pedneo.2020.03.014