Variables Affecting Results of Sodium Chloride Tolerance Test for Identification of Rapidly Growing Mycobacteria

Abstract

The sodium chloride tolerance test is often used in the identification of rapidly growing mycobacteria, particularly for distinguishing between Mycobacterium abscessus and Mycobacterium chelonae. This test, however, is frequently unreliable for the identification of some species. In this study we examined the following variables: medium manufacturer, inoculum concentration, and atmosphere and temperature of incubation. Results show that reliability is improved if the test and control slants are inoculated with an organism suspension spectrophotometrically equal to a 1 McFarland standard. Slants should be incubated at 35°C in ambient air and checked weekly for 4 weeks. Growth on control slants should be critically evaluated to determine the adequacy of the inoculum; colonies should number greater than 50. Salt-containing media should be examined carefully to detect pinpoint or tiny colonies, and colonies should number greater than 50 for a positive reaction. Concurrent use of a citrate slant may be helpful for distinguishing between M. abscessus and M. chelonae. Molecular methodologies are probably the most reliable means for the identification of rapidly growing mycobacteria and should be used, if possible, when unequivocal species identification is of particular importance.

The sodium chloride (salt) tolerance test is commonly used in the identification of some members of the genus Mycobacterium. Many species of rapidly growing mycobacteria are salt tolerant and grow readily on Lowenstein-Jensen medium with 5% NaCl (3, 7). However, Mycobacterium chelonae and Mycobacterium mucogenicum fail to grow on this medium, and salt tolerance is often used to distinguish these from other rapid growers and to differentiate Mycobacterium abscessus from M. chelonae.

Several investigators have commented on the unreliability of this test for making distinctions among these organisms (6, 10, 11). This unreliability may be due to the conflicting guidelines for test performance and to the vague criteria for test interpretation. Various inoculum concentrations are recommended by manufacturers and investigators, including a suspension equal to a 1 McFarland standard (1, 7), a barely turbid or light suspension (1b, 3, 4, 7a), and a very dilute suspension (12). The optimum incubation temperature and the optimum atmosphere are also unclear; recommendations include 28°C in ambient air (3, 7, 12), 28°C in CO₂ (1), 35°C in ambient air (1b, 4) and 35°C in 5 to 10% CO₂ (7a). Test interpretation is also not clearly defined, with some investigators noting that growth of greater than 50 colonies is considered positive (3, 7, 12), while others consider any amount of growth positive (1, 1b, 4, 7a).

The studies described here attempt to define the optimum conditions for test performance and to evaluate the usefulness of the test in relation to other available tests.

The initial study was performed to determine the effects of inoculum concentration, medium manufacturer, and incubation temperature on the results of the salt tolerance test for various rapidly growing mycobacteria. Because some manufacturers or investigators recommend the incubation of salt-containing medium in a CO₂-containing atmosphere (1, 7a) or at 30°C (10), a smaller secondary study was performed to determine the effects of incubation in these environments.

(A portion of these data was presented at the 1996 General Meeting of the American Society for Microbiology, New Orleans, La., 19 to 23 May 1996.)

MATERIALS AND METHODS

Effects of medium manufacturer, inoculum concentration, and incubation temperature. (i) Organisms.

A total of 31 isolates of five different rapidly growing mycobacterial species were examined for growth on salt-containing medium under a variety of conditions. Patient isolates included 15 isolates recovered from patients seen at the Clinical Center of the National Institutes of Health and 8 isolates referred to the Mycobacteriology Laboratory of the Maryland State Health Department (kindly provided to us by Nancy Hooper, Maryland State Health Department). Identifications of 19 of these isolates were as follows: M. abscessus, 5 isolates; M. chelonae, 3 isolates; Mycobacterium fortuitum, 5 isolates; M. mucogenicum, 4 isolates; Mycobacterium peregrinum, 2 isolates. The American Type Culture Collection (ATCC) type strains examined were M. fortuitum ATCC 6841, M. peregrinum ATCC 14467, M. abscessus ATCC 19977, M. chelonae ATCC 35752. The strains M. mucogenicum ATCC 49649 and ATCC 49651 and M. peregrinum ATCC 35755 were included as representatives of those species. An additional five isolates belonged to a genetic variant of M. abscessus (1a) and included M. abscessus ATCC 23006 and four patient isolates.

Patient isolates were identified by traditional biochemical methods (3) by using the following parameters: positive Kinyoun reaction, growth rate of less than 7 days, growth on MacConkey agar without crystal violet (Remel, Lenexa, Kans.), iron uptake reaction (Remel), nitrate reduction (Remel), and utilization of citrate, inositol, and/or mannitol as sole carbon sources (prepared in-house) (3).

Identifications of patient isolates were confirmed molecularly by PCR amplification of a 383-bp fragment of a portion of the 65-kDa heat shock protein gene (1a) with the primers described by Hance et al. (2). Restriction endonuclease analysis (REA) was performed with the PCR products (1a, 9) by digesting them with BstEII and HaeIII (New England Biolabs, Beverly, Mass.). The patterns obtained with patient isolates were compared to those obtained with the ATCC type strains.

The genetic variants of M. abscessus were identified biochemically as M. abscessus but produced a distinctive pattern by REA with the endonuclease HaeIII. This pattern did not match the patterns obtained with the M. abscessus type strain (ATCC 19977) but was identical to that obtained with M. abscessus ATCC 23006 and ATCC 23018.

(ii) Preparation of organism suspensions.

Single colonies of each organism were inoculated into 5 ml of Tween Albumin Broth (TAB; Remel) with glass beads and were incubated at 28 ± 1°C in ambient air. Broth turbidity was checked spectrophotometrically each day after vigorous vortexing. Working suspensions were prepared when the percent transmission (T) of the broth culture was ≤35.8 ± 2.0%T at 530 nm (spectrophotometrically equivalent to a 1 McFarland standard suspension) or after 8 days of incubation for organisms which did not achieve the target percent T. Suspensions with T of <33.8% were adjusted to 35.8 ± 2.0%T with TAB. Most M. abscessus isolates and two M. abscessus variant isolates did not achieve the target percent T; however, these organisms grew in clumps in TAB, and the clumps could not be completely broken apart even with vigorous vortexing.

Three different working suspensions were prepared for each organism. Each of these suspensions has been described as the appropriate inoculum for the performance of the salt tolerance test (1, 1b, 3, 4, 7, 7a, 12). Suspension A was a 1 McFarland suspension (or as close to a 1 McFarland suspension as possible) prepared as described above. Suspension B was prepared in TAB from the 1 McFarland suspension. This barely turbid suspension was standardized spectrophotometrically at 86 ± 2.0%T measured at 530 nm. Suspension C was a 1:2,500 dilution of suspension A prepared in 0.85% sterile saline.

(iii) Media inoculation.

One-tenth milliliter each of suspension A, B, and C was inoculated onto two sets of media, each set consisting of one Lowenstein-Jensen agar slant (LJ; Remel) used as a growth control and one slant each of LJ with 5% NaCl (LJ Salt) from three different manufacturers (Remel; BBL, Cockeysville, Md.; Difco, Detroit, Mich.). The medium formulations for the three different manufacturers were essentially identical. As purity check plates Middlebrook 7H11 Agar plates (Remel) and Trypticase Soy Agar with 5% Sheep Blood (BBL) were inoculated with the initial organism suspension.

Colony counts were performed on Middlebrook 7H11 Agar (Remel) from suspensions A, B, and C of five organisms (one isolate each of M. chelonae, M. peregrinum, M. mucogenicum, M. fortuitum, and M. abscessus).

(iv) Incubation temperature.

To determine the optimum incubation temperature, one set of media was incubated at 35 ± 1°C in ambient air and the other set was incubated at 28 ± 1°C in ambient air. The LJ and LJ Salt were checked for growth weekly for 4 weeks.

Effect of Incubation Temperature and Atmosphere. (i) Organisms.

A total of nine organisms were tested: two patient isolates of M. abscessus, three patient isolates of M. chelonae, M. chelonae ATCC 35752, and one patient isolate each of M. fortuitum, M. mucogenicum, and M. peregrinum.

(ii) Organism suspensions and medium inoculation.

Suspension A was prepared for each isolate as described above, and three slants each of LJ and LJ Salt from manufacturer 3 were inoculated with 0.1 ml of the suspension.

(iii) Incubation temperature and atmosphere.

An LJ and an LJ Salt inoculated with each organism were incubated at 35 ± 1°C in ambient air, 35 ± 1°C in 5% CO₂, and 30 ± 1°C in ambient air.

(iv) Test interpretation.

For both phases of the study, the salt tolerance test was considered to be interpretable if there were more than 50 colonies of any size (pinpoint colonies to confluent growth) on the control LJ. The test was considered negative if there were less than or equal to 50 colonies on the LJ Salt and positive if the LJ Salt showed more than 50 colonies of any size (3). Correct results were those which agreed with the expected results for a particular organism. The salt tolerance test was expected to be positive for M. abscessus, M. abscessus variant, M. fortuitum, and M. peregrinum; the salt tolerance test was expected to be negative for M. chelonae and M. mucogenicum.

RESULTS

All patient isolates gave the expected Kinyoun stain results and biochemical reactions (3).

One isolate of M. abscessus gave negative salt tolerance test results with all inoculum concentrations tested with media from all manufacturers and at all temperatures except at 30°C in ambient air. At 30°C growth was tiny and difficult to interpret. This isolate was identified by REA as M. abscessus; staining and biochemical results (including carbon utilization tests) were consistent with those for M. abscessus.

The failure of the initial TAB cultures of some isolates (M. abscessus and M. abscessus variant) to achieve the target percent T had no effect on the interpretability of the results for test and control inoculated with suspension A or B. Control slants prepared from these suspensions all showed far greater than 50 colonies, most with confluent or nearly confluent growth. Nine tests were considered uninterpretable because control slants showed less than 50 colonies. These had all been inoculated with suspension C, and all were M. abscessus or the M. abscessus variant.

Differences in salt tolerance test results obtained with LJ Salt from different manufacturers tested at 35°C in ambient air are presented in Table 1. The three formulations tested showed similar results with the more concentrated inocula of suspensions A and B. With the more dilute suspension, suspension C, inaccurate results occurred more frequently with all medium formulations. These results include those obtained with the salt tolerance test-negative M. abscessus isolate.

TABLE 1.

Agreement of results of sodium chloride tolerance test for rapidly growing mycobacteria with expected results for all organisms tested at 35°C in ambient air with media from various manufacturers and various inoculum concentrations^a

Medium manufacturer	No. of correct tests/no. of tests (%) for the following inoculum suspension:
	Ab	Bc	Cd
1	30/31 (97)	29/31 (94)	24/29 (83)e
2	29/31 (94)	28/31 (90)	25/29 (86)
3	30/31 (97)	30/31 (97)	26/29 (90)

^aExpected results are as defined by Kent and Kubica (3). LJ Salt from three manufacturers, BBL, Difco, and Remel, were used. Data include results for salt tolerance test-negative M. abscessus isolates. 

^bA suspension equal to a 1 McFarland standard suspension. 

^cA barely turbid suspension equal to 86 ± 2%T at 530 nm. 

^dA 1:2,500 dilution of a 1 McFarland standard suspension. 

^eResults for 2 of 31 test were uninterpretable. 

Salt tolerance test results obtained after incubation at 28°C in ambient air with media from all manufacturers are presented in Table 2. At this temperature, all four isolates of M. chelonae gave false-positive results when tested with the 1 McFarland inoculum suspension (suspension A) and the barely turbid suspension (suspension B). With suspensions A and B all other organisms (except the salt tolerance test-negative M. abscessus isolate) gave the expected results.

TABLE 2.

Agreement of results of sodium chloride tolerance test for rapidly growing mycobacteria with expected results for all organisms tested at 28°C in ambient air with various inoculum concentrations^a

Organism	Expected result	No. of organisms/no. of tests	Suspension Ab			Suspension Bc			Suspension Cd
			No. of tests pos	No. of tests neg	% Correct	No. of tests pos	No. of tests neg	% Correct	No. of tests pos	No. of tests neg	% Correct
M. fortuitum	Pos	6/18	18	0	100	18	0	100	18	0	100
M. peregrinum	Pos	4/12	12	0	100	12	0	100	12	0	100
M. chelonae	Neg	4/12	12	0	0	12	0	0	9	3	25
M. abscessuse	Pos	6/18	15	3	83	15	3	83	9	9	50
M. abscessus variant	Pos	5/15	15	0	100	15	0	100	10	2	67f
M. mucogenicum	Neg	6/18	0	18	100	0	18	100	0	18	100

^aExpected results are as defined by Kent and Kubica (3). LJ Salt from three manufacturers, BBL, Difco, and Remel, were used. pos, positive; neg, negative. 

^bA 1 McFarland standard suspension. 

^cA barely turbid suspension equal to 86 ± 2%T at 530 nm. 

^dA 1:2,500 dilution of a 1 McFarland standard suspension. 

^eIncludes a salt tolerance-negative M. abscessus isolate. 

^fResults for 3 of 15 tests were uninterpretable. 

With the dilute suspension, suspension C, at 28°C, two of the four M. chelonae isolates gave false-positive results with media from all three manufacturers; one M. chelonae isolate gave false-positive results with media from manufacturers 1 and 2, and another M. chelonae isolate gave false-positive results with media from manufacturer 3. Three of the six M. abscessus isolates (including the salt tolerance test-negative M. abscessus strain) gave false-negative results with media from all manufacturers. The result for one isolate of an M. abscessus variant on media from all manufacturers was uninterpretable, while one isolate gave false-negative results on media from two manufacturers. All other isolates gave correct results on all media.

Salt tolerance test results by species for all isolates tested at 35°C in ambient air are presented in Table 3. At this temperature, all isolates of M. chelonae gave the expected results with media from all manufacturers and with all inoculum concentrations. In addition to the salt tolerance test-negative isolate, one isolate of M. abscessus gave false-negative results with medium from manufacturer 2 with suspension A and with media from manufacturers 1 and 2 with suspension B. Additionally, with suspension B one M. peregrinum isolate gave false-negative results with the medium from manufacturer 2.

TABLE 3.

Agreement of results of sodium chloride tolerance test for rapidly growing Mycobacteria with expected results for all organisms tested at 35°C in ambient air with various inoculum concentrations^a

Organism	Expected result	No. of organisms/no. of tests	Suspension Ab			Suspension Bc			Suspension Cd
			No. of tests pos	No. of tests neg	% Correct	No. of tests pos	No. of tests neg	% Correct	No. of tests pos	No. of tests neg	% Correct
M. fortuitum	Pos	6/18	18	0	100	18	0	100	18	0	100
M. peregrinum	Pos	4/12	12	0	100	11	1	92	9	3	75
M. chelonae	Neg	4/12	0	12	100	0	12	100	0	12	100
M. abscessuse	Pos	6/18	14	4	78	13	5	72	8	7	44f
M. abscessus variant	Pos	5/15	15	0	100	15	0	100	10	2	67g
M. mucogenicum	Neg	6/18	0	18	100	0	18	100	0	18	100

^aExpected results are as defined by Kent and Kubica (3). LJ Salt from three manufacturers, BBL, Difco, and Remel, were used. pos, positive; neg, negative. 

^bA 1 McFarland standard suspension. 

^cA barely turbid suspension equal to 86 ± 2%T at 530 nm. 

^dA 1:2,500 dilution of a 1 McFarland standard suspension. 

^eIncludes a salt tolerance test-negative M. abscessus isolate. 

^fResults for 3 of 18 tests were uninterpretable. 

^gResults for 3 of 15 tests were uninterpretable. 

With suspension C, one isolate of M. peregrinum gave a false-negative result with slants from manufacturer 3, while another M. peregrinum isolate gave false-negative results with slants from manufacturers 1 and 2. In addition to the salt tolerance test-negative M. abscessus isolate, the result for one M. abscessus isolate was uninterpretable, one isolate was falsely negative with media from all manufacturers, and another isolate gave false-negative results with medium from manufacturer 2. For isolates of the genetic variant of M. abscessus, the result for one isolate was uninterpretable and two isolates each gave false-negative reactions with medium from manufacturer 1.

Table 4 compares the results of the salt tolerance tests incubated at 35°C in 5% CO₂, 35°C in ambient air, and 30°C in ambient air. The selected isolates of M. fortuitum, M. mucogenicum, and M. peregrinum gave the expected results in all environments. One of five M. chelonae isolates gave false-positive results at 35°C in CO₂, and all M. chelonae isolates gave false-positive results at 30°C in ambient air. One M. abscessus isolate gave the expected results in all three environments. The salt tolerance test-negative M. abscessus isolate gave negative results at 35°C in CO₂ and positive results at 30°C, although growth at this temperature was tiny and difficult to interpret.

TABLE 4.

Agreement of results of sodium chloride tolerance test for rapidly growing mycobacteria with expected results for organisms tested at 35°C in ambient air, 35°C in 5% CO₂, and 30°C in ambient air^a

Organism	Expected result	No. of organisms	35°C ambient air			35°C CO2			30°C ambient air
			No. of tests pos	No. of tests neg	% Correct	No. of tests pos	No. of tests neg	% Correct	No. of tests pos	No. of tests neg	% Correct
M. fortuitum	Pos	1	1	0	100	1	0	100	1	0	100
M. peregrinum	Pos	1	1	0	100	1	0	100	1	0	100
M. chelonae	Neg	5	0	5	100	1	4	80	5	0	0
M. abscessusb	Pos	2	1	1	50	1	1	50	2	0	100
M. mucogenicum	Neg	1	0	1	100	0	1	100	0	1	100

^aExpected results are as defined by Kent and Kubica (3). LJ Salt from manufacturer 3 were used. The media were inoculated with a suspension equal to a 1 McFarland turbidity standard. pos, positive; neg, negative. 

^bIncludes an M. abscessus isolate which was salt tolerance test negative except at 30°C in ambient air. 

The average colony counts from suspensions A, B, and C for the five isolates tested were 1.2 × 10⁸, 1.6 × 10⁷, and 4.9 × 10⁴ CFU/ml, respectively.

DISCUSSION

The first variable that was evaluated was the effect of medium manufacturer on the results of the salt tolerance test. This study shows that medium source does not contribute significantly to the variability of the test when tests are incubated at 35°C in ambient air (Table 1). The formulations of the three brands of medium tested were essentially identical. The differences in the results noted among the media from the various manufacturers were most significant with the dilute suspension, suspension C (Table 1). It appears that these differences are more likely related to inoculum concentration than to medium manufacturer.

An important variable in the salt tolerance test is the concentration of organisms in the inoculum. It appears that the inhibitory nature of the salt-containing medium, combined with low concentrations of organisms in the inoculum, could lead to false-negative results, especially for M. abscessus. The best results are obtained when slants are inoculated with a high concentration of organisms. Overall, suspension A gave the best results at all temperatures, while suspension B gave only slightly fewer incorrect results (Tables 2 and 3). Suspension C gave the largest number of incorrect results (Tables 2 and 3) and also produced the only tests which were considered uninterpretable due to low inocula (fewer than 50 colonies on the control slant).

It was useful to standardize the suspensions used here to ensure that sufficient numbers of organisms were present in the inoculum. In practice, organism suspensions prepared by visual comparison to a 1 McFarland standard suspension appear to be less dense than those prepared spectrophotometrically. We therefore recommend the suspension standardization procedure described herein. More than half of the organism suspensions reached the target percent T within 6 days of incubation at 28°C. Some suspensions of M. abscessus and M. abscessus variant, although displaying heavy growth in TAB, did not attain the target percent T by day 8, perhaps because of the tendency of the organisms to grow in clumps. Control slants inoculated with suspensions A and B of these organisms did show adequate growth at 35°C. In nine tests inoculated with suspension C, adequate growth was not achieved on the control slant. This failure to grow on the control slant illustrates the need for the use of a heavy inoculum prepared from a suspension which was vigorously vortexed with glass beads to distribute the organisms more uniformly.

Another important variable in the salt tolerance test is the incubation temperature. All isolates of M. fortuitum and M. mucogenicum gave the expected results at all temperatures tested (Tables 2, 3, and 4). Only one test (with slants from one manufacturer) with M. peregrinum gave a false-negative result with suspension B at 35°C (Table 3).

Two isolates of M. fortuitum biovar 3 (ATCC 49403 and ATCC 49404) also gave the expected positive results with inocula from a 1 McFarland suspension incubated at 35°C in ambient air (data not shown).

M. abscessus is generally expected to grow on salt-containing media (3), a characteristic which separates this species from M. chelonae. Overall, the results of the salt tolerance test for M. abscessus and the M. abscessus variant showed minimal deviation from the expected results when incubation was at 28 and 35°C with inocula containing high concentrations of organisms. The tendency for this organism to grow in clumps requires that inocula be prepared carefully; an adequately dense inoculum suspension will decrease the possibility of false-negative results. One of the 11 M. abscessus and M. abscessus variant isolates tested failed to grow on salt-containing media at either 28 or 35°C from any inoculum concentration (Tables 2 and 3). Wallace et al. (11) noted that some M. abscessus isolates are inhibited at temperatures higher than 28°C; it appears that some M. abscessus isolates may also be inhibited on salt-containing media at 28°C. This isolate also did not grow on salt-containing medium at 35°C in a CO₂-enriched atmosphere when suspension A was used (Table 4). A fine film of growth was noted on media incubated at 30°C with an inoculum from suspension A (Table 4). The control LJs showed adequate growth at all temperatures for all inoculum concentrations. None of the M. abscessus or M. abscessus variant isolates studied failed to grow at 35°C when a large inoculum was used.

M. chelonae is generally expected to be salt tolerance test negative (3). With inocula from suspensions A and B, all M. chelonae isolates tested gave false-positive reactions when they were incubated at 28°C. Even with the very dilute suspension, suspension C, 9 of 12 tests gave false-positive results at 28°C. All M. chelonae isolates were also salt tolerance test positive when they were tested at 30°C, and one of five isolates was positive when tested at 35°C in 5% CO₂ (Table 4). In contrast, all tests incubated at 35°C in ambient air (Table 3) gave the expected negative results. Kusunoki and Ezoki (5) previously noted these effects with a reference strain of M. chelonae. Silcox et al. (8) describes the profuse growth of M. chelonae at 28°C, and from the data presented here it appears that the stimulating effects of a lower temperature, and possibly those of a higher CO₂ content as well, encourage the growth of M. chelonae and override the inhibitory effect of the salt-containing medium for this species.

The interpretation criteria for salt tolerance test results are also unclear. In this study positive reactions showed mostly confluent or nearly confluent growth on the salt-containing media with the more dense inocula. The criterion for positivity recommended by Kent and Kubica (3) (at least 50 colonies on the salt-containing media) appears to be appropriate for accurate test result interpretation. We would like to add that the presence of greater than 50 colonies of any size should be considered a positive reaction. Some M. abscessus or M. abscessus variant strains incubated at 35°C grew as pinpoint or tiny colonies growing over the surface of the slant. Slants should therefore be carefully examined to detect weakly positive strains.

The use of the LJ control is an important part of the performance of this test since the amount of growth that it shows gives an accurate estimation of the adequacy of the inoculum. Criteria for adequate growth on the control slant as specified by others include growth (1, 1b, 4), numerous colonies (3, 7), and positive (7a) to greater than 50 colonies (12). Since inoculum density appears to be a critical factor in test performance, we recommend that the control slant show more than 50 colonies for the test to be considered interpretable, because this number is also the recommended minimum needed on the salt-containing slant for a positive result (3). If 50 or fewer colonies are visible, the inoculum density is probably inadequate and an apparently negative salt tolerance test result may merely be due to an insufficient inoculum and not to inhibition of the isolate.

Several investigators have recommended the use of the carbon utilization test because it more reliably distinguishes between M. abscessus and M. chelonae (6, 8, 10, 11). Isolates of M. chelonae utilize citrate as a carbon source, while M. abscessus does not. This test medium appears to be used mostly in reference laboratories and until recently has not been commercially available, limiting its usefulness for many clinical laboratories. In our hands, several isolates identified as M. chelonae by REA have given negative or weakly positive citrate reactions (data not shown); more study of the parameters affecting the results of the citrate utilization test seems warranted.

The performance of the salt tolerance test can be improved with the use of well-dispersed inocula with concentrations spectrophotometrically equal to a 1 McFarland turbidity standard (or the use of an inoculum incubated for 8 days for organisms which do not attain a high enough concentration otherwise). Test and control slants should be incubated at 35°C in ambient air for 4 weeks. Salt-containing media and control slants should be examined carefully for the presence of greater than 50 colonies of any size. Isolates of M. abscessus that fail to grow in the salt tolerance test, such as the one that we encountered, are presumably rare. However, the routine use of a citrate utilization test may be helpful to increase the likelihood of the correct identification of such isolates. For those laboratories with the capability of performing them, molecular methodologies such as PCR and REA (1a, 9) are probably the most unambiguous methods for the rapid and accurate identification of these organisms.