Novel statistical tools for monitoring the safety of marketed drugs.

Robust tools for monitoring the safety of marketed therapeutic products are of paramount importance to public health. In recent years, innovative statistical approaches have been developed to screen large post-marketing safety databases for adverse events (AEs) that occur with disproportionate frequency. These methods, known variously as quantitative signal detection, disproportionality analysis, or safety data mining, facilitate the identification of new safety issues or possible harmful effects of a product. In this article, we describe the statistical concepts behind these methods, as well as their practical application to monitoring the safety of pharmaceutical products using spontaneous AE reports. We also provide examples of how these tools can be used to identify novel drug interactions and demographic risk factors for adverse drug reactions. Challenges, controversies, and frontiers for future research are discussed.

The relation of pneumothorax and other air leaks to mortality in the acute respiratory distress syndrome.

BACKGROUND: In patients with the acute respiratory distress syndrome, pneumothorax and other air leaks - any extrusion of air outside the tracheobronchial tree - have been attributed to high ventilatory pressures or volumes and linked to increased mortality. METHODS: We analyzed data from a prospective trial of aerosolized synthetic surfactant in 725 patients with the acute respiratory distress syndrome induced by sepsis. We compared the ventilatory pressures and volumes in the patients without any air leaks (the highest values during the five-day study) with the pressures and volumes in those with pneumothorax or with any air leaks (the highest values during the 16- and 24-hour periods before the complication developed). RESULTS: Fifty patients (6.9 percent) had pneumothorax and 77 (10.6 percent) had pneumothorax or other air leaks. There were no significant differences between patients with air leaks and those without air leaks in any pressure or volume examined. Overall mortality at 30 days was 40.0 percent (95 percent confidence interval, 36.4 to 43.6); among the patients with pneumothorax, it was 46.0 percent (95 percent confidence interval, 32.2 to 59.8), and among those without pneumothorax, it was 39.3 percent (95 percent confidence interval, 35.6 to 43.0; P=0.35). The mortality rate was 45.5 percent (95 percent confidence interval, 34.4 to 56.6) in the group with any air leaks and 39.0 percent (95 percent confidence interval, 35.3 to 42.8) in the group without air leaks (P=0.28). CONCLUSIONS: In patients with sepsis-induced acute respiratory distress syndrome who were receiving mechanical ventilation with conventional pressures and volumes, there were no significant correlations between high ventilatory pressures or volumes and the development of pneumothorax or other air leaks. Pneumothorax or other air leaks were not associated with a significantly increased mortality rate.

Effects of aerosolized surfactant in patients with stable chronic bronchitis: a prospective randomized controlled trial.

CONTEXT: Chronic bronchitis, estimated to affect more than 13 million adults in the United States, is characterized in part by retention of airway secretions, but no approved or effective therapy for airway mucus retention in patients with chronic bronchitis has been established. Surfactant reduces sputum adhesiveness, which contributes to difficulty in clearing secretions, but surfactant has not been tested in patients with chronic bronchitis. OBJECTIVE: To examine the effects of exogenous surfactant on sputum clearance and pulmonary function in patients with stable chronic bronchitis. DESIGN: A prospective, multicenter, randomized, double-blind, parallel-group, placebo-controlled comparison of the effects of 2 weeks of treatment with 3 doses of aerosolized surfactant (palmitoylphosphadidylcholine [DPPC]) or saline (placebo). SETTING: Four US teaching hospitals. PARTICIPANTS: A total of 87 adult patients with the diagnosis of stable chronic bronchitis. MAIN OUTCOME MEASURES: Pulmonary function, respiratory symptoms, and sputum properties before treatment (day 0), after 2 weeks of treatment (day 14), and 7 days after stopping treatment (day 21). RESULTS: A total of 66 patients were randomized to surfactant treatment and 21 to saline treatment. Patient demographic characteristics between groups were similar at baseline. In patients who received a DPPC dose of 607.5 mg/d for 2 weeks, prebronchodilator forced expiratory volume in 1 second (FEV1) increased from 1.22 L (SEM, 0.08 L) at day 0 to 1.33 L (SEM, 0.09 L) at day 21 (P=.05), an improvement of 11.4%; postbronchodilator FEV1 improved 10.4% by days 14 and 21 (P=.02); and the ratio of residual volume to total lung capacity, a measure of thoracic gas trapping, decreased 6.2% by day 21 (P=.009). In the surfactant groups, there was a dose-dependent increase in the ability of sputum to be transported by cilia in vitro. CONCLUSION: Aerosolized surfactant improved pulmonary function and resulted in a dose-related improvement in sputum transport by cilia in patients with stable chronic bronchitis.

Changes in the inflammatory response of the lung during acute respiratory distress syndrome: prognostic indicators.

We studied paired bronchoalveolar lavage (BAL) in patients with sepsis-associated acute respiratory distress syndrome (ARDS). Patients were evaluated at one institution and underwent bronchoscopy with BAL within 48 h of the onset of ARDS. Patients were restudied with bronchoscopy and BAL after 4 d of treatment. Fifty-eight patients were initially studied, with 44 patients having follow-up bronchoscopy after 4 d. The overall 30-d survival for the ARDS group was 60%. In the initial lavage, there was no difference in the neutrophils between the survivors and nonsurvivors (survivors: 59 [0-98]%; Median [Range]; nonsurvivors: 55 [0-92]%). The follow-up lavage demonstrated a significant drop in the neutrophils for the survivors (36 [4-89]%, p < 0.002) which was not seen for the nonsurvivors (70 [26-95]%). Initial IL-8 concentrations in the BAL fluid were not significantly different between the two groups. In the follow-up lavage, there was a significant fall for the IL-8 concentrations for the survivors but not the nonsurvivors. We conclude that neutrophil influx in ARDS may rapidly resolve within a week of the onset of ARDS. The resolution of neutrophils was associated with a good prognosis.

Aerosolized surfactant in adults with sepsis-induced acute respiratory distress syndrome. Exosurf Acute Respiratory Distress Syndrome Sepsis Study Group.

BACKGROUND: Patients with acute respiratory distress syndrome (ARDS) have a deficiency of surfactant. Surfactant replacement improves physiologic function in such patients, and preliminary data suggest that it may improve survival. METHODS: We conducted a prospective, multicenter, double-blind, randomized, placebo-controlled trial involving 725 patients with sepsis-induced ARDS. Patients were stratified according to the risk of death at base line (indicated by their score on the Acute Physiological and Chronic Health Evaluation [APACHE III] index) and randomly assigned to receive either continuously administered synthetic surfactant (13.5 mg of dipalmitoylphosphatidylcholine per milliliter, 364 patients) or placebo (o.45 percent saline; 361 patients) in aerosolized form for up to five days. RESULTS: The demographic and physiologic characteristics of the two treatment groups were similar at base line. The mean (+/- SD) age was 50 +/- 17 years in the surfactant group and 53 +/- 18 years in the placebo group, and the mean APACHE III scores at randomization were 70.4 +/- 25 and 70.5 +/- 25, respectively. Hemodynamic measures, measures of oxygenation, duration of mechanical ventilation, and length of stay in intensive care unit did not differ significantly in the two groups. Survival at 30 days was 60 percent for both groups. Survival was similar in the groups when analyzed according to APACHE III score, cause of death, time of onset and severity of ARDS, presence or absence of documented sepsis, underlying disease, whether or not there was a do-not-resuscitate order, and medical center. Increased secretions were significantly more frequent in the surfactant group; the rates of other complications were similar in the two groups. CONCLUSIONS: The continuous administration of aerosolized synthetic surfactant to patients with sepsis-induced ARDS had no significant effect on 30-day survival, length of stay in the intensive care unit, duration of mechanical ventilation, or physiologic function.

Clinical diagnostic criteria of the adult respiratory distress syndrome in the intensive care unit.

OBJECTIVES: To determine the use of commonly used diagnostic criteria for adult/acute respiratory distress syndrome (ARDS), evaluate physiologic variables of most value in diagnosing ARDS, and assess the frequency of newly diagnosed ARDS. DESIGN: Self-administered questionnaire by mail. SETTING: Hospital intensive care units (ICUs). SUBJECTS: Intensive care physicians (n = 923) in the United States listed as ICU directors in the Society of Critical Care Medicine Membership Directory. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: A total of 448 (48.5%) physicians responded, with 301 questionnaires completed. Bilateral infiltrates on chest radiograph, the PaO2/FIO2 ratio, and pulmonary artery occlusion pressure were the most commonly used diagnostic criteria. However, the actual values used to diagnose ARDS were highly variable. The most important clinical physiologic variable used in determining the respiratory status of the ARDS patient was the PaO2/FIO2 ratio, followed by shunt fraction, alveolar-arterial oxygen tension gradient, FIO2, PaO2, respiratory system compliance, and minute ventilation. Respondents indicated that 9% of ICU beds at their institutions were occupied by a patient diagnosed with ARDS within the previous 7 days and 18.6% of all mechanically ventilated patients had ARDS by their own criteria. Based on the total number of ICU beds, the predicted incidence of ARDS would be approximately 275,000 patients per year in the United States. CONCLUSION: A wide range of diagnostic criteria are utilized by clinicians in the diagnosis of ARDS.

Safety and potential efficacy of an aerosolized surfactant in human sepsis-induced adult respiratory distress syndrome.

OBJECTIVE: To evaluate the safety and potential efficacy of aerosolized surfactant in intubated patients with adult respiratory distress syndrome (ARDS). DESIGN: A prospective, double-blind, placebo-controlled, randomized, parallel, multicenter pilot clinical trial. PATIENTS: A total of 51 patients with sepsis-induced ARDS were entered into the study within 18 hours of developing sepsis or sepsis syndrome. INTERVENTION: Patients were randomized into four treatment groups in a 2:1:2:1 ratio, as follows: 12 hours of surfactant per day, 12 hours of 0.6% saline per day, 24 hours of surfactant per day, and 24 hours of 0.6% saline per day. Surfactant or saline was aerosolized continuously for up to 5 days using an in-line nebulizer that aerosolized only during inspiration. MAIN OUTCOME MEASURES: Ventilatory data, arterial blood gases, and hemodynamic parameters were measured at baseline, every 4 or 8 hours during the 5 days of treatment, 24 hours after treatment, and 30 days after treatment, at which time mortality was also assessed. Safety was evaluated throughout the 30 days of the study. RESULTS: Surfactant was administered safely in ventilated patients when given continuously throughout the 5 days using the nebulizer system. Although there were no differences in any physiological parameters between the treatment groups, there was a dose-dependent trend in reduction of mortality from 47% in the combined placebo group to 41% and 35% in the groups treated with 12 hours and 24 hours of surfactant per day, respectively. CONCLUSIONS: Aerosolized surfactant was well tolerated when administered on a continuous basis for up to 5 days; however, at the doses given, it did not result in significant improvements in patients with sepsis-induced ARDS.

Other uses of surfactant.

Surfactant is found not only in the alveoli but in bronchioles and small airways. Along with its important role in surface-tension reduction in alveoli, surfactant has several other properties. Any condition characterized by mucus abnormality, mucociliary transport deficiency, airway obstruction, or bronchoalveolar collapse could potentially benefit from exogenous surfactant therapy. Anti-inflammatory and bactericidal properties of surfactant could provide additional benefit in a variety of diseases.

Passive smoking and middle ear effusion among children in day care.

One hundred thirty-two children who attended a research day-care center were studied to determine whether passive tobacco smoke exposure was associated with an increased rate of otitis media with effusion or with an increased number of days with otitis media with effusion during the first 3 years of life. Based on preliminary studies, a serum cotinine concentration of greater than or equal to 2.5 ng/mL was considered indicative of exposure to tobacco smoke. Otitis media with effusion was diagnosed using pneumatic otoscopy by nurse practitioners and pediatricians who reviewed the children's health status each weekday. The 87 children with serum cotinine concentrations greater than or equal to 2.5 ng/mL had a 38% higher rate of new episodes of otitis media with effusion during the first 3 years of life than the 45 children with lower or undetectable serum cotinine concentrations (incidence density ratio = 1.38, 95% confidence interval 1.21 to 1.56). The average duration of an episode of otitis media with effusion was 28 days in the children with elevated cotinine concentrations and 19 days in the children with lower cotinine concentrations (P less than .01). It is estimated that 8% of the cases of otitis media with effusion in this population and 17.6% of the days with otitis media with effusion may be attributable to exposure to tobacco smoke.

Quality of well-being before and after antibiotic treatment of pulmonary exacerbation in patients with cystic fibrosis.

General quality of life has only recently been measured with an objective tool in patients with cystic fibrosis (CF), and there have been no reported attempts to document changes in patients' overall well-being over time, as patients deteriorate or respond to intervention. We applied the Quality of Well-Being scale (QWB) in 28 patients with CF before and after a two-week course of oral ciprofloxacin used to treat pulmonary exacerbations. There were significant correlations between changes in QWB and various pulmonary function test results; QWB vs FEV1: r = 0.4, p less than 0.03; QWB vs FVC: r = 0.5, p less than 0.01; and QWB vs SaO2: r = 0.4, p less than 0.05. Thus, the QWB can track changes in general well-being in CF patients over a brief time and detect changes associated with pulmonary exacerbation and its treatment.

Pulmonary function testing reference values and interpretations in pediatric training programs.

A questionnaire was sent to all pediatric training programs to evaluate the use of pulmonary function reference standards and the interpretation of pulmonary function test results. Responses were obtained from 107 of 130 institutions, and 94 of these had pulmonary function laboratories available. Of the 94, 60 used one of three reference standards. The primary reason the reference standards were chosen was either unknown or because they came with the spirometer (24), were recommended by another person or were those used in that person's training (34), or were thought to be the best standards available or most applicable to the population to be tested (31). To define abnormality, most used an 80% predicted cutoff for forced vital capacity, forced expiratory volume in 1 second, and forced expiratory flow at 25% to 75% vital capacity. For a change in an individual through time, most used a 10% change for forced vital capacity, forced expiratory volume in 1 second, and forced expiratory flow at 25% to 75% vital capacity. Thirteen used statistical methods to define abnormal individuals and none used statistical methods to define a significant change over time. Although there are a few guidelines for reference standards and interpretations of pulmonary function tests, it appears that most laboratories are not using those guidelines and that further guidelines and education are needed.

Negative clinical trials in cystic fibrosis research.

The statistical power of 61 negative clinical trials of therapeutic regimens in patients with cystic fibrosis published from 1977 through 1988 was reviewed and the ability of the investigations to detect small, medium, and large standardized differences was calculated. Small, medium, and large standardized differences were defined as ratios of 0.2, 0.5, and 0.8, respectively, of the observed difference compared with the standard deviation. The average numbers (+/- SD) of patients in the treatment and control groups were 14.3 +/- 6.9 and 14.5 +/- 7.9, respectively. None of the studies had 80% power to detect a small or medium standardized difference and only 4 of the reports had 80% statistical power to detect a large standardized difference. The variability of cystic fibrosis causes a decrease in the standardized difference, making it more difficult to demonstrate statistical significance. Statistical power of negative clinical trials reported in the literature deserves more attention from investigators as well as physicians who treat patients with cystic fibrosis.

Longitudinal response of pulmonary function to bronchodilators in cystic fibrosis.

Previous studies have found that between 0 and 95% of patients with cystic fibrosis (CF) have a significant response to bronchodilators. These studies have been limited by small numbers and the measurement of response at one point in time. We analyzed the response to bronchodilators of patients with CF in a longitudinal and cross-sectional manner using pulmonary function data from 1980 to 1988. Overall, the proportion of patients with a positive response to bronchodilators was relatively large but not consistent over time. Of 573 tests in 127 persons, a positive response occurred in 68 tests of 51 patients. A negative response occurred in 19 tests of 17 patients. Only nine patients had a positive response in more than one third of their tests. The cross-sectional analysis showed variability similar to previous cross-sectional studies. Although a large proportion of patients with CF had a response to bronchodilators, the response was not consistent and may have been related to the number of tests performed. Continued longitudinal testing is necessary for valid decisions for bronchodilator use and for documenting the length, variability, and clinical significance of these responses.

Noncomparability of cross-sectional and longitudinal estimates of lung growth in children.

Results from longitudinal and cross-sectional studies of pulmonary function are often compared. However, previous studies in adults suggest that results from longitudinal and cross-sectional studies are different and may not be comparable. In order to evaluate these differences further, prospectively collected data in a group of children were analyzed by both longitudinal and cross-sectional methods. Spirometry was performed longitudinally over a period of 8 years on 58 healthy children. Straight-line regressions of expiratory flow-volume parameters on height were computed by averaging the individual regression lines for each child. A cross-sectional sample from these same children was analyzed and compared to the longitudinal analysis. For all expiratory flow-volume parameters, the cross-sectional analysis resulted in a significantly greater increase in growth with increasing height than the longitudinal analysis (P less than 0.005 for slope for all parameters except PEFR, P less than 0.05 and Vmax75, P less than 0.01). These differences cannot be explained by learning or horse-racing effects, loss to follow-up, or regression to the mean; however, they could be explained by cohort changes, time trends of pulmonary function, differences in the weighted averages used in the two analyses, or incorrect modeling. These observations indicate that comparisons between cross-sectional and longitudinal investigations must be made cautiously. These data do not support many of the reasons suggested for the differences in previous reports. It is recommended that studies of pulmonary function use a control group and investigate the groups at the same time and in the same manner.

Use of bronchoalveolar lavage in immunocompromised children with pneumonia.

Because pneumonia in the immunocompromised host presents a diagnostic dilemma to the pediatrician, we prospectively studied the use of bronchoalveolar lavage (BAL) with flexible fiberoptic bronchoscopy in 14 immunocompromised children with pneumonia over a 5-month period. The children received immunosuppressive agents because of organ transplants, cancer, or systemic lupus erythematosus. A diagnosis was made by BAL in 10 (71%) of the 14 children. Pathogenic organisms included Pneumocystis carinii in six, cytomegalovirus (CMV) in two, Aspergillus fumigatus in one, and a mixed infection of Candida albicans and CMV in one. There were no complications. Because BAL by flexible fiber-optic bronchoscopy is a rapid and safe procedure and a diagnosis can be made in a majority of cases, we recommend its use in immunocompromised children with pneumonia before resorting to lung biopsy.

Pulmonary function in children with sarcoidosis.

To document pulmonary function abnormalities in children with sarcoidosis, we reviewed the records of all children with this disease at the North Carolina Memorial Hospital. Spirometry was performed by 34 of 60 children at initial presentation and was repeated by 16 at least 1 yr after presentation (mean, 3.9 yr). Those with and without pulmonary function data available were similar in age, sex, race, and clinical manifestations at presentation. Results indicate that at presentation, 50% of children had the characteristic functional changes of restrictive lung disease (mean % predicted FVC, 79.8 +/- 16.0). Fifteen percent had a FVC of 70 to 80% predicted, 32% had a FVC of 50 to 70% predicted, and 3% had a FVC less than 50% predicted. The TLC and FRC supported the diagnosis of restrictive lung disease. There were 15% who had obstructive changes. Children with auscultatory abnormalities and parenchymal changes on chest radiograph had more severe decrements in pulmonary function; however, specific symptoms did not predict decrements in pulmonary function. There was significant improvement in lung function at the last pulmonary function test performed (mean % predicted FVC, 97.1 +/- 18.8).

Serum cotinine as a measure of tobacco smoke exposure in children.

To document passive smoke exposure, we measured concentrations of serum cotinine, a major metabolite of nicotine, in 38 young children and compared the results with the smoking histories of home residents. Cotinine was detected in 26 children (68%), of which ten had no household exposure according to a questionnaire. The serum cotinine concentration was significantly elevated in blacks compared with whites after controlling for the number of smokers in the home. After stratifying by race, there was a significant direct correlation between the serum cotinine concentration and the number of smokers in the home, the amount smoked by the mother, and the amount smoked by others in the home. We conclude that the serum cotinine concentration is a useful indicator of the actual exposure of young children to tobacco smoke and that unexplained racial differences in cotinine levels exist.