Recurrent respiratory papillomatosis severity score over time at a cidofovir injection concentration of 5 mg/mL.
Recurrent respiratory papillomatosis subsite score over time at a cidofovir injection concentration of 5 mg/mL.
Patient 4, just before her first cidofovir treatment at a concentration of 5 mg/mL. Her recurrent respiratory papillomatosis severity score was 20, at the following 6 subsites: aryepiglottic folds (2 on the right and 1 on the left), false vocal cords (1 on the right and 2 on the left), true vocal cords (2 on the right and 3 on the left), arytenoids (1 on the right and 3 on the left), anterior commissure (2), and posterior commissure (3).
Patient 4, 1 month after her fourth cidofovir treatment at a concentration of 5 mg/mL. Her recurrent respiratory papillomatosis (RRP) severity score was 0, as no RRP was identified at any subsite. Staging at all time points also includes rigid bronchoscopy for evaluation of the subglottis and trachea.
Percentage of improvement compared with baseline severity at a cidofovir injection concentration of 5 mg/mL.
Customize your JAMA Network experience by selecting one or more topics from the list below.
Akst LM, Lee W, Discolo C, Knott D, Younes A, Koltai PJ. Stepped-Dose Protocol of Cidofovir Therapy in Recurrent Respiratory Papillomatosis in Children. Arch Otolaryngol Head Neck Surg. 2003;129(8):841–846. doi:10.1001/archotol.129.8.841
To evaluate a stepped-dose protocol for intralesional injection of cidofovir in children with recurrent respiratory papillomatosis (RRP).
Prospective, nonrandomized case series.
Tertiary care children's hospital.
Eleven children undergoing evaluation for RRP from June 1, 2000, through December 31, 2001.
Intralesional injection of cidofovir was performed after microlaryngoscopy and carbon dioxide laser treatment. Patients received 4 monthly injections at a concentration of 5 mg/mL and returned 1 month after the last injection for follow-up. Patients with recurrent or recalcitrant disease then started a series of 4 monthly injections at a concentration of 10 mg/mL.
Papilloma stage (0-3) documented at multiple subsites by means of serial microlaryngoscopy. We calculated a severity score by summing the scores at all affected subsites.
The severity score decreased in each of the 11 patients during the course of therapy, from a mean ± SD of 13.7 ± 6.0 at enrollment to 2.1 ± 3.4 at 1-month follow-up. Six patients experienced complete resolution (stage 0) and 4 others had mild disease (stage, ≤5) after 4 treatments at the 5-mg/mL concentration. Five patients with residual or recurrent RRP subsequently started a series of 4 cidofovir treatments at a concentration of 10 mg/mL, with a mixed response.
Intralesional injection of cidofovir seems to reduce the burden of disease in children with RRP. Patients with persistent or recurrent disease may benefit from an increased cidofovir concentration of 10 mg/mL, although some aggressive papillomatous disease remains refractory to cidofovir treatment.
RECURRENT respiratory papillomatosis (RRP) remains a challenge for the patient and the treating otolaryngologist. Recurrent respiratory papillomatosis develops in 1500 to 3000 children in the United States each year.1 The cause of RRP is the human papillomavirus (HPV), particularly subtype HPV-11, although HPV-6 and, rarely, HPV-16 are also seen.2 As the disease demonstrates a predilection for the larynx, the presenting symptom is typically hoarseness. Some children may present with aphonia or airway obstruction.
The disease course is unpredictable, and in some patients multiple surgical excisions of papilloma, including repeated carbon dioxide (CO2) laser vaporization, may be necessary to maintain adequate phonation and respiration. Patients with severe cases that compromise the airway may require short-term tracheotomy or even long-term tracheostomy. Eventually most children enjoy spontaneous remission, but chronic alterations in laryngeal structure and function may result in continuing voice and airway problems.3
The difficulty of treating RRP in children has led to the development of many adjuvant therapies aimed at reducing disease burden and improving outcomes, including interferon alfa, retinoic acid, indole-3-carbinol, cimetidine hydrochloride, acyclovir sodium, ribavirin, and methotrexate.4 However, none of these treatments is considered universally curative. Moreover, concerns about their systemic use in children persist.
Cidofovir (1-[(S)-3-hydroxy-2-(phosphonomethoxy)-propyl]cytosine) is a cytosine nucleoside analogue that has shown benefit as a locally applied adjuvant therapy in the treatment of RRP in children.5,6 In contrast to acyclovir and ganciclovir sodium, which require viral activation, cidofovir is phosphorylated intracellularly to its active metabolite cidofovir diphosphate independently of viral infection.7 This active metabolite is a competitive inhibitor of viral DNA polymerase and blocks viral DNA synthesis by being incorporated into the growing DNA strand. Because viral DNA polymerase is unable to excise incorporated cidofovir diphosphate, cidofovir may have activity that is prolonged beyond the 17- to 65-hour half-life of its active metabolite.8
Van Cutsem et al9 first reported the injection of cidofovir into a squamous papilloma in the hypopharynx-esophagus of a 69-year-old woman, with good results. Snoeck et al10 later published their experience with cidofovir in laryngeal papillomatosis and noted complete resolution of papilloma in 14 of 17 patients (mostly adults), with subsequent relapse in 4 patients. A subsequent case series by Pransky et al5,6 documented the potential benefit of cidofovir in children with RRP. One case report detailed the successful treatment of juvenile laryngeal papillomatosis–related multicystic lung disease with cidofovir.11
Despite the promise of cidofovir as an adjuvant therapy for RRP in children, there are no uniform protocols for its use. Questions remain about optimal dose, interval, and duration of treatment. It also remains unknown what impact other adjuvant therapies might have on the efficacy of cidofovir. In addition, although it appears safe,6 the safety profile of cidofovir has not yet been fully established. This study does not address any of these issues directly. However, the goal of this study is to assess the effect of a stepped-dose protocol for cidofovir in children with RRP. We hope that adding to the existing knowledge about cidofovir might help to inform future uses of this promising agent.
All children undergoing evaluation for RRP at The Cleveland Clinic Foundation Children's Hospital, Cleveland, Ohio, from June 1, 2000, through December 31, 2001, were offered the opportunity to enroll in this study. The intent of the study was to examine the effect of various dosing variables for intralesional injection of cidofovir, when cidofovir is used in combination with CO2 laser treatment of RRP. No other adjuvant therapy was provided during the course of the study. Informed consent was obtained from the patient's parents or legal guardians according to the institutional policies of The Cleveland Clinic Foundation. The hospital's institutional review board approved all data collection.
All treatments were performed with the patient under general anesthesia. Suspension microlaryngoscopy was performed using an appropriately sized Benjamin or Lindholm (both from Karl Storz Endoscopy-America Inc, Culver City, Calif) laryngoscope. Patients breathed spontaneously or received jet-assisted ventilation throughout all interventions. A 0° rigid telescope was used to inspect each patient's larynx and trachea, and the findings were photographically documented. To assess disease severity, we used the staging instrument developed by Derkay et al.12 This instrument grades the involvement of each affected subsite using a 4-point scale (0 indicates none; 1, minimal; 2, moderate; and 3, severe). Each subsite is scored separately. The overall subsite score reflects the number of involved subsites, whereas the severity score is calculated by summing papilloma stage at each subsite.
After the disease severity was rated, the operating microscope was brought into the field, and all visualized disease was excised with a CO2 laser set at 5 W and a 0.2-second repeat. Cidofovir was injected into the submucosal plane at the involved subsites using a modified butterfly needle that was held in cupped forceps. The total amount of cidofovir injected, not accounting for that lost to leak above the submucosal plane, ranged from 0.2 to 3.0 mL at a concentration of 5 mg/mL, as based on the work performed by Pransky et al.5,6 All patients were observed in the postoperative recovery unit after their procedure to ensure airway adequacy and then were discharged to home on the day of surgery. The patients returned every month for the next 3 months for repeat microlaryngoscopy, disease staging, and cidofovir injection to complete a series of 4 monthly injections at a concentration of 5 mg/mL.
One month after the last injection, the patients returned for a follow-up examination. At this scheduled follow-up, most patients underwent microsuspension laryngoscopy. However, if a patient did not have any clinical evidence of disease, fiberoptic flexible laryngoscopy performed in an office setting was accepted as a satisfactory follow-up procedure to spare the patient a therapeutically unnecessary operative procedure. The severity and subsite scores were calculated at follow-up examination using the techniques described in the previous paragraphs.
Complete resolution was defined as a severity score of 0 at follow-up. Patients were considered to have mild disease if their severity score at follow-up was 5 or less. Moderate disease was defined as a severity score ranging from 6 to 15. Patients were considered to have recurrent disease if scores reached 0 before the disease returned, whereas disease that never resolved completely was termed recalcitrant.
In patients with recurrent or recalcitrant RRP noted at follow-up after the series of 4 injections, a second series of 4 injections was initiated at an increased concentration of 10 mg/mL (injection amount, 1.0-3.0 mL). Again, these injections occurred at 1-month intervals with a planned follow-up 1 month after the last injection. The severity and subsite scores were calculated during microlaryngoscopy before each injection and at the 1-month follow-up as described above.
We used a paired t test to examine the change in severity and subsite scores between the first injection and follow-up after the series of treatments at the 5-mg/mL concentration. Because of limited sample size, no statistical analysis was performed for the series of treatments at the 10-mg/mL concentration. To control for disease severity at baseline, the percentages of change in severity and subsite scores were also calculated for each patient. Because these data were not distributed normally, we used Wilcoxon signed rank tests to assess the statistical significance of the percentage of change. We considered P values of less than .05 to be statistically significant for t tests and Wilcoxon signed rank tests. A Spearman correlation coefficient was calculated for a scatterplot, in which 1 indicates perfect correlation and 0 indicates no relationship. If 95% confidence intervals (CIs) approached 0, the relationship being studied was not thought to be significant.
Eleven children elected to participate in the protocol (7 boys and 4 girls). Patient characteristics at the time of enrollment are listed in Table 1. Patient 11 underwent an unspecified previous laryngeal surgery for a weak voice, and patients 8 and 9 were enrolled in the protocol immediately after the initial diagnosis of RRP. The remaining 8 patients all underwent previous CO2 laser treatments as indicated. Except for patient 7, who had a history of patent ductus ateriosus that was closed before study enrollment, none of the children had significant comorbid illness or a chronic medical condition.
One patient experienced bradycardia on induction of general anesthesia and was intubated to secure her airway at the beginning of one of her surgical treatments. This patient was quickly stabilized, and she was extubated before continuing the surgical procedure under jet-assisted ventilation. None of the remaining patients were intubated at any time for any of their procedures.
The severity and subsite scores for each patient, which reflect the state of the upper airway before CO2 laser treatment at 5 time points (injections 1, 2, 3, and 4 and follow-up), are listed in Table 2 and Table 3. Table 2 reflects the data collected for the 11 patients who completed the protocol at the 5-mg/mL concentration, whereas Table 3 presents the data for the 5 patients with recurrent or recalcitrant RRP who subsequently began a second series of cidofovir injections at a concentration of 10 mg/mL. Two of those patients were unavailable for follow-up at the 1-month posttreatment follow-up, but each eventually returned to the office several months later with recurrent disease. The amount of cidofovir injected at each time point is listed, and the few instances in which follow-up was performed by means of fiberoptic examination in the office rather than microlaryngoscopy are indicated.
At an injection concentration of 5 mg/mL, each of the 11 patients experienced an overall decrease in the severity score during the course of therapy. Ten of the 11 patients also experienced an overall decrease in subsite involvement during the series of 4 treatments; patient 5 had the same subsite score at follow-up that she had at the beginning of the protocol. Figure 1 demonstrates the response in the severity score as a function of time, whereas Figure 2 does the same for subsite involvement. The mean ± SD improvement in the severity score across the group of 11 patients was 11.6 ± 4.1 (P<.001), whereas the mean ± SD) improvement in the subsite score was 3.1 ± 2.1 (P<.001). If the percentage of change from baseline is measured rather than absolute change, as a means of controlling for initial severity, the change in severity score (median, 100% improvement; range, 56%-100%; P = .001) and subsite involvement (median, 100% improvement; range, 0%-100%; P = .002) remain statistically significant.
Seven (64%) of the 11 patients (95% CI, 31%-89%) had a complete or near-complete response to the series of 4 CO2 laser treatments combined with cidofovir injection (severity score, 0 or 1 at 1-month follow-up). Photographs of the larynx of patient 4 at the initiation of therapy (Figure 3) and at her 1-month follow-up appointment (Figure 4) demonstrate this complete response. Of the 7 patients who initially experienced a complete or near-complete response, 6 maintained this response during serial follow-up examinations in the office (patients 3, 4, 8, 9, 10, and 11). The follow-up periods for these patients from their last time point are 19, 18, 6, 5, 5, and 3 months, respectively (mean follow-up, 9.3 months; SD, 7.2 months). None of these patients has required further therapy for RRP since concluding their series of treatments in the 5-mg/mL protocol.
Conversely, the remaining 5 patients (patients 1, 2, 5, 6, and 7) all required further treatment for RRP after their series of treatments at the 5-mg/mL concentration, and they subsequently were enrolled in a second series of treatments with the concentration of cidofovir increased to 10 mg/mL. One of these 5 patients (patient 6) experienced a complete response to the 5-mg/mL protocol but was noted to have a recurrence of his disease 3 months after his last injection. The remaining 4 patients all had decreased but persistent levels of disease after their initial series of treatments. Given the relatively small size of the group undergoing the 10-mg/mL injection, and given the wide range of clinical outcomes in this group (Table 3), no statistical analysis was performed for this group as a whole.
The 3 patients (patients 2, 5, and 6) who were enrolled in the 10-mg/mL stepped-dose protocol with relatively mild disease all experienced complete resolution of RRP during treatment or at its completion. The 2 patients who had relatively moderate or advanced disease at the start of the 10-mg/mL protocol (patients 1 and 7) had essentially stable disease scores during this course of treatment, despite the increased concentration of cidofovir. These same patients also had more aggressive disease, based on the number and frequency of previous laser treatments, than did the other patients at the time of enrollment in the study (Table 1).
These data raised the possibility that severe RRP responds more poorly to cidofovir than moderate or mild amounts of RRP might. To evaluate this possibility, we created a scatterplot to express the percentage of improvement in severity score for each patient in the 5-mg/mL series as a function of baseline severity (Figure 5). The analysis of this scatterplot suggests that the relationship between baseline severity and response to cidofovir is not strong, at least at the 5-mg/mL concentration (Spearman correlation coefficient, −0.62; 95% CI, −1.0 to −0.03).
We conducted an observational study designed to look at outcomes of children with RRP when treated with a stepped-dose protocol of cidofovir in addition to laser therapy. With regard to patient selection, we chose to offer enrollment to all children with RRP without regard to disease severity. Other articles investigating the use of cidofovir in RRP in children have offered participation only to patients with disease severe enough to require debulking at least once a month.5,6 If cidofovir is being evaluated as an adjuvant therapy to be used only in severe cases, this latter approach is certainly appropriate. Also, the thought that demonstrating the potential benefits of cidofovir relative to standard treatment may be more difficult in patients with mild disease is another reason to study only patients with severe disease. Nevertheless, if cidofovir is to become part of routine treatment for RRP in all children, studies must be designed accordingly. For this reason, study participation was offered to all patients, with no attempts made at patient stratification according to disease severity.
Because not every patient in this study required frequent laser therapy before enrollment, this study lacked the internal control seen in studies by Pransky et al,5,6 in which frequency of intervention before and after the initiation of cidofovir therapy was used to validate efficacy. Without this internal control, patient improvement in this study cannot be attributed only to the addition of cidofovir. As the severity of RRP is known to wax and wane, it is possible that patients may have improved with laser treatment in the absence of cidofovir or any treatment at all. Thus, it is impossible to attribute the improvement seen in this patient cohort only to cidofovir. Still, the study demonstrated a statistically significant benefit of the protocol for injection of cidofovir at a concentration of 5 mg/mL in addition to laser therapy performed at 1-month intervals for a course of 4 treatments. This benefit was seen in severity of RRP and in subsite involvement, and the magnitude of improvement in these areas suggests that the gain was clinically important and statistically valid.
Although the presence of severe disease was not predictive of a poor response to cidofovir in our study, as revealed in the scatterplot analysis, a poor response to cidofovir seemed limited to the patients with severe disease at baseline. If the patients with a good therapeutic response at the 5-mg/mL concentration are censored, a dichotomy in the response of patients to the 10-mg/mL concentration was anecdotally evident, as seen for patients 1 and 7. The results of this stepped-dose protocol, then, suggest that there might be different phenotypes of RRP that respond to cidofovir differently; that approximately half of RRP cases seem sensitive to cidofovir at a concentration of 5 mg/mL; that another quarter demonstrates intermediate sensitivity with response to the concentration of 10 mg/mL; and that the remaining quarter remains resistant to cidofovir even at this higher concentration. Furthermore, this last group of patients, whose disease seems resistant to cidofovir, seems to come from a patient population with more severe disease at baseline. The small sample size of this study, although on a par with the previously published findings in pediatric cohorts in the investigation of cidofovir, limits the ability to draw firm conclusions. In addition, because the natu-ral history of RRP is to wax and wane in severity, the course of 1 or 2 patients should not be used to generalize about the disease entity as a whole. However, the hypothesis that there are subgroups of patients whose RRP might respond to cidofovir differently seems to warrant further investigation, particularly if such investigation will allow stratification of patients according to likelihood of cidofovir success before treatment.
With regard to concentration, a balance must be reached between the amount of injection and the overall amount of drug delivered. If the cidofovir dose is too diluted, the volume necessary to deliver the desired amount of drug might be too large within the narrow confines of the pediatric airway, and other groups have increased their concentration of cidofovir from 2.5 to 5 mg/mL for this reason.5 On the other hand, a higher concentration than necessary might expose patients to excess amounts of a medication whose safety profile is still being evaluated. On the basis of the experience of these other groups, 5 mg/mL was chosen as the starting concentration for this study. Reserving increased concentrations of cidofovir for those patients in whom lower concentrations of therapy has failed is one benefit of a stepped-dose approach.
With regard to the location and dose of cidofovir injection, an inspection of Table 2 and Table 3 showed that the total amount of cidofovir injected at either concentration tended to increase during the study. The reason for this increase in dose was a growing realization on the part of the investigators that papilloma recurrence was not confined to those subsites of the larynx where papilloma had been previously visualized but was instead multifocal. Even as subsite involvement generally trended downward over time, it occasionally increased from one time point to the next, evidence that papilloma was presenting in previously uninvolved areas. If the disease demonstrates this field effect, it is reasonable to hypothesize that local treatment with an agent such as cidofovir must also encompass this field to be successful. The total dose of cidofovir injected, then, becomes a function of treating sites of known papilloma and other areas within an affected larynx.
This concept of field effect also has implications for the duration of cidofovir therapy necessary to maintain the desired response. In this study, patients who responded well to cidofovir at a concentration of 5 mg/mL experienced complete disease resolution after 4 treatments and have remained without disease for a mean follow-up of 9.3 months. However, the selection of 4 as the number of necessary treatments was made empirically, and, in truth, the optimal duration, frequency, and possible carryover effect (in which the results of the 5-mg/mL treatments, for instance, are not seen until after the 10-mg/mL treatments had begun) of cidofovir therapy are not yet established. Despite these uncertainties, observations from the present study might guide future treatment protocols. For instance, the concept of RRP exerting a field effect throughout the affected larynx implies that HPV might remain latent even at subsites without active disease. Recent investigation into the HPV life cycle demonstrates that it has active and latent phases, and that latency might influence the efficacy of antiviral therapies.13 As cidofovir works to block viral DNA polymerase, cidofovir might be effective only against active HPV. The duration and frequency of cidofovir application, then, may need to be tailored to the life cycle of HPV.
This investigation into a stepped-dose protocol for the intralesional injection of cidofovir in RRP in children serves as a pilot study. Its relatively small number of patients, although equal to those in previously published series on the use of cidofovir in RRP in children, and its lack of internal control make it impossible to demonstrate conclusively the efficacy of cidofovir. In the present study, however, patients demonstrated a statistically significant benefit of the treatment protocol, and cidofovir continues to be a promising new tool in the treatment of a disease that remains frustrating to the patient and otolaryngologist. A multicenter, prospective, randomized controlled trial is necessary to firmly demonstrate the benefit of cidofovir and to establish the optimal concentration and dosing variables for its use. Observations from the present study may have implications for any future trial and should be incorporated into its planning.
Corresponding author and reprints: Peter J. Koltai, MD, Department of Otolaryngology and Communicative Disorders, The Cleveland Clinic Foundation, Desk A-71, 9500 Euclid Ave, Cleveland, OH 44195 (e-mail: firstname.lastname@example.org).
Submitted for publication September 20, 2002; final revision received November 25, 2002; accepted December 12, 2002.
This study was presented as a podium presentation at American Society of Pediatric Otolaryngology; May 13, 2002; Boca Raton, Fla.