TABLE 3.

Identification of S. flexneri serotype on direct stool samples that were culture positive/ipaH positive and were previously serotyped using conventional methods^a

Serotype	No. culture positive	No. detected by PCR (% sensitivity, CI)	No. of other serotypes detected (% specificity, CI)	Sequencing confirmationb	Corrected % sensitivity/% specificity
S. flexneri 2a	93	93 (100, 96–100)	1c (99, 96–100)		100, 99
S. flexneri 3a	25	22 (88, 69–97)	1 (99, 97–100)	3 (gtrV, 5b)	100, 100
S. flexneri 6	18	17 (94, 73–100)	1c (99, 97–100)	3a, 6 (oac, gtrX)	100, 99
S. flexneri 1a	6	3 (50, 12–88)	0 (100, 98–100)	1 (oac, 1b), 2 (gtrIc, 7a)	100, 100
S. flexneri 1b	19	18 (95, 74–100)	4 (97, 95–99)	1 (wzx6, 6)	100, 100
S. flexneri 2b	20	20 (100, 83–100)	2 (99, 96–100)		100, 100
S. flexneri 3b	4	0 (0, 0–60)	0 (100, 98–100)	1 (gtrV, 5a), 3 (gtrI, 1b)	—, 100
S. flexneri 4a	15	15 (100, 78–100)	0 (100, 98–100)		100, 100
S. flexneri 5a	0		(99, 98–100)		—, 100
S. flexneri 5b	2	0 (0, 0–84)	(99, 97–100)	2 (gtrII, 2b)	—, 100
S. flexneri X	9	9 (100, 66–100)	0 (100, 98–100)		100, 100
S. flexneri 7a	10	9 (90, 56–100)	2 (99, 97–100)	2a	90, 100
All	221	206/221 (93, 89–96)	13/2,431 (99, 99–100)		220/221 (99), 2,429/2,431 (99)

^aSensitivity and specificity were calculated compared to conventional serotyping. Corrected sensitivity and specificity then were recalculated, taking into account the sequencing confirmation results.

^bThis column shows the PCR results that were discrepant from culture. These samples were tested with long amplicon PCR followed by Sanger sequencing. All detections were confirmed except for 2a.

^cAll additional detections except for these two were confirmed with long amplicon PCR followed by Sanger sequencing.