Waive Inclusion and Exclusion Criteria?

In every clinical trial there are study subjects who are very close to meeting one or more inclusion or exclusion criteria, but for one reason or another don't.

The inclusion criterion may be:

• A screening creatinine of <2 mg/dL is required, and the subject has a creatinine value of 2.2 mg/dL.

• Subjects must be >18 years old, but the volunteer is 17 years and 10 months old.

• A screening hemoglobin of >10 mg/dL is requited, but the subject has a hemoglobin value of 9.8 mg/dL.

• Dosing is to occur within 14 days of the screening visit, but the subject is dosed on day 15.

Oftentimes, the scenario goes like this: The principal investigator identifies a subject who does not meet one or more of the inclusion/exclusion criteria and contacts the sponsor to request a waiver. The sponsor provides the waiver. The subject is enrolled into the clinical trial. The principal investigator notifies the Institutional Review Board (IRB)/Independent Ethics Committee (IEC) that the sponsor has provided a waiver, and a subject who did not meet inclusion/exclusion criteria was enrolled into the trial.

To many readers, this scenario makes sense, is appropriate, and it is what they do. The purpose of this paper is to remind these readers that this scenario is not GCP-compliant. GCPs require the IRB/IEC to approve the enrollment of a subject, not meeting inclusion/exclusion criteria, prior to that subject's enrollment into the trial. This article will review the regulations and guidelines when a sponsor provides a protocol waiver to enroll a subject who does not meet one or more inclusion/exclusion criteria; the FDA's apparent position on sponsor-provided waivers (based on warning letter references); the IRB/IEC's responsibility regarding these waivers; and the risks to the subject, investigator, sponsor, and clinical trial when protocol-ineligible subjects are enrolled into a clinical without the prior approval of the IRB/IEC.

Implications of protocol waivers

The principal investigator. Before the trial begins, the principal investigator signs the FDA 1572 and, typically, signs the sponsor's protocol signature page. By signing, the principal investigator has documented his/her intention, commitment, and obligation to conduct the trial in compliance with GCPs, the applicable regulations and guidelines, and the protocol. In the absence of a life-threatening situation, US regulations and ICH guidelines stipulate that changes to the protocol are only permitted if the sponsor has alerted the IRB/IEC, and the IRB/IEC has granted prior approval.

For clinical trials involving drugs:

The investigator shall also assure that he or she will promptly report to the IRB all changes in the research activity and all unanticipated problems involving risk to human subjects or others, and that he or she will not make any changes in the research without IRB approval, except where necessary to eliminate apparent immediate hazards to human subjects. (21 CFR 312.66: Assurance of IRB Reviews)¹

The investigator should not implement any deviation from, or changes of, the protocol without agreement by the sponsor and prior review and documented approval/favorable opinion from the IRB/IEC of an amendment, except where necessary to eliminate an immediate hazard(s) to trial subjects, or when the change(s) involves only logistical or administrative aspects of the trial (e.g. change of monitor(s), change of telephone number(s)." (ICH E6: Section 4.5.2)²

For clinical trials involving medical devices:

An investigator shall notify the sponsor and the reviewing IRB (see 56.108(a) (3) and (4)) of any deviation from the investigational plan to protect the life or physical well-being of a subject in an emergency. Such notice shall be given as soon as possible, but in no event later than five working days after the emergency occurred. Except in such an emergency, prior approval by the sponsor is required for changes in or deviations from a plan, and if these changes or deviations may affect the scientific soundness of the plan or the rights, safety, or welfare of human subjects, FDA and IRB in accordance with 812.35(a) also is required. (21 CFR 812.150)³

Most protocols restate this requirement for further emphasis. A sponsor's waiver for inclusion/exclusion criteria, in the absence of IRB/IEC approval, does not supersede the FDA 1572, the protocol signature page, the regulations or guidelines, or GCPs. A waiver of inclusion/exclusion criteria granted by the sponsor does not grant permission to the principal investigator to enroll ineligible subjects without prior approval by the IRB/IEC.

The clinical trial: homogeneous population. A fundamental principle of conducting multi-center clinical trials is the ability to pool data from various clinical settings, allowing the sponsor to draw statistical and clinical conclusions based on the patient population defined by the entry criteria. This principle is only valid, however, if another fundamental principle of GCPs occurs. The protocol must be conducted in the same way at each multi-center site; a homogeneous patient population (based on criteria defined in the protocol) is treated and followed the same way (per protocol).

The ability to pool data is compromised if individual sites follow a different treatment regimen: treating subjects with a different dose, or enrolling subjects who are different from the proscribed and IRB/IEC-approved inclusion/exclusion criteria. The validity of statistical and clinical conclusions regarding safety and efficacy rests on the premise that a defined patient population was treated with the study drug and followed the evaluation schedule.

The study subjects. The trial cannot begin without documented approval for the protocol by the IRB or IEC. Per this approval, the IRB/IEC has reviewed the risk/benefit profile of the protocol, and has granted approval for the trial to begin based on this assessment and for a particular patient population (described by the inclusion/exclusion criteria). Enrolling a subject who does not meet the IRB-approved inclusion/exclusion criteria is putting the prospective patient at risk; the subject is different from the IRB/IEC-approved patient population meant to be eligible for enrollment. The informed consent form may not accurately reflect the risks/benefits for the subject, since the consent form was designed (by the sponsor) and approved (by the IRB/IEC) for the patient population described by the protocol's inclusion/exclusion criteria.

The sponsor. When the study is complete, the sponsor will need to justify and explain why ineligible subjects were enrolled into the trial. If many subjects were enrolled who did not meet inclusion/exclusion criteria (and who did not receive IRB/IEC approval prior to enrollment), the FDA (or another regulatory agency) may conclude that the intended population was not treated, as designed. Once the data is submitted to the regulatory agency, the regulatory authority reviewing the data can request the data set be re-run to exclude the ineligible subjects from the efficacy tabulations and conclusions; after all, those patients were not intended to be treated in the protocol. The exclusion of even small numbers of subjects from this data set may have devastating effects on the trial, and may even result in the trial being conducted again because the loss of these subjects has reduced the statistical power of the trial. Sponsors should be aware that each protocol waiver for inclusion/exclusion criteria might result in the data for that subject being used only for safety evaluations and conclusions, and excluded for efficacy evaluations and conclusions.

Also, when the trial is complete and the data submitted, the enrollment of a large number of ineligible subjects might confound the conclusions of the trial and negatively affect the efficacy claims and the label text. Sensitivity analyses, for example, may reveal that outcome measures are different for subjects receiving waivers than those not receiving waivers, putting the results at risk. The regulatory agency may conclude that the data are unreliable and reject the trial.

The regulatory authorities' position on waivers

The regulatory requirements of the country in which the clinical trial is conducted must be followed. As mentioned above, US regulations and EU guidelines require that subjects not meeting inclusion/exclusion criteria cannot be enrolled unless there is prior IRB/IEC approval. In India, sponsors deciding to enroll a subject not meeting inclusion/exclusion criteria must receive prior permission from India's regulatory authority, the Drug Controller General of India (DCGI).⁴

Within the United States, the FDA's stance on sponsor-provided waivers for subjects not meeting inclusion/exclusion criteria can be gleaned from warning letters which are published on the FDA website, www.fda.gov. Outside of the United States, warning letters, or their equivalent, are not publicly available. A review of warning letters⁵ to clinical investigators did not identify any letters that cited an investigator for enrolling an ineligible subject who was granted a sponsor waiver, for whom the investigator did not get IRB/IEC approval prior to enrollment. Many letters were identified that cited investigators (Branitz, Bukowski, Chaganti, Chappel, Desai, Gray, Hernandez, Hijazi, Linzer, Paulson, Stewart, Summers, and Zosa) or sponsors (Johnson & Johnson [Grosser], Feins) for enrolling subjects who did not meet inclusion/exclusion criteria; for the letters to investigators, these do not reference whether a waiver was requested by the investigator or granted by the sponsor. All of the investigators were cited for failing "to conduct the study or ensure it was conducted according to the investigational plan" (21CFR312.60). The sponsors were cited for not conducting the trial in accordance with the general investigational plan and protocols, as specified in the IND (21CFR312.50, Grosser), or failure to obtain approval prior to implementing a change to the investigational plan (21CFR812.35, Feins). For four letters (Bukowski, Croley, Hijazi, and Stewart), FDA states that the investigator had enrolled subjects without receiving a sponsor waiver; there is no mention of the need for IRB/IEC approval prior to enrollment. In one letter (Stewart), the investigators are cited for not obtaining a waiver until after the subject had been enrolled; again, there is no mention that an IRB/IEC approval had been obtained prior to the subject's enrollment. In another letter (Boden) for a trial involving a medical device, the FDA cites Cordis Corporation for not having a standard operating procedure to track the granting of waivers or protocol deviations, and for not describing the protocol deviation in the pre-market approval application. In the letter, the FDA states "As a sponsor, it is also your responsibility to ensure appropriate procedures are in place for the granting and documentation of waivers and protocol deviations and that these procedures are followed by investigators." Of these warning letters, only three (Croley, Devinney, and Feins—all subject to device regulations) specifically mentions the requirement to have IRB approval prior to the enrollment of a study subject who does not meet inclusion/exclusion criteria:

Furthermore, a waiver from the sponsor was not obtained prior to the implantation of the second study [redacted] in that subject. According to the protocol, all implants must be performed with [redacted], must have the full knowledge and consent of a recognized IRB, and must conform with the sponsor's FDA-approved investigational plan. (Croley)

You enrolled subjects that did not meet the inclusion/exclusion criteria. In addition, there is no documentation the sponsor and IRB approved of these deviations. You must obtain prior approval by the sponsor for changes in or deviations from a plan, and, in addition, by FDA and IRB when the changes or deviations may affect the soundness of the plan or the rights, safety, or welfare of human subjects (21CFR812.150(a)(4)). (Devinney)

As a sponsor, you are required to obtain FDA approval and IRB approval when appropriate, and prior to implementing a change to the investigational plan (21CFR812.35). (Feins)

Although warning letters, or their equivalent, are not available outside the United States, some guidance on the acceptability of using waivers for inclusion/exclusion criteria is available at a Q&A website⁶ sponsored by the European Medicines Agency (EMA). To the question "Can a sponsor prospectively approve deviations (so-called "protocol waivers") from the inclusion/exclusion criteria of the approved protocol without additional approval of the Ethics Committee and competent regulatory authority?" the answer (in part) was:

Adherence to the protocol is a fundamental part of the conduct of a clinical study. Any significant change to the protocol should be submitted as an amendment to the competent regulatory authority and Ethics Committee. Significant changes to the protocol include any change in inclusion and exclusion criteria, addition or deletion of tests, dosing duration of treatment etc.... Deviations from the inclusion/exclusion criteria of the protocol might erode the scientific and ethical value of the protocol and its authorization and might have an impact on the processes put in place for the care and safety of the study subjects.

Clinical practice vs. clinical research

Requesting (by the principal investigator) and granting (by the sponsor) protocol waivers for inclusion/exclusion criteria, may be based on any of the following motivations:

• The study drug is a good treatment for the subject who might not have another treatment option.

• The deviation is considered "safe" for the subject and would not influence the scientific integrity of the trial data.

• The trial needs to enroll as many subjects as quickly as possible.

• The sponsor wishes to maintain a good relationship with the investigator.

• The sponsor is interested in collecting data on the use of the drug in a population that will ultimately use the drug once it is marketed.

The first motivation blurs the line between "clinical practice" and "clinical research." It may be true that the enrollment of an ineligible subject into a clinical trial may benefit the subject. However, the purpose of clinical research is not to make subjects well; the purpose of clinical research is to collect good data. Although enrolling ineligible subjects into a clinical trial may be altruistically commendable, it has the potential to jeopardize the validity of the entire trial. When an investigator signs the FDA 1572, protocol signature page, and study contract, he/she agrees to put aside his/her clinical practice "hat" and to don his/her clinical research "hat," meaning the principal investigator will follow the IRB/IEC-approved protocol, even if following the protocol is not consistent with his/her routine medical practices. The only exception to this promise and commitment is to eliminate immediate hazards to the subjects; enrolling patients outside the trial's inclusion/exclusion criteria simply to provide a treatment that might benefit the patient does not meet this critical standard.

In the second motivation, the medical monitor might conclude that the deviation from an inclusion/exclusion criterion would be "safe" for that individual subject, thus justifying a waiver so the subject may be enrolled into the trial. However, this conclusion raises a basic and fundamental question on the reason the inclusion/exclusion criteria was written in the first place. If providing a waiver is considered a "safe" option, doesn't this suggest that the trial is excluding other potentially eligible subjects as well? Wouldn't this fact question the need for the restrictive nature of the inclusion/exclusion criterion, and would argue to amend the protocol to broaden that criterion so more patients could become subjects? The motivation for some sponsors to amend a protocol is a high incidence of waiver requests for a particular inclusion/exclusion criterion, or complaints by investigators that an specific criterion is causing many subjects to be ineligible.

The third motivation puts the sponsor's interests above the protocol's intent; the IRB/IEC's purpose; and, potentially, the study subject's rights, welfare, and well-being, and opens the sponsor to a criticism of a conflict of interest.

In the forth motivation, per the regulations (21CFR312.56), sponsors have two options when they discover that an investigator is not complying with the protocol: secure compliance or discontinue the trial at that site.

In the fifth motivation, expanding the inclusion/exclusion criteria and, by extension, the intended patient population, has the potential to confound the interpretation of the data and jeopardize its usefulness and validity. Collecting data on populations of subjects other than the one prescribed by the protocol typically is reserved for compassionate use or Phase IV protocols.

The role of the IRB/IEC

As an auditor, I have had the opportunity to evaluate the requirements imposed by IRBs/IECs on principal investigators for reporting protocol deviations, including the enrollment of ineligible subjects (based on the inclusion/exclusion criteria). These reporting requirements differ markedly between IRBs/IECs, and appear to fall into three categories:

• For some IRBs/IECs, the requirements are consistent with the regulations: principal investigators must report proposed changes to study procedures, including changes to inclusion/exclusion criteria, and obtain approval to enroll ineligible subjects (based on inclusion/exclusion criteria) before the subject is enrolled.

• Other IRBs/IECs are tolerant of sponsor-provided protocol waivers that permit the enrollment of ineligible subjects into clinical trials. One IRB states in their guidance that principal investigators need not report the enrollment of ineligible subjects (based on inclusion/exclusion criteria) if the sponsor has provided a waiver. The IRB has transferred responsibility to the sponsor for a fundamental decision reserved by an independent IRB, namely, the assessment of the risk/benefit ratio for enrolling study subjects.

• Other IRBs/IECs are unaware, per internal policies, of the frequency protocol waivers being issued and the number of ineligible subjects who are being enrolled. In web-posted instructions, one prominent central IRB states, "The [X] IRB does not require that protocol deviations be reported, unless they constitute an unanticipated problem." The policy clearly discourages investigators from even reporting that a waiver had been granted by the sponsor to enroll an ineligible subject, since the deviations are not "unanticipated." The IRB, then, would be unaware of the number of ineligible subjects who are enrolled into a trial for which it has a responsibility to provide oversight.

IRBs should be critical of sponsors that give permission to principal investigators to violate the protocol in ways that affect the rights, welfare, and well-being of study subjects, and should be critical of principal investigators who inform the IRB/IEC that a subject not meeting inclusion/exclusion criteria has been enrolled without prior IRB/IEC review and approval. IRBs/IECs should ensure their internal procedures remind principal investigators that it is the IRB's responsibility, not the sponsor's, to provide an independent assessment of the risk/benefits, including the enrollment of ineligible subjects into clinical trials. All deviations should be reported to the IRB/IEC when ineligible subjects are enrolled into a clinical trial, whether or not there is a sponsor waiver. IRBs/IECs should clearly remind principal investigators that subjects not meeting inclusion/exclusion criteria must receive IRB/IEC approval prior to their enrollment into the trial.

Enrolling subjects outside the criteria

Some sponsors and principal investigators insist that there is value in enrolling subjects outside the prescribed inclusion/exclusion criteria, and there is no time to obtain IRB/IEC approval prior to enrollment; in these cases, the principal investigator and sponsor accept the inherent risks to themselves, the subjects, and the trial. One sponsor has a policy that waivers are allowed only once for any inclusion/exclusion criterion; any other requests for waivers for that inclusion/exclusion criterion are not allowed, and the protocol would need to be amended.

As stated initially, it appears to be common practice for principal investigators to alert the IRB/IEC after the subject has been dosed, notifying the IRB/IEC that a protocol violation has occurred, and a subject who did not meet inclusion/exclusion criteria has been enrolled (viz. dosed). Although this practice is acceptable to some IRBs/IECs and has not been routinely cited as unacceptable in warning letters issued by the FDA, it is not consistent with GCPs.

The sponsor has two options for enrolling subjects into a clinical trial who do not meet the defined criteria:

Amend the protocol. As stated earlier, if a sponsor receives many requests for waivers for a particular inclusion/exclusion criterion, the sponsor should re-think the need for the restrictive language. If appropriate, the protocol could be amended, thus broadening the patient pool eligible for study enrollment.

Implement a compassionate use protocol. A cosmpassionate use protocol could have much broader inclusion/exclusion criteria. This option is particularly beneficial to obtain safety, as well as efficacy, data on a more heterogeneous population while not jeopardizing the validity of the clinical trial.

Therefore, the principal investigator has three options for treating a subject who does not meet the trial's criteria:

• Request a compassionate use protocol from the sponsor.

• Obtain a waiver from the sponsor and obtain IRB/IEC approval prior to dosing the subject.

• Provide treatment with the drug once the clinical trial is complete, the NDA has been submitted and approved, and the drug is commercially available.

Conclusion and summary

Enrolling ineligible subjects into a clinical trial without prior approval by an IRB/IEC represents risks to the subject, the principal investigator, the sponsor, and the trial. Sponsors that provide waivers for inclusion/exclusion criteria should ensure subjects are not enrolled into a trial until the enrollment of that subject is approved by an IRB/IEC. This requirement should be emphasized at the investigator's meeting and throughout the clinical trial. Sponsors should ensure documentation of the waiver is contained within the site's and sponsor's trial master file, and should have standard operating procedures for the granting and documentation of waivers and protocol deviations, to ensure that these procedures are followed by investigators.

Based on sample warning letters that involved the enrollment into clinical trial of subjects not meeting inclusion/exclusion criteria, the FDA does not appear to be strictly enforcing the GCP requirement for ensuring IRB/IEC approval prior to the enrollment of these subjects, but does appear to require documented evidence that the investigator has received a waiver by the sponsor.

Investigators and sponsors should follow the local regulations before enrolling subjects into a clinical trial who do not meet the inclusion/exclusion criteria of the IRB/IEC-approved protocol. IRBs/IECs should clearly enunciate their requirement that subjects not meeting inclusion/exclusion criteria must receive IRB/IEC approval prior to the enrollment of those subjects into the trial.

George Nauyok, Principal, Quality Auditing and Training, 167 Cresta Vista Drive, San Francisco, CA 94127, e-mail: gnsanfrancisco@hotmail.com

References

1. Responsibilities of Sponsors and Investigators, Sub-part D, 21CFR312.

2. Good Clinical Practices, Consolidated Guidance, ICH E6.

3. Investigational Device Exemptions, 21CFR812.

4. "India Halts Wyeth Vaccine Trial Over Infant Death," last modified November 14th, 2008, http://www.pharmalot.com/2008/11/india-halts-wyeth-vaccine-trial-over-infant-death//.

5. FDA Warning Letters, fda.gov/ICECI/EnforcementActions/WarningLetters/default.htm: Thomas Boden, issued July 7 2004., Bruce Branitz, issued April 9, 2009 (09-HFD-45-03-03)., Ronald Bukowski, issued March 30, 2009 (09-HFD-45-03-02).Surendra Chaganti, issued October 26, 2007., Christopher Chappel, issued February 2, 2009 (09-HFD-45-01-02)., Thomas Croley, issued May 14, 2004., Virendra Desai, issued March 2, 2009 (09-HFD-45-02-03)., Dennis Devinney, issued April 21, 2006., Neil Feins, issued May 20, 2009., William Gray, issued March 16, 2005., Karen Grosser (Johnson & Johnson), issued August 10, 2009 (09-HFD-45-07-02)., Francisco Hernandez, issued April 20, 2009 (09-HFD-45-04-02)., Saad Hijazi, issued May 19, 2009 (09-HFE-45-05-01)., Dov Linzer, issued June 12, 2009 (09-HFD-45-06-01)., Daniel Paulson, issued March 11, 2009 (09-HFD-45-03-01)., Lynette Stewart, issued January 21, 2009 (09-HFD-45-01-01)., Timothy Summers, issued February 24, 2010 (10-HFD-45-01-02)., Noli Zosa, issued June 16, 2009 (09-HFD-45-06-0).

6. "Q&A: Good Clinical Practice (GCP),"European Medicines Agency, http://www.ema.europa.eu/Inspections/GCPQaA.htm.