The Obesity Paradox: A Further Consideration in Dialysis Patients

Abstract

Several epidemiological cohorts have demonstrated that higher body mass index (BMI) is associated with lower mortality risk among patients receiving hemodialysis. However, BMI may be an inaccurate indicator of nutritional status among dialysis patients because it does not differentiate between muscle and fat mass or provide information about body fat distribution. More sophisticated methods of body composition analysis are therefore required to address the question of which component is associated with greater survival. Recent evidence has also shown that changes in body weight and body composition are more strongly associated with mortality in dialysis patients than measurement of BMI at a single time point. Given that obesity is common among the dialysis population, weight loss interventions are encouraged for obese dialysis patients who are on a transplant waiting list in order to increase the access for transplantation.

Over the past decade, a phenomenon called the “obesity paradox” has been reported in several large cohorts of chronic kidney disease patients treated with maintenance dialysis. Data on the paradoxical associations of obesity have been consistent, showing that a lower body mass index (BMI) is associated with a higher rate of mortality, whereas higher BMI, even at levels considered obese and morbidly obese, is paradoxically associated with better survival in the end-stage renal disease (ESRD) population.^1–4 The inverse relationship between BMI and mortality is consistent and robust across different ethnicities after accounting for potential time-varying confounders and biases.⁵ Recently, a meta-analysis showed that for every 1 kg/m² higher BMI, there was a lower risk of all-cause and cardiovascular mortality by 3 and 4%, respectively, in patients undergoing hemodialysis (HD). Interestingly, a similar association between BMI and risk of death was not observed in patients on peritoneal dialysis (PD).⁶

The explanation for these observations is not immediately apparent but reflection suggests that being fatter may be accompanied by better nutritional reserve and may protect against wasting in the setting of acute illness or chronic inflammation, particularly over relatively short-term periods. Such protection, outweighing the harmful effects of obesity on cardiovascular outcomes in the longer term, would be expected to be magnified among dialysis patients, who also have more comorbid illness and shorter life expectancy.⁷ However, BMI measurement is not always an accurate estimate of adiposity because BMI does not distinguish between fat and lean compartments or between subcutaneous and abdominal fat accumulation.

Misclassification of BMI in dialysis patients

Although BMI is currently considered a risk stratification tool for under- and overnutrition in the general population because of its simplicity and ease of use,⁸ it is likely to perform worse in patients receiving dialysis than in the general population. Excess fat stores and high muscle mass can both contribute to high BMI. Therefore, reliance on BMI may misclassify people; for example, those with excess adiposity but low muscle mass may incorrectly be considered as non-obese. In fact, despite the high prevalence of obesity among patients with ESRD, evidence suggests that defining obesity on the basis of a BMI ≥30 kg/m² may actually substantially underestimate the problem because of the above described type of subclinical or “masked” obesity (i.e., a low BMI but with a high percent body fat).⁹ In a cohort of prevalent HD patients in the U.S., under-identification of obesity by BMI was as high as 31% compared to percent body fat as the standard reference.¹⁰

These data highlight the limitations of BMI as a reflection of body composition. Moreover, BMI does not capture body fat distribution between subcutaneous and central fat deposits; the latter are the more metabolically active adipose tissue and are more associated with inflammation, oxidative stress, and insulin resistance.^{11, 12} Lastly, extracellular volume expansion and fluid overload could yield falsely high BMIs in ESRD patients.¹³

Associations of body composition and survival

Based on these shortcomings, studies have gone beyond assessment of BMI to further characterize the impact of a more diverse range of body composition measures on clinical outcomes among patients on dialysis. Initial attempts to determine which component – fat or lean – was more strongly related to survival in dialysis patients were reported by Huang et al.¹⁴ In the analysis of participants in the hemodialysis (HEMO) study with a median follow up of 2.5 years, both higher triceps skin-fold thickness (TSF, a proxy of body fat) and greater mid-arm muscle circumference (MAMC, a surrogate for muscle mass) were independently associated with lower risk of mortality when both variables were combined in the same multivariable model. Higher TSF and MAMC were each significantly associated with lower hazard of death in the adjusted models with the hazard ratio (HR) of a 1-SD higher fat and muscle of 0.83 [95% confidence interval, CI 0.75–0.91] and 0.90 [0.83–0.97], respectively.

In another report, Marcelli et al.¹⁵ examined a larger sample size than in the previous study and used bioimpedance spectroscopy (BIS) rather than anthropometry to estimate body composition; they found a higher mortality among patients with low lean tissue index (LTI) or low fat tissue index (FTI) when compared with the reference group (10^th −90^th percentile). Nevertheless, the combination of high LTI and FTI was not associated with statistically significantly better survival (death HR 1.91; 95% CI 0.48–7.65), although the wide confidence interval limits confidence in this result.

Overall, currently available data suggest that there is a complex relationship between fat and muscle compartments in the maintenance of body homeostasis and that the association of high BMI with better survival among HD patients may be more complicated than a simple association with higher body fat.

Despite the high prevalence of obesity, muscle wasting is common in HD¹⁶ and PD patients.¹⁷ Excess energy intake concurrent with physical inactivity, low-grade inflammation, or insulin resistance, all frequent among ESRD patients, may result in “sarcopenic obesity” - loss of muscle mass in the setting of excess adiposity.^{18, 19} Although the term “sarcopenia” was originally defined as age-related loss of muscle mass, recent consensus from experts around the world recommended using this term for the presence of both decreased muscle mass and low muscle function (strength or physical performance).^20–22

The gold standards for muscle mass measurement are magnetic resonance imaging and computerized tomography. Since these are both is resource intensive and less likely to be performed in large patient cohorts, alternative methods of estimating muscle mass with acceptable accuracy are being used instead. Well-preserved muscle mass, as assessed by serum creatinine,²³ MAMC,²⁴ adductor pollicis muscle thickness,²⁵ and BIS,²⁶ has been previously recognized as a strong indicator for survival within the ESRD population.

New evidence suggests that it is muscle function not actual muscle mass that mediates the association between muscle mass and mortality.²⁷ In a prospective cohort using data from the United State Renal Data System, HD patients with BIS-derived low muscle mass by different indexing methods (height², percentage of body weight, body surface area, and BMI) had a higher risk of death in unadjusted analysis.²⁸ However, these associations were no longer statistically significant after adjustment for covariates. In contrast, functional limitations in muscle strength or gait speed were associated with mortality even after adjusting for confounders. These findings further underscore the importance of additional potential contributors in the interpretation of the associations of muscle mass and survival among patients undergoing hemodialysis. Muscle strength and size are not solely dependent on each other, and risk factors for the reduction of muscle mass may not completely overlap with those for the loss of muscle functionality.

Importance of body fat distribution

Recent data have also found that fat is not uniformly beneficial. Measures of fat distribution and central obesity such as waist circumference and waist-hip ratio maintain a direct association with mortality not only in the general population but also among dialysis patients. Waist circumference as a proxy of visceral adipose tissue was more closely related to markers of inflammation (e.g., interleukin-6 and C-reactive protein) than subcutaneous fat, a surrogate of nutritional status, in patients with HD.²⁹

Furthermore, the negative metabolic consequences of excess visceral fat are preserved, independently of BMI, in the dialysis population. In a prospective cohort of 537 ESRD patients, a strong association between visceral fat and cardiovascular mortality was confirmed. The prognostic power of waist circumference per 10-cm for all-cause (HR 1.23; 95% CI 1.02–1.47) and cardiovascular mortality (HR 1.37; 95% CI 1.09–1.73) remained statistically significant after adjustment for comorbid conditions and other traditional cardiovascular risk factors.³⁰ Another study in an Asian hemodialysis cohort³¹ also reported that central obesity using waist circumference ≥90 cm in men and ≥80 cm in women was predictive of higher risk of cardiovascular events (HR 4.91; 95% CI 1.30–18.9) and all-cause hospitalization (HR 1.83; 95% CI 1.10–3.10).

Changes in body weight and mortality risk

Incident dialysis patients may experience rapid weight loss after beginning dialysis, and recent studies have shown that changes in body weight are more strongly associated with mortality than measurement of BMI at a single time point. In a large, contemporary, incident HD cohort,³² patients’ post dialysis weight rapidly decreased and reached a nadir at the 5^th month of dialysis, with an average decline of 2% from baseline, whereas obese patients (defined as BMI ≥30 kg/m²) lost approximately 3.8% of their initial weight at end of the first year of dialysis and did not reach a nadir. An incrementally larger weight loss during the first several months of dialysis initiation was associated with higher mortality risk, with a HR [95% CI] of 1.14 [1.07–1.22] for patients with ≥ 6% weight loss and 1.08 [1.02–1.14] for 2–6% weight loss (compared with <2% weight loss).

These data call attention to rapid weight loss as a potential indication of clinical deterioration of health status or underlying protein-energy wasting in incident HD patients. However, one of the limitations is that potential reasons for weight change could not be identified due to the observational nature of the study, making confounding by intercurrent health status likely as one would expect more spontaneous weight loss among sicker dialysis patients. Assuming that weight changes were unintentional, weight loss (>1% of initial weight at baseline) was associated with a significantly higher rate of mortality, whereas weight gain (>1%) was associated with better survival compared with stable weight (±1%) in a study of European HD patients.³³

Interestingly, the associations of weight variation and death were attenuated after stratification by BMI categories. Neither the association of weight loss with mortality (HR 0.98; 95% CI 0.74–2.14) nor the association of weight gain with survival (HR 0.95; 95% CI 0.59–1.62) was statistically significant among obese individuals. The authors hypothesized that the association between short-term weight loss and mortality was less apparent in obese individuals because of larger fat stores to counter the concurrent wasting process.

Obesity management among patients on dialysis

Although observational studies have clearly demonstrated a survival advantage among obese dialysis patients relative to those with lower BMI, and that unintentional weight loss is associated with worse clinical outcomes, one should keep in mind that targeting higher BMI levels is not necessarily recommended. First, it is entirely possible that intentional weight loss could be associated with better outcomes than maintenance of obesity, particularly for very heavy individuals. Such a situation would be analogous to hypertension in the HD population, in which observational studies^{34, 35} (that are usually not adjusted for treatment) demonstrate better survival among patients with higher blood pressure, but interventional studies demonstrate benefit of blood pressure lowering.³⁶ However, few studies have examined outcomes following intentional weight loss in the ESRD population.

Second, weight loss could increase access to transplantation or improve outcomes after transplantation among morbidly obese dialysis patients.³⁷ Many centers have adopted strict policies with regard to patients above threshold BMI of 30–35 kg/m², thus limiting the access to transplantation for obese dialysis patients.³⁸ An analysis from the French Renal Epidemiology and Information Network showed that a BMI ≥31 kg/m² at the start of dialysis was associated with a lower likelihood of receiving a KT, and the likelihood decreased as BMI increased.³⁹

Although obese transplant recipients had a higher likelihood of delayed graft function and a slightly higher risk of graft loss compared with their non-obese counterparts, there was no significant difference in overall mortality.⁴⁰ Furthermore, patient survival was significantly better among all BMI categories, including obese and extremely obese groups of KT recipients, compared with dialysis patients who remained on the waiting list.⁴¹

When pre-transplant weight management cannot be achieved through diet restriction and exercise, surgical weight loss might be considered. Use of bariatric surgery in morbidly obese transplant candidates was reported to be a safe and effective bridge to KT.⁴² Using direct matching for BMI at the time of transplantation, as well as for age, race, and allograft type, obese ESRD patients who underwent laparoscopic sleeve gastrectomy (LSG) and whose pre-transplant BMI decreased from 41.5 ± 4.4 to 32.3 ± 2.9 kg/m² had similar post-transplant outcomes including peri-operative complications, allograft, and patient survival at 1 year compared with non-LSG recipients.⁴³

For dialysis patients with sarcopenic obesity, therapeutic intervention to preserve muscle mass and reduce excess body fat is an ultimate goal for improving outcomes.⁴⁴ Data from randomized controlled trials have demonstrated that intradialytic resistance exercise training can improve muscle volume and enhance muscle strength and physical performance among HD patients.^45–47 Use of resistance exercise combined with an anabolic steroid (nandrolone decanoate) increased muscle mass and decreased body fat among patients with ESRD.⁴⁷ Furthermore, an oral androgen, oxymetholone, had a statistically significant anabolic effect to increase amount of fat-free mass and handgrip strength. However, concerns about the drug’s liver toxicity,⁴⁸ suggest that intramuscular or transdermal androgens would be better choices for further studies in the ESRD population.

In conclusion, despite the association of higher BMI with better survival in HD patients, the notion of a protective effect of excess body fat remains somewhat controversial. An accurate definition of obesity beyond the measurement of BMI is needed. Although unintentional weight loss is associated with poorer outcomes at the beginning of dialysis, intentional reduction of weight among stable obese dialysis patients should be advised if only to facilitate access to transplantation and thereby improve overall survival. Finally, the focus should not simply be on a BMI measurement at a single time point, but on changes in body composition. Interventions to preserve muscle mass and function or reduce excess body adiposity over time, particularly visceral fat, may have potentially beneficial effects on important outcomes in the dialysis population.