Heterogeneous vancomycin-intermediate among methicillin resistant Staphylococcus aureus

Abstract

Background

Hetero-resistance vancomycin intermediate Staphylococcus aureus (hVISA) is phenotype, which on in-vitro susceptibility test is vancomycin susceptible (VSSA) but has a minority population of vancomycin intermediate (VISA). hVISA is responsible for vancomycin treatment failure. Population Analysis Profile- Area under Curve (PAP-AUC) is a test for detection of hVISA; however, this test is unsuitable for clinical microbiology laboratory. Tests, such as Brain Heart Infusion Agar with 6 μg/ml vancomycin (BHIA6V), E test and Macromethod E Test (MET) are available; however reported to have variable results.

Methods

58 clinical isolates of Methicillin resistant S aureus (MRSA) having MIC of vancomycin more than 1 μg/ml by E test and agar dilution were analyzed by PAP-AUC, BHIA6V and MET.

Result

The prevalence of hVISA was 6.9%. hVISA isolates were having vancomycin E test MIC >2 μg/ml. Sensitivity of BHIA6V, MET and E test with MIC >2 μg/ml were 0.75, 0.67 and 1.0 respectively; however, positive predictive values (PPV) were 0.43, 0.4 and 0.27 respectively with PAP-AUC. PAP-AUC ratio correlated with MIC by E test and MET.

Conclusions

There is need for screening MRSA isolates showing in-vitro vancomycin susceptibility ≤2 μg/ml by agar dilution method for detection of hVISA. PAP-AUC test is unsuitable for routine laboratory testing. BHIA6V, MET and E test can be used for screening, however have low PPV.

Introduction

Vancomycin is treatment of choice for infection caused by Methicillin Resistant Staphylococcus aureus (MRSA).¹ In screening vancomycin susceptibility, the Clinical and Laboratory Standard Institute (CLSI) in 2006 has redefined S. aureus strains as vancomycin-susceptible (VSSA), vancomycin-intermediate (VISA), vancomycin-resistant (VRSA) having vancomycin MIC (Minimum Inhibitory Concentration) as ≤2, 4–8 and ≥16 μg/ml respectively by micro-dilution method.² There has been interest in use of in-vitro vancomycin MIC results to predict the outcome in patients with serious S. aureus infections being treated with vancomycin.^3,4 These studies demonstrated that infection with VSSA isolates can have vancomycin treatment failure. Later, this phenomenon was demonstrated due to hetero-resistant vancomycin intermediate S. aureus (hVISA).¹ hVISA is defined as S. aureus isolates having in-vitro susceptibility test results within vancomycin susceptible range but having proportion of population in the vancomycin-intermediate range. The resistant population is present at frequency of ≤10⁵–10⁶. As CLSI methods use inoculum of 5 × 10⁴, they are not suitable to detect hVISA.¹ Population Analysis Profile- Area under Curve (PAP-AUC) as described by Wootton, et al⁵ is a reference method of detection of hVISA. However, this method is technically demanding and not suitable for use in clinical microbiology laboratory settings. Various methods, such as Brain Heart Infusion Agar with 6 μg/ml vancomycin (BHIA6V) screening, vancomycin E test, and Macromethod E Test (MET) have been recommended; however, there is no unanimity over the vancomycin testing strategies for hVISA^1,6. hVISA may be a precursor of VISA and clinically associated with treatment failure. Various studies have reported hVISA prevalence of 0–50% from clinical isolates.^1,7,8 Infections due to hVISA are associated with longer duration of bacteraemia, higher bacterial load, longer hospital stay and treatment failure.^1,7,8 We undertook a study to know the prevalence and compare various tests in in-vitro diagnosis of hVISA.

Materials and methods

Isolates: Non-repeat clinical MRSA isolates from patients of tertiary care hospital during the period from Sep 2010 to Mar 2013 were subjected for vancomycin MIC by agar dilution method and/or E test.⁹ A total of 58 such isolates having MIC more than 1 μg/ml by any of the methods were collected and preserved at −80 °C. These isolates were revived. For each isolate various tests were carried out on same day. Isolates having borderline or indeterminate results were re-tested.

PAP-AUC method: PAP-AUC method as described by Wootton, et al⁵ was undertaken. Isolates were cultured on Trypticase soya broth for 24 h and were log diluted 10⁻³ to 10⁻⁸ by saline; 100 μl of each of suspensions were lawn cultured onto Brain Heart Infusion agar (BHI) containing 0.25, 0.5, 1, 2, 4, 6, 8 μg/ml of vancomycin and plain BHI Agar plates. Vancomycin analytical powder was commercially purchased (Hi-Media). The plates were incubated at 35 °C for 48 h. The colony counts (log10numbers of CFU/ml) were counted and were plotted against the vancomycin concentration on a graph paper. AUC was calculated for each isolate. A ratio was then calculated by dividing the AUC of the study isolate by the AUC of Mu3 strain (S. aureus ATCC 700698). Study isolates having ratio of >0.90–<1.3 were diagnosed as hVISA and those with ratio of ≥1.3 as VISA.

BHIA6V screening: In house BHI Agar plates with 6 mg/ml of vancomycin were prepared. 10 μl of 0.5 McFarland suspensions of each of the isolate was inoculated as spot of 15 mm in diameter. These plates were incubated at 35 °C for 24 and 48 h and were observed carefully in transmitted light for growth. Isolates were considered VISA/VRSA; hVISA or VSSA, if there was confluent growth, countable growth or no growth respectively after 48 h of incubation.

MET: Commercially available vancomycin E test strips (AB Biodisk) were used. A 2.0 McFarland standard suspension of MRSA isolates was prepared and 200 μl of suspension lawn cultured on the BHI agar. Vancomycin E-test strip was applied over the plate with the help of applicator within 5 min of lawn culture. Plates were incubated at 37 °C and reading was undertaken after 24 and 48 h of incubations. A tear drop zone of inhibition was observed. The zone edge intersecting the graded strip at the minimum concentration of the antibiotic is interpreted as the MIC. Those isolates with MIC >8 mg/ml were considered as hVISA. Comparison of various tests were undertaken for parameters as sensitivity, specificity, positive & negative predictive value (PPV & NPV), positive & negative likelihood ratios (LR+ & LR−) with confidence Interval at 95% against PAP-AUC as a gold standard.

Results

A total 58 MRSA isolates having vancomycin MIC more than 1 μg/ml by either agar dilution method or E test were subjected for PAP-AUC, BHIA6V Screen and Macro E Test for vancomycin. PAP-AUC ratio of study isolates ranged from 0.5 to 1.2; with a mean of 0.70 (±0.17). A total of 54 (93.1%) isolates had ratio less than 0.9 and were diagnosed as VSSA; while 4 (6.9%) isolates had ratio ≥0.9 to 1.2 and were diagnosed as hVISA.

On BHIA6V screening, 51 (87.9%) isolates were identified as VSSA; while, 5 (8.6%) and 2 (3.5%) isolates were hVISA and VISA respectively. A comparative result of PAP-AUC and BHIA6V is presented in Table 1; isolates identified as VISA and hVISA by BHIA6V were clubbed together for statistical analysis. There was disagreement on results of 5 (8.6%) isolates; amongst them, 4 were identified as S. aureus with reduced vancomycin susceptibility (i.e. VISA or hVISA) and one as VSSA by BHIA6V were actually VSSA and hVISA respectively by PAP–AUC. On MET, a total of 53 (91.4%) and 5 (8.6%) isolates had vancomycin MIC <8 and ≥8 ug/ml respectively. The comparison of MET and PAP-AUC is given in Table 2. There was agreement on results of 53 (91.4%) isolates in correctly identifying as hVISA or VISA by MET against PAP-AUC.

Table 1.

Comparison of BHIA6V Screen method against PAP AUC.

PAP AUC
		hVISA		VSSA		Total
BHIA6V Screen		No	%	No	%	No	%
	VISA & hVISA	3	5.2	4	6.9	7	12.1
	(a)VISA	1	1.7	1	1.7	2	3.4
	(b)hVISA	2	3.4	3	5.2	5	8.6
	VSSA	1	1.7	50	86.2	51	87.9
	Total	4	6.9	54	93.1	58	100.0

Table 2.

Comparison of Macro E test with PAP AUC.

	PAP AUC
		hVISA		VSSA		Total
MIC by MET		No	%	No	%	No	%
	≥8 μg/ml	2	3.4	3	5.2	5	8.6
	<8 μg/ml	2	3.4	51	87.9	53	91.4
	Total	4	6.9	54	93.1	58	100.0

We analyzed vancomycin MIC by E test of the study isolates⁹ with PAP-AUC; the comparison is given in Table 3. There were 15 isolates having MIC more than 2 μg/ml; amongst them, 4 (26.7%) and 11 (73.3%) isolates were diagnosed as hVISA and VSSA respectively by PAP-AUC test. We used two cut off i.e. number of isolates having MIC ≥4 μg/ml and MIC >4 μg/ml by E test in analyzing statistical parameters in comparison against PAP-AUC. All hVISA isolates were having vancomycin MIC by E test >2 μg/ml. Sensitivity, specificity, PPV, NPV, LR+ & LR- of various tests against PAP-AUC as gold standard is presented in Table 4

Table 3.

Comparison of vancomycin MIC by E test with PAP AUC.

	PAP AUC
		hVISA		VSSA		Total
MIC by E Test		No	%	No	%	No	%
	4 μg/ml	1	1.7	1	1.7	2	3.4
	2.5-3.5 μg/ml	3	5.2	10	17.2	13	22.4
	≤2 μg/ml	0	0.0	43	74.1	43	74.1
	Total	4	6.9	54	93.1	58	100.0

Table 4.

Statistical parameters of various tests in comparison with PAP-AUC.

	BHIA6V		MET		E test MIC ≥4 μg/ml		E test MIC >2 μg/ml
	Estimate	95% CI	Estimate	95% CI	Estimate	95% CI	Estimate	95% CI
Sensitivity	0.75	[0.301–0.954]	0.67	[0.208–0.939]	0.25	[0.046–0.699]	1.0	[0.463–0.989]
Specificity	0.93	[0.824–0.971]	0.94	[0.849–0.981]	0.98	[0.902–0.997]	0.8	[0.666–0.878]
PPV	0.43	[0.158–0.75]	0.4	[0.118–0.769]	0.5	[0.095–0.905]	0.267	[0.121–0.526]
NPV	0.98	[0.897–0.997]	0.98	[0.899–0.997]	0.95	[0.854–0.982]	1.0	[0.9–0.999]
LR+	10.13	[3.371–30.41]	12	[3.08–46.754]	13.5	[1.024–177.989]	4.30	[2.383–7.775]
LR−	0.27	[0.049–1.477]	0.35	[0.071–1.75]	0.76	[0.433–1.347]	0.13	[0.009–1.76]

Discussion

Vancomycin is currently the drug of choice in treatment of MRSA isolates. However with emergence of S. aureus with reduced vancomycin susceptibility, vancomycin may become ineffective. Various studies have shown low prevalence of VRSA and VISA.^1,9 However, there are recent reports of emergence of hVISA, a therapeutic challenge including vancomycin treatment failure.¹ We undertook this study to know the prevalence of hVISA. Vancomycin E test MIC stratification studies reported that hVISA detection is dependent on vancomycin MIC.⁷ Musta, et al⁷ found no hVISA in MRSA isolates which had MIC ≤1 μg/ml; however, detection ranged from 10 to 85% in isolates those had MIC 1.5–3.0 μg/ml respectively. Based on this findings, we defined an inclusion criteria of MRSA isolates with vancomycin MIC >1 μg/ml. We found hVISA prevalence of 6.9% by PAP-AUC amongst 58 MRSA isolates which had MIC >1 μg/ml by agar dilution or E test. hVISA prevalence in the range from 5 to 50% has been reported by various authors.^1,7,8 Low prevalence of hVISA has also been reported by Iyer et al¹⁰ from India, where only one isolate was found to be hVISA out of 50 MRSA isolates tested. Liu, et al¹¹ have reviewed 14 studies published between 1997 and 2001 and found a prevalence of 1.67% in 7920 S. aureus isolates, collected from all over the world, including Japan, Korea, Hong Kong, Thailand, France, Spain, Greece, Germany, Italy, and the United Kingdom. Prevalence appeared to vary with the setting from which the isolates were recovered. Hiramatsu, et al¹² found hVISA prevalence of 9.3%, among 129 MRSA strains collected at eight university hospitals against 1.3% prevalence among 970 strains collected at community hospitals. Horne, et al⁸ reported hVISA prevalence of 47.9%, they screened 117 isolates by PAP AUC. Higher antibiotic selection pressures at tertiary care centers may account for the higher prevalence of hVISA in these hospitals.^7,8 Thus, there is variation in prevalence of hVISA; which could be due to geographic locations, study populations, screening test employed and also type of studies i.e. retrospective or prospective. This necessitates ongoing efforts by clinical laboratories in screening hVISA.

Screening test for hVISA: Current CLSI guidelines are not suitable for diagnosed hVISA.^1,2 PAP-AUC, the gold standard in diagnosing hVISA cannot be used for hVISA detection in clinical laboratory settings.¹ We compared E test, MET and BHIA6V with PAP-AUC in screening of hVISA. We found, BHIA6V screening had a reasonably high sensitivity of 0.75 and PPV of 0.43. BHIA6V has been recommended for screening of hVISA and VISA by CDC.¹³ Wootton, et al¹⁴ reported low sensitivity and reproducibility with BHIA6V method. We found Macro E Test with breakpoint >8 μg/ml having sensitivity, specificity and PPV of 0.67, 0.94 and 0.40 respectively. Various other studies have reported sensitivity and specificity from 56 to 98.5% and 98–100% respectively.^3,15 We found E test at breakpoint >2 μg/ml had sensitivity, specificity and PPV of 1.0, 0.80 and 0.267 respectively in hVISA detection. However E test with MIC ≥4 μg/ml cut off had very low sensitivity of just 0.25 and found not suitable for screening of hVISA. As brought out earlier, high MIC by E test is associated with higher hVISA prevalence.⁷ It is well known that a single test is not suitable for screening hVISA^1,3,14,15; which is also evident from the study.

There are few limitations of the study; firstly all isolates were preserved for 6–12 months; there are reports of decline in MIC after preservation.¹⁶ The numbers of hVISA isolates by PAP-AUC were few. Further, in-vitro susceptibility results of the study were not clinically correlated. In spite of these limitations, the study has brought out, the necessity of screening of hVISA amongst MRSA isolates.

To conclude, we found hVISA prevalence of 6.9% by PAP-AUC in MRSA isolates having vancomycin MIC >1 μg/ml. All hVISA isolates had vancomycin MIC by E test more than 2 μg/ml. BHIA6V and MET had reasonable sensitivity and specificity but lacked PPV in diagnosis of hVISA. The results of the study point towards a need for surveillance of hVISA and developing laboratory tests or strategies in its diagnosis in routine clinical microbiology setting.

Conflicts of interest

This paper is based on Armed Forces Medical Research Committee Project No. 4075/2010 granted by the office of the Directorate General Armed Forces Medical Services and Defence Research Development Organisation, Government of India.