False Positives and Negatives Obtained with PCR-Based Identification of Staphylococcus aureus Clonal Complex 398

LETTER

Staphylococcus aureus belonging to multilocus sequence type (MLST) 398 (ST398) and their relatives in clonal complex 398 (CC398) have emerged recently as important zoonotic agents worldwide (1). Livestock, especially swine, are implicated in carriage and transmission of these organisms, as are individuals who work closely with livestock. However, ST398 strains have also been found in people without livestock exposure (2, 3) and have caused infections within hospitals and health care facilities (4). Since ST398 isolates are pulsed-field gel electrophoresis (PFGE) nontypeable with SmaI, belong to a wide variety of staphylococcal protein A (spa) types, and show a number of different antimicrobial resistance patterns, detection of ST398 is not a quick process. van Wamel et al. created two specific PCRs for detection of ST398 using two ST398-specific DNA sequences previously obtained from amplified fragment length polymorphism (AFLP) analyses (5). A07 and CO1 primer sets were reported to show 100% accuracy when tested against 133 isolates. Noting that the van Wamel et al. sets were targeted to a transposon, Stegger et al. developed a different PCR assay using the highly conserved sau1-hsdS1 sequence in the CC398 lineage (6). PCR validation of this target showed 100% specificity (235/235 samples) and 100% sensitivity (1,072/1,072 samples).

We screened a total of 444 human colonization and 14 human infection isolates using the A07 or CO1 primer set representing the gene SAPIG2194 or SAPIG2195, respectively. All isolates were also spa typed, and a subset, including all potential false-positive and false-negative isolates, were typed by MLST. Sensitivity was 96.5% and specificity was 98.3% (Table 1). There were three false negatives that were all t571/ST398. The six false positives included two t922/ST1 isolates, one t045/ST1 isolate, one t337/ST9 isolate, one t505/ST45 isolate, and one t002/ST105 isolate (Table 2).

TABLE 1.

Summary of the numbers of true-positive, false-positive, false-negative, and true-negative results based on the A07/C01 ST98 PCR assay results, categorized by MLST results

A07/CO1 PCR result	MLST result (no. of isolates) for:
	ST398 isolates	Non-ST398 isolates
Positive	True positive (82)	False positive (6)
Negative	False negative (3)	True negative (352)

TABLE 2.

Summary of molecular characteristics of the false-negative and false positive isolates, with results from the A07/C01 PCR assay and sau1-hsdS1 PCR assay^a

Isolate	Presence of mecA	spa type	Presence of PVL gene	Antimicrobial agent(s) to which isolate was sensitive	MLST	A07/C01 PCR result	sau1-hsdS1 PCR result
83	+	t571	−	OXA, TET	ST398	−	+
142	+	t002	+	OXA, ERY, LVX, CLI (intermed.)	ST105	+	+
303	−	t505	−	None	ST45	+	+
309	−	t337	−	TET, CLI, ERY	ST9	+	+
389	−	t571	−	GEN, ERY, CLI, TET, SXT, LVX	ST398	−	+
390	−	t571	−	GEN, ERY, CLI, TET, LVX	ST398	−	+
392	−	t922	−	None	ST1	+	−
393	−	t922	−	None	ST1	+	−
437	−	t045	−	TET, SXT	ST5	+	+

^aPVL, Panton-Valentine leukocidin; OXA, oxacillin; TET, tetracycline; ERY, erythromycin; LVX, levofloxacin; CLI (intermed.), intermediate clindamycin; CLI, clindamycin; GEN, gentamycin; SXT, trimethoprim-sulfamethoxazole.

We screened the nine false positives and false negatives from the previous assay with the sau1-hsdS1 assay. Two of the false positives were negative by this assay, while the other four were still positive. All three false negatives were positive by the sau1-hsdS1 PCR assay (Table 2).

Molecular typing of Staphylococcus aureus is an incredibly important tool for surveillance and control. Both of these PCR assays show extremely high sensitivity and specificity yet still yielded false positives with our data set. Multiple tools should be used when screening for livestock-associated Staphylococcus aureus (LA-SA) to ensure diagnostic accuracy in the lab. Various spa types have been found in livestock (7–11), and it is possible that horizontal transfer of these gene fragments may be transmitted between cocolonized hosts or between host and environment (12). Therefore, while it is important to develop and use new molecular tools, gold standards, such as MLST, should still be used to identify CC398 and to understand the epidemiology of LA-SA.