Prospective Follow-up and Pulmonary Functions From a Placebo-Controlled Randomized Trial of Ribavirin Therapy in Respiratory Syncytial Virus Bronchiolitis FREE

RESPIRATORY SYNCYTIAL virus (RSV) is a common cause of respiratory morbidity in infants around the world. In our experience in Washington, DC, approximately 1 of every 100 primary RSV infections has resulted in a hospital admission.¹ Furthermore, approximately 1 of every 4 pediatric patients hospitalized with acute respiratory tract disease has had RSV infection. The outbreaks occur in yearly cycles. Currently, preventive therapy is available for some high-risk individuals, whereas ribavirin remains the only approved modality of therapy for selected patients.

In 1983, we initiated a randomized trial of ribavirin therapy compared with placebo in hospitalized patients aged 1 month and older. These patients were admitted to the hospital with RSV infection diagnosed by indirect immunofluorescence of respiratory tract secretions. They were randomized consecutively into treatment and placebo groups. We also anticipated, on clinical grounds, that these patients would spend 3 days or more in the hospital. We excluded patients with congenital heart disease. This study, which examined whether ribavirin therapy would result in an earlier clinical cure, was conducted during a 2-year period, and the results of the therapeutic observations were previously published.² At the completion of the trial, we were interested in long-term follow-up of the children who had been enrolled in the study, which also included an additional 10 children enrolled in the following season under the same protocol. During this phase of the study we wanted to determine whether there were any obvious adverse effects associated with the use of ribavirin, such as differences in the incidence of respiratory tract illness, wheezing, pneumonia, asthma, or reactive airway disease (RAD), and whether there was any difference in long-term outcome in pulmonary function test (PFT) results between the treated and control groups. The following is a report of long-term follow-up of these children.

The follow-up study consisted of the longitudinal evaluation of patients prospectively randomized to a ribavirin or a placebo control group.

In December 1983, after obtaining informed consent, we enrolled infants (≥1 month old) admitted to the hospital with acute lower respiratory tract illness and proven RSV infection. The presence of RSV was confirmed in exfoliated respiratory tract cells present in combined nasal and oropharyngeal swab specimens with indirect immunofluorescent antibody methods. Only infants who were expected to stay 3 days or longer in the hospital were included in the research study. Although we excluded patients with congenital heart disease, patients with chronic pulmonary disease and prematurity were included.

After obtaining a complete history from the parent or legal guardian, a physical examination was done by a physician (W.J.R.) who did not know whether the infant was assigned to the drug or the placebo group. This blinded observer assigned severity rating values that ranged from 0 (normal) to 4+ (more severe). Before the daily scoring the patient and the equipment were cleaned by hospital personnel not involved in the scoring. The preparation of the ribavirin and placebo was coordinated through the hospital pharmacy so that neither medical nor paramedical personnel were aware of the type of assignment.

The methods for randomization, administration of aerosol, collection of nasopharyngeal samples, and processing of viruses were described previously.²

Follow-up after hospital discharge consisted of scheduled visits to the pulmonary clinic, once in the fall and then again in the spring for the first 3 years after joining the study. Monthly telephone calls to assess health condition, including wheezing, were made to families throughout the study, and any significant information was recorded. During the first year of follow-up, children who developed wheezing and upper respiratory tract infection were instructed to come in for collection of respiratory tract secretions, which were tested for RSV by indirect immunofluorescent antibody techniques; viral cultures were also performed. Thereafter, indirect immunofluorescent antibody testing and viral cultures of a respiratory tract specimen were done only if the child developed wheezing during the time RSV was prevalent. During years 4 and 5, the visits were scheduled once a year. At each visit a pulse oximetry was done and a questionnaire was completed by the parent about the child's health. Any other physician contact was also documented if it dealt with respiratory tract symptoms (no attempts were made to review the actual private physician records). In the fifth year, at an age when the patients were expected to be cooperative, a PFT battery was performed, and the results were analyzed by a member of the Department of Pulmonary Medicine who was unaware of the patient's previous treatment status (the members of the Department of Pulmonary Medicine were unaware of the assignment status of the patients). Parents of the patients were also questioned about exposure of the child to smoking in the household. On the basis of the results from that PFT battery, patients were categorized into groups designated as normal, mild, moderate, or severe impairment. For evaluation of PFTs, the classification essentially was as follows: normal and mild, ratio of forced expiratory volume in 1 second to forced vital capacity (FEV₁/FVC) greater than 80% and FVC greater than 80% predicted; moderate, FEV₁/FVC less than 80% and FVC less than 80% predicted; and severe, FEV₁/FVC less than 60%. Whereas patients 5 to 7 years old underwent PFTs, after 7 years of age a methacholine challenge was done by the procedure described by Chai et al.³ On the basis of their reactivity to methacholine administration, patients who underwent methacholine challenge testing were classified as follows: normal, nonreactive (ie, <20% decrease in FEV₁) at all concentrations; mild, reactive (ie, 20% decrease in FEV₁) at 25 mg/mL of inhaled methacholine; moderate, reactive (ie, 20% decrease in FEV₁) at 1.25 to 10.00 mg/mL of inhaled methacholine (4 dose steps); and severe, reactive (ie, 20% decrease in FEV₁) at less than 0.6 mg/mL of inhaled methacholine (4 dose steps). Patients with a baseline FEV₁ less than 70% of predicted value were not tested. This aspect of the study also was approved independently by the Children's National Medical Center institutional review board, and the informed consent form for this challenge was obtained separately from parents. Precautions were taken addressing medications that could affect the test with omission or modification of the medication schedule. Baseline spirometry was recorded measuring FVC, FEV₁, FEV₁/FVC, and forced expiratory flow, midexpiratory phase (FEF_25%-75%), and the best of the 3 methods was recorded. Patients inhaled 5 breaths of 1 mL buffered isotonic sodium chloride solution, followed by spirometry testing for 1.5 minutes, and FEV₁ was obtained as a control. If FEV₁ decreased less than 20%, the methacholine challenge proceeded. Patients with a 20% or greater decrease were not tested. Dilutions of methacholine and isotonic sodium chloride solution of 0.075, 0.15, 0.30, 0.60, 1.25, 2.5, 5.0, 10.0, and 25.0 mg/mL of concentration were used. The provocative dose (PD) causing a 20% fall in FEV₁ was designated PD₂₀ FEV₁. The provocative tests were also scored by a pulmonologist (R.F., blinded) who was unaware of the treatment group assignment.

Comparisons of groups and results were done by parametric t test or by nonparametric Fisher exact, Wilcoxon, Kruskal-Wallis, and χ² tests.^{4 - 6} For an observation to be considered statistically significant, P≤.05 had to be obtained.

During the therapeutic trial, 17 patients were randomized to receive placebo and 25 patients were randomized to receive ribavirin. Four of the 42 original patients had underlying conditions. Of 25 patients in the ribavirin group, 3 were premature, and 2 of these 3 patients also had bronchopulmonary dysplasia (BPD). Of 17 patients in the placebo group, 1 was premature and also had BPD. The 2 groups were comparable in terms of severity of illness, oxygen requirements, length of intensive care unit stay, and length of hospital stay.²

Of the initial 42 patients, aged 1 to 33 months, admitted to the hospital, 24 had a diagnosis of bronchiolitis, 9 had bronchiolitis and pharyngitis, 1 had laryngotracheobronchitis, and 8 had pneumonia. Two patients with severe BPD, 1 in the placebo group and 1 in the treatment group, died during hospitalization. Their deaths were considered to be caused by their primary disease. Five of the remaining 40 patients were lost to follow-up soon after hospital discharge and never made a follow-up visit. Table 1 depicts the experience spanning the whole period, from initial randomization into the treatment protocol until the final (long-term follow-up) evaluation of those available for the PFTs.

Twenty-four patients from the ribavirin-treated group and 11 patients from the placebo group joined the longitudinal follow-up study. Fifteen boys and 9 girls, aged 1 through 10 months (mean, 4 months; median, 3 months), were in the ribavirin-treated group and 8 girls and 3 boys, aged 1 through 7 months (mean, 3.3 months; median, 2.0 months), comprised the placebo group. On entry, 16 ribavirin-treated patients (67%) and 8 placebo-treated patients (73%) were younger than 6 months. Four patients (17%) in the ribavirin-treated group had a diagnosis of pneumonia compared with 2 patients (18%) in the placebo group. Three ribavirin-treated patients (12%) had underlying conditions: 2 had BPD and 1 was premature. One patient (9%) in the placebo group had BPD. This report deals with the cumulative follow-up of 35 patients followed up longitudinally by our study team who completed at least 1 visit for a total of 212 clinic visits. Of the 24 ribavirin-treated patients, 17 (71%) being followed up lived in houses in which smoking was prevalent (no information was available for 2 patients). Of the 11 placebo-treated patients, 7 (64%) also lived in such environments.

The mean arterial oxygen percent saturation determined at the visits throughout follow-up was comparable in both groups (98%-99%). During follow-up, of 22 viral cultures from 16 different patients (12 from the ribavirin group and 4 from the placebo group), 13 were negative (8 in the ribavirin group and 5 in the placebo group). Of the 9 positive cultures, 5 yielded parainfluenza virus, 1 yielded herpes simplex virus, and 3 (2 in the ribavirin group and 1 in the placebo group) were positive for RSV. Two of these RSV infections required hospitalization, 1 patient at 17 months of age (10 months after ribavirin therapy) and another at 19 months of age (1 year after placebo use). The third patient was managed with an emergency department visit 3 years after ribavirin treatment at about 40 months of age.

During the first 3 years (Table 2) there was more RAD and wheezing in 11 placebo-treated patients compared with 24 ribavirin-treated patients, but the difference did not reach statistical significance atP<.05 (P=.10 by the Wilcoxon test). When one evaluates the total study period, the total frequency of wheezing decreased (Table 2). This was still not statistically significant for the total study (same number of individuals) (P=.10 by the Wilcoxon test). During those 6 years, we documented a mean of 44% episodes of wheezing per visit in the ribavirin group (range, 0%-100%) vs 65% in the placebo group (range, 33%-100%). During that same period (Table 3), 4 (17%) of 24 ribavirin-treated patients had 2 or more episodes of wheezing compared with 6 (55%) of 11 placebo-treated patients (P=.04 by Fisher exact test). There were no significant differences in the number of wheezing episodes between those who had subsequent positive viral cultures and those who did not (P=.30). The average number of wheezing episodes for those treated with ribavirin who had subsequent positive viral cultures was 1.33 (median, 1.00; range, 0.00-3.00). During the initial 3 years we observed more episodes of pneumonia in the ribavirin group. This difference was not statistically significant (P=.40 by Fisher exact test). Similarly, no statistical significance between groups was noted for the whole period (P=.39 by Fisher exact test).

At various times during follow-up, a total of 11 patients, 6 taking ribavirin and 5 taking placebo, were lost to follow-up. Four patients in the ribavirin group moved and left no forwarding address or telephone numbers, 1 dropped out, and another repeatedly missed appointments for PFTs and ultimately dropped out. All 5 placebo patients lost to follow-up moved and left no contact information. Ultimately, 24 of the 35 patients (18 receiving ribavirin and 6 receiving placebo) were available for PFTs. Five of these patients, all in the ribavirin group, were unable to complete the tests for the following reasons: 3 patients had outright mental retardation, 1 patient was unable to cooperate because of immaturity, and 1 patient had a concurrent illness and the mother refused to bring back the child to the hospital. Thus, 19 of the original 35 patients were available for various function tests. Their mean entry severity scores before enrollment in the original treatment were 2.4 (median, 2.0; range, 2.0-3.5) for placebo-treated patients and 2.6 (median, 2.0; range, 1.0-4.0) for ribavirin-treated patients. The virus load before treatment (tissue culture infectious dose 50/0.2 mL) ranged between 1.0 and 4.0 (mean, 3.6; median, 3.0) for the placebo group and between 0.5 and 6.5 (mean, 2.9; median, 2.5) for ribavirin-treated patients. At enrollment in the study, the age of the 19 children who were ultimately available for PFTs ranged from 1 to 7 months (mean, 3; median, 2) in the placebo group and from 1 to 10 months (mean, 4 months; median, 3 months) in the ribavirin group. At the time of PFT administration ribavirin-treated patients had a mean age of 8.3 years (median, 9.0 years) compared with 7.0 years (median, 7.0 years) in the placebo group, and PFTs were corrected for age, lengths, and weights. Of the 19 patients who completed various PFTs (Table 4), 7 completed the PFT only and 12 completed the PFT and the methacholine challenge. The 19 patients were distributed as follows: 6 were in the placebo group and 13 were in the ribavirin treatment group.

Nineteen patients completed PFTs (Table 5). More patients in the placebo group (6 of 6 completed) had moderate or severe findings compared with patients in the ribavirin group (6 of 13) (P=.043 by Fisher exact test [2-tailed]).

Table 6 depicts the methacholine challenge, the concentration at which mean PD₂₀ occurred, and interpretation of the various results. If the results from the 12 patients who completed the methacholine challenge test only are analyzed in a fashion similar to that used for all PFTs, ie, grouping normal and mild vs mild and severe, the difference does not reach statistical significance (P=.20). If the pulmonary function response in the methacholine challenge tests is weighted rank for severity scoring (Table 7) and examined by the Kruskal-Wallis test (χ² approximation), then the results of the methacholine challenge are only significant in favor of less severity in the ribavirin group (χ²₁=4.38;P=.04). However, because 1 of the cells had a value less than 5, exact P and point probability for this small group missed significance (P=.07).

Finally, analysis of the whole group weighting for severity (Table 8) shows that the difference is significant in favor of the ribavirin-treated patients (χ²₁=5.08;P=.02; exact P=.02). We additionally performed analysis of variance for the whole group including asthma. After correcting for asthma, the beneficial effect seen in those who received ribavirin was still significant (P=.03). Pulmonary function testing was also analyzed in terms of actual FVC, FEV₁/FVC, and FEF_25%-75%. The percentage of predicted values shown in Table 9 is based on the sex, age, and height of each individual patient. The mean FVC%, FEV₁, and FEF_25%-75% were higher for those who had been treated with ribavirin than for those in the placebo group, but these numbers did not reach statistical significance.

In this small, prospective study, patients who had RSV bronchiolitis during infancy who were treated with ribavirin experienced fewer overt episodes of wheezing, RAD, and asthma than those in the placebo group the first 3 years after RSV bronchiolitis. The difference between groups waned further in the subsequent 3 years. At no time was the difference statistically significant. However, when patients were evaluated with PFTs and challenged at 7 years or older, there was more reactive and more severe airway disease response in the placebo group compared with those treated with ribavirin. Hiatt and coworkers⁷ evaluated the effect of ribavirin therapy on pulmonary function in 94 previously healthy children with RSV lower respiratory tract infection. Maximum flow at functional residual capacity was higher at discharge and at the 12-month follow-up visit in those treated with ribavirin (P<.01 for both).

On the other hand, our failure to demonstrate significant differences by pulmonary function variables (FVC, FEV₁, FEV₁/FVC, and FEF_25%-75%) is tantalizing. Despite the trend in differences, an FEF_25%-75% of 97.1% for ribavirin-treated patients vs 81.4% for placebo-treated patients (a deviation of placebo predicted by 16% less from the ribavirin-tested group), the actual number of patients was too small and it did not reach statistical significance. A larger group may have answered the query about differences, if any. However, when the individual PFT (done blindly) results were scored and assigned to categories ranging from normal to severe, the magnitude of the difference between the treated and placebo groups was appreciated. Our patients were all enrolled under the same treatment protocol prospectively and were not selected for risk factors on the initial enrollment. They were randomized from the start; hence, they should have been exposed to the same viral strain(s) circulating in the community pool. Patients who had bronchiolitis and pneumonia during infancy have increased risk of recurrent wheezing by as much as 50%.^{8 - 10} These observations hold true during initial follow-up after hospitalization.¹¹ Still others associate RSV bronchiolitis during infancy with subsequent sensitization to allergens (common) and development of asthma given predisposition to asthma or allergy.¹² Our observations during the first 3 years of follow-up display a trend of airway reactivity after RSV, which has been noted by others.^{8 - 9} However, the trend for less episodes of reactivity observed in ribavirin-treated patients is of interest. Children who have bronchiolitis during infancy seem to have persistent pulmonary function abnormality later in life. These problems have been characterized in terms of their response to exercise-induced decreases in peak expiratory flow rates and in the higher predisposition to bronchial reactivity.¹³ These findings have also been noted in situations in which these patients have been challenged with histamine acid phosphate.¹¹

The possibility that these subsequent pulmonary findings result from the early damage caused by RSV seems to be supported by some of the earlier long-term observations.^{14 - 15} However, a recent report¹⁶ categorizes early wheezing as falling into 2 major groups that differ in predisposing factors at birth and in their outcomes. These predisposing factors are considered unrelated to viral infections. Whether viral infections during infancy lead later to RAD is still to be determined. Observations from the basic work by Welliver and colleagues^{17 - 18} support the role of mediators in RSV-induced acute disease and suggest a subsequent association. These investigators¹⁷ noted that the degree of hypoxia correlated significantly with the titers of RSV IgE and the concentration of histamine. Respiratory syncytial virus lower respiratory tract infection is implicated with IgE elevations.¹⁸ Similarly, Volovitz et al¹⁹ noted leukotriene C₄ in approximately 4 of 5 children who had the RSV IgE-specific detection compared with 1 of 4 of those without it. Rosmer and associates,²⁰ in studying the effects of ribavirin therapy on the production of RSV-specific IgE, noted that patients treated with ribavirin had in their respiratory tract secretions lower amounts of RSV-specific IgE compared with controls.

The difference in the severity of response to challenge with methacholine and the difference in the PFT results favoring treated patients over controls that we observed need further study. Our sample sizes were small. Our observations were limited to pulmonary functions and wheezing during follow-up because we did not evaluate for cytokines, IgE, or other mediators. Our original goal was to observe treated patients for long-term safety. The loss of patients initially and during follow-up, beyond our control, affected the strength of the study. Although the initial severity assessment and viral titers remained balanced between groups, as patients dropped out, a tendency developed for those 6 placebo-treated patients who completed PFTs to be initially younger than the 13 patients in the ribavirin group (mean age, 2.3 vs 4.3 months). Also, at the time of testing, the mean age of the placebo group was 7.0 years vs 8.7 in the ribavirin group. In addition, potential bias may have been introduced by the fact that 24 of 25 ribavirin-treated patients compared with 11 of 17 placebo-treated patients participated in the longitudinal follow-up (P<.02 by the Fisher exact test). All of these differences could again affect the final observed results. Furthermore, because differences in RAD and wheezing episodes did not achieve statistical significance, we are left with speculation about the significance of the differences observed in pulmonary function, which is of interest in light of the comparable oxygen saturation in each group. A study with larger samples with pulmonary function evaluation of ribavirin-treated patients matched to controls of the same age challenged with methacholine might help answer this query. Finally, 2 recent publications must be addressed.^{21 - 22} Krilov and coworkers²¹ reported a comparison of ribavirin-treated patients and age-matched controls. These investigators conducted PFTs of the participants and did not demonstrate either beneficial or untoward adverse effects associated with ribavirin use. The results of our study, with a limited number of patients, support the safety findings. However, there are inherent differences between both studies, including the fact that our randomized patients were all part of a prospectively enrolled cohort from a single center (vs 6 centers); our patients' illnesses were comparable in severity at enrollment and furthermore presumably represent balanced exposure to the strains prevalent in the community. On the other hand, the study by Long and colleagues²² is similar to ours in representing a randomized cohort follow-up of patients initially enrolled in a treatment trial. The mean age of ribavirin-treated patients vs controls at enrollment was 2.4 vs 4.1 months, and 70% of the ribavirin-treated patients were boys. Long and coworkers²² noted a difference in peak expiratory flow rates between ribavirin-treated patients and controls of 113% vs 102% (P=.17). Although they noted a trend in better PFT results between ribavirin-treated patients and controls, this was not significant. Neither of the 2 studies^{21 - 22} used methacholine challenge to detect reactivity. Hence, to date, there are reports with findings at odds with each other. Our observations, if confirmed by larger studies, may support the use of ribavirin, regardless of its antiviral effectiveness, for its long-term effect on outcome of patients infected with RSV during infancy.

Accepted for publication October 15, 1998.

Supported in part by a grant from ICN Pharmaceuticals, Costa Mesa, Calif.

Presented in part at the Annual Meeting of the American Pediatric Society and the Society of Pediatric Resources, Washington, DC, May 10, 1996.

Members of the Ribavirin Study Group are as follows: Christine Bechtel, BSN, MSN; Susan Smith, BSN, MSN; Hollis Chaney, MD; Franco Piazza, MD; Jara Vojtova, MUC; Kantilal M. Patel, PhD; and Mark Guargliardo, PhD.

Corresponding author: William J. Rodriguez, MD, PhD, Department of Infectious Disease, Children's National Medical Center, 111 Michigan Ave NW, Washington, DC 20010-2970 (e-mail: eayala@cnmc.org).

Editor's Note: I've always been surprised that many (most?) physicians believe that the use of ribavirin is based on its safety, which has not been the case. Perhaps this study, added to several others, will convince them.—Catherine D. DeAngelis, MD