I am frequently asked if 505(b)(2) projects fail or whether any NDA submissions are rejected. My answer is that the vast majority succeeds and are eventually approved. Those that fail more often are due to money or design issues, not execution risks. Today I discuss an example of a failure based, in large part, on design.
On May 12, 2010 the Joint Meeting of the Arthritis Advisory Committee (AAC) and the Drug Safety and Risk Management Advisory Committee (DSaRM) voted 16-1 against recommending for approval NicOx’s naproxcinod for the treatment of the signs and symptoms of osteoarthritis. NicOx was hoping to show that naproxcinod, a pro-drug of naproxen, had a safer cardiovascular (reduced elevation in blood pressure) and gastrointestinal profile than naproxen with the same efficacy. Studies for GI tolerability and blood pressure failed to meet required endpoints and non-inferiority trials did not suggest any difference between the pro-drug or naproxen. Thus, NicOx S.A. ends up with an expensive “generic” to naproxen. A disclaimer: Camargo was not involved in this development and it is not certain that 505(b)(2) was actually used (but could have).
Per the FDA briefing document: Naproxcinod is a non-steroidal anti-inflammatory drug designated as 4-(nitrooxy) butyl (2S)-2-(6-methoxynaphehalene-2-yl) propanote and it has the following chemical structure:
Following absorption, this new molecular entity is metabolized to the NSAID naproxen and an organic nitrate butanediol mononitrate (BDMN) releasing moiety and subsequent downstream metabolites [gamma-hydroxybutyric acid (GHB) and butanediol (BD)]. BDMN is the moiety responsible for donating nitric oxide (NO).
Thus, the nitrate yields nitric oxide like the vasodilators nitroglycerin and isosorbide dinitrate. Per the Directors memorandum [hyperlink, page 5): “The purported benefit of the NO component of the drug is a reduction in NSAID-related gastropathy and a mitigation of the expected blood pressure elevation after NSAID initiation. According to the Applicant, NO is thought to stimulate GI protective factors that are negatively affected by NSAIDs, such as mucus production, secretion of bicarbonate and increased blood-flow in the gastric mucosa. Also according to the Applicant, the release of NO from naproxcinod is intended to counteract the rise in BP seen with NSAIDs in both normotensive and hypertensive patients, possibly via a decrease in Angiotensin II production and action through down-regulation of the synthesis of angiotensin converting enzyme (ACE) in the endothelium, as well as Angiotensin II type 1 receptors (AT1) in vascular smooth muscle cells.”
NicOx conducted 35 studies for the NDA submission:
- Twenty-six (26) Phase 1 studies consisting of single and ascending dose studies, bioavailability and bioequivalence studies, pharmacokinetics studies in healthy young subjects, elderly subjects and Japanese healthy subjects, a mass balance study, drug interaction studies, special population studies , QT study, ambulatory blood pressure monitoring studies and gastrointestinal studies.
- Five (5) Phase 2 studies consisting of dose—ranging studies, single-dose acute dental pain studies, and an endoscopy study.
- Four (4) Phase 3 studies consisting of three adequate and well-controlled efficacy and safety studies in patients with osteoarthritis of the hip or knee, and one-long term safety study.
In the 3 phase 3 studies, the FDA concluded that the non-inferiority of naproxcinod to naproxen was not demonstrated in replicated studies; in one study naproxcinod was shown inferior. The study design was a possible culprit. Generally, in a non-inferiority trial the sponsor needs to show that the difference in treatment effects are clinically insignificant. This means that that the difference should be a small percentage of the overall treatment effect. NicOx found a treatment effect size of 12mm in the WOMAC pain and function endpoints in their Phase 2 program yet used 8mm for the confidence intervals for the Phase 3 study — and still failed one study. An FDA review of the adverse events reported in these 3 studies demonstrated that naproxcinod and naproxen had the same profile expected of the NSAID class.
NicOx also conducted trials specifically looking at the effects of NO on lowering blood pressure. The reviewer in FDA Division of Cardiovascular and Renal Products concluded (page 50):
“Following treatment with naproxcinod, BP was not consistently less than baseline through the dosing interval. Although less of an effect was noted compared to naproxen in general, this was not sustained through the inter-dosing interval. More than two-fold changes in peak-trough effects were noted in some ABPM recordings.
- Based on information available to date a consistent effect on both systolic and diastolic BP is mainly present at peak.
- The effect through the dosing interval is variable.
- Typically drugs with meaningful effects on cardiac outcomes have an effect on systolic and diastolic BP that persists through the dosing interval. In this case the potential impact on cardiac outcomes is unclear.”
Could it be possible that there isn’t sufficient NO present at the required dose of naproxen?*
To show improved gastrointestinal safety, NicOx conducted three (3) endoscopy studies in (2) healthy subjects and (1) osteoarthritis patients. Note that just prior to the Advisory meeting NicOx had asked to withdraw the ulcer claim. The studies are presented in the following table:
The reviewer from FDA’s Division of Gastroenterology Products outlined the following study criteria used for approval (emphasis mine) of products (e.g., PPIs) for the reduction of NSAID-induced ulcers:
- defined an ulcer as a lesion â‰¥ 3 mm with depth as measured by endoscopic examination
- DGP has accepted Phase 3 study designs with proportion of patients who develop ulcers as a primary endpoint in these trials
- number of ulcers is not generally accepted as a primary endpoint
- DGP has not accepted the use of erosion as a primary endpoint
- in general, DGP recommends study durations of at least 6 months; The prescription proton pump inhibitors currently marketed for this indication, esomeprazole, lansoprazole, and the naproxen/lansoprazole combination product, were supported by studies of 3 months and 6 months duration
Thus, the reviewer noted that studies 0002 and 0027 did not use endpoints or duration consistent with studies used for approval of PPI’s. Study 0005 was reviewed with the caveat that it was only 6 weeks (not months) duration. The comparison of incidence of ulcers in this short study gave a point estimate (naproxcinod/naproxen) of 0.70 (95% CI, 0.48, 1.03) p=0.07. The results were not significant and the difference was judged to be not clinically meaningful.
I emphasized in the text above that the FDA used criteria that were used for PPI product approval. NicOx didn’t want treatment but just wanted labeling that showed show an improvement. Nevertheless, using NicOx’s own pre-defined study endpoints, the proposed product was demonstrated to be not any better than naproxen.
NicOx’s briefing document to the Advisory Committee tried to focus the review on the improved blood pressure profile of their product over naproxen. They conclude [page 16]:
“Naproxcinod represents an additional treatment option for physicians and OA patients, combining the proven efficacy and safety of naproxen with an NO-donating moiety that mitigates some of the known side effects of NSAIDs. Naproxcinod is less likely to increase BP than NSAIDs, which may be of particular importance to OA patients with pre-existing hypertension. The overall benefit/risk profile of naproxcinod 750 mg bid and 375 mg bid is positive and supports its use for the relief of the signs and symptoms of OA.” As I summarized above, FDA’s analysis of the BP data indicated that naproxcinod was not sufficient to label it as an antihypertensive. Ultimately the Advisory Committee agreed that the effect was not consistent.
Here is the broader issue for us in the 505(b)(2) world: we are developing improvements to existing products and in order for them to be a commercial success we must be able to promote the difference from what is often the generic RLD. In order to be able to legally promote a difference, the data must be on the label. What studies are sufficient to be included in the label? Do the data need to show clinically meaningful difference (an objective standard) or simply statistically different? Does the answer to the latter question depend on which section of the label we want to have modified? I’m uncomfortable leaving questions unanswered at the end of this posting, but the general answer is, it depends on each circumstance. Camargo meets with FDA several times per month on these issues. It is very important to document your case thoroughly and meet with FDA. Perhaps NicOx did, but then why were many study endpoints so far off from FDA’s usual standards?
* NicOx states:
“While the maximal total daily nitrate intake through naproxcinod 750 mg bid is 268 mg and is close to the WHO Acceptable Daily Intake (ADI) set for nitrate intake in food consumption (3.7 mg/kg or 222 mg for a 60-kg person or 339 mg for a 93-kg person, the average weight of patients in US pivotal studies), it is more than 4 times below the nitrate intake proposed by the high-nitrate Dietary Approach to Stop Hypertension (DASH) diet (i.e. 1,222 mg). Nitrate intake from naproxcinod corresponds to the same nitrate content found in 100 mL of fresh organic beetroot juice, or in 50 g of rocket salad or 100 g of spinach, lettuce, rhubarb or cabbage.”