Yes, the proposed stopping rules are too conservative. John R. Senior, M.D. 25 March 2010 Rethink Stopping Rules 1

Transcription

1 Yes, the proposed stopping rules are too conservative John R. Senior, M.D. 25 March 2010 Rethink Stopping Rules 1

2 Guidelines for study stop rules: ALT/AST > 8X ULN ALT/AST remains > 5X ULN over 2 wks ALT/AST > 3X ULN & Total Bili > 2X ULN or INR >1.5 ALT/AST > 3X ULN with symptoms (e.g. fatigue, N&V, RUQ pain, fever, rash) or eosinophilia 25 March 2010 Rethink Stopping Rules 2

3 WHY Too Conservative? Serum transaminase activities do not predict what will happen, only indicate what has already happened; Highest observed levels often or usually not true peaks; Even peak levels not indicative of functional loss; Serum transaminases not measures of function at all; Not specific to liver injury; caution! R/O other sources. Most transaminase elevations are transient, resolve 25 March 2010 Rethink Stopping Rules 3

4 Serum ALT, xuln Transaminase "Peak" Days on Drug depends on w hat days you draw samples Day xuln March 2010 Rethink Stopping Rules 4

5 What is the normal range of serum transaminase activities? should each local lab determine? should overweights be excluded? why are men higher than women? how much comes from muscle? how important is analysis method? can all this be standardizied? is xuln better than baseline change? 25 March 2010 Rethink Stopping Rules 5

6 What are we measuring as serum transaminase activities? the rate of conversion of an amino acid to a keto acid, in micromoles/min: e.g., alanine to pyruvate H3C CH(NH2) COOH CH3 CO - COOH or, aspartate to oxaloacetate HOOC CH2 CH(NH2) COOH HOOC - CH2 CO COOH with balanced exchange from AKG to glutamate, or balanced reverse reactions 25 March 2010 Rethink Stopping Rules 6

7 J Clin Invest Jan;34(1): PMID APPENDIX A NOTE ON THE SPECTROPHOTOMETRIC ASSAY OF GLUTAMIC- OXALACETIC TRANSAMINSE IN HUMAN BLOOD SERUM By ARTHUR KARMEN (Department of Pharmacology, New York University College of Medicine) 25 March 2010 Rethink Stopping Rules 7

8 HOOC CH 3 HOOC O O COOH oxalacetate pyruvate malate dehydrogenase DPNH (NADH), H + lactate dehydrogenase DPNH (NADH), H + HOOC CH 3 HOOC O H O H COOH malate lactate DPN (NAD) DPN (NAD) 25 March 2010 Rethink Stopping Rules 8

9 HO H 3 C N OH pyridoxine vitamin B 6 aspartate HOOC OH HOOC HO H 3 C N H 2 N O N H 2 e-palpo alanine CH 3 EC EC COOH N CH NH HC 2 N HO OPO 3 H OPO 3 H 2 2 H 3 C alpha-ketoglutarate O N e-pampo COOH COOH HOOC CH 3 COOH HOOC O O COOH H 2 N COOH oxalacetate pyruvate glutamate 25 March 2010 Rethink Stopping Rules 9

10 AST, TBL: log10(xuln) One Year on INH: 3 "Hy's Cases" Mitchell, et al., 1975 M49w M49w 30.7x 22.1x M61b M61b 14.6x M39b M39b red =AST; green = TBL 4.3x 3.3x 2.8x on INH weeks March 2010 Rethink Stopping Rules 10

11 edish hyperbilirubinemia Hy's Law range Drug X Peak TBL, xulrr x Drug C normal range 3x Temple's Corollary range Peak ALT, xulrr 25 March 2010 Rethink Stopping Rules 11

12 Treatment X ID: 7259 Time Course of Test Values male 80, caucasian X start stop ALTx dead ASTx Test Values, log10(xuln ALPx TBLx bx admit readmit Days Since X Started March 2010 Rethink Stopping Rules 12

13 Levels of DILI Severity Death/Tx Acute Liver Failure Serious: Disabled, Hospitalized Hy s Case: Detectable Slight Functional Loss Serum Enzyme Elevations Only; Many People Adapt 0 Most People Tolerate Exposure - No Adverse Effects Seen 25 March 2010 Rethink Stopping Rules 13

14 Categories of DILI Likelihood 5 definite, nearly certain (estimated range >95%) 4 very likely (estimated range >75 to 95%) 3 probable (estimated range >50 to 75%) 2 possible (estimated range >25 to 50%) 1 unlikely (estimated range 5 to 25%) 0 very unlikely (estimated range <5%) percentage ranges do not imply exactness, but may be helpful as adjectives to get more consistency between evaluators FDA categories consistent with NIH DILIN scale, but inverted) 25 March 2010 Rethink Stopping Rules 14

15 Clinical Importance of Individual Cases of Putative DILI DILI Likelihood 5: definite >95% likely 4: very likely >75 to 95% likely 3: probable >50 to 75% likely 2: possible >25 to 50% likely 1: unlikely 5 to 25% likely* 0: certainly not <5% likely** 5 fatal or transplant 4 acute liver failure Severity of Liver Injury 3 serious: 2 Hy s case 1 enzyme rises only 0 none detectable * and some other cause at least probable ; ** and another cause at least very likely 25 March 2010 Rethink Stopping Rules 15

16 In controlled clinical trials, there may be limitations to the numbers of subjects, the frequency of blood sampling, and duration of the period of observation. Transient peaks are often missed, but may not be very important. Sustained and increasing evidence of injury is worrisome, but should not be over-interpreted. Loss of true liver function is the real concern. Minor serum enzyme elevations are often reversible, whether drug is stopped (interrupted) or continued, by hepatic adaptation to xenobiotic substances. Must make effort to rule out other causes of hepatocellular injury from viral infections, fatty liver/nash, alcoholic sensitivity, cardiopulmonary problems (hypotension/shock, congestive failure, hypoxia), biliary tract diseases, autoimmune hepatitis. 25 March 2010 Rethink Stopping Rules 16

17 The normal liver is a very robust organ, and can sustain quite considerable injury. Even 70% resection is very quickly compensated for by rapid growth of hepatocytes, within a week or two in the rat, longer in humans. In addition to rapid regeneration, the liver may adapt by changing gene expression the up or down-regulate fects and become tolerant, in most individuals. Simple elevation of serum transaminases is usually reversible, causes no symptoms or illness. Only if sustained, progressive, and accompanied by functional loss (e.g., bilirubin, prothrombin) is it more alarming. We should not IGNORE elevated transaminases, but very promptly follow their course of change over DAYS (not weeks or months) to assess what s really happening. If that cannot be done at the study site, then the more conservative position is appropriate. If we do this, we may be able to avoid trashing good drug candidates but still not endanger subjects or patients. 25 March 2010 Rethink Stopping Rules 17

18 You might as well fall flat on your face as to lean too far over backward. James G. Thurber, Fables for Our Times, 1940 The Bear Who Let It Alone, New Yorker,19 April March 2010 Rethink Stopping Rules 18

19 Primum non nocere. first, not to harm Ante omnia iuvare amplius quam nocere. above all, help more than harm March June Society Rethink of Toxicologic Stopping Pathology Rules 19 19