Abstract
Purpose
To describe the bacterial findings by a targeted sequencing approach from corneal samples of patients with microbial keratitis and factors influencing culture outcome of indirectly inoculated corneal specimen.
Methods
Prospective inclusion of patients fulfilling predefined criteria of microbial keratitis. Samples from the corneal lesion were collected and dispensed in liquid transport medium, from which both culture and targeted amplification and sequencing of the V3-V4 region of the 16S rRNA gene were carried out. Additional standard corneal culture from the corneal lesions was also performed. Factors influencing culture outcome of indirectly inoculated corneal samples were identified by a multivariate regression model incorporating quantitative data from sequencing.
Results
Among the 94 included patients with microbial keratitis, contact lens wear (n = 69; 73%) was the most common risk factor. Contact lens wearers displayed significant differences in the bacterial community composition of the corneal lesion compared to no lens wearers, with higher abundance of Staphylococcus spp., Corynebacterium spp., and Stenotrophomonas maltophilia. Targeted sequencing detected a potential corneal pathogen in the highest proportional abundance among 9 of the 24 (38%) culture-negative patients with microbial keratitis. Age, bacterial density in the sample, and prior antibiotic treatment significantly influenced culture outcome of indirectly inoculated corneal samples.
Conclusion
Targeted sequencing may provide insights on pathogens in both culture negative episodes of microbial keratitis and among subgroups of patients with microbial keratitis as well as factors influencing culture outcome of indirectly inoculated corneal samples.
Similar content being viewed by others
Avoid common mistakes on your manuscript.
Introduction
Microbial keratitis caused by bacteria, fungi, or protozoa is a common ophthalmological emergency that can result in visual impairment and blindness. Bacteria are the most common cause of microbial keratitis in Europe, North America, and Australia/Oceania, being reported in 88.3–98.3% of the culture-positive cases that coincides with a high prevalence of contact lens wear [1].
The clinical presentation is usually sufficient to raise suspicion of an infectious origin. Still detection and susceptibility testing of the causative agent are vital in the management and proper treatment of the disease. Corneal culture and direct microscopy with staining have been the gold standard in the diagnostic procedure of microbial keratitis [2, 3]. However, the practice of corneal cultures has limitations. Different microorganisms require distinct substrate and incubation conditions, and only a modest proportion of cases with a clinical suspicion of microbial keratitis are culture-positive, with studies showing a median of 50% and range of 32.6–79.4% [1].
High-throughput sequencing of bacterial DNA extracted from primary samples is an attractive approach to identifying bacteria within corneal samples of patients with microbial keratitis [4, 5]. One common approach is targeted amplicon sequencing of the bacterial 16S rRNA gene present in all bacteria. This gene contains regions with high sequence diversity between bacterial families, which allows the classification of bacteria, often at genus level and sometimes at species level. Detection and classification of common bacteria at species level can also be achieved with metagenomic shotgun sequencing, but at a much higher financial cost [6].
Previous sequence-based studies have shown that the conjunctival microbiome in the healthy eye is heterogeneous, almost individual, and influenced by both age and sex [7].
We have previously reported on microbial findings from two different culture-based approaches [8]. In the present study, we used a combined quantifiable targeted sequencing approach to (i) describe bacterial findings in relation to culture results and clinical parameters in corneal samples obtained from patients with microbial keratitis, (ii) explore presumptive pathogens in the subgroup of patients with microbial keratitis where corneal cultures, directly and indirectly inoculated, are negative, and (iii) investigate variables, such as bacterial density and clinical characteristics, that influence the culture outcome of indirectly inoculated corneal samples from patients with microbial keratitis.
Methods
Study population
This prospective study was approved by the Regional Ethical Review Board of Uppsala, Sweden (ref: 2018/120). Informed written consent was collected from all participants prior to inclusion. All aspects of the management of patients participating in this study and methods applied within this study were carried out in accordance with relevant guidelines and regulations and adhered to the ethical standards of the Declaration of Helsinki. All patients aged 18 or older presenting with an episode of suspected microbial keratitis (i.e., a corneal infiltrate with an overlying epithelial defect) at the Department of Ophthalmology, Örebro University Hospital, Sweden, between 10 September 2018 and 27 January 2020 were considered for inclusion [8].
Patients were included if the corneal culture from the infiltrate showed growth of bacteria, yeast, or protozoa (all growth was considered as significant) and/or the patient fulfilled at least one of the clinical criteria for microbial keratitis, that is, infiltrate within/overlapping the central 4 mm of the cornea and/or uveitis and/or pain [9]. The exclusion criterion was corneal sampling not performed according to the study protocol.
Corneal culture according to two different approaches: standard culture and ESwab culture
The standard corneal culture procedure was in accordance with “Microbial keratitis Caused by Bacteria, Yeast, and Protozoa,” the Swedish state-of-the-art document on microbial keratitis during the study period. The corneal sampling order and culture procedure have been described previously [8]. In short, corneal samples were both directly inoculated (standard culture) and indirectly inoculated (ESwab culture) through liquid transport medium (modified Amies transport medium) of the ESwab kit (COPAN Italia S.p.A., Brescia, Italy) (see Supplementary Fig. 1). For details on culture media and incubation conditions, see Supplementary Table 1. The remaining transport medium of the ESwab kit, containing the corneal sample, was stored at − 80 °C pending further processing for DNA purification, PCR, quantitative PCR, and sequencing. Amplification and sequencing were performed after inclusion of the last patient in January 2020; hence the results of the targeted amplification and sequencing were not available during the ongoing disease episode.
Species determination was carried out with matrix assisted laser desorption/ionization time of flight-mass spectrometry (MALDI-TOF MS) with a Microflex LT and Biotyper 3.1 (Bruker Daltonik, Bremen, Germany).
DNA extraction, targeted amplicon sequencing, and quantitative PCR
DNA extraction, 16S rRNA amplicon sequencing, and quantitative PCR were performed as previously described [10]. In short, bacterial DNA was extracted from 200 µl of the ESwab transport medium on a MagNA Pure 96 instrument using the DNA and Viral NA Small Volume Kit (Roche, Mannheim, Germany) with an enzymatic pre-lysis with 3 mg lysozyme, 4 units lysostaphin, and 25 units mutanolysin per sample (Sigma-Aldrich, Merck, Germany). Amplicon libraries were prepared via a two-step PCR amplification of the 16S rRNA V3–V4 regions [11] using primers with heterogeneity spacers to increase the sequence diversity. Amplicon libraries were sequenced using an Illumina v3 600-cycle reagent kit on a MiSeq instrument (Illumina Inc., San Diego, USA). The absolute abundance of bacteria was estimated by qPCR using the same primer regions as for the amplicon sequencing and a 16S-TQM-528R TaqMan probe [12]. Details of primers, sequence pre-processing and the PCR methodology are provided as supplementary material.
Bioinformatics and statistics for the amplicon sequence data
Analyses of the amplicon sequence data were performed in R (v.4.0.2), using the following packages: DADA2 (v.1.12.1) [13] phyloseq (v.1.33.0) [14], vegan (v.2.5.6) [15], metacoder (v.0.3.5) [16], ANCOMBC (v.0.99.3) [17], and ggplot2 (v.3.3.2) [18]. The DADA2 package was used for determination of amplicon sequence variants (ASVs) and taxonomic classification using the Silva reference database (v.132) [19]. Sequence counts were agglomerated at genus level for all analyses, except when investigating bacterial richness (taxa counts) within samples, which was examined on the ASV level. Differences in richness between patient groups were tested with the Mann–Whitney U test. Differences in bacterial composition between selected patient groups (prior vs. no prior topical antibiotic treatment and contact lens wear vs. no contact lens wear) were studied by calculating sample-pairwise Bray–Curtis dissimilarities on Hellinger-transformed taxa counts followed by a statistical analysis of similarity (ANOSIM), where R (the effect size) close to 0 indicates completely similar compositions [20]. To compare differences in abundance of detected genera between contact lens wearers and non-wearers, a differential abundance analysis was performed using a composition of microbiomes with bias correction (ANCOM‐BC) modeling [17] with zero.cut set to 0.9 (exclusion of all genera present in less than 10% of the samples). The Benjamini‐Hochberg method was used to adjust for multiple testing. An adjusted p-value below 0.05 was considered to be of significance. The ANCOM-BC function was also used for testing whether any genera were more abundant within culture-negative samples than within culture-positive samples (culture-positive by either approach or both approaches). Here, zero.cut was set to 0.978, corresponding to the exclusion of genera present in less than two samples.
Statistical analysis of the clinical data and regression model
The Mann–Whitney U test was used for comparisons of age, symptom duration, visual acuity, lesion size, and absolute quantity of bacterial DNA in the corneal samples. The χ2 test, or Fisher’s exact test when appropriate, was used for comparisons on sex, laterality, and risk factors for microbial keratitis (i.e., contact lens wear or not) and whether antibiotic treatment had preceded corneal culture. A logistic regression model was used to examine associations between predictive variables such as age, sex, laterality, and 16S rRNA gene copy number (log10units) [4] and prior antibiotic treatment and culture outcome from indirect inoculation of the ESwab sample. A quantile regression model was chosen to describe the absolute amount of 16S rRNA gene copy numbers, since heteroscedasticity of the residuals with respect to the outcome was detected. The statistical analysis was performed in SPSS (v. 25).
Results
Patient characteristics
In all, 110 episodes of microbial keratitis were considered for inclusion. Of these, 16 were excluded due to not fulfilling the age criteria (n = 2), failure to adhere to randomized sampling order (n = 2), lack of written consent (n = 1), loss of the stored ESwab sample intended for molecular analysis (n = 5), and very low sequence count (i.e. < 1000; n = 2). In addition, four patients had two episodes each, where the second episode for each of these four patients was excluded (n = 4). The remaining 94 episodes of microbial keratitis in 94 patients constituted the study population. In the quantitative analysis, quantitative data for five samples were not available.
Median age in the cohort was 44 years (range 18–84), 51/94 (54%) were men, and the right eye was affected in 47/94 (50%) of cases. Topical antibiotic treatment preceded corneal sampling in 16 patients. Supplementary Table 2 contains information on preceding topical antibiotics. The most commonly identified risk factor for microbial keratitis in the cohort was contact lens wear (73%; n = 69). The contact lens wearers were significantly younger than the non-wearers (median age 42 years vs. 51 years; p = 0.003). The episodes presumably caused by contact lens wear had a significantly shorter median duration of symptoms at first visit; two days compared to six days among the episodes with other risk factors (p < 0.001). In comparison to non-wearers, contact lens wearers also displayed a significantly higher median visual acuity (Snellen) at the first visit (1.0 vs. 0.5, p < 0.001) and smaller lesions (median longest diameter: 1.0 mm vs. 1.7 mm, p < 0.001) (Table 1).
Culture results
Positive corneal cultures were detected in 70/94 (74%) of the episodes of microbial keratitis. Of these, 16 patients were culture-positive by standard culture only, 11 were culture-positive by ESwab culture only, and the remaining 43 were culture-positive by both methods. Monomicrobial growth was noted in 25 of the culture-positive patients and polymicrobial growth of 2–6 different bacteria belonging to 1–5 genera was seen in the remaining 45.
In the overall cohort, 151 different bacteria were isolated and identified to species level by culture (standard culture, ESwab culture, or both). Due to polymicrobial growth of different species belonging to the same genera among 13 patients, this corresponded at genus level to 127 distinct bacterial isolates belonging to 15 different bacterial genera: Staphylococcus (n = 46), Streptococcus (n = 2), Corynebacterium (n = 22), Micrococcus (n = 2), Nocardia (n = 1), Brachybacterium (n = 1), Haemophilus (n = 2), Moraxella (n = 3), Enterococcus (n = 3), Enterobacter (n = 1), Pantoea (n = 2), Pseudomonas (n = 3), Serratia (n = 1), Cutibacterium (n = 37), and Veionella (n = 1). An additional Gram-positive rod where no further identification was performed was also reported. Gram-positive bacteria constituted the largest group of isolated bacteria, predominantly Staphylococcus (46/127; 36%) and Corynebacterium (22/127; 17%). There were no dominant genera among the Gram-negative bacteria. Supplementary Table 3 contains information on bacteria isolated by the two culture methods.
Prediction model for culture outcome of indirectly inoculated corneal samples (ESwab culture)
We used qPCR data from the ESwab samples to create a model for predicting a positive culture outcome from the same ESwab sample, including five independent variables: age, 16S rRNA copy number (log10units), prior antibiotic treatment, laterality, and sex. The model was statistically significant (χ2 (df 5) = 23.898; p < 0.001), explained 31.6% (Nagelkerke R2) of the variance in culture outcome, and correctly classified 67.4% of the cases with a sensitivity of 74.0% and specificity of 59.0%. Positive and negative predictive values were 69.8% and 63.9% respectively. Three of the five predictive variables were statistically significant: age, 16S rRNA gene copy number (log10units), and antibiotic treatment prior to sampling, Supplementary Table 4 summarizes the regression model. The 16S rRNA gene copy number (log10units) displayed an OR of 6.3 (95% CI, 1.6–25.0; p = 0.009). The odds for a positive culture outcome from the ESwab sample also increased with increasing age (OR 1.039; p = 0.034), while prior antibiotic treatment decreased the odds for a positive culture (OR 0.177; p = 0.031) (Supplementary table 4).
Variables influencing the bacterial densities in the ESwab samples of patients with microbial keratitis
The 16S rRNA gene copy numbers displayed large dispersion within the cohort, with a range of 53–249,873 copies among the studied episodes of microbial keratitis (median copy number: 232). Univariate analysis revealed a significantly higher median copy number among the culture-positive patients compared to their culture-negative counterparts, regardless of culture method (standard culture, ESwab culture, or when all culture results were considered; Table 1). In the multivariate regression model with the five independent variables: lesion size (longest diameter), age, prior topical antibiotic treatment, risk factor for keratitis (in terms of contact lens wear or not), and sampling order (first or second) only lesion size had a significant influence on 16S rRNA gene copy number, with an increase of 65 copies for each mm increase in lesion diameter, Supplementary Table 5 summarizes the regression model.
Microbiome profiles
16S rRNA amplicon sequencing of the corneal samples revealed a median presence of 17 ASVs (range, 8–47) and 15 different bacterial genera (range, 8–30) per sample. Some samples were dominated by a single genus (e.g., Staphylococcus, Pseudomonas Moraxella, Corynebacterium, or Streptococcus), whereas others were more diverse and heterogeneous (Fig. 1). However, the ASV richness (alpha diversity) was not significantly influenced by disease severity in terms of duration of topical pharmacological treatment or surgical intervention (data not shown). No sample was PCR negative.
Overall, 13 of the 15 genera identified by culture methods could be detected by sequencing. At genus level, 96/127 (76%) of the different bacteria isolates detected by culture were also detected by sequencing within the same sample. From a patient perspective, this corresponds to a partial agreement rate of 87% between culture and targeted sequencing; that is, in 61/70 (87%) of the culture-positive patients, at least one bacterial genus that was isolated by culture was also detected by the sequencing approach. The complete agreement rate was 61%; in that 43/70 (61%) of the culture-positive patients had all of the genera isolated by culture also detected by targeted sequencing. Of the 24 culture-negative episodes, all were PCR positive, see below.
No ASVs belonging to either Nocardia or Pantoea were detected by sequencing. The ASV richness (number of unique ASVs observed within a sample) was significantly higher within the culture-positive samples (median, 19 ASVs; interquartile range (IQR), 15–25) compared to the culture-negative samples (median, 15 ASVs; IQR, 13–17; p = 0.002).
Among 9 of the 24 culture negative samples a previously described corneal pathogen [8, 22] was detected in the highest proportional abundance in the sample, in a proportional abundance of 22%-83% (Fig. 2A–C).
Of these nine patients, six received initial therapy with a fluoroquinolone alone and their samples displayed highest proportional abundance of Pseudomonas (n = 2), Staphylococcus (n = 2) and one each of Veillonella and Clostridium, respectively. A single patient received initial treatment of fortified drops i.e., vancomycin and ceftazidime, in this patient’s sample sequencing detected an 83% abundance of Staphylococcus. The remaining two patients were initially treated with a combination of fluoroquinolone and chloramphenicol, their samples displayed the highest proportional abundance of Staphylococcus and Brevundimonas, respectively. Only one patient, with the highest proportional abundance of Pseudomonas, received additional antibacterial therapy, with tobramycin due to clinical course, and only one patient, the one with initial fortified drops required surgical intervention with amniotic membrane. The total treatment duration (median; range) for respective treatment group was 17 days; 6–18 and 17, 5 days; 15–20; and 16 days for initial, empirical treatment with fluoroquinolone only, combination of fluoroquinolone, and chloramphenicol and fortified drops respectively.
Differential abundance analysis indicated that only Brevundimonas was significantly enriched within the culture-negative samples compared to the culture-positive samples (fold-change, 8.2; 95% CI, 2.1–31.3; adjusted p < 0.05).
Bacterial compositional differences (beta diversity) (R = 0.14, p = 0.01) was observed between contact lens wearers and non-contact lens wearers, with communities being more homogeneous among the contact lens wearers. Part of this observed variation was likely due to higher proportional abundances of Staphylococcus spp. (fold-change: 11.7; 95% CI, 3.7–37.4; adjusted p < 0.01), Corynebacterium spp. (fold-change, 3.6; 95% CI, 1.6–7.8; adjusted p < 0.05), and Stenotrophomonas maltophilia (fold-change, 3.0; 95% CI, 1.8–5.3; adjusted p < 0.001) within the corneal lesions of contact lens wearers compared to non-wearers (Fig. 3). S. maltophilia was present in 13/69 (19%) of the samples collected from contact lens wearers, constituting more than 20% of the sequence reads in one of the patients, but was only detected in one sample among the 25 non-wearers, at a proportional abundance below 0.1%. There was no difference in the median ASV richness between the contact lens wearers and non-wearers.
Prior antibiotic treatment had no significant influence on the composition of bacterial communities in the corneal ulcer, as the composition (beta diversity) was similar among patients who were and were not treated with antibiotics prior to sampling. However, a reduced ASV richness (alpha diversity) was observed among the antibiotic-treated patients (median, 14 ASVs; IQR, 12–17) compared to the treatment-naïve (median, 18 ASVs; IQR, 14–22; p = 0.05). There was also a significant difference in the total number of unique genera when comparing antibiotic-treated patients (median, 13 genera; IQR, 10–15) with the treatment-naïve (median, 16 genera; IQR, 12–18; p = 0.02).
Discussion
Microbial keratitis is an ophthalmological emergency and a common cause of corneal visual impairment and monocular blindness globally [23,24,25]. Corneal culture in cases of microbial keratitis has its limitations; for example, the median culture positivity rate is ≈50% of clinically cases of microbial keratitis [1]. In the literature, the clinical importance to identify the disease-causing pathogen is inconclusive [26,27,28,29].
In the present study, nine patients with a negative corneal culture displayed a previously reported corneal pathogen, in the highest abundance. Of these patients, six received topical empirical treatment with a fluoroquinolone only. Among those six patients, five displayed a corneal pathogen with a previously reported reduced sensitivity of fluoroquinolones of 20–38.8% [1, 30] One of these episodes required additional topical therapy, but none of these six patients required any surgical intervention. This favorable outcome may reflect the low prevalence of reduced susceptibility to fluoroquinolones reported previously from the setting of this present study [31].
In Europe, the USA, and Australia, episodes of microbial keratitis are dominated by a bacterial etiology [1], and an association with contact lens wear has been reported in 28.9–63% of cases [32,33,34]. In the present study, the rate of contact lens wear among patients with episodes of microbial keratitis was high, at 73%. This may be due to an awareness among contact lens wearers, opticians, and the healthcare system that unilateral redness and pain in combination with contact lens wear is enough to raise suspicion of microbial keratitis and should result in immediate contact with an ophthalmologist or equivalent. This in turn may also provide a partial explanation for the significantly shorter symptom duration, significantly higher visual acuity, and smaller lesions among the contact lens wearers in the present study, compared to those with other risk factors for microbial keratitis.
A majority of the genera isolated by culture in the cohort of patients with microbial keratitis could also be detected by applying 16S rRNA amplicon sequencing. In all, 76% of the different bacterial isolates identified at genus level were also detected by sequencing in the same sample. This is similar to previously reported concordance rates of 63–80% between sequencing and culture among patients with monomicrobial bacterial ulcers [5, 35]. In the present study, the microbiota of the corneal ulcers from the contact lens wearers were more homogenous than those of the ulcers from non-wearers, with higher proportional abundance of genera including Staphylococcus and Corynebacterium. This is in contrast to a previous report from the healthy conjunctiva [36].
We found that S. maltophilia was present in a significantly higher proportional abundance among the contact lens wearers than among patients with other risk factors for keratitis. S. maltophilia is an aerobic Gram-negative rod of the family Xanthomonadaceae. It is closely related to Pseudomonas, that can be found in the environment in water or humid settings, and is considered an opportunistic human pathogen with multidrug resistance potential [37, 38]. Trauma, contact lens wear, and penetrating keratoplasty have previously been reported as common risk factors among episodes of keratitis displaying growth of S. maltophilia [39, 40]. One explanation for the significantly higher abundance of S. maltophilia among contact lens wearers in the present study may be its ability to form biofilm. Both the contact lens itself and the lens case provide potential surfaces for biofilm growth, which gives a partial explanation for why contact lens wear is advantageous to this pathogen. This may be supported by the findings of Wiley et al., who reported that S. maltophilia is resistant to lens cleanser and was one of the dominant bacteria in biofilms detected on contact lenses [41].
In the present study, antibiotic treatment prior to sampling had no significant influence on either the composition of bacterial genera or the absolute number of bacteria (measured by qPCR) in the corneal ulcers of patients with microbial keratitis. This is supported by earlier finding by Darden et al. on the composition of the feline ocular surface microbiome after topical antibiotic treatment [42]. However, a reduced ASV richness was observed among the antibiotic-treated patients in the present study.
Brevundimonas was found to be significantly enriched among our culture-negative samples compared to the culture-positive samples. Other known corneal pathogens such as Clostridium, Staphylococcus, Veillonella, and Pseudomonas constituted at least 20% of the relative abundance among almost half of all of the culture-negative samples.
Brevundimonas spp. are Gram-negative, non-fermenting aerobe rods. Microbial keratitis presumably caused by both Brevundimonas diminuta [43] and Brevundimonas vesicularis [44] has been previously reported. In the case report of the B. diminuta keratitis, the bacteria was isolated from broth only, after at least seven days of incubation. Unfortunately, the culture media and incubation conditions were not included in the case report on the B. vesicularis keratitis. However, both B. diminuta and B. vesicularis are known to grow slowly on commonly-used nutrient media, but once isolated, can be correctly identified to species level with MALDI-TOF MS [45]. The lack of Brevundimonas isolated by culture in the present study may be due to the culture media and incubation conditions, since the fastidious anaerobe broth was discarded after seven days and no selective medium for Gram-negative bacteria was used. The findings of Brevundimonas by targeted sequencing may provide a potential causative pathogen among some of the culture-negative cases of microbial keratitis in the present study, as Brevundimonas constituted more than 10% of the reads in 5/24 samples.
The inoculation strategy (i.e., whether to transfer the corneal sample directly on culture media or indirectly via transport media) has been studied previously, but with inconclusive results [8, 46,47,48,49]. In the present study, we could explain > 30% of the variance in culture outcome of indirectly inoculated corneal samples and that the OR of a positive culture outcome increased with the 16S rRNA gene copy number and with age, and decreased with prior topical antibiotic treatment. At present, there is a lack of reports on variables influencing the culture outcome of indirectly inoculated corneal samples from patients with microbial keratitis [50,51,52].
The diagnostic and clinical applications of using qPCR to determine the absolute amount of bacteria in corneal samples of patients with microbial keratitis have been explored previously [4]. In the present study, the amount of bacteria displayed great variability (range of 53–249,873 16S rRNA gene copies) within the cohort, despite the prospective study design, strict inclusion criteria, and randomized sampling order with standardized sampling. We created a regression model in an attempt to explain this variance, but the model could only explain 0.3% of the variance in 16S rRNA gene copy number. Lesion size was the only variable with a significant effect on the copy number, highlighting the complex nature of the ocular microbiome.
Topical anesthetics have previously been reported to influence the microbiome findings from the conjunctiva [53]. This is likely due to patients’ tolerance of the pressure applied when sampling, since it has been reported that deep sampling (high pressure applied with dry swab) significantly affects the abundances of detected bacteria compared to soft sampling (minimal pressure applied with moist swab) [54]. In the present study, all patients received the same topical preservation-free anesthetics prior to sampling, the swab used for sampling in all ulcers was applied dry and was of the same model and fabrication in all patients, and samples were collected from a corneal ulcer rather than the conjunctiva. We did not standardize or control the duration or strength of pressure with which the sampling swab was applied to the corneal lesion. This may provide at least a partial explanation for the wide range of microbial DNA detected from the corneal lesions of our participants, and the difficulties in explaining this with the regression model used. There are likely also other factors, related both to the sampling process, clinical history, and patient characteristics and to the microbial characteristics and differences in pathogenicity, that influence the amount of bacterial DNA that can be detected from the corneal lesions of patients with microbial keratitis.
The strengths of the current study are the prospective design and the sample size. The main limitations are the lack of control of factors influencing the microbial DNA retrieval from the corneal ulcer, and the fact that only one sample per episode was retrieved for sequencing.
In conclusion, in 87% of the culture-positive episodes of microbial keratitis, sequencing detected at least one of the bacterial genera isolated by culture. Among more than one third of the culture-negative episodes, sequencing detected a previously described corneal pathogen in the highest proportional abundance. These findings indicate that targeted sequencing may provide valuable information in a clinical context. However, future studies, which also allow for information from molecular diagnostic methods, are needed to evaluate if this additional information significantly improves the outcome for patients with microbial keratitis, and/or provides other health economic benefits. Corneal lesions among contact lens wearers displayed a significantly different bacterial community composition compared to lesions of patients with other risk factors for microbial keratitis. Finally, quantitative data from targeted sequencing indicated that the culture outcome of indirectly inoculated corneal samples may be influenced by the absolute bacteria count in the sample.
References
Ung L, Bispo PJM, Shanbhag SS, Gilmore MS, Chodosh J (2019) The persistent dilemma of microbial keratitis: global burden, diagnosis, and antimicrobial resistance. Surv Ophthalmol 64(3):255–271. https://doi.org/10.1016/j.survophthal.2018.12.003
Jones DB, Liesegang TJ, Robinson NM, Washington JA (1981) Laboratory diagnosis of ocular infections. American Society for Microbiology, Cumitech: ch13;1–27. Washington, DC
Zemba M, Dumitrescu OM, Dimirache AE, Branisteanu DC, Balta F, Burcea M, Moraru AD, Gradinaru S (2022) Diagnostic methods for the etiological assessment of infectious corneal pathology (Review). Exp Ther Med 23(2):137. https://doi.org/10.3892/etm.2021.11060
Shimizu D, Miyazaki D, Ehara F, Shimizu Y, Uotani R, Inata K, Sasaki SI, Inoue Y (2020) Effectiveness of 16S ribosomal DNA real-time PCR and sequencing for diagnosing bacterial keratitis. Graefes Arch Clin Exp Ophthalmol 258(1):157–166. https://doi.org/10.1007/s00417-019-04434-8
Knox CM, Cevellos V, Dean D (1998) 16S ribosomal DNA typing for identification of pathogens in patients with bacterial keratitis. J Clin Microbiol 36(12):3492–3496. https://doi.org/10.1128/jcm.36.12.3492-3496.1998
Grogan MD, Bartow-McKenney C, Flowers L, Knight SAB, Uberoi A, Grice EA (2019) Research Techniques made simple: profiling the skin microbiota. J Invest Dermatol 139(4):747-752.e741. https://doi.org/10.1016/j.jid.2019.01.024
Wen X, Miao L, Deng Y, Bible PW, Hu X, Zou Y, Liu Y, Guo S, Liang J, Chen T, Peng GH, Chen W, Liang L, Wei L (2017) The influence of age and sex on ocular surface microbiota in healthy adults. Invest Ophthalmol Vis Sci 58(14):6030–6037. https://doi.org/10.1167/iovs.17-22957
Sagerfors S, Karakoida C, Sundqvist M, Ejdervik Lindblad B, Söderquist B (2021) Corneal culture in infectious keratitis: effect of the inoculation method and media on the corneal culture outcome. J Clin Med 10(9):1810. https://doi.org/10.3390/jcm10091810
Stapleton F, Keay L, Edwards K, Naduvilath T, Dart JK, Brian G, Holden BA (2008) The incidence of contact lens-related microbial keratitis in Australia. Ophthalmology 115(10):1655–1662. https://doi.org/10.1016/j.ophtha.2008.04.002
Edslev SM, Andersen PS, Agner T, Saunte DML, Ingham AC, Johannesen TB, Clausen ML (2021) Identification of cutaneous fungi and mites in adult atopic dermatitis: analysis by targeted 18S rRNA amplicon sequencing. BMC Microbiol 21(1):72. https://doi.org/10.1186/s12866-021-02139-9
Klindworth A, Pruesse E, Schweer T, Peplies J, Quast C, Horn M, Glöckner FO (2013) Evaluation of general 16S ribosomal RNA gene PCR primers for classical and next-generation sequencing-based diversity studies. Nucleic Acids Res 41(1):e1. https://doi.org/10.1093/nar/gks808
Nadkarni MA, Martin FE, Jacques NA, Hunter N (2002) Determination of bacterial load by real-time PCR using a broad-range (universal) probe and primers set. Microbiology (Reading) 148(Pt 1):257–266. https://doi.org/10.1099/00221287-148-1-257
Callahan BJ, McMurdie PJ, Rosen MJ, Han AW, Johnson AJ, Holmes SP (2016) DADA2: High-resolution sample inference from Illumina amplicon data. Nat Methods 13(7):581–583. https://doi.org/10.1038/nmeth.3869
McMurdie PJ, Holmes S (2013) phyloseq: an R package for reproducible interactive analysis and graphics of microbiome census data. PLoS One 8(4):e61217. https://doi.org/10.1371/journal.pone.0061217
Oksanen J, Blanchet FG, Friendly M, Kindt R, Legendre P, McGlinn D, Minchin P, O’Hara R, Simpson G, Solymos P et al (2019) vegan: Community Ecology Package. R package version 2.5–6. https://CRAN.R-project.org/package=vegan
Foster ZS, Sharpton TJ, Grünwald NJ (2017) Metacoder: An R package for visualization and manipulation of community taxonomic diversity data. PLoS Comput Biol 13(2):e1005404. https://doi.org/10.1371/journal.pcbi.1005404
Lin H, Peddada SD (2020) Analysis of compositions of microbiomes with bias correction. Nat Commun 11(1):3514. https://doi.org/10.1038/s41467-020-17041-7
Wickham H (2016) ggplot2: elegant graphics for data analysis. Springer-Verlag, New York
Callahan B (2018) Silva taxonomic training data formatted for DADA2 (Silva version 132). https://doi.org/10.5281/zenodo.1172783
Clarke KR (1993) Non-parametric multivariate analyses of changes in community structure. Aust J Ecol 18(1):117–143
Holladay JT (1997) Proper method for calculating average visual acuity. J Refract Surg 13(4):388–391. https://doi.org/10.3928/1081-597x-19970701-16
Bartimote CFJ, Watson, S (2019) The spectrum of microbial keratitis: an updated review. Open Ophthalmol J 13(1)100–130. https://doi.org/10.2174/1874364101913010100
Ting DSJ, Ho CS, Deshmukh R, Said DG, Dua HS (2021) Infectious keratitis: an update on epidemiology, causative microorganisms, risk factors, and antimicrobial resistance. Eye (London, England) 35(4):1084–1101. https://doi.org/10.1038/s41433-020-01339-3
Whitcher JP, Srinivasan M, Upadhyay MP (2001) Corneal blindness: a global perspective. Bull World Health Organ 79(3):214–221
Whitcher JP, Srinivasan M (1997) Corneal ulceration in the developing world–a silent epidemic. Br J Ophthalmol 81(8):622–623
Levey SB, Katz HR, Abrams DA, Hirschbein MJ, Marsh MJ (1997) The role of cultures in the management of ulcerative keratitis. Cornea 16(4):383–386
McLeod SD, Kolahdouz-Isfahani A, Rostamian K, Flowers CW, Lee PP, McDonnell PJ (1996) The role of smears, cultures, and antibiotic sensitivity testing in the management of suspected infectious keratitis. Ophthalmology 103(1):23–28
Vital MC, Belloso M, Prager TC, Lanier JD (2007) Classifying the severity of corneal ulcers by using the “1, 2, 3” rule. Cornea 26(1):16–20. https://doi.org/10.1097/ICO.0b013e31802b2e47
Ung L, Wang Y, Vangel M, Davies EC, Gardiner M, Bispo PJM, Gilmore MS, Chodosh J (2020) Validation of a comprehensive clinical algorithm for the assessment and treatment of microbial keratitis. Am J Ophthalmol 214:97–109. https://doi.org/10.1016/j.ajo.2019.12.019
Banawas SS (2022) Systematic review and meta-analysis on the frequency of antibiotic-resistant Clostridium species in Saudi Arabia. Antibiotics (Basel) 11(9). https://doi.org/10.3390/antibiotics11091165
Sagerfors S, Ejdervik-Lindblad B, Söderquist B (2020) Infectious keratitis: isolated microbes and their antibiotic susceptibility pattern during 2004–2014 in Region Örebro County, Sweden. Acta Ophthalmol 98(3):255–260. https://doi.org/10.1111/aos.14256
Ferreira CS, Figueira L, Moreira-Goncalves N, Moreira R, Torrao L, Falcao-Reis F (2018) Clinical and microbiological profile of bacterial microbial keratitis in a Portuguese tertiary referral center-where are we in 2015? Eye Contact Lens 44(1):15–20. https://doi.org/10.1097/icl.0000000000000298
Jeng BH, Gritz DC, Kumar AB, Holsclaw DS, Porco TC, Smith SD, Whitcher JP, Margolis TP, Wong IG (2010) Epidemiology of ulcerative keratitis in Northern California. Arch Ophthalmol 128(8):1022–1028. https://doi.org/10.1001/archophthalmol.2010.144
Khoo P, Cabrera-Aguas MP, Nguyen V, Lahra MM, Watson SL (2020) Microbial keratitis in Sydney, Australia: risk factors, patient outcomes, and seasonal variation. Graefes Arch Clin Exp Ophthalmol 258(8):1745–1755. https://doi.org/10.1007/s00417-020-04681-0
Kim E, Chidambaram JD, Srinivasan M, Lalitha P, Wee D, Lietman TM, Whitcher JP, Van Gelder RN (2008) Prospective comparison of microbial culture and polymerase chain reaction in the diagnosis of corneal ulcer. Am J Ophthalmol 146(5):714–723, 723.e711. https://doi.org/10.1016/j.ajo.2008.06.009
Zhang H, Zhao F, Hutchinson DS, Sun W, Ajami NJ, Lai S, Wong MC, Petrosino JF, Fang J, Jiang J, Chen W, Reinach PS, Qu J, Zeng C, Zhang D, Zhou X (2017) Conjunctival microbiome changes associated with soft contact lens and orthokeratology lens wearing. Invest Ophthalmol Vis Sci 58(1):128–136. https://doi.org/10.1167/iovs.16-20231
Looney WJ, Narita M, Mühlemann K (2009) Stenotrophomonas maltophilia: an emerging opportunist human pathogen. Lancet Infect Dis 9(5):312–323. https://doi.org/10.1016/s1473-3099(09)70083-0
Brooke JS (2012) Stenotrophomonas maltophilia: an emerging global opportunistic pathogen. Clin Microbiol Rev 25(1):2–41. https://doi.org/10.1128/cmr.00019-11
Park BC, Lim HR, Park SJ, Koh JW (2021) Clinical features and management of Stenotrophomonas maltophilia keratitis. Ophthalmol Ther 10(3):525–533. https://doi.org/10.1007/s40123-021-00348-z
Palioura S, Gibbons A, Miller D, OʼBrien TP, Alfonso EC, Spierer O (2018) Clinical features, antibiotic susceptibility profile, and outcomes of infectious keratitis caused by Stenotrophomonas maltophilia. Cornea 37(3):326–330. https://doi.org/10.1097/ico.0000000000001486
Wiley L, Bridge DR, Wiley LA, Odom JV, Elliott T, Olson JC (2012) Bacterial biofilm diversity in contact lens-related disease: emerging role of Achromobacter, Stenotrophomonas, and Delftia. Invest Ophthalmol Vis Sci 53(7):3896–3905. https://doi.org/10.1167/iovs.11-8762
Darden JE, Scott EM, Arnold C, Scallan EM, Simon BT, Suchodolski JS (2019) Evaluation of the bacterial ocular surface microbiome in clinically normal cats before and after treatment with topical erythromycin. PLoS One 14(10):e0223859. https://doi.org/10.1371/journal.pone.0223859
Pandit RT (2012) Brevundimonas diminuta keratitis. Eye Contact Lens 38(1):63–65. https://doi.org/10.1097/ICL.0b013e31821c04f7
Pelletier JS, Ide T, Yoo SH (2010) Brevundimonas vesicularis keratitis after laser in situ keratomileusis. J Cataract Refract Surg 36(2):340–343. https://doi.org/10.1016/j.jcrs.2009.07.050
Ryan MP, Pembroke JT (2018) Brevundimonas spp: Emerging global opportunistic pathogens. Virulence 9(1):480–493. https://doi.org/10.1080/21505594.2017.1419116
Kaye SB, Rao PG, Smith G, Scott JA, Hoyles S, Morton CE, Willoughby C, Batterbury M, Harvey G (2003) Simplifying collection of corneal specimens in cases of suspected bacterial keratitis. J Clin Microbiol 41(7):3192–3197
McLeod SD, Kumar A, Cevallos V, Srinivasan M, Whitcher JP (2005) Reliability of transport medium in the laboratory evaluation of corneal ulcers. Am J Ophthalmol 140(6):1027–1031. https://doi.org/10.1016/j.ajo.2005.06.042
Pakzad-Vaezi K, Levasseur SD, Schendel S, Mark S, Mathias R, Roscoe D, Holland SP (2015) The corneal ulcer one-touch study: a simplified microbiological specimen collection method. Am J Ophthalmol 159(1):37-43.e31. https://doi.org/10.1016/j.ajo.2014.09.021
Nielsen SE, Gertsen JB, Kjaersgaard M, Ivarsen A, Hjortdal J (2016) New diagnostic tool in bacterial keratitis is not superior to traditional agar plates. Acta Ophthalmol 94(7):e671–e672. https://doi.org/10.1111/aos.13051
Ting DSJ, Cairns J, Gopal BP, Ho CS, Krstic L, Elsahn A, Lister M, Said DG, Dua HS (2021) Risk factors, clinical outcomes, and prognostic factors of bacterial keratitis: the Nottingham Infectious Keratitis Study. Front Med (Lausanne) 8:715118. https://doi.org/10.3389/fmed.2021.715118
Bhadange Y, Das S, Kasav MK, Sahu SK, Sharma S (2015) Comparison of culture-negative and culture-positive microbial keratitis: cause of culture negativity, clinical features and final outcome. Br J Ophthalmol 99(11):1498–1502. https://doi.org/10.1136/bjophthalmol-2014-306414
van der Meulen IJ, van Rooij J, Nieuwendaal CP, Van Cleijnenbreugel H, Geerards AJ, Remeijer L (2008) Age-related risk factors, culture outcomes, and prognosis in patients admitted with infectious keratitis to two Dutch tertiary referral centers. Cornea 27(5):539–544. https://doi.org/10.1097/ICO.0b013e318165b200
Shin H, Price K, Albert L, Dodick J, Park L, Dominguez-Bello MG (2016) Changes in the eye microbiota associated with contact lens wearing. mBio 7(2):e00198. https://doi.org/10.1128/mBio.00198-16
Dong Q, Brulc JM, Iovieno A, Bates B, Garoutte A, Miller D, Revanna KV, Gao X, Antonopoulos DA, Slepak VZ, Shestopalov VI (2011) Diversity of bacteria at healthy human conjunctiva. Invest Ophthalmol Vis Sci 52(8):5408–5413. https://doi.org/10.1167/iovs.10-6939
Funding
Open access funding provided by Örebro University. This study was supported by the Region Örebro County Council Research Committee.
Author information
Authors and Affiliations
Contributions
All authors contributed to the study conception and design. Material preparation, data collection, and analysis were performed by SS, SE, and BL. The first draft of the manuscript was written by SS and SE, and all authors commented on previous versions of the manuscript and read and approved the final manuscript.
Corresponding author
Ethics declarations
Ethics approval
This study adhered to the ethical standards of the Declaration of Helsinki. Ethical approval was obtained by the Regional Ethical Review Board of Uppsala, Sweden (ref: 2018/120).
Consent to participate
Informed written consent was obtained from all participants in the study.
Conflict of interest
The authors declare no competing interests.
Disclaimer
The funding organization had no role in the design or conduct of this research.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary Information
Below is the link to the electronic supplementary material.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Sagerfors, S., Edslev, S., Lindblad, B.E. et al. In the eye of the ophthalmologist: the corneal microbiome in microbial keratitis. Graefes Arch Clin Exp Ophthalmol 262, 1579–1589 (2024). https://doi.org/10.1007/s00417-023-06310-y
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00417-023-06310-y