Lecture: Microbial Keratitis: Diagnosis and Management

Today, I will be talking to you about microbial keratitis, diagnosis, assessment, and clinical management. This is an intermediate lecture. We’ll assume you know what corneal infections are and we’ll briefly discuss terminology and epidemiology. And then talk about fundamentals as well as intermediate level knowledge on how to diagnose, clinically assess and based on that, how to manage infections. So here are my disclosures. So, first let’s do a quick rapid deep dive into the epidemiology of infections. These are really very important diseases that are neglected in terms of research, in terms of development of new diagnostics and therapeutics and they’re underappreciated in epidemiologic surveys. Corneal blindness affects 6 million people worldwide and it’s the fifth leading cause of blindness and infections are the biggest cause. They can cause from 1.5 to 2 million new cases of unilateral blindness every year. They’re a burden for various reasons and it’s likely underestimated. For infectious corneal ulcers, even the most basic data like incidence, prevalence, contribution to disability, adjusted life years, and loss of productivity are all unknown. The only data available comes from outdated population surveys that suggest that south and southeast Asia are the epicenters of disease. But including infectious corneal ulcers with a broad category of corneal opacity has not provided the details required for accurate epidemiologic study. So a group of infectious keratitis experts in recent years have proposed that infectious corneal ulceration is a group of neglected group of tropical diseases. The World Health Organization has defined blindness as visual acuity that is 20/400 or worse in the better seeing eye. What that means is that infectious keratitis and the burden of that disease is ignored. The mean age of onset for infections is 45 years. They tend to affect younger patients who have a lot of earning potential and the need to support their families. They sufficient from social stigma from the corneal scarring and that has severe consequences for their quality of life. The surveys that do exist that try to estimate the incidence or prevalence show highly different results depending on the region of the world. In Minnesota, a survey in the 1950s showed 2.5 cases per 100,000 patients. A later study in the 80s, showed 11 per 100,000. In Scotland in 1995, it was 1.6 per 100 in. Nottingham, 34.7 per 100,000 and period of timer mouth, 40.3. When you go to the other side of the world you see the numbers are very different. In India in 1996, it was 113 cases per 100,000 patients. 710, orders of magnitude greater than in Europe and North America. And 339 in Bhutan. And 799 in Nepal. In Africa, there is a conservative estimate, extrapolations is there are around 75,000 cases of fungal keratitis alone and it’s one of the leading causes of eye removal in Africa. So how common are corneal infections? The data is highly variable. In developed countries, the incidence and prevalence might be lower but it’s still a major unmet need in terms of diagnostics and therapeutics and preventing blindness. In developing countries, the numbers are far worse. So for these reasons, the corneal ulceration in the developing world is a silent epidemic and we’re only seeing the tip of the iceberg as far as the burden of this devastating blinding disease. We will talk about microbes in cultural positive microbial keratitis. That is fungi or bacteria or parasites. When you look at the regions and sort by what proportion are fungal versus bacterial, there is a bit of a distribution here. China and India and other parts of Asia have more fungal keratitis. Sometimes exceeding 50%. And other parts of the world have more overwhelmingly bacterial keratitis with a small portion of Acanthamoeba keratitis. Contact lens wearers is a big one in developed countries. Contact lens-associated keratitis has had a stable incidence since the 1980s but given there are millions of contact lens wearers in the United States, there are many casings of infectious keratitis. Daily connect lances are associated with lower risk and less severe cases. Ocular trauma is a risk factor. Bacterial or fungal keratitis can be in farm workers in India or Africa or in developed countries as well. Someone has an occupational injury from working in the fields or manual labor and that results in infectious keratitis. Fungal keratitis is notorious for being caused by trauma in farmers and laborers. Ocular trauma more commonly associated with bacterial keratitis caused by gram positive commensal organisms. Recent eye surgery is a risk factor well known for keratitis that is associated with procedures. Mycobacterial Chelona infection. Congenital skew layer surface disease from the bacteria on the surface and the eyelid margin. And topical steroids are another risk factor and immunocompromised and Candida keratitis is another example. There are some broad categories of infectious keratitis. Bacteria, fungi, protozoa and viruses and we’ll cover all four of those in terms of diagnosis, assessment, and targeted treatment. The epidemiology of bacterial keratitis. The bacterial genome is large and diverse. There are gram positives and gram negatives which are the two broad categories. So here are the most common causes of bacterial keratitis. No. 1 is coagulase negative staph. This is across many surveys. This is the most common cause of bacterial keratitis. No. 2, tends to be pseudomonas species and there are staph, strep, nocardia and H flu. So the epidemiologic trends is there is increased incidence from the 1950s to the 1980s. By the 1980s, over 30 percent are associated with contact lens wear. And 30,000 new cases in the U.S. alone. Over $20 million a year is spent treating bacterial keratitis in the US alone. Newer trends is there is some reduction in contact lens cases and more increase in ocular surface disease-related infections. There are numerous reports of antibiotic resistance and it’s association with worse clinical outcomes. Fungal keratitis. There are two categories. Filamentous and yeast. Filamentous tends to occur more in tropical or humid climates but there is plenty found in more humid and hot parts of the U.S. including the mid-Atlantic where I practice and Georgia and South Carolina or Florida. Temperate climates tend to have more yeast keratitis. In our center, Candida keratitis is one of the most common causes of keratitis and that is true in other northern centers in America as well. The most common forms of filamentous are Fusarium, aspergilli and Curvularia. For yeast, Candida and Cryptococcus and sporidia. The burden of disease is substantial. 1 million new cases globally. 60 percent of eyes with fungal keratitis go blind despite treatment and 10 to 25 percent of eyes perforate or require surgical removal. This is a severe unmet need in terms of therapeutics and diagnostics. Fungal keratitis may be increasing globally. There is an increase in Thailand, Nepal and Ghana and India. Increased numbers in the U.S. and UK. In our own center we see increases in the proportion caused by fungi. There may be some relation to climate change. There is a correlation with increased systemic fungal infection and humidity levels are correlated to increased fungal keratitis incidence. There are trends to be noted. The burden of disease is 1 to 33 million contact lens wearers per year are at risk. And the infection rate is 17 to 21 cases per 100,000 contact lens wearers in the UK. What is notable is this risk is 10 to 15 times higher in the UK than the U.S. In the US, there is a 10-fold increased incidence since the original outbreaks were reported. In the UK, there is a 3-fold incidence from 2000 to 2010. There are more severe presentations and more treatment resistance and treatment failures. And contact lens wearing middle-income countries may also be a factor at play. In viral keratitis, risk factors are older age, immunocompromised, white race and atopy. HSV keratitis no major increases in 1976 to 2007. With HZ, there is a 23-fold increase over 14 years. Increased incidence among all age groups even the young. There is younger age of HZO presentation. And possible reductions in population-wide immunity but that theory is controversial and not proven. Here I think there is one theme I can try to impart it’s the importance of precision and use of terminology and clarity of thinking and use of words and diagnostic techniques. The term corneal ulcer is roughly synonymous with corneal infection. I would be more precise with the terminology. Corneal ulcers, if you think about the definition of ulcers, they are just removal of the surface layer of tissue whether it’s epithelium or epithelium and keratin yum. They can be caused by infections, inflammation, or even in the case of cornea, nerve damage leading to persistent epithelial defects. That can be a lot of things. Corneal ulcers do not necessarily indicate infection. Rather than saying corneal ulcer you may say infectious keratitis. But you can use more precise terminology to improve the communication with your colleagues. Infectious keratitis is often a term used in the absence of confirmatory microbiologic testing. You can say it’s presumed infection or you can say it’s micro biologically confirmed infectious keratitis. If it’s micro biologically confirmed you can say it’s confirmed as bacterial, fungal or amebic. If it’s viral, it’s more clinically diagnosed. So to put that into a flowchart for you, if someone comes in with what we might refer to as a corneal ulcer that we think is infectious, we can say presumed infectious keratitis. Based on the testing you do or don’t do, that can determine the terminology. If you test and have a confirmatory result there are a group of terms you can use for that diagnosis. If you do testing and have a negative or unclear result or if you don’t do the testing at all, all that can determine what terminology you use. If you have a confirmatory result, I encourage you to label the disease with specific words that say what testing was done and what the result was. You can say microscopy confirmed gram negative bacterial keratitis or filamentous fungal keratitis or culture confirmed bacterial keratitis. You can also say, if it’s culture negative, you can say culture negative presumed infectious keratitis. On the other hand, if it’s viral and clinically diagnosed you can say clinically diagnosed viral epithelial keratitis or herpetic disk form keratitis. More improved terminology improves your clarity of thought on what testing is done and whether you know what you’re treating and that sets the stage for more precise treatment. So then we go to diagnosis. The clinical history is important. We discussed risk factors and you should ask every patient about this when you first see them. Corneal scraping and first and second line testing. When do you perform microbiology testing. The short answer is almost always. We never regret performing testing. We sometimes regret not performing testing. But generally speaking, there’s no downside to performing scraping with microbiology testing there. Are select situations like a small presumed infectious keratitis or marginal keratitis that doesn’t need scraping. Phenotypically obvious case that doesn’t need anything beyond a clinical diagnosis. Or perhaps a very severely thinned case of keratitis where you’re concerned about causing a perforation. So you could defer testing in those situations but generally according to the preferred practice guidelines you should perform microbiologic testing when they are large, central, mid or deep stromal if there is any melt or chronic. If there is anything atypical beyond run-of-the-mill bacterial keratitis. Occurring after surgery, not improving or worsening despite treatment. Those are indications to perform microbiologic testing. There are a few best practices with scraping and they would improve the quality of the diagnostic yield and clarity of the results if we did them more. You should place a speculum in the eye. These patients are often in pain and photophobic. A very slow expansion of the speculum will reduce the pain. A sudden place. Of the speculum will cause the patient to flinch. A single numbing drop is often sufficient to provide anesthesia for the scaping. Seldom need to put in multiple rounds of numbing. The preservative within the drops can reduce the diagnostic yield by killing the micro organisms on the surfaces. You have to make sure the patient is comfortable. But usually a single numbing drop is sufficient. Arrange all of the slides and plates and vials beforehand so you’re not fussing to find what is what. Scraping can be done with different techniques. You can use the Kimura spatula. We often use the bard Parker blades because we’re not allowed to have active blades in our clinic. So we scrape with the flat of the blade and plate each blade on a different culture plate. You can use calcium alginate swabs. You must scrape deeply and thoroughly. If you’re concerned about perforation, that might be a different matter and then I recommend taking them to the OR and scraping under the microscope you have equipment on hand in case there is a perforation. For most presumed infectious keratitis you must scrape deeply and thoroughly to get a good specimen. The confirmatory testing will greatly enhance the likelihood of a successful therapeutic outcome. If there is limited specimen, you can be intentional about the order of testing. Perhaps a small but deep infiltrate where you’re concerned about fungal. In that case you can scrape deeply within the small area and put the first inoculum on a fungal culture plate. If you’re concerned there may not be enough live organism, you can put that small scraping on a plate for KOH or white staining with microscopy. You can think about where to prioritize your specimen placement if there is limited specimen size. Then the next question is what to test with the scraping. In general, these are the three most commonly performed tests with scraping. We’ll start with microscopy which in developing countries is a very valid and great way to detect bacterial and fungal keratitis and with the right stains Acanthamoeba keratitis. The turnaround time for microscopy is shorter than culture. The sensitivity is high particularly for fungal keratitis. And with the trained microbiologist or pathologist you can have almost species-level diagnosis. This is something that is on par with and in some cases superior to culture. Culture has certain advantages and allows more definitive speciation and do antimicrobial resistance testing. If you’re able to do a microscopy, if you have sufficient specimen, send for culture as well for those reasons. The positivity is highly variable. 40 to 70 percent for bacterial. And 50 to 80 percent for fungal. But getting a good specimen and inoculating it on the growth plate is the key to getting a positive culture. Culture can be done for Acanthamoeba keratitis using a non-nutrient agar with overlay and that has comparable results to PCR. PCR, single plex or multiplex. Single plex is you must prespecify which organism you’re trying to diagnosis. Multiplex is not widely commercially available but it’s going to be more available in the future. And you can have a panel of different common causes of corneal infection which are tested for in that one scraping specimen. So the sensitivity and specificity or PCR for bacterial, fungal, and viral and amebic keratitis are quite high. If that is available, I urge you to do so. Particularly for atypical organisms or confirming viral keratitis. AMR testing is underrated and I think it should be part of the microbiology protocol. The key limitation which must be admitted for AMR is minimum inhibitory concentration or MIC cut offs are based on serum levels. There is no break point. Higher serum MIC levels are associated with worse clinical outcomes. So the trial showed each 2-fold MIC increase is associated with 1 Snellen letter mean vision loss at 3 weeks or one line of vision lost per 32-fold MIC increase. MIC mediates the organism to outcome relationship and accounts for 13 percent effect. MIC predicts scar size. And higher MIC levels are associated with delayed healing and 4 percent risk of enucleation and 5 percent surgical intervention. Antimicrobial resistance is a factor in viral keratitis and you should test for it. You can have a close relationship with your lab to test for these and you commonly use antifungals. Metagenomic deep sequences is not widely available and it’s expensive. This compares the samples DNA and RNA against a genomic library of all possible organisms. It allows unbiased detection because you don’t have to specify which microbe you’re testing for with a hypothesis like single plex PCR. There is a high risk of contaminants. You can pick up microbial DNA or RNA from the environment. There is high cost associated with sequencing and you have very intensive computational expertise needed in the center to perform this. But you’ll see this coming up more and more in the future. Confocal microscopy is worth mentioning as a first-line diagnostic test for certain types of infectious keratitis. Filamentous fungi, the sensitivity is 85 percent and specificity is 81 percent and intermediate kappa is .88. People agree strongly when looking at the photo whether or not there is fungal hyphae. Acanthamoeba keratitis, if you see typical features on confocal, then that constitutes a microbiologic equivalent diagnosis. The sensitivity is approximately 88 percent with 98 percent specificity. And decent kappa statistic of .72 meaning people tend to generally agree as to whether or not the confocal image showed Acanthamoeba. There is a double walled cyst and trophozoite. It’s worth noting that confocal machines are not widely available at all centers. They are user dependent. Some are no longer being serviced. But the user dependency is a major limiting factor because it requires specialist expertise and experience. There is substantial subjectivity and interpretation of confocal images. So here is a case where we’ll start a poll. A patient is referred to tertiary care center for presumed infectious keratitis. No prior culture performed. They were treated before they saw you. In the last few weeks before they were seen by you, steroids. And this treatment overall was six weeks. This keratitis is at least 6 weeks old. The initial scraping was negative. Confocal is inconclusive. What are the next steps in repeat the scraping while continuing broad spectrum antimicrobials. Stop all the antimicrobials for a 24-hour period and repeat the scraping. Confocal microscopy. Corneal biopsy. Or given the severity should you perform a therapeutic keratoplasty. Great. Over half of you said what I would have done. You should stop the antimicrobials for 24 hours and then repeat the scraping. This is called an antimicrobial washout period. This is a protocol published for Massachusetts Eye and Ear and it tended to increase the diagnostic yield with a second scraping. This is very important to utilize especially working at a tertiary care center where you get a lot of referrals where the ocular surface has been carpet bombed with antimicrobials for a long period of time. All the bugs on the surface are dead but there is presumed infection going on. The washout period can reduce the ocular surface toxicity, and enable re-epithelialization and allow bugs to propagate again so you can more likely detect something on the next scraping. So in our center, we do initial corneal scraping and send for microscopy or culture or PCR. Usually microscopy and culture. In the case of viral that we want confirmed we send for PCR. If there is an atypical organism, we send for PCR as well. If that is negative, we consider, especially if they’re overtreated with antimicrobials we consider a washout period and repeat the scaping. If it’s a new infection, we just do the scraping then and there and send for all the same testing. If that is still negative, we consider procedural interventions like corneal biopsy which we can send for histopathology and microbiology testing. Both have different yields depending on the situation. With the biopsy, you want to use a small derm punch to partial thickness trephinate at the margin so you get some infiltrate and adjacent normal cornea. You take half and submit for biologic testing and the other half for histo. It’s a great opportunity to do more deep, careful, thorough scraping to improve the yield of the scraping. Just because you’re doing a biopsy doesn’t mean you shouldn’t do a scraping. The other thing to consider when talking about procedures and their use as diagnostic opportunities. If you’re performing a keratoplasty for severe and non-improving infectious keratitis, consider microbiology testing at that point. Before you go into the eye with trephination, take a small 27-gauge needle and aspirate aqueous or if there is endothelial plaque you can aspirate that as well and send that for testing. You can send part of the TPK button. When doing TPK, that trip to the OR is another opportunity to perform a good scraping and try to confirm the diagnosis. Next clinical assessment. The big elephant in the room which I wish we talked about more is corneal infections are difficult to measure. We need to acknowledge there are many different ways to measure and different people use different ways. I caught myself, I will admit, sometimes using different techniques on the same patient on different days and that is not the right thing to do. You need to be very precise and standardized within yourself as you measure the infection across different visits when seeing the same patient but also across your colleagues. When you document the exam findings for corneal infections think you’re communicating to your colleagues as well as to your future self and let that be the motivator for good documentation. Here are what good slit lamp photos look like. We have a photography protocol where every infection every visit gets standardized photography. Here we should have taken the photo without fluorescein dye. We usually take it with the photograph focused on the margins of the infiltrate. We don’t over or under expose the image. We try to standardize the angle of incidence of light. Next put in fluorescein dye. Within the next few minutes we blot the eye to wipe away excess and also have the eye looking primary gaze. When we take the cobalt blue light photo, we focus on the edges of the epithelial defect. You would be surprised how this simple change has substantially improved co-management and clinical outcomes at our center. Talking about colleagues around the world, that seems to be universal. I recommend instituting a universal photography protocol for all infectious keratitis at every visit. Here are examples of bad photos. Try to get a good photo with the eye in primary gaze and focused on the lesion of interest. The epithelial defect edges in blue light. Here is serial photos. Here is the same patient several weeks later. This patient had delayed wound healing because she was over treated with antimicrobials before seeing us. She had a kitchen sink keratitis so we did a washout period and confirmed this was a gram-positive bacterial keratitis and then allowed her to heal. But because of the weeks and weeks of eye drop toxicity, the stroma was irregular and uneven and difficult to epithelialize and took over a month to do it. These are the photos that we used to track the healing. We’ll go through specific examination findings and what to look for but mistakes to avoid for each. I will touch upon conjunctival injection. I would use diffuse illumination to standardize what is going on as far as the redness level. It can be a crude semi-qualitative biomarker for inflammation in the eye. And steroids can mask infection. But at the same time for certain cases like Acanthamoeba, we try to look how well the infection is being controlled. Here is a nephron scale used to grade infection. Most infectious keratitis is on the higher end of the scale. There is a scleritis scale which is relevant for different scleritis. This is something you can also try to standardize. Here, you instill phenylephrine and wait and then take the photographs to try to quantify how much redness there is in the sclera itself. The epithelial defect is the most precisely qualifiable biomarker correlating with infection resolution. The keyword is correlating with. We urge you to use consistent methods for the epithelial defect. Uterus photos at every visit. There are exceptions to that rule of correlation. When re-epithelialization lags, I mentioned the kitchen sink and neurotrophic keratopathy can occur and that can delay wound healing and be a risk factor for infection. There are exceptions when re-epithelialization precedes complete resolution. The infection goes deep in the stroma and the epithelium is intact. So that is not a biomarker for infection treatment. And viral stromal infection or obvious cases where the epithelium is intact. Important to use clearly specified methods. This sounds mundane but it’s commonly a problem. So that’s why we’re talking about. Here you can call it, measure length by perpendicular width. In is the convention used in the herpetic eye disease study. This is the most commonly used convention in the clinic. Don’t just say 8.2 by 5.3. Say 8.2 long by 5.3W or width. Height by width. Here I would not say width, I would use horizontal. 7.8 mm horizontal and 5.4 mm vertical and specify which is which and don’t get length and width confused with vertical and horizontal. Your colleagues are likely to measure it a little differently than you and that may confound your measurements of whether the patient is epithelializing or not. Stromal infiltrates. You can see the many different phenotypes of infectious infiltrates. This is the cardinal biomarker for assessing how bad is the infection and is it getting better or not. The slit lamp unfortunately only allow you to measure in 2D planar dimensions. Here you can assess a few things. What is the diameter and surface area of this. Use the same convention that you would for epithelial defect and be consistent and precise and clear in what mechanism you used to measure. You can talk about how dense the infiltrate is. You can talk about the margins, are they sharp with infections that are turning into scars or poorly demarcated which is true in early-stage infection and inflammation. You can also talk about the area of surrounding haze as you can see in this case. And also you can talk about are there feathery edges or multifocal lesions or satellite lesions. Are there ring, ring infiltrate. These are qualitative terms that can be quite useful. A picture is worth a thousand words and for beginners it’s important to become well versed and precise in the use of terminology. When it comes to co-managing with yourself or other colleagues, pictures are worth a thousand words and you should always get a photo. I want to talk more about depth of the infiltrate. I think this is not universally measured and every description of infiltrate must contain surface area in 2D and the depth. Depth is one of the strongest risk factors for perforation and therapeutic keratoplasty. It’s important to try to measure depth down to the nearest 10 percentage points. If you’re having troubling which people do, OCT is a great option. It can tell you if there is a full thickness infiltrate or dense anterior or mid stromal infiltrate. This is a powerful prognostic sign and at our center we’re using anterior segment OCT at baseline and once a week for tracking longitudinally the infiltrate in infectious keratitis. So here is another case that bridges the epithelial defect and the infiltrate part of the conversation. Here you can answer a question as well. This patient presents with slowly worsening eye irritation and slightly blurry vision. Denies eye pain and photophobia. Endorses sleeping once in contact lenses. What should you do next. Connect lens holiday. Should you treat with topical moxifloxacin. Here is their epithelium. Treat with topical moxifloxacin. Does this look like herpetic eye disease. Treat can topical steroids given the inflammation. Or do a diagnostic corneal scraping and if so, what specific organisms should you try to test for in your scraping? The — half the audience answered contact lens holiday which is correct. But not fully correct. About a quarter said topical moxifloxacin and 19 percent said corneal scraping. I’m glad that 17 percent of you are not giving the answer I would have. I think this is a very under appreciated clinical situation where you can stop a patient from going blind. This is Acanthamoeba at the epithelial stage of disease. This is a situation where you should learn to recognize the clinical hallmark of early epithelial stage Acanthamoeba keratitis and act rapidly to debride the epithelial surface and remove the amoeba before it spreads. And that epithelial debridement is therapeutic. You will cure the patient of Acanthamoeba in that eye. You should treat with PHM for several weeks but this is a case where an epithelial debridement will save the patient. It attaches to the epithelium and reaches Bowman’s membrane and deep into the stroma where it can cause much more eye pain and difficult to eradicate infections. That’s the stage you have peri– neuritis and scleritis. Generally, does not go inside of the eye itself. There are different phenotypes. If there is one thing I can impart on you today, it’s the importance of recognizing the epithelial stage and not treating like contact lens overwear or herpetic keratitis. The first two pictures at the top are isolated epithelial stage. Note the granular appearing epithelium with little or no stromal infiltration. Generally, the infiltrates are not impressive. Back to the photo. This looks like a pseudo-dendrite but it’s not all negative staining. Some is positive staining. Some is negative staining. This is not pooling, it’s straining. This roughed up epithelium and looped fractal-ized longing long arcs and loops with intervening spaces of normal cornea as well. This is a hallmark of epithelial stage amoeba keratitis. That classic admonition to look for pain applies to late stage, when it’s too late to prevent them from being in severe pain. They’re due for a long course of therapy. One other finding which is worth pointing out is this, it’s blurry in this photo. But this is the circumferential limb bites. This is early stage limbitis. If you have this one to 2 mm radius of almost salmon colored vessel that are hyperemic and not much redness elsewhere, that is limbitis and that occurs in approximately 10 to 20 percent including some early stages. One of the clues that can tilt you more toward an amebic diagnosis. This is epitheliitis with radial neuritis. This reached Bowman’s membrane and went into the nerve layer. You see the inflammatory nodules in a linear arrangement along the corneal nerve. This is pathognomonic for Acanthamoeba keratitis. This is anterior stromal disease in panel D. This reached Bohmanns. This patient is going to be on months-over PHD. Here is deep stromal keratitis. A ring infiltrate. And panel F, the ring infiltrate. These can be dense and classic or faint. You should keep an eye out for those. The rings are not pathognomonic for amoeba, they can be from herpetic disease or medication toxicity or other inflammatory diseases. If you have a patient come in with Acanthamoeba and a ring infiltrate, that was a delayed diagnosis and it’s going to be a long course of treatment. Earlier you diagnose, the betterment depth is a good prognosticator of eradication. Early diagnosis equals less depth. It’s useful to quantify thinning of the cornea. You can try to quantify to the nearest 10 percent. Try to specify tissue loss and you should say, when you say 10 percent, you mean that 10 percent of the stroma has melted or that 10 percent is remaining. Those can mean very different things and you should specify. OCT has a role to play. It can measure and track thinning. Our own work showed it can better detect perforations and anterior abnormalities. So here is an example of an old healed keratitis. Here you can see there are extensive corneal synechia. For completely opaque corneas, you can see if there is collapsed chamber, focus iris plugging of a perforation that is old and that can help determine surgery planning or candidacy. Hypopyon is an important biomolecular and measuring it is very important. It is associated with a risk of endophthalmitis. A risk factor for needing surgery. Height matters so you should measure it. There is a dose response relationship between height and adverse outcomes. It can be difficult to defect and measure through dense infiltrates that are overlying it. It’s possible to defect and quantify on OCT. Here is our own published study showing you can detect hypopyon far better on OCT and measure in millimeters using a caliber tool. Endothelial plaque. This is a finding that we’re learning more and more about with time. Because of its substantial prognostic significance and it’s under detected. It’s often missed on slit lamp exam. There is substantial improvement in diagnosis on OCT. It’s an independent predictor of surgical intervention and independent predictor of perforation can high odds ratios. The thing to know about endothelial plaque besides the association with needing surgery or having a perforation, in the case of fungal keratitis, it represents an intraocular invasion from the cornea inside of the eye. If it’s not treated, it’s an indicator to perform keratoplasty. There are some morphological subtypes of this plaque. This data is incoming but there are two subtypes at least. One is the plaque that is contiguous with the stromal infiltrate. There is invasion of the same fungal organism into the endothelial plaque. It’s an extension of the fungal organism into the anterior chamber of the eye. This is a plaque on the bottom that is distinct and has a linear margin between the infill rate and plaque. That is less likely to be fungal but we need more studies to validate. The other two morphologies of plaque are flat versus round plaque. The data is still incoming. But there is substantial phenotypic differences. With flat more common in bacterial infections and round are more common in fungal infection. There is a need for more studies for flat versus round or contiguous with the stroma or distinct. OCT can detect this better than slit lamp where we’re under detecting it now. The presence of any plaque has diagnostic significance and there is probably additional diagnostic significance to the specific subtypes of plaque that needs to be discovered. Brief word on AI for segmenting and detecting lesions. AI can segment corneal exam findings. Deep learning can differentiate infectious keratitis and non-infectious and dystrophy. However, also differentiate obvious active infection and healed scar. However, the problem with these techniques is the pipeline in the clinic does not exist yet. As of today, they are academic exercises and there needs to be a high throughput efficient workflow friendly pipeline to take imaging of the cornea and sending it for interpretation with the algorithm and getting a recommendation back as to the most likely diagnosis. That does not yet exist but there are teams working on it. Finally, we’ll talk about treatment. So here we’re going to talk more about individual subtypes of keratitis, bacterial, fungal, Acanthamoeba and viral. This is centered on medical treatment and we’ll finish with surgical treatment. Bacterial keratitis, antibiotics are still used. The mistake to try to avoid is indiscriminate prolonged use of broad spectrum antibiotics like vancomycin and tropomyosin. In general, you need to tailor the therapy once the microbe is identified and monitor for surface toxicity that can delay re-epithelialization. For mild to moderate, moxi are broad spectrum enough. The trial was followed by SCUT2. Showed there is no benefit to adding topical steroids. Nocardia keratitis did worse with adding topical steroids. In general, topical steroids are to be avoided in nocardia which is a bacteria that behaves more like a fungus in terms of appearance and response to steroids and indolence of infection. There were better outcomes in severe keratitis. Cases that are baseline counting fingers and severe ulcers and early starting of steroids within 2 to 3 days instead of waiting 4 or more days. There was a trial done that showed that cross linking is not associated with increased scar size. SCUT2 showed that adjuvant steroids were not associated with improved outcomes. Finally, fungal keratitis. So here there are two trials to note. The mycotic ultrasound treatment trial one showed for filamentous fungi, topical natamycin has superior outcomes. MUTT did not enroll yeast. But this is one of the few randomized trials where one drug is demonstrated to be better than another drug. In our center, this is evidence strong enough to recommend natamycin. MUT2 showed that adding adjuvant oral Viconazole did not confer a benefit but did increase the risks of systemic side effects. For severe keratitis, some centers treat with oral but the key to remember is, they cause systemic side effects including hepatotoxicity so it’s important to check liver function every two weeks. There is ongoing controversy regarding intra-stromal and intra-cameral antifungal injections. High quality evidence with randomized control trials is limited. There are reports showing that there are improvements in the fungal infection with reoccurring intra-stromal injections but some studies show increased association with melt or adverse outcomes and the data is not yet in. With Rose Bengal photodynamic activated therapy. There are encouraging reports this is a salvage therapy for particularly fungus but also amoeba. The REAGIR trial did not show a clear visual benefit. But when you look at subgroup analysis, there may be different effects by species. So there was better 6 month best corrected visual acuity in Fusarium and worse with aspergillus. It depends which species you’re treating. Acanthamoeba, early diagnosis is critical. Don’t confuse with herpetic keratitis. For a long time, the mainstay of treatment was biguanide therapy. And PHMB is roughly equivalent to chlorhexidine. And then that was shown in one trial. And then, one study. The other study that assessed PHMB slowed it was roughly equivalent to dual therapy. That evidence has changed in the last few years. PHMB0.8 is higher than.2. And.8 is tolerated well by the ocular surface without too much toxicity. The ODAK trial showed that dual therapy with the old standard of 0.2 percent is equivalent to mono-therapy with 0.08 percent. So 89 percent is non-inferior to the standard dual therapy. The cure rates exceeded 86 percent in both arms. We use 0.08 percent at the first line mono-therapy. Oral miltefosine has a limited role. But the thing to note is that you have to monitor for intense inflammatory reaction that sets in week two to three. Treated with topical steroids but often so severe it causes necrotizing scleritis and corneal perforation and sometimes loss of the eye. So you have to be cautious because of that severe inflammatory reaction which can be refractory to topical or oral steroids. There are reports of cross linking as a salvage therapy. Here is one where this seemed to resolve the case which was not resolving after dual antiamebic therapy. But again, a longer trial has not been done to show there is a therapeutic role for cross linking as salvage therapy. Let’s talk about medical treatments for herpetic keratitis. Here are the current treatments. Epithelial keratitis used to be treated with topical antivirals and in many countries it still is. In the U.S. we switched to oral antivirals. Epithelial keratitis is treated with only antiviral therapy. Stromal keratitis is treated with antivirals and topical corticosteroids. Iridocyclitis and endothelialitis is topical steroids. Stromal keratitis is PO acyclovir. This trial is now decades old. There are limitations in the evidence. In the U.S. we replaced topical trifluridine with acyclovir and valacyclovir. There is no evidence of the gancyclovir gel. And the HEDSs results were extrapolated to herpes zoster. And we have the zoster eye disease study that showed in a trial of 527 adults with keratitis or iritis in the past year, 1 gram a day for a year, those patients were given that for a year and followed after 18 months. They did not have benefit in terms of time to new or worsening keratitis but there was a 28 reduced hazard of new or recurring keratitis at 18 months. So secondary analysis seem to support considering one year of suppressive valacyclovir. We’ll finish up since we’re running over time, we will skip to therapeutic keratoplasty. That means I gave away the answer to this question. Here is a culture confirmed fusarium fungal keratitis with peripheral spread and infiltrate and enlarging endothelial plaque. What are the next steps in management? Good. I may have biased the results by giving away the answer. But the therapeutic keratoplasty is a valid option. I will admit there are some differences of opinion in how others would manage this condition. I don’t want to pretend that TPK is the only. You can add another antifungal. So here is a close up pick of that infiltrate. You will note a few things and I’m trying to convey the 3 dimensional — here the cornea is bulging forward with a thick and dense full thickness infiltrate. Here in the red is an endothelial plaque. You can’t tell the depth because it’s a photo. But are the is a dense endothelial plaque growing along the inside of the cornea. When you look at OCT it’s the phenotype that we talked about. A full thickness infiltrate but you can visualize better on OCT it’s contiguous with this dense endothelial plaque which is spreading over time despite antifungal therapy. So here we go. So we, because this infiltrate was not improving on antifungal therapy. It was full thickness where we thought that topical therapy was unlikely to get deep enough to treat it. We knew this was few say yum and based on studies in the past, we know that endothelial plaque is much more likely to contain live organism. Because the endothelial plaque was growing over time, because we felt that repeated infra-stromal injections would only briefly increase the drugs and not sustain the drug to kill the plaque where the bulk of the infection was. Because this infection was central and there was still an opportunity to cut the whole thing out, we performed a therapeutic keratoplasty in this patient. With TPK, the goals should be made clear. You’re restoring integrity or trying to resect the infection or both. Here the goal was more to resect the infection itself. There is often a dense plaque that is adherent to or even invading the lens. So here we made it a point to trephinate large enough to get the entire infiltrate. If it’s a large diameter trephination, that’s okay. The goal not to help the patient see better, it’s to cut out the infection and save the eye. The other thing to note is we sent the specimen for histopathology and resent the aspirate for microbiology testing. And then the histopathology showed septate fungal forms with hyphal vesicles suggesting fusarium involving the corneal stroma and retro corneal inflammation. What’s also interesting is this patient has bilateral fungal keratitis. He had Fusarium keratitis in the other eye. I will say there is a small plaque but in general we felt this was treatable with eye drops alone. So we had him take intensive antifungals in the form of natamycin. And postop, we continued antifungal medications on the right and the left eye but for different reasons. Here it was to make sure there was no infection coming back and here it was to treat the infection that was there. We did with the right eye we covered with antibiotics until it re-epithelialized. We gave oral voriconazole but stopped it: We weren’t using topical steroids so we gave him cyclosporin which is shown to have weak antifungal properties and may reduce the severity of fungal keratitis. But more studies are needed. Here is what it looked like at month one. Clear of infection. We stopped the oral voriconazole. And kept the cyclosporin going. We stopped all of the antifungals and started on steroids for the right eye. He is seeing close to 20/20 range in the right eye. The left eye was medically treated but the left eye is 20/40 with a contact lens. Even though this was the more severe case buzz we did early TPK and cured the infection, his more severe fungal keratitis was the better seeing eye in the end compared to the eye with the corneal scar. The last thing I will say is the goals of TPK must be clear. What is the goal? To restore tectonic integrity. That is achievable in almost 1 hushing percent of eyes. Is it to remove the infection. That is achievable in 70 to 90 percent of eyes. The goal can also in some cases be to possibly maintain graft clarity and improve vision but you must counsel the patient that is not the top priority. The top priority is to save the eye and remove the infection. Otherwise you can save the eye but have the patient be disappointed they’re not seeing better. The graft failure does occur in 45 to 65 percent because you’re operating on an inflamed eye. You can always do an OPK later. Risk factors are notable for infection recurrence after TPK. You failed to achieve the goal. Hypopyon, perforation, limbal extension, lens infection. Larger size and delayed surgery. All of these are evidence that you waited too long. You looked at the bio Markers like thinning, thinning precedes perforation. Infiltrate towards the limbus. Endothelial plaque precedes lens infection. Infiltrate size is tracked over time. If you delay surgery on top of these risk factors that’s such a bad infection you can’t resect it. I advocate for measuring the anatomic findings. Cal it what it is. If it’s getting worse or not improving after at least two or three weeks of sustained antifungal therapy, then declare failure of medical management and make the difficult decision to do a TPK. Perforation is not the only reason for TPK. The others are important as well. Do not delay until the infection spreads to the limbus. Conjunctival flaps for refractory infectious keratitis. It can be a salvage or stabilize the situation before future optical PKP. One of the larger case studies was in the last year or so where they did 43 percent complete flaps and 57 percent partial flaps and partial had higher success. In is a greater vascular supply to the area. There is another earlier series in 2020 where 71 percent were for active infectious keratitis and 97 percent is ocular surface stabilization. This is a salvage therapy that can be used to further improve infection control and stabilize the situation and live to fight another day you can do optical rehabilitation later. In conclusion, microbial keratitis diagnose and management can be challenging due to ambiguities in diagnosis, clinical assessment and treatment. Achieving clarify on all of these three can help improve clinical outcomes. I hope you use this lecture to think more clearly about diagnosis, assessment and the goals of treatment and the best approaches to try to achieve those goals. Some things I discussed were elementary but that is why they are so important. Because they are fundamental to having clarity of thought as you approach care of the complex patients and ambiguous situations that are high stakes. I hope that was helpful, thank you for watching. So now what I’m going to do, there are 32 questions. I will go in chronological order. And answer all of them. We have 20 minutes left. And Orbis folks let me know if I’m taking too long. How do you perform evisceration. That is beyond the scope of this lecture. I will table that for you. I refer you to the Orbis lectures on this and there are papers and techniques available online and videos on You Tube. How about a glue patch in impending perforation. Absolutely. We didn’t have time to talk about surgical interventions but yes, a patch graft as well as glueing and amniotic membrane and combinations can be an excellent resource to use when you don’t have fresh corneal tissue available. I would highlight the utility of gamma irradiated sterilized corneal tissue which can be off the shelf and kept in a freezer or dry and used for patch grafting. Depending on the size of the perforation, if it’s small, glueing, or glueing and AMT or corneal patch graft. If it’s larger, you may have to use live tissue. Related to the medical therapy, what is the time for PKP. When should you call it medical treatment failure and how long after you treat the infection you do an optical keratoplasty. When you look at how fast the biomarkers change for bacterial keratitis, it’s usually one to two weeks. One to two weeks you get a sense of is this patient getting better or worse. For fungal it can be two to four weeks and Acanthamoeba it’s on the order of a month. But in general, you should be patient enough to wait one to two weeks for bacterial before calling it treatment failure. You should be patient enough to wait two to four weeks for fungal unless it’s very severe. But generally, if it’s ambiguous, wait two to four weeks before you have therapy failure with medical management. For Acanthamoeba, wait a bit longer. You can have this prolonged concentration for many weeks and then the infiltrate starts to reseed and the inflammation goes away and the patient gets better. Optical keratoplasty, we wait 6 months after resolution of the infection before we do optical keratoplasty. If there is neovascularization, we wait longer. How often to use prophylactic acyclovir. I think long-term prophylaxis for patients with herpetic keratitis. The study for oral valacyclovir they treated only for a year. That doesn’t mean you only treat for a year. For herpetic keratitis, we generally treat forever with suppressive acyclovir if we’re concerned about recurrence over their life span. I answered the question how long must the treatment Acanthamoeba keratitis. You need to counsel them on diagnosis, they have going to have several months of therapy. We tell them at least 6 months of intensive anti-amebic therapy. And after which maybe up to 6 more months of lower dose antiamoebic therapy to eradicate any cysts or trophozoites lingering. That is up to a year of therapy. Six months intense and frequent eye drops and six months of QID eye drops to eradicate the amoeba. Can we do conjunctival flap. I answered that question. Yes, absolutely. If it’s a macro level perforation you may have to send them to a place with enough tissue. You can also incorporate Tenon fascia. From which tissue is patch graft taken? Recently deceased donor corneal tissue that is not viable from the eye bank. Sterilized or irradiated tissue as well. Amnion and hypopyon, yes, as long as you can measure it over time. The best management for — keratitis. A great question. Taper the steroids very slowly, probably slower than you think you need to, to avoid recurrence. Be patient. It can take several months. Slow taper to prevent recurrence. How do you take a sample? Scrape. Can the secretion be a sample for the diagnosis? I’m not sure what that question means but generally scrape deeply. Use a speculum. Only use one numbing drop, more if the patient needs it but be sparing. A conjunctival flap. There is going to be more neovascularization in the cornea after the conjunctival flap. In general, we do the conjunctival flap and take down and allow the patient to re-epithelialize for a few weeks to months and consider the optical PKP. Hypothetically, the answer is yes, there is increased neovascularization. There is not great data to my knowledge to answer that question. Any early Acanthamoeba keratitis. Why does the patient not experience pain. Is it a common feature? Yes. The one thing I was trying to impart, that early stage epithelial amoeba, they are not in pain. They don’t have radial keratitis rat toe neuritis. No ring infiltrate and they don’t meet the textbook definition of keratitis teaching and that is why they’re under diagnosed. In early AK, you do not need to have a lot of pain. You can have this vague irritation in the eye that is kind of getting slowly worse over time. Maybe worse after starting topical steroids but that vague indolent pain and granular epitheliopathy. In the contact lens wearer that is misdiagnosed as contact lens overwear is the typical phenotype and history. Is there a role of cyclosporin in the management of keratitis? Potentially. We incorporate it into the fungal keratitis protocols as a steroid sparing therapy. We can’t treat with steroids because it can make the fungal keratitis worse. As I mentioned, they can somewhat reduce inflammation. And they also showed to have weak antifungal properties in veto. We need longer term studies with greater sample size of patients and a randomized control trial to figure it out. Comment on the role of chlorhexidine. I think in general there is one RCT I believe evaluating chlorhexidine. Natamycin was superior. If you’re able to get natamycin for filamentous fungal keratitis, the data is unambiguous. It’s one of the few situations where you have multiple RCTs. Showing that natamycin is better than voriconazole and there is one trial showing that it’s superior to other antiseptic medications and natamycin should be the first line of therapy for fungal keratitis. How to taper down steroids? What is the use of steroids to decrease the scar? We really are hopeful that SCUT1 and 2 would show reduction in scar with use of topical steroids. Thus far the papers have not shown a qualifiable difference. All I will say is when we start steroids in our practice, and there are some differences in how people do this. We start gingerly. We don’t start them right away with bacterial keratitis. We never start in fungal or suspected fungal. We start after a few days of antibiotics and non-nocardial bacterial keratitis that is confirmed. Start BID or TID to make sure they don’t get worse and then up to four times a day and taper down to BID and then stop. For Acanthamoeba steroid use is controversial. There is a trial on going that is trying to assess whether adding adjuvant topical steroids after 4 weeks might reduce inflammation and improve visual outcomes. At the eye hospital, they feel comfortable starting topical steroids after a few weeks. We’re enrolling to randomize patients one way or the other. I do feel comfortable at our center doing THMB for 4 weeks and considering starting topical corticosteroids. There is not great evidence to answer that question. What frequency can intra-stromal injections be given? Very divergent opinions. Some people think they’re not helpful and harmful and other people are big fans and take patients to the OR for them. I think in general, without disagreeing too much with these strong opinions, I will say one of the biggest failure modes in keratitis is doing the same thing over and other again and expecting different results. That’s not helpful for the patient. If you do repeated injections over and over and don’t see meaningful improvement in the biomarkers, it’s probably not going to work another time. You can go to the OR another time but if you have repeated trips to the OR or repeated procedures with intra-stromal injection and it’s not working, then don’t keep doing the same thing over and over again and expect different results. Which is better, oral or topical antifungal? Topical. I’m less than a third done. When to start prof laxative treatment in a setting of fungal keratitis. This is where culture and microscopy are important. We generally don’t start prophylactic antifungals because we have data showing that 85 percent are bacterial. 10 percent are fungal. 5 percent are amoeba. So based on the data at our center, we don’t see a role for prophylactic antifungal unless there are extremely strong clinical suspicion but I prefer testing before doing antifungal treatment. You will find different opinions. In south India and north India, there is a close to 50% or greater than 50% of keratitis that is fungal. In that case, it’s more justifiable to start prophylactic antifungal treatment and anti-bacterial treatment. It depends on the local epidemiology. What kind of steroid do you use for steroid ulcer? Medium potency. Low is too low. And I don’t like high. Durezol seems to be associated with precipitating corneal melt and the data from Durezol is not improve as far as changing outcomes. If we’re going to do a treatment with steroids, we’ll use prednisolone acetate which is roughly equivalent to prednisolone phosphate that we used in the SCUT one study. I don’t have strong opinions. I’m just telling you what we do. The symptoms and signs of — micro sporidia keratitis including treatment. At the epithelial stage, there are stuck on lesions that are spaced out. In deeper infections you can have multifocal infiltrates with more immune reaction appearance. That can be very faded edges. There are very good review articles discussing particular hallmarks of micro sporidia keratitis like intervening clear space and inflammatory lesions. I would encourage you to look as those. It’s a case where we didn’t have a lot of time to get into but the phenotype is worth knowing. How to taper down antibiotics when used as prophylaxis. That is an important question. In the early stage, you should treat intensively. When you see the patient and diagnose them with their presumed infectious keratitis and giving them medications for the first time or the medications, have them do a loading dose every five minutes, put in drops to get the concentration up in the cornea before they go home. And patient compliance and ability to adhere to the recommendation not withstanding we have them do the eye drops every hour for severe keratitis and maybe every two hours. The question is how to treat overnight. We try to have them take it every one to two hours overnight in the first few days. We need to be humane. They’re suffering and robbing them of sleep is more inhumane. We have to keep sleep deprivation in mind. Over the next coming days unless the keratitis is severe, we try to have them taper to taking two or three instillations overnight. As it improves we say take once or twice every night so they get chunks of three or four hours of sleep. There is not a standardized protocol for tapering the antimicrobial therapy. I would say there is a few considerations. The biomarkers are the epithelial defect closing. Is the stromal infiltrate becoming less opaque and more purulent and more gray and more transparent and faint is a rough texture meaning it’s turning into a scar. Are the edges not as fuzzy. Those are things we use to try to taper down the antibiotic. One thing I will say is, antibiotics in the last phase are not tapered down to twice a day or once a day. They’re tapered down to QID and when you feel the infection is gone, you stop it. That’s it. You don’t have to do QID, TID, BID, once a day like with steroids. You start with frequent use of eye drops and go down gradually in a way that I’m having trouble summarizing for you because it’s patient dependent based on the biomarkers and the last phase you do QID for a few weeks and then you just stop all together. If there is a persistent epithelial defect we keep it going for 4 times a day until it closes to prevent bacterial superinfection. An important thing to note, important failure to avoid is excessive fortified antimicrobial use beyond the necessary period. If you’re seeing those biomarker changes that I described that are indicative of infection resolution and you’re keeping on every hour or every two hours fortified antibiotics or antifungals you end up with kitchen sink keratitis with mushed stroma and trouble resolving the epithelial defect. We encourage tapering. I was not able to answer all of the questions. I will try to look through them and you can reach out to me with question and I’m happy to answer them. Thank you so much. One final question. Okay. Let me choose one that is particularly — they’re all really great questions. I will answer one because — how wise is it to wait and just do a saline wash for a few days. If you suspect infection, I would not try to wash it out and hope the host immune response resolves its. I would treat with antimicrobials empirically at least. That’s the last question I’m able to answer today. But again, thank you for your time.