Cascade Screening for FH: the U.S. experience Paul N. Hopkins, MD, MSPH Professor of Internal Medicine Cardiovascular Genetics University of Utah
Disclosures Consultant Genzyme, Amgen, Regeneron Research funding Regeneron, Merck
A striking family history Roger R. Williams, MD Aug 11, 1944 Sep 2, 1998 Roger was shocked as a teenager when his neighbor died from from MI on the golf course at age 42. And more shocked to learn the family had expected it! Several generations of males in this family had died from MIs before meeting their grandchildren.
Initial efforts in Utah Characterization of Coronary Prone Pedigrees Roger R. Williams, MD, principal investigator NIH grant 1977 1992 Goal: to understand the genetics and epidemiology of different types of early familial coronary disease...
Syndromes in 97 Utah Multiplex Families (2+ cases with early CAD and risk factor) Hypertension 24% FDH (MetS) 12% Diabetes 8% Homocysteine 12% Cigarette Smoking 37% High Lp(a) 23% FCHL 12 60% High TG Only 14% High TG, Low HDL 7% All Low HDL 23% Type III 3% High LDL Not FH 5% FH 4% Williams RR, et al. Arch Intern Med 1990; 150:582 and unpublished
Familial dyslipidemias in premature CHD kindreds (n = 102) No identifiable familial abnormality 43% L(a) excess 18.6% Familial dyslipidemia 14.7% Familial apo A-1 deficiency 1.0% Familial high TG 1.0% FH 3.0% Familial hyperapobeta 5.0% Familial hypoalpha 4.0% FCHL 13.7% Genest JJ, et al. Circulation 1992; 85:2025
Familial Syndromes of Early CAD in Seattle, Boston, and Utah Seattle (N = 366) Boston (N = 102) Utah (N = 97) No diagnosis 60% 35% 25% Familial combined hyperlipidemia 13% 18% 36% Familial hypertriglyceridemia 6% 18% 21% Isolated low HDL 4% 15% all low HDL 35% 20% Elevated Lp(a) 19% 23% Familial hypercholesterolemia 4% 3% 4% Type III hyperlipoproteinemia 2 3% 0% 3% Hypertension 24% Diabetes 8% Smoking 37% Goldstein JL, et al. J Clin Invest 1973; 52:1544 Genest JJ Circulation 1992 Williams RR Arch Int Med 1990
MEDPED A humanitarian project to find and help persons with familial hypercholesterolemia
Brief Historical Review of MEDPED MEDPED was officially started at the University of Utah by Roger Williams in 1989 First 5 years funding from the CDC established MEDPED methods refined ascertainment methods since then
Major findings from CDC study Documented Need for More Effective Diagnosis and Treatment of Familial Hypercholesterolemia According to Data from 502 Heterozygotes in Utah 101 FH probands with 502 identified relatives 50% had been told they had very high cholesterol 31% of those had been told they had FH 42% taking a cholesterol lowering medication 23% had total C or LDL C below the 90 th percentile Cascade screening most cost effective way to find new probands Williams RR, Am J Cardiol 1993; 72:18D
0.01 0.01 Non-FH FH Probability 0.01 0.00 0.00 0.00 0.00 0.00 False negatives 80% probability of FH False positives 0 50 100 150 200 250 300 350 400 450 500 LDL cholesterol (mg/dl) 1993 LDL cholesterol levels mean SD FH 289 63.0 (OLD) Non-FH 130 31.4 (OLD) First degree relatives Hopkins PN. Clin Lipidol 2010; 5:339
1993 MEDPED Criteria for Diagnosing Heterozygous FH in U.S. Population Cut points chosen for 80% probability that an individual is affected. ADDITIONAL CRITERIA were applied to Dx a newly identified family. Shown are total C (LDL C) in mg/dl Degree of Relation to Closest FH Age First Second Third General 100% <20 220 (155) 230 (165) 240 (170) 270 (200) (240) 20-29 240 (170) 250 (180) 260 (185) 290 (220) (260) 30-39 270 (190) 280 (200) 290 (210) 340 (240) (280) 40+ 290 (205) 300 (215) 310 (225) 360 (260) (300) Williams RR, et al. Am J Cardiol 1993; 72:171
General philosophy of MEDPED diagnostic approach FH is a dominantly (or co dominant) transmitted disease. To provide a clinically definite diagnosis of FH, there should be evidence of dominant transmission in the family. LDL C levels in family members besides proband were used to diagnose FH using additional criteria or rules. NOTE: other diagnostic approaches do not explicitly utilize actual LDL levels in the family for diagnosis.
Examples of additional criteria to diagnose new FH probands/families For DEFINITE HETEROZYGOTE diagnosis by gene testing (rare) For CLINICALLY DEFINITE HETEROZYGOTE Tendon xanthoma present and LDL C meets at least first degree relative criteria. Isolated pediatric case (<20 years old) with LDL cholesterol 240 mg/dl (>99.9 percentile). Two first degree relatives meet general population criteria for LDL cholesterol. ETC!! (including rules for probable, etc. WHO Consultation on FH http://libdoc.who.int/hq/1998/who_hgn_fh_cons_98.7.pdf
Industry funding begins in 1994 5 centers added Cincinnati (Stein) Baltimore (Kwiterovich/Kafonek) Boston (Lees) Portland (Illingworth) San Francisco (Kane/Malloy)
U.S. MEDPED Registry Growth Number of Cases Found 9000 8000 7000 6000 5000 4000 3000 2000 1000 0 Utah USA + Utah 1989 1991 1993 1995 1997 1999 2001 Year
MEDPED Case Finding (FH Only) by Center Coord. Center Utah Cincinnati Portland San Francisco Lipid Clinics Baltimore Cambridge relatives probands 0 1000 2000 3000 4000 Number registered
Total C Levels in the Treatment Support Program Total cholesterol (mg/dl) 400 380 360 340 320 300 280 260 240 220 200 <1993 (14) 1993 (21) 1994 (43) 1995 (75) Baseline Final (by 2000) 1996 (176) 1997 (56) Year (number) Stephenson SH, et al. J Clin Lipidol 2009; 3:94
Multiple lipid phenotypes in a Utah FH kindred with a novel LDLR mutation (D92K) Affected Chol Trig HDL LDL Age 64 Ages 24 50 302 106 47 237 443 278 40 286 222 94 50 159 Ages 4 21 294 370 45 168 223 164 48 153 226 240 34 143 200 316 36 101 259 160 37 195 131 59 45 74 292 193 40 204 200 103 57 127 268 149 48 185 298 240 50 185 324 95 54 257 326 98 58 232 252 103 64 157 228 222 107 44 139 184 119 44 121 213 172 45 129 248 183 41 164 223 109 42 170 166 66 56 101 204 55 48 148 Wu LL, et al. J Hum Genet 2000; 45:154
Utah FH Pedigree
What are some of the challenges of FH screening? Identifying FH patients DNA (gold standard only if found) versus clinical? Who pays for DNA diagnosis? How can clinical Dx be made available to anyone? There is urgency to finding FH Previously, the Dx usually came after MI Earlier, aggressive Rx better outcomes Universal screening of teenagers cost effective for CAD prevention
Additional challenges in the U.S. No current data on: How many FH are currently diagnosed What percent are treated (any Rx) How well FH are treated (current LDL C) CURRENT CAD RISKS OR OUTCOMES
Rethinking practical aspects of MEDPED Prior approach to diagnosis was labor intensive and relatively costly. telephone contacting and data collection diagnosis by MEDPED physician Diagnostic rules for initial FH pedigree or proband diagnosis lacked theoretical rigor, generalizability, or specific probabilities. MEDPED cutpoints had low sensitivity generally. Both general population and FH total C and LDL C mean ±SD needed updating.
Diagnostic approach using complex segregation analysis and likelihood theory..., Indices comprising all combinations of genotypes. Each combination defines a string. Likelihood products for probability of given set of genotypes. Separated into founders (prior probabilities) and offspring (transmission probabilities). Penetrance function(s). Probability of an observed phenotype given the genotype. MEDPED patent pending
Diagnostic approach using complex segregation analysis and likelihood theory..., M F C 0 0 0 0 0 1 0 1 0 0 1 1 1 0 0 1 0 1 1 1 0 1 1 1 MEDPED patent pending
Diagnostic approach using complex segregation analysis and likelihood theory..., M F C 0 0 0 0 0 1 0 1 0 0 1 1 1 0 0 1 0 1 1 1 0 1 1 1 0.998 or 0.002 Mendelian transmission probabilities MEDPED patent pending
Diagnostic approach using complex segregation analysis and likelihood theory..., M F C 0 0 0 0 0 1 0 1 0 0 1 1 1 0 0 1 0 1 1 1 0 1 1 1 0.998 or 0.002 Mendelian transmission probabilities updated means, SD penetrance functions MEDPED patent pending
Diagnostic approach using complex segregation analysis and likelihood theory Actual family with LDL C values known M F C 0 0 0 0 1 1 MI death age 39, smoked prior LDL C 216 age 24 age 11 LDL C 198 (treated) 2 most informative product strings (0.998)(0.998)(1.0)(2.34E 06)(3.04E 05) (0.998)(0.002)(0.5)(7.19E 03)(4.25E 03) = 99.3% child has FH MEDPED patent pending
What if? Clinical situation: FH Dx % we only knew daughter s LDL C 58.9% her untreated LDL C was 220 98.6% we only knew dad had MI age 39 93.3% the dad had a tendon xanthoma (almost regardless of dad s LDL C) 99.8% MEDPED patent pending
Accurate Dx of FH using this method will depend on good underlying data Prior probability of FH in founders Limited US data may need to be re assessed LDL C mean ±SD for FH and non FH general population LDL C has dropped old mean for FH was too high (ascertainment bias) CAD risk in FH and non FH some reasonable estimates exist Xanthoma by age in FH (available) almost unheard of in non FH MEDPED patent pending
Cumulative Probability of Fatal and Non Fatal CAD in 116 FH Pedigrees Cumulative Probability of Clinical CAD 0.60 0.50 0.40 0.30 0.20 Non FH Men FH Men 0.60 0.50 0.40 0.30 0.20 Non FH Women FH Women 0.10 0.10 0.00 0.00 Age Age Stone NJ, et al Circulation 1974; 49:476
Cumulative Probability of Non Fatal CAD in Utah FH vs a Random U.S. Population Cumulative Probability of Clinical CAD 0.30 0.25 0.20 0.15 0.10 0.05 Non FH Men FH Men 0.05 0.04 0.03 0.02 0.01 Non FH Women FH Women 0.00 0.00 Age Age Hopkins PN. Clin Lipidol 2010; 5:339
Cumulative probability of CAD 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 FH men FH women non FH men non FH women 20 30 40 50 60 70 80 90 Age (years
Cumulative probability of CAD 0.25 0.20 0.15 0.10 0.05 FH men FH women non FH men non FH women 0.00 20 25 30 35 40 45 Age (years
Prevalence (%) in 346 Utah FH Heterozygotes 60 50 40 30 20 10 Xanthoma Full Arcus 0 <30 30-39 40-49 50-59 60+ Age (years) Hopkins PN. Current Treatment Options in Cardiovascular Medicine 2002; 4:121
60 Tendon xanthoma prevalence by age in Utah FH 50 Prevalence of xanthoma 40 30 20 10 0 y = 0.8205x 1.8113 R² = 0.9936 prevalence of xanthoma expected to be 0 at age 2.2 0 10 20 30 40 50 60 70 Age (years)
Conclusions By 2004, about 8,000 FH had been diagnosed in the MEDPED registry about 1.3% of US FH We have a long way to go! New automated approach will help!