Sunday, February 11, 2018

The Problem of the Drinking Spouse

Any physician treating alcoholism knows this scenario very well.  You have finally convinced a person that they have a problem with alcohol.  They have been in treatment and either using or not using MAT (medication assisted treatment - naltrexone or acamprosate).  They are at the point where they are abstinent many more days than they are drinking.  This is a critical point for many people who are daily drinkers.  As you work with them in trying to define critical factors for continued use they identify: "My husband/wife refuses to stop drinking.  They say it is my problem.  They like drinking and they refuse to stop.  They have alcohol at the house."

There are many variations on that theme.  Some spouses will keep all of the alcohol locked up and imbibe only when the sober spouse is sleeping.  Some will not have any alcohol at all, but continue to drink in social situations accompanied or unaccompanied by the sober spouse.  Some will just resent the sober spouse and the sudden restriction in the couple's social life.  Many couples start drinking to increase their social activity and expand their social contacts.  In many businesses, this level of socialization and the associated drinking is expected.  The associated level of emotionality in the marriage can increase precipitously based on the new expectations of the sober spouse about how things should be to support their sobriety.  The combination of the environmental cues from alcohol and increased emotionality greatly increase the risk for continued alcohol use and make all of these patterns untenable.  Convincing the drinking spouse that their behavior does not facilitate sobriety in the marriage is a difficult task - if it is attempted at all.

Are there any large scale studies that back up those clinical observations?  A certain portion of drinking spouses may respond to clear scientific evidence if they cannot respond to the advice of a counselor or physician.  It turns out that there are and a lot of that work has been done by Kendler and co-authors. 

The most recent paper in JAMA Psychiatry (1) looks at the issue of spousal resemblance for alcohol use disorder.  In the study, subjects were obtained from a generational sample of all people born between 1960 and 1990 in Sweden who were married before December 31, 2013.  They were identified as having alcohol use disorders (AUD) through several databases that looked at medical diagnoses, medication assisted treatment prescriptions (disulfiram, naltrexone, acamprosate) and convictions or suspicions of at least two alcohol related crimes.  That resulted in marital pairs - 5883 where the husband first developed an AUD and 2679 where the wife first developed an AUD.  They note that in marital pairs, first onset AUD was much greater in pairs where a spouse had an AUD than when they did not.

They analyzed the data by two methods.  First, they looked at hazard ratios of developing an AUD relative to a control group matched by sex, birth year, year or marriage, family history of AUD, and parental educational level.  Second, they looked at intraindividual hazard ratios across subsequent marriages and divorces.

In the first analysis, the hazard ratio of AUD in the wife after the husband had an initial AUD was 13.82 dropping to 2.75 over the first two years.  In the case of  husbands after a wife's first registration of AUD the hazard ratio was 9.21 falling more slowly to 3.09 after 3 years.

In the intraindividual comparisons - for husbands moving from a spouse with no AUD to one with an AUD resulted in a HR of 7.02.  Moving from a spouse with an AUD to one without and AUD decreased the risk to a HR of 0.50 for AUD.  The protective effects persisted in the same direction in second and third marriages.  They produced a comprehensive tables of 20 possible combinations of spouses +/- AUDs and list the protective and predisposing combinations.  In each case, whether or not the prospective spouse has an AUD predicts the the probands status.

The authors conclude that this is tentative evidence that a spouses alcohol use status has a causal effect on their spouses drinking.  They suggest the likely processes and suggest that assortative mating is a factor in the large increase in drinking that can occur when a man or woman without an AUD marries a man or woman with an AUD.  Assortative mating has been previously studied by Kendler (4) and is defined as mate selection that depends on similarity across traits - in this case drinking patterns and risk factors for AUD.  It is an interesting concept because it suggests at least part of the mechanism of greatly increased risk in the spouses of drinkers.  A non-drinking spouse with those characteristics may have more credibility as a protective effect, but those specifics are not clear at this time. 

The limitations are discussed in the original paper and I won't belabor them here.  Clearly the study design is an issue.  It is likely that cases were missed.  I have not seen it studied, by my experience with diagnoses and the American insurance system suggests that many people will do what they can to stay off of a database.  I can't imagine that is not also true in Sweden.  They did a comparison of the AUD prevalence of their data to Norway and found the prevalence was lower.  This methdology also focuses on more severe AUD.  I based that on the fact that the DSM-5 committee eliminated legal problems as a diagnostic criteria for AUD based on it not adding much to the criteria because it was associated with most of the other criteria.              

An observation about the study.  It could not have occurred in the United States - at least not on the same scale.  In the US, treatment for alcohol or substance use problems comes under the the auspices of §CFR 42, limiting access to information for research purposes.  Advocates for these restrictions will of course say they are necessary and that people can still release information like they can for any other medical condition - but like most of these regulations there is general confusion and intimidation of clinicians to the point that the extra hurdles necessary to do research are seldom breached.  In the US, in the case of non-public programs like Medicare or Medicaid, all of the data is aggregated by health care system.  In Scandinavian countries all patients are on a single national database.  In the Swedish study, the researchers assigned unique serial numbers to all of the subjects and the ethics committee approval waived consent because of this procedure.

This study gets back to a philosophy of life and the issue of sobriety or at least self-correcting abstinence.  Couples do have conversations about drinking.  They do make conscious decisions about drinking and substance use.  They observe one another when they have become too intoxicated and had significant embarrassment or hangover effects.  If there are no baseline agreements about the use of intoxicants early in the marriage there should be a discussion about self correcting abstinence.  When do we agree to stop whatever we are doing as a couple and reassess our use of intoxicants.  Things do not have to get to the level of an actual alcohol or substance use disorder.

Finally, what about the approach to the couple when there is a clearly defined alcohol or substance use problem?  The couple's dynamic does need to be identified and addressed.  For any physician or counselor approaching the problem is fraught with difficulty.  Spouses tend to be defensive, resentful, and in some cases openly hostile to the idea that they need to stop drinking.  The drinking spouse may see the physician or counselor as affiliated with the nondrinking spouse and that can amplify the resentment and negative emotion.  There are programs with a more neutral response that treats the drinking spouse in an entirely different context and provides the necessary education.  Al-Anon is the prototypical self help program for spouses that attempts to address anger, resentment, and provide a focus on positive strategies.  I am still waiting to see an explicit manual, pamphlet, or book that is focused on why the drinking spouse needs to stop drinking.  If I missed that please send me a link to that resource.   

Before you send a comment on the couple where one person is sober and the other person drinks, I can assure you that I am aware that the situation exists.  I typically see it where the spouses are independent and often have separate social and recreational outlets.  In many cases, one of the spouses works excessively and alcohol use is incorporated into work activities or becomes a ritual on the way home.  The situation I hope to address here is one where both spouses are drinking - usually too much and one of them wants to quit.

I have not seen a lot written about the problem or the solution.

George Dawson, MD, DFAPA


1: Kendler KS, Lönn SL, Salvatore J, Sundquist J, Sundquist K. The Origin of Spousal Resemblance for Alcohol Use Disorder. JAMA Psychiatry. 2018 Feb 7. doi: 10.1001/jamapsychiatry.2017.4457. [Epub ahead of print] PubMed PMID: 29417130

Full text available on line.  Please read it.

2: Kendler KS, Lönn SL, Salvatore J, Sundquist J, Sundquist K. Effect of Marriageon Risk for Onset of Alcohol Use Disorder: A Longitudinal and Co-Relative Analysis in a Swedish National Sample. Am J Psychiatry. 2016 Sep 1;173(9):911-8. doi: 10.1176/appi.ajp.2016.15111373. Epub 2016 May 16. PubMed PMID: 27180900.    

3: Kendler KS, Lönn SL, Salvatore J, Sundquist J, Sundquist K. Divorce and theOnset of Alcohol Use Disorder: A Swedish Population-Based Longitudinal Cohort and Co-Relative Study. Am J Psychiatry. 2017 May 1;174(5):451-458. doi: 10.1176/appi.ajp.2016.16050589. Epub 2017 Jan 20. PubMed PMID: 28103713; PubMed Central PMCID: PMC5411284.

4: Maes HH, Neale MC, Kendler KS, Hewitt JK, Silberg JL, Foley DL, Meyer JM,Rutter M, Simonoff E, Pickles A, Eaves LJ. Assortative mating for major psychiatric diagnoses in two population-based samples. Psychol Med. 1998 Nov;28(6):1389-401. PubMed PMID: 9854280.

Graphics Credit:

Photo at the top is from Shutterstock per their licensing agreement.


  1. Cue Melody Beattie and her Codependency writings? Abstinence is the only model for recovery, sorry if that is rigid and inflexible in presentation, but, if there was a controlled use model of genuine effectiveness, well, most Sub Abuse Programs would be irrelevant, eh?

    1. Agree - the debate of controlled drinking versus abstinence dates back 20-30 years and seems confusing - except the controlled drinking intervention outcomes were abysmal. Abstinence based outcomes typically include a days to relapse or number of days drinking per month, but those results are much better if abstinence in the goal. If the drinking pattern is episodic there is generally a marked improvement in medical outcomes and better quality of life.

      Harm prevention is another option when there is a lot of chronicity. I have seen too many people put in the harm reduction model of treatment when they did not have an adequate course of primary treatment.