Did semaglutide work differently in men vs. women in the FLOW trial?

No statistically significant difference was found. The hazard ratio was 0.74 in men and 0.81 in women, with an interaction p-value of 0.56. The wider confidence interval in women reflects lower enrollment (32% female), not a reduced drug effect.

Was semaglutide's kidney benefit limited to patients with obesity?

No. Patients with BMI below 30 and those with BMI 30 or above had identical hazard ratios (0.76 for both). The kidney protection from semaglutide was independent of baseline weight status.

Does semaglutide still protect the kidneys in patients already taking an SGLT2 inhibitor?

The subgroup on SGLT2 inhibitors at baseline (n = 529) showed a numerically larger benefit (HR 0.65), though the interaction was not statistically significant (p = 0.35). The data are consistent with additive benefit but not conclusive.

Did older patients respond differently to semaglutide in FLOW?

No. The hazard ratio was 0.73 in patients under 65 and 0.80 in those 65 and older, with a non-significant interaction p-value of 0.53.

What eGFR range showed the strongest kidney protection with semaglutide?

Numerically, patients with eGFR <30 (HR 0.73) and eGFR 30, 44 (HR 0.72) showed the largest reductions. However, the interaction across eGFR strata was not significant (p = 0.67), so these differences may reflect event rate patterns rather than true effect modification.

Were there racial or ethnic differences in semaglutide's kidney benefit?

Point estimates were consistent across White (HR 0.78), Black (HR 0.68), and Asian (HR 0.72) subgroups, with an interaction p-value of 0.90. Black participants were underrepresented (6.5%), limiting the precision of that estimate.

Does baseline albuminuria level change semaglutide's effectiveness for CKD?

Both low (<300 mg/g UACR) and high (≥300 mg/g) albuminuria subgroups showed benefit, with no significant interaction (p = 0.77).

Why was the FLOW trial stopped early, and how does that affect subgroup interpretation?

FLOW was stopped at a pre-planned interim analysis after meeting its primary efficacy endpoint at a median of 3.4 years. Early stopping can inflate effect sizes and reduce the power of subgroup analyses. Subgroup differences should be interpreted as hypothesis-generating rather than definitive.

Should clinicians use FLOW subgroup data to decide who gets semaglutide for CKD?

No subgroup showed a lack of benefit. The data support offering semaglutide to all eligible patients with T2D and CKD, regardless of age, sex, BMI, race, or baseline eGFR, rather than restricting it to specific subpopulations.

What guidelines now recommend GLP-1 RAs for kidney protection in type 2 diabetes?

Both the KDIGO 2024 Clinical Practice Guideline for Diabetes Management in CKD and the ADA Standards of Care include GLP-1 RAs (specifically semaglutide, citing FLOW) as recommended therapy for patients with T2D and CKD who have not met glycemic targets on first-line agents.

FLOW Subgroup Analyses: Who Responded Most and Least to Semaglutide for Kidney Protection

Q: Were there racial or ethnic differences in semaglutide's kidney benefit?

Point estimates were consistent across White (HR 0.78), Black (HR 0.68), and Asian (HR 0.72) subgroups, with an interaction p-value of 0.90. Black participants were underrepresented (6.5%), limiting the precision of that estimate.

Q: Does baseline albuminuria level change semaglutide's effectiveness for CKD?

Both low (<300 mg/g UACR) and high (≥300 mg/g) albuminuria subgroups showed benefit, with no significant interaction (p = 0.77).

Q: Why was the FLOW trial stopped early, and how does that affect subgroup interpretation?

FLOW was stopped at a pre-planned interim analysis after meeting its primary efficacy endpoint at a median of 3.4 years. Early stopping can inflate effect sizes and reduce the power of subgroup analyses. Subgroup differences should be interpreted as hypothesis-generating rather than definitive.

Q: Should clinicians use FLOW subgroup data to decide who gets semaglutide for CKD?

No subgroup showed a lack of benefit. The data support offering semaglutide to all eligible patients with T2D and CKD, regardless of age, sex, BMI, race, or baseline eGFR, rather than restricting it to specific subpopulations.

Q: What guidelines now recommend GLP-1 RAs for kidney protection in type 2 diabetes?

Both the KDIGO 2024 Clinical Practice Guideline for Diabetes Management in CKD and the ADA Standards of Care include GLP-1 RAs (specifically semaglutide, citing FLOW) as recommended therapy for patients with T2D and CKD who have not met glycemic targets on first-line agents.

By HealthRX Medical Team

Published May 25, 2026Updated May 25, 2026Last reviewed May 25, 2026

At a glance

| Detail | Value | |---|---| | Trial | FLOW (Evaluate Renal Function with Semaglutide Once Weekly) | | N | 3,533 | | Intervention | Subcutaneous semaglutide 1.0 mg once weekly | | Comparator | Matched placebo | | Duration | Median 3.4 years (stopped early for efficacy) | | Primary endpoint | Composite of onset of kidney failure, sustained ≥50% eGFR decline, sustained eGFR <15 mL/min/1.73 m², or renal or cardiovascular death | | Key result | HR 0.76 (95% CI 0.66, 0.88), p = 0.0003 |

Why Subgroup Data Matters More Than the Headline

A 24% relative risk reduction in a kidney composite endpoint is a strong topline result. But clinicians do not treat averages. They treat a 72-year-old woman with a BMI of 26, or a 55-year-old man with an eGFR of 32 and heavy albuminuria. The FLOW trial enrolled a broad population of adults with type 2 diabetes and chronic kidney disease, so the pre-specified subgroup analyses offer a practical map for real-world prescribing decisions.

This page walks through what those analyses showed, where the point estimates diverged (even when interaction tests did not reach significance), and what the data do and do not tell us about tailoring semaglutide therapy.

Pre-Specified Subgroup Design

FLOW's statistical analysis plan defined subgroup analyses by the following baseline variables:

Age: <65 vs. ≥65 years
Sex: male vs. female
Race: White, Black, Asian, other
Region: North America, Europe, Asia, other
BMI: <30 vs. ≥30 kg/m²
Baseline eGFR: <30, 30, <45, 45, <60, ≥60 mL/min/1.73 m²
Baseline UACR: <300 vs. ≥300 mg/g (below vs. at/above severely increased range)
HbA1c: <7.5% vs. ≥7.5%
Insulin use at baseline: yes vs. no
Background RAAS inhibitor use: yes vs. no
Background SGLT2 inhibitor use: yes vs. no

Each subgroup was tested for interaction with the treatment effect on the primary composite. All interaction p-values were reported as two-sided, without multiplicity adjustment, consistent with ICH E9 guidance on exploratory subgroup analysis in confirmatory trials.

Subgroup-by-Subgroup Results

Age

| Subgroup | n | HR (95% CI) | Interaction p | |---|---|---|---| | <65 years | 1,745 | 0.73 (0.58, 0.91) |, | | ≥65 years | 1,788 | 0.80 (0.66, 0.97) | 0.53 |

The point estimate slightly favored younger patients, but the confidence intervals overlap broadly and the interaction test was non-significant (p = 0.53). This is reassuring: semaglutide's kidney benefit does not appear to attenuate in older adults, a group that makes up the majority of CKD patients seen in practice. Age alone should not drive the prescribing decision.

Sex

| Subgroup | n | HR (95% CI) | Interaction p | |---|---|---|---| | Male | 2,400 | 0.74 (0.63, 0.88) |, | | Female | 1,133 | 0.81 (0.62, 1.06) | 0.56 |

Women comprised roughly 32% of the trial population, reflecting real-world enrollment challenges in kidney trials. The female subgroup's confidence interval crosses 1.0, but the interaction p-value (0.56) provides no evidence that the treatment effect differs by sex. The wider interval in women is a power issue, not a biological signal. The FDA label for semaglutide does not recommend sex-based dose adjustment.

Race and Ethnicity

| Subgroup | n | HR (95% CI) | Interaction p | |---|---|---|---| | White | 2,290 | 0.78 (0.66, 0.92) |, | | Black | 230 | 0.68 (0.39, 1.17) |, | | Asian | 716 | 0.72 (0.53, 0.97) |, | | Other | 297 | 0.78 (0.48, 1.27) | 0.90 |

Black participants made up only 6.5% of the cohort, limiting interpretability. The point estimate (HR 0.68) was numerically the most favorable of any racial subgroup, but the confidence interval is wide. The Asian subgroup showed a statistically significant benefit (HR 0.72, upper bound 0.97). The overall interaction p-value of 0.90 indicates treatment consistency across racial groups. This matters because CKD progression rates differ by race due to a mix of genetic, socioeconomic, and systemic factors, yet semaglutide's relative benefit did not vary meaningfully.

BMI

| Subgroup | n | HR (95% CI) | Interaction p | |---|---|---|---| | BMI <30 kg/m² | 1,456 | 0.76 (0.62, 0.94) |, | | BMI ≥30 kg/m² | 2,077 | 0.76 (0.63, 0.91) | 0.97 |

The hazard ratios are nearly identical (both 0.76), producing an interaction p-value of 0.97. This is one of the most clinically meaningful subgroup findings in FLOW. GLP-1 receptor agonists are often perceived as "weight loss drugs," which can create a prescribing bias where leaner patients are less likely to receive them. These data argue against that bias: kidney protection from semaglutide is independent of baseline obesity status.

Baseline eGFR

| Subgroup | n | HR (95% CI) | Interaction p | |---|---|---|---| | eGFR <30 | 410 | 0.73 (0.54, 0.99) |, | | eGFR 30, <45 | 1,324 | 0.72 (0.59, 0.89) |, | | eGFR 45, <60 | 1,346 | 0.80 (0.63, 1.02) |, | | eGFR ≥60 | 453 | 0.89 (0.57, 1.39) | 0.67 |

The numerically strongest effect appeared in the lowest eGFR strata (eGFR <30: HR 0.73; eGFR 30, <45: HR 0.72). As baseline kidney function improved, the point estimate moved toward 1.0. The interaction test was non-significant (p = 0.67), so these differences could be attributable to event rates rather than differential drug effect. Still, the pattern is clinically intuitive: patients closest to kidney failure have the most events to prevent, and semaglutide delivered in that window.

This gradient is worth watching in future analyses. The KDIGO 2024 guidelines now recommend GLP-1 RAs for patients with T2D and CKD who do not reach individualized glycemic targets on metformin and SGLT2 inhibitors, but the eGFR stratum data from FLOW may eventually support earlier use in advanced CKD.

Baseline Albuminuria (UACR)

| Subgroup | n | HR (95% CI) | Interaction p | |---|---|---|---| | UACR <300 mg/g | 853 | 0.73 (0.52, 1.02) |, | | UACR ≥300 mg/g | 2,680 | 0.77 (0.66, 0.90) | 0.77 |

Both subgroups showed benefit, and the interaction p-value (0.77) was non-significant. The lower UACR subgroup's confidence interval crosses 1.0 because fewer events accrued in that stratum. Semaglutide's albuminuria-lowering effect has been documented independently in the SUSTAIN 6 post-hoc kidney analysis, suggesting the drug acts on multiple renal pathways beyond protein excretion alone.

Background SGLT2 Inhibitor Use

| Subgroup | n | HR (95% CI) | Interaction p | |---|---|---|---| | SGLT2i at baseline: Yes | 529 | 0.65 (0.44, 0.97) |, | | SGLT2i at baseline: No | 3,004 | 0.79 (0.68, 0.92) | 0.35 |

This subgroup is the one that generates the most practical questions. About 15% of participants were on SGLT2 inhibitors at baseline. The point estimate was numerically better in the SGLT2i-treated group (HR 0.65 vs. 0.79), though the interaction test was not significant (p = 0.35) and the on-SGLT2i group was small.

The additive (or potentially complementary) benefit of combining a GLP-1 RA with an SGLT2 inhibitor is biologically plausible. SGLT2 inhibitors reduce glomerular hyperfiltration via tubuloglomerular feedback, while semaglutide likely acts through anti-inflammatory and anti-fibrotic pathways. The EMPA-KIDNEY trial established SGLT2 inhibitor benefit across a similar CKD population. Combining data from both trials supports a dual-therapy approach, though a dedicated combination trial has not been completed.

Post-Hoc and Exploratory Findings

Beyond pre-specified subgroups, the FLOW investigators explored effect modification by:

Duration of diabetes: Patients with longer diabetes duration (≥15 years) showed numerically similar benefit to those with shorter duration.
Cardiovascular disease history: The primary kidney composite reduction was consistent regardless of prior CV events, reinforcing that semaglutide's renal effect is not merely a downstream benefit of cardiovascular protection.
Insulin use: The ~60% of patients on insulin at baseline had comparable kidney protection to those not on insulin, suggesting the benefit is not mediated purely through glycemic control improvements.

What the Subgroup Data Do Not Show

Three cautions are worth flagging for clinicians interpreting these results.

First, none of the interaction tests reached statistical significance. This means no subgroup was identified where semaglutide definitively works better or worse. The absence of interaction is not proof of uniform effect. It may reflect inadequate power in smaller subgroups.

Second, FLOW was stopped early (median 3.4 years of a planned 5-year trial) after crossing efficacy boundaries at a planned interim analysis. Early stopping can inflate overall effect estimates and compress subgroup differences. The FLOW trial authors acknowledged this as a limitation.

Third, subgroup analyses are hypothesis-generating. A clinician should not conclude that semaglutide "works best" in patients with eGFR <30 or in those already on SGLT2 inhibitors based on these data alone. These point estimates are signals for future study, not prescribing algorithms.

Practical Takeaways for Prescribers

The consistency across subgroups carries a straightforward clinical message. There is no demographic or clinical characteristic in the FLOW data that should deter a prescriber from offering semaglutide to a patient with T2D and CKD who meets eligibility criteria. The BMI-independence finding is particularly worth internalizing: patients with a BMI under 30 benefited equally. Similarly, the benefit in the eGFR <30 stratum supports use in patients approaching kidney failure, a group historically excluded from or undertreated with newer therapies.

For combination therapy, the SGLT2i subgroup data are encouraging but not definitive. Current ADA Standards of Care and KDIGO guidelines support GLP-1 RA use alongside SGLT2 inhibitors in T2D with CKD. FLOW's subgroup data are consistent with, though not sufficient to confirm, additive benefit.

Frequently asked questions

›

References

Perkovic V, Tuttle KR, Engvig A, et al. Effects of semaglutide on chronic kidney disease in patients with type 2 diabetes. N Engl J Med. 2024;391(2):109-121. PubMed
Kidney Disease: Improving Global Outcomes (KDIGO) Diabetes Work Group. KDIGO 2024 Clinical Practice Guideline for Diabetes Management in Chronic Kidney Disease. Kidney Int. 2024;105(4S):S1-S127. PubMed
Marso SP, Bain SC, Consoli A, et al. Semaglutide and cardiovascular outcomes in patients with type 2 diabetes. N Engl J Med. 2016;375(19):1834-1844. PubMed
The EMPA-KIDNEY Collaborative Group. Empagliflozin in patients with chronic kidney disease. N Engl J Med. 2023;388(2):117-127. PubMed
American Diabetes Association Professional Practice Committee. Standards of Care in Diabetes, 2024. Diabetes Care. 2024;47(Suppl 1):S1-S321. PubMed
Semaglutide injection prescribing information. Novo Nordisk. Revised 2024. FDA Label