Advertisement

A longitudinal analysis of chronic MRSA and Pseudomonas aeruginosa co-infection in cystic fibrosis: A single-center study

Open ArchivePublished:November 20, 2015DOI:https://doi.org/10.1016/j.jcf.2015.10.014

      Abstract

      Background

      Few studies have examined the association between chronic methicillin-resistant Staphylococcus aureus (MRSA) and Pseudomonas aeruginosa (PA) co-infection and health outcomes despite evidence that these pathogens alone contribute to higher morbidity and mortality in cystic fibrosis (CF). This study examines outcomes among CF patients with chronic MRSA and PA co-infection compared with patients with either or neither of these organisms.

      Methods

      CF patients attending the care center in Atlanta, GA from 2007–2013 comprised the study cohort. Chronic co-infection was defined as >50% PA+ cultures and >50% MRSA+ cultures and modeled as time-varying. The rate of decline in lung function (FEV1) and the rate of IV treatments were the main outcomes.

      Results

      Among all patients (N = 354), chronic co-infection was associated with a significantly more rapid rate of FEV1 decline compared with patients with chronic PA alone [adjusted difference: −0.60% predicted/year (−1.13, −0.08)] and chronic MRSA alone [adjusted difference: −0.89% predicted/year (−1.56, −0.22)]. Rate of IV treatments was significantly higher among patients with chronic co-infection compared with patients with chronic PA alone [adjusted IRR: 1.24 (1.01, 1.52)] and chronic MRSA alone [adjusted IRR: 1.34 (1.03, 1.74)].

      Conclusions

      Data from the Atlanta Care Center suggest that chronic MRSA and PA co-infection may be associated with increased rate of lung function decline and rate of intravenous antibiotics compared with patients with either pathogen alone.

      Keywords

      1. Introduction

      Respiratory disease in cystic fibrosis (CF) is characterized by the triad of decreased mucociliary clearance, chronic bacterial infection, and persistent inflammation ultimately resulting in respiratory failure [
      • Flume P.A.
      Pulmonary complications of cystic fibrosis.
      ]. Staphylococcus aureus (SA) and Pseudomonas aeruginosa (PA) are the two most common pathogens to chronically infect the CF airway, often co-existing [
      Patient Registry annual data report 2012.
      ]. Evidence suggests that co-infection with SA, regardless of resistance, and PA promotes the selection of more virulent SA small-colony variants and may have an additive inflammatory effect [
      • Hoffman L.R.
      • Deziel E.
      • D'Argenio D.A.
      • Lepine F.
      • Emerson J.
      • McNamara S.
      • et al.
      Selection for Staphylococcus aureus small-colony variants due to growth in the presence of Pseudomonas aeruginosa.
      ,
      • Mitchell G.
      • Seguin D.L.
      • Asselin A.E.
      • Deziel E.
      • Cantin A.M.
      • Frost E.H.
      • et al.
      Staphylococcus aureus sigma B-dependent emergence of small-colony variants and biofilm production following exposure to Pseudomonas aeruginosa 4-hydroxy-2-heptylquinoline-N-oxide.
      ,
      • Sagel S.D.
      • Gibson R.L.
      • Emerson J.
      • McNamara S.
      • Burns J.L.
      • Wagener J.S.
      • et al.
      Impact of Pseudomonas and Staphylococcus infection on inflammation and clinical status in young children with cystic fibrosis.
      ]. Similar clinical studies have shown co-infection compared with infection with one or none of these pathogens to be associated with lower survival among young children with CF and rapidly declining lung function among adults [
      • Hudson V.L.
      • Wielinski C.L.
      • Regelmann W.E.
      Prognostic implications of initial oropharyngeal bacterial flora in patients with cystic fibrosis diagnosed before the age of two years.
      ,
      • Rosenbluth D.B.
      • Wilson K.
      • Ferkol T.
      • Schuster D.P.
      Lung function decline in cystic fibrosis patients and timing for lung transplantation referral.
      ].
      There has been a dramatic increase in the prevalence of methicillin-resistant S. aureus (MRSA) over the past decade from 9% in 2002 to 26% in 2012 among patients in the CF Foundation (CFF) Patient Registry and this may be associated with significantly poorer outcomes. For example, the risk of death for patients with chronic MRSA infection has been found to be two-fold higher than that of patients without MRSA infection [
      • Dasenbrook E.C.
      • Checkley W.
      • Merlo C.A.
      • Konstan M.W.
      • Lechtzin N.
      • Boyle M.P.
      Association between respiratory tract methicillin-resistant Staphylococcus aureus and survival in cystic fibrosis.
      ]. Evidence regarding lung function decline has been mixed with some studies showing chronic MRSA infection to be associated with more rapid decline [
      • Cox D.W.
      • Kelly C.
      • Rush R.
      • O'Sullivan N.
      • Canny G.
      • Linnane B.
      The impact of MRSA infection in the airways of children with cystic fibrosis; a case–control study.
      ,
      • Dasenbrook E.C.
      • Merlo C.A.
      • Diener-West M.
      • Lechtzin N.
      • Boyle M.P.
      Persistent methicillin-resistant Staphylococcus aureus and rate of FEV1 decline in cystic fibrosis.
      ,
      • Vanderhelst E.
      • De Meirleir L.
      • Verbanck S.
      • Pierard D.
      • Vincken W.
      • Malfroot A.
      Prevalence and impact on FEV(1) decline of chronic methicillin-resistant Staphylococcus aureus (MRSA) colonization in patients with cystic fibrosis. A single-center, case control study of 165 patients.
      ] and others not [
      • Hubert D.
      • Reglier-Poupet H.
      • Sermet-Gaudelus I.
      • Ferroni A.
      • Le Bourgeois M.
      • Burgel P.R.
      • et al.
      Association between Staphylococcus aureus alone or combined with Pseudomonas aeruginosa and the clinical condition of patients with cystic fibrosis.
      ,
      • Muhlebach M.S.
      • Miller M.
      • LaVange L.M.
      • Mayhew G.
      • Goodrich J.S.
      • Miller M.B.
      Treatment intensity and characteristics of MRSA infection in CF.
      ,
      • Sawicki G.S.
      • Rasouliyan L.
      • Ren C.L.
      The impact of MRSA on lung function in patients with cystic fibrosis.
      ]. With this increase in the prevalence of MRSA, the question arises whether co-infection with MRSA and PA has worse health outcomes than infection with either alone. In a two-year study in France, MRSA and PA co-infection was associated with greater lung function decline compared with patients without any SA or PA infection [
      • Hubert D.
      • Reglier-Poupet H.
      • Sermet-Gaudelus I.
      • Ferroni A.
      • Le Bourgeois M.
      • Burgel P.R.
      • et al.
      Association between Staphylococcus aureus alone or combined with Pseudomonas aeruginosa and the clinical condition of patients with cystic fibrosis.
      ]. In addition, co-infection was associated with more need for intravenous (IV) antibiotics compared with patients with MRSA alone or with methicillin-sensitive SA (MSSA) and PA co-infection, suggesting that an interaction may occur between MRSA and PA [
      • Hubert D.
      • Reglier-Poupet H.
      • Sermet-Gaudelus I.
      • Ferroni A.
      • Le Bourgeois M.
      • Burgel P.R.
      • et al.
      Association between Staphylococcus aureus alone or combined with Pseudomonas aeruginosa and the clinical condition of patients with cystic fibrosis.
      ]. Finally, a recent study among young CF children found that patients with low lung function at initial acceptable spirometry had significantly more positive cultures for both MRSA and PA compared with patients with high lung function [
      • Com G.
      • Carroll J.L.
      • Castro M.M.
      • Tang X.
      • Jambhekar S.
      • Berlinski A.
      Predictors and outcome of low initial forced expiratory volume in 1 second measurement in children with cystic fibrosis.
      ]. However, neither of these studies specifically examined the effect of chronic co-infection. Thus, there is little understanding on the association of chronic MRSA and PA co-infection with health outcomes in CF, particularly in the United States where the prevalence of MRSA is the highest in the world [
      • Goss C.H.
      • Muhlebach M.S.
      Review: Staphylococcus aureus and MRSA in cystic fibrosis.
      ]. This study aims to investigate the association between chronic co-infection with MRSA and PA in CF and (i) rate of lung function decline and (ii) frequency of IV antibiotics compared with infection with either or neither. Some of the results of this study have been previously reported in the form of an abstract [
      • Maliniak M.L.
      • McCarty N.A.
      • Stecenko A.A.
      Chronic methicillin-resistant Staphyloccus aureus and Pseudomonas aeruginosa co-infection and lung function decline In cystic fibrosis [abstract].
      ].

      2. Methods

      This was a retrospective cohort study utilizing data from the patient registry affiliated with Emory University and Children's Healthcare of Atlanta, Inc., between January 1, 2007 and December 31, 2013. The study was approved by the Emory University Institutional Review Board (IRB) [IRB00070522].

      2.1 Patients

      Patients were included if they were ≥6 years of age, did not receive a lung transplant before or during the study period, were followed for at least 2 years, had ≥4 respiratory cultures at baseline to establish chronicity of infection, one spirometry measurement at baseline, and ≥3 spirometry measurements spanning at least 2 years during follow-up to estimate lung function decline [
      • Proesmans M.
      • Balinska-Miskiewicz W.
      • Dupont L.
      • Bossuyt X.
      • Verhaegen J.
      • Hoiby N.
      • et al.
      Evaluating the “Leeds criteria” for Pseudomonas aeruginosa infection in a cystic fibrosis centre.
      ]. A one-year lead-in period was established to determine baseline MRSA and PA infection status and confounders.

      2.2 Infection groups

      Infection status was determined from respiratory tract cultures obtained from sputum (70%), throat/nasal culture (29%), or bronchoscopy (1%). PA infection was classified using a modification of the Leed's criteria [
      • Proesmans M.
      • Balinska-Miskiewicz W.
      • Dupont L.
      • Bossuyt X.
      • Verhaegen J.
      • Hoiby N.
      • et al.
      Evaluating the “Leeds criteria” for Pseudomonas aeruginosa infection in a cystic fibrosis centre.
      ,
      • Lee T.W.
      • Brownlee K.G.
      • Conway S.P.
      • Denton M.
      • Littlewood J.M.
      Evaluation of a new definition for chronic Pseudomonas aeruginosa infection in cystic fibrosis patients.
      ,
      • Pressler T.
      • Bohmova C.
      • Conway S.
      • Dumcius S.
      • Hjelte L.
      • Hoiby N.
      • et al.
      Chronic Pseudomonas aeruginosa infection definition: EuroCareCF Working Group report.
      ]. MRSA was considered chronic if detected in ≥3 respiratory cultures as has been previously defined [
      • Dasenbrook E.C.
      • Merlo C.A.
      • Diener-West M.
      • Lechtzin N.
      • Boyle M.P.
      Persistent methicillin-resistant Staphylococcus aureus and rate of FEV1 decline in cystic fibrosis.
      ,
      • Vanderhelst E.
      • De Meirleir L.
      • Verbanck S.
      • Pierard D.
      • Vincken W.
      • Malfroot A.
      Prevalence and impact on FEV(1) decline of chronic methicillin-resistant Staphylococcus aureus (MRSA) colonization in patients with cystic fibrosis. A single-center, case control study of 165 patients.
      ,
      • Muhlebach M.S.
      • Miller M.
      • LaVange L.M.
      • Mayhew G.
      • Goodrich J.S.
      • Miller M.B.
      Treatment intensity and characteristics of MRSA infection in CF.
      ]. Thus, infection status at baseline was classified as chronic co-infection (>50% PA+ and ≥3 MRSA+ cultures), chronic PA alone (>50% PA+ and <3 MRSA+ cultures), chronic MRSA alone (≤50% PA+ and ≥3 MRSA+ cultures), intermittent infection (≤50% PA+ and <3 MRSA+ cultures but ≥1 MRSA+/PA+), and no MRSA/PA (0 MRSA+/PA+ cultures). We also examined infection status as a time-dependent variable allowing infection status to change each year. Because patients may not have had ≥4 respiratory cultures per year after baseline, we re-defined MRSA infection during follow-up as a percentage of respiratory cultures using the same modification of the Leed's criteria used for PA [
      • Lee T.W.
      • Brownlee K.G.
      • Conway S.P.
      • Denton M.
      • Littlewood J.M.
      Evaluation of a new definition for chronic Pseudomonas aeruginosa infection in cystic fibrosis patients.
      ]. If a patient had 0 respiratory cultures in a given year, the infection status of the previous year was used.

      2.3 Statistical analysis

      Descriptive statistics were calculated to summarize patient characteristics. Continuous variables were compared using analysis of variance or Kruskal–Wallis tests as appropriate. Categorical variables were compared using chi-square tests. Post-hoc analyses using the Tukey–Kramer adjustment or Kruskal–Wallis tests with multiple average comparisons were conducted to clarify significant differences between infection groups [
      • Elliott A.C.
      • Hynan L.S.
      A SAS((R)) macro implementation of a multiple comparison post hoc test for a Kruskal–Wallis analysis.
      ].
      The primary physiologic outcome of interest was annual rate of decline in lung function. Lung function was measured by forced expiratory volume in the first second (FEV1, given as % predicted), calculated using the equations of Hankinson et al. and Wang et al. [
      • Hankinson J.L.
      • Odencrantz J.R.
      • Fedan K.B.
      Spirometric reference values from a sample of the general U.S. population.
      ,
      • Wang X.
      • Dockery D.W.
      • Wypij D.
      • Fay M.E.
      • Ferris Jr., B.G.
      Pulmonary function between 6 and 18 years of age.
      ]. In the Atlanta clinic, baseline lung function is defined as the average of the two highest FEV1 values in a given year and annual rate of decline was calculated using these baseline values. Repeated measures mixed effects models with a random slope and intercept were used to determine the effect of infection status on rate of decline in FEV1% predicted, adjusting for known confounders. Confounders were assessed during the baseline year and included age, sex, genotype, baseline FEV1% predicted, number of episodes requiring IV antibiotics, body mass index (BMI) for adults, BMI percentile for pediatric patients, positive respiratory tract cultures for other pathogens (MSSA, Burkholderia cepacia complex [B. cepacia], and Haemophilus influenza [H. flu]), pancreatic insufficiency, and CF-related diabetes (CFRD). See the online data supplement for definitions of confounders and method for inclusion in the final model.
      The primary clinical outcome of interest was frequency of episodes requiring intravenous antibiotics as an indicator of clinician-identified exacerbations. We used multivariate negative binomial regression models to compare the incidence rate (IR) of IV treatments between infection groups, adjusting for baseline age, sex, FEV1% predicted, genotype, CFRD status, and number of IV treatments at baseline.
      All statistical analyses were conducted in SAS Version 9.3 (Cary, NC). Statistical significance was accepted at P < 0.05.

      3. Results

      A total of 354 patients met the inclusion criteria and were included in the analysis. At baseline, 39 (11.0%) patients had chronic MRSA and PA co-infection, 106 (29.9%) had chronic PA alone, 42 (11.9%) had chronic MRSA alone, 65 (18.4%) had intermittent MRSA/PA, and 102 (28.8%) did not culture MRSA or PA. Patients were followed for an average of 4.4 years (SD: 1.6) after the lead-in period with no differences in the average follow-up time between infection groups (P = 0.77). Patients had an average of 5.1 respiratory cultures (SD: 1.2) and 7.6 FEV1 measurements (SD: 3.5) during the baseline year. Baseline characteristics by infection group are presented in Table 1.
      Table 1Patient characteristics by infection status during the one-year lead-in period.
      CharacteristicChronic co-infection

      (N = 39)
      Chronic PA

      (N = 106)
      Chronic MRSA

      (N = 42)
      Intermittent infection

      (N = 65)
      No MRSA/PA

      (N = 102)
      N (%)N (%)N (%)N (%)N (%)P value
      Age, year
      Average (standard deviation).
      19.9 (8.9)21.5 (11.3)12.5 (8.2)12.6 (7.5)11.3 (7.0)<0.001
      Sex, female20 (51.3)54 (50.9)20 (47.6)28 (43.1)41 (40.2)0.54
      FEV1% predicted
      Average (standard deviation).
      Calculated as the average of the two highest values of FEV1 recorded during the one-year lead-in period.
      74.5 (23.4)79.2 (24.8)92.1 (20.2)96.8 (18.9)98.5 (19.3)<0.001
      Homozygous ΔF50822 (56.4)57 (53.8)18 (42.9)33 (50.8)54 (52.9)0.75
      BMI (adult), kg/m2
      Average (standard deviation).
      BMI calculated for patients >20years of age (n=76) and BMI percentile for patients ≤20years of age (n=278).
      22.2 (5.0)20.7 (2.6)22.7 (3.1)23.7 (3.6)22.1 (2.9)0.22
      BMI pctl (pediatric)
      Average (standard deviation).
      BMI calculated for patients >20years of age (n=76) and BMI percentile for patients ≤20years of age (n=278).
      33.0 (27.2)37.0 (27.7)46.5 (26.6)43.9 (24.5)50.7 (26.7)0.007
      Pancreatic insufficient39 (100)103 (97.2)41 (97.6)62 (95.4)93 (91.2)0.12
      CF-related diabetes14 (35.9)33 (31.1)7 (16.7)11 (16.9)6 (5.9)<0.001
      Mucoid PA
      Cultured positive for pathogen ≥1 time during baseline period.
      33 (84.6)103 (97.2)4 (9.5)16 (24.6)0 (0)<0.001
      B. cepacia complex
      Cultured positive for pathogen ≥1 time during baseline period.
      1 (2.6)4 (3.8)1 (2.4)3 (4.6)7 (6.9)0.69
      H. influenzae
      Cultured positive for pathogen ≥1 time during baseline period.
      0 (0)1 (0.9)2 (4.8)5 (7.7)9 (8.8)0.04
      MSSA
      Cultured positive for pathogen ≥1 time during baseline period.
      12 (30.8)48 (45.3)14 (33.3)51 (78.5)90 (88.2)<0.001
      IV antibiotic episodes
      Average (standard deviation).
      2.2 (1.7)1.2 (1.2)1.0 (1.2)0.8 (1.1)0.3 (0.7)<0.001
      Abbreviations: B. cepacia, Burkholderia cepacia; BMI, body mass index; CF, cystic fibrosis; H. influenzae, Haemophilus influenzae; IV, intravenous; MRSA, methicillin-resistant Staphylococcus aureus; PA, Pseudomonas aeruginosa; pctl, percentile; MSSA, methicillin-sensitive Staphylococcus aureus; SD, standard deviation.
      a Average (standard deviation).
      b Calculated as the average of the two highest values of FEV1 recorded during the one-year lead-in period.
      c BMI calculated for patients >20 years of age (n = 76) and BMI percentile for patients ≤20 years of age (n = 278).
      d Cultured positive for pathogen ≥1 time during baseline period.
      Infection groups were significantly different in age (P < 0.001), FEV1% predicted (P < 0.001), MSSA status (P < 0.001), H. flu status (P = 0.04), CFRD status (P < 0.001), and number of IV antibiotic treatments (P < 0.001) at baseline. Pediatric patients (≤20 years) were significantly different in BMI percentile based on infection status (P = 0.007) with chronically co-infected patients having the lowest BMI percentile; BMI did not significantly differ for adults. Patients with chronic co-infection were most similar in age to patients with chronic PA alone (19.9 vs. 21.5 years, respectively) while patients with chronic MRSA alone, intermittent infection, and no MRSA/PA were significantly younger (all P < 0.05). Patients with chronic co-infection had the lowest baseline FEV1% predicted (average: 74.5; SD: 23.4), which was significantly lower compared with all other infection groups except patients with chronic PA alone (P = 0.77). The majority of patients with chronic co-infection and PA alone cultured mucoid PA at least once compared with only one-quarter of patients with intermittent infection.
      Changes in infection status from baseline through follow-up are presented in Table 2. Examining change in infection status among patients with at least 4 respiratory cultures during follow-up (N = 344), only 54% of patients retained their baseline infection status with the most change occurring among those with intermittent infection and no MRSA/PA at baseline.
      Table 2Changes in classification of infection status from the baseline year to follow-up among patients with at least 4 respiratory cultures during follow-up (n = 344).
      Baseline infection

      status
      Baseline and follow-up infection status defined as chronic co-infection (>50% PA+ and ≥3 MRSA+ cultures), chronic PA alone (>50% PA+ and <3 MRSA+ cultures), chronic MRSA alone (≤50% PA+ and ≥3 MRSA+ cultures), intermittent infection (≤50% PA+ and <3 MRSA+ cultures but ≥1 MRSA+/PA+), and no MRSA/PA (0 MRSA+/PA+ cultures).
      Follow-up infection status
      Baseline and follow-up infection status defined as chronic co-infection (>50% PA+ and ≥3 MRSA+ cultures), chronic PA alone (>50% PA+ and <3 MRSA+ cultures), chronic MRSA alone (≤50% PA+ and ≥3 MRSA+ cultures), intermittent infection (≤50% PA+ and <3 MRSA+ cultures but ≥1 MRSA+/PA+), and no MRSA/PA (0 MRSA+/PA+ cultures).
      Chronic co-infection

      N (%)
      Chronic PA alone

      N (%)
      Chronic MRSA alone

      N (%)
      Intermittent infection

      N (%)
      No MRSA/PA

      N (%)
      Chronic co-infection29 (74)3 (8)6 (15)1 (3)0 (0)
      Chronic PA alone15 (15)82 (80)1 (1)4 (4)0 (0)
      Chronic MRSA alone4 (9)0 (0)37 (88)1 (2)0 (0)
      Intermittent infection5 (8)10 (16)18 (29)20 (32)10 (16)
      No MRSA/PA2 (2)3 (3)13 (13)56 (57)24 (24)
      Abbreviations: MRSA, methicillin-resistant Staphylococcus aureus; PA, Pseudomonas aeruginosa.
      a Baseline and follow-up infection status defined as chronic co-infection (>50% PA+ and ≥3 MRSA+ cultures), chronic PA alone (>50% PA+ and <3 MRSA+ cultures), chronic MRSA alone (≤50% PA+ and ≥3 MRSA+ cultures), intermittent infection (≤50% PA+ and <3 MRSA+ cultures but ≥1 MRSA+/PA+), and no MRSA/PA (0 MRSA+/PA+ cultures).

      3.1 Rate of lung function decline

      The unadjusted average rate of decline in FEV1 for all study patients was −2.14% predicted per year (95% CI: −2.46, −1.83). Differences in the rate of FEV1 decline between patients with chronic co-infection and patients of all other infection groups differed depending on whether the model included baseline or time-varying infection status. Because nearly half of patients changed infection status from baseline to follow-up, the time-varying model was more representative of our data and is presented in Table 3. See the Online supplement for results of the baseline model (Table E1).
      Table 3Differences in the adjusted mean rate of decline in FEV1% predicted by infection status.
      Infection status
      Infection status defined during each study year and modeled as time-varying.
      Regression coefficient
      Model adjusted for age, baseline FEV1, number of episodes of IV antibiotics during the lead-in period, sex, CF-related diabetes, presence of Burkholderia cepacia complex, and presence of H. influenzae.
      95% CIP value
      Chronic co-infection vs. chronic PA alone0.601.13, −0.080.03
      Chronic co-infection vs. chronic MRSA alone0.891.56, −0.220.008
      Chronic co-infection vs. intermittent infection0.841.45, −0.220.007
      Chronic co-infection vs. no MRSA/PA1.121.74, −0.50<0.001
      Abbreviations: CI, confidence interval; FEV1, forced expiratory volume in the first second; MRSA, methicillin-resistant Staphylococcus aureus; PA, Pseudomonas aeruginosa.
      a Infection status defined during each study year and modeled as time-varying.
      b Model adjusted for age, baseline FEV1, number of episodes of IV antibiotics during the lead-in period, sex, CF-related diabetes, presence of Burkholderia cepacia complex, and presence of H. influenzae.
      Adjusting for age, sex, baseline FEV1% predicted, CFRD, presence of B. cepacia complex, H. flu, and baseline number of IV treatments, the average rate of FEV1 decline was −2.78% predicted/year for patients with chronic co-infection compared with −1.66% predicted/year for patients without MRSA/PA infection (adjusted difference: −1.12% predicted/year; 95% CI: −1.74, −0.50). Additionally, patients with chronic co-infection had a significantly more rapid rate of decline compared with patients with intermittent infection (adjusted difference: −0.84% predicted/year; 95% CI: −1.45, −0.22), chronic MRSA alone (adjusted difference: −0.89% predicted/year; 95% CI: −1.56, −0.22), and chronic PA alone (adjusted difference: −0.60% predicted/year; 95% CI: −1.13, −0.08).
      Among patients ≤20 years of age (N = 278) and additionally controlling for BMI percentile, the rate of FEV1 decline among patients with chronic co-infection was significantly different compared with all other infection groups except for that of chronic PA alone (adjusted difference: −0.58; 95% CI: −1.27, 0.12) (Table E2). Because of small sample size, particularly for the chronic MRSA alone and no MRSA/PA groups, we were not able to examine associations restricted to patients >20 years of age.

      3.2 Rate of Intravenous antibiotic treatments and mortality

      There were 1983 episodes treated with IV antibiotics during the study period (overall IR: 0.97 per person-year). The rate of IV treatments differed by infection group (Table 4).
      Table 4Average number of episodes of intravenous antibiotics per year by infection status.
      Infection status
      Infection status defined during each study year and modeled as time-varying; see Table E3 for analysis of baseline infection status.
      Adjusted IRR
      Model adjusted for age, baseline FEV1, sex, CF-related diabetes, number of episodes of IV antibiotics during the lead-in period, and genotype (categorized as homozygous ΔF508 and other).
      95% CIP Value
      Chronic co-infection vs. chronic PA alone1.241.01, 1.520.04
      Chronic co-infection vs. chronic MRSA alone1.341.03, 1.740.03
      Chronic co-infection vs. intermittent infection1.561.21, 2.01<0.001
      Chronic co-infection vs. no MRSA/PA2.001.54, 2.61<0.001
      Abbreviations: CI, confidence interval; IRR, incidence rate ratio; MRSA, methicillin-resistant Staphylococcus aureus; PA, Pseudomonas aeruginosa.
      a Infection status defined during each study year and modeled as time-varying; see Table E3 for analysis of baseline infection status.
      b Model adjusted for age, baseline FEV1, sex, CF-related diabetes, number of episodes of IV antibiotics during the lead-in period, and genotype (categorized as homozygous ΔF508 and other).
      Patients with chronic co-infection had a significantly higher rate of IV antibiotics compared with patients without MRSA/PA, adjusting for age, baseline FEV1% predicted, sex, CFRD, genotype, and baseline number of IV treatments (adjusted IRR: 2.00; 95% CI: 1.54, 2.61). Chronic co-infection was also associated with a 24% higher rate of IV treatments compared with chronic PA alone (adjusted IRR: 1.24; 95% CI: 1.01, 1.52) and a 34% higher rate compared with chronic MRSA alone (adjusted IRR: 1.34; 95% CI: 1.03, 1.74). Results were similar for patients ≤20 years of age and adjusting for BMI percentile (Table E4).
      There were 29 deaths during follow-up. This included 9 or 23% of the patients with chronic co-infection at baseline, 11 (10%) patients with chronic PA alone, 5 (12%) patients with chronic MRSA alone, 3 (5%) patients with intermittent infection, and 1 (1%) patient without MRSA/PA (Fig. 1).
      Figure thumbnail gr1
      Fig. 1Proportion of patients who died during follow-up by infection status at baseline. Abbreviations: MRSA, methicillin-resistant Staphylococcus aureus; PA, Pseudomonas aeruginosa.
      Average age at death was 24.9 years (SD: 12.0). Of the nine patients with chronic co-infection who died, two-thirds were female and all but one had CFRD. These proportions were much higher compared with patients with chronic PA alone who died (female: 45.5%; CFRD: 54.6%) or chronic MRSA alone (female: 40%; CFRD: 40%).

      4. Discussion

      In this retrospective cohort study of adult and pediatric CF patients attending a large CF center in Atlanta, GA, chronic co-infection of MRSA and P. aeruginosa was associated with a more rapid rate of decline in lung function compared with patients with either pathogen alone when modeled as a time-dependent variable to reflect yearly changes in infection status. Because we estimated rate of decline using an average of the two best values of FEV1% predicted recorded in each study year, this decline in lung function is not likely temporary or readily reversible.
      Additionally, patients with chronic co-infection had twice the rate of IV treatments than patients without MRSA/PA, adjusting for a number of confounders including baseline number of IV treatments, the strongest predictor of future episodes of IV treatments [
      • Block J.K.
      • Vandemheen K.L.
      • Tullis E.
      • Fergusson D.
      • Doucette S.
      • Haase D.
      • et al.
      Predictors of pulmonary exacerbations in patients with cystic fibrosis infected with multi-resistant bacteria.
      ,
      • VanDevanter D.R.
      • Pasta D.J.
      • Konstan M.W.
      Treatment and demographic factors affecting time to next pulmonary exacerbation in cystic fibrosis.
      ]. The higher incidence of IV treatments may be in part due to clinicians' concern about co-infection, possibly setting a lower threshold for IV treatment based on their infection status. The type of treatment received during a pulmonary exacerbation may also be a contributing factor to recurrent IV treatments. In a recent national study of U.S. CF centers, 98% of respondents reported using oral and/or IV antibiotics to treat MRSA during an exacerbation with the vast majority also reporting that co-infection with PA did not change how they treated MRSA; however, the type and dosing of antibiotics used to treat MRSA varied considerably [
      • Zobell J.T.
      • Epps K.L.
      • Young D.C.
      • Montague M.
      • Olson J.
      • Ampofo K.
      • et al.
      Utilization of antibiotics for methicillin-resistant Staphylococcus aureus infection in cystic fibrosis.
      ]. While we did not have adequate medication information to examine this at our center, future analyses should examine whether the type of antibiotics used to treat MRSA affects associations between chronic co-infection and health outcomes in CF.
      It is important to emphasize the baseline differences between infection groups. Patients with chronic co-infection were older (19.9 years), had the lowest FEV1 (74.5% predicted), highest prevalence of CFRD (36%), and highest average number of IV treatments (2.2 per patient) at baseline compared with the other infection groups. While we did adjust for these confounders in all analyses, residual confounding may be present. However, compared with patients with chronic PA alone, who were similar in baseline characteristics, patients with chronic co-infection had a significantly more rapid rate of FEV1 decline and rate of IV treatments when analysis included all patients. When analysis was restricted to patients ≤20 years of age, differences in the rate of FEV1 decline were non-significant compared with chronic PA alone. While it is possible that the association between chronic co-infection and rate of FEV1 decline differs by age, the loss of significance may be due to a loss of power as the estimated rate of decline did not change considerably from the main analysis (−0.60% predicted/year vs. −0.58% predicted/year). Nevertheless, a larger study is needed to explore interactions between covariates, particularly infection status and age.
      Finally, we observed a higher mortality during follow-up among patients with chronic co-infection compared with other groups. Of note, there was a strikingly high prevalence of CF-related diabetes among patients with chronic co-infection who died (88.9%), suggesting that the presence of CFRD may impact the association between chronic co-infection and mortality. However, this finding could be by chance alone as we did not have a large enough sample size to test for differences. A larger study is needed to examine differences in mortality.
      Our findings support those of Hubert and colleagues who also found MRSA and PA co-infection to be associated with a more rapid rate of lung function decline and more pulmonary exacerbations requiring IV antibiotics in their two-year study conducted in France; however, there are some distinct differences between the studies [
      • Hubert D.
      • Reglier-Poupet H.
      • Sermet-Gaudelus I.
      • Ferroni A.
      • Le Bourgeois M.
      • Burgel P.R.
      • et al.
      Association between Staphylococcus aureus alone or combined with Pseudomonas aeruginosa and the clinical condition of patients with cystic fibrosis.
      ]. Specifically, the authors did not examine the effect of chronic co-infection but rather any co-infection and their reference group included patients without any S. aureus (methicillin-sensitive or resistant) or PA [
      • Hubert D.
      • Reglier-Poupet H.
      • Sermet-Gaudelus I.
      • Ferroni A.
      • Le Bourgeois M.
      • Burgel P.R.
      • et al.
      Association between Staphylococcus aureus alone or combined with Pseudomonas aeruginosa and the clinical condition of patients with cystic fibrosis.
      ]. Additionally, we found the proportion of patients in our cohort to have any co-infection, regardless of chronicity, to be considerably higher than the French cohort as expected from previous reports on the prevalence of MRSA in the U.S. compared with European countries [
      • Hubert D.
      • Reglier-Poupet H.
      • Sermet-Gaudelus I.
      • Ferroni A.
      • Le Bourgeois M.
      • Burgel P.R.
      • et al.
      Association between Staphylococcus aureus alone or combined with Pseudomonas aeruginosa and the clinical condition of patients with cystic fibrosis.
      ,
      • Goss C.H.
      • Muhlebach M.S.
      Review: Staphylococcus aureus and MRSA in cystic fibrosis.
      ]. In the present study, 157 (44%) patients cultured both MRSA and PA at least once during the same year of the study period.
      The present study has several limitations, including the use of retrospective registry data, potential misclassification of infection status, and potential selection bias from excluding patients with too few respiratory cultures or FEV1 measurements at baseline. Because the study was limited to registry data collected between 2007 and 2013, we were not able to observe incidence of infection as patients may have been infected for years before the study began. Additionally, nearly half of patients in our cohort changed infection status over time, particularly younger patients who were free of MRSA or PA at baseline. Proesmans and colleagues reported similar findings in their study examining different levels of PA infection [
      • Proesmans M.
      • Balinska-Miskiewicz W.
      • Dupont L.
      • Bossuyt X.
      • Verhaegen J.
      • Hoiby N.
      • et al.
      Evaluating the “Leeds criteria” for Pseudomonas aeruginosa infection in a cystic fibrosis centre.
      ]. For these reasons, we chose a model that included infection status as a time-dependent variable. While this model may have reduced potential for misclassification of infection over time, we had to modify our definition of chronic MRSA infection as there were patients with fewer than four respiratory cultures in a year and some without any for which the previous year's infection status was used, potentially introducing misclassification bias. Notably, defining infection status at baseline alone generated different results than the main analysis (Tables E1 and E3); specifically, there were non-significant associations between chronic co-infection and both outcomes compared with chronic PA alone and chronic MRSA alone. Waters and colleagues also found different results and conclusions depending on whether infection was defined at baseline or varied with time in their study examining the effect of chronic Stenotrophomonas maltophilia infection on survival in CF [
      • Waters V.
      • Atenafu E.G.
      • Lu A.
      • Yau Y.
      • Tullis E.
      • Ratjen F.
      Chronic Stenotrophomonas maltophilia infection and mortality or lung transplantation in cystic fibrosis patients.
      ]. The present study provides further evidence that examining the effect of microbiologic pathogens on health outcomes in cystic fibrosis is complex with many factors to consider that can modify results and interpretation.
      Because this analysis was conducted among patients attending a single CF center, the results of this study may not be generalizable to all CF patients. Further research in other CF care centers should be conducted to determine whether this association persists in the broader CF population.

      4.1 Conclusion

      In this single-center study, patients with chronic co-infection were found to have a significantly more rapid rate of decline in lung function and a higher incidence of episodes requiring IV antibiotics compared with patients with either pathogen alone when infection status was examined annually. The current data and the growing prevalence of MRSA and co-infection in this population suggest that future research into the impact of co-infection is warranted.

      Acknowledgments

      The authors thank Emory University and Children's Healthcare of Atlanta, Inc. for their support, the Cystic Fibrosis Foundation for providing the data presented in this study, and Drs. William McClellan and David Kleinbaum for advisement in the statistical analysis and critical review of the study design.

      Appendix A. Supplementary data

      References

        • Flume P.A.
        Pulmonary complications of cystic fibrosis.
        Respir Care. 2009; 54: 618-627
      1. Patient Registry annual data report 2012.
        Cystic Fibrosis Foundation, Bethesda, Maryland2013
        • Hoffman L.R.
        • Deziel E.
        • D'Argenio D.A.
        • Lepine F.
        • Emerson J.
        • McNamara S.
        • et al.
        Selection for Staphylococcus aureus small-colony variants due to growth in the presence of Pseudomonas aeruginosa.
        Proc Natl Acad Sci U S A. 2006; 103: 19890-19895
        • Mitchell G.
        • Seguin D.L.
        • Asselin A.E.
        • Deziel E.
        • Cantin A.M.
        • Frost E.H.
        • et al.
        Staphylococcus aureus sigma B-dependent emergence of small-colony variants and biofilm production following exposure to Pseudomonas aeruginosa 4-hydroxy-2-heptylquinoline-N-oxide.
        BMC Microbiol. 2010; 10: 33
        • Sagel S.D.
        • Gibson R.L.
        • Emerson J.
        • McNamara S.
        • Burns J.L.
        • Wagener J.S.
        • et al.
        Impact of Pseudomonas and Staphylococcus infection on inflammation and clinical status in young children with cystic fibrosis.
        J Pediatr. 2009; 154: 183-188
        • Hudson V.L.
        • Wielinski C.L.
        • Regelmann W.E.
        Prognostic implications of initial oropharyngeal bacterial flora in patients with cystic fibrosis diagnosed before the age of two years.
        J Pediatr. 1993; 122: 854-860
        • Rosenbluth D.B.
        • Wilson K.
        • Ferkol T.
        • Schuster D.P.
        Lung function decline in cystic fibrosis patients and timing for lung transplantation referral.
        Chest. 2004; 126: 412-419
        • Dasenbrook E.C.
        • Checkley W.
        • Merlo C.A.
        • Konstan M.W.
        • Lechtzin N.
        • Boyle M.P.
        Association between respiratory tract methicillin-resistant Staphylococcus aureus and survival in cystic fibrosis.
        JAMA. 2010; 303: 2386-2392
        • Cox D.W.
        • Kelly C.
        • Rush R.
        • O'Sullivan N.
        • Canny G.
        • Linnane B.
        The impact of MRSA infection in the airways of children with cystic fibrosis; a case–control study.
        Ir Med J. 2011; 104: 305-308
        • Dasenbrook E.C.
        • Merlo C.A.
        • Diener-West M.
        • Lechtzin N.
        • Boyle M.P.
        Persistent methicillin-resistant Staphylococcus aureus and rate of FEV1 decline in cystic fibrosis.
        Am J Respir Crit Care Med. 2008; 178: 814-821
        • Vanderhelst E.
        • De Meirleir L.
        • Verbanck S.
        • Pierard D.
        • Vincken W.
        • Malfroot A.
        Prevalence and impact on FEV(1) decline of chronic methicillin-resistant Staphylococcus aureus (MRSA) colonization in patients with cystic fibrosis. A single-center, case control study of 165 patients.
        J Cyst Fibros. 2012; 11: 2-7
        • Hubert D.
        • Reglier-Poupet H.
        • Sermet-Gaudelus I.
        • Ferroni A.
        • Le Bourgeois M.
        • Burgel P.R.
        • et al.
        Association between Staphylococcus aureus alone or combined with Pseudomonas aeruginosa and the clinical condition of patients with cystic fibrosis.
        J Cyst Fibros. 2013; 12: 497-503
        • Muhlebach M.S.
        • Miller M.
        • LaVange L.M.
        • Mayhew G.
        • Goodrich J.S.
        • Miller M.B.
        Treatment intensity and characteristics of MRSA infection in CF.
        J Cyst Fibros. 2011; 10: 201-206
        • Sawicki G.S.
        • Rasouliyan L.
        • Ren C.L.
        The impact of MRSA on lung function in patients with cystic fibrosis.
        Am J Respir Crit Care Med. 2009; 179 ([author reply 5]): 734-735
        • Com G.
        • Carroll J.L.
        • Castro M.M.
        • Tang X.
        • Jambhekar S.
        • Berlinski A.
        Predictors and outcome of low initial forced expiratory volume in 1 second measurement in children with cystic fibrosis.
        J Pediatr. 2014; 164: 832-838
        • Goss C.H.
        • Muhlebach M.S.
        Review: Staphylococcus aureus and MRSA in cystic fibrosis.
        J Cyst Fibros. 2011; 10: 298-306
        • Maliniak M.L.
        • McCarty N.A.
        • Stecenko A.A.
        Chronic methicillin-resistant Staphyloccus aureus and Pseudomonas aeruginosa co-infection and lung function decline In cystic fibrosis [abstract].
        Pediatr Pulmonol. 2014; 49: 354
        • Proesmans M.
        • Balinska-Miskiewicz W.
        • Dupont L.
        • Bossuyt X.
        • Verhaegen J.
        • Hoiby N.
        • et al.
        Evaluating the “Leeds criteria” for Pseudomonas aeruginosa infection in a cystic fibrosis centre.
        Eur Respir J. 2006; 27: 937-943
        • Lee T.W.
        • Brownlee K.G.
        • Conway S.P.
        • Denton M.
        • Littlewood J.M.
        Evaluation of a new definition for chronic Pseudomonas aeruginosa infection in cystic fibrosis patients.
        J Cyst Fibros. 2003; 2: 29-34
        • Pressler T.
        • Bohmova C.
        • Conway S.
        • Dumcius S.
        • Hjelte L.
        • Hoiby N.
        • et al.
        Chronic Pseudomonas aeruginosa infection definition: EuroCareCF Working Group report.
        J Cyst Fibros. 2011; 10: S75-S78
        • Elliott A.C.
        • Hynan L.S.
        A SAS((R)) macro implementation of a multiple comparison post hoc test for a Kruskal–Wallis analysis.
        Comput Methods Prog Biomed. 2011; 102: 75-80
        • Hankinson J.L.
        • Odencrantz J.R.
        • Fedan K.B.
        Spirometric reference values from a sample of the general U.S. population.
        Am J Respir Crit Care Med. 1999; 159: 179-187
        • Wang X.
        • Dockery D.W.
        • Wypij D.
        • Fay M.E.
        • Ferris Jr., B.G.
        Pulmonary function between 6 and 18 years of age.
        Pediatr Pulmonol. 1993; 15: 75-88
        • Block J.K.
        • Vandemheen K.L.
        • Tullis E.
        • Fergusson D.
        • Doucette S.
        • Haase D.
        • et al.
        Predictors of pulmonary exacerbations in patients with cystic fibrosis infected with multi-resistant bacteria.
        Thorax. 2006; 61: 969-974
        • VanDevanter D.R.
        • Pasta D.J.
        • Konstan M.W.
        Treatment and demographic factors affecting time to next pulmonary exacerbation in cystic fibrosis.
        J Cyst Fibros. 2015; 14: 763-769
        • Zobell J.T.
        • Epps K.L.
        • Young D.C.
        • Montague M.
        • Olson J.
        • Ampofo K.
        • et al.
        Utilization of antibiotics for methicillin-resistant Staphylococcus aureus infection in cystic fibrosis.
        Pediatr Pulmonol. 2015; 50: 552-559
        • Waters V.
        • Atenafu E.G.
        • Lu A.
        • Yau Y.
        • Tullis E.
        • Ratjen F.
        Chronic Stenotrophomonas maltophilia infection and mortality or lung transplantation in cystic fibrosis patients.
        J Cyst Fibros. 2013; 12: 482-486