Practice Point: Don’t forget to consider the harms of inaction when first doing no harm. Untreated Attention Deficit Hyperactivity Disorder (ADHD) increases the risk of death two-fold, and medication provides an overall mortality benefit that is measurable as early as two years after initiating treatment.
EBM Pearl: When randomized trials are not feasible (often due to cost, size, or ethical concerns), investigators may use the target trial emulation approach to avoid making errors of assumption that could result in erroneous causal conclusions, which are one of the greatest sources of bias in observational studies. This two-step process involves articulating a causal question by designing a hypothetical ideal trial and then carefully emulating the components of that trial using the available observational data.
Yes, Attention Deficit Hyperactivity Disorder (ADHD) is both over- and under-diagnosed, and yes, stimulants have potential for misuse. But for people who suffer from hyperactivity, inattention, and impulsivity, ADHD can be debilitating and even fatal. People with ADHD are twice as likely to die prematurely, mostly due to unnatural causes. A recent study found that medication initiation was associated with a significant reduction in all-cause mortality and particularly death due to unnatural causes among people with ADHD in the first two to five years after starting treatment.
Registry data from 148,578 patients aged 6-64 years (median age 17 years, 59 percent male) living in Sweden who had an ADHD diagnosis but no treatment in the eighteen months prior to the study period were evaluated in cohorts based on either treatment initiation or non-initiation. All patients had neuropsychological testing, which is required for an ADHD diagnosis in Sweden.
During the study period, 57 percent were started on a medication (most often a stimulant). The primary outcomes were all-cause and cause-specific mortality at two years, with a sensitivity analysis extending outcomes to five years. Cause-specific mortality was divided into “unnatural” causes (which included suicide, accidental poisoning or injuries, and other external injuries) and “natural” causes (such as, heart attack or cancer). The authors designed a target trial protocol and adjusted for baseline confounders to emulate randomization in an attempt to reduce bias.
At two years, 632 people (0.43 percent) had died, and by five years 1,402 (0.94 percent) had died, with 67 percent of all deaths occurring from “unnatural” causes. The incidence rate of all-cause mortality at two years comparing treated vs. untreated patients was 39 vs. 48 per 10,000 individuals (hazard ratio [HR] 0.79, 95% CI 0.7-0.88). Further, the two-year incidence rate of death due to unnatural causes was 26 vs. 33 per 10,000 individuals (HR 0.75, 95% CI 0.66-0.86), but there was no significant reduction in death from “natural” causes. When data were extended to five years in a sensitivity analysis, the reduction in all-cause mortality risk was less pronounced and no longer significant, but the reduction in death due to unnatural causes remained significantly reduced (HR 0.89, 95% CI, 0.81-0.97). Subgroup analyses stratified by age (ages 6-24 and ages 25-64) showed similar results.
We often worry about the potential harms of prescribing stimulant medications, particularly harms associated with cardiovascular risk and the potential for substance misuse or diversion, but this study suggests that we might be underestimating the risk of inaction when it comes to treating ADHD. The act of doing something (prescribing medication) naturally lends itself more to thinking about harms than does inaction, but the risk of high blood pressure or even a heart attack in twenty years is a moot point for the people with ADHD who die prematurely.
So, while we are valiantly considering all-cause mortality, we should think of the all-cause harms too, which includes the harms of action as well as inaction. According to this study, the balance seems to clearly lie in favor of pharmacologic treatment for ADHD.
For more information, see the topic Attention Deficit Hyperactivity Disorder (ADHD) in Adults in DynaMed.
Reference: JAMA. 2024 Mar 12;331(10):850-860
