Here’s the part of the conversation your patients usually find clarifying: the brain doesn’t really care whether the reinforcer is ethanol, an opioid, or a frosted donut at 9 p.m. It cares about the speed and reliability of the signal. Sugar, particularly when it shows up intermittently and in concentrated form, hits that signal cleanly.

The animal literature is where this gets sharpest. In the Avena line of work, rats given intermittent access to sugar solutions developed binge-like intake, signs that looked like withdrawal when sugar was removed, and shifts in both dopamine and opioid signaling that paralleled what’s seen with drugs of abuse ⁵. The authors put it plainly:

sugar “releases opioids and dopamine and thus might be expected to have addictive potential” ⁵.

That’s not a casual analogy. It’s a specific neurochemical claim about a specific intake pattern.

The intermittent piece matters more than the total dose. A rat with steady, predictable access to sugar doesn’t develop the same profile as one that gets it in bursts after deprivation. Translate that to your caseload and the pattern is familiar: the patient who white-knuckles through the morning, skips lunch, then collapses into a sweet at 4 p.m. is essentially running the experimental protocol on themselves. The dopamine response gets larger, not smaller, with that kind of scheduling.

Newer reviews echo the mechanism. High-sugar food triggers dopamine release in reward pathways, and repeated activation appears to sensitize those pathways in vulnerable individuals ⁴. For someone whose mesolimbic system has already been reshaped by substance use, this is not a neutral input. It’s a familiar key fitting a familiar lock.

That framing changes the clinical conversation. You’re not asking a patient to resist a cookie. You’re asking a sensitized circuit to tolerate a smaller, less predictable reward while it recalibrates. Worth saying out loud in session.

Now the caveat your patients deserve to hear, especially the clinically literate ones who will ask.

Most of the strongest neurochemical evidence comes from rodents on engineered schedules, not from humans eating birthday cake. The reviews that frame sugar as addiction-like are careful about this. The NIH-hosted synthesis notes that while reward pathway sensitization, tolerance, and withdrawal-like signs have all been described, formal diagnostic criteria for sugar addiction in humans remain debated ⁴. There is no DSM entry. There is no validated screening instrument with the kind of consensus you’d see for alcohol or opioid use disorder.

What you can say honestly: the biological plausibility is real, the behavioral overlap in some patients is real, and the reinforcer is powerful enough to deserve clinical attention, particularly in dual-diagnosis contexts. What you should not say: that a patient “has” sugar addiction in the same diagnostic sense they have opioid use disorder. The category isn’t settled, and overclaiming it can erode trust the first time a patient reads the same caveats you skipped.

The useful middle ground is to treat sugar as a high-potency reinforcer operating on an already-sensitized system, with addiction-adjacent features in a subset of patients. That framing lets you apply the behavioral toolkit you already trust without picking a fight with the literature. It also keeps the door open for the patient who responds well to structure but bristles at another label.

Hold the nuance. The rest of this piece assumes you can.

If you’ve used implementation intentions in relapse prevention, you already know the form. “If I see my old using buddy at the gas station, then I drive to the next exit.” The structure pre-loads a response so the patient doesn’t have to negotiate with their cravings in real time. The same technique transfers cleanly to sugar.

A systematic review of implementation-intention interventions across dietary contexts concluded that if-then planning can produce small-to-moderate improvements in eating behaviors, including reduced intake of unhealthy foods ³. Small-to-moderate is honest framing. It’s not a miracle. It’s a reliable nudge in the right direction, which is exactly what you want for a behavior the patient has to perform several times a day.

The adolescent SSB trial is worth knowing about because the pattern shows up in a higher-risk population. Adolescents with overweight or obesity who received implementation-intention training combined with inhibitory-control practice cut their sugar-sweetened beverage consumption more than controls ². Two skills, layered: a plan for what to do at the cue, and a rehearsed pause before the automatic reach.

In session, the plans that hold up are specific to the patient’s actual day. Not “if I want something sweet, then I’ll have water.” That’s a wish. Something closer to: “If I finish my last patient note at 4:30 and I’m reaching for the vending machine, then I walk to the lobby coffee station and refill my water bottle first.” The cue is named. The behavior is named. The location is named. The patient can rehearse it before leaving your office.

For a dual-diagnosis caseload, the relapse-prevention parallel is your selling point. Patients who have already built if-then plans for substance use cues don’t need to learn a new framework. They need permission to apply the one they have.

Mindfulness work in this space isn’t about savoring a raisin for ten minutes. It’s about teaching the patient to sit with a craving long enough to watch it crest and fall, without staging an intervention against themselves.

A mindfulness-based intervention study tracking participants over the course of the program found that increases in mindful eating were associated with decreased eating of sweets and lower fasting glucose levels in the mindfulness group ⁹. Association, not causation, and a modest sample, but the direction matters. The patients getting better at noticing were also, on average, eating fewer sweets and showing better glycemic markers.

For someone already using mindfulness for PTSD flashbacks or anxiety spikes, this is a small skill extension rather than a new modality. The instruction set is familiar: notice the urge, name the body sensation, observe without acting, let the wave pass. You’re just applying it to a different stimulus.

What this skill does well is decouple the craving from the automatic behavior. What it does not do well is shrink the craving in the moment. Patients should know that distinction before they try it. A craving they can observe is still a craving. The win is that they’re no longer being driven by it without consent. Over weeks, the frequency and intensity tend to soften, which is what the intervention data suggests ⁹.

Pair this with the if-then plan from the previous skill and you have something close to a complete in-the-moment response: notice, pause, execute the pre-loaded behavior.

If you want a short list of techniques to actually bring into a session, the 2025 systematic review on behavior change techniques (BCTs) for sugar reduction is the most useful single source in the supplied literature. The authors identified four techniques with the clearest signal across the included trials: goal setting, self-monitoring, feedback, and environmental restructuring ¹¹. They note that the results “will enable clinicians to provide more effective dietary advice when supporting dietary behavior change” ¹¹.

Worth walking through what each looks like in practice for a dual-diagnosis patient.

Goal setting

works when it’s specific and the patient helped write it. “Reduce sugar” is not a goal. “No sweetened drinks before 6 p.m. on weekdays” is. The narrower the target, the easier it is to either succeed or learn from a miss.

Self-monitoring

is the technique most patients underestimate and most clinicians can deploy in five minutes. A note app, a photo log, a tally on the back of an appointment card. The mechanism is awareness, not punishment. Patients who track honestly for two weeks almost always describe a pattern they didn’t know they had.

Feedback

is where outpatient counseling adds value the patient can’t replicate alone. You see the log. You reflect what you see. You name the cue you noticed that they didn’t. This is the BCT that turns self-monitoring from a journaling exercise into a clinical signal.

Environmental restructuring

is the highest-leverage one for early recovery. If the sweet isn’t in the house, the 9 p.m. craving has nowhere to go. If the coworker’s candy bowl is on a different floor, the workday cue weakens. You’re not asking the patient to outmuscle their reward system. You’re rearranging what the reward system gets to vote on.

The four techniques compound. A goal without monitoring is a wish. Monitoring without feedback is a diary. Feedback without environmental change asks the patient to win every cue in real time, which they won’t. Used together, they’re the structural backbone of the planning interventions covered next.

When patients ask whether any of this actually works in a study setting, the Sugar Habit Hacker trial is the cleanest answer in the supplied evidence. It’s a randomized test of a planning-based digital intervention built around action planning and coping strategies, the same BCT family covered above ¹⁰.

The headline numbers: the intervention produced large effects on reducing cravings (d = 0.59) and on reducing psychological distress (d = 0.68) compared with controls ¹⁰. Two things worth pulling out of that pair.

First, the distress effect was larger than the craving effect. That’s the part that should interest a dual-diagnosis clinician. A structured sugar-reduction plan didn’t just lower the urges. It lowered the psychological distress markers the participants reported alongside them. The mechanism is plausible: if cravings have been an ambient source of self-criticism and perceived loss of control, getting a working response to them removes a chronic stressor.

Second, d = 0.59 for cravings is a respectable signal for a brief, structured, mostly self-directed intervention. It tells you that the same planning techniques you can teach in an outpatient session, formatted into a coherent protocol, move the dial on the symptom most patients are most worried about.

Caveats the patient should hear once and not have to relitigate: the trial was time-limited, and long-term maintenance plus applicability to patients with severe mental illness or active substance use disorders are open questions ¹⁰. The effect sizes are an argument for trying the protocol, not a guarantee it will hold for a year. Use them to motivate engagement, not to set expectations the next relapse will undercut.