Combination of 16S rRNA and procalcitonin in diagnosis of neonatal clinically suspected sepsis.

OBJECTIVE: To investigate the application of 16S rRNA in diagnosing patients with neonatal sepsis.
METHODS: We studied 60 consecutive neonatal patients with clinically suspected sepsis and 20 non-infective cases as controls. All patients were diagnosed with sepsis by clinical and experimental criteria. Clinical characteristics were recorded and 16S rRNA sequencing was conducted for all patients. The sensitivity, specificity, and accuracy of the detection methods were analyzed.
RESULTS: The detection limit of 16S rRNA sequencing was 1 × 102 CFU/mL. For suspected sepsis, the positive rate of 16S rRNA detection was 93.3%, which was similar to that of procalcitonin detection (85%), and was significantly higher than that of bacterial culture (51.7%). The specificity of procalcitonin detection (74.1%) was significantly lower than that of 16S rRNA detection (100%). Moreover, the combination of 16S rRNA and procalcitonin detection showed a sensitivity of 100%, specificity of 74.1%, and accuracy of 92.0%. For proven sepsis, the sensitivity and specificity of 16S rRNA detection were both 100.0%, and those for procalcitonin were 87.1% and 87.0%, respectively.
CONCLUSION: Detection of 16S rRNA has high sensitivity and specificity in diagnosing sepsis. The combination of 16S rRNA and procalcitonin has even better sensitivity with acceptable specificity.

Introduction

Sepsis is a life-threatening infection leading to severe organ dysfunction, which is caused by a dysregulated host response to infection.[1,2] Sepsis affects millions of people and is the most frequent cause of mortality in intensive care units worldwide.[3,4] The mortality rate for sepsis can be even higher than 25%.[5] Because of the high mortality rate and severity, early diagnosis for sepsis and appropriate management in the initial hours after sepsis are important for improving patients’ outcomes.

The widely accepted diagnostic method for sepsis is the criteria of systemic inflammatory response syndrome (SIRS), which mainly compromises temperature, heart rate, respiratory rate, and the white blood cell count.[6] During decades of development, many potential biomarkers are currently also used in diagnosing sepsis, such as C-reactive protein (CRP), interleukin (IL)-6, and procalcitonin (PCT).[7,8] However, similar to classical SIRS criteria, biomarkers such as CRP and IL-6, are too sensitive in many patients. Many non-infectious clinical processes can cause changes in these biomarkers, such as severe trauma, burns, pancreatitis, and ischemic reperfusion events.[9-11] Blood culture is an accurate way of determining sepsis, but the sensitivity is insufficient and the results often take too long to obtain. Therefore, diagnostic methods that are faster, more sensitive, and specific are required.

16S ribosomal RNA (rRNA) is effective for identifying several diseases.[12,13] However, clinical evidence for 16S rRNA in diagnosing sepsis is still rare. In the present study, we investigated application of 16S rRNA for diagnosing patients with neonatal sepsis. This study could provide more clinical evidence for application of 16S rRNA detection in diagnosis of sepsis.

Material and methods

Subjects

The present study included neonatal patients with clinically suspected sepsis who were admitted to our hospital during January 2017 to March 2018. All patients who met the inclusion criteria were consecutively enrolled during the study period. The patients were diagnosed with clinical sepsis by clinical and experimental diagnosis.[14,15] The diagnostic criteria for sepsis included the following: 1) symptoms of infection, including an unstable temperature, apnea, or decrease in heart rate; and 2) positive for at least two of the following experimental tests: total white blood cell count <5 × 106/L or >20 × 106/L), immature-to-total neutrophil ratio ≥0.16, platelet count ≤100 × 106/L, CRP levels ≥6 mg/L, and PCT levels ≥0.5 ng/mL. Blood samples were obtained within 24 hours after admission and none of the patients received any antibiotics before the study. Bacterial culture was conducted for all samples by the Automatic Blood Culture Instrument (BD BACTEC 9120; BD Diagnostics, CA, USA). Serum PCT levels were evaluated by enzyme-linked immunosorbent assay using a commercial kit (Abcam, Cambridge, MA, USA). A positive blood culture was still used as the gold standard for proven sepsis. Clinical characteristics, such as gestational age, sex, birth weight, and mode of delivery, were recorded. Additionally, non-infective neonates who were born in the hospital were randomly enrolled as controls using a computer-generated number list with the medical card number. Informed consent was obtained from the parents of all patients. The present study was approved by the Ethics Committee of the Affiliated Wuxi Maternity and Child Health Care Hospital of Nanjing Medical University.

16S rRNA sequencing

Blood DNA was extracted using the Blood DNA Mini Kit (Qiagen Inc., Valencia, CA, USA) according to the manufacturer’s instructions. For 16S rRNA sequencing, the universal primer was designed and synthesized by GeneChem Corp. (Shanghai, China). The sequences of the forward and reverse primers was as follows: 27 F, 3′-AGTTTGATCCTGGCTCAG-5′ and 1492 R, 3′-GGTTACCTTGTTACGACTT-5′. For polymerase chain reaction reaction, the products were amplified and the reaction was performed using SYBR Green Master Mix (Solarbio Science & Technology Co., Ltd., Beijing, China) in an Exicycler™ 96 (Bioneer Corporation, Daejeon, Korea). The thermocycling conditions were as follows: 94°C pre-degeneration for 5 minutes, denaturation at 94°C for 30 seconds, annealing at 56°C for 30 seconds, and extension at 72°C for 1 minute, with a total of 35 cycles, followed by a final elongation step at 72°C for 10 minutes. The sample containing all PCR reagents, except for the DNA, was used as the blank control. DNA that was extracted from pure bacterial cultures, Staphylococcus aureus and Escherichia coli (both purchased from ATCC, Rockville, Maryland, USA), was used as a positive control.

Sensitivity and specificity of 16S rRNA determination

To investigate the sensitivity and specificity of 16S rRNA for detection of bacteria, seven pure clinical common bacteria (Staphylococcus aureus, Staphylococcus haemolyticus, Pseudomonas aeruginosa, Enterococcus faecalis, Klebsiella pneumoniae, E. coli, and Salmonella enteritidis) were used for detection. A mono band with 1380 bp was obtained from all bacteria after amplification (Figure 1a). The standard E. coli (ATCC 25922) was suspended in normal saline with a concentration of 108 CFU/mL using Maxwell turbidimetry. Different concentrations of 1 × 10−2–1 × 107 CFU/mL E. coli were used for detecting 16S rRNA and the final detection limit was 1 × 102 CFU/mL (Figure 1b). For specificity analysis, hepatitis B virus DNA (provided by the Immunology Department of our hospital) and human genomic DNA extracted from healthy individuals were used as negative controls.

Figure 1.

(a) Sensitivity of 16S rRNA detection. M indicates a DNA marker. Numbers 1–10 represent Staphylococcus aureus, Staphylococcus haemolyticus, Pseudomonas aeruginosa, Enterococcus faecalis, Klebsiella pneumoniae, Escherichia coli, Salmonella enteritidis, hepatitis B virus DNA, human genomic DNA, and blank control, respectively. (b) Specificity of 16S rRNA detection. M indicates a DNA marker. Numbers 1–10 represent 1 × 107–1 × 10−2, respectively.

Statistical analysis

Measurement data are expressed as mean ± standard deviation. Chi-square analysis was used for comparison of rates. Comparisons were conducted using the Student’s t test when comparing two groups. A P value <0.05 was considered to be statistically significant. All calculations were performed using SPSS 18.0 (Chicago, IL, USA).

Results

Basic characteristics of the subjects

The present study included a total of 60 neonatal patients with clinically suspected sepsis, with a mean gestational age of 37.4 ± 3.4 weeks (male: female 39: 21). Among all patients, 38 (63.3%) were born by vaginal delivery and 22 (36.7%) were born by cesarean section. A total of four (6.7%) patients died during the study period. No significant differences were observed in basic characteristics between patients with sepsis and controls (Table 1).

Table 1.

Basic characteristics for all of the subjects.

Variables	Patients, n = 60	Control, n = 20	P value	95% CI
Gestational age, weeks (mean±SD)	37.4±3.4	38.1±3.7	0.401	0.91 (−2.57–1.04)
Sex, male: female	39: 21	14: 6	0.940	0.93 (0.14–6.38)
Body weight, kg (mean±SD)	3.4±0.7	3.3±0.6	0.911	0.17 (−0.33–0.37)
Delivery mode, n (%)			0.386	1.37 (0.75–2.32)
Vaginal delivery	38 (63.3)	17 (56.7)	–	–
Cesarean section	22 (36.7)	13 (43.3)	–	–
Mortality, n (%)	4 (6.7)	–	–	–
Onset, n (%)		–	–	–
Early	19 (31.7)	–	–	–
Late	41 (68.3)	–	–	–

SD: standard deviation; CI: confidence interval.

As mentioned above, the final detection limit for 16S rRNA was 1 × 102 CFU/mL. Specificity analysis showed that only the target bacterial DNA was detected and the negative controls and blank control were not detected (Figure 1). These results suggested the 16S rRNA method showed high sensitivity and good specificity for the target bacterial DNAs

Combination of 16S rRNA and PCT significantly enhances the sensitivity for diagnosis of sepsis

We then examined the sensitivity and specificity of diagnosis for clinical cases (Table 2). All samples received 16S rRNA detection, PCT detection, and bacterial culture. For clinically suspected sepsis, the positive rate of 16S rRNA detection was 93.3% (56/60), which was similar to that of PCT detection (85%, 51/60), and was significantly higher than that of bacterial culture (51.7%, 31/60; P < 0.01 by chi-square test) (some data not shown in Table 2). However, the specificity of PCT detection (74.1%) was significantly lower than that of 16S rRNA detection (100%, P < 0.05 by chi-square test). Interestingly, when 16S rRNA and PCT detection were combined, 16S rRNA (+)/PCT(+) detection achieved a sensitivity of 100% and specificity of 74.1% for diagnosis. For proven sepsis, the sensitivity and specificity of 16S rRNA detection were both 100.0%, and those of PCT were 87.1% and 87.0%, respectively. Among patients with proven sepsis, the etiology included Staphylococcus aureus (n = 10), Pseudomonas aeruginosa (n = 9), Enterococcus faecalis (n = 7), Klebsiella pneumoniae (n = 3), E. coli (n = 1), and Acinetobacter baumannii (1). Additionally, we selected six samples with 16S rRNA (+) and bacterial culture (−) for gene sequencing analysis. The gene sequencing results were consistent with the 16S rRNA results, with two cases of Pseudomonas aeruginosa and one case of E. coli (Figure 2).

Table 2.

Sensitivity and specificity of 16S rRNA, PCT, and blood culture for patients with clinically suspected sepsis.

Group	No. of cases	16S rRNA		PCT, ng/mL		Blood culture		16S rRNA and PCT
		(+)	(−)	≥0.5	<0.5	(+)	(−)	(+)	(−)
Suspected sepsis	60	56	4	51	9	31	29	60	0
Controls	20	0	20	7	13	0	20	7	13
Sensitivity		93.3%		87.0%		67.4%		100%
Specificity		100%		74.1%		100%		74.1%
Accuracy		95.2%		83.3%		73.4%		92.0%
Proven sepsis	31	31	0	27	4	–	–	31	0
Controls	20	0	20	3	13	–	–	3	13
Sensitivity		100%		87.1%		–		100%
Specificity		100%		87.0%		–		87.0%
Accuracy		100%		89.5%		–		96.2%

rRNA: ribosomal RNA; PCT: procalcitonin.

Figure 2.

Gene sequencing results. (a) Two cases were Pseudomonas aeruginosa. (b) One case was Escherichia coli.

Discussion

Although sepsis has been studied for many years, an accurate diagnosis for sepsis is still a challenge in the clinic. Traditional diagnostic methods for sepsis mainly include diagnosis of symptoms and blood biomarkers, or blood culture. However, all of these methods have disadvantages, which limit application of these methods.[16] Detection of 16S rRNA is effective in diagnosis of many diseases. However, evidence for application of detecting 16S rRNA in diagnosis of sepsis is still lacking. In the present study, we found that detection of 16S rRNA had high sensitivity and specificity in diagnosing sepsis, and combination of 16S rRNA and PCT provided better diagnostic results for patients with sepsis.

Many biomarkers are potential diagnostic biomarkers for sepsis. CRP and IL-6 have been used as biomarkers in diagnosis of sepsis for many years. In 2010, Celik et al. reported that the optimum cut-off value in the diagnosis of neonatal sepsis was 24.65 pg/mL for IL-6 and 4.82 mg/L for CRP.[11] Recently, a study showed that the sensitivity and specificity of CRP in diagnosis of acute neonatal sepsis was 76.92% and 53.49%, respectively.[17] Application of PCT in the diagnosis of sepsis has also been reported in many studies. Meynaar et al. showed that serum PCT levels were more useful in differentiating between sepsis and SIRS than CRP and IL-6, with better sensitivity and specificity.[18] In a meta-analysis, Tang et al. concluded that PCT cannot reliably differentiate sepsis from other non-infectious causes.[19] These authors’ conclusion supports our view that the specificity of application of simple blood biomarkers is insufficient for sepsis. In the present study, we also found that PCT was an effective biomarker in diagnosis of sepsis with good sensitivity. However, the specificity of PCT was not perfect, which is consistent with the above-mentioned previous reports.[17-19]

16S rRNA can be used in diagnosis of many diseases, including sepsis. In a brief report, El et al. showed that detection of 16S rRNA in neonatal sepsis had a sensitivity of 62.5% and specificity of 86.9%.[14] Punia et al. found that 16S rRNA was positive in all cases of blood culture proven sepsis with 95.6% sensitivity and 100% specificity.[20] A multi-year, prospective cohort study of the Alberta Sepsis Network reported that common bacterial DNA profiles in the blood of septic patients were often associated with the patients’ primary source of infection and could be further developed as a tool for clinical diagnostics for bloodstream infections.[21] In our study, we found that 16S rRNA showed high sensitivity and specificity in diagnosis of sepsis for neonates. Moreover, the combination of 16S rRNA and PCT had better sensitivity than each alone, with acceptable specificity.

The present study has some limitations. These limitations include the small sample size and single center. Therefore, more studies are required to confirm our results.

In conclusion, detection of 16S rRNA has high sensitivity and specificity in diagnosis of sepsis. Additionally the combination of 16S rRNA and PCT has even better sensitivity than 16S rRNA or PCT alone, with acceptable specificity. This study provides more clinical evidence for application of 16S rRNA detection in the diagnosis of sepsis.