TEMPUS trial finds no short-term liver fat advantage for time-restricted eating or supervised exercise
A 12-week randomized trial in adults with obesity found no measurable edge for time-restricted eating, supervised exercise, or the combination when the target was intrahepatic fat. The primary endpoint, change in intrahepatic fat fraction measured by magnetic resonance imaging, moved only modestly across all active arms and did not separate from usual care, which consisted of Mediterranean diet and physical activity education.
The result is clinically important because intrahepatic fat is not a cosmetic biomarker. It sits at the center of insulin resistance, dyslipidemia, and the metabolic stress that drives fatty liver progression. A behavioral intervention can be elegant on paper and still fail to clear the biological threshold needed to shift liver fat over a short interval. In this trial, the liver did not meaningfully distinguish one lifestyle prescription from another.
Trial design and population
The TEMPUS study used a parallel-group randomized controlled design in Granada, Spain, and enrolled adults with obesity between May 2023 and April 2024. The trial was completed on 5 September 2025 and included 187 randomized participants, with women making up 50% of the cohort. All participants were retained in the primary intention-to-treat analysis, which strengthens the interpretability of the null finding and reduces the chance that attrition explains the result.
Participants were assigned to one of four groups: usual care, 8-hour time-restricted eating, supervised exercise, or the combined intervention. Usual care was not a passive comparator. It included Mediterranean diet and physical activity education, both of which can influence weight regulation and cardiometabolic risk. Against that baseline, any additional benefit from eating-window restriction, structured exercise, or their pairing had to overcome a relatively active control condition.
Blinding was partial, as is often unavoidable in lifestyle studies. Primary outcome assessors and statisticians were blinded to allocation, while intervention personnel were not. That is a sensible compromise for a behavioral trial, but it also means the core readout hinges on an objective imaging endpoint rather than subjective reporting.
Intrahepatic fat fraction did not separate across arms
The primary outcome was change in intrahepatic fat fraction after 12 weeks. None of the active strategies outperformed usual care. Compared with usual care, the difference in IHFF change was −0.4% for time-restricted eating, with a 95% confidence interval from −2.0 to 1.2 and a P value of 0.93. Exercise produced a difference of −0.8%, with a 95% confidence interval from −2.4 to 0.8 and a P value of 0.54. The combined time-restricted eating plus exercise arm showed a difference of −1.1%, with a 95% confidence interval from −2.7 to 0.5 and a P value of 0.28.
The intervention arms also did not differ meaningfully from one another. In plain clinical terms, the hepatic fat signal was flat.
Why a biologically plausible intervention can still miss the mark
The negative result should not be read as a verdict against either strategy in every setting. Time-restricted eating can improve caloric timing, reduce late-night intake, and alter daily insulin exposure. Exercise can improve substrate oxidation, mitochondrial efficiency, and peripheral insulin sensitivity. In theory, both should support hepatic fat mobilization. Yet biology often demands more than a plausible mechanism; it demands enough intensity, enough duration, and enough net energy deficit to move liver triglyceride stores in a durable way.
Twelve weeks may simply be too short for those pathways to show a differential effect when compared with a structured education-based control group. Liver fat is responsive, but not infinitely so. If the intervention does not create a substantial enough shift in energy balance, hepatic lipid handling, or adherence, the imaging signal may stay within statistical noise.
There is also a practical issue in lifestyle medicine that this trial makes hard to ignore. Patients can follow a named strategy and still end up with similar metabolic exposure if the underlying behavior change is modest. An 8-hour eating window and supervised exercise sound distinct, yet their downstream effect can converge when real-world adherence, compensatory intake, or insufficient training dose narrows the gap between groups.
What the null finding means at the bedside
For clinicians treating obesity or early metabolic liver disease, the study argues against assuming that time-restricted eating or supervised exercise will reliably reduce intrahepatic fat over the short term simply because the intervention is popular or physiologically appealing. The safest interpretation is narrower: in this cohort, with this comparator, over 12 weeks, no intervention produced superior IHFF reduction.
A single concrete implication stands out. If the therapeutic goal is to lower liver fat, broad lifestyle advice may need to be paired with stronger, more individualized weight-loss strategies, longer follow-up, or a different behavioral intensity than what was tested here. The trial does not answer which subgroup, if any, might benefit more than the average participant. It also does not establish whether a longer intervention horizon would eventually separate the arms.
The absence of a liver-fat advantage does not erase other possible benefits of diet timing or exercise, but those outcomes were not the focus of the primary report available here. For hepatometabolic practice, the message is restrained and direct: the imaging endpoint did not budge enough to justify a claim of superiority for time-restricted eating, supervised exercise, or their combination over usual care.
The broader lesson from TEMPUS
Behavioral interventions often enter practice with a strong mechanistic story and a weaker clinical effect size than hoped. This trial fits that pattern. When the endpoint is measured carefully by MRI and the comparison arm already includes structured lifestyle education, short-term gains may compress across groups and erase any apparent hierarchy between interventions.
That reality should shape how future studies are designed. Longer duration, tighter adherence measurement, and more detailed metabolic phenotyping will matter if investigators want to know whether eating-window restriction and supervised exercise can truly lower hepatic fat beyond standard counseling. For now, the TEMPUS data leave the field with a clean but inconvenient conclusion: the three active strategies tested here did not outperform usual care on the primary liver-fat outcome, and the combined approach did not rescue the signal.