Lymphangioleiomyomatosis Mortality in Patients with Tuberous Sclerosis Complex

To the Editor:

Lymphangioleiomyomatosis (LAM) is a rare diffuse cystic lung disease that almost exclusively affects women (1). LAM can occur sporadically, or be associated with tuberous sclerosis complex (TSC), a heritable tumor suppressor syndrome associated with hamartomas in multiple organs, seizures, and cognitive impairment (2–4). Cystic changes consistent with LAM have been reported to occur in 26–49% of women with TSC (2, 5–7), and most female patients with TSC have at least a few pulmonary cysts by age 40 years (5). Current Tuberous Sclerosis Alliance guidelines recommendations include screening women with TSC for LAM with high-resolution computed tomography starting at the age of 18 years (8). The National Heart, Lung, and Blood Institute (NHLBI) LAM Registry Committee reported that the 5-, 10-, 15-, and 20-year transplant-free survival in LAM is 95%, 85%, 75%, and 64%, respectively (9). TSC-LAM is believed to be milder and less progressive than sporadic LAM, but the natural history of TSC-LAM and its impact on mortality is less well understood (9). The purposes of this study were to better understand the long-term prognosis of LAM in patients with TSC and to establish the leading causes of death in this population.

Methods

The study was approved by the Cincinnati Children’s Hospital Medical Center (CCHMC) Institutional Review Board (IRB; IRB No. 2016-3430). The study included all patients who had attended the TSC clinic at CCHMC in the period January 1, 1998 to November 30, 2016. Diagnoses were classified according to the International Classification of Diseases-9 and -10 diagnostic codes. We reviewed the electronic medical record to determine the patients’ vital status and prevalence of TSC-associated disease complications, including LAM, renal angiomyolipomas (AMLs), and seizures. The diagnosis of LAM was based on the presence of characteristic cystic change on chest computed tomography, as determined by an expert thoracic radiologist and/or a pulmonologist familiar with LAM (10, 11). Vital status data were also obtained by querying the National Death Index. In cases where vital status was still unclear after electronic medical record review and National Death Index query, the patient or next of kin was contacted by phone and interviewed using an IRB-approved script regarding the date and cause of death.

Chi-square and Fisher’s exact test were used to study the association of LAM with categorical variables, and Wilcoxon and Mann-Whitney tests were used to investigate the association of LAM with continuous variables. Age at diagnosis of LAM for individuals who were missing diagnosis date (18/66 [27%]) was imputed using the midpoint between age 18 years and age at censoring. Kaplan-Meier curves were used to examine survival. Restricted mean survival time, calculated by numerically integrating the Kaplan-Meier curve (12), was used to estimate average life expectancy of the TSC cohort (from birth) and according to LAM status in adult women from the cohort (from age 18 yr) (13). Life expectancy difference was taken as the difference between restricted mean survival times (14).

Results

A total of 702 unique patients with confirmed diagnosis of TSC based on consensus clinical and/or genetic criteria (15) were seen in the CCHMC TSC clinic between January 1, 1998 and November 30, 2016. Of these, vital status was unable to be determined for 135 patients, and was available on 567 patients. Of these, 541 were documented to be alive and 26 were found to be deceased at the censor date of this study. There were 69 patients with known LAM (Table 1): 66 women and 3 men. There was a total of 247 women age 18 years or older in our cohort. Of these 247 women, 176 were determined to be alive, 13 were deceased, and 58 had an unknown vital status (Table 2).

Table 1.

Demographic and clinical characteristics of the overall cohort

Variable	LAM (n = 69)	No LAM (n = 593)
Age, range, yr*	38 (21–75)	19 (5–84)
Sex, n (%)
Female	66 (96)	298 (50)
Male	3 (4)	295 (50)
Seizure disorder, n (%)	41 (59)	436 (74)
Renal AMLs, n (%)	63 (91)	388 (65)
Outcome, n (%)†
Alive	48 (69)	483 (81)
Deceased	8 (12)	15 (3)
Unknown status	13 (19)	95 (16)
Cause of death, n (%)
LAM	6/8 (75)	—
ARDS	1/8 (12.5)	—
Encephalopathy	1/8 (12.5)	—
SUDEP	—	9/15 (60)
ESRD	—	1/15 (6.7)
Aspiration	—	1/15 (6.7)
Breast cancer	—	1/15 (6.7)
Sepsis	—	1/15 (6.7)
Multiple causes	—	1/15 (6.7)
Unknown	—	1/15 (6.7)

Definition of abbreviations: AML = angiomyolipoma; ARDS = acute respiratory distress syndrome; ESRD = end-stage renal disease; LAM = lymphangioleiomyomatosis; SUDEP = sudden, unexplained death due to epilepsy.

^*Age represents the age at last known follow-up or death.

^†There is a statistically significant difference in vital status in patients with tuberous sclerosis complex with and without LAM (P = 0.0003).

Table 2.

Demographic and clinical characteristics of the adult women in our cohort

Variable	LAM (n = 66)	No LAM (n = 164)
Age, range, yr*	38 (21–75)	28 (18–84)
Seizure disorder, n (%)	40 (61)	94 (57)
Renal AMLs, n (%)	62 (94)	127 (77)
Outcome, n (%)†
Alive	45 (68)	124 (76)
Deceased	8 (12)	5 (3)
Unknown status	13 (20)	35 (21)
Cause of death, n (%)
LAM	6/8 (75)	—
ARDS	1/8 (12.5)	—
Encephalopathy	1/8 (12.5)	—
SUDEP	—	2/5 (40)
Breast cancer	—	1/5 (20)
Sepsis	—	1/5 (20)
Multiple causes	—	1/5 (20)

Definition of abbreviations: AML = angiomyolipoma; ARDS = acute respiratory distress syndrome; LAM = lymphangioleiomyomatosis; SUDEP = sudden, unexplained death due to epilepsy.

^*Age represents the age at last known follow-up or death.

^†There is a statistically significant difference in vital status in adult women with tuberous sclerosis complex with and without LAM (P = 0.03).

The estimated life expectancy in patients with TSC was 70 years. The most common causes of death were sudden, unexplained death due to epilepsy (SUDEP) and LAM. SUDEP was the most common cause of mortality overall, with a death rate of 16.5 per 1,000 person-years, typically at a younger age (median age of death, 29 yr; range, 6–38 yr) than occurred in LAM (median age of death, 44.5 yr; range, 30–60 yr). Patients with LAM had a life expectancy of 63 years, and were more likely to die from their pulmonary disease than other causes (Tables 1 and 2). Adult women with TSC-LAM were at increased risk of death compared with adult women with TSC, but without LAM (Figure 1); corresponding life expectancies were 63 years and 70.5 years (difference: 7.5 yr; 95% confidence interval = 0.1–15.2). There were three male patients with LAM in our cohort; all three had mild disease and were alive at the time of data abstraction.

Figure 1.

Kaplan Meier curve for entire tuberous sclerosis complex (TSC) cohort from birth (left), and for the adult women after further dividing them into women with lymphangioleiomyomatosis (LAM; red), without LAM (blue) and unknown (green). The curve in the right panel was developed using a time-varying extended Kaplan Meier estimator (23).

Discussion

The major findings from our study are the following: 1) estimated life expectancy in patients with TSC is 70 years, which falls below the current average life expectancy of 79 years for people living in the United States (13); 2) SUDEP is the most common cause of mortality in patients with TSC; 3) death due to SUDEP occurs at a relatively young age in this population; and 4) LAM is the second leading cause of death in patients with TSC overall, and the leading cause of death among adult women with TSC.

To our knowledge, there are only two previously published studies that have examined rates and causes of mortality in TSC (16, 17), and neither have cited LAM among the most common causes of death. Shepherd and colleagues (17) performed a retrospective analysis of 355 patients with TSC evaluated at the Mayo Clinic and documented 49 deaths (13.8% of the cohort). In a more recent analysis, Amin and colleagues (16) studied 284 patients with TSC seen at a referral clinic based in the United Kingdom and recorded 18 deaths (6.3%). In both studies, complications from renal AMLs were the leading cause of death (16, 17). Although the overall mortality rate in our cohort (26 deaths out of 567 [4.6%]) and the dominance of SUDEP deaths are consistent with those of Amin and colleagues, the causes of death otherwise varied, with higher frequency of deaths due to LAM and fewer renal deaths in our patients. It is possible that this discrepancy is related to a regional bias toward more aggressive interventional approaches, including embolization procedures and earlier treatment with mechanistic target of rapamycin (mTOR) inhibitors for patients with renal AMLs.

SUDEP was a substantial cause of mortality in our cohort, consistent with the observations of Amin and colleagues (16). Several features of TSC are associated with an increased risk of SUDEP, including structural brain abnormalities, intellectual disability, multiple anticonvulsant medications, and intractable seizures (18). The risk of SUDEP in patients with TSC is substantially higher than in those with general epilepsy (18), and warrants a high level of vigilance and an aggressive treatment approach in this population.

LAM was the second leading cause of death in the entire TSC cohort, and the leading cause of death among adult women with TSC, indicating that the diagnosis of LAM has major implications for the overall prognosis of patients with TSC. TSC-LAM is generally believed to be a milder, less-progressive pulmonary disease compared with sporadic LAM (19). However, in a recent report from the NHLBI intramural cohort, the rate of decline of lung function was similar in severity-matched patients with TSC-LAM and sporadic LAM (20). Similar rates of lung function decline among patients with TSC and sporadic LAM were also seen in another recent analysis of the NHLBI LAM Registry (9). The results from our analysis further establish the prognostic implications of LAM in patients with TSC, and highlight the need for appropriate screening and close monitoring of pulmonary function as recommended by the TSC guidelines (8).

There are several limitations to this study. We were unable to determine vital status on a large proportion (19%) of the overall cohort. The effect of these “unknown status” subjects and missing diagnosis dates in a subset of patients with LAM on the overall mortality rates is unknown. This is a single-center analysis, and biases associated with referral and center-specific practice patterns need to be considered when extrapolating these findings to the general TSC population. During the course of this study, the TSC clinic at CCHMC was actively involved in multiple clinical trials evaluating the role of mTOR inhibitors for the treatment of TSC (21, 22). The U.S. Food and Drug Administration approved the use of mTOR inhibitors for the treatment of subependymal giant cell astrocytomas in 2010, AMLs in 2012, LAM in 2015, and epilepsy in 2018. The potential impact of mTOR inhibitor use on mortality may partially explain the overall low mortality in our cohort compared with the earlier results obtained by Shepherd and colleagues (17). We were unable to adjust for the potential effect of mTOR inhibitor use on overall outcomes in our analysis, and submit that this remains a future research priority.

Conclusions

LAM contributes substantially to the morbidity and mortality in patients with TSC. Close observation and monitoring of patients with TSC-LAM, similar to patients with sporadic LAM, is warranted.