Predicting the course of nutrition and lung disease in infants and children with cystic fibrosis

Abstract

As CF lung disease is a progressive condition and there is heterogeneity among patients in the impact of disease that cannot be attributed solely to their CFTR genotype, we have long sought to understand factors that predict clinical outcomes in infants and children with CF. In this issue of the Journal of Cystic Fibrosis, three articles from three prospective cohort studies offer new insights, while raising key questions. Atteih [[1]Atteih S.E. Raraigh K.S. Blackman S.M. Cutting G.R. Collaco J.M. Predictive effects of low birth weight and small for gestational age status on respiratory and nutritional outcomes in cystic fibrosis.J Cyst Fibros. 2020; 19: 888-895Abstract Full Text Full Text PDF PubMed Scopus (2) Google Scholar] evaluated prevalence of and risk factors for low birth weight (LBW, <2500 g) or small for gestational age (SGA, < 10th percentile) status, and effects of LBW and SGA on nutritional status and lung function over time. They studied 1677 children with CF, born in 2000–2013, who were enrolled in the US CF Twin and Sibling Study [[2]Vanscoy L.L. Blackman S.M. Collaco J.M. Bowers A. Lai T. Naughton K. Algire M. McWilliams R. Beck S. Hoover-Fong J. et al.Heritability of lung disease severity in cystic fibrosis.Am J Respir Crit Care Med. 2007; 175: 1036-1043Crossref PubMed Scopus (155) Google Scholar]. The study enrolled 2086 children from 1018 families attending US CF Centers. Birth weight and gestational age were collected from medical records or by parental recall; data were missing in 409 participants. Linear regressions were used to assess relationships between LBW or SGA status to FEV1 percent predicted (ppFEV1) and BMI z-score, adjusting for possible confounders and clustered by family. Linear and logistic regression analyses were used to explore contributors to LBW and SGA status, excluding twins given the known association between birth weight and multiple gestation. Outcome measures were maximum ppFEV1 measured at age 6–7, 12–13 and 18–19 years, and BMI percentile based on CDC values at age 2–3 years and again during the same time periods. Measurements were excluded after lung transplant or implementation of CFTR modulator therapy. There are several notable results from this study. Overall, LBW and SGA rates were similar to those in the US general population, in contrast to other recent reports [3Darrah R. Nelson R. Damato E.G. Decker M. Matthews A. Hodges C.A. Growth deficiency in cystic fibrosis is observable at birth and predictive of early pulmonary function.Biol Res Nurs. 2016; 18: 498-504Crossref PubMed Scopus (11) Google Scholar, 4Ramos K.J. Sack C.S. Mitchell K.H. Goss C.H. Starr J.R. Cystic fibrosis is associated with adverse neonatal outcomes in Washington State, 1996–2013.J Pediatr. 2017; 180: 206-211.e1Abstract Full Text Full Text PDF PubMed Scopus (6) Google Scholar, 5Schlüter D.K. Griffiths R. Adam A. Akbari A. Heaven M.L. Paranjothy S. Nybo Andersen A.M. Carr S.B. Pressler T. Diggle P.J. et al.Impact of cystic fibrosis on birthweight: a population based study of children in Denmark and Wales.Thorax. 2019; 74: 447-454Crossref PubMed Scopus (16) Google Scholar]. For example, a study from Denmark and Wales [[5]Schlüter D.K. Griffiths R. Adam A. Akbari A. Heaven M.L. Paranjothy S. Nybo Andersen A.M. Carr S.B. Pressler T. Diggle P.J. et al.Impact of cystic fibrosis on birthweight: a population based study of children in Denmark and Wales.Thorax. 2019; 74: 447-454Crossref PubMed Scopus (16) Google Scholar] that showed a more than two-fold increase in prematurity and LBW in infants with CF. However, there are important differences in both methodology and populations between these studies. The birth weight data sources for the Danish/ Welsh population study were linked directly to hospital records, birth certificate data were used in a study from Washington State [[4]Ramos K.J. Sack C.S. Mitchell K.H. Goss C.H. Starr J.R. Cystic fibrosis is associated with adverse neonatal outcomes in Washington State, 1996–2013.J Pediatr. 2017; 180: 206-211.e1Abstract Full Text Full Text PDF PubMed Scopus (6) Google Scholar], and medical record review or parental interview were used in a single center study from Ohio [[3]Darrah R. Nelson R. Damato E.G. Decker M. Matthews A. Hodges C.A. Growth deficiency in cystic fibrosis is observable at birth and predictive of early pulmonary function.Biol Res Nurs. 2016; 18: 498-504Crossref PubMed Scopus (11) Google Scholar], similar to the current report. While CFTR genotype was not related to birth weight in the report of Atteih, pancreatic insufficiency was associated with a 220 g lower birth weight; CFTR genotype is highly predictive of pancreatic status, and without complete data on these data the studies cannot be directly compared. Nevertheless, authors of the Danish/Welsh study concluded that the prognostic value of birth weight required further assessment given the marked influence of nutritional indices on pulmonary function [[6]Konstan M.W. Butler S.M. Wohl M.E. Stoddard M. Matousek R. Wagener J.S. Johnson C.A. Morgan W.J. Fibrosis IaCotESoCGrowth and nutritional indexes in early life predict pulmonary function in cystic fibrosis.J Pediatr. 2003; 142: 624-630Abstract Full Text Full Text PDF PubMed Scopus (319) Google Scholar] and survival [[7]Yen E.H. Quinton H. Borowitz D. Better nutritional status in early childhood is associated with improved clinical outcomes and survival in patients with cystic fibrosis.J Pediatr. 2013; 162: 530-535Abstract Full Text Full Text PDF PubMed Scopus (209) Google Scholar,[8]McColley S.A. Schechter M.S. Morgan W.J. Pasta D.J. Craib M.L. Konstan M.W. Risk factors for mortality before age 18 years in cystic fibrosis.Pediatr Pulmonol. 2017; 52: 909-915Crossref PubMed Scopus (54) Google Scholar, and the current study sheds considerable light on this issue. Low birth weight was associated with lower BMI percentile at age 2, 6, and 12 years, but not at 18 years. Of note, the BMI percentile at 18 was only available in 88 participants, and the p value was 0.013 with an a priori significance value of 0.0125 due to appropriate adjustments for multiple comparisons. This suggests potential influence of LBW on BMI at age 18, with a limitation of sample size and interpretation based on p-value, a current topic of robust discussion by statisticians [[9]Goodman S.N. Why is getting rid of p-values so hard? Musings on science and statistics.Am Stat. 2019; 73: 26-30Crossref Scopus (27) Google Scholar]. Small for gestational age also had a negative effect on BMI earlier in life but not at 18 years, but only 109 singleton births were SGA. Among LBW infants, prematurity had a negative effect on BMI z-score at 6 years but not at other time points. Each kilogram increase in birthweight was associated with a 12.8 percentile greater BMI at age 2, but only 5 percentile points at age 18 years. Interestingly a post-hoc analysis showed that LBW and SGA status were associated with lower height z-scores at all ages. There was no association between LBW or SGA status and maximum ppFEV1 at any time point. Each 1 kg increase in birth weight increased ppFEV1 by 2.9% at age 6 years, but there was no effect at other ages. The authors note that CF Center nutritional interventions may contribute to improved BMI over time, but not influence linear growth. The relevance of nutritional status at birth to long term outcomes shown in Atteih may change over time due to implementation of newborn screening. The cohort was recruited during a time period that overlapped with widespread implementation of newborn screening in the United States, which was completed in late 2009. A major benefit of newborn screening is improved nutrition from the time of diagnosis through childhood [[10]Farrell P.M. Kosorok M.R. Rock M.J. Laxova A. Zeng L. Lai H.C. Hoffman G. Laessig R.H. Splaingard M.L. Early diagnosis of cystic fibrosis through neonatal screening prevents severe malnutrition and improves long-term growth. Wisconsin Cystic Fibrosis Neonatal Screening Study Group.Pediatrics. 2001; 107: 1-13Crossref PubMed Scopus (402) Google Scholar], and the impact of contemporary newborn screening on childhood outcomes in larger populations is only currently being elucidated [[11]Schlüter D.K. Southern K.W. Dryden C. Diggle P. Taylor-Robinson D. Impact of newborn screening on outcomes and social inequalities in cystic fibrosis: a UK CF registry-based study.Thorax. 2020; 75: 123-131Crossref PubMed Scopus (16) Google Scholar]. Early reports of newborn screening outcomes showed less improvement in the trajectory of lung disease, and described airways inflammation and obstruction in the first year of life [[12]Sly P.D. Brennan S. Gangell C. de Klerk N. Murray C. Mott L. Stick S.M. Robinson P.J. Robertson C.F. Ranganathan S.C. et al.Lung disease at diagnosis in infants with cystic fibrosis detected by newborn screening.Am J Respir Crit Care Med. 2009; 180: 146-152Crossref PubMed Scopus (435) Google Scholar], but another report in this issue of the Journal has more optimistic findings. Davies et al. [[13]Davies G. Thia L.P. Stocks J. Bush A. Hoo A.F. Wade A. Nguyen T.T.D. Brody A.S. Calder A. Klein N.J. et al.Minimal change in structural, functional and inflammatory markers of lung disease in newborn screened infants with cystic fibrosis at one year.J Cyst Fibros. 2020; 19: 896-901Abstract Full Text Full Text PDF PubMed Scopus (14) Google Scholar] studied 65 (of 76 eligible) infants born 2009–2011 who were diagnosed by newborn screening and receiving care at one of six CF centers in London. All received prophylactic flucloxacillin. Infant pulmonary function testing, including multiple breath washout, plethysmography and raised volume thoracic compression, volume-controlled chest CT, and flexible bronchoscopy with bronchoalveolar lavage (BAL) were undertaken around 1 year of age at a time of clinical stability. Infants were diagnosed at a median of 3.4 weeks (interquartile range, 3.0–4.6) and while the vast majority (61/65) were pancreatic insufficient, were well nourished with weight and height z-scores greater than zero at the time of assessment. During the first year of life, wheeze or crackles were detected in 34% and Pseudomonas aeruginosa (PA) infection detected by cough swab in 32% of children, and 26% had received an intravenous antibiotic, with 4/17 receiving two and 4/17 receiving three courses. One had chronic PA by Leeds criteria [[14]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: 29-34Abstract Full Text Full Text PDF PubMed Scopus (510) Google Scholar]. In spite of this overall severe phenotype, results of the comprehensive pulmonary assessment showed surprisingly mild disease. At the time of BAL, 74% of infants had no growth of pathogenic bacteria, and only 3 (5%) had PA isolated. In a subset of 43 infants who had evaluable samples, 10 (23%) had detectable free neutrophil elastase (NE). The median IL-8 in 45 infants was 286 pg/ml (IQR, 82–1209). Pulmonary function testing showed mild elevations in LCI and FRCpleth and mildly reduced FEV0.5, while the median air trapping score on CT was zero out of a maximum 27 (range, 0–15.5) using the Brody-II CF-CT scoring system [[15]Brody A.S. Kosorok M.R. Li Z. Broderick L.S. Foster J.L. Laxova A. Bandla H. Farrell P.M. Reproducibility of a scoring system for computed tomography scanning in cystic fibrosis.J Thorac Imaging. 2006; 21: 14-21Crossref PubMed Scopus (143) Google Scholar]. There was a weak correlation between CT score and LCI. Overall, there was little or no association between infection, inflammation, and CT or pulmonary function parameters. The LCI z-score was elevated in infants with significant pathogens on BAL, but not those with a history of infection. The authors note that pro-inflammatory cytokine levels [[16]Pillarisetti N. Williamson E. Linnane B. Skoric B. Robertson C.F. Robinson P. Massie J. Hall G.L. Sly P. Stick S. et al.Infection, inflammation, and lung function decline in infants with cystic fibrosis.Am J Respir Crit Care Med. 2011; 184: 75-81Crossref PubMed Scopus (220) Google Scholar] and rates of infection [[17]Sly P.D. Gangell C.L. Chen L. Ware R.S. Ranganathan S. Mott L.S. Murray C.P. Stick S.M. Investigators A.C. Risk factors for bronchiectasis in children with cystic fibrosis.N Engl J Med. 2013; 368: 1963-1970Crossref PubMed Scopus (405) Google Scholar] were similar to those found in the AREST-CF study, while overall CT and pulmonary function abnormalities were less severe. They note that there were differences in pulmonary function equipment and reference ranges, as well as CT scoring protocols; nevertheless, the abnormalities in the current study were overall milder. They also noted that prior findings of correlations between LCI and chest CT abnormalities [[18]Owens C.M. Aurora P. Stanojevic S. Bush A. Wade A. Oliver C. Calder A. Price J. Carr S.B. Shankar A. et al.Lung Clearance Index and HRCT are complementary markers of lung abnormalities in young children with CF.Thorax. 2011; 66: 481-488Crossref PubMed Scopus (156) Google Scholar] were not replicated in this study of infants diagnosed by NBS. A third report in this issue [[19]Horati H. Janssens H.M. Margaroli C. Veltman M. Stolarczyk M. Kilgore M.B. Chou J. Peng L. Tiddens H.A.M.W. Chandler J.D. et al.Airway profile of bioactive lipids predicts early progression of lung disease in cystic fibrosis.J Cyst Fibros. 2020; 19: 902-909Abstract Full Text Full Text PDF PubMed Scopus (10) Google Scholar] evaluated bioactive lipids in bronchoalveolar lavage fluid and chest CT longitudinally who were 1 or 3 years old at initial evaluation and had a maximum of 2 more evaluations at age 3 and 5 years. The initial PRAGMA-CF [[20]Rosenow T. Oudraad M.C. Murray C.P. Turkovic L. Kuo W. de Bruijne M. Ranganathan S.C. Tiddens H.A. Stick S.M. CF ARESTfCFA. PRAGMA-CFA quantitative structural lung disease computed tomography outcome in young children with cystic fibrosis.Am J Respir Crit Care Med. 2015; 191: 1158-1165Crossref PubMed Scopus (160) Google Scholar] scores CT scores were low with a mean disease score of 1.75% (range, 0.16–5.39) and bronchiectasis score of 0.07% (range, 0–0.62). There was mild progression after 2 years including 4/9 infants without initial bronchiectasis developing it over the 2-year period. Initial PRAGMA-CT scores were associated with NE and myeloperoxidase (MPO) activity, and negatively with alveolar macrophage percentage, but were not associated with sex, CFTR mutation, pancreatic status, bacterial culture results or antibiotic treatment. Bioactive lipid families were measurable in all BALF samples by high performance liquid chromatography with mass spectroscopy [[21]Scholte B.J. Horati H. Veltman M. Vreeken R.J. Garratt L.W. Tiddens H.A.W.M. Janssens H.M. Stick S.M. CF ARESTfCFAOxidative stress and abnormal bioactive lipids in early cystic fibrosis lung disease.J Cyst Fibros. 2019; 18: 781-789Abstract Full Text Full Text PDF PubMed Scopus (20) Google Scholar], and were correlated with structural lung damage by chest CT. In particular, percent disease (%Dis) scores were correlated with the lysolipid principal components lysophosphatidic acid (LPA), lysophosphatidylcholines (LPC) and the sphingosine principal component, comprising the ceramide precursors sphingomyelin, sphinganine, and sphingosine. Lipid levels correlated with inflammatory markers, including NE, MPO, and a panel of cytokines and growth factors that control proinflammatory and tissue remodeling signaling. The LPC and ceramide precursor compound variables and Prostaglandin E2 measured in BALF at the first evaluation correlated significantly with disease progression as measured by the%Dis score on CT. This suggests that some of the airway bioactive lipids are associated with progression of structural lung disease in infants. Finally, the investigators report that treatment with an LPA agonist enhances cytokine and growth factor shedding in human bronchial epithelial cells in vitro in a series of experiments demonstrating the relationship between bioactive lipids and pro-fibrotic and pro-inflammatory factors. How do clinicians and investigators apply these findings to practice and to future studies? First and foremost, findings reinforce that current standards of care, from application of best practices in nutrition [[22]Borowitz D. Baker R.D. Stallings V. Consensus report on nutrition for pediatric patients with cystic fibrosis.J Pediatr Gastroenterol Nutr. 2002; 35: 246-259Crossref PubMed Scopus (465) Google Scholar] to widespread newborn screening and early treatment of affected infants [[23]Borowitz D. Robinson K.A. Rosenfeld M. Davis S.D. Sabadosa K.A. Spear S.L. Michel S.H. Parad R.B. White T.B. Farrell P.M. et al.Cystic fibrosis foundation evidence-based guidelines for management of infants with cystic fibrosis.J Pediatr. 2009; 155: S73-S93Abstract Full Text Full Text PDF PubMed Scopus (277) Google Scholar] are associated with milder disease during childhood. The study of Attheir demonstrates that BMI recovery occurs in LBW/SGA infants, albeit with a persistent reduction in height z-score; the latter is not trivial, since height is also a predictor of survival [[24]Keogh R.H. Seaman S.R. Barrett J.K. Taylor-Robinson D. Szczesniak R. Dynamic prediction of survival in cystic fibrosis: a landmarking analysis using UK patient registry data.Epidemiology. 2019; 30: 29-37Crossref PubMed Scopus (20) Google Scholar]. The studies of Davies and Horati note overall mild lung disease, as assessed by chest CT during early life, and noted associations between inflammatory markers and pulmonary outcomes in infants and children diagnosed by newborn screening, while the Horati study further adds information on lipodomic predictors of progression. Comparing data from these studies to each other, and to previous publications, is difficult. Mode of diagnosis was either clinical or by newborn screening in Atteih and Horati, as in some early reports from AREST-CF [[16]Pillarisetti N. Williamson E. Linnane B. Skoric B. Robertson C.F. Robinson P. Massie J. Hall G.L. Sly P. Stick S. et al.Infection, inflammation, and lung function decline in infants with cystic fibrosis.Am J Respir Crit Care Med. 2011; 184: 75-81Crossref PubMed Scopus (220) Google Scholar]. The age at diagnosis for the studies of Davies and a study with similar design evaluating AREST-CF infants diagnosed by NBS [[12]Sly P.D. Brennan S. Gangell C. de Klerk N. Murray C. Mott L. Stick S.M. Robinson P.J. Robertson C.F. Ranganathan S.C. et al.Lung disease at diagnosis in infants with cystic fibrosis detected by newborn screening.Am J Respir Crit Care Med. 2009; 180: 146-152Crossref PubMed Scopus (435) Google Scholar] were only slightly different; Davies had a median age of diagnosis of 23.8 days while those reported from AREST-CF had a median age of diagnosis at 28 days, making onset of care an unlikely explanation for apparently different outcomes. Details of care regimens are incompletely described, but both studies note that treatment was given for initial isolation of PA. Data on other demographic and environmental factors that influence the course of lung disease [[25]McGarry M.E. Williams W.A. McColley S.A. The demographics of adverse outcomes in cystic fibrosis.Pediatr Pulmonol. 2019; 54: S74-S83PubMed Google Scholar] are largely missing from these reports, with only the Atteih study indicating the presence of children who are racial or ethnic minorities. This is a particular concern given a recent report from the UK noting that NBS does not remediate the influence of social deprivation [[5]Schlüter D.K. Griffiths R. Adam A. Akbari A. Heaven M.L. Paranjothy S. Nybo Andersen A.M. Carr S.B. Pressler T. Diggle P.J. et al.Impact of cystic fibrosis on birthweight: a population based study of children in Denmark and Wales.Thorax. 2019; 74: 447-454Crossref PubMed Scopus (16) Google Scholar]. Variation in precise age at study, equipment and CT scoring systems add to the difficulty in making direct comparisons, and since volume-controlled CT scans require both anesthesia and radiation, longitudinal studies utilizing more frequent measurements to look at short term variation are unfeasible. From the standpoint of ongoing and new CF research, these studies certainly demonstrate the value of and need for ongoing and new prospective cohort studies. Cystic fibrosis care and outcomes are evolving, and as treatment continues to be refined, clinical care and anticipatory guidance to people with CF and their families must be guided by the best available evidence. An obvious example is studying disease progression as more infants and young children are treated with CFTR modulator therapies [26Davies J.C. Cunningham S. Harris W.T. Lapey A. Regelmann W.E. Sawicki G.S. Southern K.W. Robertson S. Green Y. Cooke J. et al.Safety, pharmacokinetics, and pharmacodynamics of ivacaftor in patients aged 2–5 years with cystic fibrosis and a CFTR gating mutation (KIWI): an open-label, single-arm study.Lancet Respir Med. 2016; 4: 107-115Abstract Full Text Full Text PDF PubMed Scopus (222) Google Scholar, 27Rosenfeld M. Wainwright C.E. Higgins M. Wang L.T. McKee C. Campbell D. Tian S. Schneider J. Cunningham S. Davies J.C. et al.Ivacaftor treatment of cystic fibrosis in children aged 12 to <24 months and with a CFTR gating mutation (ARRIVAL): a phase 3 single-arm study.Lancet Respir Med. 2018; 6: 545-553Abstract Full Text Full Text PDF PubMed Scopus (143) Google Scholar, 28Rosenfeld M. Cunningham S. Harris W.T. Lapey A. Regelmann W.E. Sawicki G.S. Southern K.W. Chilvers M. Higgins M. Tian S. et al.An open-label extension study of ivacaftor in children with CF and a CFTR gating mutation initiating treatment at age 2–5 years (KLIMB).J Cyst Fibros. 2019; 18: 838-843Abstract Full Text Full Text PDF PubMed Scopus (68) Google Scholar]. However, in order to increase sample sizes and reproducibility of studies, methods that increase data sharing enabled by mass data storage and analyses must be considered. For example, multisource electronic medical record data can be de-identified and transformed to common data models for large multisite data sets, as in the US-based patient-centered clinical research network (PCORNet) [[29]Fleurence R.L. Curtis L.H. Califf R.M. Platt R. Selby J.V. Brown J.S. Launching PCORnet, a national patient-centered clinical research network.J Am Med Inform Assoc. 2014; 21: 578-582Crossref PubMed Scopus (324) Google Scholar,[30]Corley D.A. Feigelson H.S. Lieu T.A. McGlynn E.A. Building data infrastructure to evaluate and improve quality: PCORnet.J Oncol Pract. 2015; 11: 204-206Crossref PubMed Scopus (37) Google Scholar, and raw data from medical imaging studies such as CT scans can be securely sent to cloud storage [[31]Kagadis G.C. Kloukinas C. Moore K. Philbin J. Papadimitroulas P. Alexakos C. Nagy P.G. Visvikis D. Hendee W.R. Cloud computing in medical imaging.Med Phys. 2013; 40070901Crossref PubMed Scopus (98) Google Scholar]. Sharing of data via online libraries can accelerate metabolomic, including lipodomic, research [[32]July 4. Lipid Maps Lipidomic Gateway. <www.lipidmaps.org>. Accessed 2020 July 4.Google Scholar]. While more complete reporting of demographic data, such as social determinants of health, may be easily achievable in reports from single or multisite studies, data about social and environmental exposures are available in multiple data sources; using life course research strategies [[33]Hanson H.A. Hay W.W. Tobin J.N. Barkin S.L. Atkins M. Karagas M.R. Dozier A.M. Wetmore C. Konstan M.W. Heubi J.E. Opportunities for life course research through the integration of data across clinical and translational research institutes.J Clin Transl Sci. 2018; 2: 156-162Crossref PubMed Google Scholar], combining diverse data sets can be used to shed more light on changes and variation of disease progression in CF. Of course, data storage and analytics can be quite costly, a consideration for investigators and funding agencies alike. Finally, while assessing risk factors for disease progression, especially modifiable ones, it is critical for clinicians and investigators to recognize that the course of CF is unpredictable in individuals. More sophisticated analyses of individual course over time [[34]Szczesniak R.D. Su W. Brokamp C. Keogh R.H. Pestian J.P. Seid M. Diggle P.J. Clancy J.P. Dynamic predictive probabilities to monitor rapid cystic fibrosis disease progression.Stat Med. 2020; 39: 740-756Crossref PubMed Scopus (7) Google Scholar] may be clinically applicable in the future in clinical and research settings. Vigilance in monitoring people with CF closely and providing interventions to remediate growth defects, infection and airway inflammation remain the most essential role of those of us who treat and study CF. The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.