Evercore ISI Presentation- Madrigal

Forward-Looking StatementsAny statements, other than statements of historical facts, made in this presentation regarding our clinical studies and our research and development programs; our ability to advance product candidates into clinical studies; our anticipated clinical development milestones; the timing or likelihood of regulatory filings and approvals for our product candidates; the timing and success of our development and commercialization of our product candidates; and the potential of our product candidates to achieve clinical benefit and safely treat cardiovascular, metabolic, and liver diseases, including non-alcoholic steatohepatitis and dyslipidemia, are, or may be deemed, forward-looking statements within the meaning of the Private Securities Litigation Reform Act of 1995. In some cases, you can identify forward-looking statements by terms such as “may,” “will,” “could,” “should,” “would,” “anticipate,” “believe,” “estimate,” “continue,” “design,” “expect,” “intend,” “plan,” “potential,” “predict,” “seek” or the negative of these words and similar expressions and their variants may identify forward-looking statements.

These forward-looking statements reflect management’s current expectations, which are based on certain assumptions and involve certain risks and uncertainties, which could change over time. Actual events or results may differ materially from the events or results discussed in these forward-looking statements due to various factors. Important factors that may cause actual events or results to differ materially from those discussed in these forward-looking statements include, but are not limited to, uncertainties associated with the outcomes, cost and timing of our product candidate development activities and clinical trials; uncertainties inherent in clinical testing; the timing, cost, and uncertainty of obtaining regulatory approvals for our product candidates; our ability to successfully progress or complete further development of our product candidates; our ability to commercialize our product candidates; our ability to protect our intellectual property; our cash resources and ability to obtain working capital to fund our proposed operations; changes in the regulatory landscape, including changes in regulatory policies or positions, or the imposition of regulations that affect our product candidates; our reliance on third-parties to meet our obligations; and other factors that are described in our filings with the Securities and Exchange Commission, including our Annual Report on Form 10-K for the year ended December 31, 2017, and subsequent Quarterly Reports on Form 10-Q, including the risk factors set forth in those filings. These forward-looking statements represent management’s expectations as of the date hereof only, and we specifically disclaim any duty or obligation to update forward-looking statements as a result of subsequent events or developments, except as required by law.

MGL-3196: Phase 2 Data and Beyond

We believe the Phase 2 results:

� Establish a minimum effective dose, once a day oral, for statistically significant NASH resolution with 2 pt reduction in NAS; the endpoint is agreed with Madrigal by FDA for Phase 3 registration study

� In the extension study, demonstrated that the higher doses (80, 100 mg) will achieve levels of response similar to the ~39% NASH resolution (MRI-PDFF responders) not 27% which included all MGL-3196 patients dosed, including those with suboptimal exposures

� Presented strong fibrosis biomarker and exploratory data, especially in advanced F3 patients, suggesting 3196 has the potential to achieve the 1 -pt fibrosis reduction endpoint in a larger patient data set in Phase 3

� Established MRI-PDFF as a non invasive method to predict NASH Resolution� Demonstrated safety including benefits in liver enzymes� Demonstrated robust lipid lowering, a unique profile, that allows potential benefit in MACE as

well as liver endpoints in Phase 4 — Lipid benefits including reduction in liver fat are critical in this population — Reduction in LDL, ApoB (the atherogenic particle)—with add-on of TG, ApoCIII and Lp(a)

3

Power of the Phase 3 for NASH Resolution

� The extension study demonstrated that with the Phase 3 doses, 80 and 100 mg, approximately 90% of high risk F2/F3 patients safely achieve an MRI-PDFF response, and average 50% hepatic fat reduction—predicting an overall 35% response on NASH resolution

� NASH resolution, no worsening of fibrosis, and at least a 2-point difference in NASH has been accepted by FDA

� Even with a very conservative 23% MGL-3196 responder rate relative to 10% placebo for NASH resolution, a 900 patient study is powered >90% to achieve 0.001; in the more likely outcome, MGL-3196>25%, placebo 6% the study is powered 99%

� Power assuming drop-outs for 900 patient study:

4

Placebo Response Rate↓ MGL-3196 Response

.21 .22 .23 .24 .25.06 99 99 99 99 99.07 99 99 99 99 99.08 95 97 98 99 99.09 90 94 97 98 99.10 82 88 93 96 98

Development Path Across the Spectrum of NAFLD/NASH

5

F3

F4 Study to show metabolic benefits in cirrhosis, combination

Phase 3 NASH study (Phase 4, both MACE and liver endpoints)

F2

F1B

F1

F0

NAFLD with dyslipidemia,

diabetics, metabolic

syndrome

CV Benefits

Fatty liverLDL-CApoBTriglyceridesLp(a)

Phase 3 Lipid study(no liver biopsy requirement)

1.6 million

2 million

15 million

5 million

2.5 million

.5 million

NASH/NAFLD Spectrum

PatientNumbers (US)

18

6075

11

61

30

68 77

22

68

0

20

40

60

80

100

12 36 12 36 12 36 12 36 12 36p<0.0001 p<0.0001 ND p=0.009

>= 30% Fat Reduction (% )

F2/F3

Week

-1.6

-7.6 -8.8

-2.3

-8.5 -9.4

-11

-9

-7

-5

-3

-1

p<0.0001 p<0.0001

Absolute Change MRI-PDFF (% )

Week 12 36 12 36 12 36

-10

-36-42

-8

-37-49

-50

-40

-30

-20

-10

0

p<0.0001 p<0.0001

Relative Change MRI-PDFF (% )

Week 12 36 12 36 12 36

Week 36: Sustained Reduction in Liver Fat on MRI-PDFF

6

Relative Fat Reduction (%)

≥30% Fat Reduction (%)

Main, 36 Week Study

n Sustained statistically significant reduction in hepatic fat Week 12 to Week 36

n Placebo response generally related to weight loss ≥5%

P value, placebo compared to MGL-3196; MGL-3196, n=78; placebo, n=38; prespecified high exposure (High Exp) n=44; F2/F3, placebo, n=19; MGL-3196, n=33

PlaceboMGL-3196 (all)MGL-3196 (high exp)

Absolute Fat Reduction (%)

Extension Study of 36 Week Phase 2 Trial

7

-21 -23 -29 -31 -21 -33

120

-40

-30

-20

-10

0

10

20

30

40

ApoB LDL-C TGs ALT AST GGT SHBG

Perc

ent C

hang

e

120

Extension Study

n The Extension study includes 14 former placebo patients with persistently mildly to markedly elevated liver enzymes from the Main 36 Week study, ~ two thirds F2/F3

n Noninvasive end points, onlyn To optimize exposure, all patients in the

Extension study received 80 or 100 mg per day of MGL-3196, a higher average dose than in the 36 Week study to move all patients into the “high exposure” category

n Highly significant reduction in lipids including LDL-C, ApoB and triglycerides

n Well tolerated, few AEs, improvement in liver enzymes from baseline

0.6

-10

-14

-12

-10

-8

-6

-4

-2

0

2

7

87

0

20

40

60

80

100-1.9

-50

-60

-50

-40

-30

-20

-10

0

Extension Study: Reduction in Liver Fat on MRI-PDFF

8

≥30% Fat Reduction (%)

Relative Fat Reduction (%)

Absolute Fat Reduction (%)

Main Extension

-22.3 -21.9

-36 -36.8 -36.5

-50

-40

-30

-20

-10

0

LDL-C(BL>100)

ApoB TGs Lp(a)(BL>=10)

ApoCIII

p<0.0001 p<0.0001 p<0.0001 p<0.001 p<0.0001

Week 36: Sustained Robust Lipid Lowering

Significant sustained lowering effect in multiple atherogenic lipids

9

Lipids (% Change from Baseline)

Placebo corrected; p value, placebo compared to MGL-3196; MGL-3196, n=79; placebo, n=39

-20

-10

0

10

20

0 12 36 38

% C

hang

e

Week

0

10

20

30

40

50

60

0 12 36 38

U/L

Week

01020304050607080

0 12 36 38

U/L

Week

Week 36: Liver Enzymes

ALT

AST

GGT

Statistically significant reductions in ALT, AST and GGT versus placebo; no change in bilirubin or alkaline phosphatase

(10.0)

0.0

10.0

20.0

30.0

0 12 24 36 38

U/L

WeekPlacebo MGL-3196

10Baseline elevated ALT =45 male, 30 female. GGT

shown as % change from baseline, females and males have different normal GGT ranges

n Week 36, 40% reduction in ALT in patients with elevated baseline (p=0.01), and all MGL-3196 relative to placebo patients (p=0.002)

n At Week 36, 60% of MGL-3196 patients with ALT <30 vs 37% of placebo (p=0.03)

n Week 36, statistically significant AST reduction in MGL-3196 vs placebo (% change and absolute change) p=0.002

n Week 36, statistically significant GGT reduction MGL-3196 vs placebo (% change and absolute change) p=0.002

-0.40-0.48

-0.62 -0.63

-1.00

-0.80

-0.60

-0.40

-0.20

0.00

12 36 36 36

p=0.009 p=0.02 p=0.007 p=0.05

-59

-171-230

-363

-600

-500

-400

-300

-200

-100

0

12 36 36 36

NS p=0.003 p=0.0004 p=0.02

NA -25% -33% -38%

-2.7 -7.7 -8.6

-21.4

-32.1

-50.0

-40.0

-30.0

-20.0

-10.0

0.0

12 36 12 36 36

p=0.08 p=0.007 p=0.02 p=0.003 p=0.01

-18% -38% -41% -88% -119%

Baseline >= 17.5 ng/ml

Week 36: Reduction of Fibrosis, Biomarkers

11

ELF, CK-18 and Pro-C3 scores, biomarkers correlated with liver fibrosis stage, were statistically significantly

reduced in MGL-3196 treated, especially in patients with advanced fibrosis at baseline

BL, baseline; compared with placebo; all, placebo n=38; MGL-3196 n=78;

ELF≥9 placebo n=21; MGL-3196 n=40; Pro-C3 BL≥17.5, placebo n=12;

MGL-3196 n=29

*Liver Int. 2015 Feb;35(2):429-37; Journal of

Hepatology 2013 vol. 59 j 236–242;

MGL-3196 (all) MGL-3196 (high exp) F2/F3

ELF BL≥9 CK-18 (M30) U/L Pro-C3 (ng/ml)

week

% change

120

32

47

0

10

20

30

40

50

All F3

p=0.03 p=0.05

% b

iops

ies

Placebo MGL-3196

Week 36: Change in Fibrosis Score on Liver Biopsy

12

n Second Harmonic Generation (SHG) microscopy provides automated fully quantitative

assessment of fibrosis on liver biopsy slides based on unique architectural features of collagen

n SHG score was generated and aligned with the pathologist baseline score (baseline, r=0.76),

(left panel), blinded to treatment code

n Using SHG, MGL-3196 treated compared with placebo showed a statistically significant ≥1-pt

reduction in fibrosis score at Week 36. Based on pathology score, fibrosis was reduced by ≥ 1

point in 29% of MGL-3196 treated patients vs. 23% in placebo

≥1 pt reduction in fibrosis on liver biopsy (SHG)

Pathologist Score

SHG

(qfib

rosi

s)

https://doi.org/10.1371/journal.pone.0199166

Week 36 pathology scores and treatment code were not provided to SHG readers.

3

21

0

SHG Score

0 1A 1B 2 30.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

3251

65 61

18

47

0

10

20

30

40

50

60

70

p=0.09 p=0.006 p=0.02 p=0.02

% o

f bio

psie

s <5% Weight loss

Week 36: NASH Liver Biopsy Endpoints

13

2-pt reduction in NAS in placebo patients was correlated with body weight loss

MRI Responder; ≥ 30% fat reduction on Week 12 MRI-PDFF High Exp,, n=44; 2-pt NAS reduction; MGL-3196, n=73, placebo n=34; NASH Resolution, prespecified endpoint: at least 2-pt reduction in NAS; ballooning=0, inflammation=0, 1, <9.5% weight loss;

PlaceboMGL-3196 (all)MGL-3196 (high exp)MGL-3196, MRI responder

In MGL-3196 treated patients with NASH resolution, 50% had fibrosis resolution (F=0)

2-Point NAS Reduction

with at least a 1-pt reduction in

ballooning or inflammation

(% of liver biopsies)

NASH Resolutionballooning=0,

inflammation =0, 1 with at least 2-point

reduction in NAS(% of liver biopsies)

6

2739

6

2537

05

1015202530354045

p=0.02 p=0.001 p=0.03 p=0.003

% o

f bio

psie

s

no fibrosis worsening

Diet and Exercise Counseling in NASH trials

� Phase 2 trial (MGL-3196-05) employed mandated diet and exercise counseling at screening and each visit, recorded in the clinical trial record

� Previous large NASH trials did not have specific diet and exercise counseling at each study visit, and the effect of diet and exercise on NASH endpoints was only known from published literature on NASH endpoints in weight loss trials e.g. Vilar-Gomez Gastroenterology 2015;149:367–378

� Because extensive literature showed that body weight loss reduces and resolves NASH, weight loss subgroups and an exclusion for extreme weight loss were prespecified in MGL-3196-05 statistical analysis plan

— In Vilar-Gomez, extreme weight loss (10%) and Harrison et al (HEPATOLOGY, Vol. 49, No. 1, 2009) >=9% weight loss in NASH patients was correlated with marked reduction in steatosis, ballooning, inflammation on biopsy, resolution of NASH in ~90%

— In Villar Gomez, lesser degrees of weight loss >=5% were associated with NASH improvement— MGL-3196-05 prespecified in the statistical analysis plan, evaluation of subgroups with >=5% weight

loss on NASH and MRI-PDFF endpoints— MGL-3196-05 prespecified in the statistical analysis plan, in the secondary liver biopsy endpoint,

elimination of patients with ~10% (>9.5%) weight loss to avoid confounding effects of weight loss on the results to help with powering the Phase 3

� Clarification: Patients with extreme weight loss (>9.5%) was an exclusion in the Phase 2 study for NASH resolution, not an endpoint. Patients with weight loss were included in all other analyses of the Phase 2 data, and the Phase 3 endpoint has no exclusion for weight loss

14

Weight loss in MGL-3196 Phase 2 Study

� There were no differences in average weight loss between MGL-3196 and placebo patients

� All 3 placebo patients with extreme weight loss (>9.5%) and all placebo patients with NASH resolution (6%) came from a single site, resulting in a statistically relevant outlier site that permits removal of the site using standard clinical trial criteria

— Removal of the single site including all MGL-3196 and placebo patients showed that MGL-3196 compared to placebo maintained statistical significance for NASH resolution with no worsening of fibrosis (p=0.019), all patients; and p=0.001, MRI-responder

— Inclusion of the site and the 3 patients with weight loss in the NASH resolution post-hoc analysis resulted in p=0.15 (all MGL-3196 patients); p=0.014, MRI-Responders

� Since the 3 patients with >9.5% weight loss were placebo patients, in a posthoc analysis, patients with >7% weight loss were also excluded (7% results in NR of ~65%), thus removing a similar number of placebo and 3196 patients (4 each), with no change in the results: Compared with placebo, NASH resolution in all MGL-3196 patients, p=0.03-0.05; MRI-PDFF responders p=0.002-0.003

� Given the number of sites in Phase 3, unbalanced weight loss or site effects are a non-issue and no patients will be eliminated from analysis for weight loss. Sites will be automatically notified during the study if a patients exceeds 9.5% weight loss or gain so that investigators can review patient behavior and counseling

15

NASH Phase 3 Endpoints

� No NASH biopsy surrogate endpoint is currently validated as “reasonably likely to be

predictive of clinical benefit”, the ultimate arbiter of provisional regulatory approval

— Clinical benefit based on a NASH endpoint had not been shown

— Each company negotiates their endpoint and trial design with FDA

— The definition of NASH resolution (NR) ballooning 0, inflammation 0,1 without

worsening of fibrosis adopted for Phase 3 by some companies does not require a

reduction in NAS. A 2-pt reduction in NAS would, nonetheless, be observed in most NR

responders. Unfortunately, we believe this endpoint was not carefully considered and

declares “NASH resolution” in patients who are not actually NASH responders, thereby

introducing a lot of noise

— Madrigal definition, which was prespecified in Phase 2, incorporates the same NASH

Resolution definition ballooning 0, inflammation 0,1 without worsening of fibrosis with a

more stringent requirement for a 2-pt NAS reduction

� We believe our NASH resolution surrogate endpoint is reasonably likely to predict clinical

benefit, and FDA agreed (recent receipt of written agreement)

— “NASH resolution (ballooning 0, inflammation 0,1) without worsening of fibrosis and at

least a 2-pt reduction in NAS” is the endpoint for registration

16

Ballooning Reduction Alone Does Not Predict NASH Benefit

17

LiverFat

NAS-4Steatosis 2Inflammation 1Ballooning 1

Ballooned hepatocyte

Inflammation

LiverFat

NAS 3-4Steatosis 2 or 3Inflammation 1Ballooning 0

A true NASH resolution response in patients with NAS=4 at baseline:

LiverFat

NAS-4Steatosis 2Inflammation 1Ballooning 1

ALT, biomarkers improve

No improvement in ALT or biomarkers

LiverFat

NAS 1-2Steatosis 0-1Inflammation 1Ballooning 0

Literature examples: the NASH response in placebo patients (weight loss, bariatric surgery), FLINT, Pivens, Belfort, LEAN

In placebo patients with NR the average residual NAS is <=2, and average decrease in NAS>2

Independent of a weight loss or steatosis reduction, a ballooning reduction does not predict a reduction in NASH components, such as inflammation, or ALT, which has been shown to correlate with histologic response (Loomba et al Gastroenterology Sept 2018)

If a 1- point ballooning reduction is the only change leading to “NASH resolution” there is no true benefit (allowed in the relaxed NR definition); Ballooning =1, is a “few ballooned cells” ballooning is highly variable and subject to noise:

The less stringent NR definition is only problematic in patients with ballooning=1 inflammation=1 at baseline

Importance of Steatosis Reduction in NASH Resolution

� A growing data set indicates that reduction of hepatic fat in NASH patients leads to histologic response, including NASH resolution

� Not considering steatosis dismisses the critical role of lipotoxicity in driving NASH fibrosis. Patients with NASH do not have “simple” steatosis, they have inflammatory fat

— Based on substantial published data in placebo and drug-treated patients, converting inflammatory fat into noninflammatory fat is not likely. Unless liver fat is reduced via decreased fat production and/or increased breakdown of fat, patients likely still have NASH

� Evidence in support of the importance of steatosis response in predicting NASH outcome— Improvement in steatohepatitis in clinical trials and longitudinal studies is generally associated

with a reduction in liver fat (Arun Sanyal, Stephen Harrison)— FLINT, LEAN, Belfort, Pivens, placebo patients with weight loss (Vilar-Gomez)

— Steatosis on biopsy is too crude a measure to predict improvement in other components of NASH. MRI-PDFF is a sensitive measure of hepatic fat that correlates with steatosis score and predicts an NAS response and NASH Resolution

— Reduction of >=30% liver fat is reasonably called a PDFF responder— Reduction of liver fat on MRI-PDFF predicts improvement in ballooning and

inflammation (NGM, MGL-3196), not only in drug-treated but in placebo patients with weight loss

18

-100-80

-60-40-20

02040

60

-3 -2 -1 0 1 2% C

hang

e in

MR

I-PD

FF (W

eek

12)

Change in Ballooning Plus Inflammation

MGL-3196

37

405

10152025303540

p=0.001

MRI-PDFFResponder

MRI-PDFFNon-

Responder

MGL-3196-treated

NASH Resolution (%)

Correlation of Decrease in Hepatic Fat (MRI-PDFF) with Improvement in Ballooning and Inflammation on Liver Biopsy

19

n Patients who were not MRI-PDFF Responders (≥30% fat reduction) had a low rate of NASH

resolution (left panel)n In both MGL-3196 (correlation coefficient 0.42) (right panel) and placebo (correlation

coefficient 0.58) % relative change in MRI-PDFF was correlated with reduction in ballooning

plus inflammation scores on liver biopsy (steatosis score removed)

NASH Resolution (%) MGL-3196-treated

MRI-PDFF Week 12, % Relative Change: Correlation with Change in

Ballooning Plus Inflammation Scores