Tesamorelin HIV-Associated NAFLD Randomized Trial: NCT02196831

NCT02196831 tested tesamorelin against placebo for 12 months in adults with HIV and elevated liver fat, followed by a six-month open-label extension. The randomized phase found a modest average reduction in magnetic-resonance-measured hepatic fat and fewer instances of fibrosis progression among participants with paired biopsies. It did not show reversal of existing fibrosis, a significant improvement in the overall NAFLD activity score, or better clinical outcomes such as fewer liver complications. The trial is informative about its named tesamorelin formulation, selected population, and protocol; it is not evidence for other formulations, other growth-hormone-releasing peptides, catalog materials, or unsupervised use.[1][2]

Study at a glance

The protocol describes a fixed historical amount and daily exposure. Those details identify the experiment and are not dosing, preparation, administration, or self-use guidance.

Design, eligibility, and masking

Participants were recruited at Massachusetts General Hospital and the US National Institutes of Health. Eligible adults had documented HIV and liver fat of at least 5% on proton magnetic resonance spectroscopy. The trial therefore selected people with both HIV and imaging-defined fatty liver; it was not a general-population NAFLD study.[1][2]

Allocation was 1:1 and stratified by site and vitamin E use, using permuted blocks of varying size. Participants, treating personnel, investigators, and outcome assessors were masked during the randomized phase. Liver fat was measured by imaging, while optional or available paired liver biopsies supplied histology data. These two forms of evidence answer different questions: an imaging change in fat is not the same as histologic resolution, prevention of cirrhosis, or improved health.

Participant flow requires three different counts

The paper describes 61 participants as enrolled and randomly assigned: 31 to tesamorelin and 30 to placebo. One person assigned to tesamorelin withdrew at baseline before receiving study drug. Thus, 60 began their assigned exposure, 30 in each group. ClinicalTrials.gov displays 30 participants as starting each arm, while retaining an overall enrollment of 61. Calling both numbers simply “randomized” hides the source difference.[1][2]

Follow-up liver imaging was available for 26 tesamorelin-assigned and 28 placebo-assigned participants. The paper’s repeated-measures primary analysis used those available observations. Registry participant flow lists 21 and 26 as completing the 12-month randomized phase, respectively. Attrition therefore was not identical across groups, and “completed,” “had follow-up imaging,” and “included in a complete-case registry table” are not interchangeable.

The registry’s posted primary-outcome table uses 20 active and 25 placebo complete cases. It explicitly notes that this population differs from the manuscript. The publication’s repeated-measures analysis, using 26 and 28 participants with outcome data, is the prespecified headline analysis. Both should remain visible rather than selecting whichever count produces the simpler story.

Primary liver-fat result

At 12 months, the estimated net absolute treatment effect on hepatic fat fraction was −4.1 percentage points (95% confidence interval −7.6 to −0.7; reported p about .02). Expressed relatively, the estimated reduction was 37% versus placebo (95% CI 7% to 67% reduction). The absolute estimate is usually easier to interpret because it retains the measurement scale.[1]

Nine of 26 participants with active-group imaging, or 35%, had hepatic fat below 5% at month 12, compared with one of 28, or 4%, under placebo (p about .007). That threshold outcome is based on small counts and available imaging, not all 61 assigned participants.

Statistical significance means that the prespecified analysis was inconsistent with a zero average difference under its model. It does not by itself establish the clinical importance of losing roughly four percentage points of liver fat, durability after therapy, or a reduction in liver-related illness. The confidence interval permits effects from less than one to more than seven percentage points, which is meaningful uncertainty.

Histology: progression signal without fibrosis reversal

Among participants with paired biopsies, fibrosis progression occurred in 2 of 19 in the tesamorelin group and 9 of 24 in the placebo group (p=.04). This was a secondary finding in 43 participants, not the full randomized cohort. The absolute counts are important: seven fewer progression events arose in a small subset, so replication is needed.[1]

The trial did not show improvement of existing fibrosis: improvement was seen in two active and three placebo participants, with p=.71. Nor was there a significant between-group change in the total NAFLD activity score; the estimated difference was −0.3 points (95% CI −1.0 to 0.5). Lobular inflammation and hepatocellular ballooning also lacked significant between-group effects. A reduced progression count cannot be rewritten as “reversed fibrosis” or “treated NASH.”

Secondary biomarkers and null findings

Overall alanine aminotransferase and gamma-glutamyl transferase changes did not differ significantly. A subgroup with elevated baseline alanine aminotransferase had an estimated −29 U/L treatment effect (95% CI −55 to −3; p=.03), but subgroup findings are less secure than a prespecified whole-cohort primary result. C-reactive protein had a small reported difference, while lipid measures were not significantly different.[1]

The study recorded no clinical liver endpoint such as decompensation, cirrhosis, transplantation, hospitalization, cancer, or mortality benefit. Imaging, histology, enzymes, and inflammatory markers are intermediate measures; none can substitute automatically for how a person feels, functions, or survives.

Glucose safety and adverse events

The primary safety focus was glucose homeostasis. Fasting glucose and hemoglobin A1c did not differ significantly between groups at 12 months. Hyperglycemia was recorded in 12 of 31 tesamorelin-assigned participants and 11 of 30 placebo-assigned participants, and two active-group participants discontinued because of hyperglycemia. A nonsignificant average laboratory comparison does not mean glucose-related events were absent.[1][2]

At least one adverse event was reported for 29/31 active-assigned and 29/30 placebo-assigned participants. Serious events affected 4/31 and 2/30, respectively; the study was not sized to compare uncommon harms. Four active-assigned participants and one placebo-assigned participant met an adverse-event discontinuation criterion. Local complaints were generally more frequent with tesamorelin, although bruising counts were similar. The separate six-month open-label safety tables cannot be used as though they preserve a concurrent placebo comparison.

Sponsor, supply, and disclosures

Massachusetts General Hospital was the registered sponsor and the National Institute of Allergy and Infectious Diseases was a collaborator. Funding included NIAID award U01 AI115711, NIH intramural and related academic support. Theratechnologies supplied tesamorelin at reduced cost. Per the paper, the company received serious-adverse-event reports and notice of publication intent under contract but otherwise had no role in study design or conduct, manuscript preparation, or submission. The paper also discloses several unrelated industry relationships and that Steven Grinspoon advised Theratechnologies, had an unrelated investigator grant, and was an inventor on a patent application concerning tesamorelin effects in hepatic disease.[1][2]

These relationships do not invalidate the data, but they matter when judging study design, interpretation, and the need for independent replication.

Statistical significance, effect, uncertainty, and importance

• Statistical significance: the liver-fat primary result and paired-biopsy fibrosis-progression comparison met conventional thresholds; several histology, enzyme, lipid, and glucose comparisons did not.
• Effect size: the modeled liver-fat difference was −4.1 percentage points, while fibrosis progression differed by 2/19 versus 9/24 in the biopsy subset.
• Uncertainty: confidence intervals were wide, outcome denominators varied, attrition occurred, and histology involved fewer people than imaging.
• Clinical importance: no liver complication, symptom, function, hospitalization, transplantation, or survival benefit was tested or established.

A secondary trial analysis is not a new real world cohort

A 2024 report examined the subgroup taking integrase inhibitor based antiretroviral therapy in an existing randomized liver fat trial. Thirty-eight participants met that baseline criterion; fifteen active and sixteen placebo participants completed twelve months. The report found favorable body composition differences. Russo and colleagues, primary abstract

This addresses an important population question, but it does not constitute a newly recruited observational cohort or independent replication of the parent study. Restricting a trial to a smaller subgroup also changes precision and makes balance and missing follow-up more important to examine. The abstract cannot settle every question about the subgroup’s analysis plan.

Important limitations

• The cohort was small, drawn from two specialist US centers, and restricted to adults with HIV and imaging-defined fatty liver.
• The registry and paper use different participant sets for primary-outcome presentation.
• Biopsy analyses involved only those with paired samples and were secondary.
• The study was not designed as a definitive NASH, fibrosis-reversal, or clinical-outcomes trial.
• Multiple secondary and subgroup analyses increase the chance of isolated positive findings.
• Twelve randomized months do not establish long-term persistence of benefit or long-latency safety.
• The study lacked power for uncommon adverse events.

The trial question and EGRIFTA labeling are different records

The NAFLD trial evaluated liver-fat and histology outcomes in a selected population with HIV. By comparison, the FDA EGRIFTA WR label revised March 2025 describes an indication for reducing excess abdominal fat in adults with HIV-associated lipodystrophy and states that the product is not indicated for weight-loss management. The publication of the NAFLD trial does not itself amend that labeled indication.

The same label expressly states that EGRIFTA WR and EGRIFTA SV are not substitutable, identifying differences in their strengths, preparation, and storage requirements. This provides a concrete reason to retain formulation details when interpreting tesamorelin evidence. A shared ingredient name or nominal amount is insufficient to establish that another preparation matches either approved product or the historical study intervention. These are evidence and product-identity distinctions, not instructions for using either formulation.

What this study does not prove

NCT02196831 does not prove that tesamorelin prevents cirrhosis, liver cancer, hospitalization, transplantation, or death; reverses established fibrosis; improves the overall NAFLD activity score; or benefits people without HIV. It does not turn statistically significant liver-fat change into established clinical benefit.[1][2]

It also does not validate another tesamorelin formulation, another growth-hormone-releasing peptide, a blend, a compounded preparation, or a catalog vial. Name similarity cannot establish identity, purity, stability, exposure, or comparability to the finished intervention studied here.

Frequently asked questions

Were 61 or 60 people randomized?

The paper says 61 were assigned, 31 active and 30 placebo; one active-assigned participant withdrew before exposure. The registry flow therefore starts 30 per arm while reporting 61 enrolled. Both source conventions should be stated.[1][2]

What was the primary efficacy endpoint?

It was change in magnetic-resonance-measured hepatic fat fraction through month 12, not fibrosis reversal or a clinical liver outcome.[1]

How large was the liver-fat difference?

The modeled net difference was −4.1 percentage points, with a 95% confidence interval from −7.6 to −0.7.[1]

Did tesamorelin reverse fibrosis?

No significant fibrosis improvement was shown. The secondary paired-biopsy signal concerned fewer progression events, 2/19 versus 9/24, not reversal of existing fibrosis.[1]

Did glucose measures prove there was no metabolic risk?

No. Mean fasting glucose and hemoglobin A1c did not differ significantly, but hyperglycemia events occurred in both groups and prompted two active-group discontinuations.[1][2]

Does the six-month extension preserve the randomized comparison?

No. All extension participants received active study treatment, so that phase can describe follow-up observations but cannot maintain the original concurrent placebo contrast.

Does this trial validate another tesamorelin or peptide product?

No. Its findings are bounded to the finished study intervention, population, comparator, measurements, and follow-up; they do not authenticate another formulation, supplier, lot, or peptide.

Primary and official sources

1. Primary full article, PMCID PMC6981288
2. ClinicalTrials.gov: NCT02196831
3. PubMed: PMID 31611038
4. Publisher version of record, DOI 10.1016/S2352-3018(19)30338-8