Cystic fibrosis related diabetes is not independently associated with increased Stenotrophomonas maltophilia infection: Longitudinal data from the UK CF Registry

Journal of Cystic Fibrosis 18 (2019) 294–298

Contents lists available at ScienceDirect

Journal of Cystic Fibrosis journal homepage: www.elsevier.com/locate/jcf

Cystic fibrosis related diabetes is not independently associated with increased Stenotrophomonas maltophilia infection: Longitudinal data from the UK CF Registry Freddy Frost a,⁎, Dilip Nazareth a, Matthew Shaw b, Martin J Walshaw a a b

Adult Cystic Fibrosis Centre, Liverpool Heart and Chest Hospital, Liverpool L14 3PE, UK Research Department, Liverpool Heart and Chest Hospital, Liverpool L14 3PE, UK

a r t i c l e

i n f o

Article history: Received 9 August 2018 Revised 11 October 2018 Accepted 15 October 2018 Available online 25 October 2018

a b s t r a c t Introduction: Stenotrophomonas maltophilia is common in the sputum of people with cystic fibrosis related diabetes (CFRD), raising the question as to whether this is a risk factor for its acquisition. We investigated this at a population level. Methods: We analysed national Cystic Fibrosis Registry data 2011–2015 for 8047 people with CF N age 6 years, looking at demographics, diagnosis of CFRD, lung function and sputum microbiology; using descriptive and multivariate strategies to establish independent predictors for S. maltophilia culture and associated outcomes. Results: S. maltophilia was present in 1148 (14.1%). Although univariate analysis confirmed it was more prevalent in those with CFRD, when adjusted for other clinical parameters there was no longer a relationship. Markers of more severe lung disease were independent risk-factors for S. maltophilia. Conclusion: Although S. maltophilia is more common in people with CFRD, it is not an independent risk-factor for S. maltophilia acquisition. © 2018 Published by Elsevier B.V. on behalf of European Cystic Fibrosis Society.

1. Introduction Treatment for people with cystic fibrosis (CF) has advanced dramatically over the last few decades and as survival improves the spectrum of organisms implicated in pulmonary infection is also shifting. The increasing prevalence of organisms such as Stenotrophomonas maltophilia, Achromobacter spp, and Non-Tuberculous Mycobacteria spp (NTM) has led them being grouped together as potential “emerging pathogens” [1]. S. maltophilia (previously known as Pseudomonas maltophilia and Xanthomonas maltophilia) is a Gram-negative, obligate aerobic nonfermenting rod with intrinsic multi-drug resistance [2]. Its prevalence in people with CF varies from centre to centre but has been reported as high as 31% [3]. The clinical significance of S. maltophilia remains controversial as no clear evidence for more rapid pulmonary decline following S. maltophilia acquisition has been found. However, strains isolated from people with CF have increased biofilm forming ability and a specific immune response has been observed and associated with increased ⁎ Corresponding author. E-mail addresses: [email protected] (F. Frost), [email protected] (D. Nazareth), [email protected] (M. Shaw), [email protected] (M.J. Walshaw).

https://doi.org/10.1016/j.jcf.2018.10.011 1569-1993/© 2018 Published by Elsevier B.V. on behalf of European Cystic Fibrosis Society.

exacerbations, suggesting S. maltophilia has potential to be more than a simple coloniser of the lungs in CF [4–6]. It is therefore of interest to understand which factors affect the acquisition of S. maltophilia and recently CF-related diabetes (CFRD) has been implicated. Lehoux Dubois et al. [7] reported increased S. maltophilia in respiratory samples of patients with dysglycaemia compared to their normoglycaemic comparators. CFRD is modifiable and optimisable and therefore any potential association with S. maltophilia warrants further investigation. Equally, should an association between CFRD and S. maltophilia exist; S. maltophilia “first growth” would dictate further investigation for CFRD. With this in mind, we sought to use data from the national UK CF Registry to investigate the relationship between CFRD and S. maltophilia. 2. Methods We utilised the UK Cystic Fibrosis Registry which collects pseudoanonymised longitudinal data from over 10,000 people with CF in the UK. With patient consent, data are inputted annually at every UK specialist CF centre including baseline demographics, co-morbidities, spirometry, blood test results, medications, intravenous antibiotics days and respiratory sample microbiology results. The registry has excellent coverage of the CF population in the UK (N99%). The UK CF Registry

F. Frost et al. / Journal of Cystic Fibrosis 18 (2019) 294–298

Research Steering Committee approved the application and subsequent release of data for this study. 2.1. Data extraction Data were extracted for all 10,552 individuals with records in the registry for the study period 2011–2015 to look for an association between CFRD and S. maltophilia, where those with any growth in the study period were considered to be positive (SM+). 2.2. Missing data and standardisation Although missing data accounted for only 5.9% of all data points, it occurred more frequently in those ≤6 years age than the remainder (1.7 vs. 0.55 missing data-points per person respectively), so the younger age group were removed from the analysis. In the remaining 8047 individuals missing data accounted for only 3.6% of all data-points (see Table 1). Hba1c values were standardised by conversion to the International Federation of Clinical Chemistry (IFCC) reference as per published formulae. [8] Hba1c values were missing in the majority of patients and we therefore did not include Hba1c in any multivariate analyses. BMI (adults) and BMI percentile (children) are automatically calculated by the registry based on the height and weight entered. In order to include BMI in a population wide multivariate analysis, we standardised adult and paediatric BMI into “underweight”, “normal”, “overweight” and “obese” based on WHO criteria. [9]. The registry does not formally define diagnostic criteria for CFRD and therefore for the purposes of the study a diagnosis of CFRD was considered to be made if the local CF team had indicated accordingly in the annual review proforma. During the study period UK guidelines recommended CFRD was a clinical diagnosis with OGTT and/or home blood glucose monitoring taken into consideration. [10] Lung function was taken as the % predicted forced expiratory volume in 1 s (FEV1) recorded at the time of the annual review, calculated using the Global Lung Initiative reference formulae [11].

295

2.3. Statistical analysis Categorical and dichotomous variables were summarised as absolute number and percentages and non-normally distributed continuous data as median and interquartile ranges (IQR). The prevalence of missing data in the registry for baseline and clinical measurements reported. Where categorical comparisons were made between groups the χ2 test was used; for similar comparisons between continuous variables, the Wilcoxon rank-sum test was used. To examine the relationship between patient demographics and risk-factors with S. maltophilia, a logistic regression model for the presence of S. maltophilia was created. A purposeful model selection strategy was used to fit the final model. Variables were selected for inclusion in the model when univariate p-value was b0.2 or where there was biologic plausibility. Continuous variables age, FEV1, IV days were separated into clinically meaningful dummy variables so that each strata of the variable could be assessed independently. The final model included the following independent demographic and risk-factor variables: age b 18 years, female gender, underweight BMI, overweight BMI, obese BMI, CFRD, pancreatic supplements, FEV1 50–79% predicted, FEV1 b 50% predicted, Phe508del homozygous genotype, ‘other’ genotype, N14 IV days, presence of S. aureus, P. aeruginosa, Aspergillus spp. or Burkholderia cepacia complex (BCC). Statistical analyses were performed with SAS version 9.3 (SAS Institute, Cary, NC). In all cases, p b .05 was considered statistically significant. Categorical data are presented as count (%), averages are presented as mean (SD) or median (IQR) as appropriate depending on normality. Odd ratios are presented with 95% confidence intervals. Unadjusted p-values are presented throughout. 3. Results 3.1. Patient characteristics Cohort demographics are presented in Table 1. Briefly, mean ± SD age was 22 ± 14 years, mean FEV1 76.0 ± 21.1%predicted, 1815

Table 1 Baseline demographics and clinical characteristics of the study cohort at study entry and univariate analysis of the differences between those who did and did not have a positive respiratory culture for S. maltophilia.

Age Age b 18 years Age 18–35 years Age N 35 years Female BMI category Underweight (BMI b18.5) Normal weight (BMI 18.5–25) Overweight (BMI 25–30) Obese (BMI N30) Pancreatic supplements FEV1 (% predicted) FEV1 ≥ 80 (normal to mild COPD) FEV1 50–79 (moderate COPD) FEV1 b 50 (severe COPD) Genotype Homozygous for F508D Heterozygous for F508D Other IV days IV days = 0 IV days = 0–14 IV days N14 CFRD HbA1c HbA1c in CFRD patients S. aureus P. aeruginosa Aspergillus BCC

All patients (n = 8047)

No S. maltophilia (n = 6909)

S. maltophilia (n = 1138)

P value

Missing Data

22.1 (15, 31) 2784 (34.6) 3913 (48.6) 1350 (16.8) 3736 (46.4)

22.4 (15, 31) 2309 (33.4) 3412 (49.4) 1188 (17.2) 3144 (45.5)

20.7 (14, 38) 475 (41.7) 501 (44.0) 162 (14.2) 592 (52.0)

b0.001 b0.001 b0.001 0.013 b0.001

0 (0)

705 (8.8) 4183 (52.0) 2368 (29.4) 791 (9.8) 6585 (81.8) 76.0 (55.6, 92.2) 3247 (40.4) 2822 (35.1) 1978 (24.6)

603 (8.7) 3569 (51.7) 2045 (29.6) 692 (10.0) 5585 (80.8) 76.0 (55.6, 92.2) 2879 (41.7) 2380 (34.5) 1650 (23.9)

102 (9.0) 614 (54.0) 323 (28.4) 99 (8.7) 1000 (87.9) 68.6 (49.3, 85.7) 368 (32.3) 442 (38.8) 328 (28.8)

0.79 0.15 0.40 0.17 b0.001 b0.001 b0.001 0.004 b0.001

262 (3.3)

3922 (48.7) 3115 (38.7) 689 (8.6)

3335 (48.3) 2670 (38.7) 607 (8.8)

587 (51.6) 445 (39.1) 82 (7.2)

0.038 0.77 0.078

321 (4.0)

3769 (46.8) 1265 (15.7) 3013 (37.4) 1815 (22.6) 39.9 (36.6, 46.5) 51.0 (43.2, 63.9) 2925 (36.4) 3635 (45.2) 967 (12.0) 304 (3.8)

3447 (49.9) 1079 (15.6) 2383 (34.5) 1530 (22.2) 39.9 (36.6, 46.5) 51.9 (44.0, 63.9) 2413 (34.9) 3087 (44.7) 598 (8.7) 272 (3.9)

322 (28.3) 186 (16.3) 630 (55.4) 285 (25.0) 40.0 (36.6, 46.5) 49.7 (42.0, 62.8) 512 (45.0) 548 (48.2) 369 (32.4) 32 (2.8)

b0.001 0.53 b0.001 0.03 0.19 0.30 b0.001 0.029 b0.001 0.07

0 (0)

0 (0)

200 (2.5) 403 (5.0)

0 (0) 3312 (41.2) 322 (17.7) 0 (0) 0 (0) 0 (0) 0 (0)

296

F. Frost et al. / Journal of Cystic Fibrosis 18 (2019) 294–298

(22.6%) had a diagnosis of CFRD, 3635 (45.2%) had documented P. aeruginosa (Pa) growth, and 3736 (46.4%) were female. There were 1138 cases (14.1%) with a growth of S. maltophilia in the study period. 3.2. S. maltophilia rates Annual S. maltophilia prevalence was observed to rise over the course of the study period with a peak in 2014 at 6.9%. Overall those with a documented growth of S. maltophilia (SM+) were younger than those without (SM-) (20.2 ± 13.2 yrs. vs. 22.1 ± 14.2 yrs., p b .001) However, S. maltophilia prevalence appeared to have a bi-modal distribution with a peak prevalence in adolescence before a trough during early-mid adulthood and subsequent increasing prevalence in those aged 45 and above, see Fig. 1. 3.3. Univariate analysis Univariate analysis was performed to assess for plausible associations with S. maltophilia growth and the results are presented in Table 1. The SM+ cohort had higher prevalence of CFRD (25.0% vs. 22.2%, p = .03), were more likely to be female (52.0% vs. 45.5% respectively, p b .001) and demonstrated evidence of more severe disease with poorer lung function (68.6 ± 23.8% vs. 76.0 ± 24.7% predicted FEV1, p b .001), more annual IV antibiotics days (10 [0−31] days vs. 0 [0–28], p b .0001), and higher rates of Pa colonisation (48.2% vs. 44.7%, p = .029) than SM- cohort. They also had higher prevalence of S. aureus (45.0% vs. 34.9%, p b .001) and Aspergillus spp. (32.9% vs. 9.3%, p b .001), but overall no significant associations were seen for BCC. 3.4. Multivariate A multivariate logistic regression model was then created to test for independent association between covariates and SM+ (Fig. 2). Age b 18 years (OR 1.62 [1.38–1.89], p b .001), female sex (OR 1.24 [1.08– 1.42], p = .002), severe lung function (FEV1 b 50% predicted OR 1.35 [1.12–1.64]. p = .002), pancreatic supplementation (OR 1.352 [1.09– 1.67], p = .006), N14 annual IV days (OR 2.00 [1.73–2.32], P b .001), S. aureus growth (OR 1.57 [1.37–1.80], p b .001) and Aspergillus spp. growth (OR 4.55 [3.89–5.31], p b .001) were all independently associated with an increased likelihood of S. maltophilia growth, whereas BCC was associated with reduced likelihood of S. maltophilia growth (OR 0.64 [0.43–0.94, p = .02]. However CFRD was not found to be independently associated with S. maltophilia growth (OR 1.08 [0.92–1.28]). 3.5. Sensitivity analyses To test the robustness of our findings we performed two sensitivity analyses. Firstly, we performed the same analysis but excluded those

individuals where S. maltophilia was seen in only one year, thereby only including those with recurrent growths in the SM+ group, See Table S1 and Fig. S1. Secondly, we examined whether inclusion of young children b10 years of age (below the age of recommended screening) impacted the outcomes of our primary analysis, See Table S2 and Fig. S2. Both sensitivity analyses were consistent with the primary analyses. 4. Discussion We utilised UK CF Registry data to investigate the association between CFRD and S. maltophilia. Although higher rates of CFRD occurred in people with S. maltophilia, when adjusted for other variables there was no evidence of an independent association. Increased S. maltophilia in people with CFRD has been reported in a number of studies. Stanojevic et al. [12] found significantly higher rates of CFRD in their SM+ group when investigating factors that influence the acquisition of S. maltophilia. Marshall et al [13] noted increased S. maltophilia in people with CFRD and Vidigal et al. [14] reported a higher prevalence of CFRD in SM+ subjects. Conversely, a number of studies have not observed a relationship between CFRD and S. maltophilia, although the aims of these studies were often targeted towards outcomes of S. maltophilia infection rather than its acquisition. [2,4,5] All previous studies have been limited by sample size, the largest including only 150 cases with S. maltophilia [12]. By utilising national registry data we were able to apply multivariable logistic regression analysis to a large CF population which included over 1000 cases of S. maltophilia. Although we confirmed the apparent increased prevalence of S. maltophilia in people with CFRD, this association disappeared once adjusted for markers of severe lung disease including poorer lung function and increased IV antibiotic exposure. These findings are in keeping with those recently reported by Lehoux Dubois et al. [7] but go a step further in suggesting the increased S. maltophilia seen in CF people patients with dysglycaemia is associated with their poorer clinical condition rather than the dysglycaemia itself. In addition to the association between pulmonary function and IV days, we observed associations between S. maltophilia and other organisms. For example, we observed increased S. aureus and Aspergillus spp. in the S. maltophilia population and less BCC. Interestingly the SM+ population had over 4 times the odds of an Aspergillus spp. growth than their SM- counterparts and this relationship persisted on multivariate analysis. A similar relationship was previously reported by Marchac et al. [15] and whilst it could be inferred that a synergism or probiosis exists between the two organisms there is also recent evidence S. maltophilia and Aspergillus spp. are antagonistic of each other in mixed biofilms. [16]. Although we have shown that younger age was an independent risk factor for S. maltophilia growth (those SM+ were on average 2 years

Prevalence (%)

40 30 20 10 0 0

20

40

60

Age (years) Fig. 1. S. maltophilia prevalence across the 5-year study window for each year of age.

F. Frost et al. / Journal of Cystic Fibrosis 18 (2019) 294–298

297

Odds Ratio (95% CI)

P value

Age < 18 years

1.615 (1.383, 1.887)

<0.001

Female

1.239 (1.084, 1.416)

0.002

0.809 (0.635, 1.031)

0.09

Covariate

1.00

Underweight (BMI <18.5) Normal weight (BMI 18.5 - 25 )

Referent category

-

Overweight (BMI 25 - 30)

1.075 (0.920, 1.256)

0.36

Obese (BMI >30)

1.022 (0.804, 1.300)

0.86

CFRD

1.084 (0.917, 1.282)

0.34

Pancreatic supplements

1.352 (1.093, 1.673)

0.006

FEV1 ≥80

Referent category

-

FEV1 50-79

1.250 (1.060, 1.473)

0.008

FEV1 <50

1.352 (1.117, 1.637)

0.002

Genotype = Phe508del homozygous

0.901 (0.780, 1.040)

0.15

Genotype = Other

0.941 (0.724, 1.223)

0.65

IV days > 14

2.004 (1.730, 2.322)

<0.001

S. aureus

1.569 (1.371, 1.795)

<0.001

P. aeruginosa

0.900 (0.777, 1.042)

0.16

Aspergillus

4.545 (3.891, 5.310)

<0.001

BCC

0.02 0.4

0.6

0.8

1

2.0

2

3

4

5 5.5

0.636 (0.431, 0.938)

Odds Ratio for association with S. maltophilia Fig. 2. Forest plot from multivariate analysis of factors associated with a positive respiratory culture for S. maltophilia. Abbreviations: BMI=Body Mass Index; FEV1 = Forced expiratory volume in 1 s; IV=Intravenous antibiotics; CFRD = Cystic fibrosis related diabetes; BCC = Burkholderia cepacia complex.

younger), there was a bi-modal distribution of S. maltophilia prevalence across the age groups and to our knowledge this is the first time this has been reported. A recent study of the US CF registry found older patients to be at significantly higher risk of NTM [17] and hence as life expectancy increases, further work may be required to investigate the risk factors for acquisition of the emerging pathogens and other organisms in the older CF population, which may be different particularly in the presence of a “survivor effect”. There are several limitations to this study. Firstly, the retrospective observational design means we cannot exclude a confounder that exists outside our model. For example, S. maltophilia and Aspergillus may each thrive in a particular niche in the lung such as cavities or severe bronchiectasis which we are unable to account for in our multivariate analysis. Secondly, SM+ individuals not classified as CFRD may still have early glucose handling abnormalities. The registry does not define diagnostic criteria for CFRD and in practice there may be heterogeneity between centres as to how CFRD is classified and/or treated. Lehoux Dubois et al. [7] noted increased S. maltophilia not only in those with CFRD based on the oral glucose tolerance test (OGTT) but also those with impaired glucose tolerance that did not meet the OGTT criteria. Therefore, those with early dysglycaemia (often characterised only by post-prandial glucose excursions and missed by OGTT) [18] may have increased prevalence of S. maltophilia but would not be classified as CFRD in the registry. Further work is required to understand the impact of early glucose abnormalities on S. maltophilia growth. Thirdly, if hyperglycaemia were related to S. maltophilia acquisition then anti-diabetic agents such as insulin may have an antagonistic effect and given no data is available on insulin prescriptions, dosing, or adherence we cannot exclude a confounding treatment effect. Fourthly, we excluded children under the age of 6 from our analysis due to increased missing data, and although the generalizability of our

results to that cohort is uncertain, CFRD increases with age and is uncommon in young children. Although screening is recommended from the age of 10, many centres screen for CFRD from a younger age, [19] a sensitivity analysis confirmed that our results were robust even when the 6–10 years age-group were excluded. Finally, the number of annual positive respiratory samples or type of sample that was positive for S. maltophilia (and other emerging pathogens) was not recorded in the UK CF Registry during the study period and hence misclassification of those individuals with no sputum samples as SM- may falsely inflate that population and equally the SM+ population may be falsely inflated by one-off growths or misclassification. Indeed, the overall 14.1% prevalence of S. maltophilia we report is higher than 9.9% and 10.4% in the historical studies reported by Goss et al. [20] and Waters et al. [21] respectively. However, given the increasing detection of S. maltophilia globally as well as the introduction of novel technologies such as MALDI-TOF for microbial identification, the prevalence data in our contemporaneous study is unlikely to be significantly inflated. Furthermore, our sensitivity analysis confirmed our primary analysis was not unduly affected by those with just a single year of documented S. maltophilia growth. In conclusion, this is by far the largest study looking at the association between S. maltophilia and CFRD, where multiple datasets in many thousands of individuals were compared over a number of years. We have shown that the previously documented association between S. maltophilia and CFRD is no longer apparent once adjustment for other clinical parameters has been made. Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi. org/10.1016/j.jcf.2018.10.011.

298

F. Frost et al. / Journal of Cystic Fibrosis 18 (2019) 294–298

References [1] Parkins MD, Floto RA. Emerging bacterial pathogens and changing concepts of bacterial pathogenesis in cystic fibrosis. J Cyst Fibros 2015;14(3):293–304. [2] Waters V, Atenafu EG, Salazar JG, Lu A, Yau Y, Matukas L, et al. Chronic Stenotrophomonas maltophilia infection and exacerbation outcomes in cystic fibrosis. J Cyst Fibros 2012;11(1):8–13. [3] Ballestero S, Virseda I, Escobar H, Suarez L, Baquero F. Stenotrophomonas maltophilia in cystic fibrosis patients. Eur J Clin Microbiol Infect Dis 1995;14(8):728–9. [4] Waters V, Yau Y, Prasad S, Lu A, Atenafu E, Crandall I, et al. Stenotrophomonas maltophilia in cystic fibrosis: serologic response and effect on lung disease. Am J Respir Crit Care Med 2011;183(5):635–40. [5] Wettlaufer J, Klingel M, Yau Y, Stanojevic S, Tullis E, Ratjen F, et al. Longitudinal study of Stenotrophomonas maltophilia antibody levels and outcomes in cystic fibrosis patients. J Cyst Fibros 2017;16(1):58–63. [6] Pompilio A, Pomponio S, Crocetta V, Gherardi G, Verginelli F, Fiscarelli E, et al. Phenotypic and genotypic characterization of Stenotrophomonas maltophilia isolates from patients with cystic fibrosis: genome diversity, biofilm formation, and virulence. BMC Microbiol 2011;11:159. [7] Lehoux Dubois C, Boudreau V, Tremblay F, Lavoie A, Berthiaume Y, Rabasa-Lhoret R, et al. Association between glucose intolerance and bacterial colonisation in an adult population with cystic fibrosis, emergence of Stenotrophomonas maltophilia. J Cyst Fibros 2017;16(3):418–24. [8] NGSP. IFCC Standardization of HbA1c 2010. [Accessed online 28/6/2018]. Available from: http://www.ngsp.org/ifcc.asp. [9] WHO. Growth reference 5–19 years. [Accessed online 28/6/2018]. Available from: http://www.who.int/growthref/who2007_bmi_for_age/en/. [10] Littlewood JB, Bridges N. Management of Cystic Fibrosis Related Diabetes Mellitus. London: Cystic Fibrosis Trust; 2004 June 2004 (Report No). [11] Quanjer PH, Stanojevic S, Cole TJ, Baur X, Hall GL, Culver BH, et al. Multi-ethnic reference values for spirometry for the 3-95-yr age range: the global lung function 2012 equations. Eur Respir J 2012;40(6):1324–43.

[12] Stanojevic S, Ratjen F, Stephens D, Lu A, Yau Y, Tullis E, et al. Factors influencing the acquisition of Stenotrophomonas maltophilia infection in cystic fibrosis patients. J Cyst Fibros 2013;12(6):575–83. [13] Marshall BC, Butler SM, Stoddard M, Moran AM, Liou TG, Morgan WJ. Epidemiology of cystic fibrosis-related diabetes. J Pediatr 2005;146(5):681–7. [14] Goncalves Vidigal P, Schmidt D, Stehling F, Mellies U, Steinmann E, Buer J, et al. Development of a quantitative immunofluorescence assay for detection of Stenotrophomonas maltophilia antibodies in patients with cystic fibrosis. J Cyst Fibros 2013;12(6):651–4. [15] Marchac V, Equi A, Le Bihan-Benjamin C, Hodson M, Bush A. Case-control study of Stenotrophomonas maltophilia acquisition in cystic fibrosis patients. Eur Respir J 2004;23(1):98–102. [16] Melloul E, Luiggi S, Anais L, Arne P, Costa JM, Fihman V, et al. Characteristics of Aspergillus fumigatus in Association with Stenotrophomonas maltophilia in an In Vitro Model of Mixed Biofilm. PLoS One 2016;11(11):e0166325. [17] Adjemian J, Olivier KN, Prevots DR. Nontuberculous mycobacteria among patients with cystic fibrosis in the United States: screening practices and environmental risk. Am J Respir Crit Care Med 2014;190(5):581–6. [18] Jones GC, Sainsbury CA. A practical approach to glucose abnormalities in cystic fibrosis. Diabetes Ther 2016;7(4):611–20. [19] Wickens-Mitchell KL, Gilchrist FJ, McKenna D, Raffeeq P, Lenney W. The screening and diagnosis of cystic fibrosis-related diabetes in the United Kingdom. J Cyst Fibros 2014;13(5):589–92. [20] Goss CH, Otto K, Aitken ML, Rubenfeld GD. Detecting Stenotrophomonas maltophilia does not reduce survival of patients with cystic fibrosis. Am J Respir Crit Care Med 2002;166(3):356–61. [21] Waters V, Atenafu EG, 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(5):482–6.