Quality Assurance for Antimicrobial Susceptibility Testing of Neisseria gonorrhoeae in Canada, 2003 to 2012

Abstract

Data from the Canadian National Gonococcal Antimicrobial Susceptibility Comparison Program, including results from 25 proficiency panels distributed between 2003 and 2012, were analyzed. The average MIC agreement between the participating laboratories ranged from 85.6% to 98.8% over the 10-year period, with the interpretation agreement ranging from 85.7% to 98.1%.

TEXT

Neisseria gonorrhoeae, the bacterial agent causing gonorrhea infections, is the second most commonly reported bacterial pathogen in Canada with >13,000 reported cases (39.2 cases/100,000 population) in 2013 (1). Worldwide, the number of cases is estimated to be 106 million (2). Antimicrobial resistance (AMR) in N. gonorrhoeae is a serious concern as this pathogen has developed resistance to all classes of the antimicrobials used for treatment, including sulfonamides, penicillins, tetracyclines, and fluoroquinolones. More recently, isolates with resistance to azithromycin and reduced susceptibilities to the third-generation cephalosporins have emerged (3). In 2012, the WHO published a global action plan to control the spread and impact of antimicrobial resistance in N. gonorrhoeae (2). The recommendations included enhancing surveillance programs and strengthening laboratory capacity. Quality assurance (QA) systems are crucial to ensuring that the antimicrobial susceptibility data generated are accurate, standardized, and comparable nationally and internationally. The WHO also provided updated reference cultures with elevated MICs to the third-generation cephalosporins, azithromycin, ciprofloxacin, and spectinomycin, for laboratories to use in internal and external quality assurance programs (4).

The National Microbiology Laboratory (NML) of the Public Health Agency of Canada has monitored antimicrobial susceptibilities in N. gonorrhoeae isolates as part of the National Neisseria gonorrhoeae Antimicrobial Surveillance Program since 1985. To standardize the susceptibility testing data and to maintain the comparability of data generated from each province, the NML offers a proficiency testing program, which provides participating laboratories the opportunity to identify discrepancies and improve testing protocols. Accurate results lead to effective treatment options and enhanced public health prevention programs.

This report presents a compilation of data through the Canadian National Gonococcal Antimicrobial Susceptibility Comparison Program from 2003 to 2012 (25 panels). Each panel usually included 4 N. gonorrhoeae isolates currently circulating in Canada and 1 blinded reference strain. Each participating laboratory tested the isolates with the antimicrobials and methodology (agar dilution [5] or Etest [6]) routinely employed in their laboratory. The NML also tested all isolates.

In addition to the test panel isolates, control strains(ATCC 49226, WHO C, WHO F, WHO B [2003-2010], and WHO K [2010 TO 2012] were tested for quality control purposes (4). The analysis of the data for the control strains is not included in the results presented here; however, the percentage of MICs within the acceptable range was >90% for all participants. Results were submitted to the NML and analyzed using SAS v7 until 2010 after which the LabWare Laboratory Information Management System v6.0 with IBM Cognos v10.2 was used. The Etest MIC results were rounded up to the closest 2-fold dilution value for comparison with the agar dilution results. Modal MICs for each panel isolate/antibiotic combination were calculated from the modified data excluding any results with a less-than-or-equal-to (≤) or greater-than-or-equal-to (≥) sign. The percent agreement of the MICs with the modal MICs was determined for each of the participating laboratories as well as for each of the antibiotics. MICs within one 2-fold dilution of the modal MICs were considered to be in agreement.

Categorical interpretations were applied to the MIC results (including MICs with a less-than-or-equal-to [≤] or greater-than-or-equal-to [≥] sign) submitted by the participating laboratories as follows: penicillin, tetracycline, ciprofloxacin, and spectinomycin results from the CLSI (5); cefixime and ceftriaxone decreased susceptibility according to the WHO guidelines (2); erythromycin results from Ehret et al. (7); and azithromycin results from the CDC (8). Modal categorical interpretations were determined for each panel isolate/antibiotic combination. The percent concordance between categorical interpretations was determined for each participating laboratory as well as for each antibiotic.

The laboratories participating in this quality program from 2003 to 2012 achieved an overall average agreement between participants' MIC results and modal MICs of >90%, which is considered the acceptable standard; only 1 laboratory had <90% (9, 10). Concordance between the participating laboratories' interpretations and the modal interpretations was >90% for all but 2 laboratories. Table 1 outlines the involvement and performance of the laboratories participating in the program.

TABLE 1.

Performance of participating laboratories in the National Gonococcal Antimicrobial Susceptibility Comparison Program between 2003 and 2012

Laboratory code	No. of panels completed	Test method used	Antibiotics testeda								Agreement between MICs of participant and modal MICs (% [no.])	Concordance between interpretations of participant and modal interpretations (% [no.])
			PEN	SPEC	TET	CX	CIP	CE	AZI	ERY
1	25	Agar dilution	x		x	X	X				97.52 (354/363)	98.13 (367/374)
2	20 (2003–2012)	Agar dilution	X	X	X	X	X	X	X	X	90.73 (499/550)	91.80 (593/646)
3	25	Agar dilution (9 panels)/Etest (19 panels)	X	X	X	X	X	X	X	X	96.15 (749/779)	95.03 (917/965)
4	25	Etest	X	x	X	X	X	X	X		97.99 (536/547)	94.86 (701/739)
5	25	Etest	X		X	X	X	X	X		96.78 (542/559)	93.42 (696/745)
6	25	Agar dilution	X	X	X	X	X	X	x	X	95.44 (272/285)	96.26 (720/748)
7	25	Etest	X	X	X	X	X	X	X	X	94.29 (694/736)	89.19 (751/842)
8	18 (2004–2012)	Etest	X		x	X	X	X	X		98.77 (400/405)	95.87 (511/533)
9	14 (2005–2011)	Agar dilution (9 panels)/Etest (5 panels)	X	X	X	X	X				85.62 (250/292)	85.71 (294/343)
10b	25	Agar dilution (25 panels)/Etest (14 panels)	X	X	X	X	X	X	X	X	96.00 (1,272/1,325)	95.63 (1,138/1,190)

^aA lowercase “x” indicates that the antibiotic was not tested for all panels. PEN, penicillin; SPEC, spectinomycin; TET, tetracycline; CX, ceftriaxone; CIP, ciprofloxacin; CE, cefixime; AZI, azithromycin; ERY, erythromycin.

^bLaboratory 10 is the NML, which started using the Etest in addition to agar dilution in 2006.

In Table 2, the MICs and interpretations are compared to their modes for each antibiotic. The percentages of agreement between the MIC results and their modes for each of the antibiotics were all >90%. This is also true for the proportion of agreement between interpretations for all of the antibiotics except tetracycline, partially due to the high percentage of tetracycline modal MICs at interpretative breakpoints. It is important to note that there were ≤1.5% false susceptible interpretations for all antibiotics combined.

TABLE 2.

MIC and categorical interpretation agreement among participating laboratories for each antibiotic

Antibiotic	Total no. of isolates	Total no. of pairwise comparisons	Modal MIC range (mg/liter)	% participant MICs that differ from the modal MICs by the following no. of 2-fold dilutions					% agreement (no. of comparisons with concordant results/total no. of comparisons) by:		Modal MICs at categorical interpretation breakpoints
				≤−2	−1	Same	1	≥2	Modal MIC (±1 log2)	Categorical interpretation	%	No.
Penicillin	106	838	0.016–64	4.1	15.9	66.0	11.9	2.2	93.79 (786/838)	93.32 (1,006/1,078)	38.67	41/106
Tetracycline	107	880	0.125–64	3.8	19.1	56.6	16.5	4.4	91.82 (808/880)	84.23 (871/1,034)	67.29	72/107
Erythromycin	103	434	0.063–32	3.5	13.4	68.7	12.7	1.8	94.70 (411/434)	90.89 (549/604)	50.49	52/103
Spectinomycin	112	571	8–32	1.2	11.0	70.4	16.1	1.2	97.55 (557/571)	98.67 (744/754)	0.00	0/112
Ceftriaxone	117	978	0.0005–0.25	3.6	17.7	63.2	14.1	1.4	94.99 (929/978)	96.07 (1,027/1,069)	23.93	28/117
Ciprofloxacin	102	817	0.002–16	2.9	15.7	63.7	14.9	2.8	94.25 (770/817)	94.48 (1,129/1,195)	23.53	24/102
Cefixime	102	528	0.002–0.5	1.3	15.2	71.8	10.2	15.2	97.16 (513/528)	98.10 (927/945)	11.76	12/102
Azithromycin	105	779	0.032–16	1.5	16.3	68.8	13.0	0.4	98.07 (764/779)	96.16 (877/912)	22.85	24/105

Table 3 shows the agreement between agar dilution and the Etest for the test isolates of the panels of this study. The agar dilution modal MICs and Etest modal MICs as well as their interpretations were calculated separately for each isolate/antibiotic combination for comparison. The modal MICs for the 2 methods were considered in agreement if they were within one 2-fold dilution of each other. The percent agreements between the agar dilution and the Etest modal MICs for all of the antibiotics except tetracycline were >90%. The differences between the means of the agar dilution modal MICs and the Etest modal MICs for tetracycline, ceftriaxone, and cefixime were significant (P < 0.05), according to the matched-pair t test (http://www.socscistatistics.com/Default.aspx). The percent concordances between the Etest and agar dilution modal interpretations for each of the antibiotics except tetracycline and erythromycin were >90%. More than 50% of the Etest MICs for these 2 antibiotics were 1 dilution lower than the agar dilution MICs. The interpretation agreements were reduced due to the high proportion of modal MICs at interpretation breakpoints.

TABLE 3.

Percentage of 2-fold dilution differences between the agar dilution modal MICs and the Etest modal MICs for proficiency panel isolates between 2003 and 2012

Antibiotic	No. of isolatesa	No. of results (agar/Etest)b	% isolates (agar dilution modal MIC − Etest modal MIC) that differ by the following no. of 2-fold dilutions					% agreement of modal MICs	Matched-pair t test, P < 0.05 of modal MIC means			% concordance of interpretations	Modal MICs at categorical interpretation breakpointsc
			≤−2	−1	Same	1	≥2		t	P	Significant		%	No.
Penicillin	74	400/554	4.1	35.1	55.4	5.4	0.0	96.0	1.96	0.09525	No	94.5	43.6	34/78
Tetracycline	82	433/586	15.9	69.5	13.4	1.2	0.0	84.1	5.20	<0.00001	Yes	79.3	67.9	57/84
Erythromycin	30	252/260	0.0	56.7	40.0	3.3	0.0	100.0	−0.45	0.65312	No	69.6	73.3	22/30
Spectinomycin	66	324/314	3.0	42.4	53.0	1.5	0.0	97.0	−0.15	0.87878	No	100.0	34.8	23/66
Ceftriaxone	90	398/641	3.3	35.6	52.2	8.9	0.0	96.7	4.27	0.00005	Yes	93.9	26.7	24/90
Ciprofloxacin	80	403/545	1.3	43.8	47.5	7.5	0.0	98.8	−1.52	0.13336	No	93.2	27.5	22/80
Cefixime	27	272/310	0.0	40.7	51.9	7.4	0.0	100.0	2.39	0.02457	Yes	94.9	31.4	11/35
Azithromycin	62	275/626	0.0	12.9	71.0	16.1	0.0	100.0	−0.04	0.96567	No	95.6	25.8	16/62

^aNumber of isolates that have modal MICs for both the agar dilution method and the Etest method and can therefore be compared.

^bTotal number of MIC values used to determine the modal MICs for both the agar dilution method and the Etest method.

^cThe numbers of modal MICs at breakpoints for both Etest and agar dilution were calculated. The highest number of the 2 calculations was used.

The laboratories participating in the National Gonococcal Antimicrobial Susceptibility Comparison Program had more than acceptable results, similar to those for other quality assurance programs. Countries participating in the European Gonococcal Antimicrobial Susceptibility Program participate in the European Neisseria gonorrhoeae Antimicrobial Resistance External Quality Assurance Programme, which involves testing 10 N. gonorrhoeae isolates for susceptibility annually. The participating laboratories, using their routine testing methods and up to 7 antimicrobial agents, had overall concordance for MICs and interpretations of ≥90% for most antibiotics in 2011 (18 laboratories) (11) and 2012 (19 laboratories) (12).

The Indian Gonococcal Antimicrobial Surveillance Program External Quality Assurance schemes (2001 to 2007) included up to 6 participating laboratories. They found an overall interpretation concordance of 82% for 5 antibiotics. As this program was using the disc diffusion method, the MIC concordance was not calculated (13).

Enhanced surveillance programs monitoring the antimicrobial resistance in N. gonorrhoeae are essential for informing treatment guidelines and combating the threat of untreatable gonorrhea. External quality assurance programs are an integral part of effective surveillance programs to ensure that their data are reliable. The results of the National Gonococcal Antimicrobial Susceptibility Comparison Program are encouraging. The high level of agreement between laboratories indicates that the AMR data generated are accurate and can be considered comparable internationally. There is confidence that emerging resistance and changing levels in antibiotic resistances will be appropriately detected, and treatment guidelines can be modified to effectively manage gonorrhoea infections.