Determinants of Progression and Mortality in Lymphangioleiomyomatosis

BackgroundLymphangioleiomyomatosis is a progressive diffuse cystic lung disease with approximately 85% survival at 10 years. The determinants of disease progression and mortality after the introduction of sirolimus therapy and vascular endothelial growth factor D (VEGF-D) as a biomarker have not been well defined.Research QuestionWhich factors, including VEGF-D and sirolimus therapy, influence disease progression and survival prognosis in patients with lymphangioleiomyomatosis?Study Design and MethodsThe progression dataset and the survival dataset included 282 and 574 patients, respectively, from Peking Union Medical College Hospital, Beijing, China. A mixed-effects model was used to compute the rate of decline in FEV1, and generalized linear models were used to identify variables affecting FEV1 decline. A Cox proportional hazards model was used to explore the association between clinical variables and the outcomes of death or lung transplantation in patients with lymphangioleiomyomatosis.ResultsVEGF-D levels and sirolimus treatment were associated with FEV1 changes and survival prognosis. Compared with patients with VEGF-D of < 800 pg/mL at baseline, patients with VEGF-D of ≥ 800 pg/mL lost FEV1 faster (SE, –38.86 mL/y; 95% CI, –73.90 to –3.82 mL/y; P = .031). The 8-year cumulative survival rates of patients with VEGF-D of ≥ 2,000 pg/mL and < 2,000 pg/mL were 82.9% and 95.1%, respectively (P = .014). The generalized linear regression model also demonstrated the benefit of delaying the decline of FEV1 by 65.56 mL/y (95% CI, 29.06-102.06 mL/y) in patients treated with sirolimus compared with those without sirolimus (P < .001). The 8-year risk of death was reduced by 85.1% (hazard ratio, 0.149; 95% CI, 0.075-0.299) after sirolimus treatment. After inverse treatment probability weighting, the risks of death in the sirolimus group were reduced by 85.6%. CT scan results of grade III severity were associated with worse progression than results of grades I or II severity. Patients with baseline FEV1 of 70% predicted or St. George’s Respiratory Questionnaire Symptoms domain 50 or higher predicted a higher risk of worse survival.InterpretationSerum VEGF-D levels, a biomarker of lymphangioleiomyomatosis, are associated with disease progression and survival. Sirolimus therapy is associated with slower disease progression and better survival in patients with lymphangioleiomyomatosis.Trial RegistryClinicalTrials.gov; No.: NCT03193892; URL: www.clinicaltrials.gov Lymphangioleiomyomatosis is a progressive diffuse cystic lung disease with approximately 85% survival at 10 years. The determinants of disease progression and mortality after the introduction of sirolimus therapy and vascular endothelial growth factor D (VEGF-D) as a biomarker have not been well defined. Which factors, including VEGF-D and sirolimus therapy, influence disease progression and survival prognosis in patients with lymphangioleiomyomatosis? The progression dataset and the survival dataset included 282 and 574 patients, respectively, from Peking Union Medical College Hospital, Beijing, China. A mixed-effects model was used to compute the rate of decline in FEV1, and generalized linear models were used to identify variables affecting FEV1 decline. A Cox proportional hazards model was used to explore the association between clinical variables and the outcomes of death or lung transplantation in patients with lymphangioleiomyomatosis. VEGF-D levels and sirolimus treatment were associated with FEV1 changes and survival prognosis. Compared with patients with VEGF-D of < 800 pg/mL at baseline, patients with VEGF-D of ≥ 800 pg/mL lost FEV1 faster (SE, –38.86 mL/y; 95% CI, –73.90 to –3.82 mL/y; P = .031). The 8-year cumulative survival rates of patients with VEGF-D of ≥ 2,000 pg/mL and < 2,000 pg/mL were 82.9% and 95.1%, respectively (P = .014). The generalized linear regression model also demonstrated the benefit of delaying the decline of FEV1 by 65.56 mL/y (95% CI, 29.06-102.06 mL/y) in patients treated with sirolimus compared with those without sirolimus (P < .001). The 8-year risk of death was reduced by 85.1% (hazard ratio, 0.149; 95% CI, 0.075-0.299) after sirolimus treatment. After inverse treatment probability weighting, the risks of death in the sirolimus group were reduced by 85.6%. CT scan results of grade III severity were associated with worse progression than results of grades I or II severity. Patients with baseline FEV1 of 70% predicted or St. George’s Respiratory Questionnaire Symptoms domain 50 or higher predicted a higher risk of worse survival. Serum VEGF-D levels, a biomarker of lymphangioleiomyomatosis, are associated with disease progression and survival. Sirolimus therapy is associated with slower disease progression and better survival in patients with lymphangioleiomyomatosis. ClinicalTrials.gov; No.: NCT03193892; URL: www.clinicaltrials.gov Take-home PointsStudy Question: Which factors, including vascular endothelial growth factor D (VEGF-D) level and sirolimus therapy, influence disease progression and survival prognosis in patients with lymphangioleiomyomatosis?Results: Based on the progression dataset (n = 282) and survival dataset (n = 574) from a cohort with lymphangioleiomyomatosis, patients with baseline VEGF-D levels of < 800 pg/mL lost FEV1 faster than those with VEGF-D levels of ≥ 800 pg/mL, whereas the beneficial effects of delaying FEV1 reduction were observed in patients treated with sirolimus. The risk of death was increased in patients with VEGF-D levels of ≥ 2,000 pg/mL and was reduced after sirolimus therapy.Interpretation: Serum VEGF-D levels may predict disease progression and survival. Sirolimus therapy likely deters disease progression and prolongs the survival of patients with lymphangioleiomyomatosis. Study Question: Which factors, including vascular endothelial growth factor D (VEGF-D) level and sirolimus therapy, influence disease progression and survival prognosis in patients with lymphangioleiomyomatosis? Results: Based on the progression dataset (n = 282) and survival dataset (n = 574) from a cohort with lymphangioleiomyomatosis, patients with baseline VEGF-D levels of < 800 pg/mL lost FEV1 faster than those with VEGF-D levels of ≥ 800 pg/mL, whereas the beneficial effects of delaying FEV1 reduction were observed in patients treated with sirolimus. The risk of death was increased in patients with VEGF-D levels of ≥ 2,000 pg/mL and was reduced after sirolimus therapy. Interpretation: Serum VEGF-D levels may predict disease progression and survival. Sirolimus therapy likely deters disease progression and prolongs the survival of patients with lymphangioleiomyomatosis. Lymphangioleiomyomatosis is a chronic progressive cystic lung disease that predominantly occurs in women.1Xu K.F. Xu W. Liu S. et al.Lymphangioleiomyomatosis.Semin Respir Crit Care Med. 2020; 41: 256-268Crossref PubMed Scopus (16) Google Scholar Lymphangioleiomyomatosis presents as a sporadic form or a type associated with tuberous sclerosis complex (TSC).1Xu K.F. Xu W. Liu S. et al.Lymphangioleiomyomatosis.Semin Respir Crit Care Med. 2020; 41: 256-268Crossref PubMed Scopus (16) Google Scholar Pulmonary functions decline gradually, with an annual loss of FEV1 of 47 to 134 mL.2McCormack F.X. Inoue Y. Moss J. et al.Efficacy and safety of sirolimus in lymphangioleiomyomatosis.N Engl J Med. 2011; 364: 1595-1606Crossref PubMed Scopus (818) Google Scholar, 3Hayashida M. Yasuo M. Hanaoka M. et al.Reductions in pulmonary function detected in patients with lymphangioleiomyomatosis: an analysis of the Japanese National Research Project on Intractable Diseases database.Respir Invest. 2016; 54: 193-200Crossref PubMed Scopus (18) Google Scholar, 4Johnson S.R. Tattersfield A.E. Decline in lung function in lymphangioleiomyomatosis: relation to menopause and progesterone treatment.Am J Respir Crit Care Med. 1999; 160: 628-633Crossref PubMed Scopus (178) Google Scholar, 5Taveira-DaSilva A.M. Stylianou M.P. Hedin C.J. Hathaway O. Moss J. Decline in lung function in patients with lymphangioleiomyomatosis treated with or without progesterone.Chest. 2004; 126: 1867-1874Abstract Full Text Full Text PDF PubMed Scopus (203) Google Scholar, 6Gupta N. Lee H.S. Young L.R. et al.Analysis of the MILES cohort reveals determinants of disease progression and treatment response in lymphangioleiomyomatosis.Eur Respir J. 2019; 53: 1802066Crossref Scopus (32) Google Scholar It is essential to identify the factors that influence disease progression and survival. A 15-year prospective longitudinal study of the National Heart, Lung and Blood Institute (NHLBI) registry reported that the 10-year and 20-year survival rates were 85% and 64%, respectively.7Gupta N. Lee H.S. Ryu J.H. et al.The NHLBI LAM Registry: prognostic physiologic and radiologic biomarkers emerge from a 15-year prospective longitudinal analysis.Chest. 2019; 155: 288-296Abstract Full Text Full Text PDF PubMed Scopus (51) Google Scholar The factors determining survival included menopause, baseline FEV1, and diffusion capacity of the lungs for carbon monoxide (Dlco).7Gupta N. Lee H.S. Ryu J.H. et al.The NHLBI LAM Registry: prognostic physiologic and radiologic biomarkers emerge from a 15-year prospective longitudinal analysis.Chest. 2019; 155: 288-296Abstract Full Text Full Text PDF PubMed Scopus (51) Google Scholar Serum vascular endothelial growth factor D (VEGF-D) was found to be associated with disease progression in the Multicenter International Lymphangioleiomyomatosis Efficacy of Sirolimus (MILES) study,6Gupta N. Lee H.S. Young L.R. et al.Analysis of the MILES cohort reveals determinants of disease progression and treatment response in lymphangioleiomyomatosis.Eur Respir J. 2019; 53: 1802066Crossref Scopus (32) Google Scholar,8Young L. Lee H.S. Inoue Y. et al.Serum VEGF-D a concentration as a biomarker of lymphangioleiomyomatosis severity and treatment response: a prospective analysis of the Multicenter International Lymphangioleiomyomatosis Efficacy of Sirolimus (MILES) trial.Lancet Respir Med. 2013; 1: 445-452Abstract Full Text Full Text PDF PubMed Scopus (149) Google Scholar but not in the NHLBI cohort.7Gupta N. Lee H.S. Ryu J.H. et al.The NHLBI LAM Registry: prognostic physiologic and radiologic biomarkers emerge from a 15-year prospective longitudinal analysis.Chest. 2019; 155: 288-296Abstract Full Text Full Text PDF PubMed Scopus (51) Google Scholar A study from the LAM Foundation suggested that the estimated 10-year survival is 86% and the median survival times from the onset of symptoms and the time of diagnosis are 29 and 23 years, respectively.9Oprescu N. McCormack F.X. Byrnes S. Kinder B.W. Clinical predictors of mortality and cause of death in lymphangioleiomyomatosis: a population-based registry.Lung. 2013; 191: 35-42Crossref PubMed Scopus (63) Google Scholar Those studies on progression and survival mainly were based on data from patients before the advent of the mechanistic target of rapamycin inhibitor sirolimus, the first effective drug for lymphangioleiomyomatosis.2McCormack F.X. Inoue Y. Moss J. et al.Efficacy and safety of sirolimus in lymphangioleiomyomatosis.N Engl J Med. 2011; 364: 1595-1606Crossref PubMed Scopus (818) Google Scholar Whether lymphangioleiomyomatosis survival has improved after sirolimus administration is a critical question that should be addressed. Sirolimus has been reported as an effective treatment for maintaining lung function; improving oxygen levels, 6-min walking distance (6MWD), and St. George’s Respiratory Questionnaire (SGRQ) scores; and reducing the serum levels of VEGF-D.2McCormack F.X. Inoue Y. Moss J. et al.Efficacy and safety of sirolimus in lymphangioleiomyomatosis.N Engl J Med. 2011; 364: 1595-1606Crossref PubMed Scopus (818) Google Scholar,10Xu K.F. Tian X. Yang Y. Zhang H. Rapamycin for lymphangioleiomyomatosis: optimal timing and optimal dosage.Thorax. 2018; 73: 308-310Crossref PubMed Scopus (16) Google Scholar A study compared mortality before and after the introduction of sirolimus in a cohort and suggested that the mortality of patients with lymphangioleiomyomatosis likely has decreased in recent years,11Harari S. Torre O. Elia D. et al.Improving Survival in Lymphangioleiomyomatosis: a 16-year observational study in a large cohort of patients.Respiration. 2021; 100: 989-999Crossref PubMed Scopus (4) Google Scholar although it was not found that sirolimus was associated with disease progression or survival in a more recent study based on a cohort observed over 7 years.12Yoon H.Y. Kim H.J. Song J.W. Long-term clinical course and outcomes in patients with lymphangioleiomyomatosis.Respir Res. 2022; 23: 158Crossref PubMed Scopus (2) Google Scholar The Peking Union Medical College Hospital (PUMCH) lymphangioleiomyomatosis cohort established a large database of patients with lymphangioleiomyomatosis in China with more than 600 patients in a regular follow-up plan, including many patients taking sirolimus. In this study, we aimed to investigate the clinical, radiologic, laboratory, and functional biomarkers that could serve as determinants of disease progression and survival based on this cohort. We focused on investigating whether sirolimus could prolong the survival of patients with lymphangioleiomyomatosis and the role of VEGF-D in predicting survival. The cohort was retrieved from the LAM Registry Study at PUMCH between June 2011 and December 2019. The diagnoses of all recorded cases were re-evaluated according to the diagnosis criteria of definite lymphangioleiomyomatosis from the American Thoracic Society/Japanese Respiratory Society guidelines.13Gupta N. Finlay G.A. Kotloff R.M. et al.Lymphangioleiomyomatosis diagnosis and management: high-resolution chest computed tomography, transbronchial lung biopsy, and pleural disease management. An official American Thoracic Society/Japanese Respiratory Society clinical practice guideline.Am J Respir Crit Care Med. 2017; 196: 1337-1348Crossref PubMed Scopus (128) Google Scholar The patients were followed up regularly in our hospital according to the LAM Registry Study protocol. They were evaluated on a yearly basis as well as during routine or emergency visits. All patients signed informed consent forms. The study protocols were approved by the institutional review board of PUMCH (Identifiers: S-379 and JS-1323). The LAM Registry Study protocol was a single-center protocol before 2017 that has been upgraded to a national lymphangioleiomyomatosis registry since 2017. The National Lymphangioleiomyomatosis Registry Study was registered on ClinicalTrials.gov.14National Institutes of Health Clinical CenterA national registry on Chinese patients with lymphangioleiomyomatosis. NCT03193892. ClinicalTrials.gov. Peking Union Medical College Hospital.2017https://clinicaltrials.gov/ct2/show/NCT03193892Google Scholar The materials included in this report were from PUMCH only. The research was carried out in accordance with the tenets of the Declaration of Helsinki. The data were retrieved from our registry database based on the following criteria. Two datasets were retrieved, namely, the progression dataset and the survival dataset, to investigate potential clinical, radiologic, functional, and laboratory biomarkers for disease progression and survival (Fig 1). For the progression dataset, the inclusion criteria were as follows: (1) a definitive diagnosis of lymphangioleiomyomatosis based on the American Thoracic Society/Japanese Respiratory Society criteria13Gupta N. Finlay G.A. Kotloff R.M. et al.Lymphangioleiomyomatosis diagnosis and management: high-resolution chest computed tomography, transbronchial lung biopsy, and pleural disease management. An official American Thoracic Society/Japanese Respiratory Society clinical practice guideline.Am J Respir Crit Care Med. 2017; 196: 1337-1348Crossref PubMed Scopus (128) Google Scholar and (2) at least two lung function tests separated by 3 months or longer. The exclusion criteria were current chylothorax or pneumothorax on baseline evaluation and incomplete baseline data and evaluation, including clinical evaluations, chest and abdominal radiologic evaluations, and spirometry tests. For the survival dataset, the inclusion criteria were as follows: (1) a definitive diagnosis of lymphangioleiomyomatosis based on the American Thoracic Society/Japanese Respiratory Society criteria13Gupta N. Finlay G.A. Kotloff R.M. et al.Lymphangioleiomyomatosis diagnosis and management: high-resolution chest computed tomography, transbronchial lung biopsy, and pleural disease management. An official American Thoracic Society/Japanese Respiratory Society clinical practice guideline.Am J Respir Crit Care Med. 2017; 196: 1337-1348Crossref PubMed Scopus (128) Google Scholar and (2) a baseline evaluation, including a clinical evaluation and chest and abdominal radiologic evaluations. Patients with an incomplete baseline evaluation were excluded. Lung function, chylothorax, or pneumothorax were not part of the exclusion criteria for the survival dataset. We collected the following information: (1) age, occurrence of menopause, symptoms of dyspnea, cough, history of pneumothorax or chylothorax, kidney angiomyolipomas, and TSC; (2) serum VEGF-D level at baseline15Xu K.F. Zhang P. Tian X. et al.The role of vascular endothelial growth factor-D in diagnosis of lymphangioleiomyomatosis (LAM).Respir Med. 2013; 107: 263-268Abstract Full Text Full Text PDF PubMed Scopus (47) Google Scholar; (3) arterial blood gas analysis on room air; (4) severity grade (I, II, and III) on chest CT scan (grade I, cysts involving less than one-third of the lungs; grade II, cysts involving one-third to two-thirds of the lungs; and grade III, cysts involving more than two-thirds of the lungs)16Avila N.A. Dwyer A.J. Rabel A. Moss J. Sporadic lymphangioleiomyomatosis and tuberous sclerosis complex with lymphangioleiomyomatosis: comparison of CT features.Radiology. 2007; 242: 277-285Crossref PubMed Scopus (103) Google Scholar; (5) lung function, including FEV1, FEV1 % predicted, FVC, FVC % predicted, FEV1 to FVC ratio, Dlco, and Dlco % predicted17Wanger J. Clausen J.L. Coates A. et al.Standardisation of the measurement of lung volumes.Eur Respir J. 2005; 26: 511-522Crossref PubMed Scopus (2046) Google Scholar,18Graham B.L. Brusasco V. Burgos F. et al.2017 ERS/ATS standards for single-breath carbon monoxide uptake in the lung.Eur Respir J. 2017; 49: 1600016Crossref PubMed Scopus (409) Google Scholar; (6) 6MWD19ATS Committe on Proficiency Standards for Clinical Pulmonary Function Laboratories. ATS statement: guidelines for the six-minute walk test.Am J Respir Crit Care Med. 2002; 166: 111-117Crossref PubMed Scopus (8440) Google Scholar; (7) sirolimus treatment; and (8) SGRQ score.20Jones P.W. Quirk F.H. Baveystock C.M. The St George’s Respiratory Questionnaire.Respir Med. 1991; 85 (discussion 33-27): 25-31Abstract Full Text PDF PubMed Scopus (1493) Google Scholar The measurement methods were same as those published previously.21Hu S. Wu X. Xu W. et al.Long-term efficacy and safety of sirolimus therapy in patients with lymphangioleiomyomatosis.Orphanet J Rare Dis. 2019; 14: 206Crossref PubMed Scopus (19) Google Scholar,22Zhan Y. Shen L. Xu W. et al.Functional improvements in patients with lymphangioleiomyomatosis after sirolimus: an observational study.Orphanet J Rare Dis. 2018; 13: 34Crossref PubMed Scopus (15) Google Scholar The FEV1 value used in this study was the value obtained before the bronchodilator test, and the changes in FEV1 were expressed as milliliters per year. The follow-up started from the diagnosis of lymphangioleiomyomatosis and ended at death; receiving lung transplantation; being lost to follow-up; or December 31, 2019, whichever came first. The patients were followed up annually according to the LAM Registry Study protocol. They were also followed up via regular or emergency clinic visits, telephone calls, text messages, e-mail, and the patient follow-up platform Xinshulin MedClip app (Xinshulin Information Technology [Beijing] Limited Company). Patients receiving sirolimus continued treatment unless severe infections, surgery, pregnancy, or severe adverse events occurred. Adverse effects were mild or moderate and none of the patients stopped sirolimus permanently because of adverse effects. Continuous variables are reported as the mean ± SD or the median (interquartile range [IQR]), and categorical variables are described as the proportional percentage, No. (%). We used Student t test or the Wilcoxon rank-sum test for continuous variables and the χ2 test or Fisher exact probability test for categorical variables. Using the progression cohort, we used a mixed-effects model with a random intercept and random slope to estimate the decline in FEV1 over time and used the PRISM version 8.0 software program (GraphPad) to visualize the results. In the mixed model, we used baseline FEV1 as the fixed effect and the follow-up time and age at baseline as the random effects. Furthermore, we used a multivariate generalized linear regression model to explore the factors related to disease progression, with the annual average change in FEV1 as the dependent variable. For survival, Kaplan-Meier curves were drawn to describe survival and the log-rank test was used for comparisons between groups. Factors related to death were identified by a multivariate Cox proportional hazards model with stepwise selection of candidate variables. The significance level for entering an explanatory variable into the model was 0.10 and that for removing an explanatory variable from the model was 0.15. Cutoff values of continuous variables were selected according to the restrictive cubic spline function and clinical significance. The variables included in the Cox proportional hazards model all met the proportional hazards assumption. We imputed missing data by multiple imputation methods based on Monte Carlo simulation, and no significant difference was found in the data distribution before and after imputation (e-Tables 1, 2). To explain further the impact of sirolimus treatment on patient survival, we conducted inverse probability of treatment weighting analyses.23Austin P.C. Stuart E.A. Moving towards best practice when using inverse probability of treatment weighting (IPTW) using the propensity score to estimate causal treatment effects in observational studies.Stat Med. 2015; 34: 3661-3679Crossref PubMed Scopus (2237) Google Scholar The weight of inverse probability weighting was based on the propensity score, and the independent variables for fitting the propensity score are shown in e-Table 3. We weighted each patient being treated with sirolimus by the inverse of the probability of sirolimus treatment and weighted each patient not being treated with sirolimus by the inverse probability of no sirolimus treatment. The standardized mean difference (SMD) of other characteristics of the two groups was improved significantly after inverse probability weighting (e-Fig 1). Two-sided tests were performed, and a P value of < .05 was considered to be statistically significant. All data were analyzed using SAS version 9.4 software (SAS Institute, Inc.) and R version 3.5.3 software (R Foundation for Statistical Computing). As described in Table 1, 282 patients in the progression dataset were divided into a no sirolimus treatment group (n = 167) and a sirolimus group (n = 115). All participants were female, with an average age of 40 ± 9 years; 16.7% of them were postmenopausal. Most of the participants (91.5%) had sporadic lymphangioleiomyomatosis. Many of the participants (65.2%) were classified as having grade III severity based on chest CT scan results. At baseline, the mean FEV1 and Dlco values were 76.2% predicted (IQR, 54.5% predicted-92.7% predicted) and 55.3% predicted (IQR, 36.1% predicted-74.5% predicted), respectively. Many differences were found between the sirolimus and no sirolimus groups. Compared with patients not treated with sirolimus, patients treated with sirolimus sought treatment at a younger age and showed worse lung function (FEV1 and Dlco) and higher VEGF-D levels (Table 1). The median follow-up time of the patients was 18 months (IQR, 12-30 months) (e-Fig 2), and 24 patients (8.5%) were followed up for < 1 year (12 ± 3 months). The median number of pulmonary function tests included in our study was around two (IQR, 2-3).Table 1Baseline Characteristics of Patients in the Progression DatasetVariableAll (n = 282)No Sirolimus Treatment (n = 167)Sirolimus Treatment (n = 115)P ValueAge, y40 ± 941 ± 938 ± 9.020aStudent t test.Follow-up time, mo18 (12-30)20 (12-36)18 (12-28).598bWilcoxon rank-sum test.Duration of sirolimus treatment, yNANA1.5 (1-2)NAMenopause reached47 (16.7)36 (21.6)11 (9.6).008cχ 2 test.History of pneumothorax74 (26.2)39 (23.3)35 (30.4).184cχ 2 test.History of chylous ascites8 (2.8)4 (2.4)4 (3.5).719dFisher exact test.Kidney angiomyolipomas88 (31.2)46 (27.5)42 (36.5).110cχ 2 test.Retroperitoneal lymphangioleiomyomas58 (20.6)28 (16.8)30 (26.1).057cχ 2 test.TSC24 (8.51)8 (4.8)16 (13.9).007cχ 2 test.VEGF-D, pg/mL1,649.50 (924.00-3,142.40)1,369.20 (712.60-2,892.00)2,199.00 (1,025.60-3,691.20)< .001bWilcoxon rank-sum test.CT scan severity grade.020cχ 2 test. I45 (16.0)31 (18.6)14 (12.2) II53 (18.8)38 (22.7)15 (13.0) III184 (65.2)98 (58.7)86 (74.8)FEV1 mL2,110 (1,460-2,510)2,220 (1,620-2,610)1,920 (1,370-2,380).003bWilcoxon rank-sum test. % Predicted76.2 (54.5-92.7)82.2 (61.6-96.9)68.6 (50.1-86.8)< .001bWilcoxon rank-sum test.FVC mL2,985 (2,550-3,320)3,020 (2,550-3,330)2,940 (2,560-3,320).550bWilcoxon rank-sum test. % Predicted93.9 (80.6-106.6)94.3 (82.4-108.0)93.7 (78.1-102.7).099bWilcoxon rank-sum test.Dlco mM/kpa/min4.5 (3.0-6.0)4.8 (3.4-6.4)3.8 (2.6-5.6).001bWilcoxon rank-sum test. % Predicted55.3 (36.1-74.5)62.3 (42.6-77.2)45.7 (31.4-64.9)< .001bWilcoxon rank-sum test.Data are presented as No. (%), mean ± SD, or median (interquartile range), unless otherwise indicated. Dlco = diffusion capacity of the lungs for carbon monoxide; NA = not available; TSC = tuberous sclerosis complex; VEGF-D = vascular endothelial growth factor D.a Student t test.b Wilcoxon rank-sum test.c χ 2 test.d Fisher exact test. Open table in a new tab Data are presented as No. (%), mean ± SD, or median (interquartile range), unless otherwise indicated. Dlco = diffusion capacity of the lungs for carbon monoxide; NA = not available; TSC = tuberous sclerosis complex; VEGF-D = vascular endothelial growth factor D. The overall FEV1 change over time in the PUMCH cohort was –63.24 mL/y (IQR, –131.81 to –7.04 mL/y), and the average annual change rate was –3.1% (IQR, –7.6% to –0.3%). The annual FEV1 decline in the sirolimus group was significantly slower than that in the no sirolimus group (–27.63 mL [IQR, –79.66 to 37.24 mL] vs –91.37 mL [IQR, –172.01 to –28.37 mL]; P < .001), and the average annual change rate in patients not being treated with sirolimus was –4.7% (IQR, –10.1% to –1.3%) vs –1.5% (IQR, –4.2% to 1.9%) in the sirolimus group (Fig 2). We constructed a generalized linear model with the difference between the 2-year follow-up FEV1 % predicted by the mixed-effects model and the baseline FEV1 as the dependent variable. The change in FEV1 in patients treated with sirolimus increased by 65.56 mL/y (95% CI, 29.06-102.06 mL/y) compared with patients not treated with sirolimus (P < .001). Compared to those in patients with CT scan grade I and II severity at baseline, the FEV1 changes in patients with CT scan grade III severity decreased by an average of 67.11 mL (95% CI, –99.92 to –34.30 mL; P < .001). Compared with patients with VEGF-D of < 800 pg/mL at baseline, patients with VEGF-D of ≥ 800 pg/mL at baseline showed significantly decreased FEV1 changes by an average of 38.86 mL (95% CI, –73.90 to –3.82 mL; P = .031). We did not find significant associations between menopause, renal angiomyolipomas, pneumothorax, or baseline FEV1 and FEV1 changes. In addition, we found that FEV1 predicted at baseline modified the association between sirolimus treatment and FEV1 changes (P = .003 for interaction) (Table 2).Table 2Generalized Linear Regression Model of Annual Average Change of FEV1VariableEstimation (SE)95% CIP ValueSirolimus (with vs without)65.5629.06-102.06< .001CT scan severity grade (III vs I/II)–67.11–99.92 to –34.30< .001VEGF-D (≥ 800 pg/mL vs < 800 pg/mL)–38.86–73.90 to –3.82.031Menopause reached (yes vs no)27.58–8.75 to 63.90.138Renal AML (yes vs no)25.24–3.20 to 53.68.083Pneumothorax (yes vs no)–12.04–42.16 to 18.09.434FEV1 (< 70% predicted vs ≥ 70% predicted)–3.14–42.33 to 36.04.875Interaction (sirolimus × FEV1 % predicted)83.4229.13-137.72.003AML = angiomyolipoma; VEGF-D = vascular endothelial growth factor D. Open table in a new tab AML = angiomyolipoma; VEGF-D = vascular endothelial growth factor D. A total of 574 patients were included in the survival analysis. During the follow-up, 38 patients died, three received lung transplantation, and eight were lost to follow-up. The baseline characteristics of the patients in the survival dataset are shown in Table 3. The mean ages at the onset of symptoms and at the initial evaluation in our hospital were 34 ± 11 years and 38 ± 10 years, respectively. Like the progression dataset, compared with patients not treated with sirolimus, patients treated with sirolimus sought treatment at a younger age, had worse lung function (FEV1 and Dlco), and higher VEGF-D levels. The sirolimus group also included more patients with a history of pneumothorax, chylothorax, and TSC; more severe CT scan severity grades; lower oxygen levels; worse SGRQ scores; shorter 6MWD; and higher Borg dyspnea scores.Table 3Baseline Characteristics of Patients in the Survival DatasetVariableAll (n = 574)No Sirolimus Treatment (n =