By Steve Jolley, VP and Director of Pharmacovigilance, Patni Life Sciences

Pharmacovigilance — the detection, assessment, understanding and prevention of adverse effects — may sound simple, but it is a complex science. Pharmaceutical companies today face an increasingly detailed set of international regulations dictating pharmacovigilance practices that are essential to minimizing risk in drug development and commercialization. Yet the path to good pharmacovigilance remains unclear. While European regulations provide fairly straightforward guidelines, the new FDA regulations do not. Drug manufacturers serving U.S. consumers are left to identify the best approach to managing this risk on their own.

Pharmaceutical companies are under immense public pressure to improve drug safety. Manufacturers who let medications with serious unexpected side effects slip out into the general market risk harming patients, and their bottom lines. These are not mere theoretical risks. Many a pharmaceutical company has fallen from grace when one of its products caused irreparable harm to the public good. The withdrawal of Vioxx from the market, for example, resulted in a $27 billion reduction in Merck's market capitalization overnight, the loss of $2.5 billion in annual sales, and potential legal liabilities of up to $5 billion [1] — not to mention the immeasurable costs to patients and their families.

Seeking Guidance

The FDA Amendment Act (FDAAA) passed at the end of 2007 outlines the need for companies to create risk evaluation and mitigation strategies (REMs). While risk assessment encompasses signal detection methods, the way to do so remains unclear. The FDAAA does not spell out detailed requirements for signaling, nor does it require REMs for all products. The only requirement for signaling in the U.S. is the March 2005 guidance, "Good Pharmacovigilance Practices and Pharmacoepidemiologic Assessment."

Manufacturers can look to the recently passed initiatives from the International Conference on Harmonization (ICH) and European Agency for the Evaluation of Medicinal Products (EMEA). These guidance documents emphasize the need for proactive pre-market risk surveillance and the development and implementation of risk management programs as tools to ensure patient safety. Specifically, these guidelines state that it is imperative to collect risk information generated during clinical trials and long-term controlled safety studies, taking into account all currently available safety data. This allows for a comprehensive evaluation of dose effects, drug interactions and misadministration, providing insight relative to class effects and the structure-toxicity relationship. In addition, pharmacoepidemiologic assessments are needed to identify and analyze potential populations at risk of adverse events from a medication.

Taking the First Steps

Developing the right game plan for pharmacovigilance starts with proper planning. This involves three steps:

1. Create an objective profile of the company's strengths and weaknesses in drug safety, and perform a detailed assessment that maps out the current environment in relation to best practices and applicable regulations.

2. Establish practices to monitor the risk.

3. Establish a formal methodology for signaling that includes a comparison of internal safety databases and external databases to assess the significance of any signals generated.

Signaling is an important process, used to define the initial safety profile of a drug. During pre- and post-marketing stages of drug development, adverse event incidents are collected and stored in databases. Medical/safety expertise is routinely applied to detect and investigate potential and true signals and other events of interest. This involves several steps, including: centralizing reporting in clinical trials; establishing the appropriate triage and response activities to medical inquiries and product complaints; and recording adverse events reported for post-marketed products. When performing signal analysis, organizations need to find the answers to critical questions.

Understanding Adverse Events

Most organizations are already collecting and reporting adverse events. Parsing this information, however, can be a daunting task. Signal intelligence should add scientific efficiency to the work of traditional safety/medical experts. As such, signal analysis tools should be comprehensive and at the same time streamlined for easy execution. The efficiency of routine signal analysis by those who assess and evaluate safety data can be significantly increased. For example, for one individual to manually assess and evaluate just 1,000 patient records, containing thousands of adverse events, can take weeks or even months. Using an appropriate signal analysis application can significantly reduce this review and assessment time.

Analyzing the relevance of adverse events is another challenge. Drug developers need tools that will quickly identify trends and patterns so that they may proactively identify and address any adverse issues. This is crucial for understanding not just what's wrong with a product, but what's right with it; these tools can be useful in making approved marketing claims. For example, during Pfizer clinical trials of a drug for hypertension, an unusual side effect inspired the development of Viagra.

Signal analysis tools can be used to generate descriptive signal patterns and statistical signal scores during the pre- and post-approval stages of product development. These tools provide a scientific approach to understanding a product's risk/benefit profile. On one hand, they can identify potential side effects that could harm public safety. On the other hand, they can also help to better understand if adverse events are restricted to a certain population. This knowledge is integral to the success of a medication.

Consider this hypothetical scenario: An anti-epileptic drug initiates hepatotoxicity in some patients, resulting in their hospitalization and in some cases death. The product is quickly pulled from the market, but not soon enough to avoid the resulting in litigation expenses, tarnished company image, lost revenues, and tumbling stock values.

However, a follow-up evaluation finds that the adverse events were only associated with a certain patient population experiencing liver conditions. Having a good signaling program in place could have turned this entire scenario around. First, and most importantly, early monitoring of elevated liver enzymes in that selected patient population could have saved lives and prevented countless illnesses. Moreover, early warnings about a selected population's use of the medication could have prevented the drug from being pulled from the market entirely, and averted the financial fallout that resulted.

Benefiting from Signal Intelligence

Through sophisticated visual analytics, drug manufacturers can fully assess the risks and benefits of a particular medication. The use of statistical analysis tools and external safety data sources, along with visualization tools, has made it much easier to evaluate the meaning of data within the greater context of the general patient population.

For example, a company developing an injectable contraceptive sought to build a risk management program into drug development. The only competing product had been withdrawn from the market due to safety concerns related to injection site reactions.

Using signaling software (in this case, Patni's Signal Intelligence-PV software), the company developed a series of analyses to assess adverse events reported around the world, with an additional focus on injection site reactions. These analyses were outlined in the company's risk management plan submitted to FDA, and helped the product gain approval.

Statistical signals must be evaluated to decipher which signals are potential or true signals and if and how they translate to potential or real risks, their clinical significance and health impact. Based on medical expert causality assessment of the signal analysis results, events are prioritized for further evaluation. Especially serious adverse events or adverse drug reactions and events that may lead to study drug discontinuation are given the greatest are evaluated first.

Employing Good Pharmacovigilance Practices

As illustrated previously, good pharmacovigilance practices require the inclusion of external data sources, which provide a useful comparison to other products in the same therapeutic area. Merging safety, clinical and sales data along with external data from FDA's AERS database and from the World Health Organization's Vigibase is most useful. These practices should be combined with additional analyses, regarding clinical and laboratory manifestations of the event, demographic characteristics of patients, exposure duration, dosage amounts, use of concomitant medications, the presence of co-morbid conditions (particularly those known to cause the adverse event, such as underlying hepatic or renal impairment) and changes in event reporting rates over the product lifecycle.

This strategy mitigates operational risk by dedicating resources to high-risk areas. For example, organizations can use analysis software to examine alternative scenarios to understand the causal relationship between the medication in question and reported adverse events.

The earlier this is started in the development life cycle, the better the results. Ideally, a risk management plan should be started in the pre-clinical phase, and certainly early on in clinical development. By the time the product is submitted for approval, a comprehensive assessment will have been made of the product's known and unknown risks (those that might be expected in this disease state or patient population but have yet to be observed). Potential risks found in a product's pre-clinical studies should be added to the clinical program and studies developed to assess and quantify the nature of these risks.

Conclusion

Drug developers are facing an array of new regulations. They are looking for clear guidance in furthering their pharmacoviligance practices, both to protect public safety and sustain financial value. Systematic efforts should be put in place to try to continue to identify new risks or any changes in the known safety profile of the product. Automation of signal analysis is an additional pharmacovigilance tool that can add critical value to risk assessment and risk management processes, allowing manufacturers to make sense of the wealth of data at hand, address adverse events quickly, identify and analyze potential populations at risk and reduce medical errors.

References

1. "Merck Agrees to Pay $4.85 billion in Vioxx Settlement." http://www.reuters.com/article/governmentFilingsNews/idUSL0929726620071109