Why Do So Many Biopharmaceuticals Fail?

James R. Zabrecky

December 1, 2008

16 Min Read

Biopharmaceutials and the processes used to make them are exceedingly complex, and the path to developing new therapeutics is a high-risk endeavor. The emphasis today is on controlling product quality, safety, and efficacy through understanding biological mechanisms, key product attributes, and process parameters. Such information is also crucial for guiding development efforts to improve chances of success in the clinic and for gaining regulatory approval. Analytical methods provide the foundation for acquiring such knowledge. Efforts devoted to developing meaningful and reliable analytical methods early in development will pay huge dividends and can make the difference between success and failure.

Drug development is a tough business. Costs are astronomical, and the odds of failure are about nine out of 10 for every new drug that enters clinical trials. Current estimates for bringing a new drug to market are over US$1 billion spent on an average timeline of 12–15 years. High development costs and the likelihood of failure are major reasons why the new generation of drugs, especially biologics, are so costly. If not for potentially high rewards, innovation would grind to a halt because few companies would consider investing the time and expense required to develop new therapies for treatment of unmet human medical needs.



Why is it that so many promising drug candidates fail? Finding answers to that question should help improve efficiency of the process and speed the transition of new medicines from laboratories to market. Although much discussion and effort have been put into trying to identify some root causes, the reasons vary widely, and there have been few significant improvements in the way drugs are developed. Very few drug candidates now in preclinical development will ever see the light of day on the market.

Most drug candidates are based on fundamentally sound scientific foundations, but they can fail in the clinic for a variety of reasons whether or not they represent truly viable therapeutics. Reasons range from flawed clinical trial design (patient populations, treatment regimens, doses, endpoints, and so on), unfavorable pharmacokinetics and pharmacodynamics, to unanticipated placebo effects, poor compliance with the treatment, lack of follow-up, or faulty data interpretation. But an often overlooked factor that may play a significant role in success is whether a drug ever gets to its target, in a consistent, active form and at the right dose, so that it has a chance to accomplish its intended task.

That is why product characterization during preclinical and early stage clinical development plays such an important role. This critical facet of the drug development process is often abbreviated or deferred because of naïve notions that drug product attributes and underlying biological mechanisms are understood well enough to go forward. That is especially true for smaller biotech companies that often are under fiscal constraints or pressure from investors to enter into human clinical trials as soon as possible. It is far too common for a promising drug candidate to be driven into the clinic prematurely, long before its properties and manufacturing processes are understood well enough to give it a reasonable chance of success. Moreover, once human trials are initiated, it becomes increasingly difficult and risky to make the critical changes that become necessary as more product and process knowledge are accumulated.

The consequences of cutting corners in product characterization are most apparent in biological therapeutics. Small–molecular-weight drugs are simpler chemical entities and thus are amenable to characterization by an arsenal of analytical tools that can explicitly define their structure and the attributes necessary for biological activity. In addition, we know a great deal about how to monitor their fate once they are delivered to patients. Biologics, on the other hand, are large, complex macromolecules made up of many components that contribute to their structure, function, and stability. Despite considerable advances in analytical technologies, it may be a long time before we are capable of fully understanding all the chemical characteristics that govern the properties of such proteins. But still there is much we can learn about the structure–function relationships and other critical properties of proteins and other biomolecules. There is an opportunity to greatly improve the odds of success through concerted efforts to develop an in-depth understanding of the biological mechanisms involved in each biomolecule’s activity and to characterize key attributes that affect those mechanisms.

It is no coincidence that monoclonal antibodies have emerged as such an important class of biopharmaceuticals. Antibodies have the unique properties of being nearly identical to one another while having the capacity for untold specificity and diversity. We now have accumulated considerable understanding of antibody structure and behavior. Unlike other therapeutic proteins, antibodies are easily purified using powerful affinity techniques. Their structures, including posttranslational modifications, have been extensively characterized. Most important, they are relatively stable and soluble, allowing for a broad range of delivery options and treatment regimens. The developmental scenario for other proteins and complex biologics is not as straightforward because of more limited historical knowledge, complicated purification schemes, and stability issues.

Mechanism of Action

In efforts to improve efficiency in drug development processes, we are becoming increasingly reliant on mechanism of action- (MOA) based drug discovery. Examples include ligands, which interact at receptors to elicit a biological response, and monoclonal antibodies targeted to specific antigens. Having a clearly defined target and a strong understanding of the underlying biology of how a drug affects its therapeutic response clearly aids in product development. There is no requirement that one have a full understanding of the MOA to gain approval, but the more that is known, the better. This allows for the development of analytical methods that focus on product attributes that are related to a drug’s intended purpose.

Many new therapeutics are made up of complex mixtures of biomolecules, and it is not always easy to identify the principal active component(s). Other entities or cofactors, in addition to the primary active ingredient, may contribute to the overall activity. In such a case, a company must be very careful when optimizing purification processes because it is possible to inadvertently remove or inactivate key components. Activity assays may not have the precision or specificity to detect changes that could alter product efficacy.


A series of analytical procedures must be developed and qualified early to address the key product attributes of identity, strength, purity, potency, and safety. These must be incorporated into a quality system for product release to ensure quality and consistency. Appropriate specifications should be established that balance product needs, manufacturing capabilities, industry standards, and regulatory expectations. When selecting release tests, it is important to focus on attributes that address properties of a biomolecule and relate to its function and safety. Some overlap is useful to aid in failure investigations, but companies should avoid too much redundancy and assays that do not provide relevant information.

Product specifications are usually set with broad limits early in development, then narrowed as processes and the level of product understanding are refined. It may not be necessary to validate analytical methods until later, but they at least should be qualified to provide a high level of confidence that the results are dependable. Making critical decisions based on unreliable data can have catastrophic consequences.


A meaningful and reliable potency assay is probably the most important tool in drug development. Not only are potency assays necessary to ensure that released drug product has the best chance of working for a patient, but they can be indispensable for assessing stability and comparability, for process and formulation development, and for process validation. Many biologics are originally developed using in vivo models, but such assays typically are not suitable over the long-term for product development because they are labor intensive and time consuming, and they lack sufficient precision. The goal should be to develop a simple, reliable, and robust potency assay the earlier in the development cycle the better. Such assays can take the form of physicochemical methods, receptor/antigen binding, ELISAs, enzymatic activity, or in vitro cell based assays. Effort should be directed toward showing correlation between potency assay results and in vivo responses.

A big challenge for biologics is to find a single potency assay that provides sufficient information to meet the needs of product development and satisfy regulatory requirements of biological relevance and quantitative ability. This is especially difficult for complex biologics or those with poorly understood MOAs. Often a single potency assay will not suffice, and a developer must consider multiple assays that address various product attributes. For example, a company might combine methods that assess biological activity with assays that measure structural integrity or the concentration of important product components or cofactors. Product release can be based on meeting individual specifications for each method. The advantage is that simple pass–fail criteria may be sufficient for less precise biological assays, and you can rely on other methods (such as physicochemical assays that may not be as biologically relevant) for the quantitative aspect of potency. Using such a combined approach (referred to as an assay matrix) can provide high assurance of product activity and manufacturing consistency.

Sometimes in the case of complex biologics, especially vaccines, the identity of the one or multiple active agents is not known with certainty, which surely can confound efforts to design meaningful potency assays. Patient-specific and autologous therapies present another difficult situation in which a drug may be active in only one or a small subset of patients. In such a case, surrogate assays can be used that focus on general attributes of the product known to be responsible and necessary for its activity. The challenge is to accumulate scientific evidence to build a strong case that a surrogate assay(s) is biologically relevant to the drug’s intended purpose.

Characterization and In-Process Methods

The analytical toolbox should include an array of methods that can be applied to the characterization of a product. For many reasons, not all analytical methods are appropriate as release tests, but such assays may serve to provide valuable insight into the molecular and biological properties of the product. Characterization methods can address product attributes such as size, structure, purity, chemical modifications, and biological activity. Sometimes their results can help predict stability, especially in conjunction with forced degradation studies or when correlated with adverse events or immunogenicity issues. Application of as many characterization methods as possible improves the general level of product understanding, provides greater assurance to regulatory agencies that a product is as it is defined, and better prepares a company for surprises that may occur.

In-process analytical tests are essential to understanding and maintaining process control. They are used for monitoring product quality and consistency, and for making critical processing decisions during manufacturing. In-process tests can include on-, in-, and at-line measurements that aid in the control of a manufacturing process to ensure that the resulting product consistently meets predefined quality standards.

In-process assays can be used to optimize process variables so as to build in process control and robustness. The emphasis is now on proactively designing in product quality and process control through better understanding of the underlying science and manufacturing design space. Insight into important product attributes, gained through knowledge of product biology and structure-function relationships, can be leveraged for better definition of critical manufacturing parameters. Such efforts rely highly on development of suitable analytical methods used to set acceptable limits around key process variables.


Because biologics are large, complex macromolecular structures, they inherently possess many more potential pathways for denaturation and degradation, and thus are more prone to stability related issues than small molecule drugs. Biologics are susceptible to a number of environmental influences including temperature, pH, light, oxidation, ionic strength, chemical modification, and drying. Denaturation usually involves changes to the three-dimensional structure or aggregation state of a molecule leading to loss of activity or altered properties. Degradation, on the other hand, often involves cleavage of chemical bonds producing new molecular species. It is important to have an understanding of how different denaturation or degradation pathways affect the activity of a product. In addition, stability-related changes can result in safety issues such as immunogenicity and unwanted side effects or toxicity from degradation products.

Stability-indicating assays should be developed and qualified early on, and stability programs should be designed to focus on product attributes that are critical to activity and safety. This again points to the importance of having meaningful and reliable potency assay(s). There are advantages to developing physicochemical methods as stability indicators because they are usually simpler and can provide greater precision than bioassays. Such assays sometimes can reveal trends that are early predictors of stability issues. Accelerated stability and forced degradation studies can be useful tools to acquire insight on product stability in a shortened time frame.

Stability assay development should be integrated along with a formulation development program to identify, as soon as possible, stabilizing excipients and optimal storage conditions. An appropriate formulation is necessary to ensure that a product is stable throughout the duration of a clinical study and that all patients are administered material of uniform quality and quantity. Variations in activity and dose resulting from stability or delivery issues can severely confound the interpretation of clinical data and make the difference between meeting or failing to achieve statistical significance of clinical endpoints. Ideally, formulation development should happen before human clinical trials are initiated. Later formulation changes will require comparability studies, and they always carry the risk of unanticipated effects on the clinical outcome.


Process Changes and Comparability

Although very risky, manufacturing process changes are inevitable during the course of clinical development, whether for scale-up or refinements to improve manufacturability for commercialization. Process changes often are implemented to improve efficiency, consistency, or robustness or to incorporate technological advances. Demonstrating comparability following a process change can be especially challenging if appropriate analytical methods are not in place. It is essential to demonstrate comparable or improved potency, stability, and especially safety (e.g., impurities and contaminants) following any significant process change. Process changes resulting in a seemingly improved product may appear to be a good thing, but they are not without pitfalls. Making changes can lead to unforeseen consequences due to changes in dose, product pharmacokinetic (PK) and pharmacodynamics (PD), side effects, and so on.

In the absence of full product understanding, the biologics world is still bound by the old concept of “the process defines the product.” This notion has evolved into modern-day FDA initiatives such as process analytical technology (PAT) and quality by design (QbD). These are intended to promote the idea of controlling the quality of final products through process and product understanding and building in better process control using in-process analytics and knowledge of key variables. For complex biologics, manufacturing processes may be optimized to yield what is believed to be an active ingredient, but a company may end up, sometimes unknowingly, with reduced activity due to removal of a key component. This is where meaningful potency assays play such an important role.

Reference Standards

A reference standard is an indispensable resource during early product development and throughout the product lifecycle. For most new biologics, it is unlikely that a recognized reference standard already exists, so an in-house primary reference must be made and characterized. It is essential to set aside some material for this purpose as soon as possible. Establishment of a reference standard does not have to wait for a finalized process or extensive knowledge of product stability. As soon as, or even before, a process begins to resemble what will be used to generate clinical trial material, a batch should be prepared and stored in aliquots for use as reference. The safest approach is to place the material at −70 °C and assume it remains stable (this has to be tested) until further formulation development efforts can be undertaken. The primary reference standard should be as representative of the final product as possible and it should be evaluated by all available analytical tools.

A reference standard has many uses. It will serve as a valuable benchmark during process development to assess comparability and consistency. It is essential in the development of analytical methods and to monitor their performance over time. In certain cases, it may be used as a working assay standard to generate standard curves or as a control to set assay acceptance criteria. It also will be used to assess stability, track trends in product attributes, and to ensure against product “drift” or “creep” over time.

Reference standards commonly change during product development when they run out or after significant process changes have been made. In such a case it is important to conduct appropriate crossover studies using all available analytical methods.


Suitable analytical methods are needed for measuring the form, concentration, and fate of an administered drug in body fluids, tissues, and at its target site of action. PK and PD studies are necessary to gain a thorough understanding of the relationship between dose and in vivo concentration, comparing the intended response and unfavorable side effects. This can include development of assays for biomarkers and their validation as reliable predictors of drug response. Such efforts typically are conducted throughout all phases of clinical development and, together with modeling studies, they can be used to optimize dose regimens and other aspects of clinical trial design to improve outcomes.

Know Your Product

There is simply no substitute for having a sound scientific understanding of your product and the processes used to produce it. That can be achieved only through the diligent application of a broad range of analytical methods to product characterization throughout all phases of the development cycle. Confidence in and reliability of analytical results come from thorough assay optimization, qualification, validation, and tracking of performance over time.

The past couple of decades have brought numerous advances in analytical technologies, providing powerful new tools that can be applied to the characterization of biological pharmaceuticals. Regulatory agencies continue to evolve their thinking as we accumulate more experience with biologics and knowledge of key factors that affect efficacy and patient safety. The path to approval for new biologics, especially MAbs, clearly is much better laid out now than it was 20 years ago when the first such products were developed.


You’ll never know everything there is to know about a complex biopharmaceutical, so you must make a number of assumptions and take calculated risks during the course of clinical development. However, the more effort you put into characterizing and understanding your product and process early in development, the better prepared you will be to make informed decisions and deal effectively with unanticipated problems that inevitably will arise.

Finally, given the enormous complexity of biological systems, there is always a certain element of luck involved. However, it is very risky to rely on luck, and it is important to remember the words of Louis Pasteur: “Chance favors the prepared mind.” It is essential to put as much effort as possible, early on, into understanding a product, its manufacturing process, and the fundamental biology behind its mechanism of action to give it the best possible chance at succeeding in the clinic.


1.) 2006. Tufts Center for the Study of Drug Development. Impact Report 8.

2.) 2007. Tufts Center for the Study of Drug Development. Impact Report 9.

3.) 2004..

4.) 2004..

5.) 1999..

6.) CDER/CVM/ORA 2004..

You May Also Like