Weight Loss for Patients with Obesity: An Analysis of Long-Term Electronic Health Record Data

Abstract

Background:

Numerous studies have reported that losing as little as 5% of one’s total body weight (TBW) can improve health, but no studies have used electronic health record (EHR) data to examine long-term changes in weight, particularly for adults with severe obesity (BMI ≥35 kg/m²).

Objective:

To measure long-term weight changes and examine their predictors for adults in a large academic healthcare system

Research Design:

Observational study

Subjects:

We included 59,816 patients aged 18-70 who had at least 2 body mass index (BMI) measurements 5 years apart. Patients who were underweight, pregnant, diagnosed with cancer, or had undergone bariatric surgery were excluded.

Measures:

Over a 5-year period: 1) ≥ 5% total body weight (TBW) loss; 2) weight loss into a non-obese BMI category (BMI < 30 kg/m²); and 3) predictors of %TBW change via quantile regression.

Results:

Of those with class 2 or 3 obesity, 24.2% and 27.8%, respectively, lost at least 5% TBW. Only 3.2% and 0.2% of patients with class 2 and 3 obesity, respectively, lost enough weight to attain a BMI < 30 kg/m². In quantile regression, the median weight change for the population was a net gain of 2.5% TBW.

Conclusions:

Although adults with severe obesity were more likely to lose at least 5% TBW compared to overweight patients and patients with class 1 obesity, sufficient weight loss to attain a non-obese weight class was very uncommon. The pattern of ongoing weight gain found in our study population requires solutions at societal and health systems levels.

Introduction

Obesity, defined as having a body mass index (BMI) of at least 30.0 kg/m², is a significant public health problem. According to data from the Centers for Disease Control (CDC), over one-third of United States (U.S.) adults are obese.¹ Obesity increases the risk of multiple comorbidities, such as type 2 diabetes (T2DM), hypertension, and hyperlipidemia.² It is also a major risk factor for multiple types of cancer, including endometrial, breast, and kidney.³ As a result, obesity is the second leading cause of preventable deaths in the U.S.³ Most studies of obesity trends use national databases or surveys, such as the National Health and Nutrition Examination Survey (NHANES), for population-level estimates.^4–6 While there are numerous studies examining obesity trends using electronic health record (EHR) data in children and adolescents,^7–9 there are limited studies using EHR data in adults.^10,11 The peer-reviewed literature on adult obesity using EHR analysis is limited by cross-sectional study designs,^12,13 small sample sizes,^14,15 and short-term data.^12–15

For individuals with obesity, a weight loss goal of 5% total body weight (TBW) is often recommended. Numerous studies have reported that losing as little as 5% of one’s total body weight can improve metabolic and cardiovascular health.^16–18 This goal has been supported by the CDC,¹⁹ the U.S. Surgeon General,²⁰ and the National Institutes of Health (NIH).^3,21 In 2013, an expert panel composed of members of the American College of Cardiology, American Heart Association, and The Obesity Society concluded that 5% TBW loss was a marker for “clinical significance”, as this quantity of weight loss demonstrated health benefits.²² While greater weight loss is associated with increased improvements in comorbidities,^16,23 the clinical significance of modest weight loss, particularly for adults with severe obesity, is unclear. It is also unknown if individuals who lose at least 5% TBW are able to attain sufficient weight loss to join a non-obese BMI category.

The objective of this study was to use EHR data from a large healthcare system in the U.S. to evaluate longitudinal weight changes for adult patients who had not undergone bariatric surgery. We aimed to evaluate the prevalence of 5% TBW loss and quantify how often weight loss into a non-obese BMI category occurred over a 5-year period for adults with obesity. Further, we sought to identify patient characteristics associated with weight changes.

Methods

Data Source

We used EHR data from the University of Wisconsin Hospital and Clinics from 6/1/2008-12/31/2016, which contains information on more than 500,000 patients.²⁴ Patient information was stored on a secure server through the University of Wisconsin Health Information Services and the Institute for Clinical and Translational Research. Epic Clarity Database was the data source. This study was approved by the University of Wisconsin Minimal Risk Institutional Review Board (IRB), and the need for informed consent was waived.

Study Population

All patients between the ages of 18-70 who had a BMI measurement at least 6 months after their initial in-person encounter in the EHR from 6/1/2008-12/31/2011 were included. We used 12/31/2011 as the latest initial encounter to ensure all patients had at least 5 years of follow-up data. All included patients had a second encounter with a BMI measurement a minimum of 5 years (± 6 months) after their initial measurement. Underweight patients (BMI < 18.5 kg/m²) were excluded. Patients with any pregnancy or cancer diagnoses at any time during the study period were excluded using ICD-9/-10 codes. Bariatric surgery patients were excluding using our institutional bariatric surgery registry.

We used a 5-year criterion to maximize both follow-up time and sample size. The maximum follow-up time of our cohort was 8 years, with a median of 5.3 years. We found that 54.6% of our total cohort had 5-year follow-up. Only 19.7% and 0.8% of patients had follow-up of 7 and 8 years, respectively. While we lost about half of the initial sample size with a 5-year inclusion criterion, the size of our final cohort allowed us to maximize both the duration of follow-up and sample size.

Data Validation and “Cleaning”

To minimize the likelihood of including incorrect heights and weights due to data entry errors in the EHR, recorded heights and weights were “cleaned.” Biologically implausible heights and weights were removed according to the algorithm by Cheng.²⁵ These included: 1. weights above 1000 or below 55 pounds; 2. weights greater than 70% of the range from the mean when the range was greater or equal to 50 pounds; 3. weights greater than 1 standard deviation (SD) from the mean weight when the SD was greater than 20% of the mean; 4. heights above 90 or below 44 inches; and 5. heights greater than 1 SD from the mean height when the SD was greater than 2.5% of the mean. For encounters with missing height data, height was imputed with the most recent previous validated height. Any remaining missing heights were filled with the most recent subsequent valid height. Missing BMI values were calculated with cleaned heights and weights.

Study Variables

Baseline BMI, age, sex, race/ethnicity, and insurance type were retrospectively identified from the EHR. Baseline insurance type was defined as the insurance type most frequently used in the 6 months prior to or after the initial baseline encounter. ICD-9/-10 codes (see Supplementary Digital Content 1, which contains the full list of ICD-9/-10 codes) were used to identify 12 obesity-related comorbidities (anxiety, cardiovascular disease, cerebrovascular disease, chronic pain, depression, gastroesophageal reflux disease [GERD], hyperlipidemia, hypertension, non-alcoholic fatty liver disease [NAFLD], obstructive sleep apnea [OSA], osteoarthritis, and type 2 diabetes mellitus [T2DM]). Baseline comorbidities were identified based on the occurrence of an ICD-9/-10 code in the 6 months prior to or after the initial baseline encounter. Healthcare utilization was defined as the total number of all inpatient visits, outpatient visits, and telephone calls with a healthcare provider in the 6 months prior to or after their initial encounter in the EHR.

Study Outcomes

The primary outcomes of interest were ≥ 5% TBW loss over a 5-year period and weight loss to BMI < 30 kg/m². When evaluating weight loss to a non-obese BMI class, we only included patients with an initial BMI ≥ 30 kg/m².

Statistical Analysis

Baseline demographics and prevalence of obesity-related comorbidities were compared between patients with and without ≥ 5% TBW loss, as well as those with and without sufficient weight loss to attain a BMI < 30 kg/m², using Mann-Whitney tests for binary variables, Kruskal-Wallis tests for variables with multiple categories, and Pearson’s correlation coefficient for continuous variables.

Binomial linear regression analysis was initially conducted to determine predictors of ≥ 5% TBW loss while adjusting for baseline patient demographics and obesity-related comorbidities. Attempts to fit a linear regression model resulted in large and skewed residuals, suggesting that a continuous quantile model would be a better fit. Quantile regression analysis of the median was used to determine associations between the same set of covariates with %TBW change as a continuous outcome. We included all significant differences in patient demographics and obesity-related comorbidities in the quantile regression model, as recommended by Kleinbaum’s textbook when building a predictive model.²⁶ For our quantile regression analysis, the reference was a female patient, with BMI of 29.3 kg/m² (mean BMI), age 45.9 years (mean age), with 2 visits (median healthcare utilization), white, non-Hispanic, with commercial insurance, and with none of the 12 obesity-related comorbidities. The %TBW change for each independent variable should be compared to the reference %TBW change.²⁷

R version 3.4.2 was used to conduct the analysis. The quantreg R package was used for quantile regression analysis.

“Non-Responder” Analysis

Baseline demographics and prevalence of obesity-related comorbidities were compared between patients who had ≥ 2 BMI measurements and those who did not using Mann-Whitney tests for binary variables, Kruskal-Wallis tests for variables with multiple categories, and Pearson’s correlation coefficient for continuous variables. We also used a logistic regression model with the outcome being ≥ 2 BMI measurements. We included patient characteristics, obesity-related comorbidities, Charlson comorbidity index (CCI), healthcare utilization, and whether patients had a primary care provider (PCP) as independent variables. The reference was a female patient, with BMI of 29.0 kg/m² (mean BMI of patients with and without ≥ 2 BMI measurements), age 43.9 years (mean age), with 4 visits (median healthcare utilization), white, non-Hispanic, with commercial insurance, with none of the 12 obesity-related comorbidities, and with Charlson Comorbidity Index (CCI) score of 0.

Results

Patient Characteristics

Nearly 300,000 patients had at least 1 in-person encounter with a BMI measurement during their initial baseline period (Figure 1). After applying exclusion criteria, 59,816 patients were included in the analysis. The mean age was 45.9 years (Table 1). Within our study cohort, 54.4% of patients were female, and 90.8% of patients were white. The majority of patients were commercially insured (86.7%). The mean baseline BMI of our study cohort was 29.3 kg/m². Of all included patients, 17.1% had severe obesity (BMI ≥ 35.0 kg/m²) at baseline. Hyperlipidemia and hypertension were the most prevalent comorbidities, affecting 18.2% and 16.4%, respectively, of patients.

Figure 1.

Strengthening the Report of Observational Studies in Epidemiology (STROBE) Diagram

Table 1.

Characteristics and Baseline Demographics of Patients

	n = 59,816 patients with 2 BMI measurements (n,%)
Age (years)	mean 45.9 (SD 12.9)
18-30	8,151 (13.6)
31-40	10,788 (18.0)
41-50	16,005 (26.8)
51-60	16,103 (26.9)
61-70	8,769 (14.7)
Sex
Male	27,278 (45.6)
Female	32,538 (54.4)
Race/ethnicity
White, non-Hispanic	54,322 (90.8)
Black, non-Hispanic	1,879 (3.1)
Asian, non-Hispanic	1,378 (2.3)
Native American, non-Hispanic	251 (0.4)
Hispanic	1,453 (2.4)
Other/unspecified	533 (0.9)
Baseline BMI category	mean 29.3 (SD 6.8) median 28.1 (IQR 24.4-32.7)
Normal	17.159 (28.7)
Overweight	19,971 (33.4)
Class 1 obesity	12,475 (20.9)
Class 2 obesity	5,851 (9.8)
Class 3 obesity	4,360 (7.3)
Insurance type
Commercial	51,959 (86.9)
Medicare	2,371 (4.0)
Medicare (under 65)	1,993 (3.3)
Medicaid	1,483 (2.5)
Other/unspecified	2,010 (3.4)
Healthcare utilization (visits)	mean 3.9 (SD 4.8) median 2 (IQR 1-5)
Obesity-related comorbidities
Anxiety	2,551 (4.3)
Cardiovascular disease	1,274 (2.1)
Cerebrovascular disease	519 (0.9)
Chronic pain	683 (1.1)
Depression	4,881 (8.2)
Gastroesophageal reflux	3,665 (6.1)
Hyperlipidemia	10,880 (18.2)
Hypertension	9,827 (16.4)
Nonalcoholic fatty liver disease	255 (0.4)
Obstructive sleep apnea	1,307 (2.2)
Osteoarthritis	2,957 (4.9)
Type 2 diabetes mellitus	2,439 (4.1)

Trends in Weight Change Over a 5-Year Period

Of patients with class 2 and 3 obesity, 24.2% and 27.8%, respectively, lost ≥ 5% TBW; 15.6% of overweight patients and 20.8% of patients with class 1 obesity lost ≥ 5% TBW (Table 2). 11.3% and 14.4% of patients with class 2 and 3 obesity, respectively, lost 10% TBW or more. Only 3.2% and 0.5% of class 2 and 3 obesity patients, respectively, were able to attain a BMI < 30 kg/m². Of patients who lost ≥ 5% TBW, only 13.2% of patients with class 2 obesity and 1.7% of patients with class 3 obesity attained a non-obese BMI. Across all BMI strata, the majority of patients remained in the same weight class (Figure 2). Of those with class 2 or 3 obesity, 52.3% and 80.9%, respectively, stayed in the same weight class while only 21.6% and 15.0%, respectively, decreased 1 weight class.

Table 2.

Weight Change Over 5-Year Period

Initial BMI class	Normal (18.5-24.9 kg/m2) (n,%)	Overweight (25.0-29.9 kg/m2) (n,%)	Class 1 obesity (30.0-34.9 kg/m2) (n,%)	Class 2 obesity (35.0-39.9 kg/m2) (n,%)	Class 3 obesity (≥ 40 kg/m2) (n,%)
Number of patients	17,159 (28.7)	19,971 (33.4)	12,475 (20.9)	5,851 (9.8)	4,360 (7.3)
Lost any weight	5,457 (31.8)	7,533 (37.7)	5,358 (42.9)	2,671 (45.7)	2,159 (49.5)
Decreased 1 BMI category	337 (2.0)	2,103 (10.5)	2,383 (19.1)	1,452 (24.8)	833 (19.1)
Lost ≥ 5% TBW	2,024 (11.8)	3,120 (15.6)	2,594 (20.8)	1,414 (24.2)	1,212 (27.8)
Lost ≥ 10% TBW	536 (3.3)	1,117 (5.6)	1,070 (8.6)	664 (11.3)	629 (14.4)
Ending BMI < 30 kg/m2	--	--	2,383 (19.1)	186 (3.2)	20 (0.5)

Figure 2.

Change in BMI Over 5-Year Period

Legend: Each dot represents one patient. The x-axis represents baseline BMI and the y-axis represents BMI after 5 years.

Characteristics of ≥ 5% TBW Loss

In unadjusted analysis, ≥ 5% TBW loss was associated with older age (47.8% of patients who lost ≥ 5% TBW were 51 years of age or older vs. 40.3%, p<0.001) and female sex (57.5% of female patients lost ≥ 5% TBW vs. 53.8%, p>0.001) (see Supplementary Digital Content 2, which contains the table of the unadjusted analysis comparing patients who lost ≥ 5% TBW vs. those who did not). Patients who lost ≥ 5% TBW had higher mean baseline BMIs compared with those who did not (31.3 kg/m² vs. 28.9 kg/m², p<0.001). Patients with Medicare or Medicaid were also more likely to lose ≥ 5% TBW (14.7% vs. 8.7%, p<0.001).

When comparing mean %TBW change, patients demonstrated weight gain across nearly all baseline demographics and comorbidities (see Supplementary Digital Content 2, which contains the table of unadjusted mean %TBW changes). Patients who had T2DM or were insured with Medicare had mean losses of 0.4% and 0.3% TBW, respectively, but these losses were not statistically significant.

Characteristics of Weight Loss to BMI < 30 kg/m²

Older patients were more likely to attain a BMI < 30 kg/m² (48.3% of patients who attained a BMI < 30 kg/m² were 51 years of age or older vs. 45.5%, p<0.001) (see Supplementary Digital Content 3, which contains the table of the unadjusted analysis comparing patients who attained a BMI < 30 kg/m² vs. those who did not). The distribution across sexes and races/ethnicities was similar amongst those who attained a BMI < 30 kg/m² versus those who did not. Patients who attained a BMI < 30 kg/m² had lower baseline BMIs (31.9 kg/m² vs. 36.6 kg/m², p<0.001). Of those who attained a BMI < 30 kg/m², 8.0% had class 2 or 3 obesity. In contrast, 49.8% of patients who did not attain a BMI < 30 kg/m² had class 2 or 3 obesity (p<0.001). Patients with Medicare or Medicaid (9.3% vs 7.1, p<0.001) were more likely to attain a BMI < 30 kg/m², while patients who had hyperlipidemia (23.5% vs. 25.8%, p=0.01), hypertension (20.1% vs. 26.7%, p<0.001), OSA (6.1% vs. 6.8%, p<0.001), or T2DM (5.9 vs. 8.1%, p<0.001) were less likely to attain a BMI < 30 kg/m².

Predictors of Weight Changes/Quantile Regression

In quantile regression with median percent change in TBW as the outcome, the median weight change for the study cohort over a 5-year period was a 2.5% TBW increase. Asian patients (−1.5% TBW change, 95% CI [−1.8, −1.1]), those with greater baseline BMIs (−0.6% TBW change per 5 kg/m² in BMI, 95% CI [−0.7, −0.6]), males (−0.6,% TBW change, 95% CI [−0.7, −0.4]), and older patients (−0.6% TBW change per 5 year increase in age, 95% CI [−0.7, −0.6]) gained less weight, but still had net positive weight gain over this period when compared to the reference +2.5% TBW change (Figure 3). Patients with T2DM also gained less weight (−1.4% TBW change, 95% CI [−1.8, −0.9]), while patients with hypertension (+0.5% TBW change, 95% CI [0.2, 0.7]) and OSA (+1.1% TBW change, 95% CI [0.5, 1.6]) gained more weight.

Figure 3.

Quantile Regression Analysis of %TBW Change

Legend: Each covariate included in the analysis is shown on the y-axis. The x-axis represents median %TBW change. The reference is a female patient, with BMI of 29.3 kg/m2 (overall mean BMI), age 45.9 years (overall mean age), with 2 visits (overall median healthcare utilization), white, non-Hispanic, with commercial insurance, and with none of the 12 obesity-related comorbidities. “+X” indicates increments of X units.

“Non-Responder” Analysis

When comparing patients who had complete 5-year BMI data versus those who did not, patients who were older (mean age of 45.9 years vs. 41.4, p<0.001), female (45.6% vs. 47.5%, p<0.001), white, non-Hispanic (90.8% vs. 87.3%, p<0.001), had higher baseline BMIs (mean BMI of 29.3 kg/m² vs. 28.6 kg/m², p<0.001), were commercially insured (83.8% vs. 78.7%, p<0.001), and had anxiety, depression, GERD, hyperlipidemia, hypertension, OSA, osteoarthritis, or T2DM were more likely to have ≥ 2 BMI measurements (see Supplementary Digital Content 4, which contains the table of the unadjusted analysis comparing patients who had ≥ 2 BMI measurements vs. those who did not).

In adjusted analysis, we found that predictors of having ≥ 2 recorded BMIs included older age (OR 1.12, 95% CI [1.11, 1.13]), anxiety (OR 1.16, 95% CI [1.08, 1.25]), GERD (OR 1.10, 95% CI [1.03, 1.16]), hyperlipidemia (OR 1.24, 95% CI [1.18, 1.29]), and T2DM (OR 1.12, 95% CI [1.03, 1.21]) (see Supplementary Digital Content 5, which contains the figure of the logistic regression analysis with having ≥ 2 BMI measurements as the outcome).

Discussion

Our findings suggest that patients with severe obesity were more likely to lose ≥ 5% TBW compared to overweight patients or those with class 1 obesity, but very rarely attained a BMI < 30 kg/m². Overall, patients in our health system gained weight during the 5-year study period, with youngest adults most likely to gain weight. Higher baseline BMI and presence of type 2 diabetes were associated with less weight gain in adjusted analyses.

We found that while nearly one in four patients with severe obesity lost at least 5% TBW, less than 3% attained a non-obese BMI. Similar results were seen in Fildes’ cohort study using EHR data of adults in the United Kingdom (U.K.), which showed the likelihood of attaining ≥ 5% TBW loss increased with BMI, but the likelihood of attaining a normal BMI decreased with greater baseline BMI.²⁸ Likewise, Calderón-Larrañaga’s EHR cohort study of 42,248 Spanish adults demonstrated that patients with greater BMIs were more likely to experience weight loss.²⁹ Higher baseline BMI as a predictor of greater weight loss was also reflected in Teixeira’s review article of psychosocial predictors of weight loss³⁰ and Jackson’s retrospective cohort study of U.S. and U.K. adults.³¹

Given that very few patients with severe obesity lost sufficient weight to join a non-obese BMI class, the appropriateness of using 5% TBW loss as a marker of weight loss “success” is unclear. No studies have demonstrated significant resolution of obesity-related comorbidities with 5% TBW loss. Additionally, the literature suggests that 5% TBW loss is difficult to maintain. In a systematic review and meta-analysis by Franz that included 19 randomized clinical trials of non-surgical weight loss interventions, only 2 studies reported 5% TBW loss at 1 year; no studies reported sustained weight loss after 1 year.²³ Although the CDC, the Surgeon General, and the NIH advocate for the benefits of 5% TBW loss,^3,19–21 the literature suggests that long-term maintenance of 5% TBW loss is unlikely. For patients with severe obesity, perhaps a more clinically relevant weight loss goal should be established as a marker for weight loss success, such as weight loss to a non-obese BMI or resolution of obesity-related comorbidities. Prospective trials evaluating weight loss and complete resolution of comorbidities would be informative.

We also found that our patient population gained about 2% TBW over the study period. Concerningly, younger patients were the most likely to gain weight. Kelly’s pooled analysis of worldwide data estimated that by 2030, 20% of the world’s population will be obese and 38% will be overweight.³² A systematic analysis of children and adults worldwide showed that rates of severe obesity continue to trend upwards.³³ Multiple studies have also demonstrated these trends of increasing weight gain specifically in younger populations; Poobalan’s systematic review and Lim’s retrospective cohort study identified similar trends of obesity in young adults in developing countries and Korea, respectively.^34,35 Additionally, Noël’s retrospective cohort study of U.S. Veterans showed that younger Veterans were less likely to lose weight regardless of receipt of weight loss counseling.³⁶

Our finding that the youngest adults are the highest risk cohort for weight gain is concerning, as this has long-term repercussions for this population. In a cohort analysis by Zheng using U.S. national databases, weight gain during early adulthood was found to significantly increase the risk of major chronic diseases and decrease the odds of healthy aging.⁵ A retrospective study published in 2019 by Sung utilizing U.S. cancer registries also showed that the risk of developing obesity-related cancers has increased in successively younger birth cohorts.³⁷ While these increased risks have multifactorial etiologies, the authors speculate that increasing obesity trends in younger populations plays a significant role. The pattern of weight gain in younger patients is alarming, as it puts future populations at higher risk of cancer and chronic health conditions. Thus, early preventive measures and weight loss interventions targeting young adults will be critical.

In our adjusted analysis, patients with T2DM were less likely to gain weight. Similarly, Calderón-Larrañaga’s EHR cohort study also found that patients with diagnosis of T2DM were less likely to gain weight. Bramlage’s cross-sectional study of German adults demonstrated that patients with diabetes and diabetes-related complications were more likely to use weight loss interventions.³⁸ In addition, Waring’s cohort study of U.S. adults with obesity showed that diabetics were more likely to receive weight management counseling.³⁹ As weight loss is important for glycemic control,^23,40,41 diabetic patients may receive more weight loss counseling than the general population. A meta-analysis by Rose found that weight loss counseling from primary care providers is critical for weight loss success.⁴² Furthermore, many diabetes medications may cause weight loss, such as semaglutide.^43,44 To better understand why diabetic patients were less likely to gain weight, qualitative studies would be helpful to identify patient factors associated with weight loss success.

Our study has several limitations. First, there may be selection bias given our inclusion criteria that all patients were required to have been in our healthcare system for at least 5 years. Our “non-responder” analysis identified that patients who had 5-year follow-up were slightly older and had more comorbidities, but there were no clinically significant differences in baseline BMI between the groups. Additionally, we excluded patients who were pregnant or had cancer, which may introduce selection bias. Second, we were unable to determine if weight loss was intentional, what motivated changes in weight, or if patients received weight loss counseling or non-surgical obesity treatments. Third, given that our study was observational, there may be unmeasured confounding that biased our findings. Fourth, there may also be errors in data entry by healthcare providers. We attempted to remove biologically implausible outliers with our height and weight cleaning and imputing algorithm. Fifth, there may be errors in entry of ICD-9/-10 codes by healthcare providers, which is a limitation of EHR datasets. Moreover, we were unable to assess for resolution of comorbidities using ICD-9/-10 codes. Sixth, we used BMI as a measurement of obesity and its related risks, as we do not have waist circumference data available. Finally, our analysis was a single institution study. Despite this, our previous study using this dataset demonstrated similar rates of obesity in our EHR compared with national data from the U.S.²⁴ Our EHR data also contains a sample size that is nearly 20 times larger than what is used by NHANES, which produces national U.S. obesity estimates.⁴⁵

In summary, using EHR data, we found that although more than one-fourth of patients with severe obesity were able to reach at least 5% TBW loss, weight loss into a non-obese category was exceedingly rare. We also identified increasing trends in weight gain, especially for younger populations. Further studies evaluating the impact of at least 5% TBW loss versus attaining a non-obese BMI category are indicated so recommendations of optimal weight loss goals can be made for obese patients. Earlier preventive measures and weight loss interventions targeted towards younger populations are necessary to combat increasing obesity trends.