Identification of Causative Pathogens in Eyes with Bacterial Conjunctivitis by Bacterial Cell Count and Microbiota Analysis

Identification of Causative Pathogens in Eyes with Bacterial Conjunctivitis by Bacterial Cell Count and Microbiota Analysis Rumi Aoki, MD,1,2 Kazumasa Fukuda, PhD,1 Midori Ogawa, PhD,1 Takako Ikeno, PhD,1 Hiroyuki Kondo, MD,2 Akihiko Tawara, MD,2 Hatsumi Taniguchi, PhD1 Purpose: To determine the causative pathogens in eyes with bacterial conjunctivitis. Design: Evaluation of diagnostic test or technology. Participants: Thirteen eyes diagnosed clinically with bacterial conjunctivitis and 12 eyes with normal conjunctival sac were studied. Methods: The bacterial cell numbers were counted in the samples stained by ethidium bromide (EtBr). The microbiota was determined by the clone library method using polymerase chain reaction (PCR) amplification of the 16S ribosomal RNA (rRNA) gene with universal primers. In addition, examinations of smears and cultures of samples were performed. Main Outcome Measures: Bacterial cell numbers determined by the EtBr staining method and microbiota analysis based on 16S rRNA gene of samples from eyes with bacterial conjunctivitis. Results: The bacterial cell numbers in the eyes with bacterial conjunctivitis were significantly higher than those in the normal conjunctival sacs. Ten of 13 samples from the eyes with bacterial conjunctivitis had positive PCR results. The remaining 3 samples and all 12 samples from the normal conjunctiva had negative PCR results. In 5 of the 10 PCR-positive samples, the predominant species accounted for 84.5% or more of each clone library. In the remaining 5 samples, the predominant and the second dominant species accounted for 27.4% to 56.3% and 19.0% to 26.8%, respectively, of each clone library. The number of detected species in the clone libraries was between 8 and 20 (average ⫾ standard deviation, 7.5⫾5.8). Bacteria were detected in 8 of the 10 bacterial conjunctivitis samples and in 5 of the 12 normal samples in the cultures. The number of species detected by cultures was 1 in the eyes with bacterial conjunctivitis and between 1 and 3 (mean ⫾ standard deviation, 1.6⫾0.9) in the normal conjunctiva. Conclusions: These results showed that the bacterial cell number in a sample is a good method of determining bacterial conjunctivitis. The microbiota analysis detected a diverse group of microbiota in the eyes with bacterial conjunctivitis and showed that the causative pathogens could be determined based on percentages of bacterial species in the clone libraries. The combination of bacterial cell count and microbiota analysis is a good method for identifying the causative pathogens in cases of monomicrobial and polymicrobial conjunctivitis. Financial Disclosure(s): The author(s) have no proprietary or commercial interest in any materials discussed in this article. Ophthalmology 2013;120:668 – 676 © 2013 by the American Academy of Ophthalmology.

Bacterial conjunctivitis is a common infectious disease, and most cases have a benign course. The causative pathogens of bacterial conjunctivitis have been determined conventionally by both the smear method and the culture method. However, the results of these methods are not always conclusive. For example, the isolation rate by direct smear examination in eyes with bacterial conjunctivitis has been reported to be between 51% and 58.8%.1,2 The rate for the culture method has been reported to be between 47.5% and 97.8% in eyes with bacterial conjunctivitis1,3– 6 and 36.7% to 90.6%6 –9 in normal conjunctival sacs. In addition, the bacterial species detected in eyes with bacterial conjunctivitis also have been found in normal conjunctival sacs.6 –10 Thus, it is unclear whether the bacteria detected by cultures are the causative pathogens. In addition, cultures require 4 to 30 days, and the findings do not necessarily reflect the entire set of microbiota because the growth conditions are species dependent.

668

© 2013 by the American Academy of Ophthalmology Published by Elsevier Inc.

Because of these limitation, there has been an increase in the number of studies using molecular methods, for example, polymerase chain reaction (PCR) with species-specific primers,11–13 amplification of the 16S ribosomal RNA (rRNA) gene by PCR using universal primer sets followed by direct sequencing,14 –17 denaturing gradient gel electrophoresis,18 and pyrosequencing.19 However, indigenous bacteria, such as Corynebacterium macginleyi, are difficult to distinguish as the causative or not causative pathogens using PCR. Antibacterial eye drops frequently are overprescribed because of a misidentification of the causative pathogen, and the number of antibiotic-resistant bacteria recently has increased in ophthalmic samples, including those from eyes with conjunctivitis.3–5,10,20 Thus, the purpose of this study was to determine a more accurate method of identifying the causative pathogen in eyes with bacterial conjunctivitis. To accomplish this, the authors performed bacterial cell counts with the ethidium ISSN 0161-6420/13/$–see front matter http://dx.doi.org/10.1016/j.ophtha.2012.10.001

Aoki et al 䡠 Identification of Causative Pathogens in Bacterial Conjunctivitis bromide (EtBr) staining method and analysis of the microbiota by the clone library method using PCR amplification of the 16S rRNA gene with universal primers. The microbiota is composed of all of the microorganisms that live in a bodily organ.

was concentrated and replaced by 30 ␮l TE buffer using Montage PCR centrifugal filter devices (Millipore). The number of remaining bacterial cells after the DNA extraction was determined using the EtBr staining method.

Polymerase Chain Reaction

Patients and Methods Samples Conjunctival swabs were obtained from 25 eyes of 25 patients from April 2009 through August 2011 at University Hospital, Kyushurousai Hospital Moji Medical Center, and Inatsuki Hospital, Fukuoka, Japan. The protocol of this study was approved by the human and animal ethics review committees of the 3 hospitals. The nature of the study and procedures to be used were explained fully to all participants, and informed consent was obtained before beginning the examinations and collection of the samples. The conjunctival swabs of 13 eyes of 13 patients with clinically diagnosed bacterial conjunctivitis, that is, conjunctivitis without follicle and papilla, and those from 12 eyes with normal conjunctival sacs as control were examined. Only 1 sample from each case was collected. Conjunctival swabs were obtained with sterile cotton swabs and were kept at 4°C in sterile tubes until they were analyzed. For the analyses, the cotton swabs were placed in a sterile tube with 2 ml ultrapure water and were vibrated for 2 minutes with a vortex mixer. The solutions were used as sample solutions for all of the examinations. Milli-Q Integral3 ultrapure water (Millipore Corporation, Billerica, MA) was used in all of the experiments.

Bacterial Cell Counting with Epifluorescent Staining Method One hundred microliters of the sample solution were added to 900 ␮l EtBr solution, which consisted of 100 ␮g EtBr/ml in 0.1 M phosphate buffer, pH 8.5, 5% NaCl, and 0.5 mM ethylenediaminetetraacetic acid-2Na. The mixture was allowed to stand for 10 minutes at room temperature. Then the mixture (1.0 ml) was filtered through a 0.2-␮m pore filter (Millipore, Bedford, MA). Objects shaped like bacteria on the filter were counted at 60 randomly selected fields of view (60/slide) with an Olympus BX40 microscope (Olympus Optical, Tokyo, Japan). The number of bacteria per milliliter of sample solution then was calculated.21–24

The 16S rRNA gene was amplified with a GeneAmp PCR system 9700 thermocycler (Applied Biosystems, Foster City, CA). The reaction mixtures contained the universal primers set (E341F, 5=-CCTACGGGAGGCAGCAG-3=; and E907R, 5=-CCGTCAATTCMTTTRAGTTT-3=)25 and AmpliTaq Gold DNA polymerase LD (Applied Biosystems) were incubated in a thermocycler at 96°C for 5 minutes, followed by 30 cycles at 96°C for 30 seconds, at 53°C for 30 seconds, at 72°C for 1 minute, and then 1 cycle for the final elongation step at 72°C for 7 minutes. The primer set used was a 100% match to approximately 92% of the bacterial 16S rRNA gene registered in the Ribosomal Database Project II database.21,22 The PCR products (approximately 580 base pairs) were confirmed with 2% agarose gel electrophoresis. Throughout the DNA extraction process, ultrapure water, instead of a sample solution, was used to exclude the possibility of false-positive PCR results as a negative control.

Clone Library Construction and Determination of Nucleotide Sequence The PCR products were cloned with a TOPO TA Cloning kit (Invitrogen, Carlsbad, CA) according to the manufacturer’s instructions. A total of 96 white colonies from each clone library were selected randomly for sequencing analysis.21 The partial fragment of the cloning vector (pCRII; Invitrogen) containing an inserted PCR product was amplified with AmpliTaq Gold DNA polymerase and a primer set (M13 forward, 5=-GTAAAACGACGGCCAG-3=; M13 reverse, 5=-CAGGAAACAGCTATGAC-3=). The primers and deoxyribonucleotide triphosphate were removed from the PCR mixture with ExoSAP-IT (GE Healthcare UK Ltd., Buckinghamshire, England) according to the manufacturer’s instructions. Then, 1 ␮l was used as a template for the sequencing reaction. The sequencing reactions were performed with primers M13 forward and the BigDye Terminator Cycle Sequencing Kit version 3.1 (Applied Biosystems). The nucleic acid sequences were determined with a 3130xl Genetic Analyzer (Applied Biosystems).

Homology Search DNA Extraction DNA was extracted from 900 ␮l of the sample solutions, to which 100 ␮l of 30% sodium dodecyl sulfate solution (final concentration, 3.0%) was added. The mixture was shaken vigorously, and then approximately 0.3 g of a mixture of glass beads that consisted of equal weights of 0.1-mm and 1-mm diameter beads in a 2.5-ml polypropylene tube (tube A) was added. The mixture then was shaken vigorously at 4500 rpm for 5 minutes on a Micro Smash MS-100 apparatus (Tomy Seiko Co., Ltd., Tokyo, Japan). The supernatant was collected after centrifugation at 20 000⫻g for 5 minutes at room temperature. After the supernatant was transferred to a fresh tube (tube B), 1 ml of 3% sodium dodecyl sulfate solution was added to the pellet in tube A. The mixture was shaken vigorously and centrifuged. The supernatant then was transferred to tube B. This DNA extraction process was repeated 3 times, and DNA was collected in tube B. The supernatant (approximately 3 ml) was treated with an equal volume of phenol-chloroformisoamyl alcohol (25:24:1, vol/vol). The DNA in the aqueous phase

Highly accurate sequences selected by Phred quality value were trimmed from the primer and vector regions. Only the sequences having good quality were used for analyses. The remaining sequences were compared with an in-house database containing the 16S rRNA gene sequences of 5870 type strains using the basic local alignment search tool algorithm. The 16S rRNA gene sequences of type strains were obtained from DNA Data Bank of Japan (http://www.ddbj.nig.ac.jp/) and the Ribosomal Database Project (http://rdp.cme.msu.edu/). The clones with 97% or higher sequence similarity to the reference type strain were presumed to be members of the same species. The sequences with less than 97% similarities were defined as unclassified bacteria.

Statistical Analyses To quantify the diversity of each microbiota, the Simpson index of diversity was used. Unclassified bacteria clones in each clone library were not used for the analyses.

669

Ophthalmology

Volume 120, Number 4, April 2013 Table 1. Characteristics Clone Library Analysis of 16S Ribosomal RNA Gene

Sample Number

Age/Sex

Bacterial conjunctivitis

1

84/F

2.2 ⫻ 106

2

84/M

4.8 ⫻ 105

3

82/M

4.1 ⫻ 105

4

89/F

2.0 ⫻ 105

5

76/F

1.5 ⫻ 105

6

85/F

1.1 ⫻ 105

7

95/F

5.6 ⫻ 104

8

80/F

1.5 ⫻ 104

9

90/M

1.2 ⫻ 104

10

75/F

1.2 ⫻ 104

11 12 13 14 15 16 17 18 19 20 21

83/F 74/F 76/F 88/F 79/F 75/M 70/M 63/F 66/F 71/M 63/F

8.9 ⫻ 103 ⬍3.0 ⫻ 103 ⬍3.0 ⫻ 103 3.0 ⫻ 103 3.0 ⫻ 103 3.0 ⫻ 103 3.0 ⫻ 103 3.0 ⫻ 103 3.0 ⫻ 103 3.0 ⫻ 103 3.0 ⫻ 103

22 23

66/F 77/F

⬍3.0 ⫻ 103 ⬍3.0 ⫻ 103

(⫺) (⫺)

(⫺) (⫺)

24 25

81/F 76/F

⬍3.0 ⫻ 103 ⬍3.0 ⫻ 103

(⫺) (⫺)

(⫺) (⫺)

Normal conjunctival sac (control)

Cell Number/ ml (EtBr)*

Predominant (clones/clones, %†)

Diagnosis

Prevotella oris (45/80, 56.3%) Staphylococcus aureus (26/67, 38.8%) Corynebacterium macginleyi (82/82, 100%) Corynebacterium macginleyi (83/84, 98.8%) Corynebacterium macginleyi (49/58, 84.5%) Haemophilus influenzae (70/73, 95.9%)

Corynebacterium macginleyi (63/68, 92.6%) Capnocytophaga ochracea (16/56, 28.6%) Propionibacterium acnes (18/58, 31.0%)

Propionibacterium acnes (17/62, 27.4%) (⫺)¶ (⫺) (⫺) (⫺) (⫺) (⫺) (⫺) (⫺) (⫺) (⫺) (⫺)

2nd Dominant (clones/clones, %) Fusobacterium canifelinum (17/80, 21.3%) Propionibacterium acnes (15/67, 22.4%) (⫺)§ (⫺) Propionibacterium acnes (3/58, 5.2%) Streptococcus intermedius (1/73, 1.4%) Curvibacter delicatus (1/73, 1.4%) Acinetobacter johnsonii (1/73, 1.4%) Proteus morganii (2/68, 2.9%) Micromonas micros (15/56, 26.8%) Corynebacterium macginleyi (11/58, 19.0%) Curvibacter delicatus (11/58, 19.0%) Streptococcus sinensis (14/62, 22.6%) (⫺) (⫺) (⫺) (⫺) (⫺) (⫺) (⫺) (⫺) (⫺) (⫺) (⫺)

CFU ⫽ colony-forming units; CL ⫽ colistin; CMX ⫽ cefmenoxime; CP ⫽ chloramphenicol; F ⫽ female; GFLX ⫽ gatifloxacin; GPC ⫽ Gram positive cocci; GPR ⫽ Gram positive *Total number of cells was counted by ethidium bromide (EtBr) staining method. † Positive clones/tested clones. ‡ Direct smear examination. § Not detected. 㛳 Not used. ¶ Not calculated.

Direct Smear Examinations, Cultures, and Identification of Cultured Bacteria After making the sample solutions, the conjunctival swabs were brushed onto glass slides and underwent Gram staining. In addition, 100 ␮l of the sample solutions were plated onto nutrient agar medium in room air, on sheep blood agar, and on chocolate agar in

670

5% CO2 for 48 to 120 hours at 37°C. The number of colonies was counted, and the colony-forming units were calculated. Bacterial strains obtained by cultures were identified by PCR amplification and direct sequencing of the 16S rRNA gene. DNA was extracted from the bacterial strains as described in the DNA Extraction section. The PCR product was amplified with AmpliTaq Gold DNA polymerase, low DNA (LD), and a

Aoki et al 䡠 Identification of Causative Pathogens in Bacterial Conjunctivitis of Samples Clone Library Analysis of 16S Ribosomal RNA Gene

Cultivation of Bacterial Species (CFU/ml)

Number of Detected Species

Simpson’s Index of Diversity

Antimicrobial Treatments

9

0.612

NT

NT

10

0.671

Staphylococcus aureus (1.2 ⫻ 104) Corynebacterium macginleyi (2.4 ⫻ 10 )

GFLX (⫺)㛳

Gram Stain‡

LVFX

1

0

GPC GPR GPR

1

0

GPR

Corynebacterium macginleyi (1.2 ⫻ 104)

7

0.281

NT

NT

GFLX

4

0.080

NT

NT

(⫺)

3

0.088

GPR

Corynebacterium macginleyi (30)

(⫺)

12

0.796

(⫺)§

CP, CL 4

Enterococcus faecalis (1.5 ⫻ 103)

(⫺)

Corynebacterium macginleyi (1.3 ⫻ 10 )

(⫺)

(⫺)

(⫺)§ (⬍10)

CMX

Corynebacterium macginleyi (4.3 ⫻ 102) Staphylococcus aureus (60) (⫺) (⬍10) Corynebacterium macginleyi (30) (⫺) (⬍10) (⫺) (⬍10) Corynebacterium macginleyi (50) (⫺) (⬍10) (⫺) (⬍10) (⫺) (⬍10) Corynebacterium macginleyi (30) Staphylococcus epidermidis (40) Staphylococcus hominis (80) Microbacterium testaceum (2.9 ⫻ 102) Propionibacterium acnes (20) Corynebacterium macginleyi (30) (⫺) (⬍10) (⫺) (⬍10)

(⫺) (⫺) (⫺) CMX (⫺) (⫺) (⫺) (⫺) (⫺) (⫺) (⫺)

8

0.766

GPR

20

0.822

0 0 0 0 0 0 0 0 0 0 0

(⫺)¶ (⫺) (⫺) (⫺) (⫺) (⫺) (⫺) (⫺) (⫺) (⫺) (⫺)

(⫺) (⫺) (⫺) (⫺) (⫺) (⫺) (⫺) (⫺) (⫺) (⫺) (⫺)

0 0

(⫺) (⫺)

GPR (⫺)

0 0

(⫺) (⫺)

(⫺) (⫺)

3

(⫺) (⫺) (⫺) (⫺)

rods; LVFX ⫽ levofloxacin; M ⫽ male; NT ⫽ not tested.

primer set (E27F primer, 5=-AGAGTTTGATCMTGGCTCAG-3=; E1492R primer, 5=-TACGGYTACCTTGTTACGACTT-3=).25 The reaction mixtures were incubated in a thermocycler at 96°C for 5 minutes, followed by 30 cycles at 96°C for 40 seconds, 53°C for 1 minute, 72°C for 2 minutes, and then 1 cycle for the final elongation step at 72°C for 7 minutes. After ExoSAP-IT treatment, 1 ␮l was used as a template for the sequencing reaction. The

sequencing reactions were performed with primer sets (E27F, E1492R, E907R, E530F; 5=-GTGCCAGCMGCCGCGG-3=)25 and BigDye Terminator Cycle Sequencing Kit version 3.1. A homology search of the assembled sequences (approximately 1300 base pairs) was performed using the National Center for Biotechnology Information basic local alignment search tool (http://blast.ncbi.nlm. nih.gov/Blast.cgi).

671

Ophthalmology

Total bacterial cell number (cells/ml)

1.0㬍107

Volume 120, Number 4, April 2013 Bacterial Cell Count Using the Epifluorescent Staining Method

* P=0.0003

The results of the bacterial cell count are shown in Table 1. Bacterial cells were detected in 11 of the 13 swabs from the eyes with bacterial conjunctivitis. In the remaining 2 swabs, the number of bacterial cells was less than the limit of this technique (3.0⫻103 cells/ml). The number of bacterial cells in each swab of the 13 patients ranged from less than 3.0⫻103 to 2.2⫻106, with an mean ⫾ standard deviation of 2.8⫾6.0⫻105 cells/ml. Swabs with less than the detection limit (3.0⫻103 cells/ml) were considered to be 3.0⫻103 cells/ml for the statistical analyses. The number of bacteria in the swab from the 12 controls was 3.0⫻103 cells/ml or less. The difference in the number of bacterial cells in the 2 groups was significant (P ⫽ 0.0003, Mann–Whitney U test; Fig 1). All of the remaining bacterial cell numbers after the DNA extraction were less than the limit of this technique. The DNA extraction had high-efficiency cell lysis.

1.0㬍106

1.0㬍105

1.0㬍104

1.0㬍103

Bacterial conjunctivitis n=13

Control n=12

Figure 1. Graph showing the total bacterial cell numbers as determined by ethidium bromide staining method, compared between the bacterial conjunctivitis and control groups. A less-than-detected limitation was regarded as 3.0⫻103 cells/ml. The P value was obtained using Mann– Whitney U test.

Detection of Bacterial 16S rRNA Gene by Polymerase Chain Reaction The solutions of 10 of 13 swabs from the eyes with bacterial conjunctivitis demonstrated positive PCR results for the 16S rRNA gene, and all 12 swabs from the 12 normal patients demonstrated negative PCR results for the 16S rRNA gene (Table 1). The average bacterial cell number of the 10 swabs with PCR amplification of the 16S rRNA gene was 3.6⫾6.7⫻105 cells/ml, with a range of 1.2⫻104 to 2.2⫻106 cells/ml. The average bacterial cell number in the 15 swabs with negative PCR results for the 16S rRNA gene, including the 3 samples from the eyes with bacterial conjunctivitis, was 3.4⫾1.5⫻103 cells/ml, with a range of less than 3.0⫻103 to 8.9⫻103 cells/ml.

Results Patient Demographics The average age of the 13 patients with bacterial conjunctivitis was 82.5 years, with a range of 74 to 95 years. Twelve patients with a normal-appearing conjunctival sac acted as controls, and their average age was 72.9 years, with a range of 63 to 88 years. Antibacterial eye drops (cefmenoxime, levofloxacin, gatifloxacin, chloramphenicol, colistin) were being used by 6 of the 13 patients (Table 1).

10

27.4

22.6

4.8

32.3

Prevotella oris

12.9

Haemophilus influenzae 9

19.0

10.3

31.0

19.0

8.6

12.1

Corynebacterium macginleyi Corynebacterium tuberculostearicum

8

5.4

5.4

26.8

28.6

Sample (n)

7

14.3

14.3

92.6

6

5.4 4.4 2.9

95.9

4.1

Staphylococcus aureus Staphylococcus epidermidis Fusobacterium canifelinum Propionibacterium acnes Micromonas micros

5

84.5

5.2

Capnocytophaga ochracea Streptococcus sinensis

4

98.8

3

100

2

10.3

4.5

38.8

4.5

1.2

Streptococcus parasanguinis Enterococcus faecalis Curvibacter delicatus

22.4

9.0

20.9

Porphyromonas endodontalis Others*

1

56.3 0%

10%

20%

30%

21.3 40%

50%

60%

70%

12.5 80%

8.8 1.3 90%

unclassified†

100%

Percentage of Each Species Figure 2. Graph showing the percentage of bacterial species in each clone library of 10 polymerase chain reaction (PCR)-positive bacterial conjunctival swabs (samples 1 through 10). The stacked bars represent the bacterial composition of the clone libraries of 10 PCR-positive bacterial conjunctival swabs. The x- and y-axes represent percentage of each species and the sample number, respectively. * ⫽ the species with 2 clones or less in each of the clone libraries appear as “others”; † ⫽ the sequences with less than 97% similarity to the reference type strain were defined as “unclassified” bacteria.

672

Aoki et al 䡠 Identification of Causative Pathogens in Bacterial Conjunctivitis Bacterial Composition of Clone Libraries of 10 Bacterial Conjunctival Samples with Positive Polymerase Chain Reaction Results The predominant and second dominant bacterial species and the number of detected species in the clone libraries of the 10 bacterial conjunctival swabs with PCR amplification of the 16S rRNA gene are shown in Table 1. The detected bacterial species and the respective percentages in each of the clone libraries are shown in Figure 2. The species with 2 clones or less in each of the clone libraries are designated as “others” in Figure 2. The number of detected species in the clone libraries ranged from 8 to 20, with a mean of 7.5⫾5.8 species (Simpson index of diversity, 0 – 0.822). In 5 of the 10 bacterial conjunctival samples (samples 3 through 7), the percentage of the predominant bacterial species was 84.5% or more in each sample. The percentage for the second dominant bacterial species was 5.2% or less in each sample. The predominant species were C. macginleyi in 4 samples and Haemophilus influenzae in 1 sample. Of the remaining 5 samples from the eyes with bacterial conjunctivitis (samples 1, 2, 8, 9, and 10), the percentage of the predominant bacterial species ranged from 27.4% to 56.3% in each sample. The second dominant bacterial species ranged from 19.0% to 26.8%. The predominant species were Prevotella oris, Staphylococcus aureus, Capnocytophaga ochracea, and Propionibacterium acnes. The sequences that had less than 97% similarity to the reference type strain were defined as unclassified bacteria. All unclassified bacteria were classifiable if the clones with 89% or more sequence similarity to the reference type strain were presumed to be members of the same species; most of the unclassified bacteria belong to the detected bacterial species in each microbiota. In the remaining unclassified bacteria that did not belong to the detected species in each microbiota, Porphyromonas catoniae (10 clones) was the third dominant species in the microbiota of sample 2. Even after considering unclassified bacteria, the microbiota were only slightly different.

Smear Examination by Gram Staining Ten of the 13 samples from the eyes with bacterial conjunctivitis and all 12 normal samples were tested by the smear examination (Table 1). Five of the 10 bacterial conjunctival samples and 1 of the 12 normal samples demonstrated positive staining results for bacteria. Four of 5 bacterial conjunctival samples showed only Gram-positive rods, and the remaining bacterial conjunctival swab showed both Gram-positive rods and Gram-positive cocci. In the normal samples, only 1 sample showed Gram-positive rods.

Cultures Ten of the 13 bacterial conjunctival samples and all 12 normal samples were tested by cultures (Table 1). Eight of the 10 bacterial conjunctival samples and 5 of the 12 normal samples demonstrated positive culture results. The number of detected species in the bacterial conjunctival samples was 1, ranging between 1 and 3 species (mean, 1.6⫾0.9 species) in normal samples. The bacterial species of the 8 bacterial conjunctival samples were C macginleyi, S aureus, and Enterococcus faecalis, whereas those of the normal samples were C macginleyi, Microbacterium testaceum, P acnes, Staphylococcus epidermidis, and Staphylococcus hominis. The number of each cultivated bacteria in the samples from the eyes with bacterial conjunctivitis was 3.0⫻10 to 2.4⫻104 colony forming units/ml, and it was 2.0⫻10 to 2.9⫻102 colony-forming units/ml in the normal samples.

Discussion This study identified the causative pathogen in eyes with bacterial conjunctivitis using a combination of bacterial cell count and microbiota analysis. Ethidium bromide, a phenanthridinium dye that intercalates between the base pairs of the double helix of DNA, is a fluorescent dye that has been used extensively to stain both prokaryotic and eukaryotic cells.24,26,27 Objects shaped like bacteria generally can be distinguished from eukaryotic cells by 3 features: their shape and arrangement, size, and fluorescence. Bacterial cells have specific shapes and arrangements, for example, diplococci, staphylococci, streptococci, tetrad, and rods. Most bacteria are 0.2 to 0.8 ␮m in diameter or width and 0.5 to 8 ␮m in length. Bacterial cells have sharp outlines and intense fluorescence. The bacterial cell number determined by the EtBr staining method was significantly higher in the samples collected from eyes with bacterial conjunctivitis than in those from normal conjunctival sacs. In addition, the bacterial numbers determined by cultures in eyes with bacterial conjunctivitis also were higher than in those from normal conjunctival sacs. However, cultures detected a smaller number of bacteria than the EtBr staining method. Therefore, culture has the risk of missing noncultivated bacteria, probably because the growth conditions are species dependent, and there are viable but nonculturable bacteria. Moreover, bacterial cell counting with the EtBr staining method requires only approximately 1 hour, whereas the culture process requires 2 to 5 days. We suggest that the bacterial cell count by the EtBr staining method is a rapid and simple method of determining bacterial conjunctivitis. In addition, there was a significant correlation between the bacterial cell numbers determined by the EtBr staining method and those obtained using PCR amplification of the 16S rRNA gene. Considering that all of the normal samples demonstrated negative PCR results, we suggest that PCR should be considered for determining the pathogen in eyes with bacterial conjunctivitis. However, samples 11, 12, and 13 from eyes with purulent conjunctivitis unexpectedly demonstrated negative PCR results for the 16S rRNA gene. There was no difference between the bacterial cell numbers of the normal samples and those of samples 12 and 13. The bacterial cell number of sample 11 was higher than that of normal samples, but the difference was not significant. This patient recovered from conjunctivitis without antibacterial eye drops. It is difficult to determine completely whether bacterial conjunctivitis is caused just by the bacterial cell numbers and PCR results. However, based on these results, it was most likely that these eyes did not have bacterial conjunctivitis. The conjunctivitis may have been caused by virus, allergy, or fungi. The limited amount of sample solution prevented more detailed examinations. Thus, 1.2⫻104 bacterial cells/ml was set as the critical value for diagnosing bacterial conjunctivitis. It is difficult to determine bacterial conjunctivitis with only clinical signs. In addition, viral and allergic conjunctivitis make up more than one half of the cases of acute conjunctivitis.28 The bacterial cell number counted using the EtBr staining method has given a criterion for diagnosing bacterial conjunctivitis.

673

Ophthalmology

Volume 120, Number 4, April 2013

Bacterial conjunctivitis occasionally has been reported to be polymicrobial by cultures and denaturing gradient gel electrophoresis.1,3,4,6,18 Several studies of the detection of bacteria by the clone library method using the 16S rRNA gene have reported on the microbiota of environmental and clinical samples.21–23,29 –35 The clone library method can detect the ratios of detected bacteria in clone libraries, and this method was applied to the microbiota analysis of bacterial conjunctivitis. The clone library method detected a greater number of bacterial species than cultures. An unexpected diverse microbiota was found in eyes with bacterial conjunctivitis using the clone library method. Analyses of 5 of the 10 bacterial conjunctival samples (samples 3 through 7) showed that the percentage of the predominant species was more than 84.5% in the microbiota, and the remaining 5 samples (samples 1, 2, 8, 9, and 10) showed that the percentages of the predominant and the second dominant bacterial species were 27.4% to 56.3% and 19.0% to 26.8% in each sample, respectively. There are no standard criteria for evaluating the pathogenicities of bacterial species detected by the clone library method. However, all normal samples demonstrated negative PCR results; therefore, it is most likely that the detected bacterial species played significant pathogenic roles. Based on the percentages of bacterial species in clone libraries, it is most likely that samples 3 through 7 were bacterial conjunctivitis caused by the predominant species and that samples 1, 2, 8, 9, and 10 were bacterial conjunctivitis caused by a mixture of bacteria. Our understanding of the percentages of bacterial species in clone libraries has contributed to the quantitative identification of the causative pathogens in bacterial conjunctivitis. It is interesting that, anaerobes36 – 40 were detected in the clone libraries of all bacterial conjunctivitis caused by mixture of bacteria. Anaerobes play a significant role in bacterial conjunctivitis.6,41 The diversity of microbiota in eyes with bacterial conjunctivitis determined by the clone library method also indicates the possibility of significant polymicrobial infections, including anaerobes. Further large-scale studies will be required to clarify the significance of polymicrobial infections, including anaerobes in bacterial conjunctivitis. The combined EtBr staining and clone library methods can be useful in determining bacterial conjunctivitis caused by indigenous bacteria that the culture method cannot determine quantitatively. For example, based on the results of high bacterial cell numbers and dominant percentage in the clone libraries, samples 3, 4, 5, and 7 were determined to be bacterial conjunctivitis caused by C macginleyi, which is an indigenous bacterium. C macginleyi recently was reported to be a causative pathogen of bacterial conjunctivitis, endophthalmitis, and keratitis.10,42– 46 Kawanami et al47 quantitatively identified indigenous bacteria as the causative pathogen of severe pneumonia by the clone library method and were successful in the treatment based on the information. The clone library method can contribute to better treatment of eye infections. If the culture conditions can satisfy the bacterial requirements, the culture is a sensitive method for identifying the bacteria in eyes with bacterial conjunctivitis and a normal conjunctival sac.6 Our culture-based isolation rates of eyes with bacterial conjunctivitis and a normal conjunctival sac were 80.0% and 41.7%, respectively. This indicates that the

674

bacteria detected by the cultures were not necessarily the causative pathogens. However, the cultures cannot identify the entire set of microbiota because the growth conditions are species dependent. The culture method failed to detect P acnes, which is an anaerobe, in sample 2. There was a lack of concordance between the predominant species and the culture results in samples 8 and 9. E faecalis and C macginleyi, which were the third and the second dominant species, respectively, in the microbiota of these samples, may have been detected by the cultures, because the predominant and the second dominant species (C ochracea, Micromonas micros, and P acnes) in each sample were difficult to cultivate under our culture conditions.36,37,48 Polymerase chain reaction and sequencing-based detection is suitable for detecting bacteria in ocular samples containing few bacteria.15,17,18 However, the extreme sensitivity of PCR enhances the possibility of false-positive results. Schabereiter-Gurtner et al18 restricted their amplification cycles to 30 to exclude improper amplifications of the 16S rRNA gene, and the normal conjunctival microbiota could not be detected under these conditions. The authors also observed that increasing the PCR cycle to 35 cycles led to amplification of contaminating DNA in some of the reagent-only controls. Thus, the amplification cycles were restricted to 30, and all reagent-only controls showed negative PCR results under the laboratory conditions. The culture method still plays an important role in the diagnosis and treatment of bacterial conjunctivitis because antibiotic sensitivity tests of bacteria need cultures and cannot be determined using molecular methods. However, the culture method alone is insufficient for more accurate diagnosis of bacterial conjunctivitis. These results showed that an accurate diagnosis of bacterial infectious disease, including differentiating mixed bacterial infection from other types of conjunctivitis such as viral, allergic, and fungal conjunctivitis, is important for selecting the appropriate treatment of the infectious conjunctivitis. Thus, the bacterial cell count and the microbiota analysis can be used as supplemental methods to cultures. This is especially true for infections caused by bacteria with unusual culture requirements, indigenous bacteria, and mixed infections and for patients who have been treated unsuccessfully with antibiotics.

References 1. Mukai N, Katsumura K, Shimizu K, et al. Correlation of conjunctival smears with cultures in bacterial conjunctivitis [in Japanese]. Atarashii Ganka 2004;21:375–7. 2. Katsumura K, Mukai N, Shimizu K, et al. Correlation of conjunctival smears with cultures in bacterial conjunctivitis— comparative study of direct isolation and enrichment culture [in Japanese]. Atarashii Ganka 2005;22:367–9. 3. Mahajan VM. Acute bacterial infections of the eye: their aetiology and treatment. Br J Ophthalmol 1983;67:191– 4. 4. Matsumoto H, Inoue Y, Ohashi Y, Usui M. Multicenter study of isolates from bacterial conjunctivitis in Japan [in Japanese]. Atarashii Ganka 2007;24:647–54. 5. Cavuoto K, Zutshi D, Karp CL, et al. Update on bacterial conjunctivitis in South Florida. Ophthalmology 2008;115:51– 6.

Aoki et al 䡠 Identification of Causative Pathogens in Bacterial Conjunctivitis 6. Perkins RE, Kundsin RB, Pratt MV, et al. Bacteriology of normal and infected conjunctiva. J Clin Microbiol 1975;1:147–9. 7. Singer TR, Isenberg SJ, Apt L. Conjunctival anaerobic and aerobic bacterial flora in paediatric versus adult subjects. Br J Ophthalmol 1988;72:448 –51. 8. Hara J, Yasuda F, Higashitsutsumi M. Preoperative disinfection of the conjunctival sac in cataract surgery. Ophthalmologica 1997;211(suppl):62–7. 9. McNatt J, Allen SD, Wilson LA, Dowell VR Jr. Anaerobic flora of the normal human conjunctival sac. Arch Ophthalmol 1978;96:1448 –50. 10. Eguchi H, Kuwahara T, Miyamoto T, et al. High-level fluoroquinolone resistance in ophthalmic clinical isolates belonging to the species Corynebacterium macginleyi. J Clin Microbiol 2008;46:527–32. 11. Burton MJ, Adegbola RA, Kinteh F, et al. Bacterial infection and trachoma in the Gambia: a case control study. Invest Ophthalmol Vis Sci 2007;48:4440 – 4. 12. Elnifro EM, Cooper RJ, Klapper PE, et al. Multiplex polymerase chain reaction for diagnosis of viral and chlamydial keratoconjunctivitis. Invest Ophthalmol Vis Sci 2000;41:1818–22. 13. Lietman TM, Dhital SP, Dean D. Conjunctival impression cytology for vitamin A deficiency in the presence of infectious trachoma. Br J Ophthalmol 1998;82:1139 – 42. 14. Rudolph T, Welinder-Olsson C, Lind-Brandberg L, Stenevi U. 16S rDNA PCR analysis of infectious keratitis: a case series. Acta Ophthalmol Scand 2004;82:463–7. 15. Chiquet C, Lina G, Benito Y, et al. Polymerase chain reaction identification in aqueous humor of patients with postoperative endophthalmitis. J Cataract Refract Surg 2007;33:635– 41. 16. Chiquet C, Cornut PL, Benito Y, et al, French Institutional Endophthalmitis Study (FRIENDS) Group. Eubacterial PCR for bacterial detection and identification in 100 acute postcataract surgery endophthalmitis. Invest Ophthalmol Vis Sci 2008; 49:1971– 8. 17. Daroy ML, Lopez JS, Torres BC, et al. Identification of unknown ocular pathogens in clinically suspected eye infections using ribosomal RNA gene sequence analysis. Clin Microbiol Infect 2011;17:776 –9. 18. Schabereiter-Gurtner C, Maca S, Röllenke S, et al. 16S rDNAbased identification of bacteria from conjunctival swabs by PCR and DGGE fingerprinting. Invest Ophthalmol Vis Sci 2001;42:1164 –71. 19. Dong Q, Brulc JM, Iovieno A, et al. Diversity of bacteria at healthy human conjunctiva. Invest Ophthalmol Vis Sci 2011; 52:5408 –13. 20. Marangon FB, Miller D, Muallem MS, et al. Ciprofloxacin and levofloxacin resistance among methicillin-sensitive Staphylococcus aureus isolates from keratitis and conjunctivitis. Am J Ophthalmol 2004;137:453– 8. 21. Akiyama T, Miyamoto H, Fukuda K, et al. Development of a novel PCR method to comprehensively analyze salivary bacterial flora and its application to patients with odontogenic infections. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 2010;109:669 –76. 22. Kawanami T, Fukuda K, Yatera K, et al. A higher significance of anaerobes: the clone library analysis of bacterial pleurisy. Chest 2011;139:600 – 8. 23. Morotomi N, Fukuda K, Nakano M, et al. Evaluation of intestinal microbiotas of healthy Japanese adults and effect of antibiotics using the 16S ribosomal RNA gene based clone library method. Biol Pharm Bull 2011;34:1011–20. 24. Roser DJ. Ethidium bromide: a general purpose fluorescent stain for nucleic acid in bacteria and eucaryotes and its use in microbial ecology studies. Soil Biol Biochem 1980;12:329 –36.

25. Lane DJ. 16S/23S rRNA sequencing. In: Stackebrandt E, Goodfellow M, eds. Nucleic Acid Techniques in Bacterial Systematics. New York: Wiley; 1991:115–75. 26. McDonough KA, Kress Y. Cytotoxicity for lung epithelial cells is a virulence-associated phenotype of Mycobacterium tuberculosis. Infect Immun 1995;63:4802–11. 27. Mansour JD, Schram JL, Schulte TH. Fluorescent staining of intracellular and extracellular bacteria in blood. J Clin Microbiol 1984;19:453– 6. 28. Stenson S, Newman R, Fedukowicz H. Laboratory studies in acute conjunctivitis. Arch Ophthalmol 1982;100:1275–7. 29. Acinas SG, Anton J, Rodriguez-Valera F. Diversity of freeliving and attached bacteria in offshore western Mediterranean waters as depicted by analysis of genes encoding 16S rRNA. Appl Environ Microbiol 1999;65:514 –22. 30. Fredricks DN, Fiedler TL, Marrazzo JM. Molecular identification of bacteria associated with bacterial vaginosis. N Engl J Med 2005;353:1899 –911. 31. Harris JK, De Groote MA, Sagel SD, et al. Molecular identification of bacteria in bronchoalveolar lavage fluid from children with cystic fibrosis. Proc Natl Acad Sci U S A 2007;104:20529–33. 32. Kassinen A, Krogius-Kurikka L, Makivuokko H, et al. The fecal microbiota of irritable bowel syndrome patients differs significantly from that of healthy subjects. Gastroenterology 2007;133:24 –33. 33. Munson MA, Banerjee A, Watson TF, Wade WG. Molecular analysis of the microflora associated with dental caries. J Clin Microbiol 2004;42:3023–9. 34. Paju S, Bernstein JM, Haase EM, Scannapieco FA. Molecular analysis of bacterial flora associated with chronically inflamed maxillary sinuses. J Med Microbiol 2003;52:591–7. 35. Gophna U, Sommerfeld K, Gophna S, et al. Differences between tissue-associated intestinal microfloras of patients with Crohn’s disease and ulcerative colitis. J Clin Microbiol 2006; 44:4136 – 41. 36. Goodfellow M, Kämpfer P, Vos PD, et al. Genus Parvimonas. In: De Vos P, Garrity GM, Jones D, et al, eds. Bergey’s Manual of Systematic Bacteriology. 2nd ed. Vol. 3. New York: Springer; 2009:1135– 6. 37. Cummins CS, Johnson JL. Genus Propionibacterium. In: Sneath PA, Mair NS, Sharpe ME, Holt JG, eds. Bergey’s Manual of Systematic Bacteriology. Vol. 2. Baltimore, MD: Williams & Wilkins; 1986:1346 –53. 38. Summanen P, Finegold SM. Genus Porphyromonas. In: Krieg NR, Staley JT, Brown DR, et al, eds. Bergey’s Manual of Systematic Bacteriology. 2nd ed. Vol. 4. New York: Springer; 2010:62–70. 39. Shah HN, Chattaway MA, Rajakurana L, Gharbia S. Genus Prevotella. In: Krieg NR, Staley JT, Brown DR, et al, eds. Bergey’s Manual of Systematic Bacteriology. 2nd ed. Vol. 4. New York: Springer; 2010:86 –102. 40. Gharbia SE, Shah HN, Edwards KJ. Genus Fusobacterium. In: Krieg NR, Staley JT, Brown DR, et al. Bergey’s Manual of Systematic Bacteriology. 2nd ed. Vol. 4. New York: Springer; 2010:748 –58. 41. Brook I. Ocular infections due to anaerobic bacteria. Int Ophthalmol 2001;24:269 –77. 42. Funke G, Pagano-Niederer M, Bernauer W. Corynebacterium macginleyi has to date been isolated exclusively from conjunctival swabs. J Clin Microbiol 1998;36:3670 –3. 43. Joussen AM, Funke G, Joussen F, Herbertz G. Corynebacterium macginleyi: a conjunctiva specific pathogen. Br J Ophthalmol 2000;84:1420 –2. 44. Riegel PR, Rummy R, de Briel D, et al. Genomic diversity and phylogenetic relationships among lipid-requiring diphtheroids from human and characterization of Corynebacterium macginleyi sp. nov. Int J Syst Bacteriol 1995;45:128 –33.

675

Ophthalmology

Volume 120, Number 4, April 2013

45. Ferrer C, Ruiz-Moreno JM, Rodríguez A, et al. Postoperative Corynebacterium macginleyi endophthalmitis. J Cataract Refract Surg 2004;30:2441– 4. 46. Suzuki T, Iihara H, Uno T, et al. Sutured-related keratitis caused by Corynebacterium macginleyi. J Clin Microbiol 2007; 45:3833– 6. 47. Kawanami T, Fukuda K, Yatera K, et al. Severe pneumonia with Leptotrichia sp. detected predominantly in bronchoalveo-

lar lavage fluid by use of 16S rRNA gene sequencing analysis. J Clin Microbiol 2009;47:496 – 8. 48. Zbinden R, von Graevenitz A. Actinobacillus, Capnocytophaga, Eikenella, Kingella, Pasteurella, and other fastidious or rarely encountered Gram-negative rods. In: Versalovic J, ed-in-chief, Carroll KC, Jorgensen JH, Funke G, et al, eds. Manual of Clinical Microbiology. 10th ed. Vol. 1. Washington, D.C.: ASM Press; 2011:574 – 87.

Footnotes and Financial Disclosures Financial Disclosure(s): The author(s) have no proprietary or commercial interest in any materials discussed in this article.

Originally received: June 25, 2012. Final revision: October 1, 2012. Accepted: October 1, 2012. Available online: December 12, 2012.

Manuscript no. 2012-929.

1

Department of Microbiology, School of Medicine, University of Occupational and Environmental Health Japan, Kitakyushu City, Japan.

2

Department of Ophthalmology, School of Medicine, University of Occupational and Environmental Health Japan, Kitakyushu City, Japan.

Correspondence: Rumi Aoki, Department of Ophthalmology, School of Medicine, University of Occupational and Environmental Health Japan, Kitakyushu City, Fukuoka, Japan. E-mail: [email protected].

Erratum With apologies from the authors, Table 1 in the publication entitled, “Clinical Applications of Cost Analysis of Diabetic Macular Edema Treatments” (Ophthalmology 2012;119:2558 – 62) had errors. Corrected Table 1 (with corrected numbers in bold) is printed below.

Table 1. Utilization Protocol, Outcomes, and Cost Parameters for Diabetic Macular Edema Treatment Modalities Laser, Surgical, or Intravitreal 1 Year Mean Age Dollars Per Dollars Per f/u Visits OCTs FA Inj. Treatments Cost ($) (yrs.) Lines Saved Line Saved Line-Year Saved QALY Grid Laser5 IVTA9,10 Poor VA (20/200-20/320) Laser IVTA Laser vs IVTA vs Ranibizumab (pseudophakic subset)15,16 Prompt laser IVTA ⫹ laser Ranibizumab Bevacizumab PACORES13 BOLT12 Ranibizumab DCRC14 READ17 Aflibercept*19 2q4 2q4⫻3, q8

3 3

3 3

1 0

2 2.3

1758 1427

52 63

0.26 0.42

6761 3749

228 188

7600 6267

3 3

4 4

1 1

2.9 3.3

2333 1907

63 63

1.4† 3.0†

1666 636

83 32

2767 1067

5 5 11

6 6 6

1 1 1

3 3⫹3 8.5

2613 3301 21289

63 62 63

.8 1.6 1.5

3266 2063 14192

158 100 706

5267 3333 23533

8 8

9 9

0 0

3 9

2490 4135

60 65

2.02 2.1

1233 1969

54 106

1800 3533

11 11

8 12

0 0

8.5 8

21265 21709

63 62

1.46 1.79

11372 11609

549 561

18300 18700

11 5

12 6

0 0

11 6.8

25913 15785

62 63

1.9 1.4

13689 11275

654 561

21800 18700

Each patient was assumed to have a level 4 new patient examination. FA ⫽ fluorescein angiography; f/u ⫽ follow-up; IVTA ⫽ intravitreal triamcinolone; OCT ⫽ optical coherence tomography. *Six month data published; usage extrapolated to 1 year and visual effect assumed stable. † Data only available for median (used average of 2 doses for IVTA)

676