Correlation of Growth of Aerobic Blood Cultures in Hypertonic Broth with Antibiotic Therapy

JOURNAL OF CLINICAL MICROBIOLOGY, Nov. 1982, p. 890-894 0095-1137/82/110890-05$02.00/0 Copyright 1982, American Society for Microbiology Vol. 16, No. 5 Correlation of Growth of Aerobic Blood Cultures in Hypertonic Broth with Antibiotic Therapy JAN ENG* AND ARILD MAELAND Microbiological Laboratory, Ullevdl Hospital, Oslo 1, Norway Received 22 April 1982/Accepted 28 July 1982 The aim of this study was to elucidate the mechanisms by which sucrose improves growth in a hypertonic medium for isolating aerobes from blood. Clinical blood cultures were made routinely in duplicate in plain broth consisting of brain heart infusion broth with sodium polyanetholesulfonate, gelatin, and penicillinase and the same broth with 20% sucrose added. The growth patterns of Staphylococcus aureus and Enterobacteriaceae from plain and from hypertonic broth were correlated with the presence or absence of antimicrobial therapy in patients when the blood cultures were collected. In S. aureus bacteremias, 58.7% of the positive cultures collected during treatment of patients with,b-lactam antibiotics showed earlier growth or growth only in hypertonic broth, compared with 16.7% of the cultures taken during treatment with other antimicrobial agents (P < 0.05) and 17.6% of the cultures made in antibiotic-free intervals (P < 0.01). In the group of cultures yielding growth of Enterobacteriaceae, growth occurred earlier or solely in hypertonic broth in 28.9% of the cultures taken during treatment with,b-lactam antibiotics, compared with 15.7% of the cultures taken during treatment with other antimicrobial agents and 21.6% of the cultures collected in antibiotic-free intervals (differences not statistically significant). It is concluded that treatment with 1-lactam antibiotics is an important reason for the improved growth of S. aureus from hypertonic broth, but other factors are also involved. In recent years, several workers have reported that the addition of 10 to 30% sucrose in broth media for clinical blood cultures enhances recoveries (3, 7, 9, 10, 16; L. R. McCarthy and P. D. Ellner, Abstr. Annu. Meet. Am. Soc. Microbiol. 1973, M48, p. 81), although negative findings have also been published (6, 18). The mechanisms by which sucrose seems to promote recoveries are, however, obscure. A leading hypothesis is that the sugar offers osmotic support to cell-wall-defective bacterial variants induced by 3-lactam antibiotics or immune substances (9, 14, 16). An inactivating effect of sucrose upon penicillins has also been reported (17). Theoretically, two possible mechanisms of growth improvement by sucrose in blood cultures from patients under treatment with 1- lactam antibiotics may thus be conceived. Thus, we found it of interest to determine the correlation between growth patterns in blood culture media with and without sucrose and the presence or absence of antimicrobial therapy in patients when the blood cultures were collected. Cultures yielding growth of Staphylococcus aureus or Enterobacteriaceae were selected for this study since we have found that the recoveries of these two groups of bacteria were significantly improved by adding 20% sucrose to the broth media used for clinical blood cultures in this laboratory (7). The blood cultures examined in the present study formed part of the material presented previously (7). MATERIALS AND METHODS All aerobic clinical blood cultures in this hospital were made in duplicate in plain (P) broth and hypertonic (H) broth media. Details of the techniques employed have been given previously (7). Briefly, P broth was made up of brain heart infusion broth with 0.05% sodium polyanetholesulfonate, 1.2% gelatin, and 20 IU of penicillinase per ml. Two types of H broth were in use at different times during the study: (i) P broth with 20o sucrose added and (ii) P broth with 20% sucrose and 0.1% magnesium sulfate. The two H broth media have been found to yield closely parallel results (7), and they are treated together in the present study. The penicillinase employed was Penase concentrate (Difco Laboratories, Detroit, Mich.) derived from Bacillus cereus. It is stated that 1 ml of the stock preparation inactivates 500,000 IU of penicillin as measured by a test described by the producer (5). The penicillinase was sterilized by filtration through a Millex-GS filter (Millipore Corp., Bedford, Mass.), diluted in sterile water, and added aseptically to the blood culture bottles in amounts calculated to inacti- 890

VOL. 16, 1982 ANTIBIOTICS AND GROWTH IN HYPERTONIC BROTH 891 TABLE 1. Growth patterns in P broth and H broth of 166 clinical blood cultures yielding growth of S. aureus Antibiotic treatment No. of cultures with growth pattern:0 Difference' of patients P > H P = H H > P No. of cultures % of totalc 1-Lactam 1 17 28 <0.01 27 58.7 Other 1 8 3 NSd 2 16.7 None (antibiotic- 3 83 22 <0.01 19 17.6 free interval) P p > H, Growth occurred earlier in P broth or in P broth only; P = H, growth occurred simultaneously in both media; H > P, growth occurred earlier in H broth or in H broth only. b Number of (H > P) cultures minus number of (P > H) cultures. c For cultures from patients in all treatment groups, 28.9o (48 of 166). d NS, Not significant. vate 20 IU of penicillin per ml of the final medium. The actual capacities of the ready-made blood culture media to inactivate penicillin G were measured in several batches of P broth and H broth by a technique previously described (5). A 1-ml sample of broth was found to inactivate between 40 and 160 IU of penicillin G (after incubation of the reaction mixture for 2 h at room temperature) as measured in different experiments; no differences were observed between the two types of blood culture broth in their penicillinase activity. The measurements were repeated at intervals in batches of medium stored at 4 C, and the penicillinase activity was found to be fairly stable for at least 3 months. The media were distributed in 50-ml amounts to 100- ml glass bottles. The blood cultures were made by inoculating 5 ml of venous blood into each of the two bottles. The cultures were incubated unvented at 37 C for 4 days and were repeatedly subcultured onto chocolate agar plates (7). Five different growth patterns were recorded: growth simultaneously in both media, earlier growth (by 21 days) in P broth or in H broth, and growth in P broth or in H broth only. The bacterial strains isolated were identified by standard methods (4). Strains of S. aureus were tested for 1-lactamase production by the cloverleaf method (11). The clinical records of a consecutive series of 214 patients whose blood cultures yielded growth of S. aureus (60 patients) or Enterobacteriaceae (154 patients) were scrutinized in retrospect to map any antibiotic therapy given in relation to the time when blood cultures were collected. The patients with S. aureus bacteremia yielded 166 positive blood cultures, and those with enterobacterial bacteremia had a total of 342 positive blood cultures. The cultures were grouped according to whether they had been collected during antibiotic treatment or in antibiotic-free intervals. Different cultures from the same patient were in several instances placed in different treatment groups. Cultures in the treated group were subdivided on the basis of treatment with 13-lactam antibiotics and treatment with other antimicrobial agents. A culture was considered to be taken in an antibiotic-free interval if at least 3 days had elapsed since the last dose of an antimicrobial agent had been administered to the patient, but in most cultures in this group the antibioticfree interval was 1 week or more. In the treated group were included all cultures taken less than 24 h after the last dose of an antimicrobial agent. The statistical procedure used was the Fisher exact 2 x 2 test. RESULTS The growth patterns in P and H broths of cultures from the different treatment groups are shown in Table 1 (S. aureus cultures) and Table 2 (Enterobacteriaceae cultures). The two parameters of better growth in one or the other medium (earlier growth and growth in that medium only) are grouped together in the tables. It is seen from both tables that the number of cultures yielding earlier growth or growth only in H broth (H > P) significantly exceeded the corresponding number of cultures yielding the opposite growth pattern (P > H), both overall and in each subgroup of material (except for cultures with growth of S. aureus during treatment with other antimicrobial agents, which form too small a group for a statistical analysis). In addition, the difference between the number of cultures showing the growth pattern H > P and the corresponding number of cultures showing P > H was calculated as a percentage of the total number of cultures in the group concerned, and the resulting percentage was considered to represent a numerical index of the rate of growth improvement in H broth. The rate of growth improvement of S. aureus in H broth in blood cultures taken during treatment with,b-lactam antibiotics was 58.7%, whereas the corresponding rate in cultures taken in antibiotic-free intervals was 17.6% (Table 1). This difference is statistically significant (P < 0.01). Also, a comparison of the rates of growth improvement in H broth in the two groups of cultures collected during antimicrobial therapy (p-lactam antibiotics, 58.7%; other antimicrobial agents, 16.7%) (Table 1) shows a statistically significant difference (P < 0.05). For cultures yielding growth of Enterobacteriaceae (Table 2), the rate of growth im-

892 ENG AND MAELAND J. CLIN. MICROBIOL. TABLE 2. Growth patterns in P broth and H broth of 342 clinical blood cultures yielding growth of Enterobacteriaceae Antibotictreament No. of cultures with growth pattern: a Difference' of patients P > H P = H H > P No. of cultures % of total' f3-lactam 5 49 29 <0.01 24 28.9 Other 2 39 10 <0.05 8 15.7 None (antibiotic- 15 133 60 <0.01 45 21.6 free interval) a See Table 1, footnote a. b Number of (H > P) cultures minus number of (P > H) cultures. c For cultures from patients of all treatment groups, 22.5% (77 of 342). provement in H broth was 28.9% in cultures taken during treatment with P-lactam antibiotics, 15.7% in cultures taken during treatment with other antimicrobial agents, and 21.6% in cultures collected in antibiotic-free intervals. None of the differences between these rates of growth improvement in H broth is statistically significant. Table 3 presents further details on the 46 cultures from which S. aureus was isolated during treatment of the patients with P-lactam antibiotics. These cultures were collected from 17 patients altogether. The strains isolated from 12 of the patients were P-lactamase producers, and those from 5 patients were not. Eleven patients were under treatment with 13-lactam antibiotics sensitive to S. aureus 13-lactamase (penicillin G, penicillin V, or ampicillin), whereas six patients received 13-lactam antibiotics resistant to that enzyme (methicillin, cloxacillin, cephalothin, or cephalexin) either alone or in combination with penicillin G. The treatment regimens varied considerably, from oral treatment with regular doses to intensive parenteral antibiotic therapy. The growth patterns in P broth and H broth in each separate blood culture obtained from these patients are also shown. In 12 patients, S. aureus was isolated from blood cultures during treatment with other antimicrobial agents (Table 1), namely, doxycycline (5 patients), sulfonamides (2 patients), co-trimoxazole (4 patients), and nalidixic acid (1 patient). Of the 508 positive blood cultures examined altogether, 192 (37.8%) were collected while the patients were under antimicrobial treatment, the separate figures being 134 cultures (39.2%) yielding growth of Enterobacteriaceae and 58 cultures (34.9%) yielding growth of S. aureus. DISCUSSION The main findings of the present investigation were that the isolation of S. aureus from blood cultures was significantly improved in H broth both in cultures taken during antibiotic-free intervals and in cultures taken during treatment of the patients with P-lactam antibiotics; furthermore, growth was significantly more enhanced in the latter group of cultures (Tables 1 and 3). Also, Enterobacteriaceae (Table 2) grew better in H broth both in cultures collected during antibiotic-free intervals and in cultures collected during treatment with P-lactam antibiotics, but the rate of growth improvement did not differ significantly between the two groups. These findings are relevant to the following hypotheses for the mode of action of sucrose in blood culture media discussed in the literature. (i) Media supplemented with sucrose may possess nutritional advantages in the isolation of certain bacteria. This obvious possibility has, to our knowledge, been only slightly examined. Henrichsen and Bruun (10), however, who found that blood culture media containing 10% sucrose enhanced the growth of Enterobacteriaceae and S. aureus, found no correlation between the effect of sucrose on the isolation of Escherichia coi and the ability or inability of the strains to ferment sucrose. In this laboratory, H broth has been found to speed up the isolation of Salmonella species, which do not ferment sucrose. (The material published previously [7] comprised 70 blood cultures with growth of strains of Salmonella; of these, growth was detected earlier in H broth in 9 cases and earlier in P broth in 1 case [P < 0.02]). In the present study, the significant differences in the rates of growth promotion of S. aureus by sucrose between cultures taken during P-lactam antibiotic therapy and cultures taken during other therapy or antibiotic-free intervals cannot be explained by a nutritional effect of sucrose. However, in general the possibility cannot be excluded that a simple nutritional effect of sucrose may manifest itself in some blood cultures. (ii) Muschel and Larsen (15) demonstrated that sucrose counteracts bactericidal reactions through an anticomplementary effect. All the media used in the present examination, however, contained 0.05% sodium polyanetholesulfonate, which inactivates complement (12).

VOL. 16, 1982 ANTIBIOTICS AND GROWTH IN HYPERTONIC BROTH 893 TABLE 3. Antibiotic treatment administered, 3-lactamase production of the strain isolated, and growth patterns in P broth and H broth in 46 clinical blood cultures yielding growth of S. aureus during treatment of the patients with P-lactam antibiotics Patient TreatmentNoofpstv Route of Dosage per Duration of S. aureus blood cultures, with No. Sex Age (yr) Antibiotic (days)2 administration' 1. M 22 Penicillin V P0 6 x106 IU 1-1 P >H, 1H >P 2. M 64 Pivampicillin PO 3. M 75 Penicillin G IV 4. F 76 Penicillin G PE 5. F 64 Penicillin G IV 1.4 g 20 x 106 IU 2-7 6 x 106 IU 1 6 x 106 IU Unknown <1-1 P= H,2H >P - 4 P= H,2H>P - 2 H > P - 2 H > P 6. M 17 Ampicillin IV 4 g 1 + 1P=H 7. M 54 Penicillin V PO 3 x 106 IU 1 + 3 H > P 8. M 42 Ampicillin IV 9. F 72 Ampicillin IV 10. M 54 Pivampicillin PO 11. M 62 Penicillin G IV Cephalothin IV 4 g 1 + 3P= H 4 g Unknown + 4P= H,2H>P 2.8g 1 30 x 106 IU 1 6 g 1 + 2P= H,1 H>P + 2H > P 12. M 71 Ampicillin Unknown 4 g Unknown + 2 H > P 13. F 47 Methicillin IV 8 g 4 14. M 14 Penicillin G IV 30 x 106 IU 3 Methicillin IV 4 g 3 15. F 53 Cephalexin PO 2 g 16. M 49 Penicillin G IV Cephalothin IV 30 x 106 IU 2 6 g 1 + 1H>P + 2P= H,1 H >P 3 + 1H>P + 2H > P 17. M 56 Cloxacillin IV 8 g 3 + 4H > P a PO, Oral; IV, intravenous; PE, parenteral. I Duration of antibiotic treatment before blood cultures were made. I For abbreviations, see Table 1, footnote a. (iii) Simberkoff et al. (17) reported that penicillins are rapidly inactivated by sucrose at alkaline ph. It has been suggested on the basis of this finding that sucrose inactivation of penicillins in samples may permit partially inhibited bacteria to grow only in hypertonic media (13). The sucrose inactivation of penicillins was, however, found to occur at ph above 8, whereas penicillin is quite stable at ph 7.2 to 7.4 (17), which is the ph level of the blood culture media employed in the present study. Moreover, the addition of blood to blood culture media has been found to cause a marked decrease in ph (2). Our media also contained penicillinase in amounts which, according to our measurements, were sufficient to inactivate as much as 400 IU of penicillin G per ml of blood in 2 h at room temperature. The,B-lactam antibiotics most frequently given to our patients were benzylpenicillin, phenoxymethylpenicillin, and ampicillin (Table 3), all of which are sensitive to 3-lactamase derived from B. cereus (1), which was used in our media. For these reasons, we consider it less likely that the enhanced growth promotion of S. aureus in H broth in the cultures collected during P-lactam antibiotic therapy (Table 1) can be explained by the inactivation of penicillins by sucrose. The improved growth in H broth of cultures taken

894 ENG AND MAELAND during antibiotic-free intervals (Tables 1 and 2) obviously indicates that other mechanisms must be involved in these cultures. (iv) Sucrose may offer osmotic support to cellwall-defective bacterial forms. Variants of this kind have been produced experimentally with,blactam antibiotics as inducing agents, and, moreover, the formation of spheroplasts through the combined action of serum lysozyme and the complement system has been demonstrated in vitro (14). The latter finding may provide a theoretical explanation of our finding of improvement by sucrose in the growth of S. aureus and Enterobacteriaceae in blood cultures collected in antibiotic-free intervals (Tables 1 and 2). A transitory stage of cell wall defect may be a regular phase in the process of bacterial killing in vivo, whether by P-lactam antibiotics or by lysozyme and other immune substances. Hence, the blood of bacteremic patients may contain, alone or together with classical bacteria, a number of short-lived bacterial cells with various degrees of cell wall defects, and these cells, when trapped in the blood cultures, may be stabilized by the osmotic support of sucrose and be enabled to revert to classical bacteria and subsequently to grow out in the culture. A priori, this phenomenon can be expected to be more pronounced in blood cultures collected when cell-wall-defective forms have been induced by,b-lactam antibiotics in addition to (or possibly in cooperation with) immune induction than it would be in cultures taken in antibioticfree intervals, when immune induction alone is responsible. Our finding of enhanced growth improvement by sucrose of S. aureus in blood cultures collected during P-lactam antibiotic therapy thus harmonizes well with this hypothesis. It is noteworthy that successful isolations of S. aureus in the L form from the blood of patients with chronic staphylococcal infections have earlier been reported (8). It is evident from Tables 1 and 2 that only some cultures displayed growth enhancement by sucrose; most cultures showed simultaneous growth from P broth and H broth. This fact may, in terms of the hypothesis discussed here, be related to differences in the relative numbers of classical bacteria and cell-wall-defective forms in the samples of blood to be cultured. We conclude that the hypothesis that sucrose acts by stabilizing cell-wall-defective bacterial forms fits our findings well. This conclusion does not exclude the possibility that other mechanisms may also be operative in hypertonic blood cultures. J. CLIN. MICROBIOL. ACKNOWLEDGMENTS We gratefully acknowledge the assistance of Ingar Holme of the Life Insurance Companies' Institute for Medical Statistics at the Oslo City Hospitals, who performed the statistical analysis of the data. LITERATURE CITED 1. Abraham, E. P., and S. G. Waley. 1979. f-lactamases from Bacillus cereus, p. 311-338. In J. M. T. Hamilton- Miller and J. T. Smith (ed.), Beta-lactamases. Academic Press, Inc., London. 2. Beaman, K. D., B. L. Kasten, C. L. Corlett, and T. L. Gavan. 1979. Effects of blood on blood culture medium. J. Clin. Microbiol. 10:488-491. 3. Coleman, R. M., W. W. Laslie, and D. W. Lambe, Jr. 1976. Clinical comparison of aerobic, hypertonic, and anaerobic culture media for the radiometric detection of bacteremia. J. Clin. Microbiol. 3:281-286. 4. Cowan, S. T. 1974. Cowan and Steel's manual for the identification of medical bacteria, 2nd ed. The Cambridge University Press, London. 5. Difco Laboratories. 1953. Difco manual, 9th ed., p. 283. Difco Laboratories, Detroit, Mich. 6. Ellner, P. D., T. E. Kiehn, J. L. Beebe, and L. R. McCarthy. 1976. Critical analysis of hypertonic medium and agitation in detection of bacteremia. J. Clin. Microbiol. 4:216-224. 7. Eng, J. 1981. Evaluation of sucrose and magnesium sulfate as additives in aerobic blood culture medium. J. Clin. Microbiol. 14:247-251. 8. Godzeski, C. W., G. Brier, R. S. Griffith, and H. R. Black. 1965. Association of bacterial L-phase organisms in chronic infections. Nature (London) 205:1340. 9. Gross, P. A., R. Fryda, and K. Reilly. 1977. Comparative evaluation of different types of blood culture media for isolation of aerobes. J. Clin. Microbiol. 6:362-366. 10. Henrichsen, J., and B. Bruun. 1973. An evaluation of the effects of a high concentration of sucrose in blood culture media. Acta Pathol. Microbiol. Scand. Sect. B 81:707-710. 11. KJellander, J., and K. E. Myrback. 1964. A simple test for penicillinase-production. Acta Pathol. Microbiol. Scand. 61:494. 12. Kocka, F. E., T. Magoc, and R. L. Searcy. 1972. Action of sulfated polyanions used in blood culture on lysozyme, complement and antibiotics. Ann. Clin. Lab. Sci. 2:470-473. 13. McGee, Z. A., R. G. Wittier, H. Gooder, and P. Charache. 1971. Wall defective microbial variants: terminology and experimental design. J. Infect. Dis. 123:433-438. 14. Muschel, L. H. 1968. The formation of spheroplasts by immune substances and the reactivity of immune substances against diverse rounded forms, p. 19-29. In L. B. Guze (ed.), Microbial protoplasts, spheroplasts and L- forms. The Williams & Wilkins Co., Baltimore. 15. Muschel, L. H., and L. J. Larsen. 1970. Effect of hypertonic sucrose upon the immune bactericidal reaction. Infect. Immun. 1:51-55. 16. Rosner, R. 1972. A quantitative evaluation of three blood culture systems. Am. J. Clin. Pathol. 57:220-227. 17. Simberkoff, M. S., L. Thomas, D. McGregor, I. Shenkein, and B. B. Levine. 1970. Inactivation of penicillins by carbohydrate solutions at alkaline ph. N. Engl. J. Med. 283:116-119. 18. Washington, J. A., H, M. M. Hall, and E. Warren. 1975. Evaluation of blood culture media supplemented with sucrose or with cysteine. J. Clin. Microbiol. 1:79-81.