Unraveling the natural history of lipodystrophy syndromes: insights from the prospective LD-Lync study

Abstract

Context

Lipodystrophy syndromes represent a diverse group of rare disorders characterized by deficiency and abnormal distribution of adipose tissue that may be generalized (GL) or partial (PL), leading to significant metabolic abnormalities.

Objective

To elucidate the natural history of critical comorbidities of lipodystrophy.

Methods

Ongoing, prospective, multicenter, registry study (LD Lync; NCT03087253). This analysis reports 266 patients with lipodystrophy enrolled until October 11, 2023.

Results

The median age was 39 years ([28-54], range: 6-76 years; male/female: 50/216). Hypertriglyceridemia (n = 202, 75.9%) was the most common comorbidity, followed by diabetes (n = 190, 72.0%), hepatic steatosis (n = 182, 68.4%), and hypertension (n = 130, 48.8%). Pancreatitis was detected in 21.8% of patients (n = 58), and cirrhosis in 7.1% of patients (n = 19). The median age at diagnosis was significantly earlier in patients with GL compared to those with PL for both diabetes (16 [14-32] years in GL vs 35 [25-50] years in PL) and hypertriglyceridemia (20 [15-26] years in GL vs 30 [22-49] years in PL) (P < .0001). Seven (2.6%) patients died during follow-up. Multiple medications were used to treat metabolic disease. Eighty-one (30.4%) patients received at least one dose of metreleptin during clinical studies or commercially during follow-up.

Conclusion

Our findings confirm that patients with lipodystrophy face a significant risk of comorbidities, with an earlier onset of diabetes, hypertriglyceridemia, and hepatic steatosis in those with GL compared to PL. This registry will serve as a crucial resource for gaining insights into the burden of disease in lipodystrophy.

Lipodystrophy syndromes encompass a group of rare multisystemic disorders characterized by varying degrees of deficiency and abnormal distribution of adipose tissue [1-3]. These syndromes are either inherited or acquired, with genetic mutations underlying inherited forms, while acquired forms are often associated with autoimmune or other systemic conditions [4]. Recent research estimates the prevalence of this disease between 1.3 and 4.7 cases per 1 million individuals, excluding HIV-associated cases [5]. Although traditionally classified as generalized or partial based on the degree of fat loss, these syndromes exhibit substantial heterogeneity beyond this distinction. Generalized lipodystrophy (GL) refers to the near-total absence or progressive loss of adipose tissue throughout the entire body, while partial lipodystrophy (PL) refers to regional loss of adipose tissue affecting particular body regions (eg, the extremities) depending on the specific subtype [3, 6]. Four major subgroups of lipodystrophy syndromes are thus defined: congenital generalized lipodystrophy (CGL), familial partial lipodystrophy (FPLD), acquired generalized lipodystrophy (AGL), and acquired partial lipodystrophy (APL) [7].

Due to functional adipose tissue deficiency, patients with lipodystrophy may experience severe metabolic disorders, including insulin resistance, poorly controlled diabetes, hypertriglyceridemia, ectopic steatosis, reproductive dysfunction, and reduced adipokines (eg, leptin, adiponectin) [8-10]. Patients may also experience life-threatening consequences such as recurrent attacks of pancreatitis, various cardiac complications (ischemic events, heart failure, and arrhythmias), cirrhosis, and stroke [11-13]. In addition, psychosocial comorbidities such as poor quality of life, anxiety, depression, pain, and eating disorders affect this population [14-16]. The severity of complications depends on the subtype, the degree of fat loss, and other clinical components that increase patient mortality and morbidity. However, a significant gap persists in the literature concerning the natural history of these orphan syndromes, emphasizing the need for further data collection in large numbers. This prospective, multicenter study seeks to thoroughly investigate the natural history of lipodystrophy syndromes, providing precise prevalence estimates for lipodystrophy-associated comorbidities and uncovering the clinical distinctions across subtypes. By examining the natural progression, morbidity, and mortality associated with lipodystrophy, this study aims to enhance early diagnosis, optimize management, and guide the development of targeted therapies. This article describes the fundamental hypotheses and aims of the study, which is envisioned to continue for at least another decade, and summarizes data collected from baseline visits and available follow-up.

The LD-Lync registry was established as a long-term international natural history study designed to systematically characterize the clinical spectrum, disease burden, complications, and longitudinal outcomes of lipodystrophy syndromes. Given the rarity and heterogeneity of these disorders, an important objective of the present report is to introduce this resource to the endocrinology community and provide a comprehensive baseline characterization of enrolled participants together with currently available longitudinal observations. Accordingly, this manuscript should be viewed as a foundational report from an ongoing prospective cohort rather than a definitive assessment of long-term disease trajectories, treatment effectiveness, or mechanistic pathways.

Research design and methods

Study cohort

This prospective, international cohort study was designed in collaboration with the patient foundation Lipodystrophy United and recruited subjects from several sites, including the University of Michigan, the National Institutes of Health (NIH), Turkey, and Brazil (from different centers). All participants or their parents/legally authorized representatives provided written informed consent. Each treatment center received local institutional review board approvals before data collection. A total of 266 participants were enrolled in the study at participating clinical centers concurrently from February 27, 2018, to October 11, 2023. Patients older than 3 years who met one of the 3 clinical criteria for a lipodystrophy syndrome diagnosis (excluding HIV-related and drug-induced localized lipodystrophy) and fulfilled the inclusion criteria (Table 1) were invited to participate in this study. Participants were classified into lipodystrophy subtypes using a conventional diagnostic approach that incorporated genetic, phenotypic, and clinical information. When a known pathogenic or likely pathogenic variant consistent with an established inherited lipodystrophy syndrome was present, genetic diagnosis was prioritized for subtype assignment. In patients without a confirmatory genetic diagnosis, classification was based on the pattern and distribution of adipose tissue loss, age at symptom onset, family history, associated metabolic abnormalities, and the presence of autoimmune or inflammatory features suggestive of acquired disease. Subtype assignment was reviewed by investigators with expertise in lipodystrophy at participating centers. Patients classified under the “other/unknown” subgroup had clinically evident partial or generalized loss of adipose tissue accompanied by metabolic and/or immunologic features suggestive of lipodystrophy, but did not fulfill the typical phenotypic or genetic criteria for established lipodystrophy subtypes (Table 1). Specifically, these individuals lacked pathogenic variants in known lipodystrophy-related genes, and their clinical presentation was not consistent with FPLD1 or other recognized categories. Thus, this subgroup represents patients with atypical, rare, or currently unclassified forms of lipodystrophy, rather than individuals without diagnostic validation. The research team gathered demographics, comprehensive medical history, and detailed family history from participants. Relevant laboratory and imaging results were recorded from medical records when available. Because laboratory testing was performed at local clinical laboratories rather than a central laboratory, formal cross-site assay calibration could not be performed. To minimize variability, laboratory values were harmonized by standardizing units, reviewing implausible values, and applying uniform clinical definitions across sites. Baseline laboratory distributions were additionally compared across centers as a descriptive assessment of intersite consistency. Adipokine measurements, including leptin, were obtained from available clinical records and were not performed using a centralized assay platform. Because assay methods, reference ranges, and data availability varied across sites and during historical data collection, formal subgroup analyses evaluating adipokine levels in relation to metabolic complications were not performed in the present study. Anthropometric measurements and physical examinations were performed at each visit. Participant-filled questionnaires, which were chosen upon guidance from the patient foundation Lipodystrophy United, are listed in Table 2 [17-27].

Table 1.

Refined diagnostic and eligibility criteria for lipodystrophy syndromes in the LD-Lync study

Clinical diagnosis of lipodystrophy syndromes
Presence of biallelic known disease-causing variants in the genes for autosomal recessive lipodystrophy syndromes; OR
2. Presence of a known (or de novo loss of function) disease-causing variant in the genes for autosomal dominant lipodystrophy syndromes; OR
3. Fat loss together with at least one additional metabolic/immunological criterion
Fat lossa	Midthigh skinfold thickness; or (≤10 mm in men, and ≤22 mm in women)
	Imaging modality showing symmetric, selective absence of fat from the entire body or regionally; or
	Presence of multifocal areas of fat loss and/or inflamed fat (panniculitis) at multiple areas in the body
Hypertriglyceridemia	Triglycerides ≥200 mg/dL; or
	Pancreatitis due to hypertriglyceridemia
Severe insulin resistance	Fasting insulin ≥30 mIU/L
	Insulin use of ≥2 units/kg/day
	Presence of hyperandrogenism and/or PCOS in women
Presence of diabetes mellitus (ADA criteria)
Metabolic dysfunction-associated steatotic liver disease (MASLD) (imaging or biopsy)
Presence of systemic immune dysregulation and/or hypocomplementemia by genetic and/or laboratory and/or in vitro testing

^a Fat loss is objectively measured by meeting at least one of these criteria.

Abbreviations: MASLD, metabolic dysfunction-associated steatotic liver disease; PCOS, polycystic ovary syndrome.

Table 2.

Key assessment domains captured during the LD-Lync registry study, including demographic, clinical, laboratory, genetic, and treatment-related data

Assessment category	Parameters
Demographics and clinical history	Date of birth, gender, ethnicity, detailed medical history, family history, current and prior medications
Patient-reported questionnaires	PHQ-9, GAD-7, RAND SF-36 Quality of Life Questionnaire, Disease Distress Screening Scale, International Physical Activity Questionnaires, Brief Pain Inventory (Short Form), Michigan Body Map, Modified Binge Eating Scale, Visual Analog Scale, Perceived Financial Burden of Lipodystrophy Scale, and KINDL health-related quality of life surveys (minors and parents/guardians)
Anthropometric measurements	Height, weight, BMI, skinfold measurements, hip/waist/neck circumferences
Vital signs	Blood pressure, respiratory rate, temperature, pulse
Physical examination	Fat distribution, musculature description, organomegaly, and systemic physical examination findings (eg, cardiovascular, respiratory, neurological)
Laboratory assessments	Complete blood count, comprehensive metabolic panel including blood glucose, HbA1c, lipid profile (LDL, HDL, triglycerides), liver enzymes (ALT, AST); coagulation parameters (eg, INR); and other clinically indicated laboratory tests, such as insulin, leptin, and relevant hormone levels

Abbreviations: ALT, alanine transaminase; AST, aspartate transferase; BMI, body mass index; GAD-7, Generalized Anxiety Disorder 7-item scale; HDL, high-density lipoprotein; INR, International Normalised Ratio for prothrombin time test; LDL, low-density lipoprotein; PHQ-9, Patient Health Questionnaire.

Prospective data collection

Data were prospectively collected across multiple centers using a standardized algorithm, at baseline, at annual follow-up visits through year 5 after enrollment, and at 2-year intervals thereafter. For those unable to visit the study site, contact was maintained via telephone and email. When necessary, legal caregivers or family members assisted minors or cognitively impaired adults in completing the questionnaires. Clinical evaluations and standardized parameters assessed at each visit are in Table 2 [3, 28-31]. Data were transferred to the registry DCC through a secure, 21 CFR Part 11 compliant, web-based electronic data collection system, OpenClinica [32], from initial launch until March 1, 2024, and data were transitioned to RedCap [33] Cloud starting June 3, 2024. Data were extracted using the Electronic Medical Record Search Engine, a validated clinical information retrieval system designed to search free-text medical records for research [34]. Gastroesophageal reflux was identified based on information obtained from patients and/or their families and through review of available medical records. Diabetes was defined according to the recommendations of the American Diabetes Association [35]. Dyslipidemia was defined by a total cholesterol level ≥200 mg/dL or triglycerides (TG) ≥150 mg/dL, or high-density lipoprotein (HDL) <50 mg/dL for females and <40 mg/dL for males, or low-density lipoprotein (LDL) ≥130 mg/dL [36]. Hypertension was defined according to the 2024 ESC Guidelines [37]. The present manuscript primarily reports on baseline data, complemented by available prospective follow-up outcomes; additional longitudinal follow-up data are still being collected and will be analyzed in a future dedicated manuscript from the LD-Lync registry.

Given the multicenter design of the registry, laboratory measurements were obtained through local clinical laboratories rather than a single central laboratory. To reduce intersite variability, laboratory data were harmonized during data cleaning by standardizing units, reviewing implausible values, and applying uniform clinical definitions for major metabolic outcomes and milestone events across sites.

Follow-up completion was assessed based on visit eligibility, recognizing that participants entered the registry at different calendar times. Therefore, patients were considered eligible for a given annual follow-up visit only if sufficient time had elapsed from enrollment to the data cutoff; visits that were not yet due were not classified as missing. Although the LD-Lync protocol was primarily designed for in-person assessments, the COVID-19 pandemic limited in-person follow-up at several sites. When in-person visits were not feasible, follow-up assessments were performed virtually or by telephone/email, according to local feasibility and study procedures. In addition, people were allowed to continue at any point even though they may have missed one or more annual visits so long as they continued (and continue to) to come to the respective study sites.

Statistical analysis

Statistical analysis was performed using SAS® 9.4 (Statistical Analysis System) software. Clinical and demographic data are reported as mean ± SD, median [IQR], or frequencies (%) for the overall population and subgroups. For GL vs PL comparisons, as well as for the most common genetic subgroups, the time to diagnosis of complications was analyzed using Kaplan–Meier curves and defined as the median age of occurrence (in years and interquartile range). Kaplan–Meier analyses were conducted using 2 complementary sources of time-to-event information: retrospectively ascertained age-at-onset data and prospectively observed incident events. First, for complications diagnosed before enrollment, age at onset or diagnosis was derived from available medical records and participant-reported history collected at the baseline visit, allowing reconstruction of pre-enrollment event timing. Second, for complications occurring after enrollment, incident events captured during prospective registry follow-up were added to the updated survival models. Participants without the outcome of interest were censored at the age of their most recent available follow-up (Fig. S1) [38]. This framework allowed integration of both retrospective and prospective observations and enabled characterization of the cumulative burden and timing of major complications within the registry. Follow-up completion was analyzed conditional on visit eligibility. For each annual follow-up visit, participants were considered eligible only if sufficient time had elapsed since enrollment to reasonably expect completion of that visit by the October 2023 data cutoff. Participants whose follow-up window had not yet elapsed were not classified as missing. This approach allowed separation of true loss to follow-up from visits that were not yet due at the time of analysis.

Missing data were handled using a complete-case approach for each variable-specific analysis. Denominators therefore varied across outcomes depending on the availability of clinical, laboratory, medication, or follow-up data. For longitudinal analyses, participants were considered eligible for a given annual visit only when sufficient time had elapsed since enrollment for that visit to be due by the data cutoff. Participants whose follow-up window had not yet elapsed were not classified as missing. This approach allowed incomplete follow-up to be distinguished from visits that were not yet expected by the time of analysis.

Depending on the data distribution, Student's t-test or Mann–Whitney U test was used to compare variables between 2 groups at baseline. Multiple group comparisons were done using ANOVA. Baseline categorical variables were analyzed by χ² or Fisher’s exact test. When comparisons were undertaken between multiple groups, appropriate models were constructed based on the study biostatistician’s guidance. Nominal P-values are reported. For subgroup comparisons involving multiple baseline variables, Bonferroni correction was applied to reduce the risk of type I error; the corrected significance threshold was calculated by dividing 0.05 by the number of comparisons within the relevant table.

Modified lipodystrophy severity score analysis

A modified Lipodystrophy Severity Score (LDS) approach was applied to participants with FPLD2 to descriptively assess longitudinal multisystem disease burden. Of the 8 domains in the previously published LDS framework, 6 domains could be calculated from available LD-Lync data: diabetes/insulin resistance, microvascular complications of diabetes, lipids, atherosclerotic cardiovascular disease, liver disease, and renal function [39]. Reproductive and other-condition domains were not included because the required variables were not uniformly available across study visits. Domain scores were calculated at baseline and available follow-up visits and summarized as median [IQR]. Because the LD-Lync study forms did not include a single dedicated item for autonomic neuropathy, this component was derived using available proxy variables, including reported constipation or diarrhea, tachycardia, ataxia, difficulty swallowing, and use of medications for neuropathy. Because heart failure severity was not systematically captured as part of the study design, participants with documented heart failure were assigned a score of 4, corresponding to moderate severity, for this component.

Results

Study cohort and demographics

Between 2018 and the end of 2023, 266 patients were enrolled in the LD-Lync study conducted at 3 main clinical centers: the University of Michigan (146 patients), various centers in Turkey (72 patients), and the NIH (48 patients). The cohort was comprised of 226 proband cases and 40 of their relatives (Tables S1A and S1B [38]). Patients from Brazil were added after the data cut for this manuscript and thus were not included. Detailed demographic characteristics are presented by lipodystrophy type in Table 3. The median age was 39 [28-54] years (range, 6-76 years); females represented 81.2% (n = 216), and most patients were Caucasian (n = 234), followed by African American (n = 11) and Asian (n = 5). The distribution of lipodystrophy subtypes by sex is summarized in Table S2 [38]. The GL group comprised 53 patients, while the PL group included 198 patients. Most individuals in the GL group had CGL (n = 37; 69.8%), whereas most patients in the PL group had FPLD (n = 173; 87.4%). Males constituted 18.8% (n = 50) of the total cohort, including 26.4% (n = 14) of the GL group, 15.6% (n = 31) of the PL group, 33.3% (n = 5) of the other/unknown subgroup. In the entire cohort, median body mass index (BMI) was 25.7 [22.0-29.6] kg/m². Median BMI was significantly lower in CGL (19.9 [17.5-24.2] kg/m²) than in other subgroups (P < .001). FPLD was the most prevalent subtype across all participating centers, although the distribution of lipodystrophy subtypes varied among centers. Detailed phenotype/subtype distribution by center is in Table 4.

Table 3.

Demographic and clinical characteristics by lipodystrophy type

	Generalized lipodystrophy (GL) (n = 53)		Partial lipodystrophy (PL) (n = 198)
	Congenital generalized lipodystrophy (CGL) (n = 37)	Acquired generalized lipodystrophy (AGL) (n = 16)	Familial partial lipodystrophy (FPLD) (n = 173)	Acquired partial lipodystrophy (APL) (n = 25)	Other/unknown subgroup (n = 15)	Total (n = 266)	P-valuea
Age at baseline (years)	27 [20-33]	28 [23-40]	42 [33-56]	47 [35-56]	44 [19-53]	39 [28-54]	<.001b
Age at the first symptom of lipodystrophy (years)	9 [2-14]	14 [6-30]	16 [12-29]	24 [8-45]	13 [11-25]	15 [10-29]	<.001 b
Age at diagnosis of lipodystrophy (years)	13 [5-21]	20 [6-41]	38 [25-50]	37 [30-47]	33 [13-45]	33 [18-47]	<.001 b
Sex (female %)	29 (78.4%)	10 (62.5%)	147 (85%)	20 (80.0%)	10 (66.7%)	216 (81.2%)	.109
BMI (kg/m2)	19.9 [17.5-24.2]	22.0 [19.8-24.9]	26.7 [23.4-31.3]	24.7 [20.9-28.5]	26.1 [23.1-27.4]	25.7 [22.0-29.6]	<.001 b
Hospitalization	15 (40.5%)	11 (68.8%)	117 (67.6%)	15 (60.0%)	8 (53.3%)	165 (62.0%)	.067
Age at first hospitalization	22 [12-29]	26 [13-35]	30 [20-38]	26 [22-37]	25 [12-39]	28 [19-37]	.138
Hypertension	11 (29.7%)	5 (31.3%)	98 (56.6%)	14 (56.0%)	2 (15.4%)	130 (49.2%)	.005
Age of onset	16 [12-33]	34 [7-45]	30 [20-40]	33.50 [26-44]	14 [14-14]	30 [20-39]	.110
Myocardial infarction	0 (0.0%)	0 (0.0%)	9 (5.2%)	1 (4.0%)	0 (0.0%)	10 (3.8%)	.799
Age of onset			38 [35-51]	42 [42-42]		40 [35-51]	.861
Heart failure	1 (2.7%)	2 (12.5%)	4 (2.3%)	2 (8.0%)	0 (0.0%)	9 (3.4%)	.279
Age of onset	42 [42-42]	29 [24-35]	59 [27-71]	27 [12-42]		35 [24-59]	.368
Stroke	0 (0.0%)	0 (0.0%)	10 (5.8%)	2 (8.0%)	1 (7.7%)	13 (5.0%)	.372
Age of onset			49 [35-60]	37 [32-42]	43 [43-43]	43 [35-53]	.548
Asthma/emphysema	3 (8.1%)	2 (12.5%)	39 (22.5%)	0 (0.0%)	2 (13.4%)	46 (17.2%)	.010
Age of onset	11 [6-16]	8 [6-10]	15 [4-28]		3 [1-5]	12 [4-25]	.447
Myopathy	5 (13.5%)	4 (25.0%)	34 (19.6%)	0 (0.0%)	3 (20.0%)	46 (17.2%)	.145
Age of onset	18 [18-24]	22 [2-42]	33 [21-40]		32 [30-44]	30 [18-40]	.397
Gastroesophageal reflux disease	7 (18.9%)	2 (12.5%)	65 (37.6%)	7 (28.0%)	3 (20.0%)	84 (31.5%)	.138
Age of onset	15 [3-21]	33 [9-58]	27 [20-35]	43 [22-52]	41 [39-50]	27 [20-40]	.062
Depression	11 (29.7%)	4 (25.0%)	78 (45.1%)	9 (36.0%)	4 (26.6%)	106 (39.8%)	.002
Age of onset	20 [13-25]	38 [26-47]	23 [16-34]	29 [16-46]	15 [13-40]	22 [16-35]	.265
Anxiety	11 (29.7%)	3 (18.8%)	84 (48.6%)	6 (24.0%)	5 (33.3%)	109 (40.9%)	.015
Age of onset	25 [14-32]	35 [17-54]	25 [15-35]	33 [15-48]	16 [15-22]	25 [15-35]	.675
Cancer	1 (2.7%)	4 (25.0%)	13 (7.5%)	5 (20.0%)	1 (6.6%)	24 (9.0%)	.028
Age of onset	40 [40-40]	14 [7-23]	48 [43-52]	32 [20-46]	29 [29-29]	43 [29-49]	.028

Data are presented median [IQR]. Abbreviation: BMI, body mass index.

^a The P-value represents the analysis of the comparison across 5 subgroups, including the “other/unknown” category.

^b P-values remained statistically significant after Bonferroni correction for 27 comparisons, using a corrected significance threshold of P < .00185.

Table 4.

Demographics and clinical characteristics by center

Characteristic	University of Michigan (n = 146)	Turkey (n = 72)	NIH (n = 48)	Total (n = 266)	P-value
Age at baseline (years)	43 ± 15	37 ± 15	36 ± 17	40 ± 16	.002
Female (n, %)	117 (81.8%)	55 (77.5%)	40 (83.3%)	212 (80.9%)	.669
Baseline BMI (kg/m2)	28.26 ± 6.49	22.95 ± 3.84	24.09 ± 4.86	26.18 ± 6.13	<.001
Race, White (n, %)	131 (91.0%)	72 (100%)	30 (75.0%)	233 (91.0%)	<.001
Lipodystrophy types/subtypes					<.001
Acquired generalized lipodystrophy (AGL)	13 (8.9%)	0 (0.0%)	3 (6.5%)	16 (6.1%)
Acquired partial lipodystrophy (APL)	13 (8.9%)	10 (13.9%)	2 (4.3%)	25 (9.5%)
Congenital generalized lipodystrophy (CGL)	5 (3.4%)	22 (30.6%)	10 (21.7%)	37 (14.0%)
CGL1, AGPAT2 mutations	1 (0.7%)	12 (17.1%)	6 (13.3%)	19 (7.5%)
CGL2, BSCL2 mutations	2 (1.5%)	4 (5.7%)	1 (2.2%)	7 (2.8%)
CGL3, CAV-1 mutations	0 (0.0%)	1 (1.4%)	1 (2.2%)	2 (0.8%)
CGL4, CAVIN-1 mutations	0 (0.0%)	1 (1.4%)	0 (0.0%)	1 (0.4%)
Other/unknown	2 (1.5%)	4 (5.7%)	2 (4.4%)	8 (3.2%)
Familial partial lipodystrophy (FPLD)	105 (71.9%)	38 (52.8%)	30 (65.2%)	173 (65.5%)
FPLD1, Kobberling variety	45 (32.8%)	1 (1.4%)	0 (0.0%)	46 (18.3%)
FPLD2, Dunnigan variety, LMNA mutations	36 (26.3%)	26 (37.1%)	10 (22.2%)	72 (28.6%)
FPLD3, PPARG mutations	3 (2.2%)	4 (5.7%)	4 (8.9%)	11 (4.4%)
FPLD4, PLIN1 mutations	0 (0.0%)	0 (0.0%)	1 (2.2%)	1 (0.4%)
Other/unknown	21 (15.3%)	7 (10%)	15 (33.3%)	43 (17.1%)
Progeroid lipodystrophy	2 (1.4%)	0 (0.0%)	0 (0.0%)	2 (0.8%)
Other/unknown	8 (4.1%)	2 (2.8%)	1 (2.2%)	13 (4.2%)
Lipodystrophy diagnosis age	37 ± 17	29 ± 16	25 ± 19	33 ± 18	<.001
Taken leptin treatment	36 (24.7%)	17 (23.6%)	19 (48.7%)	72 (28.0%)	.007
Individuals with a family history of lipodystrophy	63 (43.2%)	45 (62.5%)	19 (41.3%)	127 (48.1%)	.001

Among patients with CGL (n = 37), CGL1 (AGPAT2) was the most common subtype (n = 19), followed by CGL2 (BSCL2) (n = 7), while CGL3 (CAV1) and CGL4 (CAVIN-1) were identified in 2 and 1 patient, respectively. Among patients with FPLD (n = 173), FPLD2 (LMNA) was the most common subtype in 72 patients, followed by FPLD1 in 46 patients. Eleven patients had FPLD3 (PPARG), and only one had FPLD4 (PLIN1). There were 25 patients with APL and 16 with AGL. Table S3 [38] summarizes the potential underlying causes and coexisting conditions observed in patients with APL and AGL. Rarer forms included 5 patients with MFN-2 pathogenic variants causing multiple symmetric lipomatosis, and 2 with progeroid lipodystrophy due to POLD1 variants. Another 15 patients (classified as the other/unknown subgroup) could not be subclassified due to atypical phenotypic features and/or unidentified genetic mutations; however, all exhibited partial loss of subcutaneous fat, consistent with a clinical suspicion of PL. The available phenotypic and genetic characteristics of all patients classified as “other or unknown” are summarized in Table S4 [38]. This subgroup included 3 patients with POLD1-associated progeroid lipodystrophy and 4 patients with multiple symmetric lipomatosis/Madelung disease associated with pathogenic MFN2 variants. Additional cases included 1 patient with atypical PL associated with an EBF2 variant, 1 with atypical PL associated with a WNT10A variant, and 1 with progressive multifocal lipoatrophy with atypical distribution and AXIN2/SREBF2 variants under mechanistic investigation. One participant had atypical lipoatrophy, negative genetic lipodystrophy screening, and a history notable for autoimmune disease and suspected immune dysregulation. In the remaining participants, genetic testing was negative, not performed, or nondiagnostic despite clinical features suggestive of atypical PL.

The first symptoms of lipodystrophy were detected earliest in patients with CGL, at a median age of 9 [2-14] years, whereas the median age at diagnosis in this group was 13 [5-21] years. In contrast, the first symptoms of lipodystrophy were recognized at the latest in APL, with a median age of 24 [8-45] years. FPLD had the highest median age at diagnosis, at 38 [25-50] years.

Metabolic complications and organ manifestations

Comorbidities and other complications by study site are shown in Table S5 [38]. To evaluate intersite consistency in laboratory measurements, baseline laboratory parameters were compared across participating centers using standardized units and are summarized in Table S6 [38].

Diabetes and insulin resistance

Diabetes was reported in 190 (72.0%) of 266 individuals, with a mean age of onset of 29 ± 14 years (range: 3-62 years) (Fig. 1A). Diabetes developed in 40 (75.4%) patients with GL and 145 (73.2%) with PL (Fig. 1A). The prevalence of diabetes was highest in CGL (n = 29 of 37; 78.4%), followed by FPLD (n = 130 of 173; 75.1%), AGL (n = 11 of 16; 68.8%), and APL (n = 15 of 25; 60.0%) (P = .17) (Fig. 1B). CGL had the earliest median age at diabetes diagnosis at 16 [13-18] years, followed by AGL at 17 [10-32] years, APL at 29 [22-49] years, and FPLD at 31 [23-43] years (P < .001). The entire cohort had a median HbA1c value of 6.7% [5.6-8.4], with no significant difference among subgroups (P = .365). Median leptin levels were lower in GL (0.60 [0.60-9.40] ng/mL) compared to PL (9.30 [5.50-20.00] ng/mL) (P = .001). Acanthosis nigricans was observed in 45.1% of patients (n = 119), most of whom belonged to FPLD (n = 82) (Fig. 1A and 1B). Regarding microvascular complications of diabetes, peripheral neuropathy was reported in 35.4% (n = 93, median age 35 years, range 30-46 years) and diabetic retinopathy in 10.3% (n = 27, median age 31 years, range 27-46 years). Neuropathy and retinopathy occurred earliest in CGL, with median ages of 26 [17-30] and 27 [27-27] years, respectively. Retinopathy was noted latest in FPLD, with a median age of 40 [31-54] years (P = .03).

Figure 1.

Frequency of clinical characteristics and median age at diagnosis for the study cohort. (A) Frequency of clinical characteristics and median age at diagnosis for the overall cohort and subgroups with generalized lipodystrophy (GL) and partial lipodystrophy (PL). Prevalence is represented by the length of the bars, with data segmented by cohort. Median ages at diagnosis are expressed next to the bars. There is no corresponding age variable for Acanthosis. *Significant differences in proportion between GL and PL. ^†Significant differences in median age between GL and PL. (B) Frequency of clinical characteristics in subgroups with congenital generalized lipodystrophy (CGL), acquired generalized lipodystrophy (AGL), familial partial lipodystrophy (FPLD), and acquired partial lipodystrophy (APL). Prevalence is represented by the length of the bars. Median ages at diagnosis are expressed next to the bars. There is no corresponding age variable for Acanthosis. *Significant differences in proportion between lipodystrophy subtypes. ^†Significant differences in median age between lipodystrophy subtypes.

Dyslipidemia and pancreatitis

Hypertriglyceridemia was the most prevalent complication affecting 202 (75.9%) individuals, with significantly higher prevalence in CGL (31 patients, 83.8%) and AGL (13 patients, 81.3%) compared to other subgroups (P = .013 for all comparisons). The median age at onset of hypertriglyceridemia (24 [17-33] years, n = 202) was earlier than that of diabetes (27 [18-40] years, n = 190). CGL had the earliest median age at diagnosis of hypertriglyceridemia (16 [12-21] years), followed by AGL (21 [13-31] years), APL (24 [20-32] years), and FPLD (27 [19-36] years) subgroups (P < .001). At least one episode of pancreatitis was reported in 58 patients (21.8%), with the highest prevalence observed in FPLD (42 cases, 24.6%) and CGL (8 cases, 21.6%). TG and total cholesterol levels were comparable between the GL group (median TG: 225 [146-742] mg/dL; median total cholesterol: 164 [130-231] mg/dL) and the PL group (median TG: 238 [149-564] mg/dL; median total cholesterol: 189 [157-221] mg/dL). However, LDL levels were higher in the PL group compared to the GL group (median LDL levels: 99 [71-121] mg/dL vs 73 [59-115] mg/dL, respectively) (P = .04). In addition, the median HDL levels were significantly lower in the GL group (31 [25-38] mg/dL) compared to the PL group (37 [29-44] mg/dL) (P = .003). BMI did not predict triglyceride levels in the entire cohort or in the PL group (Fig. 2). The median (IQR) TG/HDL-C ratios were 6.73 (4.98-30.08) for AGL (n = 15), 3.81 (2.89-4.98) for APL (n = 13), 8.50 (3.54-32.50) for CGL (n = 27), 7.09 (3.85-18.48) for FPL (n = 146), and 7.02 (3.91-45.91) for the other/unknown group (n = 12), with an overall median of 6.80 (3.79-20.00) for the entire cohort (n = 213). TG/HDL-C ratios across lipodystrophy types demonstrated a statistically significant difference (χ² = 9.75, df = 4, P = .0449).

Figure 2.

Regression analysis evaluating the relationship between BMI and log-transformed triglyceride (logTG) levels across different patient cohorts. (A) Entire cohort (n = 209). (B) Familial partial lipodystrophy (FPLD) subgroup (n = 143). (C) LMNA mutation subtype subgroup (n = 52). The solid line represents the fitted regression line. The shaded area indicates the 95% confidence intervals, while the dashed lines indicate the 95% prediction intervals.

Liver complications

Despite treatment with standard therapies, most patients developed end-organ abnormalities (Fig. 1). Metabolic dysfunction-associated steatotic liver disease (MASLD) was reported in 182 (68.4%) patients, with a median age of 29 [19-42] years. Twenty patients (11%) within this group were reported to have undergone liver biopsy. Liver disease progressed to cirrhosis in 19 patients (7.1%), at a median age of 42 [32-53] years. The prevalence of cirrhosis was 12.5% in AGL, 8.0% in APL, 2.7% in CGL, and 7.5% in FPLD. At the baseline visit, the GL group had median AST and ALT levels of 24.5 [18-42] U/L and 29 [20-43] U/L, respectively, while the PL group had median AST and ALT levels of 25.5 [19-37] U/L and 30 [21-48] U/L, respectively (P > .05 for both).

Cardiovascular disease

Hypertension was present in 130 patients (49.2%), with a mean age of onset of 30 ± 13 years. Irregular heartbeat was the most common cardiac symptom reported in 46 cases (17.2%), and a cardiac pacemaker was used due to arrhythmia in 5 patients. History of heart murmur was reported in 34 cases (12.9%). Twenty-six patients reported experiencing angina at least once in their lives; myocardial infarction was reported in 10 (3.8%), and heart failure in 9 (3.4%). Thirteen patients (5.0%) had a stroke. Claudication was observed in 80 patients (30.4%), with the highest prevalence (n = 59) reported in FPLD. Detailed subgroup analyses are provided in Table 3.

Kidney complications

The kidney was another commonly affected organ, with involvement reported in 85 patients (32.2%), occurring most frequently in CGL (n = 16/37, 43.2%) and FPLD (n = 56/173, 32.3%). Proteinuria, one of the first manifestations of kidney disease, was present in 72 patients (27.0%), and hematuria in 24 (9.0%), with mean onset ages of 27 ± 15 and 27 ± 18 years, respectively. Seven patients had glomerulonephritis (mean age of 31 ± 9 years), occurring in 1 with progeroid lipodystrophy, 2 with CGL, and 4 with FPLD (Table S7 [38]). Four patients were undergoing hemodialysis due to end-stage renal disease (mean age of onset of 31 ± 8 years).

Additional complications and clinical conditions

Beyond the typical metabolic complications such as diabetes, pancreatitis, and hyperlipidemia, patients also experienced a variety of other comorbidities and end-organ dysfunctions. Gastroesophageal reflux disease (GERD) was reported in 84 patients (31.7%), with a median age of diagnosis of 27 [20-40] years. GERD was most frequently reported in FPLD (65 of 173 patients, 37.6%). Respiratory comorbidities were also present, including sleep apnea in 77 individuals (29.2%) (CGL: n = 5; AGL: n = 0; FPLD: n = 61; APL: n = 6, other/unknown subgroup: n = 5, P = .024), and asthma and/or emphysema in 46 (17.4%). Based on baseline visit data, 72 of the 216 female patients (35.1%) reported being postmenopausal. Additionally, 28.6% (n = 59) of all females reported experiencing irregular menstrual cycles at some point in their lives. Only 32.2% (n = 66) of all female patients reported having regular menstrual cycles in their most recent periods. A total of 112 patients reported at least one episode of menstrual absence lasting ≥60 days. PCOS was diagnosed in 77 patients, representing 37.2% of the female population, with a mean age at onset of 21 ± 7 years. Seven patients reported being unable to conceive a child. Psychiatric history revealed depression in 40.2% (n = 106) and anxiety in 41.3% (n = 109) of patients, with median ages of onset at 22 [16-35] years for depression and 25 [15-35] years for anxiety. Depression and anxiety were more prevalent in FPLD compared to those with CGL, AGL, and APL (P = .002 for depression and P = .015 for anxiety). A total of 24 patients had been previously diagnosed with malignancy (Table S8 [38]). Additionally, according to the Michigan Body Map assessment, 77.0% of participants (n = 104) reported experiencing chronic pain, with no statistically significant differences observed among subgroups (CGL: 14/19, 73.6%; AGL: 5/8, 62.5%; FPLD: 68/87, 78.1%; APL: 9/12, 75.0%; other/unknown subgroup: 8/9, 88.9%; P = .447 for all comparisons). Among participants reporting chronic pain, 12.6% (n = 17) indicated pain in 1 or 2 anatomical regions, whereas 64.4% (n = 87) reported pain in 3 or more regions. Chronic pain involving 3 or more anatomical regions constituted a substantial clinical burden, affecting over 50% of individuals across all subtypes, with the highest prevalence in the FPLD subgroup (67.8%). We will report on patient-reported questionnaires in a future dedicated manuscript separately.

Overall comparison of patients with acquired and congenital or familial cases showed that they share broadly similar comorbidity patterns, although differences were observed for BMI, the presence of high TG, xanthomas, claudication, sleep apnea, anxiety, and cancer (Table S9 [38]).

Concomitant medications and metreleptin treatment

At the baseline visit, 81 of 266 patients had used at least one dose of metreleptin in their lifetime, either during a clinical trial or commercially, including 26 (70.2%) in CGL, 10 (62.5%) in AGL, 37 (21.3%) in FPLD, 2 (8.0%) in APL, and 6 patients in other/unknown subgroup. At the time of the baseline visit, 28 patients with GL (CGL and AGL) and 16 patients with FPLD were taking metreleptin. Among participants with available metreleptin treatment dates, both timing of first exposure and duration of treatment were summarized by treatment status. Current users had a mean interval of 5.8 ± 4.0 years from metreleptin initiation to the corresponding study visit date, whereas participants who discontinued treatment had a mean interval of 6.2 ± 3.7 years from metreleptin initiation to the study visit date and a mean documented treatment duration of 2.0 ± 3.2 years. Table 5 presents the metabolic parameters of patients with GL and PL based on their metreleptin therapy status at the time of the baseline visit, with corresponding female and male comparisons shown in Tables S10 [38] and S11 [38]. As expected, patients with GL who are current users of metreleptin demonstrate better metabolic control compared to those who are not current users. Among the FPLD cohort, data from the study baseline demonstrate that metreleptin users overall have more metabolic disease burden and, thus, worse control.

Table 5.

Comparison of laboratory parameters based on metreleptin therapy status

Laboratory results	Generalized lipodystrophy (AGL + CGL)				Familial partial lipodystrophy (FPLD)
	Currently on metreleptin (n = 28)	Previously on metreleptin (n = 5)	Never on metreleptin (n = 17)	P-value	Currently on metreleptin (n = 16)	Previously on metreleptin (n = 17)	Never on metreleptin (n = 136)	P-value
	Median	Median	Median		Median	Median	Median
Hemoglobin A1c level (%)	6.4 [5.5-7.6]	6.8 [6.3-8.2]	6.0 [5.1-6.9]	.55	7.7 [7.0-10.5]	8.0 [6.7-8.9]	6.6 [5.7-8.3]	.007
Triglycerides level (mg/dL)	204 [162-653]	420 [130-1992]	254 [107-1038]	.99	290 [157-625]	256 [198-416]	253 [151-652]	.81
Cholesterol level (mg/dL)	160 [130-197]	131 [105-362]	207 [129-262]	.57	184 [155-266]	206 [157-223]	189 [158-220]	.91
LDL level (mg/dL)	70 [60-112]	52 [44-88]	88 [59-133]	.38	96 [68-107]	116 [72-129]	100 [69-120]	.51
HDL level (mg/dL)	30 [26-36]	27 [21-40]	30 [24-47]	.94	36 [11-37]	37 [29-40]	36 [28-44]	.985

Data are presented as median [IQR].

Insulin use was documented in 11 patients (44.0%) within the CGL subgroup and in 7 patients (63.6%) within the AGL subgroup. In the PL group, insulin usage was more prevalent in FPLD (79 of 173, 59.4%) compared to APL (5 of 25, 35.7%). Metformin was used by 142 patients (53.4%), including 101 with FPLD, 21 with CGL, 5 with AGL, and 10 with APL. Forty-seven patients received sodium-glucose transport protein 2 inhibitors, 16 were on dipeptidyl peptidase 4 therapy, 39 were on glucagon-like peptide-1 therapy, and 48 were on thiazolidinediones. Regarding dyslipidemia management, fenofibrate was administered to 79 patients, while 17 were treated with gemfibrozil. Statins were the most frequently prescribed lipid-lowering agents, with 100 patients receiving treatment, primarily atorvastatin (n = 57) and rosuvastatin (n = 25). Omega-3 supplementation was reported in 53 patients, whereas niacin and cholestyramine were used by 2 and 1 patient, respectively. Four patients used proprotein convertase subtilisin/kexin type 9 inhibitors, including evolocumab (n = 3) and alirocumab (n = 1).

Time to comorbidities and mortality

The Kaplan–Meier estimates of the median age of diagnosis for important comorbidities comparing PL vs GL and the most common genetic subgroups are presented in Fig. 3. The median age at diagnosis of diabetes was 16 [14-32] years in GL and 35 [25-50] years in PL (P < .0001) (Fig. 3A), while the median age of hypertriglyceridemia was 20 [15-26] years in GL and 30 [22-49] years in PL (P < .0001) (Fig. 3C). In CGL compared to FPLD, the median age at diagnosis was lower for diabetes (16 [14-24] vs 35 [25-46] years), hypertriglyceridemia (17 [14-23] vs 30 [20-42] years) and MASLD (22 [16-40] vs 40 [26-57] years) (P < .0001 for all comparisons) (Fig. 3B, 3D, and 3H, respectively). The median age of the first cardiac condition was similar between CGL and FPLD: 31 [21-36] and 37 [24-53] years, respectively (Fig. 3I and 3J). More time and events are needed to obtain accurate estimates for pancreatitis and kidney disease. Seven patients died during the follow-up period (range: 36-78 years) (Table 6).

Figure 3.

Kaplan–Meier estimates of the median age at diagnosis for each comorbidity comparing lipodystrophy subgroups. Panels show diabetes-free (A, B), hypertriglyceridemia (HighTG)-free (C, D), pancreatitis-free (E, F), metabolic dysfunction-associated steatotic liver disease (MASLD)-free (G, H), heart disease-free (I, J), and baseline (BL) hospitalization-free (K, L) survival probabilities over time (years). Survival probability over time (y-axis), disease-free survival time, in years (x-axis). Abbreviations: GL, generalized lipodystrophy; PL, partial lipodystrophy; CGL, congenital generalized lipodystrophy; FPLD, familial partial lipodystrophy; MASLD, metabolic dysfunction-associated steatotic liver disease; HighTG, hypertriglyceridemia.

Table 6.

Clinical characteristics and reported causes of death in patients with lipodystrophy

Age (years)	Sex	Lipodystrophy type	Comorbid conditions	Causes of death
36	F	FPLD1	Diabetes mellitus	Multiorgan failure following acute exacerbation of chronic pancreatitis
			Hypertriglyceridemia
			PCOS
			MASLD
			History of pulmonary embolus
51	F	FPLD1	Diabetes mellitus	No available cause of deatha
			Hypertriglyceridemia
			Arrhythmia
			Cirrhosis
			MASLD
			Melanoma
			Pancreatitis
			Proteinuria
64	F	FPLD1	Coronary artery disease	Aspiration pneumonia, coronary artery disease
			Diabetes mellitus
			Hyperlipidemia
37	F	FPLD2	Diabetes mellitus	Stage-4 endometrial cancer
			Hypertension
			Hypertriglyceridemia
			PCOS
			Hirsutism
			Endometrial cancer
74	M	FPLD2	Coronary artery disease	Myocardial infarction
			Paroxysmal atrial fibrillation
			Chronic congestive heart failure
			Diabetes mellitus
78	M	FPLD2	Coronary artery disease	Decompensated cirrhosis
			Hypertension
			Hemochromatosis
			Psoriasis
			MGUS
42	F	AGL	Diabetes mellitus	Acute respiratory failure due to gastrointestinal bleeding, hemophagocytic lymphohistiocytosis, bronchiectasis, endocarditis
			Hypertriglyceridemia
			Hypothyroidism
			Granulomatous Lung Disease
			CVID
			Hepatosplenomegaly

Abbreviations: AGL, acquired generalized lipodystrophy; CVID, common variable immunodeficiency; F, female; FPLD1, familial partial lipodystrophy type-1; FPLD2, familial partial lipodystrophy type-2; M, male; MASLD, metabolic dysfunction-associated fatty liver disease; MGUS, monoclonal gammopathy of undetermined significance; PCOS, polycystic ovary syndrome.

^a Cause of death information was not available for this patient due to the absence of accessible medical records.

Prospective follow-up outcomes

Prospective follow-up completion

Of the participants eligible for the first annual follow-up visit, 111 of 242 (45.9%) completed the assessment. At year 2 follow-up, 55 of 175 eligible participants (31.4%) completed the visit. At year 3, 51 of 109 eligible participants (46.8%) completed follow-up, and at year 4, 37 of 88 eligible participants (42.0%) completed the visit. Most longitudinal analyses presented below are derived from the FPLD subgroup, which had the greatest availability of follow-up data. During prospective follow-up, incident clinical outcomes were documented across metabolic, hepatic, cardiovascular, renal, and malignancy-related domains.

Incident clinical outcomes during prospective follow-up

During follow-up, incident diabetes was diagnosed in 6 participants overall. Diabetes-related microvascular complications were also documented during the first 3 years of prospective follow-up, including 6 new cases of peripheral neuropathy and 3 new cases of diabetic retinopathy. Hypertriglyceridemia and fatty liver disease each developed in 9 participants, with all incident cases identified within the first 3 years. Three new cases of cirrhosis were recorded during this period, and pancreatitis occurred in 9 additional participants overall. Cardiovascular morbidity accumulated over time, with 3 new cases of heart failure, 2 myocardial infarctions, and 3 strokes documented within the first 2 years. Eight new diagnoses of hypertension were observed. Notably, all myocardial infarctions (n = 2) and strokes (n = 3) during the prospective follow-up period occurred in participants with FPLD. Five participants underwent coronary angiography, and 3 required coronary artery bypass surgery during follow-up, all of whom belonged to the FPLD subgroup. Five additional cases of kidney disease were detected within 4 years of follow-up. Four incident cancer cases were identified during the first 2 years, 2 of which occurred in participants with FPLD. Acanthosis nigricans was newly documented in 124 participants, and 3 new cases of GERD were recorded. Newly diagnosed psychiatric comorbidities included depression in 8 participants and anxiety in 6 participants. During follow-up, 15 participants required hospitalization due to acute pancreatitis, 12 of whom belonged to the FPLD subgroup. Cardiovascular morbidity during prospective follow-up was further reflected by hospitalizations and coronary interventions. Four participants were hospitalized for angina pectoris within the first 2 years, and 2 participants were hospitalized for myocardial infarction within the first 4 years. Coronary angiography was performed in 5 participants, all from the PL group. Three new cases of coronary artery bypass grafting and 3 new cases of coronary angioplasty with stent placement were also documented during follow-up. Metabolic complications and temporal characteristics among patients with cardiovascular events are summarized in Table S12 [38]. Among patients with available age-at-onset data, cardiovascular events occurred before the formal diagnosis of LD in approximately half of those with angina (11/20, 55.0%), myocardial infarction (6/11, 54.5%), stroke (8/14, 57.1%), and coronary angiography (9/17, 52.9%). Diabetes, hypertriglyceridemia, and hypertension frequently preceded major cardiovascular events and coronary interventions. Specifically, diabetes and hypertriglyceridemia preceded most myocardial infarctions (7/8, 87.5% for both), heart failure events (6/7, 85.7% for both), and coronary artery bypass grafting (CABG) procedures (5/6, 83.3% for both). Among patients who underwent percutaneous coronary intervention (PCI)/stent placement with available temporal data, diabetes was documented before the intervention in 2 of 2 patients, hypertriglyceridemia in 3 of 3 patients, and hypertension in 2 of 2 patients.

By the October 2023 data cutoff, 84 female participants were classified as postmenopausal, compared with 72 at baseline. Clinical characteristics and comorbidity burden by menopausal status through the observation period are summarized in Table S13 [38]; no comparison remained statistically significant after Bonferroni correction.

The Kaplan–Meier analyses were prepared to incorporate both retrospectively reported age-of-onset data and prospectively observed incident events accrued after enrollment. The resulting shift in time-to-event estimates after inclusion of prospective follow-up data is demonstrated in Fig. S1 [38].

FPLD-specific longitudinal outcomes

Among participants with FPLD, 4 incident cases of diabetes and hypertriglyceridemia were identified within the first year. Pancreatitis continued to emerge prospectively, with 4 new cases documented at the first follow-up visit and an additional 2 cases recorded at later visits. Five new diagnoses of hypertension were observed. Of the 3 incident cirrhosis cases, 2 occurred in participants with FPLD. Longitudinal detailed laboratory data are summarized in Table S14 [38] and Fig. S2 [38]. Longitudinal models evaluating changes in key metabolic parameters over time from baseline are summarized in Table S15 [38]. Overall, key metabolic parameters remained relatively stable during available follow-up, with no significant longitudinal changes observed for blood glucose, TG, HDL, HbA1c, total cholesterol, or LDL.

In the FPLD subgroup, poor glycemic control (HbA1c ≥ 8%) affected 32% of participants at baseline and 42% at the first follow-up visit, followed by a reduction to 33%, 39%, and 21% of participants at visits 2 to 4, respectively. Severe hypertriglyceridemia (>500 mg/dL) demonstrated a similar pattern, present in 28% of participants at baseline and 25% at visit 1, with subsequent decreases to 19%, 20%, and 5% of participants at later visits.

The LD-Lync modified LDS was applied in the FPLD2 subgroup, with domain scores summarized across baseline and available annual follow-up visits in Table S16 [38]. At baseline, the highest median scores were observed for liver disease (11 [5-15]), diabetes/insulin resistance (10 [4-13]), atherosclerotic cardiovascular disease (ASCVD) (9 [3-19]), and lipids (8 [5.5-12]). During follow-up, cardiometabolic and liver-related domains remained prominent contributors to disease burden. ASCVD scores were higher at later annual visits, reaching 21 [7-21] at year 3 and 21 [15-23] at year 4, while microvascular complication scores increased from 0 [0-5] at baseline to 7 [5-13] at years 3 and 4.

Discussion

A major strength of the present study is the establishment of a large international prospective natural history study dedicated to the systematic study of lipodystrophy syndromes. The primary purpose of this report is to characterize the enrolled population, describe the spectrum of clinical manifestations and complications, and provide initial longitudinal observations from this ongoing natural history effort. As enrollment continues and follow-up duration increases, the LD-Lync registry will provide opportunities for more detailed analyses of disease progression, genotype-phenotype relationships, treatment outcomes, patient-reported outcomes, and long-term survival. The current report, therefore, represents an important foundation upon which future longitudinal investigations can be built.

This report contains one of the richest datasets collected prospectively on the natural history of lipodystrophy syndromes. LD-Lync study is quite unique in that participants are contacted and, when possible, examined annually for the first 5 years and thereafter every 2 years. Data collection tools were designed in consultation with the patient community. Our data describe the distribution and development trajectory for diverse clinical and metabolic features under direct observation. The documentation of incident metabolic, hepatic, cardiovascular, renal, and malignancy-related outcomes during follow-up underscores the ongoing multisystem disease burden captured by the LD-Lync registry. To our knowledge, the Kaplan–Meier estimates provided for the development of these complications are unique, and no similar data were reported in real-world patients followed prospectively to date. Importantly, this study enables both retrospective reconstruction of disease onset and prospective survival tracking, offering an integrated and comprehensive characterization of complication trajectories. Our data provides insights on known or suspected organ complications, such as the development of pancreatitis and kidney disease, within our heterogeneous cohort of 266 patients representing a broad range of lipodystrophy subtypes, including rare variants. Moreover, we provide data on some less recognized aspects, such as pulmonary complications, and GERD (which was quite common). Our findings are further supported by the recently published ECLip Registry, which reported epidemiological and clinical data from a large international cohort of patients with lipodystrophy [40]. Similar to LD-Lync, ECLip identified FPLD as the most common lipodystrophy subtype and demonstrated a high burden of metabolic disease, with dyslipidemia and diabetes among the most frequent complications. ECLip also documented substantial multisystem involvement, including MASLD and cardiovascular complications, as well as mortality primarily related to cardiovascular events and cancer. The convergence of findings across these 2 large international studies strengthens the evidence for substantial cardiometabolic and multisystem morbidity in lipodystrophy and supports continued longitudinal data collection across international cohorts.

Clinical analyses reveal that patients with lipodystrophy have a high burden of disease starting at an early age. As anticipated, diabetes manifested earliest in patients with GL. The prevalence of MASLD in our cohort was 68.4%, considerably higher than the estimated global prevalence of 32.4% [41]. End-organ damage, including cirrhosis, can manifest at a relatively early age in patients with lipodystrophy. Pancreatitis, another potentially life-threatening complication, was found with similar frequency among GL and PL. In our cohort, the prevalence of cirrhosis and pancreatitis exceeded that reported in previous natural history studies. In a previous international study by Akinci et al, liver complications were the most common organ abnormality [8]. Similarly, the prevalence of diabetes was reported as 58.3% in that cohort, whereas our study demonstrated a notably higher prevalence [8]. Our findings identified hypertriglyceridemia as the most frequently observed comorbidity, particularly affecting 3-quarters of patients with GL and typically manifesting during childhood or early adulthood. Based on a recently published study on the Turkish GL cohort, the Kaplan–Meier estimate for the median time to diabetes diagnosis was reported as 16 years [12], consistent with our findings. In our study, the median time to the diagnosis of hypertriglyceridemia in GL was 20 years, whereas it was 14 years in the Turkish GL cohort [12].

It is noteworthy that many organ manifestations occurred while patients were being treated intensively, including while they had access to metreleptin. Limited data suggest that early metreleptin therapy may prevent disease onset when initiated before metabolic complications develop [42]. Many of the registry patients had access to metreleptin after developing metabolic complications. This parallels observations that bariatric surgery benefits decline after diabetes or fatty liver disease develop in stage-3 obesity [43-45]. Although lipodystrophy and stage-3 obesity share many common multisystemic features, patients with lipodystrophy tend to develop complications at lower BMIs, likely due to their markedly reduced lipid storage capacity and the inability of dysfunctional adipose depots to expand.

Renal involvement emerged as one of the most frequently affected organ systems in our cohort; its prevalence was slightly lower than the 40.4% rate reported in a previous international chart review study by our group, possibly reflecting differences in cohort characteristics [8]. While our findings align with previous studies highlighting renal involvement in lipodystrophy, they also exhibit notable differences. For instance, Javor et al reported a prevalence of albuminuria in 88% of patients with GL [46]. The comparatively lower prevalence of proteinuria observed in our cohort may be attributable to the inclusion of patients with more severe metabolic complications in their study [46].

Upon investigating potential new clinical associations of lipodystrophy, asthma/emphysema, stroke, and GERD emerged as notable conditions. Stroke was reported in 5% of patients, underscoring the possible atherosclerotic and vascular consequences of the disease. In addition, among patients with available age-at-onset data, temporal characterization of cardiovascular events showed that diabetes, hypertriglyceridemia, and hypertension often preceded major cardiovascular events and coronary interventions, supporting the relevance of cumulative cardiometabolic risk in this population. The higher frequency of pulmonary issues (asthma and emphysema) observed in patients with FPLD may suggest a possible predisposition within this subgroup. Data from the Third National Health and Nutrition Examination Survey demonstrate a high prevalence of obesity among individuals with asthma (32.8%) [47]. Given the relatively higher BMI observed in FPLD, these findings may partly explain the increased frequency of pulmonary complications in FPLD. Although the link between lipodystrophy and pulmonary pathology has not been well characterized, a Brazilian study on CGL1 patients reported pulmonary fibrosis and infection leading to respiratory failure as potential causes of death [48]. These findings emphasize the need for further studies to clarify the full spectrum of pulmonary involvement in lipodystrophy. Additionally, more than 30% of patients reported symptoms of GERD, with a particularly high prevalence observed in FPLD. This is markedly higher than the pooled global prevalence of 14.8% reported in recent meta-analyses, suggesting a potentially increased burden in lipodystrophy [49].

The association of cancer with the presence of insulin resistance in the general population has attracted attention. The statistically significant higher prevalence of cancer in APL and AGL (exceeding 20% in each) suggests that either the cancer itself or the associated immunomodulators and chemotherapy regimens may contribute to the pathophysiological mechanisms underlying lipodystrophy in these groups. Supporting this, a recent US claims study found increased malignancy rates, particularly breast and prostate cancers, among adults with non-HIV lipodystrophy compared to matched controls [50].

Our findings also suggest a high prevalence of psychiatric comorbidities in patients with lipodystrophy, with over 40% experiencing depression and anxiety, and more than 60% reporting chronic multiregional pain. Statistically significant differences in the prevalence of depression and anxiety were observed between the groups, with FPLD exhibiting the highest prevalence.

Our findings suggest that patients with GL tend to experience a heavier disease burden at an earlier age; however, severe cases of PL with life-threatening complications are also present. The fact that most of the patients who died during the prospective period belonged to FPLD may point to multiple factors contributing to these outcomes. Primarily, the late diagnosis of FPLD likely indicates that many patients had already developed chronic complications at the time of diagnosis. Also, the therapeutic efficacy of metreleptin for most patients with FPLD is limited. Collectively, these factors emphasize the critical need for earlier diagnosis and more effective therapeutic interventions in the management of this patient population.

In our natural history study, GL was associated with more severe metabolic disease compared to PL. However, among patients receiving active metreleptin therapy, those with GL demonstrated a better metabolic profile than those with PL. This observation can be attributed to several factors. First, metreleptin therapy is an approved treatment in the United States and a reimbursed option in Turkey for GL but not approved for PL in these countries. In the United States, patients with GL can initiate metreleptin therapy soon after diagnosis. In contrast, access to metreleptin therapy for PL is limited, primarily restricted to clinical trials and expanded access programs, which predominantly include patients who are unresponsive to standard-of-care treatments for metabolic abnormalities. Second, clinical studies indicate that metreleptin therapy results in more pronounced metabolic improvements in GL compared to PL. Although our study was not designed to evaluate the effectiveness of metreleptin, baseline observations suggest that patients with GL on metreleptin therapy maintain better metabolic health than those with PL. Third, our findings are likely affected by multiple confounding factors, including differences in baseline disease severity. While it can be inferred that patients with GL receiving metreleptin therapy exhibit a more favorable response, the lack of pre- and post-treatment response evaluations limits the strength of our conclusions. Furthermore, comparing treated vs untreated patients introduces bias due to the differing severity of the disease. Similarly, lower HbA1c observed in the untreated group likely reflects an earlier disease stage or absence of metabolic complications, rather than a treatment effect. These findings highlight the need for larger-scale studies to clarify better the efficacy of metreleptin therapy, particularly in PL. The ongoing randomized controlled METRE-PL study is expected to shed further light on this issue. Moreover, large-scale real-world studies are crucial to comprehensively assess the effectiveness of metreleptin therapy in lipodystrophy. An additional important observation in our study was the extensive use of concomitant medications to manage metabolic abnormalities in lipodystrophy, with a higher prevalence in FPLD. Given that the metabolic improvements in patients with PL receiving metreleptin therapy were less pronounced than those observed in GL, it is plausible that a combination of multiple pharmacological agents may be necessary to achieve optimal metabolic control in PL.

Between enrollment and October 2023, 7 deaths were recorded. Although insufficient for formal mortality modeling, these events occurred in participants with advanced disease and suggest a potential link between cumulative complication burden and survival. Continued prospective follow-up will be essential to define mortality risk.

Study limitations

Several limitations should be considered when interpreting these findings. As an international observational registry of an ultra-rare disease, LD-LYNC is subject to potential referral bias, variable follow-up duration, incomplete availability of selected variables, and heterogeneity in clinical practice across participating centers. Historical information used for age-at-onset analyses may also be affected by recall bias and incomplete documentation. These limitations are inherent to many large natural history studies of rare disorders and should be interpreted in the context of the primary descriptive objectives of the present investigation. Despite these challenges, the registry provides one of the largest prospectively followed cohorts of individuals with lipodystrophy assembled to date and offers a unique opportunity to characterize disease burden and longitudinal outcomes across diverse lipodystrophy subtypes.

Although our cohort included almost all subtypes of lipodystrophy syndrome, our study is not exempt from limitations. The current analysis is primarily based on baseline data, with limited prospective follow-up outcomes available at the time of data cutoff. While the registry was designed to capture longitudinal outcomes prospectively, this follow-up data will be collected continuously and analyzed periodically. We anticipate annual or biannual manuscripts on each interesting aspect of our long-lived study. More extended longitudinal follow-up and accrual of additional incident events will be required to generate robust estimates of complication incidence, progression, and predictors. The planned long-term registry follow-up will further expand the characterization of comorbidity trajectories and disease progression over time. Although the LD-Lync modified LDS analysis provided an opportunity to descriptively assess multisystem disease burden over time, it was limited to the domains that could be derived from available registry variables and should not be interpreted as equivalent to the full previously published LDS instrument [39]. The Kaplan–Meier analyses should also be interpreted in light of the mixed retrospective and prospective data structure. Although this approach allowed us to integrate age-at-onset information collected at enrollment with incident events observed during registry follow-up, a substantial proportion of historical age-at-onset data was reconstructed from medical records and participant reports. Therefore, estimates may be affected by recall bias, incomplete documentation, lead-time bias, and differential clinical surveillance across subtypes. In particular, earlier clinical recognition of GL may have led to earlier ascertainment of metabolic and organ-specific complications compared with PL, which is often diagnosed later. Accordingly, these Kaplan–Meier estimates should be interpreted as descriptive measures of complication timing rather than definitive causal estimates of disease progression. Furthermore, retrospective data concerning the duration, age at onset, and severity of complications are derived from available medical records and patient self-reports, which may introduce recall bias and potentially impact our results. Another key limitation of this multicenter study is the potential heterogeneity in clinical assessments and laboratory measurements across participating sites. Although standardized data collection procedures were implemented, differences in local laboratory assays, diagnostic workflows, and clinical practice patterns may have introduced measurement variability, including in lipid parameters such as LDL-C. Accordingly, laboratory-based comparisons were interpreted in the context of clinical diagnoses, medication history, and longitudinal clinical records where available. Laboratory and adipokine measurements were obtained from local laboratories rather than a centralized assay platform; therefore, formal cross-site assay calibration was not possible. Although values were harmonized using standardized units, adipokine analyses should be interpreted cautiously, and future studies using centralized assays are needed to evaluate relationships between adipokine profiles and metabolic complications. Additionally, individuals enrolled in the registry were likely self-selected based on their ability and motivation to participate, which may have introduced a selection bias toward those with higher socioeconomic status, health literacy, and potentially more complex disease, partly explaining the higher comorbidity rates observed compared to earlier studies. Follow-up assessments were also partially disrupted by the COVID-19 pandemic, resulting in missed or delayed annual visits for a subset of participants. Although remote follow-up was implemented when feasible, the absence of standardized in-person evaluations may have introduced gaps in longitudinal assessments and potentially affected the ascertainment of incident outcomes. We also acknowledge that inclusion of both probands and affected relatives may introduce nonindependence due to shared genetic and environmental factors. However, the number of relatives was limited, and the study was not designed or powered for family-based or proband-only stratified analyses. Future studies with larger family clusters will be needed to evaluate intrafamilial variability and genotype-specific disease expression. Lastly, we did not detail the rich data collected from patient-reported questionnaires and hope to report on these very important data dimensions in a future dedicated manuscript very shortly.

Conclusion

To conclude, our study provides a comprehensive overview of natural history and clinical manifestations associated with lipodystrophy syndromes, offering unique insights into both generalized and partial forms of the disease. While significant progress has been made in understanding the role of metreleptin therapy and its metabolic benefits, especially in GL, the limited efficacy observed in PL highlights the need for further investigation and development of targeted therapies. The findings emphasize the importance of early diagnosis, tailored treatment strategies, and long-term follow-up to mitigate disease burden and improve outcomes for these patients. Future studies, particularly randomized controlled trials and real-world investigations, are essential to advance our understanding of therapeutic efficacy and optimize clinical management in this complex and heterogeneous patient population.

Abbreviations

AGL

acquired generalized lipodystrophy

APL

acquired partial lipodystrophy

BMI

body mass index

CGL

congenital generalized lipodystrophy

GERD

gastroesophageal reflux disease

GL

generalized lipodystrophy

HDL

high-density lipoprotein

LDL

low-density lipoprotein

MASLD

metabolic dysfunction-associated steatotic liver disease

NIH

National Institutes of Health

PL

partial lipodystrophy

TG

triglycerides

Funding

This study was previously funded by Aegerion Pharmaceuticals; however, the company or its beneficiaries no longer provide financial support for the study. Current support is provided by the UM Lipodystrophy Fund (PI: Elif A. Oral) and by the National Center for Advancing Translational Sciences (UM1TR004404).

Author contributions

The study was conceived by E.A.O. together with the patient community and A.S. The study structure was organized by E.A.O. and C.S. B.A., M.F.F., and E.A.O. launched the study together at the UM site. M.C.G., B.A., and E.A.O. wrote the initial version of the manuscript and contributed to data review. E.A.O., R.J.B., and B.A. provided critical revisions and editorial input. M.Y. and S.K. conducted statistical analyses using pooled data from all study sites. All authors contributed to patient recruitment, ongoing data collection, and entry. All authors reviewed and approved the final version of the manuscript.

Disclosures

E.A.O. received consultancy fees from Amryt Pharmaceuticals, Chiesi, and formerly Aegerion Pharmaceuticals. E.A.O. has served as a consultant to Ionis, Rhythm, and Regeneron Pharmaceuticals through payments to the University of Michigan and receives grant support from all 3 companies. She also has ongoing clinical trial contracts with Morphic Medical (formerly GI Dynamics), Fractyl, and Novo Nordisk. E.A.O. has received personal consulting fees from Third Rock Ventures and Amryt Pharmaceuticals and serves on the Scientific Advisory Board of Rejuvenate Bio. She has IP related to the use of metreleptin in lipodystrophy and is also entitled to royalty fees. B.A. has run projects for and/or served as a consultant, board member, steering committee member, and/or speaker to Amryt Pharmaceuticals (wholly owned subsidiary of Chiesi Farmaceutici S.p.A.), Alnylam, Regeneron, ThirdRock Ventures, AstraZeneca, Novo Nordisk, Boehringer Ingelheim, Sanofi, Bilim Ilac, ARIS, and Servier. R.J.B. has received support from Pfizer, Regeneron, Marea Therapeutics, and Chiesi Farmaceutici. C.M.V. has received speaker/consultancy fees from Chiesi Pharmaceuticals, PTC Therapeutics, Lilly, Merck, BI-Lilly, and Novo Nordisk; was principal investigator on a Chiesi trial; and was sponsored by Novo Nordisk, Lilly, Chiesi, and PTC for conference attendance. R.M.J. has received speaker and/or consultancy fees from Servier, Nestle, Chiesi Pharmaceuticals, EMS, PTC Therapeutics, Brace Pharma, Hypera Mantecorp, Lilly, Merck, and Novo Nordisk, and was principal investigator on Servier, Chiesi, Sanofi, Ionis, Novo Nordisk, Novartis, and Jansen clinical trials. The remaining authors declare no conflicts of interest.

Data availability

Some or all datasets generated during and/or analyzed during the current study are not publicly available but are available from the corresponding author on reasonable request. Supplemental Data File [38] is openly available in Figshare.

Clinical trial information

This prospective observational registry study is registered at ClinicalTrials.gov under ID NCT03087253.