Italian Survey on Comparative Levofloxacin Susceptibility in 334 Clinical Isolates of Pseudomonas aeruginosa

Abstract

A national survey on susceptibility patterns of 334 Pseudomonas aeruginosa isolates from intensive care units and hematology and oncology wards from 13 Italian hospitals compared the in vitro activity of levofloxacin, an injectable oral fluoroquinolone, to those of ciprofloxacin, ofloxacin, ceftazidime, imipenem, amikacin, and gentamicin. Amikacin and imipenem had the best susceptibility profiles. The activity of levofloxacin was superior to those of the other quinolones and was comparable to that of ceftazidime. The effect of levofloxacin in vitro on P. aeruginosa clinical isolates suggests that further clinical investigations are warranted.

Levofloxacin, the optical S(−) isomer of ofloxacin, is a fluoroquinolone with a broad spectrum of antimicrobial activity, including activity against gram-negative and gram-positive bacteria and atypical respiratory pathogens (3). The mechanism of action of levofloxacin, like that of other fluoroquinolones, involves inhibition of DNA gyrase (19) and topoisomerase IV (6), resulting in inhibition of bacterial DNA replication and transcription. Recent investigations have produced evidence for almost double the activity of levofloxacin in vitro with respect to that of ofloxacin, due to ofloxacin’s containing only 50% of the bacteriologically active S(−) isomer (14).

Although Pseudomonas aeruginosa is included in the levofloxacin activity spectrum, some reported susceptibility values were controversial and were sometimes higher than those reported for other fluoroquinolones (5, 20, 23).

Thus, we found it interesting to assess the in vitro activity of levofloxacin against a large number of P. aeruginosa clinical isolates in comparison with those of other fluoroquinolones and antipseudomonal agents. We collected strains in critical environments, such as intensive care units of hematology and oncology wards, to initially match levofloxacin with the worst clinical scenario in terms of bacterial resistance. To reduce variability, the susceptibility determinations were centralized.

(These results were partly presented at the 8th International Congress of Microbiology and Infectious Diseases, Lausanne, Switzerland, 25 to 28 May 1997.)

The study compounds were obtained from the following manufacturers: levofloxacin and ofloxacin, Hoechst Marion Roussel (Romainville, France); imipenem, Merck Sharp & Dhome (Rome, Italy); amikacin, gentamicin, and ceftazidime, Sigma (Milan, Italy); and ciprofloxacin, Mast (Merseyside, United Kingdom).

Microbiology laboratories from 13 general hospitals distributed throughout Italy (6 in the northern region, 4 in the southern region, and 3 in the central region) were requested to select at least 25 unique P. aeruginosa clinical isolates from intensive care units and hematology and oncology wards, regardless of the site of isolation. The strains were then collected and reidentified by conventional methods and tested with the study drugs. P. aeruginosa ATCC 27853 was examined for quality control at the beginning of the study and was reexamined periodically throughout the study.

MIC determinations were performed by conventional broth microdilution procedures in 0.1-ml volumes of Mueller-Hinton broth supplemented with the cations calcium and magnesium to approach free physiological concentrations. A final inoculum of 5 × 10⁵ CFU/ml was used, as suggested by the National Committee for Clinical Laboratory Standards (NCCLS) (15). Following 16 to 20 h of aerobic incubation at 37°C overnight, the trays were examined for growth. MIC results were recorded as the dilution value at which no visible growth occurred. The determination of all MICs was performed in two separate sets of experiments during the study period. The NCCLS breakpoints for susceptibility, intermediate susceptibility, and resistance were followed for all drugs, including levofloxacin (17).

A total of 334 P. aeruginosa clinical isolates were collected between September 1996 and February 1997 from intensive care units (261 [78.1%]), oncology wards (49 [14.7%]), and hematology wards (24 [7.2%]), and the MICs were determined (Table 1). Respiratory isolates were the most frequently encountered (188 [56.2%]), and 55.3% of these strains were obtained from bronchial or tracheal aspirates. Levofloxacin and ciprofloxacin MIC frequency distributions are illustrated in Fig. 1.

TABLE 1.

In vitro antimicrobial susceptibility of 334 P. aeruginosa clinical isolates according to NCCLS criteria^a

Antimicrobial agent	MIC (μg/ml)b			% of isolates susceptible
	Range	50%	90%
Amikacin	<0.125–>64	2	32	86.2
Gentamicin	<0.125–>64	2	>64	63.2
Ofloxacin	0.125–>64	8	>64	28.4
Ciprofloxacin	0.125–>64	4	>64	17.1
Levofloxacin	0.06–>32	4	>32	48.2
Imipenem	<0.125–64	0.5	8	86.8
Ceftazidime	0.5–>64	8	>64	55.1

^aSusceptibility break points (micrograms per milliliter): amikacin, ≤16; gentamicin, ≤4; ofloxacin, ≤2; ciprofloxacin, ≤1; levofloxacin, ≤2; imipenem, ≤4; ceftazidime, ≤8. 

^b50% and 90%, MIC₅₀ and MIC₉₀, respectively. 

FIG. 1.

Levofloxacin and ciprofloxacin MIC frequency distribution for 334 P. aeruginosa clinical isolates. N., number.

Table 2 shows the patterns of resistance to the different antimicrobials found during the study. The results confirmed imipenem and amikacin as the most active antimicrobial agents in vitro, while ciprofloxacin showed a higher number of either intermediate or resistant strains. The activity of levofloxacin was superior to that of other 4-quinolones, ofloxacin, and ciprofloxacin and was comparable to that of ceftazidime, a reference antipseudomonal agent. Similar data were observed when the clinical isolates were divided according to the different isolation wards. The lowest overall susceptibility percentages were observed for wound isolates and blood cultures.

TABLE 2.

Resistance of isolates to various antimicrobial agents according to the different clinical specimens

Antimicrobial agent	% of isolates with resistancea
	Total (n = 334)			Respiratory tract (n = 188)			Urine (n = 53)			Blood cultures (n = 34)			Catheters (n = 10)			Wounds (n = 21)			Others (n = 28)
	S	I	R	S	I	R	S	I	R	S	I	R	S	I	R	S	I	R	S	I	R
Amikacin	86.2	4.8	9.0	88.9	2.1	9	85	9.4	5.6	79.4	8.8	11.8	80		20	71.4	9.5	19.1	92.9		7.1
Gentamicin	63.2	6.3	30.5	67	8	25	60.4	3.8	35.8	58.8	2.9	38.3	70		30	42.9	9.5	47.6	60.7	3.6	35.7
Ofloxacin	28.4	18.3	53.3	29.2	20.2	50.6	28.3	7.6	64.1	29.4	14.7	55.9	20	20	60	14.3	14.3	71.4	32.1	21.4	46.5
Ciprofloxacin	17.1	21.5	61.4	20.7	22.3	57	11.3	17	71.7	17.6	26.5	55.9	20	10	70	14.3	4.7	81	14.3	25	60.7
Levofloxacin	48.2	10.2	41.6	51	13.3	35.7	39.6	3.8	56.6	44.1	11.8	44.1	40		60	28.6	4.7	66.7	57.1	3.6	39.3
Imipenem	86.8	5.7	7.5	83.5	6.4	10.1	92.4	1.9	5.7	91.2	2.9	5.9	100			85.7	14.3		89.3	7.1	3.6
Ceftazidime	55.1	7.2	37.7	58	3.2	38.8	64.2	9.4	26.4	32.4	8.8	58.8	40	20	40	47.6	9.5	42.9	53.6	21.4	25

^aS, susceptible; I, intermediate resistance; R, resistant. 

Various degrees of cross-resistance were found among levofloxacin-resistant isolates and isolates resistant to gentamicin, ceftazidime, amikacin, ciprofloxacin, and imipenem, and the results are summarized in Table 3. Thirteen strains were concomitantly resistant to levofloxacin, gentamicin, ceftazidime, and amikacin, and 1 isolate’s resistance pattern also included imipenem (data not shown).

TABLE 3.

Frequency of cross-resistance to antimicrobial agents in 139 levofloxacin-resistant strains

Antimicrobial agent	No. (%) of isolates with cross-resistancea
Imipenem	11 (8)
Amikacin	27 (19)
Ceftazidime	72 (51)
Gentamicin	90 (65)
Ciprofloxacin	132 (94)

^an = 139. 

P. aeruginosa is widely recognized as an important nosocomial pathogen with increasing resistance to various antimicrobial agents and is frequently associated with severe infections in hospitalized patients (18). Continuous clinical usage in intensive care units and hematology and oncology wards of antipseudomonal agents even in empirical treatment has probably contributed to the emergence and spread of multiresistant clinical isolates with a combination of old and new resistance mechanisms, such as efflux (11, 12).

The aim of the present study was to survey the in vitro activity of levofloxacin, a new parenteral and oral fluoroquinolone on P. aeruginosa clinical isolates obtained from all over Italy in comparison to those of other antimicrobial agents. We centralized the strains and followed NCCLS recommendations in order to reduce variability and systematic errors, as sometimes reported when disc tests (7) or automated methods (1) are used. The overall results showed a high level of resistance in the study strains. Organisms had the highest susceptibility to amikacin and imipenem: 86.2 and 86.8%, respectively. Because the in vitro overall susceptibilities of organisms to ceftazidime and gentamicin were about 55.1 and 63.2%, respectively, and that of organisms to ciprofloxacin was only 17.1%, a decreased in vitro activity of these drugs against P. aeruginosa among Italian intensive care units and hematology and oncology wards should be taken into account. According to the accepted selective pressure theory, it could be argued that an extensive usage of ceftazidime and ciprofloxacin in the wards examined in Italy either for therapy or antibacterial prophylaxis might have contributed to such a condition. A recent survey on P. aeruginosa carried out in Italy reported high resistance figures, but the figures were lower than the ones we found as far as amikacin, imipenem, ceftazidime, and ciprofloxacin are concerned (2).

However, that study took into account consecutive isolates from all wards (data not shown), and, indeed, the worst figures were observed in intensive care units and hematology wards.

In the present study, a notable difference between the susceptibility to levofloxacin (48.2%) and the susceptibility to ciprofloxacin (17.1%) was observed. Such a difference should be mainly addressed in terms of the different NCCLS susceptibility breakpoints (2 μg/ml for levofloxacin and 1 μg/ml for ciprofloxacin), which are supposed to integrate the distribution of MICs of a large sensitive and resistant bacterial population, the concentrations of the drug in blood and tissue, and the comparisons of in vitro and in vivo results (16). In the present investigation, the MIC at which 50% of the isolates are inhibited (MIC₅₀), MIC₉₀, MIC range, and MIC distribution of levofloxacin and ciprofloxacin were found to be comparable. The higher MIC frequency is around 2 μg/ml, which includes strains susceptible to levofloxacin and strains intermediate with susceptibility to ciprofloxacin, leading to the susceptibility difference between the two compounds. Nevertheless, a certain higher in vitro activity of levofloxacin compared to that of ciprofloxacin was confirmed, because had the breakpoints been the same (either 1 or 2 μg/ml), 10% more of the strains would have been susceptible to levofloxacin. An interesting levofloxacin in vitro susceptibility pattern was confirmed by also analyzing the isolates according to the ward and/or the site of isolation. The susceptibility to levofloxacin of the isolates in blood was higher, not only than that of ofloxacin and ciprofloxacin, but even that of ceftazidime. Also, among the respiratory isolates, half of which were obtained by aspiration procedures, levofloxacin performed well. More than half of the levofloxacin-resistant strains were also resistant to gentamicin and ceftazidime, while nearly all were resistant to ciprofloxacin. Although a deeper comprehension of the resistance mechanisms should be reached before any conclusion can be drawn, only partial cross-resistance seems to exist between ciprofloxacin and levofloxacin, because more than 35% of the ciprofloxacin-resistant strains were still susceptible or had intermediate resistance to levofloxacin (data not shown). A partial explanation may be related to the selection of different P. aeruginosa efflux-type systems exerted by quinolones, as recently reported by Kohler et al. (10).

Preliminary results from experimental infections (22) and clinical trials (4, 8, 9) provide evidence for the use of levofloxacin in pseudomonal infection, at least as an empiric treatment, although the MICs can be on the border of susceptibility. The good pharmacokinetic properties of levofloxacin (high peak and area under the concentration-time curve) (21) and its bactericidal activity against P. aeruginosa (13) probably contribute to its efficacy. Thus, on the basis of our data, we can conclude that in vitro susceptibility of P. aeruginosa to levofloxacin in Italy is better than that to ofloxacin and ciprofloxacin in critical environments, such as intensive care units and hematology and oncology wards, and that further clinical investigations on this subject are warranted.