Safety and effectiveness in an uncontrolled setting of glucagon‐like‐peptide‐1 receptor agonists in patients with familial partial lipodystrophy: Real‐life experience from a national reference network

Abstract

Aim

To describe the effects of Glucagon‐like peptide‐1 receptor agonists (GLP‐1RA) in patients with familial partial lipodystrophy (FPLD) assessed in a real‐life setting in a national reference network.

Patients and Methods

We retrospectively collected clinical and metabolic parameters in patients with FPLD in the French lipodystrophy reference network, who initiated GLP‐1RA. Data were recorded before, at one‐year (12 ± 6 months) and at the latest follow‐up on GLP‐1RA therapy (≥18 months).

Results

Seventy‐six patients (89.4% of women), diagnosed with LMNA‐related FPLD2 (n = 57), PPARG‐related FPLD3 (n = 4), PLIN1‐related FPLD4 (n = 5) or FPLD1 (n = 10) initiated GLP‐1RA therapy between 2008 and 2024. Patients were aged a median (IQR) 48 years (34.5–57), body mass index (BMI) was 26.0 kg/m² (23.9–29.5), HbA1c 8.3% (7.5–9.3), triglycerides 2.31 mmol/L (1.62–3.88). GLP‐1RA were used in addition to previously used antidiabetics, 50% of patients being insulin‐treated. After one year with GLP‐1RA therapy, BMI, HbA1c and triglycerides significantly decreased to 25.6 kg/m² (22.7–29.1), 7.3% (6.6–8.3) and 1.97 mmol/L (1.5–3.2) respectively (p < 0.001, p < 0.001 and p < 0.01, respectively), without significant changes in other antidiabetic and lipid‐lowering drugs. Gamma‐glutamyl‐transferase and alanine‐aminotransferase levels also significantly decreased. Effects on HbA1c, BMI and triglycerides persisted in the long term. One case of acute pancreatitis occurred during follow‐up, associated with severe hypertriglyceridemia in a non‐observant patient. Gastrointestinal symptoms affected 34% of patients, leading to GLP‐1RA withdrawal in six patients.

Conclusion

GLP‐1RA significantly improved BMI, HbA1c and triglycerides in a large majority of patients with FPLD. Larger and prospective controlled studies are warranted for identification of predictive factors and safety.

1. INTRODUCTION

Lipodystrophy syndromes (LD) are rare disorders characterized by the selective loss of adipose tissue, leading to metabolic complications including severe insulin resistance, dyslipidemia, hepatic steatosis, atherosclerotic events and ovarian hyperandrogenism in women. ¹ , ² , ³ Familial partial lipodystrophy (FPLD), characterized by a predominant subcutaneous peripheral lipoatrophy, presents unique therapeutic challenges. ⁴ , ⁵ , ⁶ , ⁷ The current therapeutic landscape mainly focusses on lifestyle modifications and pharmacologic interventions primarily targeting hyperglycemia and dyslipidemia. ¹ , ⁸ However, classical treatments often fall short in managing the multifaceted metabolic dysfunctions inherent to FPLD. High‐dose insulin therapy using multi‐injections and/or continuous subcutaneous infusion, even with U500‐concentrated insulin is not consistently efficient to overcome FPLD‐associated severe insulin resistance. ⁸ , ⁹ The orphan drug metreleptin, a recombinant analog of leptin, obtained a marketing authorization in Europe for patients with FPLD over 12 years of age who did not achieve adequate metabolic control with conventional treatments. ¹⁰ However, its metabolic effects, which are known to be tempered in FPLD as compared with generalized forms of lipodystrophy, did not reach statistical significance in the French real‐life cohort of patients with FPLD. ¹¹ The challenge of effective metabolic treatment is all the more important given the earlier occurrence of metabolic complications over young generations observed in families with FPLD. ¹²

Glucagon‐like peptide‐1 receptor agonists (GLP‐1RA), originally developed to treat type 2 diabetes, have garnered attention for their potential in managing lipodystrophy syndromes. GLP‐1RA enhance glucose‐dependent insulin secretion, suppress glucagon release, slow gastric emptying and promote satiety. Weight loss resulting from GLP‐1RA therapy was shown to be due to a reduction in fat mass, rather than lean mass, with some studies reporting a preferential decrease in visceral fat ¹³ , ¹⁴ and others in subcutaneous fat. ¹⁵ Recent studies have illuminated the pleiotropic effects of GLP‐1RA, extending beyond glucose metabolism to beneficial impacts on lipid profile, hepatic fat content and cardiovascular risk factors. ¹⁶ , ¹⁷ Altogether, these findings render GLP‐1RA promising candidates for addressing therapeutic issues in FPLD. Moreover, circulating levels of dipeptidyl peptidase‐4 (DPP4), which cleaves active GLP‐1 to inactive forms, could be higher in patients with LMNA‐related FPLD (FPLD2) than in healthy subjects, reinforcing the potential therapeutic interest of GLP‐1RA in affected patients. ¹⁸ Although the risk of acute pancreatitis, based on meta‐analyses of randomized placebo‐controlled trials of GLP‐1RA in type 2 diabetes, is not significantly increased, ¹⁹ , ²⁰ , ²¹ the specific risk of severe hypertriglyceridemia and acute pancreatitis associated with FPLD prompts clinicians to act with caution in using incretin‐based therapies in affected patients, in the absence of safety data in this population. ⁹ Also, the weight loss associated with GLP‐1RA therapy could worsen lipoatrophy in this specific population.

Despite the growing interest, the evidence base for the efficacy and safety of GLP‐1RA in FPLD remains limited to case reports, and to a recent retrospective study investigating 14 patients with FPLD. ²² , ²³ , ²⁴ , ²⁵ The present study aims to evaluate the therapeutic impact and safety of GLP‐1RA therapy in patients with FPLD, based on the real‐world experience within a national rare disease reference network. By analysing clinical outcomes, including glucose control, lipid profile and liver parameters and overall safety in 76 patients with FPLD, this study seeks to inform clinical practice and guide future research directions.

2. PATIENTS AND METHODS

2.1. Study design and participants

This multicentre retrospective observational study included all patients with FPLD who initiated GLP‐1RA therapy between 2008 and 2024 in adult endocrinology centres from the National Reference network for Rare Diseases of Insulin Secretion and Insulin Sensitivity (PRISIS), France.

Data collection was coordinated by the AP‐HP Saint‐Antoine PRISIS centre. Patients' files were included in the BaMaRa National Rare Disease Database (French data protection agency CNIL # 2211418). The study followed the principles of the Declaration of Helsinki and all patients gave their informed consent for data collection.

Familial partial lipodystrophy syndromes were classified based on the responsible pathogenic variant in the genes LMNA (FPLD2), PPARG (FPLD3) or PLIN1 (FPLD4). ¹ Patients with clinical characteristics of FPLD without any known identifiable monogenic cause, and with a body mass index (BMI) under 30 kg/m² were classified as having the polygenic FPLD1 form. ²⁶

GLP‐1RA therapy was initiated by the usual physician according to her/his clinical judgement. Treatment was administered by subcutaneous injection and dose adjustments were based on patients' metabolic response and tolerance as recommended. Liraglutide was administered at a daily dose of 0.6 to 1.8 mg. Dulaglutide and semaglutide were administered once a week, at doses comprised between 0.75 and 4.5 mg, or 0.25 and 1 mg, respectively.

2.2. Methods

Patients' records were reviewed using structured data collection forms. The following variables were collected at baseline (i.e, within 2 years before initiation of GLP‐1RA therapy), after 12 ± 6 months of GLP‐1RA therapy (short‐term response) and at the latest visit on GLP‐1RA therapy, after a minimum of 18 months of treatment (long‐term response):

1. Gender, weight, BMI, ongoing treatments (lipid‐lowering, antihypertensive and antidiabetic drugs), daily insulin dose and daily GLP‐1RA dose.

2. HbA1c, liver enzymes (aspartate aminotransferase [AST], alanine‐aminotransferase [ALT], gamma‐glutamyl‐transferase [GGT]), FIB‐4 index (calculated as age × AST/ platelet × [ALT]^1/2), total cholesterol, high‐density lipoprotein (HDL‐) and low‐density lipoprotein (LDL‐) cholesterol, fasting triglycerides, MDRD‐estimated glomerular filtration rate (eGFR), urinary albumin‐to‐creatinine ratio (UACR), body fat percentage evaluated with dual energy x‐ray absorptiometry (DEXA) or impedancemetry and vibration‐controlled transient elastography (VCTE) performed using a FibroScan® (M Probe; XL Probe; Echosens, Paris, France), when available.

At baseline, additional parameters were recorded, including FPLD subtype/genotype gene when available, the presence of hypertension, dyslipidaemia and/or diabetes and the age at diagnosis of diabetes and age at initiation of GLP‐1RA treatment.

Diabetes was defined by an HbA1c concentration ≥6.5% (≥48 mmol/mol) or at least one antidiabetic medication. Hypertension was defined by systolic blood pressure (SBP) ≥140 mmHg and/or diastolic blood pressure (DBP) ≥90 mmHg, or at least one antihypertensive treatment. Dyslipidaemia was defined by serum fasting triglycerides ≥1.7 mmol/L, or LDL cholesterol ≥4.88 mmol/L or HDL cholesterol ≤1.03 mmol/L (men) or ≤1.28 mmol/L (women), or at least one lipid‐lowering therapy.

We also collected and reviewed serious adverse events that occurred during GLP‐1RA therapy.

In patients with diabetes at baseline, a positive response to GLP‐1RA treatment on glucose control was defined (1) as a ≥0.5 point (pt) decrease in HbA1c without increasing insulin dose ≥20% or adding an antidiabetic class or (2) as HbA1c stability with a decrease of more than 20% in total daily insulin or discontinuation of at least one antidiabetic class, between baseline and short‐term GLP‐1RA therapy. A positive response to GLP‐1RA on triglycerides was defined, in patients with serum fasting triglycerides ≥1.7 mmol/L at baseline, as serum triglyceride levels <1.7 mmol/L, or as a decrease of more than 30% in serum triglycerides between baseline and short‐term GLP‐1RA therapy. ¹¹

2.3. Statistical analysis

Results are presented as medians and interquartile ranges (25th–75th percentile) for quantitative variables and as numbers and percentages for qualitative variables. Missing values were imputed using the last observation carried forward (LOCF) method.

Changes from baseline to short and long‐term follow‐up were assessed using either one‐sided paired t‐tests or one‐sided paired Wilcoxon tests, following visual data inspection and Shapiro–Wilk tests to evaluate normal distribution.

In order to distinguish responders from non‐responders on glucose and/or triglyceride control as defined above, we conducted two‐sided independent tests.

All descriptive and statistical analyses were performed using R (version 4.2.1, Foundation for Statistical Computing, Vienna, Austria. https://www.R-project.org/). Statistical significance was set at a type I error of 5%.

3. RESULTS

3.1. Baseline characteristics

In the French PRISIS network, we collected data from 77 patients affected by FPLD and who have been treated by GLP‐1RA between 2008 and 2024. No data were available for one patient who was excluded from the study. Among the 76 patients included in the study, follow‐up data were available for 64 at short term and 43 at long term (Figure 1).

FIGURE 1.

Study flow diagram. Short term = 12 ± 6 months of GLP‐1RA therapy. Long term = last visit, more than 18 months of GLP‐1RA therapy.

Women represented 89.5% of the cohort, and 57 patients were diagnosed with FPLD2 (75% of the cohort). Most patients with FPLD2 (i.e., 50.9%) carried a hotspot LMNA variant at position 482 (24 patients with LMNA p.[Arg482Trp], 5 with LMNA p.[Arg482Gln]). The specific founder variant LMNA p.(Thr655Asnfs*49) was identified in 15 patients originating from Reunion Island (26.3% of patients with FPLD2). ²⁷ Median age of patients at initiation of GLP‐1RA was 48 years old (34.5–57), and median BMI was 26.0 kg/m² (23.9–29.5) (Table 1 and Table 2). The percentage of total body fat was moderately low, although variable (in women, 26.1%, [21.8–33.1]), with low leptin levels (6.9 [4.35–14.1] ng/mL) (Table 2).

TABLE 1.

Characteristics of patients at GLP‐1RA initiation (n = 76).

	At baseline (n = 76)
General characteristics
Women	68 (89.47%) (n = 76)
Age at diagnosis of diabetes (y)	27 [20.75–40] (=48)
Age at onset of aGLP1R (y)	48 [34.5–57] (n = 75)
Type of lipodystrophy
FPLD1 (polygenic form of partial lipodystrophy)	10 (13.16%) (n = 76)
FPLD2 (LMNA‐related partial lipodystrophy)	57 (75.0%) (n = 76)
LMNA p.(Arg28Trp)	3 (5.26%)
LMNA p.(Arg60Pro)	1 (1.75%)
LMNA p.(Thr150Ala)	2 (3.51%)
LMNA p.(Ala287Valfs*193)	1 (1.75%)
LMNA p.(Arg401Cys)	1 (1.75%)
LMNA p.(Arg439Cys)	1 (1.75%)
LMNA p.(Asn466Asp)	1 (1.75%)
LMNA p.(Arg482Gln)	5 (8.77%)
LMNA p.(Arg482Trp)	24 (42.11%)
LMNA p.(Arg527Pro)	1 (1.75%)
LMNA p.(Arg582Leu)	1 (1.75%)
LMNA p.(Val586Met)	1 (1.75%)
LMNA p.(Thr655Asnfs*49)	15 (26.32%)
FPLD3 (PPARG‐related partial lipodystrophy)	4 (5.26%) (n = 76)
PPARG p.(Arg165Thr)	1 (25.0%)
PPARG p.(Arg194Gln)	1 (25.0%)
PPARG p.(Arg212Gln)	1 (25.0%)
PPARG p.(Arg425Cys)	1 (25.0%)
FPLD4 (PLIN1‐related partial lipodystrophy)	5 (6.58%) (n = 76)
PLIN1 p.(Val398Glyfs*166)	1 (20.0%)
PLIN1 p.(Pro403Argfs*164)	3 (60.0%)
PLIN1 p.(Leu404Alafs*158)	1 (20.0%)

Note: Values are expressed as median (25th–75th percentile) or n (%).

Abbreviation: FPLD, familial partial lipodystrophy syndrome.

TABLE 2.

Anthropometric and metabolic characteristics of patients with FPLD during GLP‐1RA therapy (n = 76).

	At baseline (n = 76)	Short‐term response (n = 64)	Long‐term response (n = 43)	P short‐term response versus baseline	P long‐term versus short‐term baseline	P long‐term response versus baseline
Anthropometry
Body mass index (kg/m2)	26.04 [23.92–29.48] (n = 75)	25.60 [22.68–29.07] (n = 61)	25.22 [22.12–27.69] (n = 38)	<0.001	0.25	<0.001
Total fat mass (%)	25.90 [21.68–32.98] (n = 38)	23.40 [20.93–35.13] (n = 26)	24.95 [22.45–35] (n = 10)	0.70	0.34	0.15
Fat mass women	26.10 [21.75–33.10] (n = 35)	25.05 [21.18–36.30] (n = 22)	25.70 [22.10–36.10] (n = 9)	0.74	0.58	0.27
Fat mass men	24.40 [22.35–28.70] (n = 3)	21.05 [20.60–23.95] (n = 4)	24.20 [24.20–24.20] (n = 1)	0.5	0.5	0.5
Serum leptin (ng/mL)	6.9 [4.35–14.1] (n = 46)
Diabetes	73 (96.05%) (n = 76)	62 (96.88%) (n = 64)	43 (100.0%) (n = 43)
HbA1c (%)	8.30 [7.50–9.40] (n = 73)	7.30 [6.55–8.28] (n = 63)	7.30 [6.64–8.08] (n = 38)	<0.001	0.49	<0.001
Other antidiabetic treatment
Insulin alone	6 (8.22%) (n = 73)	6 (9.84%) (n = 61)	4 (9.76%) (n = 41)	1	0.53	0.53
Insulin + oral antidiabetics	32 (43.84%) (n = 73)	29 (47.54%) (n = 61)	20 (48.78%) (n = 41)	0.70	0.20	0.10
Number of oral antidiabetics therapeutic classes used	1 [1–2] (n = 76)	1 [1–2] (n = 64)	1 [1–1.5] (n = 43)
Patients treated by insulin
Daily insulin dose (IU/kg/d)	1.40 [0.73–2.25] (n = 39)	1.31 [0.69–1.94] (n = 34)	1.10 [0.72–1.93] (n = 23)	0.40	0.20	0.69
Daily insulin dose ≥1 IU/kg/d	25 (64.10%) (n = 39)	24 (70.59%) (n = 34)	13 (56.52%) (n = 23)
Patients treated by metreleptin	6 (8.22%) (n = 73)	4 (6.56%) (n = 61)	4 (10.26%) (n = 39)	0.97	0.77	0.99
Hypertension	51 (69.86%) (n = 73)	42 (64.62%) (n = 65)	36 (80.0%) (n = 45)
Dyslipidemia	67 (89.33%) (n = 75)	48 (85.71%) (n = 56)	30 (85.71%) (n = 35)
Serum triglycerides (mmol/L)	2.31 [1.62–3.88] (n = 73)	1.97 [1.53–3.20] (n = 58)	2.37 [1.39–3.44] (n = 37)	<0.01	0.23	<0.01
Total cholesterol (mmol/L)	4.21 [3.61–4.88] (n = 64)	4.0 [3.33–4.82] (n = 45)	3.68 [3.33–4.53] (n = 30)	0.06	0.82	0.23
LDL cholesterol (mmol/L)	2.22 [1.55–2.61] (n = 65)	2.09 [1.48–2.62] (n = 51)	1.81 [1.33–2.68] (n = 31)	0.23	0.66	0.08
HDL cholesterol (mmol/L)	0.88 [0.75–1.02] (n = 63)	0.90 [0.79–1.0] (n = 47)	0.90 [0.75–1.11] (n = 29)	0.62	0.83	0.69
Lipid‐lowering therapy				1	0.12
Statine	41 (57.75%) (n = 71)	33 (55.93%) (n = 59)	28 (73.68%) (n = 38)			0.08
Fibrate	13 (18.31%) (n = 71)	11 (18.64%) (n = 59)	10 (27.03%) (n = 37)	1	0.48	0.42
Ezetimibe	10 (13.89%) (n = 72)	10 (16.67%) (n = 60)	6 (16.22%) (n = 37)	0.84	1	0.97
Liver parameters
Fatty liver	49 (76.56%) (n = 64)	27 (72.97%) (n = 37)	11 (55.0%) (n = 20)
FIB‐4 score	0.77 [0.55–1.21] (n = 59)	0.85 [0.62–1.21] (n = 45)	0.94 [0.65–1.77] (n = 28)	0.37	0.98	0.99
Liver stiffness (kPa)	6.3 [5.23–8.4] (n = 18)	6.35 [5.9–7.2] (n = 20)	6.55 [5.35–8.18] (n = 18)	0.14	0.15	0.11
Liver enzymes
AST (IU/L)	25.5 [20–39.75] (n = 66)	26 [21–38] (n = 49)	27 [23–36] (n = 30)	0.29	0.87	0.98
ALT (IU/L)	32 [24–54.75] (n = 66)	34 [23–44] (n = 50)	34 [24.5–47.5] (n = 30)	0.02	0.79	0.89
GGT (IU/L)	47 [32–60] (n = 63)	40 [29.5–57.5] (n = 47)	39 [28–50] (n = 29)	<0.01	0.85	0.11
Kidney parameters
eGFR MDRD (ml/min/1.73 m2)	115.5 [90–139.55] (n = 26)	93 [76.5–109] (n = 23)	80 [64.5–110.5] (n = 19)	0.01	0.41	0.08
Urinary albumin‐to‐creatinine ratio (mg/mmoL)	9.2 [2.78–27] (n = 41)	2.71 [0.97–14.35] (n = 36)	9.35 [1.9–22.5] (n = 27)	0.33	0.97	0.95

Note: Values are expressed as n (%), or median (25th–75th percentile) as indicated. Total fat mass was measured with dual energy x‐ray absorptiometry (DEXA) or impedancemetry, fatty liver was diagnosed using echography, liver stiffness was measured by transient elastography. Diabetes is defined by HbA1c ≥6.5% or at least one antidiabetic treatment. Dyslipidemia is defined by serum triglycerides ≥1.7 mmol/L, LDL cholesterol ≥4.88 mmol/L, HDL cholesterol ≤1.03 mmol/L (men) or ≤1.28 mmol/L (women), or lipid‐lowering therapy. Hypertension was defined by systolic blood pressure (SBP) ≥140 mmHg and/or diastolic blood pressure (DBP) ≥90 mmHg, or antihypertensive treatment.

Abbreviations: ALT, alanine aminotransferase; AST, aspartate aminotransferase; eGFR, estimated glomerular filtration rate; GGT, gamma‐glutamyltransferase.

All patients were diagnosed with diabetes at baseline except three women, treated with GLP‐1RA therapy in the context of glucose intolerance, food cravings and overweight. The median age at diagnosis of diabetes was 27 years (20.75–40), and median HbA1c at initiation of GLP‐1RA was 8.3% (7.5–9.4) (Table 1 and Table 2).

Half of the patients with diabetes required insulin therapy (n = 38), at relatively high daily doses (1.4 IU/kg/d [0.7–2.2]). Oral antidiabetics were used by 90% of patients, only one of them being treated with SGLT2 inhibitors at GLP‐1RA initiation. Six patients were treated with metreleptin when started on GLP‐1RA (Table 2).

Compared to women, all men initiating with GLP‐1RA had diabetes (n = 8), which was diagnosed later (median age of 39 [34.5–43.2] vs. 26 [20–39.2] years old in women). In line, GLP‐1RA treatment was initiated later in men than in women (at 58 [53.5–68] vs. 47 [53.5–68] years old), but with similar BMI (median 25.5 [24.9–27.4] vs. 26.1 [23.4–29] kg/m²) and HbA1c (8.6 [7.0–9.7] vs. 8.2 [7.5–9.0] %), respectively.

In the whole cohort, median levels of triglyceride and LDL cholesterol were 2.31 (1.62–3.88) and 2.22 (1.55–2.61) mmol/L, respectively. The majority of patients were on lipid‐lowering therapy at initiation of GLP‐1RA (70%). Among patients with dyslipidemia (n = 67, representing 88% of the studied patients), 54 (80% of them) had fasting triglycerides ≥1.7 mmol/L (Table 2).

Over three‐quarters of investigated patients had fatty liver disease at GLP‐1RA initiation (76.6%), with median values of liver enzymes in the normal‐to‐high range. FIB‐4 score, available in most patients (n = 59), was low (0.77 [0.55–1.21]) and transient elastography, performed in 18 patients, evaluated the median liver stiffness at 6.3 kPa (5.23–8.4), indicating that the risk of advanced liver fibrosis was low in most patients (Table 2). ²⁸ , ²⁹

Regarding kidney parameters, the median level of urinary albumin‐to‐creatinine ratio was within normal range, whereas the median eGFR was in favour of glomerular hyperfiltration (Table 2).

At the cardiovascular level, 35% of the patients were affected at baseline with ischemic heart disease and/or arrhythmia (n = 19 and n = 10 patients among 75, respectively, with three patients presenting both). Before initiation of GLP‐1RA therapy, six patients were implanted with defibrillators (five with LMNA pathogenic variants and 1 with FPLD1), two had coronary artery bypass surgery and two underwent heart transplantation (all with LMNA pathogenic variants).

3.2. Response to GLP‐1RA therapy in patients with FPLD

Different molecules belonging to the GLP‐1RA therapeutic class were used during follow‐up. As expected, the use of daily GLP‐1R decreased over time and weekly GLP‐1RA were most used at long term (Supplemental Figure 1).

Short‐term and long‐term follow‐up evaluations were performed after a median duration of GLP‐1RA therapy of 9 months (6–12) (n = 64), and 44 months (28.5–69.5) (n = 43), respectively.

Table 2 shows the course of anthropometric and metabolic parameters of patients treated with GLP‐1RA. BMI significantly decreased at the short term, an effect that was maintained at the long term (median 26.0 to 25.6 and 25.2 kg/m², p < 0.001 for both vs. baseline values, respectively) (Table 2 and Figure 2A). Glucose homeostasis improved, with median HbA1c level showing a significant decrease from baseline to short‐term follow‐up, with a persistent effect over time (8.3% to 7.3, then 7.3% at long‐term follow‐up, p < 0.001 for both vs. baseline) (Table 2 and Figure 2B).

FIGURE 2.

Body mass index (BMI) (A), glycated haemoglobin (HbA1c) (B) and serum triglycerides (C) during GLP‐1RA treatment in patients with familial partial lipodystrophy (FPLD) at baseline, and after short‐ and long‐term follow‐up. NS, non‐significant. **p < 0.01, ***p < 0.001. Boxplots show median values, interquartile ranges (25th–75th percentile) with dots indicating outliers.

Fasting triglycerides decreased significantly under GLP‐1RA in the cohort, with a persistent effect at long term (Table 2, p values <0.01, considering 57 data pairs available at short term and 36 pairs available at long term). At long‐term follow‐up, mean triglyceride levels were reduced by 11.4%. Median levels of total and LDL‐cholesterol levels were not elevated at baseline in patients with FPLD (4.21 [3.61–4.88] and 2.22 [1.55–2.61] mmol/L, respectively) and tended to decrease under GLP1RA therapy (Table 2). Median HDL‐cholesterol levels were low at baseline (0.88 [0.75–1.02]) and did not significantly change during GLP‐1RA therapy (Table 2). We did not observe any significant change in the use of lipid‐lowering drugs during GLP‐1RA therapy.

The number of insulin‐treated patients did not significantly change during GLP‐1RA therapy (52% at baseline (38/73), 55% at the short term (35/61) and 59% at the long term (24/41)), and total daily insulin dose was comparable at baseline, short‐term and long‐term GLP‐1RA therapy (1.40, 1.31 IU/kg, then 1.10 IU/kg). Insulin was initiated in eight patients, for most of them during the first year of GLP‐1RA therapy. Three patients discontinued insulin treatment after introduction of GLP‐1RA therapy, with subsequent HbA1c remaining stable (at 7% in two patients), or decreasing (in one patient, from 9.9 to 8.8% after five years of follow‐up). The number of oral antidiabetic therapeutic classes used by patients with FPLD did not significantly change during follow‐up on GLP‐1RA (i.e., 1 [1, 2], 1 [1, 2] and 1 [1–1.5] at baseline, short term and long term, respectively).

Six patients were treated with metreleptin when started on GLP‐1RA. Three of them discontinuated metreleptin during follow‐up on GLP‐1RA, while metreleptin was initiated during GLP‐1RA therapy in another patient.

Regarding liver parameters, AST remained within normal‐to‐high ranges during GLP‐1RA therapy, whereas medians ALT and GGT levels decreased significantly at short term follow‐up (p = 0.02 considering 46 pairs and, p < 0.01 considering 41 pairs, respectively). We did not observe any significant changes in liver fibrosis markers during GLP‐1RA therapy, as evaluated by the FIB‐4 score or measured, in a limited number of patients, by transient elastography.

In our study, an improvement in kidney function was observed with a decrease in hyperfiltration attested by the decrease of eGFR from 115.5 to 93 mL/min/1.73m² at short term (p = 0.01) and 80 mL/min/1.73m² at long term (p = 0.08 vs. baseline). We did not observe any significant changes in UACR values, which were in the normal range for most patients.

Some patients switched from metreleptin to GLP‐1RA therapy (n = 3). In all three patients, HbA1c improved from 8% to 7.5%, 7.4% to 6.8% and 9.5% to 8.4%, respectively, as did triglyceride levels (from 3.88 to 2.66 mmol/L, 2.17 to 1.53 mmol/L and 16 to 6 mmol/L).

In an attempt to determine predictive factors of treatment efficacy, we classified patients according to their response on glucose control or triglyceride levels. Data, as defined in Methods, were available in 61 patients after a median 9 months [6–12] of GLP‐1RA therapy. Around a quarter of patients had a positive response on both glucose and triglyceride control (n = 15), and one‐third (n = 19) were responders for triglyceride levels only. Fifteen patients had a positive response on glucose control only, and 12 patients did not respond for either glucose or triglyceride values (Supplemental Figure 2). Responders and non‐responders did not differ according to gender, type of partial lipodystrophy, time between onset of diabetes and initiation of GLP‐1RA therapy or presence of insulin therapy or high insulin requirements (>1 IU/kg/d) at baseline (Supplemental Table 1). Among the 35 patients with serum leptin levels and/or percentage of fat mass below the median values, 85% of them were classified as responders. The serum leptin levels and the percentage of fat mass at baseline were not different between responders and non‐responders (Supplemental Table 1). In addition, in the subgroup of patients with BMI <25 kg/m2 at baseline (n = 25), BMI and HbA1c significant decreased at short‐ and long‐term after initiation of GLP‐1RA (respectively p = 0.02 and p = 0.002, short term vs. baseline, and p = 0.005 and p < 0.001, long term vs. baseline), although triglycerides levels did not significantly changed (Supplemental Table 2). Taken together, these results suggest that the severity of clinical lipoatrophy does not predict the metabolic response to GLP‐1RA in patients with FPLD. However, although baseline BMI did not significantly differ between responders and non‐responders, BMI decreased significantly to a greater extent during GLP‐1RA therapy in responders than in non‐responders (26.3 [24.1–30.5] to 25.4 [22.4–29.6] vs. 25.9 [24.4–26.2] to 25.3 [23.7–26.8] kg/m² respectively, p = 0.04), strongly suggesting that weight reduction contributes to treatment efficacy.

3.3. Adverse events

Thirty‐four percent of patients reported adverse events under GLP‐1RA (n = 26), mainly related to gastrointestinal side effects, that is, nausea (n = 13), vomiting (n = 4), anorexia (n = 3) or to other events (n = 6, including one case of acute pancreatitis as described below, one case of ketoacidosis and four cases of non‐specific symptoms), resulting in treatment interruption in six cases (7.9%). No patient discontinued GLP‐1RA therapy due to weight loss, which was moderate in most cases.

One episode of acute pancreatitis was reported during follow‐up. It was related to severe hypertriglyceridemia in a woman with FPLD2, occurring 15 months after initiation of GLP1RA, in a context of poor therapy adherence. No other adverse event was reported.

Of note, two other women with FPLD2 were prescribed GLP‐1RA while having experienced an episode of acute pancreatitis related to hypertriglyceridemia 6 or 20 years earlier. No other episode occurred after 11 and 8 years of follow‐up on GLP‐1RA, respectively.

Twenty‐eight patients from this cohort are treated with GLP‐1RA for more than three years, with a good tolerance.

4. DISCUSSION

This study reports significant improvements in glucose and triglyceride control in patients with FPLD treated with GLP‐1RA as an add‐on therapy in a real‐life setting. Given the rarity of the disease, ³⁰ , ³¹ , ³² our study evaluating the effects of GLP‐1RA in seventy‐six patients, followed in a national reference rare disease network, provides significant insight into the management of affected patients. Our results align with those from a previous retrospective, non‐controlled study evaluating the efficacy of GLP‐1RA in 14 patients with FPLD compared with 14 patients with type 2 diabetes. ²⁵ In the latter study however, 13 patients were affected with the polygenic FPLD1 form of the disease, whereas our cohort provides data in patients affected with FPLD1 (n = 10), but also monogenic forms of FPLD, including 57 patients with LMNA‐related FPLD2, the main form of monogenic FPLD. ¹ , ³⁰ , ³² To our knowledge, only two patients with FPLD2 have been previously reported with a GLP‐1RA therapy. ²² , ²³

Importantly, although slight‐to‐moderate gastrointestinal side effects, reported in 34% of patients with FPLD, led to treatment discontinuation in six patients (7.9% of the cohort), as expected from GLP‐1RA tolerability data, ³³ only one case of acute pancreatitis occurred during follow‐up, which was attributed to hypertriglyceridemia but not to GLP‐1RA. In addition, two patients with a previous history of hypertriglyceridemia‐linked acute pancreatitis episodes did not present any relapse during several years of GLP‐1RA therapy.

Our study shows significant improvement of HbA1c in patients with FPLD treated with GLP‐1RA, with a decreased median value from 8.3% to 7.3%, an effect that was sustained at long term, without concomitant increase in other antidiabetic treatments, including insulin.

The beneficial effect on glucose control may be explained, at least partly, by the decrease in BMI, from a median 26.0 to 25.6 kg/m² after one year of GLP‐1RA therapy and 25.2 at long term. In line, BMI tended to decrease to a greater extent in patients with a positive metabolic response to GLP‐1RA as compared to non‐responders. Effects of GLP‐1RA on central satiety signals, leading to decreased caloric intake, could contribute, at least in part, to the metabolic effects observed in patients with FPLD. ³⁴ GLP‐1RA therapy could also specifically counteract the decreased endogenous levels of active GLP‐1 due to enhanced DPP4 activity reported in patients with FPLD, ²⁰ which could originate from increased visceral versus subcutaneous adipose tissue. ³⁵ GLP‐1RA are also known to improve insulin and leptin sensitization, ³⁶ fat redistribution and hepatic steatosis; all of which responding to pathophysiological features of lipodystrophy. Of note, a case reported by Banning et al described the major effect of liraglutide on insulin secretion in a patient with FPLD2. ²²

Although hypertriglyceridemia is a key feature of FPLD, ¹ only 73% of patients had more than 1.7 mmol/L of serum triglycerides in our cohort, encouraging that lipid‐lowering therapies, used by 70% of patients, could be, at least partly, efficient in patients with FPLD. The overall cohort exhibits a significant reduction in fasting triglycerides under GLP‐1RA, and the decrease was also significant for patients with baseline triglyceride levels greater than 1.7 mmol/L. This finding, consistent with other studies reporting triglyceride‐lowering effects of GLP‐1RA, ³⁷ could be particularly important in patients with FPLD who do not achieve adequate triglyceride control with standard lipid‐lowering therapy.

ALT and GGT levels showed significant reduction after one‐year GLP‐1RA therapy, suggesting a beneficial effect on hepatic steatosis. We observed a normalization of glomerular hyperfiltration, as previously reported in patients with metabolic syndrome treated with GLP‐1RA therapy. ¹⁷ , ³⁸ The stability of UACR values is in favour of the renal safety of GLP‐1RA in this patient population.

In our study, reported adverse events mainly included gastrointestinal side effects. They affected about one‐third of patients, a frequency which is comparable to that previously reported in patients treated by GLP‐1RA for diabetes and/or obesity. ³⁹ , ⁴⁰ Side effects led to treatment interruption in six patients. One patient from our study presented an episode of hypertriglyceridemia‐related acute pancreatitis, in a context of poor therapy adherence, which was not attributed to GLP‐1RA.

Our study has limitations, such as its retrospective setting. It was conducted during an extended period (2008–2024) over which GLP‐1RA and other antidiabetic drugs have evolved, as such we were not able to not compare results based on the different molecules used. Neither did we have access to precise data regarding treatment adherence or data on dietary and health measures which are potential confounding factors. However, the medical management of patients was conducted in specialized centres belonging to a national rare disease reference network, which insures coordinated and homogenized therapeutic approaches.

Larger cohort studies and double‐blind randomized controlled trials are needed to appraise metabolic efficacy of GLP‐1RA therapy in LD, to outline the predictive factors of response and to establish drug safety in this peculiar situation.

Other new antidiabetic molecules might be useful in patients with FPLD. Beneficial effects of SGLT2 inhibitors on metabolic complications has been reported in a series of 12 patients with partial lipodystrophy, ³⁶ , ⁴¹ which warrants further confirmation. Although no data are available in patients with FPLD, multi‐agonist drugs targeting the receptors for GLP‐1, glucose‐dependent insulinotropic polypeptide and glucagon, could also be promising directions in decreasing severe insulin resistance. ⁴²

In conclusion, this study provides valuable insights into the therapeutic potential of GLP‐1 receptor agonists for managing FPLD, highlighting their efficacy in improving glycemic control and lipid profiles, along with their favourable safety profile. Further large‐scale studies are needed to confirm these benefits and to explore the cardiovascular effect of this therapeutic class.

CONFLICT OF INTEREST STATEMENT

H.M. served as speaker to Amryt Pharmaceuticals (now Chiesi Farmaceutici); C.Vi. to Amryt Pharmaceuticals, Sanofi, Ipsen and Lilly; C.Va. to Abbott, Advanz Pharma, Amryt Pharmaceuticals, AstraZeneca, Lilly, Novartis, Novo Nordisk, Sanofi; MCV to Amryt and Sanofi; and C.A to Lilly, Sanofi, Novo Nordisk, Decom. SH received institutional research grants from Zeneca, Asten Santé, Air Liquide Health Science, Bayer, Boehringer Ingelheim, Eli Lilly, Homeperf, Isis Diabète LVL, Nestle Home Care, NovoNordisk, Pierre Fabre, Sanofi, Valbiotis and Vitalaire and served as consultant and/or received honoraria from Astra Zeneca, Bayer, Boehringer Ingelheim, Eli Lilly, Mundipharma, NovoNordisk, Novartis, Sanofi, Servier and Valbiotis. SB has received honoraria for board, conferences, clinical trial or congress from Amarin, Amryt, Amgen, Akcea, Chiesi, Novartis, Sanofi and Ultragenyx.

PEER REVIEW

The peer review history for this article is available at https://www.webofscience.com/api/gateway/wos/peer‐review/10.1111/dom.16175.

Supporting information

Supplemental Figure 1. Therapeutic GLP‐1RA molecules used in patients with FPLD during follow‐up.

Supplemental Figure 2. Diagram showing the different metabolic responses to short‐term GLP1 receptor agonist therapy in patients with familial partial lipodystrophy (FPLD).

Supplemental Table 1. Comparison of baseline data between responders and non‐responders to GLP‐1RA therapy.

Supplemental Table 2. Effect of GLP‐1RA therapy on Body Mass Index (kg/m²), HbA1c (%), and serum triglycerides (mmol/L) at baseline, short‐ and long‐term follow‐up for a subgroup of patients characterized by BMI.

Lamothe S, Belalem I, Vantyghem M‐C, et al. Safety and effectiveness in an uncontrolled setting of glucagon‐like‐peptide‐1 receptor agonists in patients with familial partial lipodystrophy: Real‐life experience from a national reference network. Diabetes Obes Metab. 2025;27(4):1815‐1825. doi: 10.1111/dom.16175

DATA AVAILABILITY STATEMENT

Data available on request from the authors.