Growth Inhibition of Mycobacterium tuberculosis by ... - NCBI

JOURNAL OF CLINICAL MICROBIOLOGY, Jan. 1996, p. 84–86 0095-1137/96/$04.0010 Copyright q 1996, American Society for Microbiology

Vol. 34, No. 1

Growth Inhibition of Mycobacterium tuberculosis by Polyoxyethylene Stearate Present in the BACTEC Pyrazinamide Susceptibility Test MARK A. MILLER,1* LOUISE THIBERT,2 F. DESJARDINS,1 SALMAN H. SIDDIQI,3

AND

A. DASCAL1

McGill University and Department of Microbiology, SMBD-Jewish General Hospital, Montreal,1 and Laboratoire de sante´ publique du Que´bec, Ste-Anne-de-Bellevue, Quebec,2 Canada, and Becton Dickinson Diagnostic Instrument Systems, Sparks, Maryland3 Received 17 August 1995/Returned for modification 22 September 1995/Accepted 13 October 1995

We have previously found that approximately 3.5% of 428 clinical isolates of Mycobacterium tuberculosis yield uninterpretable results in the BACTEC pyrazinamide (PZA) susceptibility test system, because of inadequate growth. We tested the hypothesis that polyoxyethylene stearate (POES), the ingredient of the reconstituting fluid for the test, was the cause of this growth inhibition. A total of 15 isolates known for their previously uninterpretable results and 100 randomly chosen clinical isolates were tested in parallel both with and without POES. Repeat testing of the isolates with previously uninterpretable results yielded results in the presence of POES in only seven (47%). In the absence of POES, all gave interpretable results but one such result showed false resistance. For the other 100 clinical isolates, interpretable results were obtained with and without POES, but growth was enhanced in the absence of POES, especially in the PZA-susceptible strains. This was evidenced by a decreased time to attain a growth index of 200 in the control vial (4.9 days without POES versus 5.8 days with POES; P < 0.001) and a higher mean growth index ratio on the day of interpretation of the test (7.4% without POES versus 2.2% with POES; P < 0.001). However, the enhanced growth without POES led to 20 susceptible strains being misinterpreted as either resistant or borderline. We suggest that isolates of M. tuberculosis which yield uninterpretable results in the BACTEC PZA test system should be retested both with and without POES. If interpretable results indicating PZA resistance are obtained only in the absence of POES, the result should be confirmed by a pyrazinamidase assay or by the conventional proportion method. Routine omission of POES from the BACTEC test for all clinical strains is discouraged because of the unacceptably high false-resistance rates. Pyrazinamide (PZA) has become one of the first-line drugs for the treatment of tuberculosis (7). The BACTEC method for mycobacterial culture, identification, and susceptibility testing now includes a radiometric test to assay for PZA susceptibility (8). Satisfactory interpretation depends on adequate growth in the acidified Middlebrook 7H12 (BACTEC 12B) medium at the time when a growth index (GI) of at least 200 is reached in the control vial within 4 to 20 days. In a previous study, it was observed that clinical isolates of Mycobacterium tuberculosis, when tested in the BACTEC radiometric PZA susceptibility system, yielded interpretable results in a median time of 5.0 days. However, 15 strains out of a total of 428 different isolates (3.5%) yielded uninterpretable results due to poor growth (6). We previously observed that some strains displaying poor growth in the PZA test medium grew well in regular 12B medium (pH 6.8) and that polyoxyethylene stearate (POES) enhanced growth of these strains at pH 6.8. However, at the low pH of the PZA test medium (pH 6.0), the presence of POES partially inhibited growth of these strains. POES is included in the reconstituting fluid of the PZA susceptibility test to prevent cell clumping. Suspecting that growth inhibition in the BACTEC PZA test method was produced by POES, we conducted a prospective study of PZA susceptibility by means of the BACTEC method for 115 clinical isolates of M. tuberculosis, both with and without the addition of POES.

tory) were used for this study. Fifteen strains had previously yielded uninterpretable results in the BACTEC PZA susceptibility test because of insufficient growth during the maximum 20 days of incubation for the test. Another 100 isolates were randomly chosen from 413 strains with previous interpretable results in the BACTEC system. The BACTEC PZA susceptibility testing was carried out in parallel, both with and without the addition of POES to the growth medium. In brief, for each isolate, the BACTEC PZA susceptibility test was performed as suggested by the manufacturer. However, because POES is contained in the reconstituting fluid for the test, an additional parallel test was set up with the exact same methodology but the lyophilized PZA was reconstituted with sterile water instead of reconstituting fluid. For each test, actively growing isolates in regular BACTEC 12B medium (containing POES) were appropriately inoculated into two BACTEC PZA test medium bottles, one of which contained the drug while the other, without the drug, served as a control (Fig. 1). Daily readings of the GI in both bottles were performed until the day of interpretation when the GI of the control vial reached at least 200. The recommended interpretive criteria for the ratio of the GI of the drug-containing vial to the GI of the control vial were followed: ,9%, susceptible; 9 to 11%, borderline; .11%, resistant. An uninterpretable result was documented if a GI of 200 in the control vial was not obtained after 20 days. All the 15 strains which gave uninterpretable results in the earlier study, as well as strains giving discrepant results in the two assays with and without POES, were tested further for PZA susceptibility at the Centers for Disease Control and Prevention (by J. O. Kilburn) by means of the conventional proportion method (PZA concentration of 25 mg/ml in Middlebrook 7H10 agar at pH 5.5). Results of a pyrazinamidase assay, using the Wayne method (4, 14), were also available for each isolate. Isolates displaying pyrazinamidase activity as well as susceptibility to PZA in the BACTEC test were considered susceptible to PZA and not further tested by the conventional proportion method. For the statistical analysis, a matched t test was used to compare mean GI ratios or mean days of interpretation for testing isolates with and without POES. A chi-square analysis was used to compare results between the two tests expressed as proportions.

MATERIALS AND METHODS RESULTS

One hundred and fifteen recent clinical isolates of M. tuberculosis from the Laboratoire de sante´ publique du Que´bec (Quebec Provincial Health Labora-

The results of the BACTEC PZA susceptibility test for 100 randomly chosen clinical isolates, with and without POES, are shown in Table 1. Six isolates yielded uninterpretable results

* Corresponding author. 84

INHIBITION OF M. TUBERCULOSIS BY POES

VOL. 34, 1996

85

FIG. 1. PZA susceptibility test protocol with and without the addition of POES. RF, reconstituting fluid.

only when POES was present, and the absence of POES led to significantly faster results overall (mean of 5.8 days with POES versus 4.9 days without POES; P , 0.001). When another 15 previously uninterpretable strains were retested in an identical manner in the presence of POES, 8 isolates (53%) again yielded uninterpretable results due to poor growth by day 20. However, seven strains yielded interpretable results during this identically repeated test (six susceptible and one resistant), as shown in Table 2. The strain which was interpreted as resistant was shown to be susceptible by the conventional method. Without the presence of POES in the test, all 15 strains grew adequately in the test system and yielded interpretable results: 10 susceptible, 1 borderline, and 4 resistant (Table 2). However, three of the strains displaying resistance as well as the single borderline strain were all PZA-susceptible by the conventional testing. Simultaneous testing of the 100 randomly chosen isolates with and without POES led to 6 strains giving uninterpretable results only when tested in the presence of POES (Table 1). These were excluded from the analysis of time to result and accuracy of result. As shown in Table 1, interpretable results were obtained with and without POES in 5.8 and 4.9 days (means), respectively (P , 0.001). The enhanced growth in the absence of POES was noticeable for the 58 PZA susceptible strains, as can be seen by examining the GI ratio at the time of

TABLE 1. Results of the BACTEC PZA susceptibility test, with and without POES No. of isolatesa BACTEC result With POES

100 randomly chosen isolates Interpretable Uninterpretable 15 previously uninterpretable isolates Interpretable Uninterpretable

94 (5.8) 6 7 8

Without POES

100 (4.9) 0 15 0

a The average numbers of days to result given in parentheses were significantly different (P , 0.001).

interpretation under the two test conditions (Table 2). The average GI ratios for the 58 susceptible isolates with and without POES were 2.2 and 7.4%, respectively (P , 0.001), while the 36 PZA-resistant isolates produced equivalent GI ratios under both conditions (90.7% with POES versus 91.6% without POES [not a significant difference]). In the absence of POES, tests could be interpreted for 48, 89, and 95% of the isolates on days 4, 5, and 6, respectively. This contrasts with interpretable results in only 32, 62, and 84% of the strains on those days in the presence of POES (x2 5 14.1; P , 0.001). However, as can be seen in Table 2, the enhanced growth without POES led to a drop in accuracy of the test due to 20 strains susceptible by the conventional method which were misinterpreted as either resistant or borderline. The accuracy of the BACTEC PZA susceptibility test in the presence of

TABLE 2. Results of BACTEC PZA susceptibility testing, with and without the addition of POES, and comparison with the conventional proportion method BACTEC test conditions

94 randomly chosen isolates All With POES Without POES 58 PZA-susceptible strains With POES Without POES 36 PZA-resistant strains With POES Without POES 15 previously uninterpretable isolates With POES Without POES a

Avg % GI

2.2 7.4c 90.7 91.6

No. of isolatesa

Accuracyb

S

R

B

54 38

39 48

1 8

96 79

54 38

3 12

1 8

93 66

0 0

36 36

0 0

100 100

6 10

1 4

0 1

86 73

S, susceptible; R, resistant; B, borderline. Percent agreement between BACTEC test results and results of the conventional proportion method. c Significantly different from the corresponding value for the test with POES. b

86

MILLER ET AL.

J. CLIN. MICROBIOL.

POES, compared with the conventional proportion method, was 96% (Table 2). DISCUSSION Conventional PZA susceptibility testing of M. tuberculosis requires growth of the organism and may lead to uninterpretable results in a proportion of strains due to poor growth in the acidified environment (5, 11, 12). Even with modifications to the test, there are still approximately 10% of isolates which fail to grow sufficiently (1). In the BACTEC PZA test system, the pH of the medium is 6.0, higher than that in the solid medium used for conventional testing via the proportion method. Still, we found that approximately 3.5% of a large number of M. tuberculosis isolates failed to grow. The results of this prospective study of PZA susceptibility testing with and without the presence of POES suggest that this behavior may be explained by growth inhibition caused by the presence of POES in the BACTEC PZA test medium. POES is one of a large number of nonionic emulsifying agents, some of which have been shown to enhance growth of mycobacteria in vitro (2, 3, 10). The growth-enhancing effect of POES in the BACTEC system has been described (9). One investigator has shown that there is dose-dependent growth stimulation of Mycobacterium paratuberculosis by one of these compounds (Tween) in a bimodal manner, with contradictory inhibition of growth at intermediate concentrations (13). Our preliminary findings of the differential effect of POES on growth of M. tuberculosis at neutral and acid pHs suggest that the inhibitory effect of POES in the BACTEC PZA test could be pH dependant. All 15 strains having previously displayed inadequate growth in the BACTEC PZA test, showed accelerated growth when POES was omitted. In addition, many of the 94 clinical isolates of M. tuberculosis which grew well in the BACTEC PZA test showed enhancement of growth in the absence of POES, as seen by the more rapid attainment of a GI of 200 in the control vial. Unfortunately, in many instances, this accelerated growth rate was also seen in the vial containing PZA, leading to a marked increase in the GI ratio on the day of interpretation of the test. Therefore, by omitting the POES, 20 strains identified as susceptible by the conventional method were falsely labeled: 12 isolates interpreted as resistant and 8 interpreted as borderline susceptible. The results of the BACTEC PZA test for PZA-resistant strains did not change with the omission of POES. In conclusion, isolates of M. tuberculosis which yield uninterpretable results in the BACTEC PZA test system should be retested both with and without POES. If interpretable results indicating PZA resistance are obtained only in the absence of

POES, the result should be confirmed by a pyrazinamidase assay (6) or by the conventional proportion method. Confirmatory tests are not required for isolates showing susceptibility to PZA in the absence of POES, since false susceptibility is not seen when POES is omitted from the test. Routine omission of POES from the BACTEC test for all clinical strains is not recommended because of the unacceptably high false-resistance rates. ACKNOWLEDGMENT We thank J. O. Kilburn (Centers for Disease Control and Prevention, Atlanta, Ga.) for performing the conventional PZA susceptibility testing. REFERENCES 1. Butler, W. R., and J. O. Kilburn. 1982. Improved method for testing susceptibility of Mycobacterium tuberculosis to pyrazinamide. J. Clin. Microbiol. 16:1106–1109. 2. Cutler, R. R., P. Wilson, and F. V. Clarke. 1987. The effect of polyoxyethylene stearate (POES) on the growth of mycobacteria in radiometric 7H12 Middlebrook TB medium. Tubercle 68:209–220. 3. Dubos, R. J., and G. Middlebrook. 1948. The effect of wetting agents on the growth of tubercle bacilli. J. Exp. Med. 88:81–88. 4. Kent, P. T., and G. P. Kubica. 1985. Public health mycobacteriology—a guide for the level III laboratory, p. 107–109. Centers for Disease Control, Atlanta, Ga. 5. Libonati, J. P., N. M. Hooper, J. F. Baker, and M. E. Carter. 1990. Evaluation of PZA susceptibility testing of M. tuberculosis by the radiometric technique, abstr. U-54, p. 150. In Abstracts of the 90th General Meeting of the American Society for Microbiology. American Society for Microbiology, Washington, D.C. 6. Miller, M. A., L. Thibert, F. Desjardins, S. H. Siddiqi, and A. Dascal. 1995. Testing of susceptibility of Mycobacterium tuberculosis to pyrazinamide: comparison of Bactec method with pyrazinamidase assay. J. Clin. Microbiol. 33: 2468–2470. 7. Perez-Stable, E. J., and P. C. Hopewell. 1988. Chemotherapy of tuberculosis. Semin. Respir. Med. 9:459–469. 8. Salfinger, M., L. B. Reller, B. Demchuk, and Z. T. Johnson. 1989. Rapid radiometric method for pyrazinamide susceptibility testing of Mycobacterium tuberculosis. Res. Microbiol. 140:301–309. 9. Siddiqi, S. H., J. P. Libonati, M. E. Corter, N. M. Hooper, J. F. Baker, and C. C. Hwangbo. 1988. Enhancement of mycobacterial growth in Middlebrook 7H12 medium by polyoxyethylene stearate. Curr. Microbiol. 17:105– 110. 10. Stinson, M. W., and M. Solotorovsky. 1971. Interaction of Tween 80 detergent with mycobacteria in synthetic medium. Am. Rev. Respir. Dis. 104: 717–727. 11. Stottmeier, K. D., R. E. Beam, and G. P. Kubica. 1967. Determination of drug susceptibility of mycobacteria to pyrazinamide in 7H10 agar. Am. Rev. Respir. Dis. 95:1072–1075. 12. Tummon, R. 1975. Growth inhibition of M. tuberculosis by oleate in acidified medium. Med. Lab. Technol. 32:229–232. 13. van Boxtel, R. M., R. S. Lambrecht, and M. T. Collins. 1990. Effect of polyoxyethylene sorbate compounds (Tweens) on colonial morphology, growth, and ultrastructure of Mycobacterium paratuberculosis. APMIS 98: 901–908. 14. Wayne, L. G. 1974. Simple pyrazinamidase and urease tests for routine identification of mycobacteria. Am. Rev. Respir. Dis. 109:147–151.