Underwriting NT-proBNP. NEHOUA Fall 2014 Michael Clark, MD FACC DBIM

Transcription

1 Michael Clark, MD FACC DBIM

2 Agenda: BNP FAQs Case 1: Deciding on BNP thresholds Case 2,3: BNP and co-morbidities Case 4,5: BNP and "offsets" And some other brain teasers. 2

3 BNP Discovery and development for mortality risk assessment Year Findings Reference 1986 Atrial natriuretic factor in humans Hypertension (Crozier) Brain natriuretic factor produced by ventricles Circ Res (Ogawa) Breathe Not Properly study N Eng J Med (Maisel) BNP in insured applicants J Insur Med (Illango) BNP for mortality risk assessment J Insur Med (Clark) 2014 Source Heart (Bettencourt)

4 BNP is elevated by cardiac STRESS Prevalence of heart failure (US population) NHANES (Bui)

5 BNP identifies both systolic and diastolic cardiac dysfunction Source: barnesjewish.org 5

6 Causes of BNP elevation (besides heart failure!) Acute coronary syndromes (unstable angina, myocardial infarction) Atrial fibrillation Valvular heart disease Left ventricular hypertrophy/hypertension Sepsis Pulmonary embolism/pulmonary hypertension COPD Renal dysfunction (NT-proBNP is excreted by the kidneys) 6

7 Case 1: Deciding on BNP thresholds 71 y.o. female smoker Normal ECG Normal TM to early Stage III Normal echo: EF 60% NT-proBNP : 228 pg/ml 7

8 Current and future clinical use of BNP Each clinical indication may require different ranges of "normal"! Source: J Am Coll Cardiol (Maisel)

9 BNP as a screening tool in an asymptomatic population Source: EHJ (Linssen PREVEND cohort) 2010 mean age 49 Mortality increased at much lower BNP ranges (dotted line) Over half of the population were at increased risk (bars) 9

10 Clinical prognostic use of NT-proBNP Source: JACC (Omland PEACE Trial) 2007 Cohort (n= 3761; aged years) stable CAD Quartile 4 >98 pg/ml (males); > 115 pg/ml (females) Low NT-proBNP threshold for adverse prognosis in CAD patients 10

11 Clinical reference ranges (heart failure): BNP vs. NT-proBNP BNP <100 pg/ml - HF unlikely >400 pg/ml - HF likely pg/ml - Use clinical judgment NT-proBNP <300 pg/ml - HF unlikely Age <50 years, NT-proBNP >450 pg/ml - HF likely Age years, NT-proBNP >900 pg/ml HF likely Age >75 years, NT-proBNP >1800 HF likely Symptomatic patients with dyspnea! Source: Heart (Bettencourt)

12 Clinical use of BNP/NT-proBNP: approach and challenges BNP/NT-proBNP measurement is valuable clinically for: Diagnosis: determining the cause of pulmonary insufficiency in the ER Prognosis in those at risk: multiple risk factors, known CAD, known CHF BNP/NT-proBNP clinical challenges Settling on a standard "cut point" value as "normal" Defining "heart failure" Monitoring treatment BNP measurement variability 12

13 NT-proBNP in an insurance population: distribution NT-proBNP level % of general population Source: EHJ (Galasko) 2005 % of insurance cohort Age 55 Age 75 > 200 pg/ml (males) <5% 1% 15% > 300 pg/ml (females) <5% 1% 13% 144,000 applicants for life insurance mostly routine requirements BNP thresholds for "normal" are different in an insurance population Source: JIM (Clark, Kaufman, Fulks, Dolan, Stout)

14 NT-proBNP in an insurance population: results Age and smoking covariates NTproBNP Number Total Mortality Ratio (pg/ml) deaths Tested (Cox) (ref) ,965 Mortality Ratio 95% CI Lower Age and CVD tertile covariates Upper 1.0 (Cox) 95% CI Lower Upper to , Table 5. Mortality ratios (Cox regression) for age 50 to 89, with no missing physical measurements, no missing egfr, and no history of heart disease Source: JIM (Clark, Kaufman, Fulks, Dolan, Stout)

15 NT-proBNP in an insurance population: results 1100% Mortality ratio 1000% 900% F 50 to % F 70 to % M 50 to % M 70 to % 400% 300% 200% 100% 0% (ref) NT-ProBNP value Figure 1. Mortality ratios for NT-proBNP values by age group and sex, cases denying history of heart disease. Source: Clark, Kaufman, Fulks, Dolan, Stout. J Insur Med

16 NT-proBNP in an insurance population: subgroups Source: Clark, Kaufman, Fulks, Dolan, Stout. J Insur Med 2014 Age/sex NT-proBNP Number Total Mortality Ratio group (pg/ml) deaths Tested (Cox) F 50 to (ref) 54 13, , , F 70 to 89 Mortality ratios (Cox regression) by age and sex for NT-proBNP bands, cases denying history of heart disease Lower Upper , , , , (ref) 34 3, , , , , (ref) 65 8, , , , ,001+ M 50 to 69 M 70 to 89 95% CI (ref) 16

17 NT-proBNP - what about very low values? No clear pattern emerged from our subgroup analysis of low BNP values Source: Clark, Kaufman, Fulks, Dolan, Stout. J Insur Med

18 Case 1: Deciding on thresholds 71 y.o. female smoker Normal ECG Normal TM to early Stage III Normal echo: EF 60% NT-proBNP : 228 pg/ml 18

19 Insurance use of BNP/NT-proBNP: approach and challenges BNP measurement strengths Predicts mortality risk proportional to BNP level The predictive ability of BNP is independent of classical cardiovascular risk factors BNP/NT-proBNP underwriting challenges BNP "cut-offs" need to be scaled based on age and gender What is the impact of "co-morbidities" and what are the "offsets" that can be used when faced with an elevated BNP result?? 19

20 Agenda: BNP FAQs Case 1: Deciding on BNP thresholds Case 2,3: BNP and co-morbidities Case 4,5: BNP and "offsets" And some other brain teasers. 20

21 Case 2: BNP with co-morbidities 1100% 1000% 900% 800% 700% 600% 500% 400% 300% 200% 100% 0% F 50 to 69 Mortality ratio F 70 to 89 Insurance lab: NT-proBNP: 1100 pg/ml M 50 to 69 M 70 to (ref) NT-ProBNP value 69 y.o. male $250,000 Medical history: Sick-sinus syndrome Pacemaker implanted ECG: atrial fibrillation Echo: EF 45%, mild LV and LA dilatation

22 Coexisting factors that impact BNP levels Impact on BNP level Non-cardiac factors Cardiac factors Increased BNP level Older age Female gender Renal failure Sepsis + Pharmacokinetics Heart failure Coronary ischemia/infarction Valvular heart disease Pulmonary heart disease Atrial fibrillation Pacemakers (ventricular) Decreased BNP level Obesity Treatment BNP variability Significant day change: 25% decline Significant weekly change: 72% decline NT-proBNP variability Significant day change: 11% decline Significant weekly change: 47% decline Source: J Am Coll Cardiol (Maisel)

23 Case 2: BNP with co-morbidities 1100% 1000% 900% 800% 700% 600% 500% 400% 300% 200% 100% 0% F 50 to 69 Mortality ratio F 70 to 89 Insurance lab: NT-proBNP: 1100 pg/ml M 50 to 69 M 70 to (ref) NT-ProBNP value 69 y.o. male $250,000 Medical history: Sick-sinus syndrome Pacemaker implanted ECG: atrial fibrillation Echo: EF 45%, mild LV and LA dilatation

24 Case 3: More about co-morbidities 72 y.o. male $750,000 Medical history: Diabetes, controlled Hypertension, controlled No other cardiac studies in the file NT-proBNP: 660 pg/ml 24

25 BNP and diabetic cardiomyopathy both systolic and diastolic dysfunction Significant diastolic dysfunction is found in >50% of asymptomatic type 2 diabetics Elevated BNP levels are associated with microalbuminuria, nephropathy, and autonomic neuropathy Source: Circ Res (Poornima)

26 Case 3: More about co-morbidities 72 y.o. male $750,000 Medical history: Diabetes, controlled Hypertension, controlled No other cardiac studies in the file NT-proBNP: 660 pg/ml 26

27 Agenda: BNP FAQs Case 1: Deciding on BNP thresholds Case 2,3: BNP and co-morbidities Case 4,5: BNP and "offsets" And some other brain teasers. 27

28 Case 4: BNP and "offsets" 62 y.o. male $1 million CTA done for family history of CAD Diffuse CAD Abnormal calcium score Normal stress test to Stage 4 Normal echocardiogram Carotid ultrasound with mild bilateral plaque BNP: 108 pg/ml 28

29 ECGs: evolution of the underwriting approach ECG finding Life Re Guide 1993 Life Guide 2013 Change in rating (%) First degree block +50 STD - 33% Left axis deviation +50 STD - 33% Major ST, T changes / % Positive treadmill % Paroxysmal atrial fibrillation +100 STD - 50% Factors impacting ECG protective value: Mortality improvement Age/amount guidelines Interpretation "flexibility" Rating "evolution" 29

30 BNP and exercise testing in CAD 355 patients (aged years) with stable CAD undergoing stress echocardiography 30% (quartiles I and II) of stable CAD patients will have normal BNP levels but positive exercise tests >50% of CAD patients with high BNPs will have negative stress tests Source: Circulation (Bibbins_Domingo)

31 BNP plus calcium scanning BNP and CAC provide complementary prognostic information A negative result on either test implies improved survival over situations where both tests are positive A positive result for both suggests increased risk Source: EISNER cohort: 2400 asymptomatic adults (mean age years) AJC (Shaw)

32 Underwriting BNP: "judgment" factors Assessment impact Echo/nuclear results J Am Coll Cardio (McKie) 2010 ECG/Exercise testing results Circulation (Bibbins-Domingo) 2003 Am J Cardiol (Mathewkutty) 2013 Coronary imaging results Coron Artery Dis (Sahinarslan) 2005 Calcium scan results Am J Cardiol (Shaw) 2009 Classic cardiac risk factors N Eng J Med (Wang) 2006 J Am Coll Cardiol (Smith)

33 Case 4: BNP and "offsets" 62 y.o. male $1 million CTA done for family history of CAD Normal stress test to Stage 4 Normal echocardiogram Carotid ultrasound with mild bilateral plaque BNP: 108 pg/ml 33

34 Case 5: BNP and "offsets" 71 y.o. male $5 million Inferior MI 5 years ago EF 40%-59% Stress echo: Inferior akinesis but no other changes to Stage 3 exercise NT-proBNP : 2220 ng/ml 34

35 BNP with other insurance testing In "healthy normal" subgroup (=no cardiovascular risk factors or echo abnormalities) there was no increased risk of death, heart failure, CVA or MI with increased BNP In stage A/B heart failure subgroup, elevated NT pro-bnp was independently associated with these events, even after adjusting for traditional risk factors Mayo/Olmstead county community cohort 703 "healthy normal" & 1288 stage A/B heart failure patients diagnosis based on echo and clinical assessment All patients underwent labs, ECGs, and treadmills Analysis: Kenneth Krause, MD 35

36 Sequential testing, 1% prevalence (BNP then TMT favored) Analysis: Kenneth Krause, MD 36

37 Decision analysis: model to assess value of NTproBNP Conclusions: Separate testing model: NT-proBNP gave the "most bang for the buck" in low or medium risk patients. (pretest probability of 1%, 5%, 10%) Treadmills become more protective/cost-effective in higher risk patients (pre-test probability >20%) Sequential model: Perform NT-proBNP first then do a treadmill if NT-proBNP abnormal Important: the value of the treadmill suffered in this analysis because of cost as compared to NT-proBNP 37

38 Agenda: BNP FAQs Case 1: Deciding on BNP thresholds Case 2,3: BNP and co-morbidities Case 4,5: BNP and "offsets" And some other brain teasers. 38

39 Case 6: Any chance? 76 y.o. female applying for $1 million WL Known "cardiomyopathy" for years "Walks 2/4 miles almost every day" ECG: LBBB NT-proBNP: 5135 pg/ml 39

40 Case 7: Any chance? 63 y.o. male $2 million WL Anterior MI stents placed at the time of MI: proximal LAD (90% stenosis) and midlcx (99%) LV function: "mildly impaired consistent with MI" ECG: LBBB Follow-up 2012 (after 4-year absence) Treadmill: 9 minutes (10 METs). ECG LBBB. SPECT: "No ischemia". EF 36% "mildly impaired" NT-proBNP: 81pg/ml 40

41 BNP a perspective BNP rating thresholds should be related to insurance distribution and scaled for age, gender, and possibly build NT-proBNP measures cardiac stress, so will be elevated in a number of cardiac conditions, including CAD, valvular heart disease, atrial fibrillation, and left ventricular hypertrophy. Unless the NT-proBNP level is extremely high (>1000 ng/ml), our impairment rating should cover the risk of a small/moderate increase with no or only a small additive rating NT-proBNP is giving us risk information that is different than ECGs, calcium scanning or exercise testing. They can be added together or used as offsets to get to the best risk assessment Very low BNP levels may be useful as rating "credits", particularly in cases of valvular heart disease or left ventricular hypertrophy of questionable severity In the final analysis, BNP underwriting requires guidelines and good judgment! 41

42 Thank you

43 Legal notice 2014 Swiss Re. All rights reserved. You are not permitted to create any modifications or derivatives of this presentation or to use it for commercial or other public purposes without the prior written permission of Swiss Re. Although all the information used was taken from reliable sources, Swiss Re does not accept any responsibility for the accuracy or comprehensiveness of the details given. All liability for the accuracy and completeness thereof or for any damage resulting from the use of the information contained in this presentation is expressly excluded. Under no circumstances shall Swiss Re or its Group companies be liable for any financial and/or consequential loss relating to this presentation. 43