- About this Journal ·
- Abstracting and Indexing ·
- Advance Access ·
- Aims and Scope ·
- Annual Issues ·
- Article Processing Charges ·
- Articles in Press ·
- Author Guidelines ·
- Bibliographic Information ·
- Citations to this Journal ·
- Contact Information ·
- Editorial Board ·
- Editorial Workflow ·
- Free eTOC Alerts ·
- Publication Ethics ·
- Reviewers Acknowledgment ·
- Submit a Manuscript ·
- Subscription Information ·
- Table of Contents
The Scientific World Journal
Volume 2014 (2014), Article ID 471463, 14 pages
Tumor Necrosis Factor-α as a Diagnostic Marker for Neonatal Sepsis: A Meta-Analysis
1Systems Sepsis Team, Soochow University Affiliated Children’s Hospital, Suzhou 215003, China
2Center for Systems Biology, Soochow University, Suzhou 215006, China
3Suzhou Zhengxing Translational Biomedical Informatics Ltd., Taicang 215400, China
Received 15 August 2013; Accepted 3 October 2013; Published 11 February 2014
Academic Editors: B. Shen, J. Wang, and J. Wang
Copyright © 2014 Bokun Lv et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Neonatal sepsis (NS) is an important cause of mortality in newborns and life-threatening disorder in infants. The meta-analysis was performed to investigate the diagnosis value of tumor necrosis factor-α (TNF-α) test in NS. Our collectible studies were searched from PUBMED, EMBASE, and the Cochrane Library between March 1994 and August 2013. Accordingly, 347 studies were collected totally, in which 15 articles and 23 trials were selected to study the NS in our meta-analysis. The TNF-α test showed moderate accuracy of the diagnosis of NS both in early-onset neonatal sepsis (sensitivity = 0.66, specificity = 0.76, Q* = 0.74) and in late-onset neonatal sepsis (sensitivity = 0.68, specificity = 0.89, Q* = 0.87). We also found the northern hemisphere group in the test has higher sensitivity (0.84) and specificity (0.83). A diagnostic OR analysis found that the study population may be the major reason for the heterogeneity. Accordingly, we suggest that TNF-α is also a valuable marker in the diagnosis of NS.
Neonatal sepsis (NS), as one of the major causes, leads to neonatal mortality and morbidity, especially in neonates born preterm . Early diagnosis and management of the newborn infant with NS play key roles in preventing severe and life-threatening complications . During the first hours of life, reliable infection markers are absent in NS. Therefore, neonatologists often begin early antibiotic treatment in newborn infants with risk factors for infection, exposing many neonates to unnecessary treatments. Due to the limitation of the treatment strategy in early diagnosis of sepsis, the isolation of causative organisms from microbiological cultures takes up to 72 h, which cannot be used to identify most infected infants .
Procalcitonin (PCT), a 116-amino acid peptide considered as a precursor in calcium homeostasis, has been proved to be a valuable marker for distinguishing sepsis from other noninfectious disease [4, 5]. C-reactive protein (CRP), another biomarker, is a protein discovered in the blood and whose levels will rise after inflammation [6–8]. To date, CRP and PCT have been proposed for inclusion as two mostly used diagnostic markers in the international definition of sepsis [9, 10]. Recently, many studies have found that some new markers also play important roles in the diagnosis of neonatal sepsis. However, the systematic research and comparison of these biomarkers for diagnosing NS are limited. For example, we have investigated the diagnosis value of serum amyloid A (SAA) in NS . Here, we continue to evaluate the value of the tumor necrosis factor-α (TNF-α), by considering it as a useful marker. TNF-α is a cytokine involved in systemic inflammation, which belongs to a member of a group of cytokines that stimulate the acute phase reaction .
Thus, the objective of this meta-analysis was to investigate the value of TNF-α for detecting NS. Although a lot of works indicate that both PCT and CRP are two superior markers for diagnosis of sepsis and infection, we suggest here the TNF-α is also a promising marker in NS. A deeper meta-analysis of these studies is thus currently needed.
2.1. Studies Retrieval and Selection
In order to perform a systematic analysis of the available evidence on the efficacy of TNF-α in NS , the common approach of literature search was performed in PUBMED, EMBASE, and the Cochrane Library for relevant citations from March 1994 to August 2013. The search terms used were “TNF-α,” “neonatal,” “neonate,” “sepsis,” “infant,” “newborn,” and “tumor necrosis factors-α”. The reference lists of all known primary and review articles were also searched. No language restriction was used, so that we have examined the references of known articles to fully retrieve the data.
If an article does not include enough data for calculating sensitivity and specificity (2 × 2 table), we asked the corresponding author to provide us with necessary data. If there was no response from the corresponding author, a reminder was sent after one week. If we still cannot achieve the data after this process, the study was excluded from meta-analysis. The selection of articles was performed by two investigators independently to ensure the high accuracy.
2.2. Data Extraction
Data collected from the studies included the first author, publication year, diagnostic cut-off point and time, test methods, and sensitivity and specificity data. So the numbers of true-positive, false-positive, false-negative, and true-negative results were extracted for each study. Accuracy data was extracted to construct 2 × 2 table at a specific time. We have requested the information from the authors, if no enough data on the criteria was found in the studies.
2.3. Statistical Analysis
We used Meta-Disc 1.4 software and Review Manager 5.0 to perform the statistical analysis . Diagnoses were grouped into two groups according to the time of TNF-α test for diagnosis of NS. One group is the time points of TNF-α measurement for the diagnosis of early-onset neonatal sepsis (EONS), which were 0–72 h of age; the other group is the time points of TNF-α measurement for the diagnosis of late-onset neonatal sepsis (LONS), in which the age of neonates is older than 72 h. We calculated the sensitivity, specificity, diagnostic odds ratio (OR), and corresponding 95% confidence intervals (CI) from each study. We also gained the pooled sensitivity, specificity, and diagnosis OR from each group.
The diagnostic OR expresses how much greater the odds of having sepsis are for newborns who have a positive test result, relatively to newborns who have a negative result . For the estimates of diagnostic OR, heterogeneity was assessed by using the Cochrane Q statistic. Normally, lies between 0% and 100%. If , then there is a lot of homogeneity among studies in meta-analysis; whereas shows there is more heterogeneity among studies. A value of 0% indicates no observed heterogeneity, and larger values show increasing heterogeneity . We explored the reasons for heterogeneity by carrying out the subgroup analysis and examined characteristics of included studies.
In order to summarize these results, we constructed summary receiver operator characteristic (SROC) curves, which showed the relationship between sensitivity and the false positives (1-specificity). values was received from the SROC curves. Meanwhile, the area under the (SROC) curves was also calculated from the SROC curves, which have been proposed as a way to assess diagnostic data in the context of a meta-analysis .
3.1. Study Selection
The literature search was completed in August 2013. We found 347 potentially relevant restudies, but only 15 articles met our inclusion criteria. Figure 1 shows the chart of literature search. Detailed information for each included study is presented in Table 1.
3.2. Accuracy of the TNF-α Test in the Diagnosis of Proven Early-Onset Neonatal Sepsis
Eleven articles and twelve trials were included to estimate the use of the TNF-α test in the diagnosis of proven early-onset neonatal sepsis (EONS). EONS was defined as the clinical sepsis in the 0–72 h after delivery, met the inclusion criteria in [18–27].
In these trials, we can get the TP, TN, FP, FN, sensitivity, specificity, DOR, PPV, and NPV from the articles. The sensitivity ranged from 20.8% to 100% and pooled sensitivity is 66.1% (95% CI 60.7%–70.1%), specificity ranged from 43.1% to 100% and pooled specificity is 75.6% (95% CI 72.2%–78.9%), and the detailed forest map is shown in Figure 2. We calculated the significant heterogeneity among studies (sensitivity, ; specificity, ); it indicated that patient selection or other relevant factors might be responsible for heterogeneity.
The value of DOR of TNF-α was 7.43 (95%CI 3.47–15.90), as shown in Figure 3. In these articles, we calculated the significant heterogeneity (). The SROC curve for TNF-α markers was plotted in Figure 4; the AUC was 0.81 with the standard error being 0.04. The pooled diagnostic accuracy () of 0.7430 with the standard error was 0.04.
3.3. Accuracy of the TNF-α Test in the Diagnosis of Proven Late-Onset Neonatal Sepsis
Six articles and eleven trials were included to estimate the use of the TNF-α test in the diagnosis of proven late-onset neonatal sepsis (LONS). LONS was defined as the clinical sepsis 72 h after birth, met the inclusion criteria in [1, 14, 24, 28–30].
In these trials, we can also get much information from the articles. The sensitivity ranged from 23.1% to 100% and pooled sensitivity is 68.0% (95% CI 62.8%–72.8%), specificity ranged from 73.2% to 100% and pooled specificity is 88.5% (95% CI 85.9%–90.7%), and the detailed forest map is shown in Figure 5. We calculated the significant heterogeneity among studies (sensitivity, ; specificity, ), which indicated that patient selection or other relevant factors might be responsible for heterogeneity.
The value of DOR of TNF-α was 37.44 (95% CI 19.07–73.48), as shown in Figure 6. In these articles, we calculated the significant heterogeneity (). The SROC curves for TNF-α markers were plotted in Figure 7; the AUC was 0.93 with the standard error being 0.017. The pooled diagnostic accuracy () of 0.8696 with the standard error was 0.02.
3.4. Intensive Study of the TNF-α Test in the Diagnosis of Proven Late-Onset Neonatal Sepsis
In the LONS study, study populations come from different countries, but in general they can be further divided into two regions: the northern hemisphere and the southern hemisphere.
Eight trials were included to estimate the use of the TNF-α test in the northern hemisphere at the diagnosis of proven late-onset neonatal sepsis [1, 14, 29, 30]. In these trials, sensitivity ranged from 61.4% to 100% and pooled sensitivity is 84.0% (95% CI 78.8%–88.4%), specificity ranged from 68.8% to 96.6% and pooled specificity is 83.3% (95% CI 79.6%–86.6%), and the detailed forest maps are shown in Figure 8. We calculated the significant heterogeneity among studies (sensitivity, ; specificity, ), which indicated that patient selection or other relevant factors might be responsible for heterogeneity.
The value of DOR of TNF-α test in the northern hemisphere at the diagnosis of proven late-onset neonatal sepsis was 44.94 (95% CI 20.71–97.50), as shown in Figure 9. In these articles, we calculated the significant heterogeneity (). The SROC curves for TNF-α markers were plotted in Figure 10; the AUC was 0.93 with the standard error being 0.017. The pooled diagnostic accuracy () of 0.8710 with the standard error was 0.02.
Three trials were included to estimate the use of the TNF-α test in the southern hemisphere at the diagnosis of proven late-onset neonatal sepsis [24, 28]. In these trials, sensitivity ranged from 23.1% to 35% and pooled sensitivity is 32.1% (95% CI 23.5%–41.7%), specificity ranged from 97.1% to 100% and pooled specificity is 98.3% (95% CI 95.8%–99.5%), and the detailed forest map is shown in Figure 11. We calculated the significant heterogeneity among studies (sensitivity, ; specificity, ), which indicated that patient selection or other relevant factors might be responsible for heterogeneity.
The value of DOR of TNF-α test in the northern hemisphere at the diagnosis of proven late-onset neonatal sepsis was 20.88 (95% CI 3.84–113.49), as shown in Figure 12. In these articles, we calculated the significant heterogeneity (). The SROC curves for TNF-α markers were plotted in Figure 13; the AUC was 0.0468 with the standard error being 0.224. The pooled diagnostic accuracy () of 0.1052 with the standard error was 0.309.
3.5. Comparison of the Diagnostic Accuracy of Markers for Neonatal Sepsis
In LONS, PCT and CRP have been proven to be useful markers for the diagnosis of NS [3, 9, 31]. In order to show the value of diagnosis of the TNF-α test for NS in LONS, we compared TNF-α with PCT and CRP in LONS. Six articles and eleven trials were used to evaluate the diagnosis of TNF-α. Compared with 55% (95% CI 45%–65%) for the CRP test and 72% (95% CI 63%–81%) for the PCT test , the pooled sensitivity for the TNF-α test was 68% (95% CI 63%–73%).
The pooled specificity for the TNF-α was slightly higher than for the CRP and PCT test (88.5% (95% CI 86%–91%) versus 85% (95% CI 81%–88%) versus 77% (95% CI 72%–81%)). Furthermore, the pooled diagnostic OR for the TNF-α was higher than CRP and PCT (37.4 (95% CI 19.1–73.5) versus 8.6 (95% CI 3.5–21.0) versus 11.6 (95% CI 5.2–26.0)). The value was slightly higher for the TNF-α than CRP and PCT (0.87 versus 0.75 versus 0.77). In SROC curve the TNF-α’s AUC is almost equal to CRP (0.93 versus 0.96). As many articles reported that PCT and CRP are good markers, the TNF-α is also a good marker for the diagnosis of NS in LONS.
3.6. Analysis of Heterogeneity
Heterogeneity is very critical in a meta-analysis, so we should try to explore the reason for the heterogeneity. Generally speaking, variations include several influence factors, for instance the cut-off value, study population, inject antibodies, and so forth.
Firstly, we consider the cut-off value. In our study, many data are so large, so we suppose that this is the possible reason for the heterogeneity. We excluded two studies of [24, 28] whose cut-off values were relatively large. But the (heterogeneity test) almost does not change (41.7% versus 41.6%). So the cut-off value is not the reason for the heterogeneity. Secondly, we consider the study population. The is reduced from 41.6% to 0% when two studies of [28, 29] were excluded from the meta-analysis. Their study population is different from others. So the study population may be the major reason for the heterogeneity.
3.7. Publication Bias
The publication bias is difficult to avoid in meta-analysis . In each study, we often choose favorable results and give up negative results; many “blindness” of test results was never reported. The limit of current available data may bias our conclusion.
NS is one of the most common diseases and life-threating disorder in neonate, and thus it can bring the high mortality and morbidity in infants. So the identification of biomarkers is very important to improve the diagnosis of NS. The clinical signs are nonspecific and laboratory indicators including blood culture are not reliable . The sensitivities of markers are not always so high . So it is necessary to find a good marker for NS.
It is well known that an excellent marker should have high sensitivity and specificity. In our meta-analysis, the TNF-α tests’ sensitivity is 0.66 for the diagnosis of early-onset neonatal sepsis, and the specificity is 0.46 and the is 0.74. At the late-onset neonatal sepsis, the TNF-α test’s sensitivity is 0.68, whereas the specificity is 0.89 and the is 0.87. In particular, TNF-α shows a higher accuracy for the diagnosis of NS in LONS. Therefore, we have further analyzed the regional issues in LONS. To this end, we have classified studies into two groups, that is, northern hemisphere group [1, 14, 29, 30] and southern hemisphere group [24, 28]. Our analysis found that the northern hemisphere group has higher sensitivity and specificity (sensitivity = 84%, specificity = 83%). The results show that the TNF-α has appropriate accuracy for the diagnosis of NS and thus is a good biomarker for the diagnosis of NS.
CRP is an excellent marker and has been applied in clinic . The sensitivity of CRP is 30%–97%, and the specificity ranged from 75% to 100% . In our meta-analysis, TNF-α’ sensitivity ranged from 23.1% to 100%, and specificity ranged from 73.2% to 100%. The study from TNF-α is similar to CRP, so the TNF-α is a useful marker in the diagnosis of NS. In addition, PCT is more excellent marker which has better accuracy than CRP for the diagnosis of NS . In our meta-analysis, the pooled sensitivity of TNF-α is slightly lower than that of the PCT test in EONS (66.1% versus 74%), the pooled specificity also in such (76% versus 86%). But in LONS, the pooled specificity of TNF-α is higher than that of the PCT test (89% versus 77%), although the pooled sensitivity is slightly lower than PCT (68% versus 72%). Generally speaking, the data of TNF-α and PCT is greater than CRP. This result again shows that TNF-α is good marker in the diagnosis of NS.
In conclusion, TNF-α shows the moderate accuracy in the diagnosis of NS, both in EONS and LONS. If we test the accuracy of TNF-α by further dividing data into two regions, the study in northern hemisphere shows a better result. Because of the relatively few testing data, the experiments results need to be further studied and the clinical validation is also needed.
Conflict of Interests
The authors declare that there is no conflict of interests regarding the publication of this paper.
Bokun Lv and Jie Huang equally contributed to this work.
This work was supported by the National Natural Science Foundation of China Grants (81272143 and 21203131), the Natural Science Foundation of Jiangsu Province (K200509), the Natural Science Foundation of the Jiangsu Higher Education Institutions of China (12KJB180014), Jiangsu Innovation Team Grant LJ201141, and Program of Innovative and Entrepreneurial Talent.
- E. Hotoura, V. Giapros, A. Kostoula, P. Spyrou, and S. Andronikou, “Pre-inflammatory mediators and lymphocyte subpopulations in preterm neonates with sepsis,” Inflammation, vol. 35, no. 3, pp. 1094–1101, 2012.
- U. K. Mishra, S. E. Jacobs, L. W. Doyle, and S. M. Garland, “Newer approaches to the diagnosis of early onset neonatal sepsis,” Archives of Disease in Childhood, vol. 91, no. 3, pp. F208–F212, 2006.
- Z. Yu, J. Liu, Q. Sun, Y. Qiu, S. Han, and X. Guo, “The accuracy of the procalcitonin test for the diagnosis of neonatal sepsis: a meta-analysis,” Scandinavian Journal of Infectious Diseases, vol. 42, no. 10, pp. 723–733, 2010.
- B. Uzzan, R. Cohen, P. Nicolas, M. Cucherat, and G.-Y. Perret, “Procalcitonin as a diagnostic test for sepsis in critically ill adults and after surgery or trauma: a systematic review and meta-analysis,” Critical Care Medicine, vol. 34, no. 7, pp. 1996–2003, 2006.
- B. Müller, M. Christ-Crain, E. S. Nylen, R. Snider, and K. L. Becker, “Limits to the use of the procalcitonin level as a diagnostic marker,” Clinical Infectious Diseases, vol. 39, no. 12, pp. 1867–1868, 2004.
- E. Bilavsky, H. Yarden-Bilavsky, S. Ashkenazi, and J. Amir, “C-reactive protein as a marker of serious bacterial infections in hospitalized febrile infants,” Acta Paediatrica, vol. 98, no. 11, pp. 1776–1780, 2009.
- A. Kordek, M. Hałasa, and W. Podraza, “Early detection of an early onset infection in the neonate based on measurements of procalcitonin and C-reactive protein concentrations in cord blood,” Clinical Chemistry and Laboratory Medicine, vol. 46, no. 8, pp. 1143–1148, 2008.
- L. Simon, F. Gauvin, D. K. Amre, P. Saint-Louis, and J. Lacroix, “Serum procalcitonin and C-reactive protein levels as markers of bacterial infection: a systematic review and meta-analysis,” Clinical Infectious Diseases, vol. 39, no. 2, pp. 206–217, 2004.
- J. P. S. Caldas, S. T. M. Marba, M. H. S. L. Blotta, R. Calil, S. S. Morais, and R. T. D. Oliveira, “Accuracy of white blood cell count, C-reactive protein, interleukin-6 and tumor necrosis factor alpha for diagnosing late neonatal sepsis,” Jornal de Pediatria, vol. 84, no. 6, pp. 536–542, 2008.
- M. M. Levy, M. P. Fink, J. C. Marshall et al., “2001 SCCM/ESICM/ACCP/ATS/SIS international sepsis definitions conference,” Intensive Care Medicine, vol. 29, no. 4, pp. 530–538, 2003.
- H. Yuan, J. Huang, B. Lv et al., “Diagnosis value of the serum amyloid a test in neonatal sepsis: a meta-analysis,” BioMed Research International, vol. 2013, Article ID 520294, 9 pages, 2013.
- L. P. Francesconi, K. M. Ceresér, R. Mascarenhas, L. Stertz, C. S. Gama, and P. Belmonte-de-Abreu, “Increased annexin-V and decreased TNF-alpha serum levels in chronic-medicated patients with schizophrenia,” Neuroscience Letters, vol. 502, no. 3, pp. 143–146, 2011.
- W. M. Bernardo, F. T. Aires, R. M. Carneiro, F. P. Sá, V. E. Rullo, and D. A. Burns, “Effectiveness of probiotics in the prophylaxis of necrotizing enterocolitis in preterm neonates: a systematic review and meta-analysis,” Jornal de Pediatria, vol. 89, no. 1, pp. 18–24, 2013.
- E. Kocabaş, A. Sarikçioǧlu, N. Aksaray, G. Seydaoǧlu, Y. Seyhun, and A. Yaman, “Role of procalcitonin, C-reactive protein, interleukin-6, interleukin-8 and tumor necrosis factor-α in the diagnosis of neonatal sepsis,” Turkish Journal of Pediatrics, vol. 49, no. 1, pp. 7–20, 2007.
- A. S. Glas, J. G. Lijmer, M. H. Prins, G. J. Bonsel, and P. M. M. Bossuyt, “The diagnostic odds ratio: a single indicator of test performance,” Journal of Clinical Epidemiology, vol. 56, no. 11, pp. 1129–1135, 2003.
- J. P. T. Higgins, S. G. Thompson, J. J. Deeks, and D. G. Altman, “Measuring inconsistency in meta-analyses,” British Medical Journal, vol. 327, no. 7414, pp. 557–560, 2003.
- S. D. Walter, “Properties of the Summary Receiver Operating Characteristic (SROC) curve for diagnostic test data,” Statistics in Medicine, vol. 21, no. 9, pp. 1237–1256, 2002.
- N. M. Fida, J. A. Al-Mughales, and M. F. Fadellah, “Serum concentrations of interleukin-1 alpha, interleukin-6 and tumor necrosis factor-alpha in neonatal sepsis and meningitis,” Saudi Medical Journal, vol. 27, no. 10, pp. 1508–1514, 2006.
- K. H. Park, B. H. Yoon, S.-S. Shim, J. K. Jun, and H. C. Syn, “Amniotic fluid tumor necrosis factor-alpha is a marker for the prediction of early-onset neonatal sepsis in preterm labor,” Gynecologic and Obstetric Investigation, vol. 58, no. 2, pp. 84–90, 2004.
- E. Layseca-Espinosa, L. F. Pérez-González, A. Torres-Montes et al., “Expression of CD64 as a potential marker of neonatal sepsis,” Pediatric Allergy and Immunology, vol. 13, no. 5, pp. 319–327, 2002.
- H. Martin, B. Olander, and M. Norman, “Reactive hyperemia and interleukin 6, interleukin 8, and tumor necrosis factor-alpha in the diagnosis of early-onset neonatal sepsis,” Pediatrics, vol. 108, no. 4, p. E61, 2001.
- R. Berner, B. Tüxen, A. Clad, J. Forster, and M. Brandis, “Elevated gene expression of interleukin-8 in cord blood is a sensitive marker for neonatal infection,” European Journal of Pediatrics, vol. 159, no. 3, pp. 205–210, 2000.
- R. C. Silveira and R. S. Procianoy, “Evaluation of interleukin-6, tumour necrosis factor-α and interleukin- 1β for early diagnosis of neonatal sepsis,” Acta Paediatrica, vol. 88, no. 6, pp. 647–650, 1999.
- U. Hallwirth, G. Pomberger, D. Zaknun et al., “Monocyte phagocytosis as a reliable parameter for predicting early-onset sepsis in very low birthweight infants,” Early Human Development, vol. 67, no. 1-2, pp. 1–9, 2002.
- R. S. Procianoy and R. C. Silveira, “Association between high cytokine levels with white matter injury in preterm infants with sepsis,” Pediatric Critical Care Medicine, vol. 13, no. 2, pp. 183–187, 2012.
- Y. Shi, C. Shen, J. Wang, H. Li, S. Qin, and R. Liu, “Role of tumor necrosis factor in neonatal sepsis,” Chinese Medical Sciences Journal, vol. 9, no. 1, pp. 45–48, 1994.
- P. C. Ng, K. Li, K. M. Chui et al., “IP-10 is an early diagnostic marker for identification of late-onset bacterial infection in preterm infants,” Pediatric Research, vol. 61, no. 1, pp. 93–98, 2007.
- C. Sherwin, R. Broadbent, S. Young et al., “Utility of interleukin-12 and interleukin-10 in comparison with other cytokines and acute-phase reactants in the diagnosis of neonatal sepsis,” American Journal of Perinatology, vol. 25, no. 10, pp. 629–636, 2008.
- P. C. Ng, S. H. Cheng, K. M. Chui et al., “Diagnosis of late onset neonatal sepsis with cytokines, adhesion molecule, and C-reactive protein in preterm very low birthweight infants,” Archives of Disease in Childhood, vol. 77, no. 3, pp. F221–F227, 1997.
- E. Hotoura, V. Giapros, A. Kostoula, P. Spirou, and S. Andronikou, “Tracking changes of lymphocyte subsets and pre-inflammatory mediators in full-term neonates with suspected or documented infection,” Scandinavian Journal of Immunology, vol. 73, no. 3, pp. 250–255, 2011.
- B. S. Naher, M. A. Mannan, K. Noor, and M. Shahiddullah, “Role of serum procalcitonin and C-reactive protein in the diagnosis of neonatal sepsis,” Bangladesh Medical Research Council Bulletin, vol. 37, no. 2, pp. 40–46, 2011.
- T. B. Knudsen, T. B. Kristiansen, L. Simon et al., “Issues pertaining to data extraction and classification and publication bias in meta-analysis of the diagnostic accuracy of markers for bacterial infection,” Clinical Infectious Diseases, vol. 40, no. 9, pp. 1372–1374, 2005.
- J. S. Gerdes, “Diagnosis and management of bacterial infections in the neonate,” Pediatric Clinics of North America, vol. 51, no. 4, pp. 939–959, 2004.
- C. Chiesa, A. Panero, J. F. Osborn, A. F. Simonetti, and L. Pacifico, “Diagnosis of neonatal sepsis: a clinical and laboratory challenge,” Clinical Chemistry, vol. 50, no. 2, pp. 279–287, 2004.
- T. Chan and F. Gu, “Early diagnosis of sepsis using serum biomarkers,” Expert Review of Molecular Diagnostics, vol. 11, no. 5, pp. 487–496, 2011.