http://www.jhltonline.org
Impact of lung function interpretation approach on pediatric bronchiolitis obliterans syndrome diagnosis after lung transplantation Paul D. Robinson, MBChB, FRCP, PhD,a,b,c Helen Spencer, MBChB, MD,b and Paul Aurora, MBChB, PhDb From the aDepartment of Respiratory Medicine, The Children’s Hospital at Westmead, Sydney, New South Wales, Australia; bDepartment of Cardiothoracic Transplantation, Great Ormond Street Hospital for Children NHS Trust, London, UK; and the cDiscipline of Pediatrics and Child Health, Sydney Medical School, University of Sydney, Sydney, New South Wales, Australia.
KEYWORDS: monitoring; pediatric; lung function; bronchiolitis obliterans syndrome; reference equations; spirometry
BACKGROUND: The diagnostic criteria for bronchiolitis obliterans syndrome (BOS) are predominantly adult-focused. The relationship between application and impact of reference equation choice on pediatric baseline lung function achieved and subsequent BOS diagnosis remains unclear. METHODS: Lung function spirometry data (FEV1, FVC and FEF25–75) from pediatric subjects transplanted at the Great Ormond Street Hospital over a 10-year period were collated. Baseline values achieved after lung transplantation and BOS rates were examined. Raw values were compared with 2 different reference equations (the “Brompton” and modern collated “All-age” equations). The impact of FEF25–75 baseline definition was investigated. RESULTS: Fifty subjects were included, 17 males and 33 females, transplanted at a median (range) age of 14.0 years (3.2 to 17.3 years, 83% 410 years old), and followed for 1,028 (388 to 2,613) days posttransplantation. Raw values underestimated baseline lung function attainment for all indices. Magnitude of baseline lung function was affected by reference equation choice. Mean FEV1 values were: Brompton 97.9% (SD 20.3%) and All-age 86.3% (SD 15.4%) of predicted (p o 0.0001). BOS 0p incidence was significantly higher for All-age predicted than for raw values (64% and 40%, respectively, p ¼ 0.027). Modification of FEF25–75 baseline (to either FEV1 or FVC baseline) led to a reduction in BOS 0p detection (p o 0.01). CONCLUSIONS: Modern collated reference equations should be used for lung function monitoring in pediatric subjects after lung transplantation. Standardization of FEF25–75 baseline definition is urgently required. These data question the utility of the FEF25–75 criterion as an early marker of BOS 0p in pediatric subjects. J Heart Lung Transplant 2015;34:1082–1088 r 2015 International Society for Heart and Lung Transplantation. All rights reserved.
Bronchiolitis obliterans syndrome (BOS) is a leading cause of morbidity and mortality after lung transplantation.1 Reprint requests: Paul D. Robinson, MBChB, FRCP, PhD, Department of Respiratory Medicine, The Children’s Hospital at Westmead, Locked Bag 4001, Westmead, Sydney, NSW 2145, Australia. Telephone: þ61-298453395. Fax: þ61-2-98453396. E-mail address: [email protected]
This clinical surrogate for delayed allograft dysfunction was introduced due to the challenges of histologic diagnosis, and based on the relative decline in lung function (e.g., forced expiratory volume in 1 second [FEV1]) from maximal values after lung transplantation (i.e., the subject’s “baseline” lung function).2,3 Originally, the earliest category of BOS (termed “BOS 1”) was defined once FEV1 had fallen by
1053-2498/$ - see front matter r 2015 International Society for Heart and Lung Transplantation. All rights reserved. http://dx.doi.org/10.1016/j.healun.2015.03.010
Robinson et al.
BOS Diagnosis Post-LTx
Z20% from the baseline value, and severity (Grades 1 to 3) was based on magnitude of FEV1 decline. To facilitate earlier detection, an additional category, BOS 0p, was subsequently introduced, which was defined as an FEV1 decline of Z10% from baseline and/or the additional criterion of a Z25% fall in mid-expiratory flows (FEF25–75) from FEF25–75 baseline value.4 FEF25–75 had been incorporated as this parameter was considered to be a better reflection of more peripheral airway changes, the primary site of the underlying pathophysiology. Several aspects of the BOS definition require clarification, particularly for pediatric subjects. First, the definition of “true” FEF25–75 baseline has been challenged, and a modification proposed to define at the point FEV1 baseline is met, due to potential false early FEF25–75 peaks in adult subjects.5 Given the high lung-volume dependence of FEF25–75 modification to forced vital capacity (FVC) at baseline may be more appropriate. Second, continuing somatic growth for pediatric subjects is not corrected for by absolute or “raw” lung function values. True lung function decline, relative to lung size, may also remain undetected. Pediatric-specific reference equations exist, generating percent-predicted values adjusted for factors such as age, height and gender. There are several different pediatric reference equations, each with their own inherent strengths and weaknesses. Centers often have to switch between reference equations as a subject ages, or use multiple reference equation combinations to cover all indices of interest, and this may have detrimental management implications.6 To overcome these issues, reference data from several large historic cohorts were recently collated to develop “All-age” equations,7 with smooth transition across a wide age range (4 to 80 years) and well-defined lower limits of normal for all ages. The impact of using these more accurate reference equations in this setting has not yet been evaluated. Lung transplantation centers may use a variety of approaches to interpret pediatric lung function data. Pediatric centers typically use percent-predicted values, but reference equation choice may vary as choice of equation has not been standardized. Adult centers transplanting adolescent subjects may use an adult approach that focuses on raw values. The aim of this study was to determine the impact of that variation in practice. Baseline lung function achieved and subsequent BOS diagnosis in children were investigated, based on longitudinal data collected at a single large pediatric transplant center. Current local pediatric practice (current reference equation values) was compared with both potential adult practice (raw values) and updated pediatric practice at the time of this analysis (All-age reference equation values).
Methods A retrospective analysis of lung function data was performed from pediatric subjects undergoing lung or heart–lung transplantation at the Great Ormond Street Hospital (GOSH) over a 10-year period, from 2002 to 2011. Subjects were included if they had Z1 year of
1083 lung function data available after lung transplantation. Medical records were reviewed in accordance with the guidelines of the research ethics committee of the Institute of Child Health and the GOSH for Children NHS Trust. Spirometry was performed according to American Thoracic Society criteria for school-age children8 and using pre-school age range-specific quality control in younger children.9 Collated raw lung function variables were FEV1, FVC, FEF25–75 and FEV1/FVC ratio. To investigate the effect of varying clinical practice, 3 different approaches were taken: the current clinical practice values were calculated using “Brompton” reference equations (termed “Brompton predicted” hereafter), which were the most commonly used equations in the UK at the time of this study10; absolute values were complied for the current adult approach (termed “raw” hereafter); and the updated approach was calculated using the recently collated “All-age” reference equations (termed “All-age predicted” hereafter).7 This latter option was considered the optimal choice. As Brompton reference equations do not include FEF25–75 reference data (only maximum expiratory flow at 50% and 25% of FVC), Brompton predicted FEF25–75 values were not generated for this study. Incorporation of a separate FEF25–75 reference equation was considered by the authors to be an overcomplication of the study design and did not reflect local management practice, which was one of the aims of the study. Baseline was defined, as recommended in the BOS guidelines, as the average of the 2 highest post-transplantation values measured Z3 weeks apart, regardless of how long the gap was between the 2 measurements. The time taken to reach baseline was defined at the time between the second value and the lung transplantation date. For FEF25–75 baseline, 2 further types of baseline were calculated. The first was based on FEV1-modified baseline as recommended by Rosen et al,5 who noted an artificial early peak in FEF25–75 values in some subjects in the initial posttransplantation period. This FEV1-modified baseline for FEF25–75 was calculated as the average of the 2 FEF25–75 values at FEV1 baseline. In addition, an FVC-modified baseline for FEF25–75 was also calculated, defined as the average of the 2 FEF25–75 values at FVC baseline. The current BOS 0p criteria can be achieved by either a persistent decrease in FEV1 of 10% from the post-transplantation baseline FEV1 value or a persistent decrease in FEF25–75 of 25% from the baseline FEF25–75 value.4 To assess the impact of the differing approaches, fulfillment of BOS 0p incidence was examined in 3 ways: by FEV1 criteria alone; by FEF25–75 criteria alone; or by both. BOS 1 was defined based on FEV1 criteria alone, as outlined in recommendations (persistent decrease in FEV1 of 20% of the post-transplantation baseline value).4 For percentpredicted data, relative change in percent-predicted change, not actual change, was used (i.e., a 10% decrease, defined as 80% to 72%, not 80% to 70% predicted). Time to reach BOS was defined as time between baseline achieved and the first value fulfilling the BOS criteria specified. Clinical data for each subject contained within the hospital record was examined across the study period to ensure that other reasons for lung function decrease were not present (e.g., infection), as specified in the BOS recommendations.4 Statistical analysis was performed using SAS, version 9.2 (SAS Institute, Inc., Cary, NC). Proportions for categorical variables were compared using Fisher’s exact test. Kaplan–Meier analyses were used to illustrate time to event occurrence (i.e., baseline achieved or time to BOS 0p or 1). Related-samples Wilcoxon’s signed-rank test was used to compare differences in median values. Cox proportional hazards regression analysis was used to generate
1084
The Journal of Heart and Lung Transplantation, Vol 34, No 8, August 2015
hazard ratios for event rates. GraphPad Prism version 5.00 for Mac O SX (GraphPad Software, San Diego, CA) was used to generate additional data.
Results Fifty-six children received a bilateral lung or heart–lung transplant during the study period, of whom 50 (89%) had 41 year of lung function data available and were included in the study. There were 33 girls (66%) and 17 boys (34%), with a median age of 14.0 (range 3.2 to 17.3) years at the time of transplantation. Median length of follow-up was 1,028 (388 to 2,613) days.
Baseline lung function achieved after lung transplantation FEV1 value did not reach a defined baseline during the study period in all subjects and was affected by the approach used. Use of reference equations to generate percent-predicted data led to a higher level of baseline achievement (Table 1 and Figure 1). For Brompton predicted vs raw values, Cox proportional hazards regression generated a hazard ratio of 1.97 (95% confidence interval [CI] 1.44 to 2.70, p o 0.0001). In other words, the event rate of reaching baseline FEV1 was 97% higher if Brompton predicted was used compared with raw FEV1 values. This was even greater when All-age predicted was used: 2.29 (95% CI 1.68 to
Figure 1 Kaplan–Meier plot of time to reach baseline for FEV1 after lung transplantation, based on raw FEV1 values (solid line), Brompton percent predicted values (dashed line) and All-age predicted values (dotted line). Censored events represent children who reached transition to adult services without reaching a demonstrated FEV1 baseline.
3.13, p o 0.0001). The same overall pattern was seen for both FVC and FEF25–75 values (refer to Figures S1 and S2, respectively, in the Supplementary Material [SuppMat]
Table 1 Comparison of Baseline Lung Function After Transplantation for FEV1, FVC and FEF25–75 Comparing Raw Values to Those Obtained Using Brompton and All-age Reference Equations Raw FEV1 (liters) Baseline reached 35 of 50 (70%) Baseline value 2.26 (0.69) Time to baseline 662 (119–2,044) FVC (liters) Baseline reached 30 of 50 (60%) Baseline value 2.71 (0.82) Time to baseline 642 (157–2,242) FEF25–75 (liters/s) Baseline reached 47 of 50 (94%) Baseline value 2.84 (1.05) Time to baseline 128 (17–1,345) Modified FEF25–75 (to FEV1 baseline, liters/s) Baseline reached 36 of 50 (72%)d Baseline value 2.59 (1.02) Time to baseline 662 (119–2,044) Modified FEF25–75 (to FVC baseline, liters/s) Baseline reached 32 of 50 (64%)d Baseline value 2.44 (1.01) Time to baseline 642 (157–2,242)
Brompton predicted
All-age predicted
44 of 50 (88%)a 97.9 (20.3) 345 (112–1,211)
46 of 50 (92%)b 86.3 (15.4)c 345 (71–1,352)
39 of 50 (78%) 100.0 (17.3) 452 (141–1,234)
44 of 50 (88%)a 89.1 (14.2)c 431 (157–1,234) 46 of 50 (92%) 99.6 (26.9) 107 (17–1,345) 47 of 50 (94%)a 86.9 (26.7) 345 (71–1,352) 44 of 50 (88%)a 77.1 (27.3) 431 (157–1,234)
Raw values expressed in liters and predicted values as percent predicted. Parametric data presented as mean (SD) and non-parametric data presented as median (range), unless noted otherwise. Time to baseline is shown in days. Brompton reference equations do not include FEF25–75 as an outcome and therefore Brompton predicted FEF25–75 values are not available. a p o 0.05 vs raw. b p o 0.01 vs raw. c p o 0.01 vs Brompton predicted. d p o 0.01 vs unmodified FEF25–75 values.
Robinson et al.
BOS Diagnosis Post-LTx
available online at www.jhltonline.org). The hazard ratios for these are given in the legends of the respective figures. Although the proportions of patients achieving baseline did not differ between the reference equations, the FEV1 baseline values achieved were significantly lower using All-age predicted, 86.3% (SD 15.4%), compared with Brompton predicted, 97.9% (SD 20.3%), with p o 0.0001 (Table 1). The same pattern was seen for FVC values. Time to reach baseline differed between lung function indices. Brompton and all-age values, baseline was reached earliest for FEF25–75, followed by FEV1 and then FVC (see Figures S3 and S4 in SuppMat), whereas, for raw values, both FEV1 and FVC were reached at a similarly late time-point (see Figure S5 in SuppMat). Hazard ratios for these are given in the legends of the respective SuppMat figures. Modification of FEF25–75 baseline (to either FEV1 or FVC baseline timepoints) led to later baseline achievement and a significantly smaller proportion of subjects reaching an identified FEF25–75 baseline for both raw and All-age-predicted values (Table 1 and Figure 2, and Figure S6 in SuppMat). For raw values, the hazard ratio for FEF25–75 baseline modified to FEV1 baseline vs unmodified was 0.36 (95% CI 0.24 to 0.54, p o 0.0001), and FEF25–75 baseline modified to FVC baseline vs unmodified was 0.26 (95% 0.16 to 0.41, p o 0.0001). For All-age predicted values, the corresponding values were similar: 0.46 (95% CI 0.32 to 0.66, p o 0.0001) and 0.34 (95% CI 0.22 to 0.52, p o 0.0001), respectively.
1085
Incidence of BOS The effects of variation in approach to BOS 0p are summarized in Table 2. The incidence of BOS 0p was lowest for raw values, but was increased when expressed using reference equations for both FEV1 and FEF25–75 components of the classification criteria. Based on FEV1 criteria alone, the increase was statistically significant for All-age predicted than for raw values (42% vs 18%, p ¼ 0.016), and trended toward statistical significance for Brompton predicted compared with raw values (36% vs 18%, p ¼ 0.07). The difference in BOS 0p diagnosis between reference equations was not statistically significant. The incidence of BOS 0p based on the FEF25–75 criterion alone was greater compared with FEV1 criterion alone for raw values (38% vs 18%, p ¼ 0.04), but not for All-age predicted values (54% vs 42%, p ¼ 0.32). This was not seen with modified FEF25–75 baseline values. The addition of the FEF25–75 criterion to the FEV1 criterion for BOS 0p diagnosis increased BOS 0p incidence from 18% to 40% for raw values (p ¼ 0.027) and from 42% to 64% for All-age predicted values (p ¼ 0.045). Use of Allage predicted values led to a doubling of BOS 0p incidence (Figure 3) compared with raw values (hazard ratio 2.25, 95% CI 1.50 to 3.36, p o 0.0001). The increase in BOS 0p diagnosis with the incorporation of the FEF25–75 criterion was much smaller if a modified FEF25–75 baseline was used (Table 2). This effect was most pronounced with use of FVC-modified FEF25–75 baseline (Table 2 and Figure 4). For time to BOS 0p with All-age predicted values, the hazard ratio, based on modified FEF25–75 baseline (to FEV1 baseline) vs unmodified FEF25–75 baseline, was 0.61 (95% CI 0.43 to 0.86, p ¼ 0.005), and modified to FVC baseline vs unmodified FEF25–75 baseline was 0.50 (95% CI 0.35 to 0.71, p ¼ 0.0001). Incidence of BOS 1 also increased from that seen with raw values (10 of 50, 20%) when Brompton (14 of 50, 28%) or All-age (12 of 50, 24%) predicted values were used, but did not reach statistical significance. The time to reach BOS 1 was significantly shorter with Brompton predicted vs raw values (hazard ratio 1.64, 95% CI 1.06 to 2.53, p ¼ 0.025). Statistical significance was not reached for All-age reference equation values when compared with raw values (hazard ratio 1.24, 95% CI 0.95 to 1.61, p ¼ 0.11) Figure 5.
Discussion
Figure 2 Kaplan–Meier plot of time to reach baseline for all age reference equation FEF25–75 values after lung transplantation. Estimates are shown for time to baseline achievement for raw FEF25–75 values (solid line), values modified to FEV1 baseline time-point (dashed line) and values modified to FVC baseline timepoint (dotted line). Censored events represent children who reached transition to adult services without reaching a demonstrated baseline.
In this study we have provided an initial detailed description of lung function trends across a pediatric cohort after lung transplantation, and have outlined several major effects when the approach to expressing data is varied. The impact of reference equation on the timing and magnitude of baseline lung function achieved is illustrated, as well as the subsequent effect on BOS incidence. Raw lung function values led to poor estimation of FEV1, FVC and FEF25–75 baseline and an underestimation of the true BOS 0p incidence. Choice of reference equation also affected the magnitude of baseline lung function achieved and subsequent BOS 0p incidence. FEF25–75 baseline definition
1086 Table 2
The Journal of Heart and Lung Transplantation, Vol 34, No 8, August 2015 Comparison of BOS 0p Diagnosis Based on Approach Utilizedb Raw
Using FEV1 criterion alone Incidence 9 of 50 (18%) Time to BOS 0p 1,072 (367–1,604) Using FEF25–75 criterion alone Incidence 19 of 50 (38%) Time to BOS 0p 562 (85–1,604) Using modified FEF25–75 baseline (to FEV1 baseline) 11 of 50 (22%) Time to BOS 0p 733 (367–1,604) Using modified FEF25–75 baseline (to FVC baseline) 10 of 50 (20%) Time to BOS 0p 895 (367–1,604) Using full criteria of either FEV1 or FEF25–75 criterion being met Incidence 20 of 50 (40%) Time to BOS 0p 540 (85–1,604) 12 of 50 (24%) Using modified FEF25–75 baseline (to FEV1 baseline) Time to BOS 0p 726 (367–1,604) Using modified FEF25–75 baseline (to FVC baseline) 11 of 50 (22%) Time to BOS 0p 718 (367–1,604) BOS 0p fulfilling FEF25–75 criterion earlier than FEV1 Incidence 11 of 20 (55%) Using modified FEF25–75 baseline (to FEV1 baseline) 3 of 12 (25%) Using modified FEF25–75 baseline (to FVC baseline) 3 of 11 (27%)
Brompton predicted
All-age predicted
18 of 50 (36%) 760 (367–1,604)
21 of 50 (42%)a 896 (173–1,541) 27 of 50 (54%) 556 (85–2,241) 17 of 50 (34%) 896 (265–2,241) 11 of 50 (22%)b 1072 (367–1,604) 32 of 50 (64%)a 631 (85–1,541) 24 of 50 (48%)a 871 (173–1,541) 21 of 50 (42%)b 896 (173–1,541) 14 of 32 (44%) 4 of 24 (17%)b 0 of 21 (0%)b
The bronchiolitis obliterans syndrome (BOS) 0p classification is shown based on individual FEV1 and FEF25–75 criteria as well as combined criteria. Parametric data presented as mean (SD) and non-parametric data presented as median (range), unless otherwise shown. Time to diagnosis shown in days. a p r 0.05 vs raw. b p r 0.05 vs unmodified FEF25–75 values.
significantly affects BOS 0p incidence. The apparent beneficial effect of using the FEF25–75 criterion was greatly reduced when FEF25–75 baseline definition was modified to FEV1 baseline, and was lost completely with a more
Figure 3 Kaplan–Meier plot of time to reach BOS 0p for raw (solid line) and All-age reference equation data (dotted line), based on either fulfilling FEV1 or FEF25–75 criteria, which was based on unmodified FEF25–75 baseline values. Censored events represent children who reached transition to adult services without reaching a demonstrated BOS 0p diagnosis.
physiologic modification to FVC baseline. These observations clearly show that the approach to lung function monitoring after lung transplantation requires urgent standardization and has significant implications for future BOS detection rates. Baseline lung function achievement after lung transplantation has been poorly described in the literature, for both children and adults. Raw values were found to increase due to ongoing somatic growth, and children may fail to reach a defined FEV1 or FVC baseline, even years after transplantation. This occurred in almost one third of our cohort for FEV1 baseline, despite a median (range) follow-up of 1,115 (500 to 2,613) days in these subjects. Acute decreases in lung function (e.g., acute rejection) occur over short periods, during which minimal somatic growth may occur, and these decreases are detectable irrespective of the different methods used to report lung function. However, lung function loss in chronic rejection typically occurs at a slower rate, and BOS identification requires accurate identification of baseline lung function achieved. Correction for somatic growth by using reference equations improved baseline identification (only 8% failed to reach an FEV1 baseline for All-age data by the end of followup). Baseline lung function may not be achieved for a significant period in children: for All-age predicted values, only 14% and 58% reached a formal FEV1 baseline by 6 months and 1 year post-transplantation, respectively. This suggests long-term growth of the donor organ, beyond simple recruitment during the post-operative period. It remains unclear whether this represents true donor organ growth or
Robinson et al.
BOS Diagnosis Post-LTx
Figure 4 Kaplan–Meier plot showing the effect of choice of FEF25–75 baseline on subsequent time to reach BOS 0p for all age reference equation data. FEF25–75 values (solid line) and those modified to FEV1 baseline time-point (dashed line) and modified to FVC baseline time-point (dotted line) are shown. Censored events represent children who reached transition to adult services without reaching a demonstrated FEF25–75 baseline.
simple graft expansion, as suggested by living related lobar transplant data.11 Reference equation choice is important and it affects BOS incidence. Comparison of a historic local reference equation (i.e., Brompton predicted) and the recently collated improved
Figure 5 Kaplan–Meier plot of time to reach BOS 1 diagnosis after lung transplantation, based on raw FEV1 values (solid line), Brompton predicted values (dashed line) and All-age predicted values (dotted line). Censored events represent children who reached transition to adult services without reaching a demonstrated BOS 1 diagnosis.
1087 All-age reference equations highlight 2 important findings. First, magnitude of baseline FEV1 (and FVC) achieved may be overestimated by historic reference equations. The Brompton equations do not adjust for age, so children who are short for age (a large proportion of children undergoing lung transplantation) have underestimated predicted values and, therefore, overestimated percent-predicted values. This difference was most noticeable in 2 age groups: younger children (r7 years) and adolescents (Figure 6). In the younger children, this probably reflects relatively low subject numbers in this age range of the original Brompton reference data set, weakening the accuracy of predicted values. Discrepancies in adolescence are significant as this the most common age group for pediatric lung transplantation.1 Second, overestimation of FEV1 percentpredicted baseline may lead to a false BOS diagnosis, due to higher starting baseline and improved accuracy (and therefore reduction) in Brompton predicted reference values as the subject ages (see Figure S7 in SuppMat). All-age equations have now been extended using a larger data set to 3 to 95 years by the Global Lungs Initiative.12 Recommendations to use modern collated reference equations for all pediatric lung transplant subjects extends to both the All-age and Global Lungs Initiative equations.7,12 The failure to show a statistical difference in BOS 0p incidence between the reference equation choices may reflect either Type II error or the fact that relatively few subjects were in the younger age range where differences in baseline estimates were most marked (9 of 50, or 18% of those aged o10 years at the time of transplantation). However, the illustrated cases highlight the potential for inappropriate BOS classification. This is clinically relevant as it may trigger management decisions such as trials of macrolide therapy, as recommended in recent BOS guidelines.3 The present data highlight the importance of the FEF25–75 baseline definition on BOS 0p diagnosis. Initial studies supported incorporation, but they used an unmodified baseline for both adults13,14 and children.15 Its value depends on it being an early marker for subsequent FEV1 decline, and the potential for supranormal values was raised in subsequent adult data, leading the authors to question its utility.16 Our pediatric data reinforce these concerns. Individual cases supporting utility do exist (see Figure S8 in SuppMat), but many subjects were affected by this artificial FEF25–75 baseline early after transplantation, leading to inappropriate BOS 0p diagnosis (see Figure S9 in SuppMat). An artificial peak in FEF25–75 most
Figure 6 Difference in baseline FEV1 between Brompton and All-age predicted values, in comparison to age at the time of FEV1 baseline attainment (by All-age values).
1088
The Journal of Heart and Lung Transplantation, Vol 34, No 8, August 2015
likely occurs due to the volume dependence of the measurement. Thoracic adaptation during this recovery period from the transplant operation may lead to an inability to exhale fully to residual volume, causing FEF25–75 to be measured in a higher lung volume range. Therefore, there are physiologic reasons why standardizing FEF25–75 baseline to establishment of FVC baseline, and not FEV1 baseline, is justified. Attainment of FVC and FEV1 baseline occur at differing time-points in most cases: 62% based on All-age values, with a median difference of 25 (range –518 to 666) days. The impact of this early FEF25–75 signal was greatly reduced when modified to FEV1 baseline, and was lost completely when the more physiologically robust modification to FVC baseline was performed. This study was not intended to validate the accuracy of BOS diagnosis, and the weaknesses of transbronchial biopsies in diagnosing BOS are well recognized. A direct comparison of pediatric BOS 0p rates in this study to available adult data is also challenging, due to the format of data presentation.16 However, median time to BOS diagnosis (either BOS 0p or BOS 1) does appear to be longer in pediatric subjects, even if raw values are used. Standardizing the approach to BOS diagnosis should help with these comparisons in the future. In conclusion, we have reviewed descriptive data of pediatric lung function patterns in a large cohort of pediatric lung transplant subjects from a single center and presented several important findings regarding post-transplantation lung function monitoring. The adult-based approach of using of raw values is not appropriate in the pediatric age group, due to ongoing somatic growth. Children may take a significant period of time to achieve a true lung function baseline. Choice of reference equation affects both baseline achieved and subsequent BOS incidence. Modern collated reference equations represent a significant advancement and help to avoid the potential misdiagnosis of BOS, particularly in younger subjects. The data presented herein reinforce the need for an urgent standardization of the FEF25–75 baseline definition and also question the added utility of this criterion in BOS 0p diagnosis.
Disclosure statement The authors have no conflicts of interest to disclose. P.R. was funded during his time at the Great Ormond Street Hospital by a generous donation from David Mackintosh to the Children’s Hospital at Westmead.
Acknowledgements The authors would like to thank Elisabeth Barnes (University of Sydney), who assisted with the statitsical analysis on this manuscript.
Supplementary material Supplementary data associated with this article can be found in the online version at www.jhltonline.org.
References 1. Benden C, Goldfarb SB, Edwards LB, et al. The registry of the International Society for Heart and Lung Transplantation: seventeenth official pediatric lung and heart–lung transplantation report—2014; focus theme: retransplantation. J Heart Lung Transplant 2014;33: 1025-33. 2. Cooper JD, Billingham M, Egan T, et al. A working formulation for the standardization of nomenclature and for clinical staging of chronic dysfunction in lung allografts. J Heart Lung Transplant 1993;12:713-6. 3. Meyer KC, Raghu G, Verleden GM, et al. An international ISHLT/ATS/ERS clinical practice guideline: diagnosis and management of bronchiolitis obliterans syndrome. Eur Respir J 2014;44: 1479-503. 4. Estenne M, Maurer JR, Boehler A, et al. Bronchiolitis obliterans syndrome 2001: an update of the diagnostic criteria. J Heart Lung Transplant 2002;21:297-310. 5. Rosen JB, Smith EO, Schecter MG, et al. Decline in 25% to 75% forced expiratory flow as an early predictor of chronic airway rejection in pediatric lung transplant recipients. J Heart Lung Transplant 2012;31:1288-92. 6. Kirkby J, Aurora P, Spencer H, et al. Stitching and switching: the impact of discontinuous lung function reference equations. Eur Respir J 2012;39:1256-7. 7. Stanojevic S, Wade A, Stocks J, et al. Reference ranges for spirometry across all ages: a new approach. Am J Respir Crit Care Med 2008;177:253-60. 8. American Thoracic Society. Standardization of spirometry, 1994 update. Am J Respir Crit Care Med 1995;152:1107-36. 9. Beydon N, Davis SD, Lombardi E, et al. An official American Thoracic Society/European Respiratory Society statement: pulmonary function testing in preschool children. Am J Respir Crit Care Med 2007;175:1304-45. 10. Rosenthal M, Bain SH, Cramer D, et al. Lung function in white children aged 4 to 19 years: I—Spirometry. Thorax 1993;48:794-802. 11. Sritippayawan S, Keens TG, Horn MV, et al. Does lung growth occur when mature lobes are transplanted into children? Pediatr Transplant 2002;6:500-4. 12. Quanjer PH, Stanojevic S, Cole TJ, 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;40:1324-43. 13. Patterson GM, Wilson S, Whang JL, et al. Physiologic definitions of obliterative bronchiolitis in heart–lung and double lung transplantation: a comparison of the forced expiratory flow between 25% and 75% of the forced vital capacity and forced expiratory volume in one second. J Heart Lung Transplant 1996;15:175-81. 14. Valentine VG, Robbins RC, Berry GJ, et al. Actuarial survival of heart–lung and bilateral sequential lung transplant recipients with obliterative bronchiolitis. J Heart Lung Transplant 1996;15:371-83. 15. Sritippayawan S, Keens TG, Horn MV, et al. What are the best pulmonary function test parameters for early detection of post-lung transplant bronchiolitis obliterans syndrome in children? Pediatr Transplant 2003;7:200-3. 16. Hachem RR, Chakinala MM, Yusen RD, et al. The predictive value of bronchiolitis obliterans syndrome stage 0-p. Am J Respir Crit Care Med 2004;169:468-72.
Impact of lung function interpretation approach on pediatric bronchiolitis obliterans syndrome diagnosis after lung transplantation Paul D. Robinson, MBChB, FRCP, PhD,a,b,c Helen Spencer, MBChB, MD,b and Paul Aurora, MBChB, PhDb From the aDepartment of Respiratory Medicine, The Children’s Hospital at Westmead, Sydney, New South Wales, Australia; bDepartment of Cardiothoracic Transplantation, Great Ormond Street Hospital for Children NHS Trust, London, UK; and the cDiscipline of Pediatrics and Child Health, Sydney Medical School, University of Sydney, Sydney, New South Wales, Australia.
KEYWORDS: monitoring; pediatric; lung function; bronchiolitis obliterans syndrome; reference equations; spirometry
BACKGROUND: The diagnostic criteria for bronchiolitis obliterans syndrome (BOS) are predominantly adult-focused. The relationship between application and impact of reference equation choice on pediatric baseline lung function achieved and subsequent BOS diagnosis remains unclear. METHODS: Lung function spirometry data (FEV1, FVC and FEF25–75) from pediatric subjects transplanted at the Great Ormond Street Hospital over a 10-year period were collated. Baseline values achieved after lung transplantation and BOS rates were examined. Raw values were compared with 2 different reference equations (the “Brompton” and modern collated “All-age” equations). The impact of FEF25–75 baseline definition was investigated. RESULTS: Fifty subjects were included, 17 males and 33 females, transplanted at a median (range) age of 14.0 years (3.2 to 17.3 years, 83% 410 years old), and followed for 1,028 (388 to 2,613) days posttransplantation. Raw values underestimated baseline lung function attainment for all indices. Magnitude of baseline lung function was affected by reference equation choice. Mean FEV1 values were: Brompton 97.9% (SD 20.3%) and All-age 86.3% (SD 15.4%) of predicted (p o 0.0001). BOS 0p incidence was significantly higher for All-age predicted than for raw values (64% and 40%, respectively, p ¼ 0.027). Modification of FEF25–75 baseline (to either FEV1 or FVC baseline) led to a reduction in BOS 0p detection (p o 0.01). CONCLUSIONS: Modern collated reference equations should be used for lung function monitoring in pediatric subjects after lung transplantation. Standardization of FEF25–75 baseline definition is urgently required. These data question the utility of the FEF25–75 criterion as an early marker of BOS 0p in pediatric subjects. J Heart Lung Transplant 2015;34:1082–1088 r 2015 International Society for Heart and Lung Transplantation. All rights reserved.
Bronchiolitis obliterans syndrome (BOS) is a leading cause of morbidity and mortality after lung transplantation.1 Reprint requests: Paul D. Robinson, MBChB, FRCP, PhD, Department of Respiratory Medicine, The Children’s Hospital at Westmead, Locked Bag 4001, Westmead, Sydney, NSW 2145, Australia. Telephone: þ61-298453395. Fax: þ61-2-98453396. E-mail address: [email protected]
This clinical surrogate for delayed allograft dysfunction was introduced due to the challenges of histologic diagnosis, and based on the relative decline in lung function (e.g., forced expiratory volume in 1 second [FEV1]) from maximal values after lung transplantation (i.e., the subject’s “baseline” lung function).2,3 Originally, the earliest category of BOS (termed “BOS 1”) was defined once FEV1 had fallen by
1053-2498/$ - see front matter r 2015 International Society for Heart and Lung Transplantation. All rights reserved. http://dx.doi.org/10.1016/j.healun.2015.03.010
Robinson et al.
BOS Diagnosis Post-LTx
Z20% from the baseline value, and severity (Grades 1 to 3) was based on magnitude of FEV1 decline. To facilitate earlier detection, an additional category, BOS 0p, was subsequently introduced, which was defined as an FEV1 decline of Z10% from baseline and/or the additional criterion of a Z25% fall in mid-expiratory flows (FEF25–75) from FEF25–75 baseline value.4 FEF25–75 had been incorporated as this parameter was considered to be a better reflection of more peripheral airway changes, the primary site of the underlying pathophysiology. Several aspects of the BOS definition require clarification, particularly for pediatric subjects. First, the definition of “true” FEF25–75 baseline has been challenged, and a modification proposed to define at the point FEV1 baseline is met, due to potential false early FEF25–75 peaks in adult subjects.5 Given the high lung-volume dependence of FEF25–75 modification to forced vital capacity (FVC) at baseline may be more appropriate. Second, continuing somatic growth for pediatric subjects is not corrected for by absolute or “raw” lung function values. True lung function decline, relative to lung size, may also remain undetected. Pediatric-specific reference equations exist, generating percent-predicted values adjusted for factors such as age, height and gender. There are several different pediatric reference equations, each with their own inherent strengths and weaknesses. Centers often have to switch between reference equations as a subject ages, or use multiple reference equation combinations to cover all indices of interest, and this may have detrimental management implications.6 To overcome these issues, reference data from several large historic cohorts were recently collated to develop “All-age” equations,7 with smooth transition across a wide age range (4 to 80 years) and well-defined lower limits of normal for all ages. The impact of using these more accurate reference equations in this setting has not yet been evaluated. Lung transplantation centers may use a variety of approaches to interpret pediatric lung function data. Pediatric centers typically use percent-predicted values, but reference equation choice may vary as choice of equation has not been standardized. Adult centers transplanting adolescent subjects may use an adult approach that focuses on raw values. The aim of this study was to determine the impact of that variation in practice. Baseline lung function achieved and subsequent BOS diagnosis in children were investigated, based on longitudinal data collected at a single large pediatric transplant center. Current local pediatric practice (current reference equation values) was compared with both potential adult practice (raw values) and updated pediatric practice at the time of this analysis (All-age reference equation values).
Methods A retrospective analysis of lung function data was performed from pediatric subjects undergoing lung or heart–lung transplantation at the Great Ormond Street Hospital (GOSH) over a 10-year period, from 2002 to 2011. Subjects were included if they had Z1 year of
1083 lung function data available after lung transplantation. Medical records were reviewed in accordance with the guidelines of the research ethics committee of the Institute of Child Health and the GOSH for Children NHS Trust. Spirometry was performed according to American Thoracic Society criteria for school-age children8 and using pre-school age range-specific quality control in younger children.9 Collated raw lung function variables were FEV1, FVC, FEF25–75 and FEV1/FVC ratio. To investigate the effect of varying clinical practice, 3 different approaches were taken: the current clinical practice values were calculated using “Brompton” reference equations (termed “Brompton predicted” hereafter), which were the most commonly used equations in the UK at the time of this study10; absolute values were complied for the current adult approach (termed “raw” hereafter); and the updated approach was calculated using the recently collated “All-age” reference equations (termed “All-age predicted” hereafter).7 This latter option was considered the optimal choice. As Brompton reference equations do not include FEF25–75 reference data (only maximum expiratory flow at 50% and 25% of FVC), Brompton predicted FEF25–75 values were not generated for this study. Incorporation of a separate FEF25–75 reference equation was considered by the authors to be an overcomplication of the study design and did not reflect local management practice, which was one of the aims of the study. Baseline was defined, as recommended in the BOS guidelines, as the average of the 2 highest post-transplantation values measured Z3 weeks apart, regardless of how long the gap was between the 2 measurements. The time taken to reach baseline was defined at the time between the second value and the lung transplantation date. For FEF25–75 baseline, 2 further types of baseline were calculated. The first was based on FEV1-modified baseline as recommended by Rosen et al,5 who noted an artificial early peak in FEF25–75 values in some subjects in the initial posttransplantation period. This FEV1-modified baseline for FEF25–75 was calculated as the average of the 2 FEF25–75 values at FEV1 baseline. In addition, an FVC-modified baseline for FEF25–75 was also calculated, defined as the average of the 2 FEF25–75 values at FVC baseline. The current BOS 0p criteria can be achieved by either a persistent decrease in FEV1 of 10% from the post-transplantation baseline FEV1 value or a persistent decrease in FEF25–75 of 25% from the baseline FEF25–75 value.4 To assess the impact of the differing approaches, fulfillment of BOS 0p incidence was examined in 3 ways: by FEV1 criteria alone; by FEF25–75 criteria alone; or by both. BOS 1 was defined based on FEV1 criteria alone, as outlined in recommendations (persistent decrease in FEV1 of 20% of the post-transplantation baseline value).4 For percentpredicted data, relative change in percent-predicted change, not actual change, was used (i.e., a 10% decrease, defined as 80% to 72%, not 80% to 70% predicted). Time to reach BOS was defined as time between baseline achieved and the first value fulfilling the BOS criteria specified. Clinical data for each subject contained within the hospital record was examined across the study period to ensure that other reasons for lung function decrease were not present (e.g., infection), as specified in the BOS recommendations.4 Statistical analysis was performed using SAS, version 9.2 (SAS Institute, Inc., Cary, NC). Proportions for categorical variables were compared using Fisher’s exact test. Kaplan–Meier analyses were used to illustrate time to event occurrence (i.e., baseline achieved or time to BOS 0p or 1). Related-samples Wilcoxon’s signed-rank test was used to compare differences in median values. Cox proportional hazards regression analysis was used to generate
1084
The Journal of Heart and Lung Transplantation, Vol 34, No 8, August 2015
hazard ratios for event rates. GraphPad Prism version 5.00 for Mac O SX (GraphPad Software, San Diego, CA) was used to generate additional data.
Results Fifty-six children received a bilateral lung or heart–lung transplant during the study period, of whom 50 (89%) had 41 year of lung function data available and were included in the study. There were 33 girls (66%) and 17 boys (34%), with a median age of 14.0 (range 3.2 to 17.3) years at the time of transplantation. Median length of follow-up was 1,028 (388 to 2,613) days.
Baseline lung function achieved after lung transplantation FEV1 value did not reach a defined baseline during the study period in all subjects and was affected by the approach used. Use of reference equations to generate percent-predicted data led to a higher level of baseline achievement (Table 1 and Figure 1). For Brompton predicted vs raw values, Cox proportional hazards regression generated a hazard ratio of 1.97 (95% confidence interval [CI] 1.44 to 2.70, p o 0.0001). In other words, the event rate of reaching baseline FEV1 was 97% higher if Brompton predicted was used compared with raw FEV1 values. This was even greater when All-age predicted was used: 2.29 (95% CI 1.68 to
Figure 1 Kaplan–Meier plot of time to reach baseline for FEV1 after lung transplantation, based on raw FEV1 values (solid line), Brompton percent predicted values (dashed line) and All-age predicted values (dotted line). Censored events represent children who reached transition to adult services without reaching a demonstrated FEV1 baseline.
3.13, p o 0.0001). The same overall pattern was seen for both FVC and FEF25–75 values (refer to Figures S1 and S2, respectively, in the Supplementary Material [SuppMat]
Table 1 Comparison of Baseline Lung Function After Transplantation for FEV1, FVC and FEF25–75 Comparing Raw Values to Those Obtained Using Brompton and All-age Reference Equations Raw FEV1 (liters) Baseline reached 35 of 50 (70%) Baseline value 2.26 (0.69) Time to baseline 662 (119–2,044) FVC (liters) Baseline reached 30 of 50 (60%) Baseline value 2.71 (0.82) Time to baseline 642 (157–2,242) FEF25–75 (liters/s) Baseline reached 47 of 50 (94%) Baseline value 2.84 (1.05) Time to baseline 128 (17–1,345) Modified FEF25–75 (to FEV1 baseline, liters/s) Baseline reached 36 of 50 (72%)d Baseline value 2.59 (1.02) Time to baseline 662 (119–2,044) Modified FEF25–75 (to FVC baseline, liters/s) Baseline reached 32 of 50 (64%)d Baseline value 2.44 (1.01) Time to baseline 642 (157–2,242)
Brompton predicted
All-age predicted
44 of 50 (88%)a 97.9 (20.3) 345 (112–1,211)
46 of 50 (92%)b 86.3 (15.4)c 345 (71–1,352)
39 of 50 (78%) 100.0 (17.3) 452 (141–1,234)
44 of 50 (88%)a 89.1 (14.2)c 431 (157–1,234) 46 of 50 (92%) 99.6 (26.9) 107 (17–1,345) 47 of 50 (94%)a 86.9 (26.7) 345 (71–1,352) 44 of 50 (88%)a 77.1 (27.3) 431 (157–1,234)
Raw values expressed in liters and predicted values as percent predicted. Parametric data presented as mean (SD) and non-parametric data presented as median (range), unless noted otherwise. Time to baseline is shown in days. Brompton reference equations do not include FEF25–75 as an outcome and therefore Brompton predicted FEF25–75 values are not available. a p o 0.05 vs raw. b p o 0.01 vs raw. c p o 0.01 vs Brompton predicted. d p o 0.01 vs unmodified FEF25–75 values.
Robinson et al.
BOS Diagnosis Post-LTx
available online at www.jhltonline.org). The hazard ratios for these are given in the legends of the respective figures. Although the proportions of patients achieving baseline did not differ between the reference equations, the FEV1 baseline values achieved were significantly lower using All-age predicted, 86.3% (SD 15.4%), compared with Brompton predicted, 97.9% (SD 20.3%), with p o 0.0001 (Table 1). The same pattern was seen for FVC values. Time to reach baseline differed between lung function indices. Brompton and all-age values, baseline was reached earliest for FEF25–75, followed by FEV1 and then FVC (see Figures S3 and S4 in SuppMat), whereas, for raw values, both FEV1 and FVC were reached at a similarly late time-point (see Figure S5 in SuppMat). Hazard ratios for these are given in the legends of the respective SuppMat figures. Modification of FEF25–75 baseline (to either FEV1 or FVC baseline timepoints) led to later baseline achievement and a significantly smaller proportion of subjects reaching an identified FEF25–75 baseline for both raw and All-age-predicted values (Table 1 and Figure 2, and Figure S6 in SuppMat). For raw values, the hazard ratio for FEF25–75 baseline modified to FEV1 baseline vs unmodified was 0.36 (95% CI 0.24 to 0.54, p o 0.0001), and FEF25–75 baseline modified to FVC baseline vs unmodified was 0.26 (95% 0.16 to 0.41, p o 0.0001). For All-age predicted values, the corresponding values were similar: 0.46 (95% CI 0.32 to 0.66, p o 0.0001) and 0.34 (95% CI 0.22 to 0.52, p o 0.0001), respectively.
1085
Incidence of BOS The effects of variation in approach to BOS 0p are summarized in Table 2. The incidence of BOS 0p was lowest for raw values, but was increased when expressed using reference equations for both FEV1 and FEF25–75 components of the classification criteria. Based on FEV1 criteria alone, the increase was statistically significant for All-age predicted than for raw values (42% vs 18%, p ¼ 0.016), and trended toward statistical significance for Brompton predicted compared with raw values (36% vs 18%, p ¼ 0.07). The difference in BOS 0p diagnosis between reference equations was not statistically significant. The incidence of BOS 0p based on the FEF25–75 criterion alone was greater compared with FEV1 criterion alone for raw values (38% vs 18%, p ¼ 0.04), but not for All-age predicted values (54% vs 42%, p ¼ 0.32). This was not seen with modified FEF25–75 baseline values. The addition of the FEF25–75 criterion to the FEV1 criterion for BOS 0p diagnosis increased BOS 0p incidence from 18% to 40% for raw values (p ¼ 0.027) and from 42% to 64% for All-age predicted values (p ¼ 0.045). Use of Allage predicted values led to a doubling of BOS 0p incidence (Figure 3) compared with raw values (hazard ratio 2.25, 95% CI 1.50 to 3.36, p o 0.0001). The increase in BOS 0p diagnosis with the incorporation of the FEF25–75 criterion was much smaller if a modified FEF25–75 baseline was used (Table 2). This effect was most pronounced with use of FVC-modified FEF25–75 baseline (Table 2 and Figure 4). For time to BOS 0p with All-age predicted values, the hazard ratio, based on modified FEF25–75 baseline (to FEV1 baseline) vs unmodified FEF25–75 baseline, was 0.61 (95% CI 0.43 to 0.86, p ¼ 0.005), and modified to FVC baseline vs unmodified FEF25–75 baseline was 0.50 (95% CI 0.35 to 0.71, p ¼ 0.0001). Incidence of BOS 1 also increased from that seen with raw values (10 of 50, 20%) when Brompton (14 of 50, 28%) or All-age (12 of 50, 24%) predicted values were used, but did not reach statistical significance. The time to reach BOS 1 was significantly shorter with Brompton predicted vs raw values (hazard ratio 1.64, 95% CI 1.06 to 2.53, p ¼ 0.025). Statistical significance was not reached for All-age reference equation values when compared with raw values (hazard ratio 1.24, 95% CI 0.95 to 1.61, p ¼ 0.11) Figure 5.
Discussion
Figure 2 Kaplan–Meier plot of time to reach baseline for all age reference equation FEF25–75 values after lung transplantation. Estimates are shown for time to baseline achievement for raw FEF25–75 values (solid line), values modified to FEV1 baseline time-point (dashed line) and values modified to FVC baseline timepoint (dotted line). Censored events represent children who reached transition to adult services without reaching a demonstrated baseline.
In this study we have provided an initial detailed description of lung function trends across a pediatric cohort after lung transplantation, and have outlined several major effects when the approach to expressing data is varied. The impact of reference equation on the timing and magnitude of baseline lung function achieved is illustrated, as well as the subsequent effect on BOS incidence. Raw lung function values led to poor estimation of FEV1, FVC and FEF25–75 baseline and an underestimation of the true BOS 0p incidence. Choice of reference equation also affected the magnitude of baseline lung function achieved and subsequent BOS 0p incidence. FEF25–75 baseline definition
1086 Table 2
The Journal of Heart and Lung Transplantation, Vol 34, No 8, August 2015 Comparison of BOS 0p Diagnosis Based on Approach Utilizedb Raw
Using FEV1 criterion alone Incidence 9 of 50 (18%) Time to BOS 0p 1,072 (367–1,604) Using FEF25–75 criterion alone Incidence 19 of 50 (38%) Time to BOS 0p 562 (85–1,604) Using modified FEF25–75 baseline (to FEV1 baseline) 11 of 50 (22%) Time to BOS 0p 733 (367–1,604) Using modified FEF25–75 baseline (to FVC baseline) 10 of 50 (20%) Time to BOS 0p 895 (367–1,604) Using full criteria of either FEV1 or FEF25–75 criterion being met Incidence 20 of 50 (40%) Time to BOS 0p 540 (85–1,604) 12 of 50 (24%) Using modified FEF25–75 baseline (to FEV1 baseline) Time to BOS 0p 726 (367–1,604) Using modified FEF25–75 baseline (to FVC baseline) 11 of 50 (22%) Time to BOS 0p 718 (367–1,604) BOS 0p fulfilling FEF25–75 criterion earlier than FEV1 Incidence 11 of 20 (55%) Using modified FEF25–75 baseline (to FEV1 baseline) 3 of 12 (25%) Using modified FEF25–75 baseline (to FVC baseline) 3 of 11 (27%)
Brompton predicted
All-age predicted
18 of 50 (36%) 760 (367–1,604)
21 of 50 (42%)a 896 (173–1,541) 27 of 50 (54%) 556 (85–2,241) 17 of 50 (34%) 896 (265–2,241) 11 of 50 (22%)b 1072 (367–1,604) 32 of 50 (64%)a 631 (85–1,541) 24 of 50 (48%)a 871 (173–1,541) 21 of 50 (42%)b 896 (173–1,541) 14 of 32 (44%) 4 of 24 (17%)b 0 of 21 (0%)b
The bronchiolitis obliterans syndrome (BOS) 0p classification is shown based on individual FEV1 and FEF25–75 criteria as well as combined criteria. Parametric data presented as mean (SD) and non-parametric data presented as median (range), unless otherwise shown. Time to diagnosis shown in days. a p r 0.05 vs raw. b p r 0.05 vs unmodified FEF25–75 values.
significantly affects BOS 0p incidence. The apparent beneficial effect of using the FEF25–75 criterion was greatly reduced when FEF25–75 baseline definition was modified to FEV1 baseline, and was lost completely with a more
Figure 3 Kaplan–Meier plot of time to reach BOS 0p for raw (solid line) and All-age reference equation data (dotted line), based on either fulfilling FEV1 or FEF25–75 criteria, which was based on unmodified FEF25–75 baseline values. Censored events represent children who reached transition to adult services without reaching a demonstrated BOS 0p diagnosis.
physiologic modification to FVC baseline. These observations clearly show that the approach to lung function monitoring after lung transplantation requires urgent standardization and has significant implications for future BOS detection rates. Baseline lung function achievement after lung transplantation has been poorly described in the literature, for both children and adults. Raw values were found to increase due to ongoing somatic growth, and children may fail to reach a defined FEV1 or FVC baseline, even years after transplantation. This occurred in almost one third of our cohort for FEV1 baseline, despite a median (range) follow-up of 1,115 (500 to 2,613) days in these subjects. Acute decreases in lung function (e.g., acute rejection) occur over short periods, during which minimal somatic growth may occur, and these decreases are detectable irrespective of the different methods used to report lung function. However, lung function loss in chronic rejection typically occurs at a slower rate, and BOS identification requires accurate identification of baseline lung function achieved. Correction for somatic growth by using reference equations improved baseline identification (only 8% failed to reach an FEV1 baseline for All-age data by the end of followup). Baseline lung function may not be achieved for a significant period in children: for All-age predicted values, only 14% and 58% reached a formal FEV1 baseline by 6 months and 1 year post-transplantation, respectively. This suggests long-term growth of the donor organ, beyond simple recruitment during the post-operative period. It remains unclear whether this represents true donor organ growth or
Robinson et al.
BOS Diagnosis Post-LTx
Figure 4 Kaplan–Meier plot showing the effect of choice of FEF25–75 baseline on subsequent time to reach BOS 0p for all age reference equation data. FEF25–75 values (solid line) and those modified to FEV1 baseline time-point (dashed line) and modified to FVC baseline time-point (dotted line) are shown. Censored events represent children who reached transition to adult services without reaching a demonstrated FEF25–75 baseline.
simple graft expansion, as suggested by living related lobar transplant data.11 Reference equation choice is important and it affects BOS incidence. Comparison of a historic local reference equation (i.e., Brompton predicted) and the recently collated improved
Figure 5 Kaplan–Meier plot of time to reach BOS 1 diagnosis after lung transplantation, based on raw FEV1 values (solid line), Brompton predicted values (dashed line) and All-age predicted values (dotted line). Censored events represent children who reached transition to adult services without reaching a demonstrated BOS 1 diagnosis.
1087 All-age reference equations highlight 2 important findings. First, magnitude of baseline FEV1 (and FVC) achieved may be overestimated by historic reference equations. The Brompton equations do not adjust for age, so children who are short for age (a large proportion of children undergoing lung transplantation) have underestimated predicted values and, therefore, overestimated percent-predicted values. This difference was most noticeable in 2 age groups: younger children (r7 years) and adolescents (Figure 6). In the younger children, this probably reflects relatively low subject numbers in this age range of the original Brompton reference data set, weakening the accuracy of predicted values. Discrepancies in adolescence are significant as this the most common age group for pediatric lung transplantation.1 Second, overestimation of FEV1 percentpredicted baseline may lead to a false BOS diagnosis, due to higher starting baseline and improved accuracy (and therefore reduction) in Brompton predicted reference values as the subject ages (see Figure S7 in SuppMat). All-age equations have now been extended using a larger data set to 3 to 95 years by the Global Lungs Initiative.12 Recommendations to use modern collated reference equations for all pediatric lung transplant subjects extends to both the All-age and Global Lungs Initiative equations.7,12 The failure to show a statistical difference in BOS 0p incidence between the reference equation choices may reflect either Type II error or the fact that relatively few subjects were in the younger age range where differences in baseline estimates were most marked (9 of 50, or 18% of those aged o10 years at the time of transplantation). However, the illustrated cases highlight the potential for inappropriate BOS classification. This is clinically relevant as it may trigger management decisions such as trials of macrolide therapy, as recommended in recent BOS guidelines.3 The present data highlight the importance of the FEF25–75 baseline definition on BOS 0p diagnosis. Initial studies supported incorporation, but they used an unmodified baseline for both adults13,14 and children.15 Its value depends on it being an early marker for subsequent FEV1 decline, and the potential for supranormal values was raised in subsequent adult data, leading the authors to question its utility.16 Our pediatric data reinforce these concerns. Individual cases supporting utility do exist (see Figure S8 in SuppMat), but many subjects were affected by this artificial FEF25–75 baseline early after transplantation, leading to inappropriate BOS 0p diagnosis (see Figure S9 in SuppMat). An artificial peak in FEF25–75 most
Figure 6 Difference in baseline FEV1 between Brompton and All-age predicted values, in comparison to age at the time of FEV1 baseline attainment (by All-age values).
1088
The Journal of Heart and Lung Transplantation, Vol 34, No 8, August 2015
likely occurs due to the volume dependence of the measurement. Thoracic adaptation during this recovery period from the transplant operation may lead to an inability to exhale fully to residual volume, causing FEF25–75 to be measured in a higher lung volume range. Therefore, there are physiologic reasons why standardizing FEF25–75 baseline to establishment of FVC baseline, and not FEV1 baseline, is justified. Attainment of FVC and FEV1 baseline occur at differing time-points in most cases: 62% based on All-age values, with a median difference of 25 (range –518 to 666) days. The impact of this early FEF25–75 signal was greatly reduced when modified to FEV1 baseline, and was lost completely when the more physiologically robust modification to FVC baseline was performed. This study was not intended to validate the accuracy of BOS diagnosis, and the weaknesses of transbronchial biopsies in diagnosing BOS are well recognized. A direct comparison of pediatric BOS 0p rates in this study to available adult data is also challenging, due to the format of data presentation.16 However, median time to BOS diagnosis (either BOS 0p or BOS 1) does appear to be longer in pediatric subjects, even if raw values are used. Standardizing the approach to BOS diagnosis should help with these comparisons in the future. In conclusion, we have reviewed descriptive data of pediatric lung function patterns in a large cohort of pediatric lung transplant subjects from a single center and presented several important findings regarding post-transplantation lung function monitoring. The adult-based approach of using of raw values is not appropriate in the pediatric age group, due to ongoing somatic growth. Children may take a significant period of time to achieve a true lung function baseline. Choice of reference equation affects both baseline achieved and subsequent BOS incidence. Modern collated reference equations represent a significant advancement and help to avoid the potential misdiagnosis of BOS, particularly in younger subjects. The data presented herein reinforce the need for an urgent standardization of the FEF25–75 baseline definition and also question the added utility of this criterion in BOS 0p diagnosis.
Disclosure statement The authors have no conflicts of interest to disclose. P.R. was funded during his time at the Great Ormond Street Hospital by a generous donation from David Mackintosh to the Children’s Hospital at Westmead.
Acknowledgements The authors would like to thank Elisabeth Barnes (University of Sydney), who assisted with the statitsical analysis on this manuscript.
Supplementary material Supplementary data associated with this article can be found in the online version at www.jhltonline.org.
References 1. Benden C, Goldfarb SB, Edwards LB, et al. The registry of the International Society for Heart and Lung Transplantation: seventeenth official pediatric lung and heart–lung transplantation report—2014; focus theme: retransplantation. J Heart Lung Transplant 2014;33: 1025-33. 2. Cooper JD, Billingham M, Egan T, et al. A working formulation for the standardization of nomenclature and for clinical staging of chronic dysfunction in lung allografts. J Heart Lung Transplant 1993;12:713-6. 3. Meyer KC, Raghu G, Verleden GM, et al. An international ISHLT/ATS/ERS clinical practice guideline: diagnosis and management of bronchiolitis obliterans syndrome. Eur Respir J 2014;44: 1479-503. 4. Estenne M, Maurer JR, Boehler A, et al. Bronchiolitis obliterans syndrome 2001: an update of the diagnostic criteria. J Heart Lung Transplant 2002;21:297-310. 5. Rosen JB, Smith EO, Schecter MG, et al. Decline in 25% to 75% forced expiratory flow as an early predictor of chronic airway rejection in pediatric lung transplant recipients. J Heart Lung Transplant 2012;31:1288-92. 6. Kirkby J, Aurora P, Spencer H, et al. Stitching and switching: the impact of discontinuous lung function reference equations. Eur Respir J 2012;39:1256-7. 7. Stanojevic S, Wade A, Stocks J, et al. Reference ranges for spirometry across all ages: a new approach. Am J Respir Crit Care Med 2008;177:253-60. 8. American Thoracic Society. Standardization of spirometry, 1994 update. Am J Respir Crit Care Med 1995;152:1107-36. 9. Beydon N, Davis SD, Lombardi E, et al. An official American Thoracic Society/European Respiratory Society statement: pulmonary function testing in preschool children. Am J Respir Crit Care Med 2007;175:1304-45. 10. Rosenthal M, Bain SH, Cramer D, et al. Lung function in white children aged 4 to 19 years: I—Spirometry. Thorax 1993;48:794-802. 11. Sritippayawan S, Keens TG, Horn MV, et al. Does lung growth occur when mature lobes are transplanted into children? Pediatr Transplant 2002;6:500-4. 12. Quanjer PH, Stanojevic S, Cole TJ, 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;40:1324-43. 13. Patterson GM, Wilson S, Whang JL, et al. Physiologic definitions of obliterative bronchiolitis in heart–lung and double lung transplantation: a comparison of the forced expiratory flow between 25% and 75% of the forced vital capacity and forced expiratory volume in one second. J Heart Lung Transplant 1996;15:175-81. 14. Valentine VG, Robbins RC, Berry GJ, et al. Actuarial survival of heart–lung and bilateral sequential lung transplant recipients with obliterative bronchiolitis. J Heart Lung Transplant 1996;15:371-83. 15. Sritippayawan S, Keens TG, Horn MV, et al. What are the best pulmonary function test parameters for early detection of post-lung transplant bronchiolitis obliterans syndrome in children? Pediatr Transplant 2003;7:200-3. 16. Hachem RR, Chakinala MM, Yusen RD, et al. The predictive value of bronchiolitis obliterans syndrome stage 0-p. Am J Respir Crit Care Med 2004;169:468-72.
