On the Reading of the Week

As we did last month with the benzodiazepine papers, the selection of this week’s Reading was made with the editorial board of the International Psychiatry Twitter Journal Club, allowing us to consider this paper here, and to continue the conversation on Twitter. And that conversation is going on today.

Bonus: the paper’s first author is participating in the Twitter discussion. And Dr. Maldonado will be presenting unpublished data, too. #Cool.

Interested? Follow @psychiatryJC.

On Alcohol

Molly. G. Purple passion.

Street drugs have a tendency to catch our attention. But in reality, what really plagues our society isn’t the latest designer drug reported on the evening news or the stuff discussed in case reports in addiction journals, but the drugs that are completely legal. In a 2002 analysis, the Canadian Centre on Substance Abuse pegged the total economic costs of substance abuse at $39.8 billion. Tobacco accounted for 42.7% of the total estimate; alcohol, 36.6%. In other words, the vast majority of our drug problem is cigs and booze.

Source: Canadian Centre on Substance Abuse

Let’s focus on alcohol.

As a clinician who spends part of his day rounding on medical and surgical patients, I see the long shadow that alcohol casts over our health-care system. Alcohol causes a near-elderly man to slip, fall, and break his femur. Alcohol is the reason that an elderly woman is so confused post-operatively, and is unable to participate in her physio. I am reminded of Tennessee Williams’ line that many people don’t touch alcohol but alcohol touches them.

Alcohol is so common and so problematic that we have gotten pretty good at treating the withdrawal, at least the mild and moderate symptomatology. CIWA (and the revised version, CIWA-Ar) is a simple scale for assessing and treating withdrawal and it is – literally – a life saver. When I entered medical school, I remember the randomness with which we seemed to treat withdrawal. Some doctors took a wait-and-see approach while others liberally ordered standing doses of benzodiazepines (based on the patient’s size, his withdrawal symptoms, and the phase of the moon – or so it seemed). Today, we have a vastly more thoughtful and effective approach. And I’m going to make a quick tip of my hat to the hard work Bob Barnes and Zehra Bana have done at my hospital in standardizing care and educating clinicians.

On This Week’s Reading

This Week’s Reading is a thin paper from Alcohol with a heavy ambition. The authors ask: can we identify patients at risk for complicated alcohol withdrawal? Then, Maldonado et al. set out to do just that.

The need is clear: alcohol withdrawal, while usually not difficult, can prove very challenging – linked to a slew of complications. There’s not much I can do when I visit the orthopedics ward and see the man with the broken femur – he came to hospital like that – but what about the elderly woman with the confusion who was lucid on admission? Now that she’s in delirium tremens, it’s a game of catch up. Couldn’t we have flagged her first as being at risk? And if we could identify patients at high risk early and accurately, at the very least there would be the opportunity to monitor them more closely and maybe even intervene early.

Here’s what the study authors did:

  • They started with a massive literature review involving 22 search terms, finding almost 6,000 papers. They then cut this number down, with several exclusion criteria and focusing on unique articles that considered predictive factors for Acute Withdrawal Syndrome. They ended up with 233.
  • Based on these papers, the authors came up with a checklist of just 10 questions (all yes or no, assigned a point for each yes answer).
  • The resulting checklist – the Prediction of Alcohol Withdrawal Severity Scale (PAWSS) – is significant if a patient scores 4 or more.

And PAWSS is pretty straight forward. There is a threshold criteria (alcohol in the last 30 days or a positive blood alcohol level on admission) and then a series of clear questions (about past DTs, blackouts, etc.).

Maldonado et al. have done incredible work in reviewing the literature, and the four or so page justification for the individual PAWSS questions is worth reading. For example, they include nearly 20 studies when discussing “previous episodes of blackouts” (Question 6).


And then they go further. The study authors do a small pilot study using PAWSS: 69 patients admitted to general medical wards.

The results? Maldonado et al. report:

  • Of 69 patients, 1 declined and 51 didn’t meet the PAWSS threshold.
  • 13 scored negative on the PAWSS – and none developed moderate to severe Alcohol Withdrawal Syndrome.
  • 4 were positive – and all developed moderate to severe AWS.

The authors conclude:

Based on our review of the literature of factors associated with alcohol withdrawal severity, we have developed a new tool, PAWSS, designed to identify patients at risk for moderate to severe AWS. The pilot study demonstrated 100% sensitivity, specificity, PPV, and NPV of the PAWSS’ ability to predict complicated alcohol withdrawal in this inpatient medical population. While the tool takes less than a minute to be administered and minimally adds to the overall cost of an inpatient stay, it has the potential to accurately identify those patients who are at high risk to develop complicated AWS.

This conclusion is sweeping in part because the number of PAWSS positive is just 4. A thin paper, to return to my earlier point, with a big ambition, but a small n, alas. So, in the end, I find the topic is interesting; the literature review is impressive; the checklist is thoughtful… and the need for a bigger study is great.

But the core idea – developing a simple checklist to flag patients at significant risk of complicated alcohol withdrawal – is important.


Reading of the Week. Every week I pick a reading — often an article or a paper — from the world of Psychiatry.