Update on methods of assessing microbiologic success or failure in patients with otic disease.
Classification criteria for success and failure
In clinical studies conducted up to 10 years ago, simple improvement in signs and symptoms was often sufficient to qualify a regimen as a clinical success. In today's environment, however, mere improvement represents a treatment failure; for a regimen to be classified as a clinical success, it is necessary that all signs and symptoms be completely resolved.
Similarly, microbiologic success 10 years ago was declared when there was only a reduction in the amount of a particular pathogen rather than an eradication. In other words, the patient was almost cured, and this was sufficient to qualify as a success. But again, in today's environment, the presence of any amount of pathogen represents a treatment failure. Of course, treatment is an obvious microbiologic failure if the pretherapy pathogen persists, but it is also considered to be a failure if the patient still has signs or symptoms because a different pathogen is present. In such a case, the presence of the new pathogen does not represent persistence; rather, it represents a superinfection or reinfection, depending on when the new pathogen appeared.
Eradication of the offending pathogen can be either documented or presumed. Eradication is documented by taking a microbiologic sample of the cured tissue. When the act of obtaining such a sample is not desirable or feasible, and if clinical signs and symptoms are absent, eradication can be presumed. For example, in the evaluation of patients who have been treated for acute otitis media with otorrhea through tympanostomy tubes (AOMT), there is no justification for obtaining a post-therapy specimen from the middle ear cavity when no otorrhea is present. Essentially, the otolaryngologist would be attempting to sample something that is not there. Therefore, in the absence of clinical signs or symptoms, microbiologic cure can be reliably presumed.
Timing of outcome assessment
Another function that has been revised is the timing of outcomes assessments. Ten years ago, clinical and microbiologic efficacy was assessed as soon as the patient completed the course of therapy. Today we assess outcomes at the test-of-cure visit, which usually takes place 3 to 7 days after the completion of therapy. This additional time allows the otolaryngologist to detect any re-emergence of pathogens that were not eradicated. In clinical trials, of a course, a physician can declare a treatment failure at any a time if the patient is not responding.
Defining a pathogen according to growth rate
Another new element in assessing microbiologic outcomes in otic studies concerns the definition of what constitutes a pathogen. With some exceptions, Dohar et al contend that a pathogen is defined as an organism that is recovered with at least a 2+ growth index. (1) Two notable exceptions pertain to Hemophilus influenzae and Streptococcus pneumoniae; any presence of these organisms--even if it is only with a 1+ growth index--is sufficient to qualify as pathogenic. Conversely, Dohar et al list approximately 20 bacterial species that are never considered pathogenic, even when they are recovered with a 3+ or 4+ growth index. But the major organisms seen in AOMT are considered pathogenic if they are recovered with a 2+ growth index. The amount of growth is usually measured 1 or 2 days after the specimen is placed on the plate in the laboratory. In some clinical studies, plates have been incubated for as long as 5 days in an attempt to recover everything that might be present.
Microbiologic recovery vs DNA detection
In some ophthalmology studies and in an AOMT study that has just gotten under way, a new protocol was implemented that involves not just the recovery of bacteria, but its detection by polymerase chain reaction (PCR) methods. With DNA-based technology, bacteria can be detected regardless of how little is present. For detection purposes, quantity is irrelevant; it is the presence that matters.
Bacterial detection has shed quite a bit of light on organisms that have not been recoverable even though they are perfectly respectable aerobes or anaerobes. Of course, the fact that an organism can be recovered does not mean it is pathogenic. This has some bearing in AOMT, because the pathogens that are recovered are those that grow very easily. This does not mean that these organisms are the only ones present, but AOMT has typically been defined by the presence of organisms that are the easiest to isolate.
One interesting discovery that we made by using DNA-based detection was that Alloiococcus otitidis is extremely common in healthy ear canals; only Turicella otitidis and Coryneform bacteria are more common. We have found A otitidis in approximately 40% of healthy ear canals. We also learned that A otitidis is not usually recoverable unless it is incubated for 5 days, and even then we find only a "haze" of growth. It grows very slowly. Our findings were confirmed by Frank et al, who studied healthy ears by PCR. (2) They found that A otitidis and T otitidis were the most common organisms.
On the other hand, A otitidis has been classified as pathogenic by Bosley et al in Finland. (3) They discovered A otitidis by DNA-based technology in patients with middle ear infections. They tried repeatedly to recover A otitidis from middle ear effusions by standard microbiologic means but were unable to do so. Nevertheless, they argued that it is indeed pathogenic, based on their DNA-detection studies. Therefore, the criteria used to classify an organism as pathogenic or not varies from study to study and often depends on the investigators' ability to recover it and on how well it grows.
Some scientists argue that an organism is not pathogenic if it does not grow well. However, in our studies of conjunctivitis in the United States, we missed probably 20% of cases of H influenzae and S pneumoniae infection simply because we were unable to recover them. And these are two well-known pathogens. Whether or not organisms are recovered and identified as pathogens is a function of microbiologic technique.
The reasons for our failure to recover H influenzae and S pneumoniae are not clear cut. It might have been that these organisms were not present in abundance. Also, they are both rather fastidious and sometimes difficult to recover. DNA-based detection, on the other hand, is much more sensitive than standard microbiologic techniques. It can detect as few as five cells on a specimen swab. For example, Gemella hemolysans is a strepoccocns-like organism found in patients with conjunctivitis. Although it has never been recovered by standard microbiologic techniques, DNA technology detected 30 cases of it.
Similarly, we have collected 12,000 bacterial isolates during the past 4 years, and standard microbiology has recovered only two cases of Corynebacterium tuberculostearicum--one from an infected ear and one from an infected eye. But when we used DNA technology on the same specimens, we detected more than 30 cases. For whatever reason--perhaps this organism has some unusual growth requirement--we generally have not been able to recover C tuberculosteuricum, even though it is certainly present.
Finally, Rhizobium radiobacter is another organism we have recovered only occasionally from swabs but have detected quite often by DNA-based technology. It is worth noting that DNA analysis rarely detected G hemolysans, C tuberculostearicum, and R radiobacter at the test-of-cure visit. They were detected primarily at the pretreatment visit, which tells us that the antibiotic is eradicating them.
Simonsiella spp are gram-negative organisms that resemble Neisseria spp, and they are found in many conjunctivitis patients. They were first described 100 years ago, but they are rarely reported today because they are rarely recovered. We recover Propionibacterium acnes in patients with ear and eye infections only occasionally; DNA-based technology detects it at least 50 times more often.
As suggested earlier, one problem with DNA-based detection is that it cannot identify the relative quantity of one species versus another in the same sample. The technology can identify which organisms are present, but it cannot distinguish the proportions of those organisms. Nevertheless, the ability to detect multiple bacteria in a single specimen is useful. We now routinely look for all Chlamydia spp in a specimen. In a recent conjunctivitis study, we recovered Chlamydia felis from two patients. The literature contains only one case of conjunctivitis that was caused by C felis, but with DNA-based technology, we have already identified two more.
Pathogens implicated in treatment failures
Data on specific microbiologic failure rates have been obtained in two phase III studies of ciprofloxacin/dexamethasone otic solution; one study was conducted in patients with acute otitis externa (AOE) and one in patients with AOMT (see page 2). All outcomes were measured by standard microbiologic techniques, and all failures were confirmed by DNA fingerprinting. Treatment successes were judged conservatively. For example, a treatment was classified as a failure if any sign of residual infection remained in a patient who did not undergo follow-up testing to determine whether the residual bacteria represented the original pathogen or a new organism. Such a circumstance was considered to represent a presumed failure. The assumption was that the remaining bacteria represented the original pathogen. This protocol was somewhat misleading because in most cases when residual infection was identified, the bacteria were found to be new. All evaluations were conducted at the test-of-cure visit.
In AOE, treatment eradicated 38 of 42 cases (90.5%) of Staphylococcus aureus infection; of the four treatment failures (9.5%), one was documented and three were presumed. In cases of Pseudomonas aeruginosa infection, treatment eradicated the pathogen in 223 of 232 cases, (96.1%); of the nine failures (3.9%), five were documented and four were presumed.
In the AOMT study, all failures were documented; none was presumed. Ciprofloxacin/dexamethasone treatment failed in two of 26 (7.7%) H influenzae infections, in six of 44 (13.6%) S pneumoniae infections, in three of 53 (5.7%) S aureus infections, and in one of 48 (2.1%) P aeruginosa infections. As was the case with the AOE study, none of the pathogens was resistant to ciprofloxacin prior to therapy, and none became resistant to ciprofloxacin during therapy.
The question is often asked as to whether the increase in antibiotic-resistant isolates has resulted in a corresponding increase in treatment failures. Thus far, no clear-cut evidence has been found that resistant strains lead to more clinical failures. Studies of the lower respiratory tract have yielded conflicting findings. Pichichero et al studied a group of children with recurrent and persistent otitis media and suggested that resistance is not a factor in treatment failures? Some of these chronic infections were caused by resistant organisms, but many were caused by susceptible organisms.
In a recent study in India (unpublished), microbiologic specimens were collected from the eyes of healthy subjects in the community and the eyes of ophthalmologists themselves. The ophthalmologists, who had access to free antibiotics, had many more resistant organisms in their healthy flora than did those from poorer areas of the country where there was no access to antibiotics. However, the ophthalmologists did not have a greater incidence of infection. Their greater incidence of resistant strains did not result in a higher rate of disease.
Finally, Waturangi et al studied a group of antibiotic-naive lizards from remote areas of Indonesia and recovered 23 strains of Escherichia coli, 70% of which were resistant to tetracycline. (5) In fact, the lizards had the same resistance genes as do humans.
Dr. Poole: With respect to DNA-based technology, the benefit of PCR lies in its ability to detect pathogens in places where they are not usually present. PCR would presumably yield a positive for an organism even if there were only one dead organism in the specimen. Therefore, I do not see how it is worthwhile to look for these organisms in mucosal areas that are in proximity to the skin or the oral cavity, because we would expect them to be there. If you examined a series of specimens taken from tears, for example, I suspect that they would all be PCR-positive, at some point or another, for every organism that's present in the air.
Second, we typically diagnose otitis and sinusitis on the basis of the presence or absence of organisms. But that's probably a naive assumption, because healthy people can also harbor pathogens in low numbers. Therefore, we should probably rethink how we define disease in the middle ear, in the nose, in the eye, etc. Perhaps we should discard our black-and-white, all-or-nothing way of thinking and consider the relative contributions to the inflammatory process of the various amounts of bacteria, viruses, and intrinsic inflammatory mediators. At any rate, the landscape is constantly changing, and PCR is helping as understand that a little better.
Dr. Manning: When you attempt to detect organisms by PCR, is it true that you can't use a cotton swab because dead bacterial DNA might be present on the swab?
Dr. Stroman: No, that's not true. Certainly, swabs do become contaminated, but we control for it. In fact, there is one particular bacterium--Delftia acidovorans--that we find quite often on swabs. And swabs supplied by certain companies are contaminated by waterborne organisms that are introduced during manufacturing.
Dr. Manning: At our children's hospital in Seattle, about 50% of neck abscesses grow nothing on routine culture. Our first step is to perform PCR for Bartonella henselae--that is, cat scratch--if the standard culture is sterile. If the B henselae PCR is negative, our pathologist will perform further PCR testing to look for other organisms. For example, we recently diagnosed a case of Legionella micdadei infection, which had never been known to cause an abscess. But we supported this diagnosis with acute and convalescent titers on serology.
Prof. Hawke: I think DNA detection is an exciting development, However, I wonder how accurate it will be because it identifies both dead and live bacteria that might not necessarily be the cause of the infection.
Dr. Stroman: Absolutely. However, I should mention that even though DNA-based detection does not distinguish between dead and live organisms, the DNA in dead organisms degrades very rapidly in the liquid environment of an incubator and in a body with an elevated temperature. So for the most part, we are detecting live organisms.
Dr. Haynes: There has been some talk about the difference between obtaining a culture sample from a tube or perforation and obtaining it by swabbing the ear canal. We know there are different types of bacteria at these different sites. Has a study ever been conducted to determine just how much of a difference?
Dr. Stroman: No. There is too much variation among physicians with respect to bow they collect culture specimens to allow us to arrive at valid conclusions. If we were to perform such a study, the results would not be applicable to all physicians.
Prof. Hawke: With respect to antibiotic treatment failures, I wonder if we could go back to the initial visit and determine whether there was anything that could be identifiable as a contributing factor, such as the presence or absence of a wick, the severity of disease, etc. Are there any clues that we might pick up by reviewing the findings on the initial evaluation?
Dr. Stroman: I don't know. We know that some parameters are not risk factors, but we don't know which are.
(1.) Dohar JE. Garner ET, Nielsen RW, et al. Topical ofloxacin treatment of otorrhea in children with tympanostomy tubes. Arch Otolaryngol Head Neck Surg 1999;125:537-45.
(2.) Frank DN, Spiegelman GB, Davis W, et al. Culture-independent molecular analysis of microbial constituents of the healthy human outer ear. J Clin Miarobiol 2003;41:295-303.
(3.) Bosley GS, Whitney SA, Pruckler JM, et al. Characterization of ear fluid isolates of Alloiococcus otitidis from patients with recurrent otitis media. J Clin Microbiol 1995;33:2876 80.
(4.) Pichichero ME, Reiner AM, Brook I, et al. Controversies in the medical management of persistent and recurrent acute otitis media. Recommendations of a clinical advisory committee. Ann Otol Rhinol Laryngol Suppl 2000; 183:1-12.
(5.) Waturangi DE, Suwanto A, Schwarz S, Erdelen W. Identification of class-1 integrons-associated gene cassettes in Escherichia coli isolated from Varanus spp. in Indonesia. J Antimicrob Chemother 2003;51:175-7.
|Printer friendly Cite/link Email Feedback|
|Publication:||Ear, Nose and Throat Journal|
|Date:||Aug 1, 2003|
|Previous Article:||Topical treatment of chronic suppurative otitis media in Aboriginal children.|
|Next Article:||Basics of biofilm in clinical otolaryngology.|