Appropriate Use of Recovery Groups in Nonclinical Toxicity Studies: Value in a Science-Driven Case-by-Case Approach

Abstract

A recovery phase—a nondosing period that follows the main dosing phase of a study—is sometimes included in nonclinical toxicity studies, and it is designed to understand whether toxicities observed at the end of the dosing phase are partially or completely reversible. For biopharmaceuticals with long half-lives, the inclusion of recovery arms can be helpful in understanding effects of prolonged exposure and assessing antidrug antibodies. This commentary discusses when to include recovery groups in nonclinical toxicity studies, the number of recovery groups to include in a given study, the number of animals to include in each recovery group, and the duration of the recovery phase. In general, the inclusion of recovery arms should follow a case-by-case approach that values rational scientific design and reflects the development needs and regulatory requirements applicable to individual nonclinical programs to ensure appropriate guidance for human studies while minimizing laboratory animal use.

Nonclinical toxicity studies are a component of the pharmaceutical drug development process, and they are required by regulatory agencies to support various stages of drug development and registration. They are conducted before the initial phase 1 clinical investigations to identify doses and exposures that cause toxicity (including dose-response relationships); to assist in setting an appropriate starting dose and a dose escalation strategy for the first human trials, as well as the maximum human dose; to identify possible monitoring strategies of potential toxicity in humans; and, in some cases, to assess pharmacologic activity. As drug development progresses, longer-term toxicity studies are conducted if needed, based on the clinical indication and clinical dosing duration. Overall, the data generated from these studies are crucial to guide safe clinical development that eventually leads to drug approval and marketing.

In addition to the evaluation of animals immediately after completion of the dosing phase of a toxicity study, some studies include cohorts of animals that undergo a dosing phase followed by a nondosing phase of a specified duration, termed a recovery phase or reversibility phase. There can be several reasons for inclusion of recovery groups in toxicity studies. The most common reason to include a recovery phase has been to assess whether effects observed at the end of the dosing phase persist, partially recover, or fully recover. Demonstration of partial or full reversibility of test article–related toxicity can be extrapolated to humans to suggest that a particular finding, if it were to occur in humans, might not be permanent. Conversely, a lack of reversibility or progression of findings may suggest that exposure to humans might not be safe, depending on the nature of the finding. Another reason for inclusion of recovery groups has been to determine whether the test article has the potential to produce delayed toxicity after dosing has ended. In some cases, toxicity developing after the end of dosing can be the result of prolonged exposure to the test article, such as that which occurs with monoclonal antibodies. It can be argued that this does not represent true delayed toxicity but rather represents a response to prolonged exposure. A third reason for inclusion of recovery groups has been specific for biopharmaceuticals, to allow for detection of antidrug antibodies, which may not be detectable in the presence of the test article, depending on the assay used. Finally, regulatory guidance (eg, addition of requirements for alternative first-in-human [FIH] approaches) has suggested use of recovery groups in some study types. Table 1 shows a listing of the current International Conference on Harmonisation (ICH) recommendations regarding the inclusion of recovery arms in nonclinical safety studies for pharmaceuticals. ^{1 –3}

Table 1.

International Conference on Harmonization General Guidance Regarding the Need to Assess Recovery/Reversibility on Nonclinical Studies to Support Clinical Development

Guidance M3 (R2): Conduct of Nonclinical Safety Studies

“The goals of the nonclinical safety evaluation generally include a characterization of toxic effects with respect to target organs, dose dependence, relationship to exposure, and, when appropriate, potential reversibility.” Alternative first-in-human approaches (exploratory clinical trials) Approach 1: Microdose  “Generally, extended single dose toxicity studies should be designed to evaluate hematology, clinical chemistry, necropsy, and histopathology data (control and high-dose only if no treatment-related pathology is seen at the high dose) after a single administration, with further evaluations conducted 2 weeks later to assess delayed toxicity and/or recovery.”  “A single dose level to assess reversibility/delayed toxicity on day 14 can support the microdose approach. The dose level used need not be the high dose but should be a dose that is at least 100 times the clinical dose.”  Approach 3: Subtherapeutic into anticipated therapeutic range (single dose)  “Generally, extended single dose toxicity studies should be designed to evaluate hematology, clinical chemistry, necropsy, and histopathology data (control and high dose only if no treatment-related pathology is seen at the high dose) after a single administration, with further evaluations conducted 2 weeks later to assess delayed toxicity and/or recovery.”  Approach 5: Repeat dose (14 days) without escalation to the maximum tolerated dose Relative to the maximum clinical dose: “In the absence of adverse effects in the clinical trial, escalation above this AUC can be appropriate if the findings in the toxicity studies are anticipated to be monitorable, reversible, and of low severity in humans.”

Guidance S9: Nonclinical Evaluation of Anticancer Pharmaceuticals

“Assessment of the potential to recover from toxicity should be provided to understand whether serious adverse effects are reversible or irreversible. A study that includes a terminal nondosing period is called for if there is severe toxicity at approximate clinical exposure and recovery cannot be predicted by scientific assessment. This scientific assessment can include the extent and severity of the pathologic lesion and the regenerative capacity of the organ system showing the effect. If a study of recovery is called for, it should be available to support clinical development. The demonstration of complete recovery is not considered essential.”

Guidance S6 (R1): Preclinical Evaluation of Biotechnology-Derived Pharmaceuticals

“Recovery from pharmacological and toxicological effects with potential adverse clinical impact should be understood when they occur at clinically relevant exposures. This information can be obtained by an understanding that the particular effect observed is generally reversible / non-reversible or by including a non dosing period in at least one study, at least one dose level, to be justified by the sponsor. The purpose of the non-dosing period is to examine reversibility of these effects, not to assess delayed toxicity. The demonstration of complete recovery is not considered essential. The addition of a recovery period just to assess potential of immunogenicity is not required.  “The addition of a recovery period just to assess for immunogenicity is not appropriate.”

While inclusion of a recovery phase in nonclinical toxicity studies can provide valuable data, there are some concerns regarding what constitutes the most appropriate use of recovery groups. One concern involves the responsible use of animals in research, as the addition of recovery groups generally leads to increased animal use. Inclusion of recovery phases also prolongs the study length and may delay the final report or result in increased work with the generation of both an interim report (without recovery data) and full final report. The increased number of animals used and the longer study durations both contribute to increased costs of the studies. Therefore, given animal welfare and resource considerations, scientists must critically evaluate when it is appropriate to include recovery groups in nonclinical toxicity studies. When recovery groups are deemed necessary, scientists should implement a science-driven case-by-case approach to determine the most appropriate study design, including potentially nontraditional designs. Factors to consider in study design are discussed below.

When to Include Recovery Groups

The general trend in nonclinical drug development study design in industry and the regulatory arena is toward a scientifically justified, judicious use of animals that includes appropriate use of recovery groups. In general, at least one nonclinical study should incorporate a recovery period to assess the reversibility of toxicity or the potential that toxicity will develop or progress after cessation of test article administration. Given the need for at least some recovery information in most programs, the primary question becomes one of timing. While some scientists prefer to include recovery groups in all GLP (good laboratory practices) toxicity studies, we believe that this is not necessary; instead, a case-by-case approach should be followed. Several strategies can be used, and the choice depends on the particular program. For example, recovery groups can be included in the FIH-enabling studies. When these studies are at least 1 month in duration, they have been shown to detect 70% of all toxicities associated with the test article when human toxicity is observed. ⁴ Thus, inclusion of recovery arms in these initial toxicity studies can provide reasonable confidence to drug development scientists that most potential toxicities for which recovery might need to be demonstrated would have been identified in these early studies. Also, data from these initial toxicity studies can then guide the need for recovery arms in longer-term studies. In addition, in many cases, the FIH package can be filed with regulatory agencies before completion of the recovery arms; therefore, this strategy does not necessarily delay the initiation of clinical trials.

Another strategy is to not include recovery arms in the FIH-enabling studies but rather to include them in next set of GLP toxicity studies—for example, 3-month studies. This strategy has the advantage of being able to use existing data to select the most appropriate recovery period and dose levels in the longer-term studies. Better dose-level selection should help mitigate the risk of mortality in main study animals as well as recovery animals, since mortality in either may negatively affect the ability to interpret the study and support clinical development.

If recovery groups have been included in the short- and/or medium-length toxicity studies (ie, up to 3-months in duration), then there is the question of whether to include more recovery groups in the chronic studies, if chronic studies are needed for the program. There are several prevalent opinions on this topic. If there has been no toxicity or if the effects observed in shorter studies have fully reversed, then some scientists prefer not to include recovery groups in the chronic toxicity studies. However, if a new finding develops and recovery data are considered necessary, an additional study may be needed; this may lead to significant program delays and additional costs. Based on this risk, some scientists prefer to include a recovery phase in most chronic toxicity studies.

Experienced pathologists can make scientific assessments of the potential for reversibility of many findings, based solely on the data at the end of the dosing period, taking into consideration the extent and severity of the change and the regenerative capacity of the affected organ system. In some cases, it is well known that certain toxicities are not relevant to humans (eg, α_2μ globulin nephropathy in male rats), while in others, it is well known that certain toxicities are reversible (eg, mild bone marrow depletion, mild necrosis in intestinal crypt epithelium). In the above scenarios, an assessment of reversibility may not be necessary.

It is also possible that a finding might be known to be irreversible but might still be considered acceptable for the given indication. Development of drugs for life-threatening indications, including advanced cancer, is one area where demonstration of reversibility may not add value in assessing clinical risk. When anticancer therapeutic candidates show compelling preclinical efficacy, it is uncommon for sponsors to be denied entry into phase 1 clinical trials just because they have not evaluated reversibility in first-in-patients enabling studies, as long as the observed toxicities are expected to recover (based on scientific assessment) and can be reasonably monitored in the clinical setting. The ability to make a scientific determination of reversibility may be more relevant to classes of drugs where there is extensive experience (eg, cytotoxic drugs and their toxicity to rapidly dividing cells). For novel mechanisms/targets, prediction of reversibility based on pathology at the end of the dosing phase may not be as clear-cut. With the flood of new oncology agents possessing novel pharmacology, it will be interesting to monitor the percentage of studies where an experienced pathologist is able to accurately assess reversibility at the end of the dosing phase.

In the past, one reason for including recovery arms in early preclinical toxicity studies of anticancer therapeutic candidates was to allow dosing in the clinic past the coverage provided by the preclinical studies. In reality, patients showing any positive response to a new drug (eg, stable disease or partial/complete response) are unlikely to be removed from a study in the absence of significant toxicity. Therefore, the value of inclusion of reversibility in early preclinical toxicity studies is debatable. Most recently, ICH S9 guidelines note that an assessment of recovery should be provided before entry into phase 1, but this assessment may be based on scientific judgment rather than demonstrated reversibility. In the case that there is severe toxicity at an exposure approaching the clinical exposure range and scientific assessment cannot reasonably predict reversibility, another study would be required to address recovery and should be made available to support clinical development.

Thus, when reversibility can reasonably be predicted, then it may be possible to reduce the number of dose levels in which there are recovery groups, to reduce the duration of recovery, to show only partial recovery, or to eliminate the recovery phase in some or all studies. However, there is a risk that regulatory agencies may require that recovery (partial or full recovery) actually be demonstrated. Engagement of regulatory agencies before finalizing study designs may be very useful in these cases.

In the end, there is not necessarily one best strategy for including recovery phases in development programs, but the scientists in the program should rationally design their nonclinical toxicity studies—including when to use recovery groups—within the overall context of the development program.

Numbers of Recovery Groups and Numbers of Animals per Group

Inclusion of recovery groups at all dose levels is often not necessary, and careful consideration of the program needs and existing data can assist in making appropriate decisions. For example, if a dose–response relationship for reversibility is needed or complete reversibility must be demonstrated, then it might be appropriate to include recovery groups at the low-and high-doses but not at the mid-dose. In contrast, for molecules that demonstrate little toxicity, such as monoclonal antibodies against soluble cytokines, it might be appropriate to include recovery groups only at the high dose or argue that a recovery group is not necessary at all. There are risks with only including recovery groups at some dose levels. If recovery is included in only the high-dose group and unexpected high mortality occurs, there may be no animals available for recovery. Additionally, if recovery is included only at the high-dose and full reversibility is not demonstrated, it may necessitate another study, depending on the nature of the finding and indication. If recovery is included only at the mid-dose or low dose, there is a risk that the toxicity of interest may not occur in those groups. Nevertheless, in most cases adequate studies can be designed without including recovery groups at all dose levels.

The number of animals in recovery groups should be sufficient to determine whether effects are reversible while using the minimum number needed. Existing pharmacology and toxicity data, including the incidence of the toxicity in question, can play a critical role in selecting the appropriate group size. Most rodent studies include 5 to 10 recovery animals per sex per group. For large animals, such as dogs and monkeys, it is generally recommended that at least 2 recovery animals per sex per group be used—and sometimes 3 or more. In general, findings with a lower incidence usually require more animals in the recovery phase to ensure that at least some animals have the finding at the end of the dosing phase. As such, even in case of toxicologic findings of lower than 100% incidence, a recovery group the size of the dosing phase group (eg, 10 rodents per sex per group) should provide a reasonable assurance that the toxicity noted in the dosing phase is represented in the recovery phase. However, it must be accepted that designing the size of a recovery group for findings of extremely low incidence is difficult given that the incidence of such findings is generally based on small sample sizes in limited number of studies; therefore, extrapolation to predict future incidence might be erroneous. Biomarkers that can demonstrate the presence of the finding in the dosing phase can be very useful in these cases and may allow the selection of appropriate animals for the recovery group. For example, if a drug candidate causes hepatocellular injury in a small subset of the dosed animals, the protocol could specify that Alanine transaminase (ALT) values at the end of the dosing phase be used to determine which animals were necropsied and which went into recovery. Based on this example, animals could be stratified from low to high ALT values, and every other animal could be placed in the recovery group, ideally resulting in a comparable mix of affected and unaffected animals in both the main study and the recovery arms.

Duration of Recovery

The duration of the recovery period should be long enough to meet the program needs. In some instances, partial reversibility may be acceptable (eg, prior experience with a particular toxicologic finding, life-threatening indications), while in other cases, full reversibility may be required (eg, novel findings, novel class of drugs, “lifestyle” indications). Consideration of existing pharmacology, toxicology, and pharmacokinetic data should play a key role in determining the duration of recovery, since failure to meet the program needs may necessitate additional studies. Also, the duration of recovery may not be the same in all species. In cases where there are limited pharmacology or toxicology data to help dictate the recovery duration, certain general principles can be applied.

For test articles with half-lives of 1 day or less, recovery periods of 2 to 4 weeks in initial toxicity studies are common. For molecules with longer half-lives, which often include biopharmaceuticals (especially monoclonal antibodies or other related molecules that have a half-life of up to several weeks), dose-free periods of 8 to 12 weeks are not uncommon. The term dose-free period is used because it should be recognized that part of this time will have continued exposure to the test article and will not represent a true recovery period. In this context, for molecules with long half-lives, the possibility exists for toxic effects to appear after cessation of the dosing period, related to the prolonged exposure. The possibility for continued test article exposure in at least part of the recovery phase may need to be factored into the duration of recovery and the interpretation of the data. In the absence of other guiding information, recovery periods of 5 to 7 half-lives can be used for molecules with half-lives of 1 week or longer. Approximately 95% of the drug is cleared after the fourth half-life (> 99% by the seventh), providing an opportunity for an evaluation of true recovery after that time. If pharmacodynamic effects of the test article last well beyond exposure to the test article, recovery periods may need to be prolonged.

Assessment of Delayed Toxicity

Knowledge of the half-life of the test article in blood or tissues can help determine whether toxicity is truly delayed or represents continued exposure for test articles with longer half-lives. A specific evaluation of delayed toxicity is generally uncommon; however, in some cases, it is explicitly required. ICH M3 (R2) (Table 1) describes specific study requirements for nonclinical toxicity studies supporting exploratory clinical trials that are generally performed in healthy volunteers. ¹ Because the duration of the dosing phase of these toxicity studies can be short (even a single dose), regulators have included a requirement for the assessment of delayed toxicity (and/or recovery). Conversely, the ICH S6 (R1) guidance notes that reversibility of effects should be evaluated. This revised guidance also specifically notes that recovery phases are for assessing reversibility and not for assessing delayed toxicity for biopharmaceuticals. ² This recommendation is based on the premise that, to date, delayed toxicity with biopharmaceuticals has generally not been observed and toxicity present after the end of the dosing phase is likely to be associated with the prolonged exposure of these agents that have prolonged half-lives.

Additional Considerations

In addition to the more standard designs for recovery studies, where similar numbers of animals are used in control and test article–dosed groups, other designs can sometimes be considered. To reduce animal use, control animals may be omitted in the recovery arm. However, this approach has the potential to compromise a rigorous scientific study design by hindering accurate interpretation of findings due to the lack of concomitant, age-matched controls. An intermediate approach is to reduce the number of animals in the control cohort, although this is difficult when considering large-animal studies that already have few animals in the recovery phase. Additionally, the statistical implications of unequal group size will need to be taken into consideration during study design and interpretation of results.

The determination of particular end points to be examined during the recovery phase (histopathology, clinical pathology, organ weights, biomarkers, etc) and the frequency of examination (of nonterminal end points) should be decided on a case-by-case basis and should take into account the toxicologic finding being analyzed in the recovery phase, the species under study, and the individual development program needs. For example, in the case of hepatocellular injury, inclusion of blood sampling to evaluate liver transaminases during the recovery period may provide insight regarding when during the recovery period the liver signal resolved, and it may also suggest the time for necropsy, to confirm the resolution of the morphologic finding. Most commonly, histopathologic analysis following the recovery phase is performed only on tissues that were affected during the dosing phase. However, it may be prudent to collect the standard array of tissues in case further analyses might be necessary.

As noted above, scientists may consider including recovery arms in post-FIH, longer-term studies, even when recovery arms were included in shorter-term toxicity studies. If no significant toxicity is observed at the end of such a study or if the toxicities observed can be justified to likely resolve (perhaps based on the recovery data from the shorter-term study), it is possible to consider terminating the recovery arm immediately or at the end of the scheduled recovery phase without further evaluation or with some type of limited evaluation (such as collection of tissues without microscopic examination) to optimize resource utilization. A scientific, case-by-case approach should be used when implementing such a strategy.

Conclusions

There is clearly value for many pharmaceutical development programs to include recovery groups in nonclinical toxicity studies to assess the reversibility of nonclinical findings. Recovery data can provide important information for the risk assessment of drug candidates, as well as for the design of future toxicity studies. Although the approach to the design of the recovery phase is often similar across industry, nonstandard designs that intentionally minimize animal use or resource outlay can provide useful information even though they may carry additional risk. The nonclinical toxicity studies in which to include a recovery phase, and the design of those studies, should use a case-by-case approach that values rational scientific design and reflects the development needs and regulatory requirements applicable to individual nonclinical programs. Such an approach should maximize the value of the obtained data, minimize the use of animals, and provide appropriate guidance for human use.