Timeline of LLFS
Visit 1 - Creation of the cohort
The Long Life Family Study (LLFS), established in 2005 in response to an NIA RFA, enrolled families enriched for exceptional longevity (EL), to discover factors that contribute to healthy aging and survival. The “family longevity selection score” (FLoSS) was developed to select families with clustering of EL (Sebastiani et al.2009). From 2006 to 2009, LLFS enrolled 539 sibships (G1, probands), their offspring (G2, offspring) and spouses (total 4,953). . The probands were at least 78 years old in the US centers and 86 years old in Denmark. The families were selected to participate in the study based on The Family Longevity Selection Score (FLoSS), a score generated according to birth year cohort survival probabilities of the proband and their siblings. Probands and their families with FLoSS score of 7 or higher, at least one living sibling, and at least one living offspring (minimum family size of 3). Comparison with a referent cohort reveals that LLFS families have strong exceptional clustering of EL. The least exceptional LLFS G1 sibship shows more clustering of longevity than 99% of the Framingham Heart Study (FHS) sibships (Sebastiani et al. 2009). LLFS participants were comprehensively phenotyped by in-home visits (V1). The G2 offspring have a variety of Healthy Aging Phenotypes (HAPs), defined as an unusually low age-specific prevalence of one or more specific conditions or risk factors, compared to population-based cohorts suggesting enrichment of shared (possibly genetic) protective effects in LLFS families. There is considerable heterogeneity of HAPs in LLFS families: e.g. some show exceptional cognition (Barral et al. 2013); others show strong linkage to metabolic HAPs (An et al. 2013). Thus, LLFS is a unique resource for studying familial EL and its relationship to HAPs. In particular, G2 participants represent an unprecedented opportunity to prospectively study the progression of HAPs as related to survival in a cohort likely to show eventual higher EL prevalence.
The in-home visit consisted of demographic, anthropometric, cognitive, activities of daily living, ankle-brachial index, blood pressure, grip strength, pulmonary function measurements. Serum, plasma, lymphocytes, red cells, and DNA are stored for future studies. An array of biochemical assays were completed for Visit 1.
Establishment of the LLFS cohort
The original recruitment and examination for LLFS was conducted over a three-year period, April 2006 – May 2009. LLFS successfully enrolled and extensively phenotyped 4,953 individuals from 539 two-generational families that demonstrated clustering for exception longevity in the upper generation from three US field sites (Boston University, Boston, MA; Columbia University, New York, NY; and University of Pittsburgh, Pittsburgh, PA) and one field site in Denmark (University of Southern Denmark, Odense, Denmark), Table 1. These families included long-lived individuals (generally 90+), their siblings, spouses, offspring, and spouses of the offspring (a referent group). The LLFS has continued to recruit family members, growing to 5,097 participants (1,704 proband generation, ~1/3, and 3,393 offspring generation, ~2/3).
Identification of Families/ FLoSS Score
The U.S. field centers used Center for Medicare and Medicaid Services lists of Medicare enrollees to mail a recruitment brochure. The initial file included people who were at least 79 years old on January 1, 2005; had no recorded date of death; were not in the end-stage renal disease or hospice programs; lived in zip codes “near” (within 3 hours driving distance) one of the 3 U.S. study centers (Boston University Medical Center in Boston, MA, Columbia College of Physicians and Surgeons in New York City, NY, and the University of Pittsburgh in Pittsburgh, PA). A pilot mailing tested the yield of families recruited from mailing to individuals in their 80’s and higher age strata. Based on these yields, subsequent mailings targeted those age 89 and older. Study participants were also recruited from local communities using mailed brochures, posters, web-based media and newspaper advertisements as well as community-based presentations. Additional mailing lists were obtained through voter registries and purchased public domain lists from various commercial vendors (Newman et al., 2011). The University of Southern Denmark field center identified individuals who would be ages 90 and above during the study recruitment period through the Danish National Register of Persons, which contains current information on names, including past names such as maiden names for women, addresses, place of birth, marriages, and vital status (Pedersen, Gotzsche, Moller, & Mortensen, 2006). Second, using information on the place of birth and the names, parish registers available in regional archives were searched to locate the parents of the elderly individuals in order to identify sibships. Based on the above information, potentially eligible families were identified and contact was made with potential probands to further assess the family’s eligibility for and willingness to participate in the LLFS using criteria parallel to that used in the United States.
The “family longevity selection score,” or FLoSS, was developed to select for families with sibships that were rare by virtue of the number of siblings achieving rare percentiles of survival, with an added bonus for being alive and therefore available for study (Sebastiani et al., 2009) The FLoSS was calculated for each family (Sebastiani et al., 2009), and included information on the current age or age at death of siblings, the sibship size and the number of living individuals available to study. Potential families were ranked by their FLoSS. To be eligible, a proband family had to have a FLoSS of 7 or greater and a minimum family size of 3 (specifically the proband, at least one living sibling, and one offspring), all willing and able to give informed consent and participate in the interview and in-home examination including providing a blood sample for serum and DNA extraction.
Components of the In-Person Examination
Interviews and examinations were conducted in the home setting with portable equipment by centrally trained and certified research assistants using a common protocol. The in-home visit consisted of demographic, anthropometric, cognitive, activities of daily living, ankle-brachial index, blood pressure, grip strength, pulmonary function measurements. All family members, regardless of age, undergo as much of the same assessment as possible. We developed proxy interview formats for individuals who are too ill cognitively and/or physically to participate in part or all of the in-person examination. If an in-home visit was not feasible, we conducted a comprehensive telephone interview and obtained biological specimens (blood or saliva sample) using an outside laboratory or physician’s office. Selected measures were designed to assess aspects of EL that: 1) have significant heritability, 2) are related to longevity and healthy aging, and 3) can be assessed in the home setting.
Generated resources include DNA, PAXgene tubes for RNA profiling, plasma and sera from whole blood, and banked lymphocytes. All participants provided informed consent approved by the field centers’ Institutional Review Boards (IRBs) for sharing genotype and phenotype on dbGaP. Genome-Wide Association data were generated using Illumina 2.5M Omni SNP array, which were imputed to 38 million SNPs using 1000 Genomes imputation, resulting in a number of publications (An et al., 2014; Bae et al., 2013; Barral et al., 2014; Feitosa et al., 2014; Lee et al., 2013; Minster et al., 2015; Thyagarajan et al., 2014) and which have been shared with the broader scientific community via dbGaP (dbGaP Study Accession: phs000397.v1.p1).
Annual follow-up has been performed annually by phone interview to update vital status and medical conditions. Expanded follow-up is conducted every year for the proband generation and includes reported physical function, activities of daily living and telephone assessed-cognitive function. Because they are generally so much younger these items are assessed only every three years for the offspring generation. However, once an offspring generation participant reaches age 70, they are given the expanded follow-up yearly.
GWAS Analysis Period
In the second funding period (2010-2013), annual reassessments continued by questionnaires. Analyses of stored V1 blood provided additional HAP measures. We conducted a 2.5 million SNP GWAS (dbGaP phs000397); developed an efficient high-throughput technique by which we sequenced ~450 candidate genes for HAPs and EL (Ramos et al. 2013); replicated many variants previously discovered for selected HAPs; and found additional ones. 54% of LLFS G1 and 92% of G2 remain alive. Subject retention has been 94%.
Visit 2 Overview
The third funding period of LLFS conducted a second in-person examination (V2) in order to prospectively study rates of change in Healthy Aging Phenotypes (HAPs) with age and identify genetic and other factors contributing to HAPs and longevity. We hypothesize that exceptional longevity (EL) and HAPs entail common and rare variants that individually have modest effects, but which in combinations strongly influence longevity and specific HAPs, and may only be detectable in family studies enriched for HAPs, such as LLFS. HAPs evaluated at the initial in-person visit show strong linkage peaks which are not explained by common haplotypes interrogated by GWAS (HLODs ranging from 6.0-45.1). These are likely driven by rare, lineage-private alleles that will only be found by sequencing specific families, and may point to important new biology.
The Specific Aims were as follows:
Specific Aim 1: to conduct a second in-home examination on all surviving LLFS participants.
Specific Aim 2: to analyze cross-sectional and longitudinal phenotypes. The goal is to identify pathways for EL and HAPs by characterizing the shared and distinct LLFS phenotypes and environmental factors. We will characterize individual longitudinal patterns of HAPs to identify subgroups showing similar patterns and exceptional phenotypes. We will test whether these HAPs are heritable, and test for differences with internal and external referent groups.
Specific Aim 3: to find genes/variants associated with cross-sectional and longitudinal phenotypes using a) Whole Exome Sequencing to comprehensively search for coding variants associated with HAPs and EL and b) Targeted Regulatory Sequencing of regions under linkage peaks for HAPs in selected families showing the strongest linkage evidence.
Specific Aim 4: to replicate our genetic and epidemiological findings in other aging study cohorts.
Second In-Person Examination, V2
The second in-person visit, collecting all information at Visit 1 plus a carotid ultrasound, additional cognitive testing, and fatigability, was completed from 2014-2017. Biological samples were repeated at Visit 2 and are also stored for future studies. A subset of biochemical assays from Visit 1 were performed at Visit 2.
Visit 2 was designed to measure longitudinal change from Visit 1. Therefore, at the Visit 2 examination, we repeated measures that were expected to change over time, updated medical history and medications and repeated a blood draw. We enhanced the Visit 2 examination by adding portable carotid ultrasound to better define vascular health. Visit 1 and Visit 2 were approximately 8 years apart on average.
LLFS Cohort Retention
LLFS participant retention has been facilitated primarily through the annual phone interviews. These have a high completion rate of 81-85%, depending on the generation. In order to combat the primary reasons for disengaging from the study (proband generation: “too sick” or offspring generation: “no longer interested”) the Field Centers send holiday greeting cards, calendars, birthday cards, and short research updates as well as an annual newsletter. As of October 2018, 73% of the proband generation have died, as have 7% of the offspring generation.
Longitudinal Change Phenotypes
Longitudinal study has been critical for characterizing rates-of-change in aging-related phenotypes. An important first step is to accurately assess personal longitudinal change for each LLFS participant for every key phenotype. Longitudinal change is more accurately estimated (and thus shows higher heritabilities) using a Random Coefficient Model (RCM) (Laird & Ware, 1982), a.k.a Growth Curve Model rather than simply calculating the traditional Ordinary Least Squares approach to longitudinal change [ (Post-Pre phenotype difference) per unit time]. The RCM simultaneously estimates each participant’s personal trajectory while also using all participant’s data. It then transforms the data from outcomes at time points to personal slopes and intercepts. These personal slopes and intercepts then become new longitudinal change phenotypes. We systematically generated individual growth curve trajectories for all key phenotypes and all participants in LLFS.
Visit 3, ongoing (2020-present)
The LLFS was funded to perform a third in-person visit on participants already enrolled and for the first time to enroll interested grandchildren of the proband generation. This was scheduled to begin in 2020, however due to the COVID-19 pandemic, no in-person visits are currently planned in the US. Our Denmark site is able to perform in-person visits, as their pandemic hot-spots are much less than in the US.
In an attempt to reconnect with our participants in the US and keep our study moving forward, we have devised a new split visit for the in-person visit. The first part will be a video (Zoom) call to collect as much phenotypic data as possible. If your residence is located in an area that is deemed low-risk of COVID-19 transmission, we will then send a phlebotomist to your house (in protective equipment and give the participant an N-95 mask to wear) to perform a blood draw. If your residence is not located in a low-risk area, we will continue to monitor conditions and obtain a blood specimen when it is safe. The second part of this visit would then be an in-person visit (when safe) to obtain measurements we are not able to do via video, such as anthropometrics, blood pressure, physical performance measures, carotid ultrasound, ankle-brachial index, and others. Additionally, during this time, our annual telephone follow-up with all of our participants will continue and we look forward to talking again with each and every one of you!