1. Sensitivity of each PCR test
1) For all patients diagnosed with scrub typhus
56-kDa N-PCR, 47-kDa Q-PCR, 16S rRNA C-PCR, and 16S rRNA Q-PCR were conducted using blood samples collected from 148 patients with a confirmed diagnosis of scrub typhus; the positive rates were 81.1%, 74.3%, 87.8%, and 91.9%, respectively. Of the 4 tests, 16S rRNA Q-PCR exhibited the highest sensitivity (Fig. 2). McNemar tests were conducted to compare the sensitivities of the 4 tests; 47-kDa Q-PCR and 56-kDa N-PCR did not exhibit any significant difference, whereas 16S rRNA C-PCR and 16S rRNA Q-PCR exhibited significantly higher sensitivities than 56-kDa N-PCR (P = 0.03, P < 0.01, respectively). 16S rRNA C-PCR and 16S rRNA Q-PCR had significantly higher sensitivities than 47-kDa Q-PCR (P < 0.01, P < 0.01, respectively). Moreover, the sensitivity of 16S rRNA Q-PCR was significantly greater than that of 16S rRNA C-PCR (P = 0.01).
2) According to the previous use of antibiotics
The sensitivity of each test was investigated for patients who received prior antibiotics treatment and those who had not. For the 112 patients without prior antibiotics treatment, the positive rates observed were 85.7%, 82.1%, 95.5%, and 99.1% for 56-kDa N-PCR, 47-kDa Q-PCR, 16S rRNA C-PCR, and 16S rRNA Q-PCR, respectively; 16S rRNA Q-PCR exhibited the highest sensitivity (Fig. 2). For those with prior antibiotics treatment, the positive rates obtained using 56-kDa N-PCR, 47-kDa Q-PCR, 16S rRNA C-PCR, and 16S rRNA Q-PCR were 66.7%, 50.0%, 63.9%, and 69.4%, respectively, with 16S rRNA Q-PCR exhibiting the highest sensitivity (Fig. 2). The length of pre-visit antibiotics treatment varied from 1 to 11 days in the 36 patients, and the differences in the positive rate that were observed were correlated with the length of treatment (Fig. 3).
2. Comparison of the copy number detected using 47-kDa Q-PCR and 16S rRNA Q-PCR according to the disease severity and presence of complications
Our findings confirmed that the copy number of each Q-PCR test increased with the severity of scrub typhus. To exclude the effects of antibiotics on Q-PCR, an analysis was conducted only on the 112 patients without prior antibiotics treatment. Those who had positive results on 16S rRNA Q-PCR and 47-kDa Q-PCR were divided into 4 severity groups, and the mean copy number of each severity group was compared using ANOVA. Using 16S rRNA Q-PCR, the mean copy numbers were 255.0, 218.4, 446.8, and 900.8,whereas they were 114.1, 113.2, 157.9, and 314.7 when 47-kDa Q-PCR was used; in both tests, the mean copy numbers increased significantly as the severity of scrub typhus increased (P = 0.03 and P = 0.04, respectively) (Table 1). Moreover, the 112 patients without prior antibiotics treatment were divided into the mild and severe groups and compared. The mean copy numbers from the 16S rRNA Q-PCR and 47-kDa Q-PCR results for the mild group were 234.4 and 130.5, respectively, whereas the severe group had copy numbers of 584.4 and 244.7, respectively; in both tests, the mean copy numbers were significantly greater in the severe group (P < 0.01 and P = 0.01, respectively) (Table 2). The copy numbers obtained from 47-kDa Q-PCR and 16S rRNA Q-PCR for the following complications were compared through t-tests: hemorrhage, pneumonia, hepatic malfunction, and renal malfunction. The copy numbers obtained using both16S rRNA Q-PCR and 47-kDa Q-PCR were significantly greater when pneumonia was present (P = < 0.001 and P = < 0.001, respectively) (Table 3). Furthermore, to investigate the correlation between the copy numbers obtained using 47-kDa Q-PCR and 16S rRNA Q-PCR and the various test results, Pearson's correlation coefficients were calculated. 47-kDa Q-PCR exhibited significant correlations with decreases in white blood cell count (WBC), C-reactive protein (CRP), and albumin (P < 0.01, P = 0.01, and P < 0.01, respectively), whereas 16S rRNA Q-PCR exhibited significant correlations with decreases in WBC, creatinine, and albumin (P = 0.01, P = 0.01, and P = < 0.01, respectively) (Table 4).
Table 1
Mean copy number detected by 16S rRNA Q-PCR and 47-kDa Q-PCR according to the severity grade divided into four levels.
|
16S rRNA Q-PCR
|
47-kDa Q-PCR
|
Severity
|
Number of Patients
|
Copy number
Mean ± SD
|
Number of Patients
|
Copy number
Mean ± SD
|
0
|
34
|
255.0 ± 324.6
|
24
|
114.1 ± 159.4
|
1
|
44
|
218.4 ± 203.3
|
30
|
113.2 ± 124.5
|
2
|
23
|
446.8 ± 711.7
|
13
|
157.9 ± 213.7
|
3
|
10
|
900.8 ± 1247.1
|
7
|
314.7 ± 285.3
|
P value
|
< 0.01
|
0.04
|
Table 2
Mean copy number detected using 16S rRNA Q-PCR and 47-kDa Q-PCR according to the severity grade divided into 2 levels.
|
16S rRNA Q-PCR
|
47-kDa Q-PCR
|
Severity
|
Number of Patients
|
Copy number
Mean ± SD
|
Number of Patients
|
Copy number
Mean ± SD
|
Mild
|
78
|
234.4 ± 261.9
|
65
|
130.5 ± 128.3
|
Severe
|
33
|
584.4 ± 911.4
|
28
|
224.7 ± 210.9
|
P value
|
< 0.01
|
0.01
|
Note. ‘Number of Patients’ indicates the number of positive patients detected by each PCR method for patients without prior antibiotics treatment. |
Table 3
Mean copy number detected by 16S rRNA Q-PCR and 47-kDa Q-PCR according to complication.
|
|
16S rRNA Q-PCR
|
47-kDa Q-PCR
|
|
|
Number of Patients
|
Copy number
Mean ± SD
|
Number of Patients
|
Copy number
Mean ± SD
|
Hemorrhage
|
No
|
106
|
348.3 ± 572.6
|
90
|
161.1 ± 164.6
|
|
Yes
|
5
|
129.9 ± 101.6
|
3
|
91 ± 23.6
|
|
P value
|
0.39
|
0.47
|
Pneumonia
|
No
|
99
|
269.5 ± 445.6
|
81
|
130.8 ± 123.7
|
|
Yes
|
12
|
907.1 ± 994.8
|
12
|
348 ± 254
|
|
P value
|
< 0.001
|
< 0.001
|
Renal impairment
|
No
|
101
|
307.9 ± 214.5
|
84
|
154 ± 159.9
|
|
Yes
|
10
|
647.1 ± 1119.9
|
9
|
204.5 ± 188.7
|
|
P value
|
0.07
|
0.46
|
Hepatic injury
|
No
|
8
|
567.7 ± 1642.1
|
7
|
182.4 ± 235.3
|
|
Yes
|
103
|
320.6 ± 471.6
|
86
|
156.9 ± 156.9
|
|
P value
|
0.23
|
0.79
|
Note. ‘Number of Patients’ indicates the number of positive patients detected by each PCR method for patients without prior antibiotics treatment. |
Table 4
Correlation between laboratory findings and copy number determined using 47-kDa Q-PCR and 16S rRNA Q-PCR
|
47-kDa real time PCR
|
16 s rRNA real-time PCR
|
|
Correlation coefficient
|
P value
|
Correlation coefficient
|
P value
|
WBC
|
0.35
|
< 0.01
|
0.27
|
0.01
|
Platelet
|
-0.17
|
0.09
|
-0.02
|
0.82
|
CRP
|
0.29
|
0.01
|
0.19
|
0.05
|
Creatinine
|
0.17
|
0.11
|
0.23
|
0.01
|
Albumin
|
-0.46
|
< 0.01
|
-0.36
|
< 0.01
|
Bilirubin
|
0.12
|
0.24
|
-0.01
|
0.93
|
APACHE score
|
0.20
|
0.06
|
0.14
|
0.14
|
WBC: white blood cell, CRP: C-reactive protein |
3. Comparison of PCR positive rates according to the duration of symptoms
The positive rates of 16S rRNA Q-PCR and 47-kDa Q-PCR were investigated in the 112 patients without prior antibiotics treatment according to the time between the onset of symptoms and the time of hospital visit. The 47-kDa Q-PCR results exhibited a low positive rate on day 3 after the onset of symptoms; the rate increased starting on day 4 and reached 100% between days 8 and 12. The rate decreased thereafter. 16S rRNA Q-PCR yielded negative results for only 1 patient tested on day 2 after the onset of symptoms and yielded positive results for all other patients, including the patient with the longest duration of symptoms, 21 days (Fig. 4).