Comparative Study on the Use of Solid Media: Lowenstein–Jensen and Ogawa in the Determination of Anti-Tuberculosis Drug Suscep

Tuberculosis (2002) 82(2/3), 63 -- 67 & 2002 Elsevier Science Ltd. All rights reserved doi:10.1054/tube.323, available online at http://www.idealibrary.com on Comparative study on the use of solid media: Lo¨wenstein–Jensen and Ogawa in the determination of anti- tuberculosis drug susceptibility

S.Tanoue, S. Mitarai, H. Shishido

Department of Respiratory Diseases,National Tokyo Hospital, 3-1-1Takeoka,Kiyose,Tokyo 204-5858, Japan

Summary A total of 53 Mycobacterium tuberculosis strains were examined to evaluate the consistency of susceptibility tests using two solid egg-based culture media,Ogawa and L˛wenstein--Jensen (L--J), in order to assessthe inter-media differencesin susceptibility tests foranti-tuberculosis drugs.Fifty-three M. tuberculosis strainswere tested for resistance against isoniazid, rifampicin, ethambutol, streptomycin and other alternative drugs on L--J and Ogawa media.Data from isoniazid (INH) revealed 31 strainswere susceptibleand11strainswereresistant in L--J medium.Onthe other hand, one strain showedresistanceagainst INH onlyon Ogawamedium.Thisgaveaconsistencyratio of 97.7% for INHbetweenthesetwomedia.Similarly,themethodshad100% compatibility with rifampicin, 86.0% with ethambutol and 88.4% with streptomycin.When 41of these strainswere tested against ethionamide, enviomycin, para-aminosalicylate, sparfloxacin, andlevofloxacin, the same numbers of strains were shown to be resistant or susceptible to L--J and Ogawa media. Similarly, the consistency ratio was 75.6% for kanamycin and 61.0% for capreomycin.There are inconsistencies between results obtained with L--J and Ogawa media,Ogawa medium had a tendency to produce results that indicated more resistant strains.However, the differenceswere statistically negligible. & 2002 Elsevier Science Ltd. Allrights reserved

INTRODUCTION for agar plates is increasingly becoming the standard for determining M. tuberculosis resistance in western coun- Mycobacterium tuberculosis infection is re-emerging in tries. However, this technology has not been introduced most parts of the world and recently the World Health in most of the developing countries in Asia and sub- Organization (WHO) has sent a stern warning to its Saharan Africa because it is relatively easy to use the members on the problems associated with the treatment conventional egg-based media for initial isolation and of tuberculosis.1 Specifically, WHO warns of the impact of consecutive susceptibility testing. Although susceptibility drug-resistant M. tuberculosis. In this situation, even the testing using egg-based media is popular in many diagnosis of tuberculosis with smear-positive specimens countries, the methods used vary widely. Because of and subsequent directly observed treatment with short- increasing international communication, the comparabil- course chemotherapy (DOTS) are not enough to control ity of susceptibility results from different methods will the disease. Thus, drug susceptibility testing for resistant become essential and will be crucial in understanding the M. tuberculosis is important in designing subsequent real trend of resistant M. tuberculosis in the world. treatment options for such patients. However, there are The present study was conducted to evaluate the currently several methods for susceptibility testing, i.e. compatibility of M. tuberculosis resistance results the absolute method, resistant ratio method and propor- obtained using Lo¨ wenstein–Jensen and Ogawa media. tion method with a variety of culture media.2 Recently, the proportion method using artificial culture media like Middlebrook 7H9 for the BACTEC/MGIT system or 7H10 MATERIALS AND METHODS Isolation of M. tuberculosis

Correspondence to: Satoshi Mitarai,The Research Institute of Tuberculosis, M. tuberculosis strains were obtained from pulmonary Japan Anti-Tuberculosis Association, 3-1-24 Kiyose,Tokyo 203-8533, Japan. Tel.:+81-424-93-3090;Fax:+81-424-92-8258;E-mail:[email protected] tuberculosis patients admitted to the National Tokyo Hospital and National Hiroshima Hospital (51 and 2 Accepted: 14 February 2002 strains, respectively) between July 1996 and January 1997. 63 64 Tanoue et al.

All the strains were recovered from sputum specimens Ta ble 1 Critical concentrations of anti-tuberculosis drugsin the proportion method collected after informed consent. The initial isolates were identified as M. tuberculosis complex using Ziehl–Neelsen Drug Concentration Drug Concentration (Z–N) staining, niacin accumulation tests and the DNA– (mg/ml) (mg/ml) DNA hybridization (DDH) method (Mycobacterium, Isoniazid 0.2 Enviomycin 20 Kyokuto Pharmaceutical Industrial Co., Ltd., Tokyo, Rifampicin 40 p-Aminosalicylate 0.5 Streptomycin 4 Capreomycin 20 Japan). A total of 53 strains of M. tuberculosis showing Kanamycin 20 Cycloserin 30 complete drug resistance to at least one of the four first Ethambutol 2 Levofloxacin 1 line drugs, i.e. isoniazid (INH), rifampicin (RFP), etham- Ethionamide 20 Sparfloxacin 1 butol (EB), and streptomycin (SM), was obtained. The absolute concentration method on 1% Ogawa media was Ta ble 2 Composition of L˛wenstein-Jensen and Ogawa media used at each laboratory for initial susceptibility testing. Thereafter, the drug susceptibilities were confirmed with Amount the proportion method in the present study. Ingredient 1% Ogawa medium L--J medium Potassium phosphate 1.0 g 1.2 g Drug susceptibility test Sodium glutamate 1.0 g F Magnesium sulphate F 0.12 g F The susceptibilities of all 53 resistant isolates of M. Magnesium citrate 0.3 g L-Asparagine F 1.8 g tuberculosis were examined against several anti-tubercu- Distilled water 100 ml 300 ml losis drugs for data confirmation and comparison. These Egghomogenate 200 ml 500 ml drugs were INH, RFP, EB, SM, kanamycin (KM), ethiona- Glycerol 6.0 ml 6.0 ml 2%Malachitegreensolution 6.0ml 10ml mide (TH), enviomycin (EVM), cycloserin (CS), capreomy- Inspissation temperature 901C/60 min 901C/45 min cin (CPM), para-aminosalicylate (PAS), levofloxacin Expected pH 6.5 6.9 (LVFX) and sparfloxacin (SPFX). The critical concentration of each drug used in the resistance assays is shown in Table 1. They were determined using the standard growth of tuberculosis bacilli was inspected in the MIC concentrations in use on the L–J medium in western tubes. For quality control, a standard susceptible M. countries, clinical experience and treatment effects.2,3 tuberculosis strain was tested in the same experiments. INH, RFP, EB, and SM were stipulated as first-line drugs for the treatment of tuberculosis. The proportion method Consistency of two different methods was used to evaluate resistances against anti-tuberculosis drugs. The drug susceptibility results of 53 strains examined The cultured M. tuberculosis strain was harvested and using L–J and Ogawa media were compared and grouped suspended in distilled water at McFarland turbidity No. 1. into two different categories, i.e. consistent and incon- The original bacterial suspension was further diluted sistent, according to the results. The consistency and the 100 Â and 10,000 Â . One hundred microlitres of each of MIC of each drug with these media were evaluated. the two suspensions were inoculated onto each drug- Conventionally, the results with the L–J medium were containing medium. If the number of colonies on the stipulated as the standard. drug-containing medium was more than that on another plain medium with a 100 Â diluted suspension of the inoculum, the strain was judged to be 1% or more Statistical analysis resistant to the drug. L–J and 1% Ogawa media were used Statistical analysis was done by chi-square or Fisher’s for the tests. A summary of the ingredients in each exact test for consistency results. Parametric (two-way medium is shown in Table 2. analysis of variance) and non-parametric (Wilcoxon) The minimum inhibitory concentration (MIC) of each analytic methods were used for MIC results. A p-value drug was determined using commercially prepared drug- of o0.05 was considered significant. containing medium (Kyokuto Pharmaceutical Industrial Co., Ltd., Tokyo, Japan). MIC is the minimum drug concentration that inhibits the growth of the tested RESULTS strain. Approximately 100,000 cfu of resistant tuberculo- Confirmation of drug resistance sis bacilli were inoculated onto the control medium and the medium of a two-fold dilution series of each drug The 53 strains that were judged as having no resistance in ranging from 0.08 to 100 mg/ml. When the control tubes initial testing were confirmed with the proportion method showed sufficient growth (basically 3–4 weeks), the in newly designated concentrations as shown in Table 1.

Ta bl e 3 M. tuberculosis susceptibility comparisons using L--J and Ogawa media Consistent Inconsistent Susceptible Resistant Ratio (%) Resistant Resistant on Ogawa on L--J INH 31 11 97.7 1 0 RFP 26 17 100 0 0 EB 31 6 86.0 6 0 SM 10 28 88.4 5 0 KM 24 7 75.6 10 0 TH 29 12 100 0 0 EVM 32 9 100 0 0 PAS 2 9 12 10 0 0 0 CPM 12 13 61.0 16 0 CS 35 4 95.1 0 2 SPFX 40 1 100 0 0 LVFX 38 3 100 0 0 Note:Forty-three and 41strains were examined for first-line and other drugs, respectively.

100 100 Isoniazid (n=25) Rifampicin (n=28)

80 80 Ogawa 60 L-J 60 L-J d 40 40 Ogawa

20 20

0 0 0.08 0.16 0.32 0.64 1.25 2.5 5 10≤ (µg/ml) 0.78 1.56 3.13 6.25 12.5 25 50 100≤ (µg/ml) t of strains inhibite 100 100 Ethambutol (n=17) Streptomycin (n=27)

80 80

60 L-J 60 L-J Cumlative percen 40 40 Ogawa Ogawa 20 20

0 0 0.08 0.16 0.32 0.64 1.25 2.5 5 10≤ (µg/ml) 0.78 1.56 3.13 6.25 12.5 25 50 100≤ (µg/ml) Fig. 1 MICs of first-line drugs against M. tuberculosis.

With L–J medium, 12 strains were resistant to INH. In the showed resistance against INH only in the Ogawa same way, 17, 6 and 29 strains were resistant to RFP, EB medium. This gave a consistency ratio of 97.7% for INH and SM, respectively. As for other drugs, 12, 7, 12, 9, 13 between these two media. If these L–J results are and 16 strains were resistant to PAS, KM, TH, EVM, CPM considered to be the standard, the sensitivity and and CS, respectively. Fourteen strains were confirmed specificity of the Ogawa medium could be calculated as susceptible to all drugs with the L–J medium. 100% and 96.9%, respectively. Regarding the other first- line drugs, the L–J and Ogawa methods had 100% compatibility in their susceptibility to RFP, 86.0% Resistance to four first line drugs to EB, and 88.4% to SM, respectively. No strain On L–J medium, 31 strains were susceptible and 11 showed resistance to L–J medium alone. The results are strains resistant to INH. On the other hand, one strain summarized in Table 3.

Resistance to second- and third-line drugs evaluating the comparability of drug susceptibility find- For second- and third-line drugs, the susceptibility ings using L–J and 1% Ogawa media. The resistant strains of 41 strains were tested on L–J and Ogawa media. were mainly used in order to clarify differences in results There was no difference in the numbers of resistant obtained using these two media. If a majority of the and susceptible strains to TH, EVM, PAS, SPFX, and strains tested was susceptible to anti-tuberculosis drugs, LVFX on the two media used. Thus, the results the consistency of these methods would be over- were completely consistent on the two media. On the evaluated. For example, if 90 of the 100 strains were other hand, 10 strains showed resistance only on Ogawa susceptible, the consistency ratio would be always over against KM, giving a consistency ratio of 75.6%. Further- 90%. It was thus necessary to use an increased number of more, 16 strains showed resistance only on Ogawa resistant strains in order to evaluate the two different against CPM giving a consistency ratio of 61.0%. The media. Ogawa method produced results indicating more resis- It has been shown that some proteins in eggs, and tant strains, although two strains only showed resistance phosphate, sometimes affect the activities of anti-tuber- against CS on the L–J medium, giving a consistency culosis drugs.7, 8 However, in this present study, the ratio of 95.1%. components of egg are the same in both media because these media were prepared by one company using the eggs from a designated farmer. The major differences of MICs of four first-line drugs these media are the source of nitrate, malachite green The cumulative graphs of MICs for the four first-line concentration, volume of egg homogenate and pH. As for drugs are shown in Fig. 1. The available number of strains the source of nitrate, L-asparagine is used in L–J media examined for each drug was different as indicated in the and sodium L-glutamate is used in Ogawa medium. These figure. The biggest dissociation in susceptibility results amino acids have the same type of side chains and it is was seen with EB. The variances of EB data were not thus unlikely that such small essential amino acids could statistically significant with parametric (two-way analysis cause different test results. It has also been reported that of variance) and non-parametric (Wilcoxon) analytic the growth-supporting ability on both L–J and Ogawa methods. Furthermore, the differences in the numbers media without drug is almost equal.9,10 Thus, either the of strains resistant to INH, RFP, and SM on the two media source of nitrate or the higher concentration of malachite were not significant. green in Ogawa would not inhibit the growth of mycobacteria more than L–J. However, the higher volume percent of egg homogenate in Ogawa will absorb drug DISCUSSION into the medium. This could explain the result that no With the re-emergence of tuberculosis in the world, strain showed resistance on L–J medium alone except for focus has moved to the study of M. tuberculosis CS. Additionally, the expected pH of 1% Ogawa medium drug resistance. Global surveillance of drug-resistant is 6.5 while that of L–J is approximately 6.9. The sodium M. tuberculosis was conducted by WHO/IUATLD world- hydroxide solution used for the treatment of clinical wide between 1994 and 1997 in order to estimate its specimens (mainly sputum) will neutralize the acid in an impact.4,5 In the report, the need for international initial isolation culture. However, the saline or distilled standardization of drug susceptibility testing (DST) for water used for bacterial suspension in the indirect testing comparative evaluation of results was emphasized. The method is not enough to neutralize the culture medium. criteria for use in determining drug resistance on This factor may explain the result that there was no strain solid media were developed and standardized, although showing resistance on L–J medium alone in this study there was no mention of the differences in resistance except for CS. results on different solid media.4,6 The BACTEC or Ethambutol showed the biggest dissociation of suscept- MGIT system is now widely used in advanced countries. ibility results between L–J and Ogawa among the four However, even culture examination of clinical specimens first-line drugs, although there was no statistically using solid media has not yet been done routinely significant difference in the results obtained between in many developing countries. Many of the deve- the two media. There are several possible reasons for the loping countries that do adopt culture methods in the EB results. First, this disassociation could have been due determination of drug susceptibility of M. tuberculosis to the fewer number of strains tested. Second, the higher strain still use solid media. Thus, it is important to volume percent of egg homogenate in the Ogawa evaluate the differences in solid media for drug suscept- medium may have contributed to the discrepancy, as ibility testing. mentioned above. Additionally, in a previous study, the A total of 53 strains tested that were resistant to at least activity of EB is reported to be stronger at pH 7.2 than at one anti-tuberculosis drug in initial testing were used in pH 6.8 using the Ogawa medium.11 It is possible that the

relatively low pH of Ogawa medium also affected the REFERENCES results. 1. Raviglione M C, Snider Jr D E, Kochi A. Global epidemiology of For resolution of discrepancies in the results between tuberculosis: morbidity and mortality of a worldwide epidemic. 1% Ogawa and L–J media, some changes in drug JAMA 1995; 273: 220–226. concentrations should be considered. This is a feasible 2. Canetti G, Fox W, Khomenko A, Mahler H T, Menon N K, approach because the composition of media is somewhat Mitchison D A, Rist N, Smelev N A. Advances in techniques of testing mycobacterial drug sensitivity, and the use of sensitivity rigid and it is difficult to introduce a new standard. As for tests in tuberculosis control programmes. Bull World Health Org INH, RFP, TH, EVM, PAS, SPFX, LVFX and CS, it seems 1969; 41: 21–43. reasonable to consider the results as equal because the 3. The Committee for Mycobacterial Examination. A new guideline consistency ratio is over 95% for each drug. However, for for tuberculosis examination. Tokyo: Japan Anti-Tuberculosis EB and SM, in which the consistency ratio is between 85% Association, 2000. 4. Pablos M A, Laszlo A, Bustreo F et al. Anti-tuberculosis drug and 95%, it is recommended to increase the concentra- resistance in the world; The WHO/IUATLD Global Project on tion of drugs. In fact, the concentrations of EB and SM in Anti-tuberculosis Drug Resistance Surveillance 1994 – 1997. the new operational standard advanced by the Japanese Geneva, 1998. Society for Tuberculosis are 2.5 and 10 mg/ml, respec- 5. Cohn D, Bustreo F, Raviglione M C. Drug resistance in tively. There is no proposed concentration for CPM in the tuberculosis: review of the worldwide situation and WHO/ IUATLD global surveillance project. Clin Infect Dis 1997; 24 present study because even the Japanese Society for (Suppl 1): S121–S130. Tuberculosis recommends the same concentration in 6. Rist N, Crofton J. Drug resistance in hospitals and sanatoria. A Ogawa and L–J at present. The biggest discrepancy in report to the Committee on Laboratory Methods and to the CPM must be studied further. Committee on Antibiotics and Chemotherapy of the In conclusion, although there was a tendency for International Union Against Tuberculosis. Bull Int Union Tuberc 1960; 30: 13–27. Ogawa medium to indicate more resistant strains than 7. Inderlied C B, Salfinger M. Antimicrobial agents and susceptibility L–J medium, the difference was not statistically signifi- tests: mycobacteria. In: Murray P R, Baron E J, Pfaller M A, cant. There were fewer ingredients in Ogawa medium Tenover F C, Yolken R H, eds. Manual of clinical microbiology, than L–J medium, making Ogawa advantageous, less Sixth Edition. Washington, DC: ASM Press, 1995: pp 1393. laborious and economical to prepare. Moreover, it is 8. Hawkins J E. Drug susceptibility testing. In: Kubica G P, Wayne L G, eds. The mycobacteria a sourcebook. New York: Marcel Dekker relatively easy to control its quality. Some laboratories Inc., 1984: pp 179–180. may go through a process of using egg-based solid media 9. Ogawa T, Koseki Y. Ogawa’s quantitative isolation for tubercle for susceptibility tests before artificial media can be bacilli primary report. Jpn J Bact 1957; 12: 703–706 (in Japanese) introduced largely because of cost. Ogawa medium would 10. Ogawa T, Sawai T, Nakajima K, Honda A. Ogawa’s quantitative likely be employed in some cases. M. tuberculosis isolation for tubercle bacilli; second report. Jpn J Bact 1958; 13: 71–76 (in Japanese). susceptibility data obtained by employing either medium 11. Tamura M, Takano S. Effects of pH of media on the minimal are compatible. Therefore, the drug resistance results inhibitory concentration of D-ethambutol to Mycobacterium obtained from different parts of the world using either tuberculosis. Kekkaku 1968; 44: 19–23 (in Japanese with English method can be evaluated using the same standard. abstract).