Clinical Impact of Severe Obesity in Patients with Lymphoedema

4% ( p < .001), hypertension 75.0% vs. 47.6% ( p < .001), and renal disease 24.8% vs. 11.9% ( p < .001). Use of diuretics in the lymphoedema SO þ group was greater: 57.6% vs. 38.0% ( p < .001). Patients with lymphoedema SO þ had higher risk of cellulitis: 34.5% vs. 13.5% ( p < .001). Speci ﬁ c lymphoedema treatment was given more often to lymphoedema SO e : 66.3% vs. 64.3% ( p ¼ .003). This was signi ﬁ cant for manual lymphatic drainage (46.6% vs. 40.0%; p < .001) and physical therapy (55.4% vs. 51.6%; p < .001), but not for compression garments (18.2% vs. 17.7%; p ¼ .38). However, more patients with lymphoedema SO þ received pneumatic compression device treatment: 20.9% vs. 13.7% ( p < .001). Conclusion: There was an increase in SO associated lymphoedema. Patients with lymphoedema SO þ have over a two and half fold increase in cellulitis incidence, with a signi ﬁ cant increase in medical resource use and cost. Despite this, patients with lymphoedema and SO receive less speci ﬁ c therapy such as compression, which has proven to reduce cellulitis incidence.


INTRODUCTION
Obesity is increasingly being recognised as a risk factor for lymphoedema.1e4 Lymphoedema is characterised by a build up of fluid in the intercellular compartment, enriched with proteins including pro-inflammatory substances. 5,6Among these, a significant increase in fibrosis and deposition of adipose tissue has been observed. 1,7,8ecent studies have found a close relationship between obesity and lymphoedema.Lymphatic trauma results in oedema, stimulating adipose tissue deposition, which causes further damage. 1Adipose accumulation from lymphatic damage was shown by higher marker levels of adipocyte deposition such as CEBP-alpha, PPAR-gamma, and adiponectin. 1,8Obesity has also been found to alter lymphatic functioning.9e11 Obese mice had altered lymph node function and on lesioning tail lymphatics mounted an augmented inflammatory response with higher CD4þ levels, adipose accumulation, and fibrosis relative to lean mice. 11A possible mechanism suggested for lymphatic alterations due to obesity has been the rise of inflammatory cells in the perilymph, including increases in T cell and nitric oxide synthase. 12besity in the United States is rising at epidemic proportions with a prevalence of 42.4% and annual healthcare cost of $147 billion. 13While lymphoedema prevalence is limited to 0.13% of the population, it was found in 74% of morbidly obese patients in a sample taken from 17 US wound centres. 14Integrating interventions for weight loss into lymphoedema treatment have already shown preliminary positive results. 15,16There may be a potential body mass index (BMI) benchmark where lymphatic damage could be irreversible, increasing the need for early weight loss intervention. 17Furthermore, obesity also increases the risk of cellulitis, which could cause secondary lymphatic dysfunction.Thus, in patients with obesity induced lymphoedema having a BMI > 40 kg/m 2 weight loss should be a primary therapy. 18here is a growing impetus to understand the relationship between lymphoedema and obesity.The study aimed to accomplish this goal through analysing the demographic, treatment, and clinical profiles of patients with lymphoedema with and without severe obesity (SO).

Study design and data source
A retrospective observational cohort design was employed to analyse data from an integrated US healthcare claims repository (IBM MarketScan Commercial Claims and Encounters [CCAE] and Medicare Supplemental and Coordination of Benefits [MDCR] databases).For this study, data spanned April 2012 to March 2020.Ethics committee approval and informed consent were not obtained as this study was done through a commercial claims database that does not have identifiable patient health information.
The CCAE database includes healthcare claims and enrolment information from employer sponsored plans throughout the United States that provide health benefits to working persons aged < 65 years annually, including the employees, their spouses, and their dependents.The MDCR database includes healthcare claims and enrolment information for retirees who are Medicare eligible and have elected to enrol in employer sponsored Medicare supplementary plans (and for which both the Medicare paid amounts and employer paid amounts are available).Medicare is the federal health insurance program for all persons > 65 years of age regardless of economic status.
Healthcare claims include medical and outpatient pharmacy claims.Data available for each facility and professional service claim include the dates and places of service, diagnoses, procedures performed or services rendered, and quantity of services (professional service claims).Data available for each outpatient pharmacy claim include the drug dispensed, dispensing date, dose, quantity dispensed, and number of therapy days supplied.Medical and pharmacy claims also include amounts paid by health plans and patients to healthcare providers for services rendered.Selected demographic and eligibility information also is available.Patient level data can be arrayed chronologically to provide a detailed longitudinal profile of all medical and pharmacy services used by each plan member.
Patients were excluded from the study population if they had ! 1 diagnosis code for lymphoedema at any time prior to their index dates; < 12 months of healthcare coverage prior to their index dates; < 365 day follow up; or ! 1 diagnosis of head or neck cancer before their index dates (presents with different symptoms of swallowing or phonation dysfunction rather than peripheral oedema).Patients with lymphoedema were categorised into two subgroups; those with vs. without SO, defined as having a BMI > 40.0 kg/m 2 .Among the various obesity categories, this study was focused on SO or class 3. 19 The ICD-9-CM 278.01 and ICD-10-CM E66.01 codes for SO were used.

Study measures
Lymphoedema treatment patterns were ascertained during follow up.The follow up duration was fixed to one year after the patient's respective index dates.Treatments included manual lymphatic drainage (MLD), physical therapy or occupational therapy (PT or OT), compression garments, simple pneumatic compression device, and advanced pneumatic compression device.
Each unique treatment course was identified, beginning with the first, and all qualifying encounters (i.e., with the same Current Procedural Terminology or Healthcare Common Procedure Coding System code) occurring within 30 days of each other was deemed to be part of the same treatment course.Patterns of treatment were characterised in terms of frequency of use, interval from index date to first evidence of treatment, intervals between treatments, and duration of treatment courses.
Use of selected procedures, drugs, and incidence of certain conditions were characterised, including diuretic therapy days, cellulitis (with and without antibiotics), and lymphangitis.Infection and lymphoedema related expenditures were also evaluated.

Baseline characteristics
Baseline characteristics of patients were ascertained during the 12 month period prior to their index dates, and included their demographic profile (age, sex, geographic region of residence, insurance type); clinical profile (lymphoedema related conditions, comorbidities); and treatment profile (diuretics, anti-inflammatory agents).Lymphoedema related conditions and comorbidities were identified based on ! 1 inpatient encounter or ! 2 outpatient encounters with a corresponding diagnosis code.

Data analysis
Baseline characteristics of patients stratified by presence (vs.absence) of SO, were described.Categorical variables were reported as counts and percentages; for continuous variables, means, 95% confidence intervals and standard deviations were reported.Use of alternative treatment modalities, treatment patterns, and clinical outcomes were similarly evaluated.Differences in baseline characteristics and study measures between patients with lymphoedema with and without SO were evaluated using an independent samples t test for continuous measures and chi square or Fisher's exact test for categorical measures.Due to the large sample size, the level of statistical significance was set to .005.
The IBM MarketScan Medicare supplementary database contains data from retirees with Medicare supplementary insurance paid by employers.The database includes the Medicare covered portion of payment, the employer paid portion, and out of pocket patient expenses.The MarketScan Medicare supplementary database provides detailed cost, use and outcomes data for healthcare services performed for inpatients and outpatients.

Baseline characteristics of patients with lymphoedema
The baseline demographics of 60 284 patients with lymphoedema were evaluated (Table 1).Among these, 10.9% of patients were designated to the SOþ cohort and the rest to the SOe cohort.The SOþ cohort was younger (57.8 vs. 60.8 years), with a higher proportion of male patients (37.7% vs. 24.9%)than the SOe cohort.The SOþ group also had a higher percentage of patients living in the south and Midwest regions (75.5%), and fewer in the northeast (15.5%) and west (8.6%) regions (Table 1).The percentage of patients diagnosed with lymphoedema from the SOþ cohort increased from 20.9% to 42.8% over the index date range (Table 2).

Treatment interventions in patients with lymphoedema
The treatment modalities chosen for the patients in the SOþ group significantly differed from those that were not (Table 3).

DISCUSSION
This study established an increase in SO associated lymphoedema.Patients with lymphoedema SOþ exhibited higher proportions of comorbidities.They also had over a two and half fold increase in cellulitis, raising their medical costs, but still received less specific therapy.
According to the recent literature, the expected interstitial fluid re-absorption via the venules simply does not occur and the return happens via the lymphatic system. 20,21th the obesity epidemic there has been growing interest in understanding its contribution to lymphoedema.Patients with a higher BMI have increased intra-abdominal pressure. 22The abnormal amount of adipose tissue and increased pressure could hinder venous return. 23Obese patients have higher venous reflux, pressure, and femoral vein diameter. 23They also have a higher incidence of chronic venous disease. 24An increased deep vein thrombosis and iliac vein obstruction incidence can occur, as seen in the lymphoedema SOþ group.Ultimately, the enlarged venous volume would increase the filtration rate and could result in lymphoedema.In patients with leg lymphoedema, phlebolymphoedema was identified as the most common reason, and morbid obesity was common and significantly associated with the lymphoedema stage. 25Given how lymphoedema can result from venous or lymphatic aetiologies, phlebolymphoedema is defined as dysfunction in both; lymphatic dysfunction co-exists with venous dysfunction at higher CEAP (Clinical, Etiological, Anatomical, Pathophysiological) levels C3 e C6, indicating that it takes longer for the lymphatic dysfunction to occur. 25,26Although obesity related lymphoedema can be clinically improved, the lymphatic damage cannot be cured. 27Lymphoedema that may be without clinical symptoms, but persistent underlying lymphatic dysfunction, is known as "systemic subclinical lymphoedema". 27It was prevalent in w16% of women with lipoedema and BMI < 30 kg/m 2 and increased with higher BMI, illustrating the exacerbating role obesity plays. 28However, as such, subclinical lymphoedema is being under recognised and undertreated.In operations on lymphoedema afflicted regions, not only fluid accumulation but also fat deposits are prominent. 21Liposuction has been successful to treat lymphoedema by decreasing the volume of the limb, improving psychological well being, and future disease management. 29,30rst, it was found that patients with lymphoedema SOþ had more comorbidities.They had higher occurrences of diabetes, hypertension, and heart failure.Obese patients are generally known to suffer from several such comorbidities and have a higher mortality rate. 31Several factors contribute to this, including lack of physical activity. 32A US healthcare database study found that patients with higher BMI had higher rates of hypertension, type 2 diabetes, and Data is expressed as % (95% confidence interval) or mean (95% confidence interval).Data are expressed as % (95% confidence interval) or mean (95% confidence interval) CONS ¼ conservative care; SPCD ¼ simple pneumatic compression device; APCD ¼ advanced pneumatic compression device; FU ¼ follow up.
cardiovascular issues. 31There is additionally a complex interplay between the lymphatic system, obesity, and comorbidities such as diabetes or heart disease.
The lymphatic system has been implicated in inflammatory processes and the development of chronic degenerative diseases. 33The lymphatic system is closely linked to lipid levels by clearing cholesterol in interstitial fluid and absorption from food via lacteals. 34A mutualistic relationship was found where pathology in the lymphatic system can predispose to obesity, and obesity can conversely further its dysfunction. 1,34The lymphatic system also influences development of atherosclerosis. 35Lymphatic impairment in diabetes has been linked to higher lymphangiogenesis, lesser lymph node uptake, and increased thoracic duct flow. 34Given the lymphatic involvement, it is not surprising that patients with lymphoedema often had diabetes, heart failure, and hypertension.This was especially true in SO patients whose risk of these comorbidities was compounded.SO patients with lymphoedema have more complex pathologies and higher comorbidity risks, necessitating specialised care and attention.It should be emphasised that weight reduction, an often ignored approach is essential not only to treat lymphoedema, 36 but all other associated comorbidities as well.
Cellulitis was more common in the patients with lymphoedema SOþ.Since lymphatics are significantly involved in the adaptive immune system, their dysfunction could explain an increased risk of cellulitis. 21A French case control study showed that this risk was 71 fold (95% CI 5.6 e 908), but unlike this study where controls included patients with lymphoedema who were not obese, their controls had conditions such as trauma, infection, abdominal surgery, and others without lymphoedema. 37Similar trends have been shown in Canada where patients with lymphoedema with an occurrence of cellulitis had higher BMI, with a third of the obese cohort having suffered from it. 38A study in Poland on patients with cellulitis or erysipelas found the recurrent condition to be correlated with obesity and lymphoedema. 39While lymphoedema could be the trigger for cellulitis, lymphoedema could also be augmented from cellulitis causing lymphatic dysfunction. 40Understanding this association may be instrumental because, as suggested in the study, therapies such as antibiotic prophylaxis could be proposed for obese patients with lymphoedema. 39espite the higher comorbidities and infection rates, patients with lymphoedema SOþ still received less targeted medical intervention.The primary treatment for lymphoedema is complex decongestive treatment (CDT), including compression, exercise, and MLD. 33Interestingly, these interventions have been shown to reduce the incidence of cellulitis in patients with oedema. 41,42Although the percent difference of targeted medical intervention was small, yet statistically significant, one would expect people with more severe disease to receive a lot more treatment and as early as possible.In this study, the severely obese cohort received less MLD and PT or OT.They also received less compression, albeit not statistically significant.One potential reason could be more misdiagnoses in obese patients from lymphoedema complicated differential diagnosis.The manifestation of lymphoedema can be similar to other oedematous conditions such as venous issues, drug related oedema, lipoedema, endocrine pathology, trauma, infection, and obesity. 43,44Specifically in obese patients, on conducting a physical examination it is hard to distinguish externally whether an obese leg is affected by lymphoedema. 17Furthermore, a diagnostic technique for lymphoedema such as a Stemmer sign can be false positive in those with obesity due to the higher adipose accumulation. 17besity related lymphoedema can also be misdiagnosed as lipoedema. 17Lipoedema is an adipose pathology prevalent in women that is distinct from obesity and presents as painful fat accumulation, primarily in the legs, and cannot be treated through diet or exercise interventions. 17,45urther differentiating it from lymphoedema, it presents with a negative KaposieStemmer sign, excludes the feet and has a normal lymphoscintigraphy, but it can progress to lipolymphoedema. 17,46Increasing awareness of the obesity correlation with lymphoedema and considering it in diagnostic protocols could allow for more efficient treatment.Apart from CDT, more well rounded treatment approaches should also be taken regarding epigenetic influences such as nutrition and lifestyle. 33,36Another reason for reduced intervention could be that obese patients pursue medical help later than others due to the stigma and poor treatment they receive by the healthcare community leading to a vicious circle. 47besity has also huge financial implications.Those with SO had twice the probability of having medical costs than the normal weight group. 48It could be attributed to higher drug related costs, in or outpatient care, and explained by the occurrence of more comorbidities with obesity like heart disease. 14,48In this study, the lymphoedema SOþ group had higher infection related and lymphoedema related costs than the lymphoedema SOe group.Recognising obesity as a risk factor and addressing it could help reduce the healthcare burden for patients with lymphoedema.
The demographics of patients with lymphoedema with SO were found to be altered when compared with those without.The lymphoedema SOþ patients were younger, with male preponderance and most living in the south and Midwest, than the lymphoedema SOe group.

Study limitations
This study has several limitations.Administrative claims are observational without the usual statistical rigour of a randomised controlled trial.Propensity score matching can be used to account for the observational nature of the data but does not substitute for a fully randomised data set.The identification of lymphoedema and associated comorbidities within the administrative claims relies on the proper use of diagnosis coding for claims algorithms.Lymphoedema, other comorbidities, and procedures can only be identified if they are coded correctly.The IBM MarketScan database provides basic characteristics of the population, which include gender, age, and geographic location of the subscriber in the United States, but in this database race, ethnicity, and income status are not provided.The IBM MarketScan database contains commercially insured patients and does not include the original (parts A and B) fee for service Medicare patients.As a result, the data set may under represent Medicare and Medicaid populations.The IBM MarketScan Medicare supplementary database, however, contains data from retirees with Medicare supplementary insurance paid by employers.This database includes the Medicare covered portion of payment (represented as Coordination of Benefits Amount or COB), the employer paid portion, and out of pocket patient expenses.The MarketScan Medicare Supplementary Database provides detailed cost, use, and outcomes data for healthcare services performed in both inpatient and outpatient settings.Furthermore, claims data do not indicate the severity or initial lymphoedema onset, therefore the study cannot examine trends in more severe patients or compare treatment timing relative to diagnosis (if outside the span of provided data).Finally, this study set the threshold to class 3 obesity, but many people fall below this and the previous categories of obesity deserve to be studied further as well.

Conclusion
Given the obesity epidemic, this study focused on understanding its contribution to lymphoedema.It was found that severely obese patients with lymphoedema were younger, with more men, and faced higher rates of comorbidities.They were also two and half fold more likely to suffer from cellulitis.However, they still received less targeted lymphoedema treatments such as PT, OT or MLD and had a higher healthcare expenditure making the disease both a health and economic burden.

CONFLICT OF INTEREST
Derek Weycker received consultative reimbursement from Tactile for his independent performance of the health economic analysis.Antonios Gasparis is a consultant to Tactile Medical.Thomas O'Donnell Jr is a consultant to Tactile Medical FUNDING Tactile Medical (Minneapolis, MN, USA)

Table 1 .
Baseline demographics of patients with lymphoedema included in this retrospective cohort study based on healthcare claims, divided into severely obese and not severely obese patient cohorts Data are expressed as n (%) or as mean AE standard deviation.A lower proportion of SOþ patients (64.3%, 95% CI 63.1 e 65.5) received lymphoedema targeted treatment than SOe patients (66.3%, 95% CI 65.9 e 66.8).Fewer SOþ patients received MLD and PTor OT than SOe patients (p < .001),but it was not significant for compression garments (p ¼ .38).However, more SOþ patients received simple and advanced pneumatic compression device interventions (p < .001).

Table 2 .
Health related characteristics of patients with lymphoedema included in this retrospective cohort study based on healthcare claims, divided into severely obese and not severely obese patient cohorts Data are expressed as n (%).

Table 4 .
Other health related and economic outcomes in patients with lymphoedema included in this retrospective cohort study based on healthcare claims, divided into severely obese and not severely obese patient cohorts

Table 3 .
Comparison of different treatment options, including conservative care (manual lymphatic drainage, physical or occupational therapy, or compression garments) and simple or advanced pneumatic compression devices used in patients with lymphoedema included in this retrospective cohort study based on healthcare claims, divided into severely obese and not severely obese patient cohorts