So far, yes, within limits. Four published studies have found no new safety problem. The largest followed 354 people. That is enough to see common effects. It is not enough to see rare ones.
The evidence comes in two kinds. One randomized trial compared semaglutide with placebo. It was reported twice, once for body fat and once for cognition. Two cohorts followed people starting semaglutide in routine HIV care. Mood is covered for the general population in the depression and anxiety guide. This page covers the HIV-specific record.
The trial: 108 people, 32 weeks
The trial enrolled people with controlled HIV-1 and a BMI of 25 or more [1]. All had lipohypertrophy, a build-up of abdominal fat. Nobody had diabetes. Participants were randomized 1:1. The semaglutide dose rose over 8 weeks and then held at 1.0 mg weekly for 24 weeks.
Visceral fat fell 30.6% against placebo. Total body fat fell 18.9%. Eight people in each arm, 15%, withdrew early.
Adverse event rates did not differ significantly between arms. Three events are named in the results. One grade 4 lipase elevation was judged related to semaglutide. Two cases of gallstones, grades 1 and 2, were judged possibly related. Pancreas and gallbladder events across the drug class are covered in the pancreatitis guide and the gallbladder evidence. The authors call for larger trials to examine serious adverse events.
Cognition in the same trial
A second report measured thinking skills in the same 108 people [2]. Against placebo, three scores rose at 32 weeks. They were visuospatial, naming and language, and delayed recall. After adjusting for sex and CD4 count, only visuospatial stayed significant. The wider question sits in the dementia evidence.
Routine care: weight and blood sugar
A US cohort followed 222 people with HIV who started semaglutide [3]. At baseline, 97% were on antiretroviral therapy. 89% had a suppressed viral load. 77% had diabetes. Mean follow-up was 1.1 years.
Weight fell 6.47 kg at one year. HbA1c fell 1.07 points in the 157 people with a follow-up test. The authors call both results comparable to the general population.
Routine care: depression scores
Depression is common in people with HIV. A second cohort asked whether semaglutide made it worse [4]. It followed 354 people across nine clinic sites. Scores came from a standard nine-item questionnaire collected at routine visits.
The overall change was -0.1 points. The 95% confidence interval ran from -0.7 to 0.5. No subgroup worsened by BMI, diabetes or antidepressant use.
That limit matches other psychiatric evidence on this site. See the bipolar cohort and the effort and reward trial.
What the four studies do not cover
The trial excluded people with diabetes. Every study group had controlled or treated HIV. The largest group was 354 people. A harm affecting one person in a thousand would likely go unseen. A prescriber who knows the full medication list is the right check before starting.