Oncology Review Division At The Head Of The Pack – But Can The Others Keep Up?

Analysis of drug review performance by division finds Cardiovascular & Renal Drugs Division to occupy the lower end of the spectrum. Staffing appears to be a vital component of review performance, giving weight to the researchers’ warning of impending wholesale staff turnover.

A survey of performance across the Center for Drug Evaluation and Research’s review divisions shows that the Office of Hematology and Oncology Products is the “exemplary” group, with a high rate of approvals with a low staffing level and high number of submissions – but the unique characteristics of the cancer setting may mean that the oncology review group is more of an outlier than a model.

The study, titled “FDA Review Divisions: Performance Levels and the Impact on Drug Sponsors,” aims to “shed new light on the regulatory review process and address sponsors’ perceptions of disparities and inconsistencies among the review divisions,” Tufts Center for the Study of Drug Development Director Kenneth Kaitin and Associate Director Christopher Milne explain in the March 2012 issue of Clinical Pharmacology and Therapeutics. Criticism of the various FDA review divisions has an extensive history.

“The exemplary review division is Oncology,” they state. The oncology division addressed a “fairly high” workload of IND and NDA submissions with one of the smallest divisional staffing levels, yet produced a higher share of NDA approvals than any other division, with the shortest average approval time, Milne and Kaitin found. Oncology also produced high rates of first-cycle approvals and met most PDUFA goals.

A Tale of Two Divisions: Oncology, “The Exemplary Review Division”

• Accounted for 26% of 2006-2010 NDA approvals; “fairly high” number of IND and NDA submissions
• Low staffing level (31 for Drug Oncology in 2011)
• High proportion of priority reviews (68%)
• Low use of advisory committee meetings (convened for 25% of products eventually approved)
• “Complete response” letters issued for only 11% of products approved
• Shortest NME approval times (10.5 months on average)

Source: “FDA Review Divisions: Performance Levels and the Impact on Drug Sponsors,” Milne & Kaitin, Clinical Pharmacology & Therapeutics, March 2012

But the Tufts CSDD researchers’ finding is not necessarily surprising: oncology also features some of the clearest risk/benefit calculations, given the severity of the illness – especially in the advanced settings where novel drugs are typically first positioned. Trials are often smaller and shorter than in fields like cardiovascular disease or diabetes, where comparatively healthy patients will take a drug chronically; oncology NDAs therefore often feature fewer case report forms and data points to evaluate.

“There are obviously a number of product characteristics such as therapeutic area, unmet medical need, and primary vs. specialty use that not only affect the nature and amount of data required of sponsors on the one hand but also impact the workload of the division,” the authors acknowledge. “We could not adjust for all these factors,” many of which were “’beyond reach’ in terms of the practical limits of data availability and assessment.”

The authors tried to “purposely limit the assessment so that differences among divisions could be detected without distraction from too many variables pointing in too many directions.” To that end, the study focused on eight review divisions characterized by high “output factors,” like a large percentage of total NDAs, high “input factors,” such as high workload, or approval times “that varied substantially from the mean” during the 2006-2010 period.

The authors ranked, roughly, six selected divisions in terms of review performance. After oncology, the middle ground is populated by Anti-Infectives (the division “tended toward the optimal side of the spectrum”) and Anesthesia, Analgesia & Addiction (“midrange, but reflecting the optimal side”). The “suboptimal region of the middle ground” is held by Neurology (its performance was “a mirror image” of Anti-Infectives) and Metabolic & Endocrine (“not easy to quantify … midrange, perhaps leaning toward suboptimal”). At the bottom of the eight divisions in terms of regulatory performance, according to Milne and Kaitin, is Cardiovascular & Renal.

A Tale Of Two Divisions: Cardio-Renal, “The Other End Of The Spectrum”

• Accounted for 9% of NDA approvals; average to below average on IND, NDA submissions
• Above-average staffing level (58 for 2011)
• Moderate proportion of priority reviews (40%)
• More advisory committee meetings than average (50% of approved products)
• Above average “complete response” letter rate (37.5%)
• Longest NME approval times (29.8 months on average)

Source: “FDA Review Divisions: Performance Levels and the Impact on Drug Sponsors,” Milne & Kaitin, Clinical Pharmacology & Therapeutics, March 2012

The Human Factor

Staffing is a vital component of review performance, the authors note, declaring that “it is likely that this factor alone would account for a significant amount of variability in performance, given the apparent disparity between workload and staff.” The coming reauthorization of the Prescription Drug User Fee Act means a regime change that provides “a natural jumping-off point for staff departures,” they state, especially as “a relatively large number” of long-term FDA senior executives are reaching retirement age.

“Review divisions that appear to have a very high workload have at most only twice the number of staff as compared to divisions with considerably smaller workloads,” Milne and Kaitin report. And divisions with dramatically different workloads – Cardiovascular & Renal (with 6 NDAs submitted in 2010, below the CDER average of 6.9) and Anesthesia, Analgesia & Addiction (14 NDAs submitted in 2010) – have almost the same number of staff: 58 for Cardio-Renal and 54 for Anesthesia in 2011. They add that Anti-Infectives had among the lowest number of staff (48 in 2011), but almost the highest number of approvals with a fairly high workload in terms of complexity and number of applications.

Drug oncology, despite having the lowest 2011 staff level counted by the study with 31 people, had the highest proportion of 2006-2010 NME approvals (26%). The largest divisions by 2011 staff levels were Metabolic & Endocrinology (63) and Neurology (69); those divisions produced 6% and 8%, respectively, of 2006-2010 NME approvals.

Of course, workload is highly variable. Though Cardio-Renal had a low number of submissions in 2010, it was coming off two years of well above-average numbers, including the highest single-division annual NDA submission count from 2006-2010 (19 in 2008).

And the amount of work a single NDA requires also is highly variable – a novel cancer drug for patients with no other options will require smaller, shorter trials and, correspondingly, a more compact NDA than the mega-trials required for common, chronic diseases to identify not only signs of efficacy but, increasingly, potential safety signals.

The authors, in their conclusion, warn of a “vicious circle of increasing requirements” emerging from “an unhealthy public and political fixation on achieving a zero-risk system of product approval and utilization.” Efforts to enhance safety assessment have “led to the creation of a set of powerful tools to look for safety signals, some of which will inevitably be false but require long, large clinical trials to discover small and often short-lived ‘truths,’” they say.

There have been legislative efforts to increase FDA staffing with workload increases, both after the FDA Amendments Act and as part of the current round of PDUFA reauthorization (Also see "Biosimilars, Generic Programs Leave Little Room For Other New FDA Drug Initiatives In FY ‘13" - Pink Sheet, 20 Feb, 2012.). But the staff level is a relatively static variable, which is “unlikely to be responsive to yearly fluctuations in workload,” Milne and Kaitin state. The number of NDAs received each year by each division fluctuates significantly, making it hard to adjust. But the distribution of NDAs is relatively consistent – six divisions were consistently among the tops in terms of applications received: Cardiovascular & Renal, Anesthesia, Analgesia & Addiction (which included Rheumatology until 2010), Metabolic & Endocrine, Anti-Infectives, Antivirals, and Oncology.

NDA Filings Fluctuate, But Approval Times Are Consistent

The study found that an average of 67% of the NDAs received by FDA were approved after one review cycle. The first-cycle approval rates fell in a band from 50% to 75% for most divisions, the authors report, bracketed by Metabolic & Endocrine and Oncology on the high end (with 100% and 90% first-cycle approval rates, respectively) and Psychiatry on the bottom, with only 17% of approvals coming after one review cycle.

However, the time from NDA filing to approval proved “relatively consistent for the majority of review divisions,” Milne and Kaitin report. “Eight of the 12 divisions had mean durations that were within ±2.5 months of the mean for all divisions: 16.1 months.” Only four had greater deviations – Anti-Virals and Oncology were shorter and Cardiovascular & Renal and Neurology were “considerably longer” than the overall mean.

An earlier study by Milne and Kaitin’s Tufts CSDD colleagues Joseph Di Masi and Laura Faden also found that it was difficult to distinguish between divisions in terms of approval time. DiMasi and Faden similarly found that antineoplastic (oncology) and anti-infective products were on the low end of average approval times, but they singled out CNS and respiratory drugs as holding down the longer approval time end of the spectrum (Also see "Fast Track Speeds Approval, Tufts Finds; Therapeutic Category Also Key" - Pink Sheet, 1 May, 2009.).

In the Cardiovascular & Renal and Neurology divisions, Milne and Kaitin found that longer approval times correlated with low first cycle approval rates. “Approximately 40% of the submissions were subjected to multicycle reviews,” the study found. And approval time in those two divisions was considerably longer than the overall average for first-cycle approvals, with four- and two-fold increases, respectively. But the association between more multicycle reviews and longer approval time is not always consistent. Psychiatry had an extremely high rate of multicycle reviews, but, “surprisingly, the delays averaged only about a 1.5 fold longer approval time than single-cycle approvals,” the authors report.

Is There An Association Between Advisory Committees And Approval Times?

Advisory committee meetings are convened, in general, for about 40% of products that are eventually approved, the authors observe, although there is “considerable variation in this respect from division to division (17-67%) and also from year to year (11-71%).”

2011 was the second year in a row in which the majority of FDA’s novel drug approvals were done without advisory committee reviews (Also see "FDA Approving More Novel Drugs Without Advisory Committee Review" - Pink Sheet, 23 Jan, 2012.). The oncology division had the highest rate of approvals without advisory committee reviews.

“Review divisions that had a lower percentage of advisory committee meetings, specifically oncology (25%) and Psychiatry (17%) … had shorter approval times overall,” Milne and Kaitin found. Divisions with more committee meetings than average, like Cardio-Renal and Neurology (both held advisory committee meetings for 50% of approved new products), “have longer approval times overall.” The effect is not entirely consistent, however; the authors point out that “the Antivirals division had the highest percentage of meetings (67%), but there was little evidence of any impact on approval time.”

FDA’s Workload Is Likely To Increase

The impending PDUFA V legislation is “likely to increase the FDA’s workload even more,” Milne and Kaitin predict, although they hold out hope that Congress will provide sufficient resources to “clarify rather than compound some of the complexity introduced by PDUFA IV.”

Industry also is adding to the workload, bringing more complex products to the agency. “Industry is a victim of its own success,” Katin and Milne comment. The pharma sector “has risen to address substantial unmet medical needs globally” while also taking out “’low-hanging fruit.’” They report that “industry submissions of first-in-class and best-in-class new molecular entities (NMEs) rose from 7 per year in 1987-1996 to 10 per year in 1997-2007.”

Innovative drugs are, however, not only “more difficult to develop” but “often more difficult to review,” the authors point out, thanks to “the lack of experience with a new class of compounds, thereby contributing to the regulatory-uncertainty problem.”

Innovations like novel therapeutics paired with companion diagnostics require even more personnel involvement and coordination. But, such drugs can also have compelling cases that lead to extra fast approvals. CDER Director Janet Woodcock cited the personalized medicine aspect as a factor in several faster-than-usual approvals in a post on the “FDAVoice” blog (Also see "Innovative Drug Development: Kalydeco Provides A Paradigm, FDA’s Spielberg Says" - Pink Sheet, 20 Feb, 2012.).

FDA review documents for one of those – Pfizer Inc.’s Xalkori (crizotinib), an oncologic that sailed through FDA with its diagnostic companion in less than its six-month priority review timetable – indicate that the feat was achieved in part through intensive communication and cooperation with both CDER and CDRH (Also see "Synchronized Regulation: Tandem Approvals Of Xalkori, Diagnostic Required Considerable Coordination" - Pink Sheet, 1 Mar, 2012.).

As to the current workload across the divisions, it is not possible to know definitively how many NDAs are currently under review, but Pharmaceutical Approvals Monthly’s listing of pending publicly disclosed applications provides a rough yardstick of current workload (see chart, (Also see "Estimated FDA User Fee Review Goals For Pending NDAs/BLAs" - Pink Sheet, 1 Apr, 2012.))

Oncology appears on track to continue in the same vein; it is currently reviewing at least five NMEs or novel BLAs for metastatic cancers, along with about twice as many supplemental applications for new uses of approved oncologics. Cardiovascular & Renal has many fewer drugs under review, but the newer wave of anticoagulants, like Pfizer/Bristol-Myers Squibb Co.‘s Eliqis (apixaban), have some of the largest trial programs of any new agents. Many of the novel antithrombotics to come through the agency have taken multiple review cycles.

The division whose current workload seems to be most outsized – even given one of the larger staff totals in CDER – is Endocrine and Metabolic Drugs. Diabetes drugs, many of which now come with large cardiovascular safety studies, are a perennial source of NDAs, and the division continues to wrestle with three NDAs for obesity therapies. Hypercholesterolemia is active, with first-in-class agents like mipomersen from Ionis Pharmaceuticals Inc. and Sanofi’s Genzyme Corp., and even quiet fields like Cushing’s disease and lipodystrophy are in play with new drug candidates under review.