ADVERTISEMENT
Symptoms and Diagnoses
Question:
"I believe one of the major challenges clinicians face today is arriving to a diagnosis that differs from the patient’s main complaint … In Psychiatric Times and other journals we have been discussing and blogging on the issue of diagnostic accuracy and it is my impression that most of us agree that, like in the field of somatic medicine, in mental health appearances can be deceiving. 'Anxiety' could be agitation, lack of attention, maybe a sign of sleep deprivation which can be due to reasons as different as excessive caffeine intake or racing thoughts from hypomania or simply worries about the current economic crisis. I could type a long list but instead I want to close by disclosing a teaching moment too close to home: For a few weeks a close relative of mine was complaining of being too tired in the morning and she was irritable for no clear reason. Recently we learned that she was not a teenager staying up late to watch movies, but that actually those complaints were the initial symptoms of lupus. So as Drs. Raison, Jain, Draud, and Maletic would say, we have to look at the whole patient and remember that ‘psychological’ and somatic conditions go hand-in-hand in many cases."
I’m breaking with our tradition of using this Community Forum to answer questions from our live events to make a few comments about this very insightful and challenging comment, posted in response to our last blog. These comments are just my impressions and don’t reflect in any way a consensus amongst my TWP collaborators.
To begin with, when it comes to symptoms, I’m reminded of the joke that ends Woody Allen’s classic movie Annie Hall: “To elderly women are at a Catskill mountain resort, and one of ‘em says, ‘Boy, the food at this place is really terrible.’ The other one says, ‘Yeah, I know; and such small portions.’” We all know that symptoms are a completely inadequate way to diagnose and treat either medical or psychiatric disorders, and yet we can’t do without them. Imperfect as they are, at this point in our science, they are our truest guides.
Before delving into a few complexities, let’s step back and take a 10,000 foot view of why symptoms are important. They are important because they tell us that something is wrong. This may seem ridiculously self-evident, but consider the fact that the most fatal cancers are often so precisely because they manifest no symptoms that would allow their detection and treatment until it is too late to do anything effective about the situation. So in a way, thank God for symptoms.
Symptoms are also important because they can tell us—albeit imperfectly—that our treatments are on the right track. Symptoms emerge from changes in brain/body functioning, so it makes sense that for them to resolve the abnormalities that created them in the first place must be reversed.
Classic evidence in support of this idea comes from a study by Frodl and colleagues.1 They recruited a group of depressed patients and a matched group of non-depressed individuals and did brain imaging to measure the volume of key brain areas involved in major depression (MDD) including the hippocampus and other limbic areas. They then let 3 years pass in which the patients with depression got whatever treatment they could and responded however they did, and then the researchers conducted a follow-up brain scan. The results were striking. The healthy controls showed no loss of brain volume, whereas the depressed group did. But here is the key point about symptoms: depressed individuals who responded to treatment and became asymptomatic had far less brain volume loss than individuals who remained depressed. The most straightforward interpretation of these results would be that when treatment successfully eliminates symptoms it does so by changing brain/body function in ways that block the destructive physiology that promotes MDD in the first place.
But these findings need to be tempered by a reminder that symptoms are far from a perfect guide to the underlying pathophysiology from which they arise. If they were perfect markers of their underlying biology, then putting people in remission from an episode of MDD should ensure that they were at no greater risk of a future episode than people who had never been depressed. We know from numerous studies that this is not true. Although people in remission are far less likely to relapse than those who remain symptomatic, over time their risk of developing a new depressive episode is far greater than it would be had they never gotten depressed. In this regard, psychiatric illnesses are akin to cancer, in which a cure following treatment is the only hope for survival, even if it leaves the individual at much greater risk than the general population for eventual relapse and death.
With these thoughts in the foreground, let’s consider the situation outlined in the comment above. Dr. Mota-Castillo correctly identifies the fact that a major challenge for our field is to find ways to more accurately map symptoms onto diagnoses. He offers great examples of how symptoms can hide as much as they reveal, which is to say that psychiatric symptoms do not map onto underlying conditions in any type of one-to-one fashion that would make them truly diagnostic. He offers anxiety as a classic example of this conundrum. If I say “I just saw an incredibly anxious 40-year-old white female” and ask what this might mean, each of us could make a near interminable list. This woman could be manic and agitated, high on drugs, withdrawing from drugs, depressed, obsessive-compulsive, terrified by paranoid delusions, hyperthyroid, responding to an occult pancreatic cancer, or—as Dr. Mota-Castillo notes—developing systemic lupus erythematosus.
I want to take a stab at radically simplifying why this matters. Trying to figure out what lies behind this patient’s anxiety only matters in two circumstances. The first is when the different syndromic explanations point to very different treatment strategies. The second is when the different syndromes have divergent likely outcomes, with or without appropriate treatment.
The simplest scenarios that highlight these “truths” can be found at the mind-body interface, where well-characterized medical illnesses cause symptoms that appear to be psychiatric. If our patient is anxious because her thyroid hormones are through the roof and we treat her with a benzodiazepine and leave it at that, we have made a terrible mistake for two reasons. First, because in this case the patient’s anxiety had nothing to do with anxiety per se, but was simply a sequelae of thyroid dysfunction. Second, because the best way to have treated this patient would have been to fix the underlying thyroid illness, which we as doctors know how to do quite well.
The situation is even graver if our patient is anxious because of signals to the brain from an occult pancreatic carcinoma and we treat her with a benzodiazepine and send her on her way. Why? Because, although the benzodiazepine might quell the symptom, this relief will have come at the cost of her life, given that early diagnosis of the cancer might have offered a curative procedure.
Here’s where I want to gently complicate our hunger for similar diagnostic rigor in psychiatry by asking a question. How much does it matter whether this patient’s anxiety is an expression of one psychiatric condition or another? I would suggest that the answer to this question is no different than the answer to why diagnoses matter in general. It matters exactly to the degree that assigning the patient’s anxiety to one or another of our current psychiatric conditions suggests different treatment approaches, and doesn’t matter much at all when the diagnoses are largely treated the same.
Let’s stick with anxiety and do a thought experiment. Would it matter if the patient’s anxiety is from generalized anxiety disorder (GAD) versus substance withdrawal? I would say yes absolutely, and for two reasons both related to treatment. First, withdrawal syndromes characterized by anxiety can occasionally be lethal (ie, alcohol withdrawal). Second, because substance abuse requires a different type of treatment than does GAD. If the abuse isn’t stopped, treating the symptom will either a) not be effective or b) won’t solve all the other problems that go along with substance abuse dependence.
OK. Would it matter if the patient’s anxiety derived from GAD versus a paranoid delusion associated with schizophrenia? We’d all say, “Absolutely!” But why do we say this? At the end of the day, doesn’t a lot of it boil down to the fact that we’d use an antidepressant/anxiolytic in the first place and an antipsychotic in the second? And making the diagnostic distinction also matters because we know that people with schizophrenia are likely to have very different life trajectories (on average) from people with GAD, which is to say that the diagnosis provides important prognostic information.
Fair enough. But now consider two other possible diagnostic dichotomies. Would it matter if the patient’s anxiety derived from GAD versus MDD? What about deriving from bipolar disorder versus schizophrenia? In each of these cases, I would suggest that it doesn’t matter nearly as much as in our prior examples. Why? Because our treatments for these conditions are similar, and because the long-term outcome for these disorders also demonstrates considerable overlap. If you doubt this claim, think back to the days before atypical antipsychotics (if, like me, you are old enough), back to the days when schizophrenia was preferentially treated with typical antipsychotics and bipolar disorder was treated with mood stabilizers. Back then, we also believed that differentiating the two conditions had important prognostic implications. Remember when you had to figure out if a psychotic patient was manic or just psychotic, because lithium would work in the first instance, but not the second? Remember when we used to tell people that patients with bipolar disorder would likely lead normal lives when not in an episode?
All gone. Nowadays we treat these conditions with most of the same medications, and we know that schizophrenia can turn out better than we thought and that bipolar disorder usually turns out worse than we thought. It’s all mashed together. Even the genetic risks for these conditions seem to overlap almost entirely. Accordingly, our alacrity in making these diagnostic distinctions has diminished. I would suggest this is a perfectly natural result of the diagnoses mattering less vis-à-vis each other.
Not that I’m suggesting this, but behind these musings is a potentially radical way to reshape the DSM. We could rate disorders by how effectively the point to unique treatments and clinical outcomes compared to other disorders with which they share symptoms. At the top of such a list would be the requirement to be as assured as possible that any given psychiatric symptom is not directly arising from a known medical condition, because in these cases the optimal psychiatric treatment is to address the underlying medical condition.
On the next rung down I might place what I call the “great pretenders,” by which I mean syndromes with specific treatments that are nonetheless often missed because patient’s present with symptoms that are not pathognomonic for the condition. My top three “great pretenders” are substance abuse (which can mimic any other psychiatric condition), obsessive-compulsive disorder (which often presents with non-specific symptoms like anxiety and insomnia because people are embarrassed to talk about their obsessions and compulsions), and psychosis (which can look like depression, anxiety, or primary substance abuse). In addition to being “pretenders” each of these conditions would be high on my list of diagnostic importance because each requires specific treatment interventions for optimal outcomes (ie, sobriety for substance abuse, behavioral therapy for OCD, and antipsychotic medications for psychosis).
At the bottom of my diagnostic hierarchy would be many of our most important disorders, including MDD, bipolar disorder, GAD, and schizophrenia, not because these conditions aren’t giant threats to health and well-being, but because their diagnostic boundaries seem to me—at least—to be increasingly blurry. If people are depressed and anxious they need to be treated, but the treatments are largely overlapping as are the conditions. If people are psychotic they need to be treated, but making all the finer distinctions about the flavor of any given psychosis seems far less relevant these days.
It would be a fascinating and potentially very useful exercise for those of us in the TWP community to consider when and why our diagnoses matter, and when they don’t matter so much, when in fact symptoms and their treatment matter more than assigning these symptoms to one syndrome or another.
References
- Frodl TS, Koutsouleris N, Bottlender R, et al. Depression-related variation in brain morphology over 3 years: effects of stress? Arch Gen Psychiatry. 2008;65(10):1156-1165.
- Judd LL, Akiskal HS, Maser JD, et al. Major depressive disorder: a prospective study of residual subthreshold depressive symptoms as predictor of rapid relapse. J Affect Disord. 1998;50(2-3):97-108.
