TO THE EDITOR:
We are compelled to respond to the criticism of Welling et al. (1) about 99mTc-ciprofloxacin (Infecton; Draximage Inc., Kirkland, Quebec, Canada) imaging because of their continued use of references that, as detailed analysis reveals, do not support their statements. We pointed these out in a previous letter to the editor of this journal (2) and in letters to the editor of the European Journal of Nuclear Medicine (3,4). No useful benefit will be served by repeating that information again in detail here; suffice it to say that, first, the amount of ciprofloxacin in Infecton is so tiny (2 mg, which is 1/200 of a single therapeutic intravenous dose of ciprofloxacin) that the risk of emergence of resistant bacteria is unlikely to be a clinical problem. Indeed, such emergence has not been observed in experiments in vitro or in patients who have undergone Infecton imaging, although we are monitoring the situation carefully. Second, mammalian topoisomerase has 100–1,000 times less affinity for binding to ciprofloxacin than does bacterial DNA gyrase (the basis of selective toxicity), and the binding is readily reversible. By comparison, Infecton remains bound to bacteria at sites of infection, giving a high target-to-background ratio, which is the basis of bacteria-specific imaging with Infecton. Third, our clinical experience with Infecton imaging has now extended to well over 1,300 patients and supports our previously published data showing that Infecton can detect a wide variety of bacterial infections with high specificity and distinguish reliably between infective and noninfective inflammation, if 24-h images are included, as previously reported (2–4).
We have a few other points to make regarding the article of Welling et al. (1). The authors used multidrug-resistant Staphylococcus aureus and Klebsiella pneumoniae in their experiments to induce infection and to study the uptake of various agents, including Infecton. Were these organisms also resistant to ciprofloxacin, and if so, what was the mechanism of resistance? If an organism does not take up ciprofloxacin, bacteria-specific imaging with Infecton will not occur.
Similarly, it is well known that ciprofloxacin is not taken up by and is inactive against fungal pathogens, including Candida albicans. This fact explains the experimental results of Welling et al. (1) for Infecton imaging of infection caused by this organism. In fact, Candida albicans acts as a good negative control for Infecton.
Some antibiotics, notably azithromycin and clindamycin, have been shown to affect the immune system, and such an effect may be beneficial by augmenting clearance of bacterial pathogens from the host. However, the clinical significance of this effect remains unknown. In addition, ciprofloxacin is unlikely to significantly affect the immune system at the tracer amount present in Infecton. The predicted peak serum concentration of ciprofloxacin after intravenous administration of Infecton is only 0.01–0.02 mg/L. The concentration of ciprofloxacin used by Riesbeck et al. (5) was far greater, ranging from 5 to 80 mg/L (i.e., 250–4,000 times greater), with the effect on interleukin-2 production being more pronounced at the higher concentrations of 20 and 80 mg/L. In this regard, it is relevant to note that the peak serum concentration after 1 h with a standard therapeutic intravenous dose of 400 mg is approximately 2–4 mg/L. One always has to be careful about extrapolating in vitro and, indeed, animal experimental data to humans. These experiments are useful as a guide to what may or may not occur in humans but by themselves cannot totally be relied on. Because of the very nature and function of the molecules, it is more likely that defensins such as antimicrobial peptides, even when designed to contain only the “antimicrobial” domains, will have a more pronounced and significant effect on the immune system than does ciprofloxacin or other antibiotics.
One must question whether Welling et al. (1) are preparing 99mTc-ciprofloxacin. They state that their preparation needs a further purification step because of 20%–40% free 99mTc. This is never seen with our 99mTc labeling of ciprofloxacin and suggests colloid formation, which may occur if the instructions are not followed precisely, such as when the reagents are combined in the wrong order.
The captions of Figures 1 and 4 indicate that the studies were obtained at 1 h. The authors appear to still not understand that any small molecule diffuses into inflammation and out again in proportion to its size and the rate of the fall in blood level. Our studies on humans show that 4- and 24-h images are required to distinguish active inflammation, for example, in joints, from septic arthritis. Larikka et al. (6) showed an increase in specificity from 41% to 95% when a 24-h image was combined with a 4-h image to distinguish an infected hip prosthesis from noninfected inflammation.
In conclusion, successful imaging of bacterial infection stems from a combination of direct binding of 99mTc-ciprofloxacin to dividing bacteria and the dynamic specificity that occurs when uptake caused by bacterial infection persists while that caused by nonspecific inflammation fades. It appears difficult to mimic the reality of the situation in humans from animal models.
We welcome fair and scientifically valid criticism of our novel imaging technique. As a matter of good scientific principle, data from studies should be critically analyzed before they are used to support or refute anything. Sadly, Welling et al. (1) appear to be repeatedly neglecting this step when commenting on Infecton imaging.
REFERENCES
REPLY:
In response to the letter of Das et al., we must first emphasize that the main point of our study was completely missed by these authors. Our study concerned the development and testing of a new class of radiolabeled compounds, namely antimicrobial peptides, for the detection of bacterial and Candida albicans infections (1). Our findings for 99mTc-ciprofloxacin (Infecton) were of importance but in the context of our article were used only as a reference.
Nevertheless, it was informative to take note of the 3 issues mentioned by Das et al. We were happy to read that the possible emergence of ciprofloxacin resistant bacteria in patients is being carefully monitored. We also agree with the statement that mammalian topoisomerase has 100–1,000 times less affinity for binding to unlabeled ciprofloxacin than does bacterial DNA gyrase and that the binding is readily reversible, but we wonder if these data are also true for radiolabeled ciprofloxacin. To explain the accumulation of ciprofloxacin in both bacterial and sterile inflamed tissues, we must emphasize that the affinity of radiolabeled ciprofloxacin for binding to bacterial as well as human DNA in vitro was equal (2). Besides our data concerning 99mTc-ciprofloxacin (2), we have never seen any radiochemical or preclinical data in the literature. Therefore, we assume that Das et al. will perform the experiments with radiolabeled ciprofloxacin in view of the possibility that this radiolabeled agent may show completely different behavior from the unlabeled molecules. In our hands, 99mTc-ciprofloxacin bound in vitro even better to human cells than to bacteria (2). The final issue that Das et al. mention is the clinical experience with Infecton in 1,300 patients and the increasing reliability of this agent to distinguish infection from sterile inflammation. Farther on, Das et al. mention that the specificity of Infecton increased from 40% to 96% when 24-h images were taken instead of 4-h images. We are surprised to hear that only 40% specificity was obtained using scintigraphic data after 4 h, because this information has never been included in reports by the London group and puts a different complexion on the reliability of previously published clinical data. In addition, one may wonder how it is possible that the abstract by Wareham et al. (3) from the same London group mentions 80% specificity whereas Das et al. mention 96% specificity.
Another aspect of late imaging with the 99mTc-labeled compound concerns, of course, the quality of the images after 4 half-lifetimes. It would be interesting to see routine clinical images and judge the readability of the scintigrams. In our experience, 99mTc-images at 24 h are difficult to interpret, and a recent report by the London group agrees (4). Also, the in vivo stability of the agent at 24 h should be considered, and one might want to see the radiochemical data on 99mTc-ciprofloxacin after incubation in human serum for 24 h.
We also must remark on the statement by Das et al. that “if an organism does not take up ciprofloxacin, bacteria-specific imaging with Infecton will not occur.” Because 99mTc-ciprofloxacin was proposed as a sensitive and specific tracer for bacterial infections, one would expect that infections with ciprofloxacin-resistant bacteria, perhaps through an aspecific binding, are also detected. However, as reported in our study, 99mTc-ciprofloxacin accumulated in a similar way in both bacterial infections and sterile inflammation (1). In agreement with our results, Sarda et al. (5) recently reported the failure of 99mTc-ciprofloxacin to discriminate between infected and noninfected prosthetic joints in rabbits. In addition, although 99mTc-ciprofloxacin is not effective against C. albicans, accumulation of 99mTc-ciprofloxacin at sites of infection with C. albicans was even more pronounced than at sites of infection with ciprofloxacin-sensitive bacteria (1).
Another remark we must make involves the paragraph on the labeling and quality control of 99mTc-ciprofloxacin. We performed this labeling according to the method published by Vinjamuri et al. (6). The London group has used this method in at least 3 subsequent studies, for which successful imaging of infections in more than 240 clinical cases was reported. We understand that in the meantime this number has increased to well over 1,300 patients. In various communications by the group from London, we thoroughly searched for quality control data on the 99mTc-ciprofloxacin preparation. Unfortunately, we could not find relevant data. Even a recent article on 99mTc-ciprofloxacin did not report the results of quality control, except for mentioning a labeling efficiency of >92% (4). For such a new and important preparation, one would like to see data from high-performance liquid chromatography analysis of preparations containing 99mTc-ciprofloxacin, including ultraviolet chromatography and the identification of radioactivity peaks. It is, in our view, surprising that detailed information is still lacking, and it would only be logical and scientifically correct to provide such data as well as the relevant preclinical data in connection with successful clinical imaging. We had to remove 20%–40% of the total activity from the mixture containing radiolabeled ciprofloxacin (1)—an amount that agrees with the results published by the London group (7). Following the London procedure, we purified the preparation using diethylaminoethyl chromatography, which removes all colloids as well as free pertechnetate. Therefore, we consider our preparations colloid free, as confirmed by our routine reverse-phase high-performance liquid chromatography analysis for quality control of radiolabeled compounds. Therefore, it is encouraging that an independent research group using the Infecton kit, prepared according to the instructions of the London group, reported relevant quality control data showing lack of specificity for 99mTc-labeled ciprofloxacin in the detection of bacterial infections (5).
In their letter, Das et al. expressed some concern about the effect of antibiotics on the immune system. They caution the reader not to extrapolate in vitro and animal experimental data to humans. The London group is right, and we are happy that we did not speculate on this in our articles (1,8). We agree that antimicrobial peptides may interact with cells of the innate and acquired immune system (9), as we reported earlier for defensins (10). We have never observed side effects from (the low) concentrations of antimicrobial peptides that have been used so far for infection detection. These data agree with the statement by Das et al. that low concentrations of radiolabeled ciprofloxacin were used for detection of infections, in contrast to the higher concentrations used in therapy.
With regard to Figures 1 and 4, Das et al. suggested that we do not understand the basic principles of the accumulation of radiopharmaceuticals in tissues and organs. This suggestion is disturbing, because the London group apparently fails to remember that in several of our publications we have provided data showing the increasing accumulation of 99mTc-ubiquicidin 29–41 at infection sites (1,8). The same publications also provide data showing that 99mTc-ciprofloxacin, but not 99mTc-labeled antimicrobial peptides, accumulated at the site of sterile inflammation. On the basis of these findings we concluded that accumulation of 99mTc-labeled antimicrobial peptides at sites of infection could not be attributed merely to “increased perfusion.”
We hope to read more evidence on the specificity of 99mTc-ciprofloxacin as a marker for bacterial infections. The London group may remember that a report by Sarda et al. (11) and a recent publication in The Journal of Nuclear Medicine (5) pointed out a lack of specificity of 99mTc-ciprofloxacin for the detection of a Staphylococcus aureus infection in a rabbit model as well as infections in patients (12). On the basis of these data and our results (1,8), we cast doubt on the conclusion that successful imaging of bacterial infection is caused by direct binding of 99mTc-ciprofloxacin to dividing bacteria. In this connection, we wonder what the authors have in mind when stating that “there might be additional localization mechanisms for Infecton” (13). Also, we welcome scientifically valid criticism on our findings, and we suggest that the London group should perhaps submit their findings with additional details as a regular manuscript to peer review.
