The following is a list of trial data presentations from this year’s American Diabetes Association (ADA) Scientific Sessions, which took place in Washington, DC on June 9-13, 2006. This list is not comprehensive; it includes only the events that we at BioMedTracker consider to be particularly relevant. To access other ADA events not included in this report, please click here and search using the keyword "ADA". Please note that we are still in the process of adding additional abstracts. 2-year Byetta Data In this open-label extension study, Byetta led to good improvements in glycemic control out to two years. Though the number was similar to that reported last year for 82 weeks of treatment in the overall group (-1.1 change in A1C, a measure of long-term glucose control), on a graph shown in an Amylin presentation, it appeared there was a slight deterioration from 82 weeks in this cohort followed for 2 years. In the presentation, 3 year data was also shown for a subgroup of patients. While good, the -1.0 reduction in A1C at 3 years was also slightly reduced from that seen at 2 years. More information is needed on the percent of dropouts and how they were handled. Otherwise, it is not entirely clear what one can expect long-term for the entire starting group (if measures for dropouts are on the commonly used last-observation carried forward, meaning the last measure before the patients dropped out, it would not include further deterioration in those dropout patients). Encouragingly, the data presented also showed a continued gradual mean weight loss out to 3 years.

In sum, while the data show good improvements in A1C sustained in up to 3 years, more information on dropouts is needed, and there is some evidence for a slight deterioration in A1C over time. It will be interesting to see three year data from a larger group, as well as longer term data, to get a better understanding of how the benefit of Byetta therapy holds up over time.

Byetta as a Substitute for Insulin – Pilot Study

This pilot study suggests that it is feasible to substitute Byetta for insulin in some patients with type 2 diabetes. While encouraging, more information is needed on A1C levels, as it is unclear how well controlled these patients were. One interesting use of off-label use of Byetta, mentioned in a company conference, was addition of Byetta instead of mealtime insulin for patients on basal insulin. This could be of benefit due to lower rates of hypoglycemia, and we think this is a promising area for potential expansion in the use of Byetta. Our revenue model for Byetta in this subpopulation is under review.

Byetta in Combination with Thiazolidenediones

Phase II Final Data Top-line data had been released previously. Since the trial is small, results are preliminary. The study found a large reduction in A1C for Exenatide LAR, with placebo subtracted improvements in A1C (a marker of long-term glucose control) of 1.8 for the 0.8 mg dose and 2.1 for the 2.0 mg dose (p values were not provided). The 2.0 mg dose also led to a good reduction in weight, while the 0.8 mg dose did not. A larger study is obviously needed, but given the good reduction in A1C in both treatment groups, we are encouraged by the results. It is also reassuring that much of the change came from a reduction in baseline for the treatment groups and not an increase in A1C in the placebo group. More detail on whether the nausea diminishes, as with Byetta, is needed. While data also released at the conference on once-daily liraglutide also showed very good efficacy, given that Exenatide LAR is given once-weekly, we think it will be more popular. Our current projections presume that LAR’s efficacy is at least as good as Byetta’s and close to (if not better than) liraglutide’s. (More data on liraglutide may also be presented at the late breaking session on Tuesday.)

Another question is the fate of exenatide and exenatide LAR if the DPP-IV inhibitors are approved. The DPP-IV inhibitors act along the same pathway, inhibiting the degradation of glucagon-like peptide (GLP-1), whereas exenatide mimics the action of GLP-1. The DPP-IV inhibitors are oral agents rather than subcutaneous, but do not generally lead to weight loss. For patients who use a DPP-IV inhibitor earlier on but start to have deterioration in their blood glucose, there may not be much benefit with exenatide, though the added weight loss could lead to a benefit that would lead some patients to switch, rather than continue the DPP-IV inhibitor. We also see a number of patients preferring a weekly LAR injection, especially if the reduction in A1C holds up in larger trials. Finally, despite the increased convenience of the DPP-IV inhibitors as oral agents, we think there will still be a number of patients and physicians who choose exenatide for its weight loss benefits. That is, we believe there is room in the market for both to be successful. Our revenue model for these agents is currently under review.

Phase IIb 14 Week Trial In this 14-week trial of type 2 diabetics treated with diet or a single oral agent, the once-daily GLP-1 analogue liraglutide led to a marked reduction in A1C levels (a marker of long term glucose control) relative to placebo. The change in A1C for the high dose group is numerically better than for twice-daily Byetta, though the data available for Byetta are in patients failing other medications. While they are not as large as seen in a study with weekly exenatide LAR, that study was quite small, and results could be different in a larger trial. Complicating interpretation, however, is the fact that this liraglutide study involved a 4 week washout period for patients on prior oral agents. While this is commonly done, A1C levels may not have stabilized, and the "true" baseline could have been higher (about 50% of the A1C represents blood glucose levels more than 30 days in the past). Since higher baseline A1C levels usually lead to larger treatment differences, this could have magnified the results somewhat, though we still think they would have been relatively good. Liraglutide also showed a good trend in lowering weight.

In another abstract from the trial, liraglutide led to significant reductions in blood pressure and triglycerides, though LDL levels were higher for liraglutide than placebo. It was interesting that nausea rates were lower in this study than those reported for Byetta, though more detail is needed, particularly for the different dose groups (and Byetta nausea rates decrease with time).

Overall, these preliminary results are encouraging for liraglutide, and we are raising our likelihood of approval by 4%. Even if the drug has similar efficacy to Byetta, its once daily dosing will give it an advantage. However, we believe that once weekly exenatide LAR will do better if its results hold up in larger studies. While both long-acting drugs have recently started Phase III development, exenatide LAR will likely be able to be approved based on comparative data with Byetta, which should expedite the process. Liraglutide could still be useful in patients who do not tolerate exenatide LAR. We are in the process of adding a revenue model for the drug. (Please see our discussion on exenatide LAR for additional thoughts on these agents if the DPP-IV inhibitors are approved.)

Continuous Infusion Study This was a small study looking at the continuous infusion of BIM51077 for 28 days. Another study, looking at infusion over 7 days, was also presented at the conference. BIM 51077 is a glucagon-like-peptide-1 (GLP-1) analogue. We would presume that the drug is being studied to develop a long-acting formulation to compete with exenatide LAR, and these are the initial pharmacodynamic studies. In this study, a 400 ug/day infusion led to a placebo-subtracted reduction in the glucose area under the curve (AUC) of about 777 mg-hr/dl (the area under a curve plotting glucose concentrations over time). In the seven day study, the top dose (800 ug/day) lead to a placebo subtracted reduction of about 937 mg-hr/dl in the AUC. While difficult to compare due to the study design and short duration, given these and the other values reported, we think the results generally look good. For comparison, in an infusion study of a different design, 0.2 ug/kg/day of Byetta resulted in a placebo-subtracted AUC of 901 mg-hr/dl, though another study found a much lower result. While data on an injectable formulation are needed, given the current preliminary results we are raising our likelihood of approval by 3%.

Please see below for a summary of Januvia and Galvus.

Late Breaking Session: Comparison with Glipizide

18-Week Monotherapy Study (with Prior Drug Washout)

24-Week Monotherapy Study (with Prior Drug Washout)

24-Week Add On to Pioglitazone Study

24-Week Add On to Metformin Study

Monotherapy in Japanese Patients

Safety Data in Patients with Renal Impairment

While there appeared to be more cardiovascular deaths in the Januvia group, there were apparently a disproportionate number of patients in the group with pre-existing coronary artery disease or heart failure. There has not been a reported cardiovascular safety signal in patients without renal failure. However, the FDA could want clearer data in this population, given that dose adjustments are needed.

Galvus, Actos, and Combination

Versus Avandia

Combination with Metformin

24 Week Monotherapy in Treatment Naive

52 Week versus Metformin

Galvus Add on to Insulin

There were a number of studies for the DPP-IV inhibitors Januvia and Galvus, presented at the conference. These agents inhibit the enzyme that breaks down GLP-1 (glucagon-like peptide-1), in contrast to drugs such as Byetta which are GLP-1 analogues.

The most striking data presented for Januvia was a comparison study with the sulfonylurea glipizide in patients failing metformin. (Enrolled patients apparently had whatever treatment they were on stopped and then were put on metformin. Those who failed were reportedly eligible for the trial.) The A1C changes (a measure of long term control) for Januvia were similar to glipizide. This was notable, as sulfonylureas are quite effective oral agents, though they tend to be more effective in patients with insufficient insulin production rather than higher insulin resistance. While the overall change in A1C for both drugs was moderate, this could reflect the relatively low starting A1C at baseline, and encouragingly efficacy proved similar across subgroups with varying baseline A1C levels. The glipizide used in the study was apparently the regular (not the long-acting) formulation, titrated up to 10 mg twice daily. This is a common dose, and while it is not the maximum dose, patients who need more treatment frequently receive an additional drug instead. However, it is unclear what the average final dose of glipizide was. Also, in another 12 week study, Januvia led to a reduction in A1C that was 0.2 smaller than glipizide.

It is interesting that in this study, Januvia led to a slight weight loss (glipizide led to a weight gain). However, there was no placebo comparator, and the drug has not shown significant weight loss in other studies. Presenting physicians at a Merck conference at the sessions downplayed the weight loss, and noted that the main advantage was no weight gain relative to certain other therapies (Galvus similarly does not lead to weight gain). As expected, Januvia also led to a significantly lower rate of hypoglycemia than glipizide. The glipizide study was particularly notable, since most of the data we have seen for Januvia and Galvus has suggested they are not particularly effective compared to existing options. For example, a 54-week study presented at the conference showed Galvus 50 mg twice daily inferior to metformin 1000 mg twice daily. Data on Galvus showing equivalence to the thiazolidinedione Avandia was presented, though Avandia had a numerically better result at the endpoint and it is possible that a greater difference could develop with time (in the aforementioned metformin study, the two appeared to have closer results at 24 weeks, as well). In another study of Galvus, presented in the late breaking session, Galvus led to a 0.3 numerically poorer reduction in A1C than the other thiazolidinedione, Actos (30 mg), though a similar percentage of patients reached A1C goals of 7% or less for both.

In any case, even if Galvus and Januvia turn out in practice to have efficacy disadvantages, from the information thus far, they are likely to have an advantage over current oral drugs in terms of lower side effects (though longer term results in greater numbers of patients are needed). As noted above, they also do not lead to weight gain, whereas the sulfonylureas and thiazolidinediones do. Hence we think they will do well in the market, likely in patients who fail metformin or front-line in some patients who cannot tolerate metformin.

The injectable drug Byetta is another potential competitor for the DPP-IV inhibitors. Of note, in a 24 week trial of Galvus as an add on to patients failing metformin, patients receiving 50 mg of Galvus twice daily (bid) achieved a respectable 1.1 reduction in A1C (a measure of long term glucose control) compared to placebo, which is similar to what was seen for Byetta at 24 weeks. Byetta has the advantage over Galvus and Januvia of greater weight loss, which could in turn potentially lead to greater benefits in glucose control over time, but has the disadvantage of requiring an injection. Our revenue models for the drugs presume Byetta will be used in patients with a strong motivation for weight loss. (Please see our above comments on exenatide LAR for more on this.)

A major question is how Januvia and Galvus stack up against each other. Given differences in the trials, the results for the two are difficult to compare (for example, trials had patients with different baseline A1C levels, and the higher the baseline A1C the greater the reduction with treatment). In add on trials to metformin presented at the conference, Galvus 50 mg twice daily has shown numerically greater reductions in A1C than Januvia 100 mg daily, though the Galvus study patients started with a somewhat higher A1C.

This raises an additional question, though, particularly given all the Galvus 50 mg twice daily studies at the conference, whether Galvus is truly effective once-daily. In a conference call after the late breaking sessions, Novartis officials sought to dispel that question with data showing comparable efficacy between the 100 mg once daily dose and the 50 mg twice daily dose. Officials stressed that they had bridging studies to demonstrate equivalence, so efficacy for the 50 mg twice daily could be extrapolated to the 100 mg daily dose. Without seeing this data and the rest available for registration, it is difficult to assess whether this will be sufficient for the FDA. We also should note we have not seen studies for either drug in certain settings, such as in patients failing sulfonylureas. If the studies have been done, we question why those data in particular have not been presented, and if there are issues that could limit treatment settings for which the drugs are indicated. (Januvia also had some small changes in certain laboratory values. While described as not clinically meaningful, it is not clear if there were outliers that could be of concern.)

Overall, we think the battle between Galvus and Januvia will come down to marketing. Galvus may have the advantage of some numerically higher results in some studies. Also, it may have an advantage in some patients, such as those with renal insufficiency, as it does not appear to require dose adjustments whereas Januvia does. However, the Januvia - glipizide comparison could help tilt the marketing balance in favor of Januvia. In addition, the weight loss seen with Januvia in the trial could be a one time artifact, it also adds to the marketing. Overall, while we think both will do well, we currently think Januvia will come out slightly ahead. Galvus also has a sulfonylurea trial that is ongoing, which could change the balance (company officials would not reveal the timing of the results). We are currently in the process of revising our revenue models.

Phase II Proof-of-Concept Study This was a short (14 day) study of mealtime dosing with the DPP-IV inhibitor PSN9301. Obviously this drug, given two to three times a day, would be less convenient than once-a-day DPP-IV inhibitors currently in NDA status (Januvia and Galvus). OSIP holds patents to the DPP-IV family and should receive a small amount of royalties on these drugs. The company is developing PSN9301, because as a short-acting agent, it may have fewer side effects. While the DPP-IV inhibitors have had very good safety records thus far, it is difficult to fully assess their safety until they are used in a large number of patients for a longer period of time. In addition, the DPP-IV inhibitors are likely to be used in a combination pill with metformin, which would require twice daily dosing (though this would be one pill twice daily).

Top line results of the current study have been presented previously. Though it is difficult to compare with existing drugs due to study differences (and the popular drugs tend to affect baseline glucose more than post meal values), the percent reduction in glucose after an oral glucose tolerance test (OGTT, in which a dose of glucose is given), appear clinically meaningful, and could be very good if further studies demonstrate an effect size in the upper end of the range. However, we think the strategy for this drug really depends on side effect issues with the longer acting DPP-IV inhibitors, so we currently project a very limited market share.

Phase II Dose-Ranging Study CS-917 blocks the second to the last step in gluconeogenesis, the process of making glucose that partly contributes to elevated glucose levels in diabetes. A putative advantage of blocking this step is that it allows inhibition of gluconeogenesis from a variety of potential substrates. The popular drug metformin also inhibits gluconeogenesis, but only by 33-36%. In today's trial (top line results have been reported earlier), CS-917 led to significant reductions in fasting glucose over 0 to 6 hours after a dose, with the exception of the 100 mg dose. The results did not exhibit a clear dose response relationship. The 50 mg dose led to a reduction around 200 mg-hours/dL (since the endpoint was the area under the glucose time curve [AUC], the units are expressed in glucose concentration x hours). The 100 mg dose resulted in a roughly 100 mg-hr/dL decrease (which was not statistically significant). The 200 mg dose was similar to the 50 mg dose, and the 400 mg dose led to a slightly greater reduction. While these results are likely clinically meaningful, for comparison, they are less than what was seen in a fasting study of Byetta, where the placebo subtracted AUC over 6 hours was over 250 mg-hr/dl.

While development of CS-917 had been halted due to a serious side-effect called lactic acidosis in a Phase I metformin interaction study, Phase IIb testing was resumed earlier this year. Metformin is thought to potentially cause lactic acidosis, though quite rarely. It is somewhat troublesome that in the current study 2 patients developed lactic acid levels over 4.5 mmol/L, when levels over 5 mmol/L are part of the definition of lactic acidosis. While lower doses may provide an adequate margin of safety, without more safety data, we are still concerned that there could be more of a risk of lactic acidosis with CS-917 than with metformin. As a result, we are leaving our likelihood of approval unchanged.

Phase II Dose-Ranging Study with Pioglitazone Comparator Rivoglitazone is an agonist for the peroxisome proliferator-activated receptor (PPAR) gamma receptor, like approved drugs Actos (pioglitazone, the comparator in this trial) and Avandia. The results of this study showed good reduction in fasting plasma glucose, with changes for the higher doses better than a moderate dose of Actos. While rivoglitazone is different from the two high profile PPAR dual agonists (targeting both the gamma and alpha receptors) that were suspended from development this year, the approved PPAR gamma drugs also have significant side effects, such as edema. Detailed safety information was not released in the abstract, and it will be important to see that the good efficacy is not compromised by unacceptable side effects. Pending more detailed safety information, we are restricting our increase in likelihood of approval to 1%.

Updated Phase II Data ISIS 113715 is an antisense inhibitor of the gene encoding PTP-1B, an enzyme that appears to interfere with insulin’s ability to regulate blood sugar. While the placebo-subtracted fasting blood glucose with ISIS 113715 reported in this study was clinically meaningful, the value reported was for the 10 week time point, and in a graph from an accompanying conference call, subsequent values were smaller. Also, though the absolute reduction in A1C (a measure of long-term glucose control) was good, the placebo group also had a sizeable reduction, making the difference fairly small. In accompanying conference call, company officials noted that the change in A1C towards the end of the study was leveling for placebo, but that for ISIS 113715 was continuing to drop. It is possible that due to the long half-life of the drug, more time is needed. Given the mixed results, we are lowering our likelihood of approval but limiting that to -1%. We do think that further study is warranted and could change the outlook for the drug.

Amylin Obesity Development Review During a company review at the American Diabetes Association meeting, officials from Amylin discussed their strategy on obesity treatment. One of the highlights was information regarding leptin from a preclinical model. While leptin has been implicated in obesity, past clinical trials of leptin replacement have been disappointing. Amylin researchers found that the hormone amylin, of which Symlin is an analogue, restored leptin sensitivity in an animal model.

In another preclinical study, while leptin alone had little effect, added to amylin, leptin increased weight loss, and weight loss was increased further with the addition of the hormone PYY3-36, of which Amylin's AC162352 is an analogue. According to company officials, the percent of body weight lost with the triple combination was apparently greater than what had been seen with the popular fen-phen combination (withdrawn from the market due to safety issues), though not as great as that seen with an obesity surgery (Roux-en-Y procedure).

Amylin acquired the rights to the leptin molecular franchise and on-going clinical program earlier this year. Amylin is also investigating second generation analogues for amylin and PYY3-36, which have shown even greater efficacy in animal models. The company's strategy is thus to ultimately use the 3 molecule types in combination for improved weight loss.

Updated Phase IIb Results Top line data for this trial have been reported before. As noted in our Obesity Indication Report, while lorcaserin has relatively good efficacy, it is not notably better than currently available Meridia. However, we think lorcaserin may have a better side effect profile. Of note, in an accompanying conference call, company officials presented data showing that the headaches seen with lorcaserin occur more frequently initially, and are transitory, with frequency diminishing with time.

Since this drug has a related mechanism of action to fenfluramine, withdrawn from the market for heart valve problems, it is being closely monitored with echocardiograms. This has not been done before in such a trial, and of course raises the risk of finding a numerical disparity that could require more study, which could be difficult to do. In this trial, there were some numerical differences in individual measures between treatment and placebo. However, company officials said that cardiologists (who had experience from fen-phen studies) did not deem there to be any safety signal.

Finally, one unfavorable finding in the study was a decrease in HDL cholesterol in all doses of the drug. While this might turn out to be different in longer term studies, it could be an unfavorable favorable factor versus Acomplia, which is being targeted as a risk reduction agent.

Overall, while these are all factors that need to be followed, today's news does not significantly change our assessment of the drug

Combined Analysis of Two Trials This was a combined analysis of 2 trials (the original PKC-DRS only had 67 patients in the Arxxant 32 mg/day group). The trial found that compared to placebo, Arxxant led to improved vision on a couple measures, including sustained moderate visual loss (SMVL; loss of =15 letters, or 3 eye chart lines, sustained for the last 6 months on study). Arxxant also reduced the progression of edema that was not at the macula (the part of the back of the eye responsible for detailed central vision) from encroaching on the center of the macula. We would like to see more data on the second trial individually. We would also like to see more data on whether the effect occurred only in patients with retinal edema at baseline, since the drug appeared to only help these patients in the first study (the FDA may also want the data to better define the target population). Nevertheless, these are encouraging data, and they bode well for the ultimate approval of the drug. As a result we are raising our likelihood of approval by 4%. However, since the first trial did not lead to statistically significant changes on all measures, it is possible the FDA will want to see additional confirmatory data.