sign in sign up
<<
Home , Azlocillin

J. Med. Microbiol. - Vol. 43 (1999, 30G309 0 1995 The Pathological Society of Great Britain and Ireland

ANTI MICROBIAL AGENTS

Mechanisms of resistance to p-lactam amongst isolates collected in the UK in 1993

H. Y. CHEN, ME1 YUAN and D. M. LIVERMORE

Department of Medical Microbiology, The London Hospital Medical College, Turner Street, London El 2AD

Summary. Antimicrobial resistance among 199 1 Pseudomonas aeruginosa isolates collected at 24 UK hospitals during late 1993 was surveyed. Three-hundred and seventy-two of the isolates were resistant, or had reduced susceptibility, to some or all of azlocillin, , , and , and the mechanisms underlying their behaviour were examined. Only 13 isolates produced secondary p-lactamases : six possessed PSE- 1 or PSE-4 enzymes and seven had novel OXA enzyme types. Those with PSE types were highly resistant to azlocillin and carbenicillin whereas those with OXA enzymes were less resistant to these . Chromosomal p-lactamase derepression was demonstrated in 54 isolates, most of which were resistant to ceftazidime and azlocillin although susceptible to carbenicillin and

. p-Lactamase-independent “ intrinsic” resistance occurred in 277 isolates and is believed to reflect some combination of impermeability and efflux. Two forms were seen : the classical type, present in 195 isolates, gave carbenicillin resistance (MIC > 128 mg/L) and reduced susceptibility to ciprofloxacin and to all p-lactam agents except imipenem; a novel variant, seen in 82 isolates, affected only azlocillin, ceftazidime and, to a small extent, meropenem. Resistance to imipenem was largely dissociated from that to other p-lactam agents, and probably reflected loss of D2 porin, whereas resistance to meropenem was mostly associated with intrinsic resistance to penicillins and . Comparison of the present results with those of a similar study in 1982 revealed significant increases in the proportions of isolates with intrinsic resistance or stable derepression (p < 0.01, x2 test). Isolates with secondary p-lactamases appeared significantly rarer than in 1982 (p < 0-01, x2 test), but this comparison was distorted by outbreak strains.

Introduction hospitals (see Acknowledgements) collected up to 100 consecutive non-replicate isolates between 1 Sept. and Most Pseudornonas aeruginosa isolates are susc- 31 Dec., 1993, and sent them to this laboratory. Of eptible to ureido- and carboxy-penicillins, ceftazidime, 21 84 isolates received, 1991 were confirmed as viable , , , imipenem and P. aeruginosa. Resistance rates were 10.4% to meropenem. Nevertheless, resistance to these agents azlocillin 16 mg/L, 11.7 YOto carbenicillin 128 mg/L, can arise by various mechanisms, including mutational 9.6% to ceftazidime 4mg/L, 2.5% to imipenem derepression of the AmpC chromosomal p- 4mg/L and 1.1% to meropenem 4mg/L. For lactamase,l, ’ acquisition of secondary plasmid- or carbenicillin and ceftazidime, these frequencies were transposon-mediated p-lactama~es,~,~reduced per- increased significantly (p < 0.05) compared to those meability5. or multi-drug efflux’’ ’ (see Discussion) found in a similar study in 1982.12Although resistance or, finally, in the case of imipenem, via loss of the D2 to azlocillin, 16 mg/L was not increased significantly porin.’, lo (p > 0.05) there was a significant rise (p < 0.01) in the During 1993 the incidence of resistance proportion of isolates resistant to the drug at amongst P. aeruginosa isolates in the UK was 32 mg/L.12 Imipenem and meropenem were not tested surveyed.” Microbiology laboratories in each of 24 in 1982. In the present study the mechanisms under- lying these resistances were investigated and their current incidences were compared to those found in Received 8 Dec. 1994; revised version accepted 25 March 1995. 1982.

300 DRUG RESISTANCE IN P. AERUGINOSA 301

Materials and methods enzyme detection tests described above, except that 0.01 M, rather than 0.1 M, phosphate buffer was Selection of isolates used. After centrifugation for 2 min at 12000 (1, the supernates were subjected to electrofocusing at Of the 1991 P. aeruginosa isolates collected in the 15 W constant power in polyacrylamide gels 1993 survey," 372 were examined, being selected (220 x 100 x 1 mm) containing ampholines (Pharmacia, because they were resistant to one or more of the Milton Keynes, Buckinghamshire) 2 % v/v of various following drug concentrations: azlocillin 16 mg/L; carbenicillin 128 mg/L; ceftazidime 4 mg/L; pH ranges.15 Enzyme activities were detected by overlaying the gel with 0.5 mM nitrocefin in 0.1 M imipenem 4 mg/L, meropenem 4 mg/L. These con- phosphate buffer, pH 7-0, and observing for the centrations were selected to facilitate discrimination of appearance of pink bands. Identification of p-lacta- normal isolates from those with abnormally reduced mases was by comparison to reference enzymes, run susceptibility, and do not invariably coincide with in tracks adjacent to the test samples. clinical breakpoints." The MIC tests allowing this categorisation were performed on IsoSensitest Agar Assays of /I-lactamase actiziity (Chipath) with inocula of lo4 cfufspot, and have been described already." Reference P. aeruginosa strains Cultures were grown overnight at 37°C in 25-ml included NCTC10662, as the control in the suscep- volumes of nutrient broth with orbital shaking at tibility testing, R20, as a control with known 150 rpm, then diluted into warm (37°C) 250-ml inducibility of AmpC P-lactamase,13and 1405-con and volumes of the same broth. Incubation was continued 2297-con, with derepressed AmpC p-lactamase ex- for 4 h under the previous conditions. Subsequently, pre~sion.'~Transconjugants of P. aeruginosa PU2 1 the cells were harvested by centrifugation at 5000y, and Escherichia coli K-12 553 and J622114served as washed once in 0.1 M phosphate buffer, pH 7.0, re- sources of reference plasmid- and transposon- suspended in 8 ml of the same buffer and subjected to mediated enzymes. two alternate cycles of freezing and thawing. Debris and residual cells were removed by ultracentrifugation Antimicrobial agents for 30 min at 100000 g and 4°C. The supernates were assayed against 0.5 mM p-lactam solutions in 0.1 M Antimicrobial agents were from suppliers as phosphate buffer, pH 7.0, by UV spectrophotometry follows : azlocillin sodium and ciprofloxacin, Bayer, at 37°C. Substrates and wavelengths were : , Newbury, Berkshire ; ampicillin sodium, carbenicillin and carbenicillin, 23 5 nm ; , disodium, clavulanate lithium and sodium, 265 nm; and , 295 nm. SmithKline Beecham, Brockham Park, Surrey; cephaloridine and oxacillin, Sigma ; and P-Lactamase induction assays imipenem, Merck, Hoddesdon, Hertfordshire ; ceftazidime, Glaxo, Greenford, Middlesex ; mero- Sonicates were prepared from logarithmic phase , Zeneca, Macclesfield, Cheshire; and nitrocefin, cells in nutrient broth cultures that had or had not BBL, Cockeysville, MD, USA. been exposed to cefoxitin 500 mg/L for 4 h and were assayed against 0.1 mM nitrocefin in 0.1 mM phosphate /3- Lac tamase de tee tion and typing buffer, pH 7.0.13 Activity was standardised against protein concentration, assayed by the method of Cultures were grown overnight on nutrient agar Lowry et plates at 37"C, then harvested by washing into 1.5-ml volumes of 0.1 M phosphate buffer, pH 7.0. The Data handling and analysis washings were transferred to Eppendorf tubes and sonicated for 2 x 30 s at an amplitude of 12 pm, with Data analysis was with the programs Statview intermediate cooling on ice. Three 25-pl samples of SE + Graphics and STATXACT. Comparisons of each sonicate were transferred to wells in a susceptibility data for different groups of isolates were microtitration plate. One well was supplemented with by x2 or Fisher's exact tests, as appropriate to the 25 pl of 0.3 mM cloxacillin in 0.1 M phosphate buffer, sample numbers. pH 7.0, one with 25 pl of 0.3 mM clavulanate in 0.1 M phosphate buffer, pH 7.0 and the third with drug-free Results buffer. After 10 min at room temperature, 25-pl of 0.3 mM nitrocefin in 0.1 M phosphate buffer, pH 7.0, Categorisation of isolates resistant to penicillins or were added to each well. The appearance of a red ceftazidime colour, indicating p-lactamase activity, was noted Nitrocefin tests in the presence and absence of in comparison to the controls (R20, 1405-con and 0.1 mM cloxacillin or clavulanate were performed on 2297-con). 344 isolates selected as resistant to one or more of azlocillin 16 mg/L, carbenicillin 128 mg/ L or Iso-electric focusing of p-lactamases ceftazidime 4 mg/L. Most (274), like the p-lactamase- P-Lactamase extracts were prepared as for the inducible control strain R20, gave a pink colour in the 302 H. Y. CHEN, M. YUAN AND D. M. LIVERMORE

Table I. Properties of isolates with secondary P-lactamases and of their enzymes

Relative hydrolysis rates tls 0.5 mM B-Lactamase antibiotic

Cb Az Caz Imp Mem PI Alignment PenG Amp Cb Oxa Cld

85 4096 512 4 2 5.7 = PSE-1 100 115 86 10 9 723 4096 512 4 0.125 5.7 = PSE- 1 100 106 93 10 10 I294 4096 512 4 1 5-7 = PSE- 1 100 127 118 11 11 1856 4096 256 2 1 5.3 = PSE-4 100 106 105 10 12 2223 4096 512 4 1 5.7 = PSE- 1 100 101 84 8 8 2312 4096 512 8 2 5.7 = PSE- 1 100 100 101 9 16 PSE- 1 * ... .., 5.7 ... 100 107 99 10 10 PSE-4* ...... 5.3 ... 100 127 89 9 11 520 512 128 4 1 0-5 7.5 = OXA-4 100 874 79 537 67 543 256 64 2 2 0.25 7.5 = OXA-4 100 914 101 453 115 556 512 128 4 1 0.5 7.5 = OXA-4 100 1293 52 549 72 OXA-4* ...... 7.5 ... 100 152 45 960 51 820 256 16 2 1 0.5 7.69 Between 100 31 20 142 37 OXA-6 & -2 898 256 4 0.25 0-25 0.0 3 6.8 & 7.67 100 237 16 173 8 1227 512 128 4 4 1 7-67 Between 100 181 38 685 36 OXA-6 & -7 2241 512 128 32 4 4 7-67 Between 100 71 07 116 14 OXA-6 & -7

Cb, carbenicillin; Az, azlocillin; Caz, ceftazidime; Imp, imipenem ; Mem, meropenem ; PI, iso-electric point; PenG, G ; Amp, ampicillin; Oxa, oxacillin ; Cld, cephaloridine. *Reference standard.

Table 11. Resistance of derepressed isolates in relation to their uninduced levels of a-lactamase activity

Uninduced Number Number Geometric mean MIC (mg/L) Group* P-lactamase activity of remaining U/mg of protein? isolates inducible5 Ceftazidime Azlocillin Carbenicillin Imipenem Meropenem

1 45-100 7 7 8.8 86-1 128-0 1-8 1.o 2 101-300 7 7 13.1 156.0 141-3 1.0 0.7 3 30 1- 1000 23 22 18.6 128.0 120.5 1.5 0.7 4 1001-3000 11 4 26.5 225.7 186-8 1.5 1.2 5 3001-1 0000 3 0 50-8 256.0 161.2 1.3 0.8 6 10 00 1-300 000 3 0 101.0 406.3 406.3 5.0 2.0 Controls R20 8.0 1 1 2 4 64 2 0.25 Survey$ ND 1924 ND 2.3 8.1 66.5 1.1 0.4

*Graded by level of uninduced B-lactamase activity. fl Unit: 1 nmole nitrocefin hydrolysed/min at 37°C and pH 7.0. $Data for all 1991 isolates collected in the survey, except those with stable derepression or secondary B-lactamases. $Inducibility was assessed with cefoxitin 500 mg/L as inducer and isolates giving less than two-fold induction were counted as uninducible. inhibitor-free test only after c. 5-10 min. These lactamase(s), as the AmpC type has the converse reactions were inhibited completely by cloxacillin, but inhibition profile.3+l7 Electrofocusing and hydrolysis not by clavulanate, and were taken to reflect the assays were used to test this inference, and to charac- chromosomal AmpC p-lactamase, which is universally terise the j3-lactamases (table I). PSE-1 enzyme was present in P. aeruginosa and which normally, without found in five isolates and PSE-4 in one (table I). The induction, is expressed ~eakly.~The remaining 70 other seven isolates had OXA enzymes, as resistant isolates gave faster nitrocefin reactions than characterised by stronger activity against oxacillin strain R20; many gave an immediate red colour and all than benzylpenicillin or carbenicillin. Three OXA- did so within 2-3min. For 57 isolates, and the producers (isolates 520, 543 and 556) were probably derepressed control strains 1405-con and 2297-con, replicates; they were from the same hospital and had these reactions were inhibited by cloxacillin but not by the same enzyme. This p-lactamase co-focused with clavulanate; for the other 13 isolates, the reactions OXA-4, but was relatively more active against were inhibited by clavulanate but not cloxacillin. ampicillin and less so against oxacillin. The remaining four isolates had unique OXA-type enzymes, none of which co-focused with any previously described type. Isolates giving strong nitrocefin reactions, inhibited by Electrofocusing revealed two p-lactamase activities clavulanate but not cloxacillin (PI 6.8 and 7-67) in one of these isolates (898), and it This pattern indicated production of secondary p- remains uncertain whether these were satellite bands DRUG RESISTANCE IN P. AERUGINOSA 303 of a single enzyme or were distinct. Further studies on reactions (see above). Of the total of 277 isolates, 195 all these OXA enzymes are in progress. were resistant to carbenicillin at 128 mg/L and had Isolates with PSE- 1 and PSE-4 enzymes were highly reduced susceptibility, compared to modal MICs, to resistant to carbenicillin and azlocillin but remained azlocillin, ceftazidime, meropenem and ciprofloxacin, susceptible to both carbapenems and, with one ex- although not to imipenem (fig. 1). This is the typical ception (isolate 23 12), to ceftazidime (table I). Isolates pattern of “intrinsic resistance” and isolates with this with OXA enzymes were less resistant to penicillins mechanism constituted the upper “tail” of a con- than those with PSE types. One OW-producer (2241), tinuous distribution, not a discrete cluster. This last was resistant to ceftazidime but the others were point is illustrated in fig. 2, which shows that the MICs susceptible to this (table I). of azlocillin, ceftazidime, meropenem and, to some extent, ciprofloxacin were related to those of carbenicillin across the whole spectrum of carbenicillin Isolates giving strong nitrocejn reactions, inhibited by susceptibility and resistance represented in the survey cloxacillin but not clavulanate collection, not just for the 195 isolates in the present This pattern - also observed for the derepressed group. The remaining 82 isolates in the present group controls 1405-con and 2297-con - indicated hyper- were resistant to one or both of ceftazidime 4 mg/L or production of the AmpC chromosomal P-lactamase, azlocillin 16 mg/L and, like isolates with the classical but it might arise via production of another enzyme form of intrinsic resistance, had reduced susceptibility with similar inhibitor susceptibility. It could also occur to meropenem. They nevertheless remained fully sus- trivially, for normal isolates, if nitrocefin tests were ceptible to carbenicillin and ciprofloxacin, as well as to over-inoculated. Electrofocusing and specific activity imipenem (fig. 1). determinations were undertaken to distinguish these possibilities. All 57 isolates had P-lactamases with iso- Isolates resistant to carbapenems electric points > 7.5; only the AmpC enzymes have such high iso-electric points and are inhibited by Resistance to carbapenems was considered separ- cloxacillin but not clavulanate. Quantitative nitrocefin ately, since imipenem resistance in P. aeruginosa is assays confirmed that 54 of the 57 isolates produced at largely independent of that to penicillins and least five-fold higher uninduced levels of P-lactamase cephalosporins.1.2v6-Of the 1991 isolates collected in than strain R20, confirming derepression (table 11). the survey, 59 were resistant to one or both The remaining three isolates had uninduced enzyme carbapenems at 4 mg/L. Of these, 39 were resistant to levels scarcely above that of strain R20 and were imipenem alone, 10 to meropenem alone and 10 to transferred to the group of organisms (below) in which both compounds. Thus, cross-resistance between the resistance was considered to be P-lactamase-indep- two carbapenems was very incomplete. Most (27 of 39) endent. In 40 of the 54 isolates, p-lactamase expression of the isolates resistant to imipenem but not remained inducible by cefoxitin, indicating that dere- meropenem had modal susceptibility to other p- pression was only partial. The remaining 14 isolates lactam agents, except meropenem, to which MICs had the highest p-lactamase specific activities, were raised to 2-4mg/L, as compared to a modal 9&2900-fold greater than that of strain R20, and were value of 0.5 mg/L. None of these 27 isolates gave a totally derepressed (table 11). Resistance to azlocillin strong nitrocefin reaction. The other 12 isolates in the was almost universal amongst the derepressed isolates, group had intrinsic resistance (n = 7) or P-lactamase with 53 of 54 of them resistant to this penicillin at derepression (n = 5) and, consequently, had broader 16 mg/L and 50 resistant at 32 mg/L. Insusceptibility resistances. Of the 20 isolates resistant to meropenem to ceftazidime was also frequent, with 51 isolates alone or to both carbapenems, 17 had intrinsic resistant at 4mg/L and 25 isolates resistant to resistance to penicillins and cephalosporins, two had 16 mg/L. Some derepressed isolates were more widely stable derepression and one was fully susceptible to resistant, 25 were insensitive to carbenicillin 128 mg/L, every other p-lactam agent. Thus, meropenem re- seven insensitive to imipenem 4 mg/L and two insen- sistance, although rarer than imipenem resistance, was sitive to meropenem 4 mg/L. Generally, however, much more strongly associated with intrinsic resist- the level of resistance to carbenicillin and the carba- ance to other P-lactam agents (17 of 20 cf. 7 of 39; was unrelated to the amount of P-lactamase p < 0.01, x2 test). produced without induction, whereas resistance to Attempts to determine whether -resist- azlocillin and ceftazidime was strongly related to ant isolates lacked D2 porin, which provides outer- enzyme activity (table 11). membrane pores permeable to carbapenems but not other p-lactam agents, were unsuccessful (not shown), because comparisons to isogenic carbapenem-suscep- Penicillin- or ceftazidime-resistant isolates that did tible parents were not available. not give strong nitrocejn reactions The 274 isolates initially placed in this category were Sources of isolates with various modes of resistance supplemented with the three initially mis-classified as giving strong, cloxacillin-inhibited, nitrocefin Table I11 summarises the distribution of resistance 304 H. Y. CHEN, M. YUAN AND D. M. LIVERMORE

003 0.06 0.125 0.25 0.5 1 2 4 8 16 32 64 128 256 512 1024 2048

b 40 -I 20 - 00 -

80 - ~ 60 - 40 - r

503 0-060.125 0-25 0-5 1 2 4 8 16 32 64 128 256 512 1024 2048

140 C I

100- 80 - 60 - 40 - 20 - n Ill111 1111 003 0.06 0.125 0.25 0.5 1 2 4 8 16 32 64 128 256 512 1024 2048 MIC (mg/L) Fig. 1. For legend see p. 305 mechanisms in relation to patient type. Chromosomal imipenem 4 mg/L and meropenem 4 mg/L were P-lactamase derepression and intrinsic resistance were examined. These concentrations served to divide significantly more frequent amongst isolates from isolates in which a specific resistance mechanism general in-patients than in those from out-patients, seemed likely from those with normal susceptibility." whereas imipenem resistance and secondary p- MICs two-fold above these values may still denote lactamases were randomly distributed amongst clinical susceptibility, except for carbenicillin. For isolates from these groups of patients. Every resistance ceftazidime, a clinical breakpoint of 16 mg/L (four- mechanism appeared commoner in isolates from in- fold above the present value) is defensible. The tensive care unit patients than in those from general in- mechanism(s) present in the resistant isolates were patients but statistical significance at p < 0-05 was investigated and the frequencies of individual achieved only with respect to imipenem resistance. mechanisms were compared to those found in a similar survey in 1982.12,13Methods of speciation and sus- ceptibility testing were identical in both surveys, but Discussion there were differences in collection strategy. Although 24 hospitals participated in each survey, only 18 Three hundred and seventy-two P. aeruginosa participated in both; moreover, the number of isolates isolates resistant to one or more of azlocillin 16 mg/L, from each hospital varied more widely in 1982, from carbenicillin 128 mg/L, ceftazidime 8 mg/L, two to 330, compared with 17-98 in 1993. DRUG RESISTANCE IN P. AERUGZNUSA 305

100- 80 - 60 - 40 - 20 - - Ill 003 006 0125 0.25 0.5 1 2 4 8 16 32 64. 128 256 512 1024 2048

140- e v) 120- scd

40-

IIIIII 003 006 0125 0.25 0.5 1 2 4 8 16 32 64 128 256 512 1024 2048

0.03 006 0125 0.25 0.5 1 2 4 8 16 32 64 128 256 512 1024 2048 MIC (mg/L)

Fig. 1. MICs of antimicrobial agents for isolates with intrinsic resistance. .,Data for 195 isolates resistant to carbenicillin 128 mg/L; 0,data for 82 carbenicillin-susceptible isolates included because of their resistance to azlocillin 16 mg/L or ceftazidime 4 mg/L, or both. Panel a, carbenicillin; b, azlocillin; c, ceftazidime; d, imipenem; e, meropenem; f, ciprofloxacin. The vertical bars indicate the modal MICs for the collection of 1991 isolates.

P-Lactamase-mediated resistance remained rare. declined insignificantly, from 14 of 24 in 1982 to 8 of 24 Secondary P-lactamases were found in only 13 of the in 1993 (p > 0.05, x2 test). The scarcity of secondary 1991 present isolates and in 47 of 1866 collected in P-lactamases in P. aeruginosa from the UK is apparent 1982. Comparison of these proportions indicates a from other studie~.l'*~~Such enzymes are commoner significant decrease (p < 0.01, x2 test) but few among P. aeruginosa isolates from Francezo and producers were obtained in either year, and analysis Catalonia,21 occurring in 7-12 % of isolates, but was prone to distortion by outbreak strains. In 1982, remain much rarer than in enterobacteria, where they one hospital contributed 10 indistinguishable isolates are produced by 30-50 70 of isolates.22Moreover, the with PSE-4 enzyme; another sent 10 with an OXA- predominant enzymes differ amongst bacterial type, possibly OXA-6, and three further hospitals sent families: PSE-1 and PSE-4 types are commonest in P. pairs of likely replicates. In the present study, one aer~ginosa,'~.20- 21,23 as confirmed here, whereas the hospital sent three isolates (520, 543 and 556: table I) TEM and SHV types predominate in enterobacteria.22 with the same novel OXA enzyme. In this situation it A scatter of OXA enzymes was found here (table I), as is better to compare the proportions of hospitals in previous surveys of P. aer~ginosa;~~*~~+~~but no sending isolates with secondary p-lactamases ; this individual member of this enzyme class is frequent in 306 H. Y. CHEN, M. YUAN AND D. M. LIVERMORE

512-

256 -

128- 64 - P 32 - n =& 16-

vE 51 8- 5 5 4- E .-0 2- z g 1- 0 0.5-

0.25 -

0-12-

006 -

003 -

I I I I I 1- 1 16 32 64 128 256 512 21024 Carbenicillin MiC (mg/L) Fig. 2. Inter-relationship of MICs of different antimicrobial agents with those of carbenicillin for all P. aeruginosa isolates collected in the survey, except those with stable derepression or secondary /?-lactamases (n = 1924). The 195 isolates with classical intrinsic carbenicillin resistance fall to the right-hand side of the graph (MIC 2 256 mg/L). Lines are: azlocillin 0,ceftazidime +, ciprofloxacin 0,imipenem 0. meropenem 0.

Table 111. Distribution of resistance mechanisms amongst P. aeruginosa isolates from various types of patients

~~ ~ ~ ____ ~ ~__ Mechanisms ICU Other in-patient Ou t-pa tient (number of isolates) (134) (1041) (797)

Secondary /?-lactamase (1 3) 4 0.056 7 0.62 2 Stably derepressed (54) 10 0.052 34 0.0 104 10 Intrinsically resistant (273*) 27 0.34 162 0.0036 84 All imipenem-resistant (49) 13 < 0-0001 19 1-0 17 Imipenem-resistant and 4 0.056 7 1.o 4 intrinsically resistant (I 5) Imipenem-resistant and stably 2 0.28 4 0.8 1 derepressed (7)

*Does not total 277, as elsewhere, since sources of four intrinsically resistant isolates were unrecorded. tIsolates from non-ICU in-patients were taken as a reference group to which isolates from other patient groups were compared by Fisher’s exact tests, with the p values thereby obtained being adjusted by the Bonferroni correction (i.e., doubled in the present cases) to compensate for the fact that two comparisons (one for ICU isolates and one for out-patient isolates) were being made to a single reference group (the in-patient isolates). The groups being compared are indicated by the arrows. the species. A difference from the 1982 survey was that Isolates with secondary p-lactamases mostly were PSE-1 replaced PSE-4 as the commonest enzyme. resistant to the penicillins but susceptible to Notably, all the 19 PSE-4 producers collected in 1982 ceftazidime and the carbapenems, reflecting the known were of serogroup 0:I1 and, although from 10 activity of PSE and OXA en~yrnes.~’One isolate with hospitals, most gave similar phage lysis a novel OXA enzyme (2241) was resistant to These findings suggested clonal spread of a resistant ceftazidime (MIC 32 mg/L) but it is doubtful whether strain, which has since disappeared. the resistance involved the P-lactamase which, when DRUG RESISTANCE IN P. AERUGINOSA 307

12 -I

Cb Azl Azl Cat Caz Imp Imp Mem Mem >128 >16 >32 >4 >16 >4 >a >4 >8 Resistance taken as MIC (mg/L) Fig. 3. Distribution of resistance mechanisms amongst isolates resistant to the antimicrobial concentrations (mg/L) indicated on the horizontal axis : secondary p-lactamase production H ; stable derepression of chromosomal p-lactamase El ; intrinsic resistance affecting all penicillins and cephalosporins 0; intrinsic resistance affecting only azlocillin and ceftazidime 0; imipenem-specific resistance probably caused by loss of D2 porin ; meropenem-specific resistance m. Abbreviations are: Cb, carbenicillin ; Az, azlocillin; Caz, ceftazidime; Imp. imipenem and Mem, meropenem. Data are analysed with regard to the low breakpoints used elsewhere in this study, and with respect to high breakpoints which may be considered as the limits of clinical susceptibility.” It is emphasised that detection of a mechanism does not prove it to be the sole cause of resistance : e.g., although jl-lactamase derepression and intrinsic resistance were present in some imipenem-resistant isolates they do not cause this resistance;’.8 likewise stable derepression does not cause meropenem resistance’ and it is doubtful whether the OXA p-lactamase found in one ceftazidime-resistant isolate (2241, see table I) was responsible for that resistance. extracted, had only feeble activity against 0.5 mM antibiotic selection pressure is greatest. The degree of ceftazidime (not shown). resistance of derepressed organisms to azlocillin and The survey yielded 54 stably derepressed isolates, ceftazidime reflected the amount of enzyme produced defined as those with at least five-fold higher un- without induction (table 11) whereas resistance to the induced AmpC P-lactamase activity than strain R20. more stable agents, carbenicillin, imipenem and Nineteen isolates were classed as derepressed in the meropenem, was not related to enzyme quantity. 1982 survey, only 17 of which would meet the present The largest group of p-lactam-resistant isolates were definiti~n.’~This increase (17 of 1866 cf. 54 of 1991) those lacking untoward p-lactamase activity and with was significant (p < 0.01; x2 test). Moreover, the inter-related resistance to penicillins, ceftazidime, change was not a consequence of multiple inclusion of ciprofloxacin and meropenem, although not to outbreak strains; the isolates from the 1982 survey imipenem (fig. 1). This antibiogram profile is well- came from only 8 of 24 participating hospitals and known in P.aeruginosa, and has been termed “intrinsic included one pair of likely replicates, whereas those re~istance”.~,’,~Its mechanism was long assumed to from the present survey were from 19 of 24 hospitals be impermeability, because other mechanisms were and included three pairs of likely replicates and one set contra-indicated and because it was difficult to en- of four. The proportion of hospitals sending de- visage anything else that could compromise such repressed isolates had increased compared to 1982 diverse drugs.’ However, recent biophysical data (p < 0.01, xz test). These increases are not surprising. suggest increased multi-drug efflux as the mechan- and extended-spectrum cephalo- ism,’~’and a multi-drug efflux operon, cloned from the sporins, used increasingly since the early 1980s, are species, was shown to enhance resistance to labile to AmpC enzymes but are weak inducers of their quinolones, tetracycline and chl~ramphenicol.~~ synthesis, and so select derepressed mutants from Nevertheless, it remains surprising that such a system inducible pop~lations.l-~, which should affect /I-lactam agents, which have targets were the main antipseudomonal p-lactams before 1980, external to the cytoplasmic membrane. In the present are more stable and less selective.’ Although de- survey, 195 of 1991 isolates were inferred to have repressed organisms remain rare it is telling that most classical intrinsic resistance, with carbenicillin MICs were from in-patients and, especially, those in intensive > 128 mg/L and with reduced susceptibility to care units (table 111). These are the patients in whom ceftazidime, azlocillin, meropenem and ciprofloxacin. 308 H. Y. CHEN, M. YUAN AND D. M. LIVERMORE

This proportion compares to 131 of 1866 in 1982 and that the occurrence of a mechanism does not prove it represents a significant increase (p < 0.01; x2 test). to be the sole cause of resistance, and cases where the Eighty-two further carbenicillin-susceptible (MIC < mechanisms found do not account for observed 128 mg/L) isolates were resistant to either or both of resistance are indicated in the legend. It is possible also azlocillin and ceftazidime and, mostly, had reduced that some isolates may have possessed further, un- susceptibility to meropenem (fig. 1). No similar group detected mechanisms, such as nitrocefin-negative p- was recognised in 1982 and their mechanism of lactamases. More generally, the MICs of p-lactam resistance, although uncertain, seems most likely to agents reflect the balance of penicillin-binding protein entail some combination of impermeability or efflux. sensitivity, P-lactamase activity, permeability and The final mechanisms considered were those efflux and ascribed “ resistance mechanisms ” are affecting carbapenems. Imipenem resistance in P. merely those factors whose change shifts the MIC aeruginosa commonly emerges via mutational loss of beyond the normal range. Finally, it should be stressed D2 outer-membrane protein.’. lo This does not affect that even “susceptible’’ P. aeruginosa strains have non-carbapenems, which cannot traverse the narrow considerable defences against /?-lactam agents, being pores formed by this specialised porin,” and generally 100-fold less susceptible than those mutants which does not raise meropenem MICs above 2-4mg/L. lack both /I-lactamase inducibility and multi-drug Most or all the 49 isolates resistant to imipenem efflux and which are hyperpermeable.’. 8, 26 4 mg/L seem likely to have lacked this porin, but this could not be confirmed in the absence of isogenic We are grateful to the following individuals and their colleagues for their participation in the study: Dr A. W. Anderson, York porin-producing controls (not shown). Although a few District Hospital; Dr P. Cockcroft, St Mary’s Hospital, Ports- imipenem-resistant isolates had derepression of AmpC mouth; Professor B. 1. Duerden, University Hospital of Wales, enzymes or intrinsic resistance, neither of these Cardiff; Dr F. X. S. Emmanuel, City Hospital, Edinburgh; Dr I. Gould, Aberdeen Royal Infirmary; Drs J. Hood and G. F. S. mechanisms generally affects imipenem.’ Resistance to Edwards, Glasgow Royal Infirmary; Dr A. J. Howard, Ysbyty meropenem 4 mg/L only partly overlapped that to Gwynedd, Bangor (Wales); Mr M. R. Howard, Manchester Royal imipenem, being seen in 10 of 49 imipenem-resistant Infirmary; Mr R. P. Human, Derriford Hospital, Plymouth; Dr H. R. Ingham, Newcastle General Hospital; Dr M. Keaney, Salford isolates and in a further 10 imipenem-susceptible Royal Hospitals NHS Trust (Hope Hospital), Salford; Dr P. G. organisms. In 17 of these 20 organisms, the Murphy, Belfast City Hospital; Professor A. Percival, Royal Liver- meropenem resistance was associated with intrinsic- pool University Hospital; Dr G. L. Ridgeway, University College Hospital, London; Dr M. H. Robertson, Herts and Essex Hospital. type resistance to other /?-lactam agents. Both the Bishop’s Stortford; Dr N. Saunders, Hammersmith Hospital, present (figs. 1 and 2) and previous studies2g6support London; Dr E. T. M. Smyth, Royal Victoria Hospital, Belfast; the view that the intrinsic resistance mechanism affects Professor D. C. E. Speller and Dr R. Jones, Bristol Royal meropenem, although frank resistance is rarely con- Infirmary; Dr D. J. Waghorn, Wycombe General Hospital, High Wycombe; Dr A. B. White, Raigmore Hospital, Inverness; Dr ferred. W. D. White, Barking Hospital; Dr P. J. Wilkinson, University The distribution of mechanisms amongst isolates Hospital, Nottingham; Professor J. D. Williams, Royal London resistant at either the low microbiological breakpoints Hospital; Dr R. Wise, Dudley Road Hospital, Birmingham. We are grateful to Zeneca Pharmaceuticals (UK) for financial sup- used throughout this Discussion or at higher port and to Mr G. Byttebier ofZeneca Pharmaceuticals NV Benelux, breakpoints, thought to represent the limits of clinical for his advice and help on statistical analyses. susceptibility,” is summarised in fig. 3. It is emphasised

References 7. Li X-Z, Livermore DM, Nikaido H. Role of efflux pump(s) in intrinsic resistance of Pseudomonas aeruginosa : resistance 1. Livermore DM, Yang Y-J. p-Lactamase lability and inducer to tetracycline, chloramphenicol, and norfloxacin. power of newer B-lactam antibiotics in relation to their Anrimicrob Agents Chemother 1 994 ; 38 : I 732- I 74 I. activity against p-lactamase-inducibility mutants of 8. Li X-Z, Ma D, Livermore DM. Nikaido H. Role of efflux Pseudomonas aeruginosa. J Infect Dis 1987; 155 : 775-782. pump(s) in intrinsic resistance of Pseudornonas aeruginosa : 2. Livermore DM, Yang Y. Comparative activity of meropenem resistance to tetracycline, chloramphenicol, and against Pseudomonas aeruginosa strains with well- norfloxacin. Antimicrob Agenrs Chemother 1994; 38: characterized resistance mechanisms. J Antimicrob 1 732- 1 74 1. Chemother 1989; 24 Suppl A: 149-159. 9. Buscher KH, Cullmann W, Dick W, Opferkuch W. Imipenem 3. Livermore DM. p-Lactamases of Pseudomonas aeruginosa. In : resistance in Pseudomonas aeruginosa resulting from Schonfeld H (series ed) Antibiotics and chemotherapy, vol. diminished expression of an outer membrane protein. 44;Homma JY, Tanimoto H, Older IA, Horby N, Doring Antimicrob Agents Chemother 1986 ; 31 : 703-708. G (vol. eds) Pseudomonas aeruginosa in human diseases. 10. Trias J, Nikaido H. Outer membrane protein D2 catalyzes the Basel, Karger. 199 1 : 2 1 5-220. facilitated diffusion of carbapenems and penems through 4. Jacoby GA, Sutton L. Activity of @-lactamantibiotics against the outer membrane of Pseudomonas aeruginosa. Pseudomonas aeruginosa carrying R-plasmids determining Antimicrob Agents Chemother 1990; 34: 52-57. different p-lactamases. Antimicrob Agents Chemother 1979 ; 11. Chen HY, Yuan M, Ibrahim-Elmagboul IB, Livermore DM. 16: 243-245. Susceptibility and resistance to antimicrobials amongst 5. Livermore DM. Penicillin-binding proteins, porins and outer- Pseudomonas aeruginosa isolates collected in the UK in membrane permeability of carbenicillin-resistant and 1993. J Antimicrob Chemother 1995; 35: 521-534. -susceptible strains of Pseudomonas aeruginosa. J Med 12. Williams RJ, Lindridge MA, Said AA, Livermore DM, Microbiol 1984; 18: 261-270. Williams JD. National survey of antibiotic resistance in 6. Masuda N, Ohya S. Cross-resistance to meropenem, , Pseudomonas aeruginosa. J Antimicrob Chemother 1984 ; and quinolones in Pseudomonas aeruginosa. Antimicrob 14: 9-16. Agents Chemother 1992; 36: 1847-1851. 13. Williams RJ, Livermore DM, Lindridge MA, Said AA, DRUG RESISTANCE IN P. AERUGINOSA 309

Williams JD. Mechanisms of p-lactam resistance in British Pseudomonas aeruginosa and susceptibility against 0- isolates of Pseudomonas aeruginosa. J Med Microbioll984; lactam antibiotics. Chemioterapia 1985 ; 4: 3M2. 17: 283-293. 21. Tirado M, Roy C, Segura C, Reig R, Hermida M, Foz A. 14. Livermore DM, Corkill JE. Effects of CO, and pH on inhibition Incidence of strains producing plasmid determined of TEM-1 and other B-lactamases by penicillanic acid p-lactamases among carbenicillin resistant Pseudomonas sulfones. Antimicrob Agents Chemother 1992 ; 36: aeruginosa. J Antimicrob Chemother 1986; 18: 453458. 1870-1 876. 22. Liu PYF, Gur D, Hall LMC, Livermore DM. Survey of the 15. Matthew M, Harris AM, Marshall MJ, Ross GW. The use of prevalence of p-lactamases amongst 1000 Gram-negative analytical isoelectric focusing for detection and bacilli isolated consecutively at the Royal London Hos- identification of beta-lactamases. J Gen Microbiol 1975; pital. J Antimicrob Chemother 1992; 30: 429-447. 88: 169-178 23. Stobberingh EE, Houben AW, Van Boven CPA. Distribution 16. Lowry OH, Rosebrough NJ, Farr AL, Randall RJ. Protein and characterization of P-lactamase-producing strains measurement with the Folin phenol reagent. J Biol Chem isolated in the southern part of the Netherlands. In: Periti 1951 ; 193: 265-275. P, Grassi GG (eds) Current chemotherapy and 17. Bush K. Characterization of #I-lactamases. Classification of immunotherapy. Proceedings of the 12th International p-lactamases: groups 1, 2a, 2b, and 2b. Classification of Congress of Chemotherapy. Washington DC, American p-lactamases: groups 2c, 2d, 2e, 3, and 4. Antimocrob Society for Microbiology. 1982: 747-748. Agents Chemother 1989 ; 33 : 259-263 ; 264-270; 27 1-276. 24. Livermore DM, Pitt TL, Jones CS, Crees-Morris JA, Williams 18. King JD, Farmer T, Reading C, Sutherland R. Sensitivity to RJ. PSE-4 8-lactamase : a serotype-specific enzyme in Pseudomonas aeruginosa. Med ; 19 : 45-53. carbenicillin and , and the beta-lactamases of J Microbioll985 25. Poole K, Krebes K, McNally C, Neshat S. Multiple antibiotic Pseudomonas aeruginosa in the UK in 1978-1979. J Clin resistance in Pseudomonas aeruginosa: evidence for in- Pathol 1980; 33: 297-301. volvement of an efflux operon. J Bacteriol 1993; 175: 19. Chen HY, Bonfiglio G, Allen M et al. Multicentre survey of the 73 63-7372. comparative in-vitro activity of / 26. Zimmermann W. Penetration of p-lactam antibiotics into their against from hospitalized patients in the British target enzymes in Pseudomonas aeruginosa : comparison of Isles. J Antimicrob Chemother 1993; 32: 247-266. a highly sensitive mutant with its parent strain. Antimicrob 20. Thabaut A, Philippon A, Meyran M. p-Lactamases of Agents Chemother; 18: 94100.