Published in the October 2015 issue of Today’s Hospitalist
DEPENDING ON WHERE YOU PRACTICE, as many as 5% of your patients will need help getting through alcohol withdrawal. The good news is that the days of winging it “with individual physicians managing symptoms based on individual preferences or what their attendings taught them “are pretty much over.
Instead, the standard of care now favors use of the revised Clinical Institute Withdrawal Assessment of Alcohol Scale (CIWA) scoring system, along with benzodiazepine (typically lorazepam) dosing protocols.
But while CIWA and benzodiazepines are now mainstays in withdrawal management, many hospitalists and researchers have issues with both. As a result, physicians are testing other scoring systems, dosing protocols and medications.
“I worry that some patients with anxiety issues end up scoring high on CIWA.”
~ Anthony Worsham, MD, University of New Mexico
Here’s a look at where withdrawal management in the hospital is heading.
CIWA: too many elements
When it comes to CIWA, physicians complain that the relatively complex scoring system contains too many subjective elements. They say that CIWA scoring “which incorporates 10 domains “and evaluation are too labor-intensive for nurses and can be interpreted too broadly.
“There’s a lot of inter-observer variability,” says Jeffrey Lyon, MD, a hospitalist who serves as patient safety officer for Essentia Health in Duluth, Minn. “People often don’t score the same patient the same way, and the score relies heavily on patient-reported symptom severity.”
Anthony Worsham, MD, an assistant professor in the hospital medicine division at the University of New Mexico in Albuquerque, points to another possible problem: Patient-reported symptoms can be related to something else.
“I worry that some patients with anxiety issues end up scoring high on CIWA,” Dr. Worsham says. “We’re medicating them for alcohol withdrawal when in reality, they’ve got something else going on.”
Another issue with CIWA: Some people in severe withdrawal might not be able to articulate how they’re doing. “Patients who have a CIWA score over 20 are so confused and agitated that they can’t describe their symptoms,” Dr. Worsham notes. “We switched away from CIWA in those patients and now use Riker scoring instead.” That system, which is used in the ICU for patients on a ventilator or who are otherwise unresponsive or agitated, is based on observation.
John Dickens, MD, MPH, hospitalist medical director of the Central Maine ACO and chief transformation officer at Central Maine Medical Center in Lewiston, Maine, thinks that CIWA works well enough most of the time. But comorbidities and nurses’ inexperience with CIWA scoring can complicate treatment decisions.
“If you have a patient with dementia who is also a drinker, that patient is never going to be oriented,” Dr. Dickens says. “You have to take into account what their baseline is when you score with CIWA, and it takes a savvy person to recognize that. A patient with a Parkinson’s tremor can’t necessarily say if the tremor is different or worse.”
Further, when nurses aren’t used to scoring withdrawal symptoms, treatment might be way off base. “Around the country, we have less experienced nurses than we did 10 or 15 years ago, and illness severity in hospitals is increasing,” Dr. Dickens points out. “I have actually worked with nurses who have never used CIWA before.”
CIWA’s reactive approach
Then there’s this problem: Patients don’t start receiving medications until their symptoms translate into a “treatable” CIWA score, typically 8 or higher. Sometimes, that’s too late to stop the onslaught of physiological and autonomic effects that can land patients in the ICU.
“We say, ‘We will monitor you and wait until you get into withdrawal, and then we will start giving you medication, ‘ ” says Austin Voigt, MD, a hospitalist at Virginia Tech Carilion in Roanoke, Va. “That can really get out of hand quickly.” Dr. Voigt is leading a randomized trial looking at whether continuing moderate alcohol consumption in the hospital prevents or minimizes withdrawal severity better than benzodiazepines. (See “Should you give patients alcohol in the hospital?“)
Those concerns prompted Allina Health System in Minneapolis to change both its scoring system and dosing approach. In 2014, two Allina hospitals switched from CIWA to an internally developed protocol based on the (Minnesota Detoxification Scale) designed by the Minnesota VA. The MINDS scale includes far fewer domains than CIWA, and it calls for early scheduled doses of benzodiazepines to head off severe withdrawal.
According to David Beddow, MD, Allina’s regional medical director for hospitalist services, the MINDS scoring system is more objective than CIWA because “it’s more focused on signs: diastolic BP, pulse, sweating and autonomic discharge. The idea is to score for very early withdrawal signs and to give large doses of oral Valium or IV if patients can’t take oral.”
High doses, early on
Under the Allina protocol, patients receive 20 to 80 milligrams of diazepam every hour for two doses. Dosing then varies based on MINDS scores, which include vital signs.
“Dosing is pretty high, pretty quick,” Dr. Beddow explains. “We also don’t use lorazepam here unless patients have advanced liver disease. Ativan has the potential for withdrawal syndrome itself.”
When Allina first introduced its protocol, some hospitalists cited concerns about the higher benzodiazepine doses. But in the year since the protocol was implemented, there has been no increase in safety issues or over-sedation. To the contrary, transfers to the ICU have gone down significantly, Dr. Beddow notes.
And Unity Hospital, which is located in Fridley, Minn., has seen a drop in “green alerts” for security personnel.
Readmissions and returns to the ER after discharge have gone down as well.
Allina’s nurses have for the most part been pleased. ”The nurses like this better than CIWA because it’s objective and dosing correlates directly with a number they get after scoring,” Dr. Beddow says. “The protocol hasn’t changed the dynamic of how we treat alcohol withdrawal, but it seems to be improving outcomes.”
Other CIWA alternatives are also gaining traction. Hospitalists at St. Joseph’s Hospital Health Center in Syracuse, N.Y., are moving away from CIWA to the Glasgow Modified Alcohol Withdrawal Scale, which was developed in the U.K. For patients at high risk of severe withdrawal, treatment starts with a fixed oral dose of diazepam: 20 milligrams every four hours for the first 24 hours. It then goes down in 5-milligram increments every four hours until symptoms resolve.
Like the MINDS protocol, the Glasgow guideline uses a simple system. A 0 to 2 point scoring system looks at five domains: tremor, sweating, hallucination, orientation and agitation. The protocol incorporates fixed and symptom-triggered dosing.
“We were looking for an easier way to do things and to prevent treatment failures and ICU transfers,” explains James Leyhane, MD, hospitalist medical director at St. Joseph’s. “We brought this to committee meetings, and it got broad support from hospitalists and nursing, as well as psychiatry and pharmacy. ”
The hospitalists would be using the Glasgow protocol now if St. Joseph’s hadn’t switched EHRs in 2014, putting the protocol on the waiting list of order sets that have to be implemented manually.
“We’re still using CIWA while we’re waiting for implementation, but we’re encouraging a fixed-plus-symptom-relief dosing approach,” says Dr. Leyhane. “We’ve thought that what we were giving on the floors was too limited.”
The hospital also recently started allowing hospitalists to give patients in withdrawal a few IV doses of lorazepam to reduce the potential for worsening symptoms and ICU transfers.
Shifting drug choices
While lorazepam remains the treatment of choice, some hospitalist programs are eyeing other alternatives. In Duluth, Dr. Lyon has studied baclofen as an alternative or adjunct to lorazepam and prefers this central-nervous system agent to lorazepam-only treatment for several reasons. For one, baclofen as an adjunct appears to decrease the total requirement for benzodiazepines.
“We found an order-of-magnitude difference, not a subtle one, in the amount of benzodiazepine use,” says Dr. Lyon, referring to his study in the October 2011 issue of the Journal of Hospital Medicine that compared lorazepam and placebo to a combination of lorazepam and baclofen (10 milligrams, three times daily). That study found total cumulative lorazepam doses of up to 1,035 milligrams in the lorazepam-placebo arm in the first 72 hours after hospitalization vs. 39 milligrams total in the baclofen group. European studies have produced similar results.
As Dr. Lyon points out, less lorazepam is associated with shorter lengths of stay, fewer ICU transfers, and fewer patients on ventilators or with aspiration pneumonia. Baclofen “activates a gaba-B receptor that isn’t down-regulated in chronic alcoholism,” he explains, “so you can use it in low doses and get the same effect as with higher benzodiazepine doses.”
Aggressive dosing finds favor
While some hospitalists are experimenting with newer agents or non-benzodiazepines, others are sticking with the tried-and-trues. The University of New Mexico program, for instance, still uses chlordiazepoxide (Librium) in starting doses of 25, 50, 75 or 100 milligrams, depending on CIWA scores, as first-line treatment. But hospitalists turn to IV lorazepam, 1 or 2 milligrams PRN, for patients who either can’t take oral Librium or whose symptoms keep spiking on Librium.
“We have experience using Ativan infusions for brief periods “six to eight hours “and we think that’s kept some patients out of the ICU,” says John Rush Pierce, MD, MPH, the hospital medicine division’s director of research and scholarly activity. “We haven’t had any safety problems, but we’ve monitored these patients very closely.”
Patients with CIWA scores in the 20-plus range receive a 4-milligram lorazepam bolus, followed by an additional 6-milligram bolus “and up to 8 milligrams “if scores stay above 20 after the second or third assessment, respectively. At that point, the Riker scale is invoked.
This aggressive protocol, which has been tweaked a few times since the group began using it in 2011, was based on real-life experience. Dr. Pierce points out that in New Mexico, the number of both DUIs and alcohol-related deaths are the highest in the country.
“Our initial thinking was that if patients had advanced liver disease, they weren’t going to metabolize these drugs as well and would develop over-sedation,” says Dr. Pierce. “Our experience was actually the opposite: People for whom we used low doses got into more trouble and developed more severe withdrawal. So we no longer modify the dose based on whether somebody has severe liver disease or not. Everybody gets the same dose.”
UNM’s protocol has been a big hit with nurses, Dr. Pierce adds, because it provides specific instructions like “evaluate in two hours,” not “evaluate in three to four hours.”
“Before, nurses were calling the physician all the time,” he says. “Now, nurses call only when they’ve used 300 milligrams of Librium in 24 hours or when the CIWA score remains over 20, and we need to decide whether to initiate IV Ativan.”
At Virginia Tech Carilion, hospitalists primarily use benzodiazepines. But individual physicians can vary the agents, notes James Franko, MD, the former hospitalist program director who is primarily focused on resident teaching.
Dr. Franko and his colleagues have used baclofen as an add-on when patients aren’t responding well to benzodiazepines and aren’t in the kind of danger that requires care in the ICU. “Another drug I have personally found successful is Tegretol [carbamazepine] for patients still having some withdrawal, but this is admittedly anecdotal success.”
Some of Dr. Franko’s colleagues have also reported success with Gabapentin to help patients who continue abstinence post-discharge.
“And most of us have used beta-blockers to blunt some of the sympathetic effects of withdrawal,” says Dr. Franko. “But what we all need is more guidance on when to escalate using or adding another drug such as Precedex. There’s not a good guideline for that.”
Dr. Lyon thinks that benzodiazepine treatment for withdrawal is here to stay because it’s safe at both low and high doses. Still, he urges hospitalists to consider using benzodiazepines in an accelerated symptom-triggered dosing protocol, in concert with adjunctive therapies.
“For patients in severe withdrawal, I would add in baclofen because I have experience with it,” he says. “And for the population that’s really resistant to treatment, consider using propofol.”
Bonnie Darves is a freelance health care writer based in Seattle.
USING ALCOHOL TO HEAD OFF severe withdrawal isn’t new, but a group of Virginia researchers is trying to gauge how effective that strategy is to prevent or minimize withdrawal.
Austin Voigt, MD, a hospitalist at Virginia Tech Carilion in Roanoke, Va., is leading a randomized trial to compare standard symptom-triggered, CIWA-based lorazepam treatment with oral alcohol. Patients with mild alcohol use disorder (based on DSM-V criteria) receive 14 grams of alcohol “a standard drink “every six hours. Those deemed moderate get a drink every four hours, while those with a severe disorder get one drink every two hours.
Alcohol is ordered as needed and patients can refuse drinks if they don’t want them. Four alcohol options are in the formulary.
Patients with a history of seizures or delirium tremens in previous withdrawal episodes are deemed moderate. For safety purposes, any patients in the alcohol arm whose CIWA scores start climbing quickly receive lorazepam.
While enrollment is in the early stages, CIWA scores have been lower in the alcohol arm than in the standard protocol group, and no patients in the alcohol group have required a transfer to the ICU. “I think we will have statistical significance when we’re done,” Dr. Voigt says.
He believes that imposing abstinence on patients admitted with coexisting alcohol disorders or withholding symptom prophylaxis is wrong-headed, both ethically and clinically. In his view, hospitalists managing withdrawal with benzodiazepines or other drugs are making presumptions about both patients’ wishes and clinicians’ appropriate role in treating alcohol-use disorders.
Individuals who go through even moderate alcohol withdrawal are at risk for “kindling” in the future, Dr. Voigt explains, resulting in a lower seizure threshold, and the potential for more severe withdrawal the next time. “It also makes it more difficult for people to stop drinking because the alcohol craving is intensified,” he says.
“I think it’s medical arrogance to say that when we hospitalize patients for conditions unrelated to their alcohol consumption, that we will save them from their alcoholism or alcohol-use disorder by mandating abrupt abstinence,” says Dr. Voight. “This is a lifelong disease for many people. We think that by drying them out for a few days, we will fix their acute medical needs and that, magically, when they’re discharged, they won’t have any drive or social pressures to drink. Not only is that impractical, it may be dangerous.