Table 3. Comparative acute-phase diagnostic likelihood ratios and post-test probabilities for accurately detecting human leptospirosis by DPP assay using finger stick blood (FSB), venous whole blood (VWB), and serum, by serum IgM-ELISA, and by clinical impression for 96 patients–Salvador, Brazil 2012.
| Test/Specimen type by patient group | Positive LR (95% CI) |
Positive Post-TP, by level of Pre-TP |
Negative LR (95% CI) |
Negative Post-TP, by level of Pre-TP |
||||
|---|---|---|---|---|---|---|---|---|
| L | M | H | L | M | H | |||
| Classic leptospirosis | ||||||||
| DPP/FSB | 4.73 (2.61–8.59) | 54.2 | 82.5 | 95.0 | 0.09 (0.02–0.35) | 2.2 | 8.3 | 26.5 |
| DPP/VWB | 3.77 (2.30–6.18) | 48.5 | 79.0 | 93.8 | 0.05 (0.05–0.34) | 1.2 | 4.8 | 16.7 |
| DPP/Serum | 6.67 (3.11–14.32) | 62.5 | 87.0 | 96.4 | 0.17 (0.07–0.42) | 4.1 | 14.5 | 40.5 |
| IgM-ELISA/Serum | 37.15 *(5.30–260.52) | 90.3 | 97.4 | 99.3 | 0.20 (0.09–0.43) | 4.7 | 16.5 | 44.1 |
| Clinical impression | 2.26 (1.61–3.18) | 36.1 | 69.3 | 90.0 | 0.06 (0.01–0.45) | 2.5 | 5.7 | 19.4 |
| Meningitis | ||||||||
| DPP/FSB | 2.27 (0.51–10.01) | 36.2 | 69.4 | 90.1 | 0.73 (0.34–1.54) | 15.4 | 42.2 | 74.5 |
| DPP/VWB | 6.80 (1.72–26.86) | 63.0 | 87.2 | 96.5 | 0.23 (0.04–1.32) | 5.4 | 18.7 | 47.9 |
| DPP/Serum | 9.60 (1.26–72.96) | 70.6 | 90.6 | 97.5 | 0.43 (0.15–1.26) | 9.7 | 30.1 | 63.2 |
| IgM-ELISA/Serum | 13.60 (1.93–95.71) | 77.3 | 93.2 | 98.2 | 0.21 (0.04–1.23) | 5.0 | 17.4 | 45.7 |
DPP = Dual Path Platform; CI = confidence interval; LR = likelihood ratio; Post-TP = post-test probability; VWB = venous whole blood; ELISA = enzyme-linked immunosorbent assay; Pre-TP = pre-test probability; the clinical Pre-TP of leptospirosis was designated as low, moderate, or high; L = low clinical Pre-TP of leptospirosis (20%); M = moderate clinical Pre-TP (50%); H = high clinical Pre-TP (80%); NA = not applicable; Undefined = contains operation divisible by zero. Values not calculated for clinical impression among meningitis patients.
*Positive LR underestimates the true value, as 45 was used as number of true negatives and one was used as false negative in order to avoid division by zero.