Reduced sensitivity to/mactam antibiotics arising during ceftazidime treatment of Pseudomonas aeruginosa infections

Transcription

1 Journal of Antimicrobial Chemotherapy (983) 2, Reduced sensitivity to/mactam antibiotics arising during ceftazidime treatment of Pseudomonas aeruginosa infections Anna King, Kevin Shannon, Susannah Eykyn and Ian Phillips Department ofmicrobiology, St. Thomas's Hospital Medical School, London SE 7EH, England Pseudomonas aeruginosa isolated from two patients with empyema and one with bronchopneumonia became less sensitive after treatment with ceftazidime, while Ps. aeruginosa persisted in a patient with an infected compound fracture of the tibia treated with ceftazidime but did not become less sensitive. The reduction in sensitivity to ceftazidime, which was small, was accompanied by resistance to azlocillin but there was little reduction in sensitivity to carbenicillin. The resistant strains produced increased amounts of the chromosomally-mediated cephalosporinase produced by most isolates of Ps. aeruginosa. Variants with reduced sensitivity to ceftazidime, which resembled those that developed in vivo, were selected in vitro from each of the initial ceftazidime-sensitive isolates. Introduction Development of resistance during therapy of Pseudomonas aeruginosa infections has been reported for a number of /Mactam antibiotics including carbenicillin (Darrell & Waterworth, 969), ticarcillin (Ervin & Bullock, 976), azlocillin (Eykyn, 982; Shannon, Phillips & King, 982) and piperacillin (Marier el al, 982). A subpopulation that was less sensitive to ceftazidime observed during ceftazidime therapy of a post-traumatic Ps. aeruginosa mediastinitis has been reported (Wardle el al., 98) as has development of resistance to ceftazidime during ceftazidime therapy of a Citrobacter freundii infection (Gozzard el al., 982). Development of resistance in vitro to carbenicillin and azlocillin has also been reported (Gwynn & Rolinson, 980). In this paper we describe the development of reduced sensitivity by Ps. aeruginosa in three patients during therapy with ceftazidime and also the selection of resistant variants in vitro with ceftazidime. Patients and methods Patients. The patients were selected from among those included in studies of the clinical efficacy of ceftazidime (Gozzard el al, 982; Eykyn & Phillips, unpubl. obs.); their case histories are given below. Identification and typing o/ps. aeruginosa. Isolates were identified as Ps. aeruginosa on the basis of characteristic colonial morphology, positive oxidase reaction and /83/OO $02.00/0 983 The British Society for Antimicrobial Chemotherapy

2 3 A. King et al. sensitivity to gentamicin, carbenicillin, azlocillin, cefsulodin and colistin. Organisms resistant to one or more of these antibiotics were identified as Ps. aeruginosa if they were oxidative, produced fluorescein (pyoverdin) and pyocyanin and grew at 2 C in addition to having characteristic colonial morphology and a positive oxidase reaction. Pyocine typing (Govan, 978) phage typing and serotyping were performed on all the isolates, the last two by the Division of Hospital Infection of the Central Public Health Laboratory, Colindale. Determination of minimum inhibitory concentrations (MICs). MICs were determined by the agar dilution method on 'Oxoid' (Basingstoke, U.K.). Diagnostic Sensitivity Test (DST) Agar (CM26) with an inoculum of about 0 colony-forming units. The MIC was taken as the lowest concentration of antibiotic that prevented visible growth after incubation overnight at 37 C. P-Lactamase studies. Organisms were grown overnight in 'Oxoid' Nutrient Broth No. 2 (CM 67), diluted :9 in broth and shaken at 37 C. After -5 h cefoxitin (as an inducing agent) (500 mg/ final concentration) was added to part of the culture. Incubation was continued for a further 2 h. The cells were collected by centrifugation, washed once with water, suspended in water and broken by ultrasonication. /?-Lactamase activity was determined in a reaction mixture that consisted of -8 ml potassium phosphate buffer (0 M, ph 7), looul of cell extract and looul of nitrocefin solution (2 mm, to give a final concentration of 00 (IM) by the monitoring of the rate of increase in absorbance at 500 nm at 37 C in a 0 mm path length cuvette with a Pye-Unicam (Cambridge, England) SP 8-00 spectrophotometer. Protein content was estimated by the biuret method. Attempts to determine crypticity were made by measurement of the rate of hydrolysis of nitrocefin by whole cells and comparison with the rate of hydrolysis by ultrasonically disrupted cells. Substrate profiles of /Mactamases were determined with the ultraviolet spectrophotometric method as described before (King et al., 980). The enzyme preparation (25 (jj) was mixed with antibiotic (00 UM final concentration) in phosphate buffer (2 ml, 00 mm, ph 7) in a cm-square quartz cell. The rate of change of absorbance at 37 C was measured at 235 nm for penicillins and at 260 nm for cephalosporins. A microbiological method was also used. The enzyme preparation (50 uj) was mixed with antibiotic (50 ul, 0 UM) in 0 M phosphate buffer ph 7 and incubated at 37 C for h. Samples (50 uj for carbenicillin and cephaloridine, louj for were then placed in wells cut in DST Agar that had been flooded with Escherichia coli NCTC 08 as the indicator organism. Appropriate standards, which also had been incubated at 37 C for one hour, were included so that residual antibacterial activity could be measured. Isoelectric focusing was performed in Agarose IEF gels (Pharmacia, Uppsala, Sweden) containing equal amounts of Pharmalyte ph range 3-0 and Pharmalyte ph range (Pharmacia) with a maximum potential difference of 600 volts for h. At the end of the run, the ^-lactamases were stained with nitrocefin (2000 mg/). Selection of resistant variants in vitro. In early experiments organisms were grown overnight in nutrient broth (0 ml), collected by centrifugation, resuspended in 0 ml of broth containing ceftazidime (2,, 8, 6 or mg/) and incubated overnight. At

3 Ceftazidime treatment of Ps. aeruginosa 365 this stage two procedures were adopted. Firstly, -u] samples from each culture were inoculated on DST agar containing ceftazidime (2 mg/) and incubated for three days. The sensitivity patterns of colonies that developed were determined. Secondly, about 00 ul from the initial ceftazidime broth cultures were inoculated into 5 ml of broth containing the same concentration of ceftazidime, incubated overnight and then subcultured to blood agar. The sensitivity patterns of the organisms were then determined. In later experiments the initial incubation in ceftazidime broth was omitted and cultures were plated directly on DST Agar containing 2 mg/ ceftazidime. Case reports Patient A. A 53 year old man, admitted as an emergency with an oesophageal tear, developed bilateral empyema as a result of an anastomotic leak after oesophageal surgery. Satisfactory drainage was difficult to establish and he continued to run a spiking pyrexia for four weeks, when after prolonged and varied antibiotics, Escherichia coli was isolated from blood and a pleural aspirate and a gentamicin-resistant isolate of Ps. aeruginosa was grown from a mouth swab, the neckline and pleural aspirates. Tobramycin and ceftazidime were given intravenously in maximum doses and the chest drains were re-positioned. There was slow but definite clinical improvement. Both antibiotics were stopped after 20 days in view of the emergence of Ps. aeruginosa with reduced sensitivity to ceftazidime in his sputum. The patient continued to improve and left the intensive care unit ten days later without further antibiotic therapy. Patient B. A 2 year old man was admitted with a pyopneumothorax with a chest drain in situ that had been inserted four days previously for a spontaneous pneumothorax. Ps. aeruginosa was isolated in pure heavy growth from the drain. At thoracotomy the following day, litre of pus was drained and several bullae oversewn. Despite drainage, his fever persisted and intravenous ceftazidime was given for ten days with some clinical improvement By this time the organism had become less sensitive to ceftazidime. Chest drainage continued for four weeks with gradual resolution of his fever. Thereafter a sinus persisted at the drainage site but the patient was well enough to leave hospital. Patient C. A 25 year old man had gastrointestinal bleeding, leakage of bile and a severe respiratory infection ten weeks after renal transplantation. Numerous antibiotics had been given during this period including ceftazidime for a klebsiella septicaemia and possible klebsiella pneumonia. After ten days without any antibiotic he again looked septic and febrile and Ps. aeruginosa and Klebsiella sp. were isolated from the sputum. He was again treated with ceftazidime with minimal clinical effect but further gastric surgery was required following which he made a slow recovery. By the fourth day of treatment, Ps. aeruginosa of diminished sensitivity to ceftazidime was isolated. Patient D. A year old man sustained a compound fracture of the tibia in a road accident. On admission the wound was cleaned and a Denham pin inserted. Nine days later external fixation and skin grafting were performed. He received flucloxa-

4 366 A. King et al. cillin and ampicillin throughout this period. Two weeks later pus was oozing from pin and graft sites and the leg was swollen and inflamed. Ps. aeruginosa was grown from all sites and he was treated with intravenous ceftazidime with a marked improvement noted after a week. Nine days later the pins were removed and the ceftazidime was stopped the next day. The organism persisted but its ceftazidime sensitivity was unchanged throughout. Results The organisms isolated before the start of therapy were sensitive to all the compounds tested and produced only small amounts of /Mactamase unless synthesis of the enzyme was induced by growth in the presence of cefoxitin (Table I). In contrast, the organisms isolated from the sputum of patients A, B and C after treatment with ceftazidime showed reduced sensitivity to ceftazidime and azlocillin, little change in sensitivity to carbenicillin and increased production of /Mactamase. However, except for isolate A.2, the /Mactamase activity was considerably less than that induced by cefoxitin (Table I). The organism isolated from a leg wound swab from patient D after treatment was still as sensitive to ceftazidime, azlocillin and carbenicillin as the initial isolate and did not show increased synthesis of /Mactamase. In early experiments selection (on agar after initial incubation in broth containing 8 or 6 mg/ of of resistant variants of all four fully sensitive initial isolates produced organisms that showed some reduction in sensitivity to ceftazidime and azlocillin and some elevation of /Mactamase levels (Table I). Many of these variants were neither as resistant to ceftazidime and azlocillin nor produced as much /Mactamase as the resistant isolates that developed in vivo (Table I, variants A.l-Rl and B.l-Rl). However, A.-R2, which was selected on agar without previous incubation in ceftazidime broth, resembled strain A.2 in producing large amounts of /Mactamase constitutively. Selection in broth, without plating on ceftazidime agar, produced resistant variants of strains C. and D. when low concentrations of ceftazidime were used. Organisms that survived the higher concentrations of ceftazidime remained sensitive to the compound and did not exhibit increased /Mactamase synthesis. We were not successful in attempts to isolate resistant variants of strains A. and B. by selection in broth. The resistant variants had maintained their ceftazidime-resistant phenotype after four subcultures on nutrient agar. For each patient, the less sensitive isolates that emerged in vivo had the same pyocine, phage and seotypes as the sensitive organisms isolated before treatment. In isoelectric focusing gels there were five bands of/mactamase activity within the ph range 7 to 8-5, though not all bands were produced by each strain. In all isolates whether sensitive or resistant, the major band was either the first or the second band from the alkaline side of the gel. Minor bands were usually more prominent in resistant variants. None of the isolates produced bands consistent with the production of the plasmid-mediated /Mactamases most commonly found in Ps. aeruginosa (TEM-, TEM-2, PSE- or PSE-; Matthew, 979). In the later experiments with direct plating on DST Agar containing ceftazidime, it was found that the proportion of the bacterial population of the ceftazidimesensitive strains isolated before treatment that was able to form colonies within two days at 37 C in the presence of 2 mg/ ceftazidime was about in 0 7.

5 Patien A B C D NCTC 0662 A. I A.2 A.l-Rl A.-R2 B.I B.2 B.l-Rl C.I C.2 C.l-Rl C.-R2 D.I D.2 D.2-R D.2-R2 Table I. Sensitivity patterns and /Mactamase production (isolated before treatment) (isolated after treatment) (isolated before treatment) (isolated after treatment) (isolated before treatment) (isolated after treatment) (selected in broth containing 2 mg/l (isolated before treatment) (isolated after treatment) (selected in broth containing 2 mg/l Ceftazidime MICs(mg/l) Azlocillin Carbenicillin 28 6 /Mactamase Uninduced activity* Induced junoles per minute per milligram of protein.

6 368 A. King et al. Attempts to measure crypticity for the oranisms from patients A and B were unsuccessful because the detectable /Mactamase activity of intact, uninduced fully sensitive strains was too low to be measureable. The relative hydrolysis rates for various substrates by the /Macatamase from strain A.2 assessed by the ultraviolet spectrophotometric method were as follows: 00 for cephaloridine, 3 for benzylpenicillin, 8-5 for azlocillin, <0 for carbenicillin and ceftazidime. Inactivation, as assessed by the microbiological method, of both ceftazidime and carbenicillin appeared to be at about 0-25% of the rate of inactivation of cephaloridine. Discussion As with the development of resistance to azlocillin that we have previously reported (Shannon et al., 982) development of resistance to ceftazidime occurred in the three patients with a sump of Ps. aeruginosa infection in their chests. However, it is interesting to note that, although the Ps. aeruginosa was not eradicated, resistance to ceftazidime did not develop in the patient with an infected leg wound. In view of the clinical improvement in all the patients, it is not clear whether the relatively small reduction in sensitivity to ceftazidime in the 'resistant' isolates has any clinical implication. Our results suggest that the changes in ceftazidime susceptibility that developed in these patients has a similar mechanism to that elicited in the laboratory in ceftazidime-sensitive strains, though some questions about the mechanism of this resistance remain unanswered. Since azlocillin is clearly a substrate, albeit a poor one, of the Ps. aeruginosa cephalosporinase, it is not difficult to explain the increased resistance to this compound of strains that produce enhanced levels of the enzyme. The decrease in sensitivity to ceftazidime is not so easy to explain. In agreement with previous findings (Labia, Beguin-Billecoq & Guionie, 98) we found that hydrolysis of ceftazidime by the Ps. aeruginosa cephalosporinase was too slow to be measurable with accuracy but that some inactivation was detactable by a microbiological method. It has been suggested that hydrolysis of other new cephalosporins by the Ps. aeruginosa cephalosporinase is physiologically efficient at the low concentrations likely to be encountered in the periplasmic space of the bacterium (Livermore et al., 982). However, it is not yet clear whether /Mactamases hydrolyse or, as suggested by Then & Angehrn (982) merely bind those compounds that are such poor substrates that inactivation can only be detected by microbiological methods. In either case it is the fairly high affinity for ceftazidime (KJ=HM, Labia et al., 98) that is important. We found little reduction in sensitivity to carbenicillin in the strains with elevated levels of/mactamase although according to Then & Angehrn's hypothesis sensitivity to this compound, which has a very high affinity for the Ps. aeruginosa cephalosporinase (K m =0-06iiM, Labia et al., 982) should be reduced by the enhanced /Mactamase synthesis. Despite the high affinity, hydrolysis can be physiologically inefficient if the K^,, is sufficiently low. It should be noted in this context that although we detected slight inactivation of carbenicillin by the microbiological method, Labia and his colleagues (982) did not.

7 Ceftazidime treatment of Ps. aeruginosa 369 The mechanism of the increased /Mactamase synthesis also remains to be answered. The isoelectric focusing results indicate that the resistant strains produce essentially the same enzyme as is induced by cefoxitin in sensitive strains so it seems unlikely that the ceftazidime-resistant strains produce an additional /Mactamase that is not present in the sensitive strains. The increased synthesis of /Mactamase cannot be ascribed to induction of the enzyme by ceftazidime since, as with the development of resistance during therapy with azlocillin (Shannon et al., 982), the continued presence of the /Mactam antibiotic is not needed and the phenomenon is stable at least over the period that we studied. In a few'cases, notably strains A.2 and A.-R2, synthesis of the enzyme seems to be fully derepressed, so, if regulation of /Mactamase synthesis is similar to that of ^-galactosidase synthesis in E. coli, it may be due to a mutation in a gene, analogous to lac I, coding for a repressor. However, another explanation needs to be sought for the majority of cases where the enzyme activity was only about 0%, or less, of that induced by cefoxitin. It is tempting to speculate that this phenotype represents operator constitutive mutants since such mutants in the lactose operon of E. coli are only partially constitutive with synthesis of 0-20% of the maximum amount of /?-galactosidase (Jacob & Monad, 96). It is possible that the phenotype of the less sensitive isolates is the result of a change in permeability to /Mactam antibiotics. However, it seems to us that this is unlikely to be the main cause since susceptibility to carbenicillin was little affected. Unfortunately we were unable to determine crypticity which may have provided useful information. Whatever the precise genetic mechanism the development of such resistance seems to occur with alarming ease both in vivo and in vitro. Though it is not yet known if all strains of Ps. aeruginosa have the ability to develop resistance to ceftazidime, our in-vitro results suggest that most can. Acknowledgement We are grateful to Dr T. L. Pitt of the Division of Hospital Infection, Central Public Health Laboratory, Colindale for serotyping and phage typing the isolates. References Darrell, J. H. & Waterworth, P. M. (969). Carbenicillin resistance in Pseudomonas aeruginosa from clinical material. British Medical Journal Hi, -3. Ervin, F. R. & Bullock, W. E. (976). Clinical and pharmacological studies of ticarcillin in Gram-negative infections. Antimicrobial Agents and Chemotherapy 9, 9-0. Eykyn, S. J. (982). Azlocillin in the treatment of serious infection with Pseudomonas aeruginosa. Journal of Antimicrobial Chemotherapy 9, Govan, J. R. W. (978). Pyocin typing of Pseudomonas aeruginosa. In Methods in Microbiology. Vol. 0 (Bergan, T. & Norris, J. R., Eds) pp Academic Press, London. Gozzard, D. I., Geddes, A. M.. Farrell, I. D., Eykvn, S. J., Phillips, I., Wise, R. & Brown, R. M. (982). Ceftazidime a new extended-spectrum cephalosporin. Lancet i, Gwynn, M. N. & Rolinson, G. (980). Selection of variants of Pseudomonas aeruginosa resistant to beta-lactam antibiotics. Infection 8, Jacob, F. & Monod, J. (96). Genetic regulatory mechanisms in the synthesis of proteins. Journal of Molecular Biology 3,

8 370 A. King et al. King, A., Warren, C, Shannon, K. & Phillips, I. (980). The in-vitro antibacterial activity of cefotaxime compared with that of cefuroxime and cefoxitin. Journal of Antimicrobial Chemotherapy 6, Labia, R., Beguin-Billecoq R. & Guionie, M. (98). Behaviour of ceftazidime towards /Mactamases. Journal of Antimicrobial Chemotherapy 8, Suppl. B, -6. Labia, R., Beguin, R., Masson, J. M. & Kazmierczak, A. (982). Cefsulodin and Pseudomonas aeruginosa: the influence of /Mactamases. In Current Chemotherapy and Immunotherapy. Proceedings of the 2th International Congress of Chemotherapy (Periti, P. & Gialdroni Grassi, G., Eds), pp American Society for Microbiology, Washington, D.C. Livermore, D. M., Williams, R. J., Lindridge, M. A., Slack, R. C. B. & Williams, J. D. (982). Pseudomonas aeruginosa isolates with modified beta-lactamase inducibility: effects on beta-lactam sensitivity. Lancet i, Marier, R. L., Marinaccio, A. T., Sanders, C. V., Aldridge, *L E. & Mitchell, J. W. (982). Susceptibility patterns of bacteria following therapy with piperacillin. Journal of Antimicrobial Chemotherapy 9. SuppL B, Matthew, M. (979). Plasmid-mediated /Mactamases of Gram-negative bacteria; properties and distribution. Journal of Antimicrobial Chemotherapy 5, Shannon, K., King, A. & Phillips, I. (982). Development of resistance to beta-lactam antibiotics during therapy of Pseudomonas aeruginosa infections. Lancet i, 66. Then, R. L. & Angehrn, P. (982). Trapping of nonhydrolyzable cephalosporins by cephalosporinases in Enterobacter cloacae and Pseudomonas aeruginosa as a possible resistance mechanism. Antimicrobial Agents and Chemotherapy 2, 7-7. Wardle, J. K,, Selkon, J. B., Ingham, H. R. & Freeman, R. (98). Ceftazidime in Gramnegative infections: three case reports. Journal of Antimicrobial Chemotherapy 8, Suppl. B, {Manuscript accepted 9 May 983)