Serology as a diagnostic tool for predicting initialPseudomonas aeruginosa acquisition in children with cystic fibrosis.

JCF-01070; No of Pages 8

Journal of Cystic Fibrosis xx (2014) xxx – xxx www.elsevier.com/locate/jcf

Original Article

Serology as a diagnostic tool for predicting initial Pseudomonas aeruginosa acquisition in children with cystic ﬁbrosis☆,☆☆ Cori Daines a,⁎, Donald VanDeVanter b , Umer Khan c , Julia Emerson d , Sonya Heltshe d , Sharon McNamara d , Michael Anstead e , Markus Langkamp f , Gerd Doring g,1 , Felix Ratjen h , Bonnie Ramsey d , Ronald L. Gibson d , Wayne Morgan a , Margaret Rosenfeld d for the EPIC Investigators a University of Arizona, Tucson, AZ 85724, United States Case Western Reserve University School of Medicine, Cleveland, OH, United States c Seattle Children's Research Institute, Seattle, WA 98121, United States d Seattle Children's Hospital, University of Washington School of Medicine, Seattle, WA 98105-0371, United States e Department of Pediatrics, University of Kentucky, Lexington, KY 40563-0284, United States f Mediagnost®, Aspenhaustr. 25, 72770 Reutlingen, Germany g Institute of Medical Microbiology and Hygiene, University of Tübingen, Tübingen, Germany Division of Respiratory Medicine, Department of Pediatrics, The Hospital for Sick Children, University of Toronto, Toronto, Canada b

Received 24 April 2014; received in revised form 14 June 2014; accepted 16 June 2014

Abstract Rationale: Pseudomonas aeruginosa (Pa) serology could potentially be a useful adjunct to respiratory culture methods for the detection of initial or early Pa infection in patients with cystic ﬁbrosis (CF). Objective: To evaluate the utility of Pa serology to predict Pa isolation from respiratory (generally oropharyngeal) cultures in the subsequent 6 or 12 months among young children with CF from whom Pa had never been previously cultured. Pa serology was also evaluated in a group of healthy controls. Methods: Children ≤ 12 years of age without prior isolation of Pa from respiratory cultures participating in the Early Pseudomonal Infection Control EPIC Observational Study (EPIC OBS) had annual serum samples for measurement of antibodies against alkaline protease, elastase and exotoxin A using a commercial kit; controls had a single serum sample. Logistic regression with generalized estimating equations was used to characterize associations between log10 serum antibody titers and ﬁrst isolation of Pa from a respiratory culture within the subsequent 6 or 12 months, with adjustment for sex and age. Receiver operating characteristic curves were used to optimize antibody titer cutpoints by age group. The diagnostic properties of each antibody were estimated using these optimized cutpoints. Results: Pa serology was evaluated in 582 children with CF (2084 serum samples) and 94 healthy controls. There was substantial overlap between serum antibody titers among controls, CF patients who did not acquire Pa (N = 261) and CF patients who did acquire Pa (N = 321). The maximum positive predictive value for ﬁrst Pa positive culture within the ensuing 6 months was 76.2% and maximum negative predictive value was 72.1% for any antigen or combination of antigens; values were similar for 12 months.

Abbreviations: CF, cystic fibrosis; Pa, Pseudomonas aeruginosa; OP, oropharyngeal; BAL, bronchoalveolar lavage; EPIC OBS, Early Pseudomonas Infection Control Observational Study; CFFNPR, CFF National Patient Registry; GEE, generalized estimating equations; ROC, receiver operating characteristic; AUC, area under the curve; CFTR, cystic fibrosis transmembrane regulator; PPV, positive predictive value; NPV, negative predictive value. ☆ Financial support: Cystic Fibrosis Foundation grants OBSERV04K0 and EPIC0K0 to M. Rosenfeld, Seattle Children's Hospital, Seattle, WA; Nemours Children's Clinic Research Program, Jacksonville, FL. ☆☆ Prior presentation: Portions of this manuscript were presented as an oral and poster presentation at the 2012 North American Cystic Fibrosis Conference. ⁎ Corresponding author at: University of Arizona Medical Center, 1501 N. Campbell Ave., PO Box 245073, Tucson, AZ 85724, USA. Tel.: +1 520 626 7780. E-mail address: [emailprotected] (C. Daines). 1 Posthumous.

http://dx.doi.org/10.1016/j.jcf.2014.06.005 1569-1993/© 2014 European Cystic Fibrosis Society. Published by Elsevier B.V. All rights reserved. Please cite this article as: Daines C, et al, Serology as a diagnostic tool for predicting initial Pseudomonas aeruginosa acquisition in children with cystic ﬁbrosis, J Cyst Fibros (2014), http://dx.doi.org/10.1016/j.jcf.2014.06.005

C. Daines et al. / Journal of Cystic Fibrosis xx (2014) xxx–xxx

Conclusions: Pa serology does not appear useful for predicting ﬁrst Pa positive oropharyngeal culture among young CF patients. © 2014 European Cystic Fibrosis Society. Published by Elsevier B.V. All rights reserved. Keywords: Cystic ﬁbrosis; Pseudomonas; Serology; ROC curves

1. Introduction Progressive obstructive lung disease due to chronic airway infection and inflammation is the leading cause of morbidity and mortality in cystic fibrosis (CF), with the bacterial pathogen Pseudomonas aeruginosa (Pa) playing a prominent role [1]. Initial Pa acquisition generally occurs in early childhood [2,3], but is often transient [2] and may be limited to the upper airways [4]. In contrast, approximately 80% of CF adults has chronic airway Pa infection [5], which is associated with more rapid lung function decline [3,6], increased morbidity [3,7] and decreased survival [8,9]. Today, children with CF are routinely treated with antipseudomonal antibiotics upon first Pa isolation in an attempt to eradicate the organism [10,11]. Eradication of early Pa infection has a roughly 80% success rate [11–13] and has been shown to reduce the prevalence of chronic Pa infection in CF cohorts compared to historical controls [14–16]. The accurate detection of early Pa infection is problematic, as young children and those with mild lung disease typically do not expectorate sputum. Surveillance respiratory cultures in these patients are typically performed on oropharyngeal (OP) swabs. While OP cultures are known to have imperfect diagnostic accuracy compared to lower airway cultures [4,17], they are nonetheless standard of care in the U.S. and many other countries, and are widely used to guide treatment decisions [18], define stages of Pa infection [19] and predict clinical outcomes [20,21]. Serum titers of antibodies against Pa antigens have been shown to be elevated in chronic Pa infection [22,23] and to distinguish intermittent from chronic Pa colonization [24,25]. However, as most chronically infected patients expectorate sputum, the clinical utility of serology in this context is limited. In contrast, it has been suggested that Pa serology could prove a useful adjunct to respiratory culture methods for the detection of initial or early Pa infection [24,26,27], as serology has the potential advantages of being more accurate than upper airway cultures and less resource-intensive and invasive than BAL. The diagnostic accuracy of Pa serology relative to concurrent respiratory cultures remains controversial [17,28–30]. Importantly, several studies have demonstrated that positive Pa serology may precede initial isolation of Pa from both upper [2,3,25,29,31] and lower [2] airway cultures. If positive Pa serology could predict subsequent isolation of Pa from respiratory cultures, eradication therapy could potentially be initiated at an earlier stage to improve outcomes; this has been advocated [24] but not yet investigated. The Early Pseudomonas Infection Control Observational Study (EPIC OBS) is a U.S. national prospective study to evaluate the risk factors for and clinical outcomes associated with isolation of Pa from respiratory cultures in a large cohort of children with CF who were Pa-culture negative at enrollment

[32]. The objective of the current analysis was to evaluate the utility of Pa serology to predict subsequent Pa isolation from respiratory (generally OP) culture among young children with CF from whom Pa had never been previously cultured. We hypothesized that Pa serology would have acceptable diagnostic accuracy in predicting first isolation of Pa from respiratory cultures within the ensuing 6 or 12 months. As part of the current analysis, we also examined Pa serology and OP cultures in a cohort of children without CF undergoing elective surgical procedures at a single institution to assess the levels of anti-Pa antibodies in the unaffected population. Portions of this work have previously been published in abstract form. 2. Methods 2.1. Study participants and samples The design of the EPIC OBS has been reported elsewhere [33,34]. Children with an established diagnosis of CF [35] ≤ 12 years of age were enrolled at 59 accredited U.S. CF care centers between 2004 and 2006. Annual serum samples were collected for serology and banking, and the results of clinical respiratory cultures were recorded in the CFF National Patient Registry (CFFNPR). Eligibility criteria for participation in the current analysis were 1) no prior isolation of Pa from respiratory cultures since CF diagnosis, confirmed with CFFNPR data, 2) no loss to follow up or isolation of Pa from a respiratory culture in the first 120 days after enrollment [36] (as these individuals may have had had Pa infection prior to enrollment), and 3) at least one serum sample collected. Written informed consent was obtained from the family of each participant and the study was approved by the Institutional Review Board at each participating site. Serum samples collected through 2009 and data collected through 2010 were included in the current analysis. 2.2. Non-CF controls Otherwise healthy children ≤ 18 years of age undergoing a clinically indicated procedure that required sedation or anesthesia at Seattle Children's Hospital, Seattle, WA, USA between September 2008 and February 2010 were recruited. Exclusion criteria included: (1) presence of indwelling catheters or devices (including myringotomy tubes) at enrollment or within the past year; (2) oral or IV antibiotic treatment within the past month; (3) presence of congenital or acquired immunosuppression; (4) history of cancer; (5) currently undergoing an otolaryngology or dental procedure; (6) immediate family member with CF; (7) blood transfusion within the past year. A serum sample for serology and an OP swab for culture were collected from each participant. The study was approved by the Seattle Children's

Please cite this article as: Daines C, et al, Serology as a diagnostic tool for predicting initial Pseudomonas aeruginosa acquisition in children with cystic ﬁbrosis, J Cyst Fibros (2014), http://dx.doi.org/10.1016/j.jcf.2014.06.005

C. Daines et al. / Journal of Cystic Fibrosis xx (2014) xxx–xxx

Hospital IRB and informed consent was obtained from all parents/guardians, as well as assent from participants as applicable. The respiratory culture results from this cohort have been previously published [34]. 2.3. Pseudomonas serology Serum samples were analyzed for titers of antibodies to the Pa antigens alkaline protease, exotoxin A, and elastase by Mediagnost® (Reutlingen, Germany) using their commercially available IgG enzyme immunoassay system [24,28,37]. Assays were batched and performed in duplicate. The lower limits of quantification were 0.41, 0.15 and 0.35 titer/ml for alkaline protease, exotoxin A and elastase, respectively (equal to 10 times the standard deviation of the blank). Intraassay variances were 4.39, 5.3 and 11.5% of the coefficient of variation (CV) and interassay variances were 5.7, 7.7 and 8.0% of the CV for alkaline protease, elastase and exotoxin A, respectively. The linearity of sample dilution has been proven for dilution range of 1:500 to 1:24,000 by the manufacturer. An independent evaluation of the test system has been conducted [24,28]. 2.4. Statistical analyses Logistic regression models were used to characterize associations between log10 serum antibody titers and first isolation of Pa from a respiratory culture within the subsequent 6 or 12 months, with generalized estimating equations (GEE) methods used to account for repeated observations per patient, and adjustment for sex, age (1–≤ 3 years, N 3–≤ 6 years, N 6 years) and time (days) between serum sample and respiratory culture. Receiver operating characteristic (ROC) curves were used to optimize antibody titer cutpoints to maximize the sensitivity and specificity of each antibody and by age group (0–b 6 years, ≥ 6 years), using the area under the curve (AUC). The diagnostic properties of each serologic assay, including sensitivity, specificity, positive predictive value, and negative predictive value were estimated based on these optimized cutpoints. Similar analyses were performed for concurrent respiratory cultures (defined as collection within 3 weeks of the serum collection date) for comparison purposes. Analyses were performed using R version 2.15.1 (R Foundation for Statistical Computing, Vienna, Austria) and SAS, version 9.2 (SAS Institute Inc., Cary, NC). 3. Results A total of 1797 children with CF were enrolled in EPIC OBS between 2004 and 2006, of which 687 had no isolation of Pa from respiratory cultures prior to or during the first 120 days after enrollment. Of these 687 children, 582 had at least one serum sample collected during the observation period and therefore comprise the cohort for the current analyses. These 582 children contributed 2084 serum samples; 261 children (44.8%) had a first Pa-positive respiratory culture during the observation period and 321 remained Pa-negative. At baseline (Table 1), participants who subsequently acquired Pa were similar to those who remained Pa negative with respect to age and sex, but the proportion of children

Table 1 Baseline characteristics of children with CF.

Age, years mean (SD) Age distribution 0–3 years, n (%) N 3–6 years, n (%) N 6–12 years, n (%) Male, n (%) CFTR functional class a High risk Low risk Not classified Missing Pancreatic sufficient b

Did not acquire Pa (N = 321)

Acquired Pa (N = 261)

Total (N = 582)

5.0 (3.5)

5.5 (3.4)

4.8 (3.4)

114 (35.5%) 84 (26.2%) 123 (38.3%) 157 (48.9%)

106 (40.6%) 66 (25.3%) 89 (34.1%) 127 (48.7%)

220 (37.8%) 150 (25.8%) 212 (36.4%) 284 (48.8%)

210 (65.4%) 53 (16.5%) 51 (15.9%) 7 (2.2%) 73 (22.7%)

210 (85%) 14 (5.4%) 26 (10%) 11 (4.2%) 23 (8.8%)

420 (72.2%) 67 (11.5%) 77 (13.2%) 18 (3.1%) 96 (16.5%)

a High risk: both alleles with mutations in functional class 1, 2 or 3; low risk: at least one allele with a class 4 or 5 mutation; not classified: functional class not able to be determined based on mutations detected [46]. b Defined by pancreatic enzyme replacement therapy use in the CFF Registry.

with a high risk CFTR genotype was higher in participants who acquired Pa (85.0% versus 65.4%, P b .0001), as was the prevalence of pancreatic insufficiency (91.2% versus 77.3%, P b .0001). A serum sample for serology and a simultaneous OP swab for culture were collected from 94 non-CF controls, who had a mean (SD) age of 8.5 (5.4) years; 54 were male. Pa was isolated from only 1 of the 94 OP cultures. 3.1. Serum antibody titers and association with concurrent or subsequent Pa isolation from respiratory cultures Among the 261 children with CF who acquired Pa during the observation period, 169 (64.8%) had serum samples collected within 6 months prior to Pa isolation and all had a serum sample within 12 months prior to Pa isolation. A total of 68 (26.1%) had serum samples collected concurrently with Pa isolation, with a median elapsed time between the recorded date of serum collection and respiratory sample collection of 0.0 day (mean = 0.94 day, SD = 3.86 days) for this group. Alkaline protease titers tended to be higher in CF patients than in controls, though these titers did not distinguish between CF patients who remained Pa negative and those who acquired Pa (Fig. 1). Exotoxin A titers tended to be highest in CF patients who were concurrently Pa positive. In general, however, there was substantial overlap between titers of all three antigens across participant groups (Fig. 1). Among the non-CF control children, 40%, 59% and 36%, respectively, had antibody titers below the limit of quantitation and 24%, 29%, and 16% had antibody titers ≥ 100 against alkaline protease, exotoxin A and elastase, respectively. Among 94 serum samples randomly selected from 94 age-matched CF children, 33%, 31% and 53%, respectively, had antibody titers below the limit of quantitation and 27%, 28%, and 21% had antibody titers ≥ 100 against alkaline protease, exotoxin A and elastase, respectively. In logistic regression models, serum exotoxin A and elastase antibody titers, but not alkaline protease antibody titers, were

Serology as a diagnostic tool for predicting initialPseudomonas aeruginosa acquisition in children with cystic fibrosis. - PDF Download Free (2024)

References