Study Enrollment When “Preconsent” Is Utilized for a Randomized Clinical Trial of Two Treatments for Acute Agitation in the Emergency Department

Acute agitation in the emergency department (ED) is common, representing up to 2.6% of all ED visits,1 and represents a danger to both patients and their caregivers.2 Medication is often needed to keep patients and caregivers safe while simultaneously facilitating a rapid assessment to rule out underlying critical illness.3 Few high-quality randomized trials have evaluated the optimal drug for this vulnerable population,4, 5 and much of the existing evidence utilized by expert consensus panels is either retrospective6-9 or observational in nature.10, 11 Thus, emergency physicians are subject to a dearth of high-quality evidence upon which to base pharmacotherapeutic choices for agitated patients.

In the United States, investigators may conduct studies without informed consent under one of two federal regulations: Waiver of Informed Consent (WIC; 46 CFR 45.116), and Exception From Informed Consent (EFIC; 21 CFR 50.24). Under WIC regulations the research must involve no more than minimal risk, meaning that the probability and magnitude of harm or discomfort anticipated in the research is not greater than encountered in routine medical examination and testing (21 CFR 50.3[k]).12 WIC has been utilized for many prospective comparative effectiveness investigations comparing standards of care, such as medical devices or clinical protocols.13-20 WIC trials have not generally been used for comparative drug investigations, such as randomized, blinded drug intervention trials. If such trials are intended, EFIC may be pursued. Per the current federal guidelines, EFIC trials must involve patients with a “life-threatening condition” where the proposed treatments are “unsatisfactory or unproven.”12 EFIC trials are more strictly regulated by federal agencies and have extensive and specific requirements that the investigators must meet.12

Quality research on treating agitation has been profoundly limited by these issues surrounding consent. Because it is well established that the majority of agitated patients in the ED are incapable of giving meaningful informed consent,21, 22 alternative consent methods must be employed. Agitation trials may not qualify for EFIC because agitation is not necessarily a “life-threatening condition.”4 WIC may not be appropriate because a blinded randomized drug intervention trial historically has been unlikely to be deemed minimal risk by local institutional review boards (IRBs).

Another option for consent in agitation trials that has been suggested by federal agencies is a “preconsent” method, where consent is sought in advance of agitation occurring. With this method, after preconsent is obtained, when patients present at a later time or date with agitation, they can be enrolled in the trial. To our knowledge, preconsent (or a similarly named method) has not been used to study agitation in the ED. We attempted this method of consent in a randomized double-dummy trial of inhaled loxapine versus intramuscular haloperidol plus lorazepam for acute agitation. The purpose of this study is to report our experience with obtaining informed consent with preconsent methodology.

Methods

Results

Over 67 days, 1,461 patients were screened in the ED at site A: 269 had bipolar I or schizophrenia, 194 of whom had a contraindication to inhaled loxapine (most commonly COPD), leaving 75 eligible patients. In 43 of these patients, we successfully obtained preconsent. An additional four patients who had not preconsented were consented in real time: three were consented by a LAR and one patient was deemed to have capacity while agitated. This resulted in a total of 47 patients consented for the study.

Of these 47 patients, 12 patients were removed from the study prior to completion of enrollment and study drug administration: 10 patients had unrecognized exclusion criteria for inhaled loxapine, one patient later asked to be removed after preconsenting, and 1 surrogate immediately revoked consent real time.

Only two patients in total received study medication; neither was consented by preconsent: one was consented via a LAR the day of enrollment, and the other was mildly agitated and was determined to have capacity to consent. The remaining patient with a valid surrogate consent did not receive study medication. Enrollment is fully delineated in Figure 1.

Discussion

In this randomized, blinded, clinical trial of two treatments for acute agitation secondary to bipolar I or schizophrenia, despite screening > 1,400 patients and obtaining preconsent in 43 patients, we were unable to successfully enroll a single patient using preconsent methods. Only two patients were enrolled in the trial: one who was deemed to have capacity to consent and one via a LAR. Because the study was slow to launch and progress for various reasons, including IRB concerns about the consent mechanism, the trial sponsor revoked funding, forcing the trial to close. These data suggest that utilization of preconsent may not be feasible for trials of agents to treat acute agitation in the ED, despite recommendations to use this consent method by federal agencies.

While patients have been successfully enrolled in randomized trials of parenteral treatments for agitation in either medical or psychiatric inpatient units using traditional methods of consent (either the patient consenting themselves or via LAR),25 enrollment in trials via these mechanisms in the ED is likely not feasible. First, while some mildly agitated patients may have capacity to consent, the preponderance of evidence suggests most agitated patients in the ED do not have capacity. A common reason is that, unlike inpatients, ED patients with acute agitation are frequently acutely intoxicated.1 Some have proposed the use of brief consent tools, such as the University of California at San Diego Brief Assessment of Capacity to Consent (UBACC) to mitigate this issue. One convenience sample found that when using the UBACC, 16 of 19 intoxicated ED patients were deemed to have capacity to consent.26 These patients had multiple ED visits and were studied as a cohort highly likely to return to the ED. The UBACC was applied when the patients were clinically sober after observation in the ED but still had elevated ethanol concentrations. This observation period limits the generalizability of these findings when applied to patients presenting with acute agitation, which, in the ED, is frequently due to acute intoxication.1 The UBACC when subsequently applied to a larger random sample of 415 acutely intoxicated ED patients shortly after a medical screening examination on ED arrival found that only 16 of 415 (3.9%) intoxicated ED patients had capacity to consent; furthermore, of these 16 patients, eight did not recall taking the assessment once clinically sober, suggesting that consent tools have limited utility in the intoxicated, agitated ED patient.27 Use of LARs to obtain consent may be possible for studies of some agitated ED patients; however, obtaining consent from a LAR for a patient with agitation who requires immediate treatment, such as in excited delirium syndrome, is not feasible because of the immediate need to treat the patient. Furthermore enrollment solely via LARs is likely to result in a smaller sample skewed toward less ill patients with good social support, severely limiting the generalizability of the data. For example, in a previous randomized trial of three parenteral agents for acute agitation in the ED, of 144 enrolled patients only three were successfully enrolled via a LAR.5

To obtain high-quality data and better inform practice, randomized trials of a broad range of agitated patients, including those who cannot consent, are needed. One mechanism to accomplish this is utilization of EFIC (21 CFR 50.24) for enrollment. EFIC trials have numerous requirements, including the fact that the condition to be studied must be life-threatening and that current treatments are unproven or unsatisfactory.21 Since EFIC trials are FDA regulated, investigators must also notify the FDA by filing an investigational new drug (IND) application after local IRB approval and completion of the other elements of EFIC, such as community consultation and public disclosure.28 FDA has been inconsistent on whether or not they deem agitation in the ED to be a life-threatening condition. Martel et al.5 conducted a trial of three common intramuscular drugs on agitated ED patients utilizing EFIC; however, when researchers from the same institution approached the FDA about a follow-up trial, FDA did not approve the IND, citing insufficient evidence that the patients could not provide informed consent4 and that the researchers did not provide adequate evidence that agitation was a life-threatening condition. Of note, the FDA suggested the researchers pursue consenting “psychiatric patients highly prone to agitation” when they have capacity at another time, effectively suggesting preconsent be applied. Setting aside the fact that a substantial portion of ED agitation is due to acute intoxication, not psychiatric disease, this method of preconsent, in our current study, we found to be infeasible. FDA's criticism of the life-threatening nature of agitation may be a more accurate rationale for not approving an EFIC trial; the life-threatening nature of agitation exists on a spectrum. While the most severe form of agitation, excited delirium syndrome, has an associated mortality rate of 8% to 16%,29 a recent prevalence study of 1,146 agitated ED patients observed no deaths in the ED.1 As currently written, there is a gap within the EFIC regulations for completion of randomized trials on therapies for agitation in the ED, with the possible exception of excited delirium syndrome, until the “life-threatening” clause is revisited.

Complicating the matter of using EFIC to study agitation further is the concept of treatments being “unproven or unsatisfactory.” Although scant comparative effectiveness trials of therapies for agitation exist, it is likely that over time supportive care has generally improved for this patient population. A parallel likely exists in a similar patient population with agitation: those with delirium tremens. The mortality rate for delirium tremens in the early 20th century was estimated at 70%.30 As demonstrated in a retrospective analysis of patients with delirium tremens at an urban city hospital, mortality from delirium tremens dropped from 52% to 14% over a 20-year period from 1915 to 1935, largely due to advances in general supportive care.30 However, it was not until a landmark double-blind study of four common agents used for alcohol withdrawal established the superiority of benzodiazepines for alcohol withdrawal31 that mortality for delirium tremens began to fall again to the 1% to 4% rate observed currently.32 It is likely the care of agitated patients in the ED has undergone a similar evolution as emergency medicine has become hypervigilant about the underlying critical illness agitation may represent.3 It is now standard of care that agitated patients are brought immediately to the ED by police and EMS and when indicated receive rapid sedation to prevent physical injury and metabolic derangements. Anecdotally, in the practice at site B, metabolic acidosis and restraint-related injuries were far more common before the more liberal use of both antipsychotics and sedatives for agitated patients.33 This overall improvement in supportive care may be the reasoning for the FDA's changed position on acute agitation as a life-threatening condition that qualifies for EFIC research. However, unlike delirium tremens, high-quality randomized trials have not yet defined which of several, commonly used treatments for agitation in the ED are superior. While medications such as benzodiazepines, antipsychotics, and ketamine all appear effective4, 11, 34 (and, perhaps, “proven and satisfactory”), some acutely agitated patients, such as those with sympathomimetic or novel psychoactive drug intoxication, still die.35, 36 Great progress has been made thus far in the care of agitated patients; however, without high-quality randomized trials to better inform practice, we will not take the next step to advance care as we once did with delirium tremens.

Recently, some comparative effectiveness trials for agitation treatments have been published that utilized WIC (46 CFR 45.116).4, 37 This method observed open-label changes to ED or EMS clinical protocols instituted by departmental leadership, in which one drug is recommended as the first-line agent for a period of time, followed by a change to a different first-line drug for a second period of time. Since this type of research is merely observation of the outcomes of the clinical protocols, which provided recommendations for all patients without regard to whether they were enrolled in the WIC study or not, an IRB may determine these studies do not exceed minimal risk and are therefore eligible for WIC. While this pragmatic method allows for gathering prospective data, it has significant limitations. First, it involves data collection in a nonblinded manner. Second, “before–after” designs are far more subject to secular trends, such as the sudden introduction of an abusable novel psychoactive substance in a community,38 that may disproportionately affect one arm in an adverse manner. Third, while the research activities (observational data collection) are minimal risk, the design of alternating clinical protocols offers no additional protections to the subjects compared to a blinded, randomized trial such as those conducted under EFIC, as in both approaches the patient receives a medication per protocol rather than by physician decision, although the protocol can be overridden in both designs. These recently published before–after studies, from a methodologic standpoint, are clearly inferior to higher-quality blinded randomized trials. A collaborative approach from investigators with expertise in emergency medicine, critical care medicine, medical toxicology, and emergency psychiatry, in conjunction with regulatory bodies such as the FDA and local IRBs, will be needed to design future research protocols to better study this patient population.

Fortunately there exists a potential pathway to conduct higher-quality comparative effectiveness trials of drugs in patients who cannot provide consent that are not necessarily in an immediate life-threatening condition, which encompasses the majority of agitated ED patients needing parenteral sedation. The 21st Century Cures Act, passed December 13, 2016, provides the FDA with the authority to permit an EFIC when proposed clinical trials poses no more than minimal risk to human subjects, provided that appropriate safeguards are in place to protect their rights, safety, and welfare.39 This new authority allows local IRBs to approve waived consent research for minimal risk comparative-effectiveness drug studies of two accepted standards of care, such as our proposed trial of inhaled loxapine versus haloperidol + lorazepam. In July 2017, the FDA issued guidance that until such regulations are finalized, FDA does not intend to object to a local IRB approving a waived consent study when the IRB finds and documents that traditional WIC criteria have been met (see Table 1).39 To our knowledge no such trial has yet been conducted under this new FDA authority, although one small randomized trial of midazolam versus haloperidol for prehospital agitation was conducted under WIC criteria prior to passage of the 21st Century Cures Act.40

Limitations

This study has several limitations. Specifically the short study duration does not reflect the overall enrollment rate had the study been able to be conducted over the full 2-year period. The termination of the study before site B could launch also limits generalizability; in terms of successful preconsent enrollments we report only single-center data. However, the study period was truncated by the sponsor partly because of low enrollment, highlighting the limitation of this method. Substantial resources were committed to this trial with no successful preconsent enrollments. Even if preconsent were viable, the resources required to enroll enough patients would likely make trials for agitation treatments infeasible.

Conclusions

Utilization of preconsent to enroll patients in a randomized trial of parenteral treatments for acute agitation in the ED requires substantial resources and may not be feasible. New FDA authority may allow for Waiver of Informed Consent (45 CFR 46.116) for minimal risk clinical trials of drugs. Trials involving the highest-risk agitated patients, such as those with excited delirium, may qualify for Exception From Informed Consent (21 CFR 50.24). Future comparative effectiveness trials on treatments for acute agitation in the ED are needed to advance care for this vulnerable patient population.