Japanese Journal of Ophthalmology

, Volume 56, Issue 5, pp 497–501

Effectiveness of 1.25 % povidone–iodine combined with topical levofloxacin against conjunctival flora in intravitreal injection

Authors

    • Department of Ophthalmology E7Osaka University Graduate School of Medicine
  • Miki Sawa
    • Department of Ophthalmology E7Osaka University Graduate School of Medicine
  • Motokazu Tsujikawa
    • Department of Ophthalmology E7Osaka University Graduate School of Medicine
  • Fumi Gomi
    • Department of Ophthalmology E7Osaka University Graduate School of Medicine
  • Naoyuki Maeda
    • Department of Ophthalmology E7Osaka University Graduate School of Medicine
  • Kohji Nishida
    • Department of Ophthalmology E7Osaka University Graduate School of Medicine
Clinical Investigation

DOI: 10.1007/s10384-012-0160-5

Cite this article as:
Ikuno, Y., Sawa, M., Tsujikawa, M. et al. Jpn J Ophthalmol (2012) 56: 497. doi:10.1007/s10384-012-0160-5

Abstract

Purpose

To compare the effect of a reduced concentration (1.25 %) of povidone–iodine (PI) eye drops, combined with 0.5 % topical levofloxacin (LVFX), with that of the standard of care (5 % PI) on conjunctival flora before intravitreal injections (IVT).

Methods

A prospective, randomized, single-blind clinical trial. One hundred eyes from 100 patients who underwent IVT were included. Eyes were randomly assigned to two groups and underwent different preparatory procedures before therapeutic IVT. LVFX drops were instilled three times every 5 min and then either 1.25 % PI drops were instilled three times every few minutes in group A or 5 % PI drops were instilled three times every few minutes in group B, the control. Conjunctival flora and the injection needles were collected for culture both before and after preparation. The number of positive cultures, the species isolated, and the minimum inhibitory concentrations (MIC) of the antibiotics were assessed.

Results

The pretreatment culture was positive for 45 eyes (86.5 %) in group A (n = 52) and for 37 eyes (77.1 %) in group B (n = 48). After the preparatory procedure, the cultures were positive for 25 eyes (48.1 %) in group A and for 27 eyes (56.3 %) in group B (P = 0.43). After preparation, the number of positive cultures decreased by 46.7 % in group A and by 33.4 % in group B (P = 0.48). No needle samples were contaminated in either group. The most common isolate in both groups after preparation for surgery was Propionibacterium acnes. Among the different antibiotics, vancomycin and oxacillin had the lowest MIC >90 for overall isolates.

Conclusions

1.25 % PI with LVFX is as effective as 5 % PI. P. acnes was the most common conjunctival flora detected. Vancomycin has been confirmed as the best choice for treating infectious endophthalmitis after IVT.

Keywords

Intravitreal injectionLevofloxacinEndophthalmitisPovidine iodine

Introduction

Intravitreal therapeutic injections (IVT) of retinal reagents, for example anti-vascular endothelial growth factor (VEGF) are increasingly used to treat a variety of retinal pathologies including age-related macular degeneration, choroidal neovascularization associated with pathological myopia, diabetic macular edema, and macular edema from retinal vaso-occlusive diseases [15].

Acute infectious endophthalmitis, a serious complication after IVT, reportedly occurs after approximately 0.02–0.05 % of injections [611]. The optimum pre-IVT preparation is controversial. Topical prophylactic antibiotics instilled before IVT reduce the number of positive cultures isolated from specimens from the conjunctival sac; however, its effectiveness in reducing the risk of post-IVT endophthalmitis is still questionable [1214]. Isolation of bacterial specimens from corneal transplantation donors is also common [15].

There is a consensus that instillation of povidone–iodine (PI) prevents postoperative endophthalmitis [1618]. This idea is supported both by a prospective study [16] and by a review of evidence-based prophylaxis after cataract surgery [17]. Topical PI is the recommended treatment in many countries, and it is suggested that application of PI immediately before IVT with a wide-spectrum antibiotic, combined with eye draping to remove the eye lashes from the surgical field, minimizes the risk of endophthalmitis [18]. We prospectively investigated the efficacy of eliminating the bacterial flora from the conjunctival sac and bacterial contamination from the injection needle by application of a lower concentration of PI (1.25 %) combined with levofloxacin (LFVX) eye drops. We compared the results with the current standard of care, instillation of 5 % PI.

Patients and methods

The institutional review board of Osaka University Hospital approved this prospective randomized trial. This trial was registered at the University Hospital Medical Information Network Clinical Trial site (http://www.umin.ac.jp/ctr/index/htm, Clinical Trial Number: UMIN000002260). We enrolled 100 eyes of 100 patients who were scheduled to undergo IVTs of ranibizumab (Lucentis, Novartis, Basel, Switzerland) or bevacizumab (Avastin, Genentech, South San Francisco, CA, USA) via the pars plana. Those patients were randomly divided into two groups. Exclusion criteria included corticosteroid IVTs, diabetes mellitus and any ocular surface problems, because of the possible effect on bacterial profiles of the ocular surface. Fifty-two eyes were assigned to group A and 48 eyes to group B.

Preparation procedures

All eyes were anesthetized before IVT with 4 % xylocaine eye drops instilled three times every 2–3 min. The eyes were randomly assigned to one of the two groups with different sterilizing preparatory procedures before IVT. In group A, 0.5 % LVFX eye drops (0.5 % Cravit, Santen Pharmaceutical, Osaka, Japan) were applied three times every 5 min, and hospital-formulated 1.25 % PI eye drops were then applied three times every 3 min. In group B (Control group), 5 % PI reagent (Betadyne, Alcon, Ft. Worth, TX, USA) was applied three times every 3 min. In both groups, the eye lids and lashes were painted with 10 % PI swabs immediately before IVT.

Conjunctival culture sampling

Culture samples were collected from the superotemporal conjunctiva, by use of sterile cotton swabs, 3.5 mm from the corneal limbus, the site of the IVT injection. The specimens were collected before preparation with either the LVFX eye drops or the 5 % PI, after preparation and immediately before IVT.

Immediately after injection the injection needles were collected for culture. Care was taken that needles did not come into contact with the lashes when the needles were retracted.

A technician who was unaware of the group assignment analyzed the samples. The cultures were stored at 4 °C until they were evaluated. The data were coded and stored at the Department of Ophthalmology, Osaka University Medical School, Osaka, Japan.

Bacteriological examination

The specimens were subjected to bacteriological examination at the Research Institute for Microbial Diseases of Osaka University. From each sample 50 μl were placed on two Columbia agar plates containing 5 % sheep’s blood and incubated under either aerobic (35 °C, 1–2 days) or anaerobic (35 °C, 1–5 days) conditions. Enrichment culture was also performed on clinical thioglycolate media plates incubated at 35 °C for 1–14 days.

Sample size calculation and statistical analysis

The sample size was calculated before the study in the following manner. Assuming a 15 % difference between the study groups and 25 % positive cultures in group A after preparation with a power of 80 % and an error of 10 %, 43 injections were required for the subjects in each group to prove the non-inferiority of the procedure studied. As a result, we recruited 100 eyes for the study. Statistical analysis was performed by using Fisher’s exact test from JMP version 8.0 (SAS Systems, Cary, NC, USA). All analysis was performed by a statistician unaware of the group assignment.

Results

Positive cultures

The pretreatment cultures were positive for 45 eyes (86.5 %) in group A and for 37 eyes (77.1 %) in group B (82 eyes; positive cultures 82.0 % overall). After preparation using LVFX and/or PI, the number of positive cultures decreased to 25 (48.1 %) for eyes in group A and 27 (56.3 %) for eyes in group B (52 eyes; positive cultures 52.0 % overall), with no significant difference between the groups (P = 0.43, Fisher’s exact test). The number of positive cultures decreased by 46.7 % in group A and 33.4 % in group B, which did not differ significantly (P = 0.48). None of the needle samples of either group proved positive.

Isolated species

The bacterial species isolated from all cases before preparation are shown in Table 1. In both groups Propionibacterium acnes (P. acnes) was the most common, with Staphylococcus epidermidis second. The bacterial species isolated after preparation are shown in Table 2 for group A and Table 3 for group B. The profiles of bacteria isolated were similar in both groups. The profiles did not change before or after preparation.
Table 1

Numbers and percentages of positive culture specimens from the conjunctival sac of all cases (n = 100) before the disinfection procedure

Isolate

No.

%

Propionibacterium acnes

50

35.5

Staphylococcus epidermidis

43

30.5

Corynebacterium spp.

21

14.9

S. aureus (MRSA)

8 (1)

5.7 (0.7)

S. lugdunensis

5

3.5

Bacillus spp.

4

2.8

Streptococcus oralis

2

1.4

St. anginosus

1

0.7

Bacillus

1

0.7

Enterococcus durans

1

0.7

E. faecalis

1

0.7

P. granulosum

1

0.7

St. gordonii

1

0.7

S. capitis

1

0.7

Kocuria rosea

1

0.7

Total

141

100

MRSA, methicillin-resistant Staphylococcus aureus

Table 2

Numbers and percentages of eyes in group A with bacterial species detected after disinfection with levofloxacin and 1.25 % povidone–iodine

Isolate

No.

%

Propionibacterium acnes

25

35.2

Staphylococcus epidermidis

22

31.0

Corynebacterium spp.

11

15.5

S. aureus (MRSA)

4 (1)

5.6 (1.4)

Bacillus spp.

3

4.2

S. lugdunensis

2

2.8

S. capitis

1

1.4

Streptococcus anginosus

1

1.4

St. oralis

1

1.4

Total

71

100

MRSA, methicillin-resistant Staphylococcus aureus

Table 3

Numbers and percentages of eyes in group B with bacterial species detected after disinfection with 5 % povidone–iodine

Isolate

No.

%

Propionibacterium acnes

10

27.8

Staphylococcus epidermidis

10

27.8

Corynebacterium spp.

7

19.4

S. aureus (MRSA)

2 (1)

5.6 (2.8)

Bacillus spp.

1

2.8

Candida parapsilosis

1

2.8

Moraxella osloensis

1

2.8

P. granulosum

1

2.8

Sphingomonas paucimobilis

1

2.8

S. cohnii

1

2.8

Total

36

100

MRSA, methicillin-resistant Staphylococcus aureus

Minimum inhibitory concentration

The accumulative MIC curve for the top 3 selected species from all cases after treatment is shown in Fig. 1, and MICs exceeding 90 % are listed in Table 4. P. acnes was sensitive to most antibiotics except ceftazidime and amikacin, and methicillin-sensitive S. aureus, Corynebacterium, and methicillin-sensitive S. epidermidis proved strongly resistant to all the antibiotics tested. Vancomycin and oxacillin had the lowest MIC, 2 mg/ml, against the isolates overall.
https://static-content.springer.com/image/art%3A10.1007%2Fs10384-012-0160-5/MediaObjects/10384_2012_160_Fig1_HTML.gif
Fig. 1

Accumulative minimal inhibitory concentration curves for aPropionibacterium acnes, b methicillin-sensitive Staphylococcus epidermidis, and cCorynebacterium. CP, chloramphenicol; AMK, amikacin; CAM, clarithromycin; LVFX, levofloxacin; CAZ, ceftazidime; ABPC, ampicillin; SB, sulbactam; VCM, vancomycin; MPIPC, oxacillin

Table 4

Minimum inhibitory concentrations (MIC) >90 for selected bacterial species isolated from all cases after treatment

Antibiotic

P. acnes

MSSE

Corynebacterium

MSSA

Overall (mg/ml)

CP

0.5

16

32

4

8

AMK

32

2

2

8

32

CAM

≤0.06

64

64

64

64

LVFX

0.5

8

64

32

4

CAZ

8

32

8

128

32

ABPC/SB

0.25

16

0.25

16

8

VCM

0.25

2

0.5

2

2

MPIPC

1

8

2

>128

2

CP, chloramphenicol; AMK, amikacin; CAM, clarithromycin; LVFX, levofloxacin; CAZ, ceftazidime; ABPC, ampicillin; SB, sulbactam; VCM, vancomycin; MPIPC, oxacillin; P. acnes, Propionibacterium acnes; MSSE, methicillin-sensitive Staphylococcus epidermidis; MSSA, methicillin-sensitive Staphylococcus aureus

Discussion

PI is the standard of care for minimizing the risk of infectious endophthalmitis after IVT [18]. The 5 % formulation is commercially available and commonly used for pre-IVT preparation. Although it is the recommended treatment, this agent tends to stimulate the ocular surface. Although efforts have been made to identify the optimum concentration of PI for application to the ocular surface, no consensus has been reached. In this study we compared the effects of a lower dose of PI, instilled in conjunction with LVFX, with those of 5 % PI, and found that the former method was not inferior to that of the standard of care. To the best of our knowledge, this is the first report of comparison of a reduced dose of PI, combined with LVFX, with the standard of care.

The optimum concentration of PI is unknown; therefore, it is uncertain whether it is more effective at the lower concentration of 1.25 % than at 5 %. Others investigated the survival of S. aureus after exposure to different concentrations of PI and claimed that 1:100 dilution was more effective than a 10 % stock solution [19]. Also, 0.25 % PI irrigation during cataract surgery significantly reduces detection of bacteria in the anterior chamber at the end of surgery [20]. These lines of observation suggest that a low dose of PI is effective (possibly more effective than a higher dose) at reducing bacterial contamination. Our results are in line with these reports. The mechanisms of bacteria death differ between LVFX and PI. LVFX inhibits gyrase activity, which is critical for DNA replication, whereas PI destroys the cellular membrane. Thus, the combination of these reagents may have a synergistic effect. Combination of antibiotic agents and PI is reported to have a synergetic effect [21, 22]. The fact that the efficacy of LVFX + 1.25 % was similar to that of 5 % PI was at least in part because of that synergetic effect.

Ocular surface disturbances are a possible complication of use of PI. We checked the corneal surface status of 40 eyes of subjects who could visit on the day after the injections. However, there were no significant differences between their corneal erosion scores (data not shown, P = 0.88). Three patients with 5 % PI reported ocular surface stimulation during the night after the injection; none within the diluted LVFX + 1.25 % PI group reported similar sensations. Therefore, the diluted PI seemed to cause less ocular surface damage, another possible advantage.

P. acnes, S. epidermidis, and Corynebacterium were the three bacteria isolated most often from both groups before preparation, which agrees with previous studies [23]. We also identified the isolates after preparation. Interestingly, the profiles of the bacterial species were similar in both groups. This finding suggests that the lower concentration of PI combined with LVFX did not affect the bacterial profile. Although P. acnes is very sensitive to antibiotics, the lower concentration was not as effective against Staphylococcus species and Corynebacterium. Vancomycin was relatively more effective overall on the species isolated, which suggests that, as many studies recommend, vancomycin should be the first choice in treating postoperative endophthalmitis [23].

Reports describe needle contamination as a result of IVT, with approximately 0.36–2 % of needles being culture positive after IVT injections [24, 25]. Both studies used PI as a preparatory method. In our study, none of the needles was contaminated during either procedure. Hence, it is suggested that the incidence of needle contamination is very low with either method.

A strength of this study is that it was prospective and randomized, which enabled evaluation of the effects of reducing conjunctival flora. In addition, we excluded eyes affected by diabetes and steroid injections, which might have affected the local immune system. A limitation was that we were unable to show a relationship between the reduced number of positive cultures and the incidence of endophthalmitis. Because the incidence of endophthalmitis is usually low, at least 4000 patients would be needed to prove a significant reduction in the incidence of endophthalmitis [12, 16]. Finally, the study was conducted at a single-center, which might have introduced some bias. In summary, a lower dose of PI combined with wide-spectrum antibiotics is as effective as 5 % PI.

Copyright information

© Japanese Ophthalmological Society 2012