We found that lower BMI was associated with higher mortality in a cohort of 109,605 maintenance HD patients. Obesity (>30 kg/m2) and that even morbid obesity (>40 kg/m2) was associated with greater survival compared to normal or low BMI ranges. The inverse association of higher BMI with survival was strongest in Blacks and weakest in Hispanic maintenance HD patients. Whereas survival was higher in most BMI categories for Hispanic patients compared to other racial/ethnic categories, the benefit associated with each 1-kg/m2 higher BMI appeared to be stronger in Blacks and also to some extent in non-Hispanic white patients. In accordance with previous findings, Black and Hispanic dialysis patients had lower mortality rates compared to non-Hispanic Whites in our cohort. These findings present an important picture of the survival paradox observed in maintenance HD and how this pattern differs across racial/ethnic groups.
Even though obesity is associated with deleterious outcomes in the general population, in maintenance HD patients the association between BMI and mortality is reversed, a phenomenon known as the “obesity paradox’ or “reverse epidemiology”.(13
) The concept of the “altered risk factor pattern” may seem counterintuitive, especially since obesity is an established risk factor for poor health outcomes in the general population.(31
) Even so, given the consistency of the observations, there may be aspects of these and other chronic disease populations that make them more resistant to poor outcomes at greater BMI levels. Suggested reasons for this effect have been published(33
), including a hemodynamic status that is more stable in obese individuals,(35
) a higher concentration of tumor necrosis factor alpha receptors in obesity,(36
) neurohormonal alterations of obesity,(37
) endotoxin-lipoprotein interaction,(38
) differences in bone and mineral and vitamin D,(24
) reverse causation,(40
) survival bias,(41
) time discrepancies among competitive risk factors (over- vs. under-nutrition),(41
) and the overwhelming effect of malnutrition inflammation complex on traditional cardiovascular risks.(42
) Since the majority of maintenance HD patients in the US die within 5 years of beginning dialysis treatment,(44
) the long-term effects of conventional risk factors on future mortality may be irrelevant in the face of their potential short-term benefits or by other risk factors intrinsic to dialysis populations, such as under-nutrition and inflammation. On this note, perhaps dialysis patients do not survive long enough to die of the consequences of over-nutrition, since they face more pressing mortality risks from other health conditions such as protein-energy wasting.(45
) Hence, obesity, by providing more nutritional reserve, may protect them against early death.(15
The incidence and prevalence of CKD and ESRD have been rising in the United States and in most other countries.(49
) Differences in incidence and prevalance of CKD by racial/ethnic lines have been present for at least the past 20 years.(50
) The annual ESRD incidence for Blacks and Hispanics in 2006 reached 1,010 and 520 per million population, respectively, which is 3.6 and 1.5 times greater than the incidence in non-Hispanic Whites.(49
) The prevalence of ESRD continues to be highest for Blacks and Hispanics, at 5,004 and 2,326 per million population, respectively, as compared to 1,194 per million among non-Hispanic Whites.(49
) Compared with non-dialyzed Medicare patients of the same age group, mortality of dialysis patients is 10–100 times greater, irrespective of race or ethnicity.(49
) Almost half of the deaths in maintenance HD patients are attributed to cardiovascular disease.(52
) Black and Hispanic dialysis patients have higher survival than their non-Hispanic White counterparts while in the general population without CKD the opposite is observed; and this so-called “CKD racial survival paradox” has existed over the past two decades.(51
) The higher dialysis survival of the minorities is rather consistent and independent of demographic or residency status or the modality of dialysis treatment (thrice-weekly maintenance HD vs. daily peritoneal dialysis), dialysis dose, or other factors related to dialysis treatment or techniques.(49
There are some potential explanations for the differences in life expectancy seen in different races among maintenance HD patients. In the U.S. general population, disparities in income, education, and health status have been implicated in the increased total mortality and shorter life expectancy of Blacks compared with non-Hispanic Whites.(57
) However, among dialysis patients Blacks have a substantially lower annual mortality rate (18%) than non-Hispanic Whites (28%).(52
) On the other hand, in a 3-year contemporary cohort of 15,859 HD patients, Blacks were the only racial group in whom a high serum LDL cholesterol level (>100 mg/ml) was associated with increased risk of cardiovascular mortality, whereas, paradoxically, LDL-hypercholesterolemia was protective in other races.(30
) Additionally, two recent studies by Wolf et al.(60
) and Kalantar-Zadeh et al.(24
) suggests that therapy with active vitamin D may potentially explain greater survival of Black maintenance HD patients. Additionally, Volkova et al. found racial differences in coronary heart disease and heart failure in incident dialysis patients.(61
) Our findings indicate that higher BMI is strongly associated with longer survival. This association may be due to increased food intake having a short-term salutary effect on survival, which may be particularly advantageous in populations with short-term life expectancies. Such an effect could in turn be due to micronutrient intake, attenuation of the malnutrition-inflammation complex syndrome, and/or neurophysiologic modulation of oxidative stress.
Despite the lesser likelihood of Hispanics developing many chronic diseases, they are nearly twice as likely to develop ESRD as non-Hispanic Whites.(62
) This striking ‘paradox-within-a-paradox’ might relate to the increased incidence and prevalence of diabetes mellitus in this group. Hispanics who are at greatest risk of developing CKD may have a genetic background and ‘thrifty genotype’ in common with Native American Indians, who as a group are also at high risk of developing diabetes mellitus and CKD.(62
) Even so, when diabetic Hispanics receive maintenance dialysis, they have a greater likelihood of surviving than non-Hispanic Whites;(66
) which is still another example of a paradox-within-a-paradox. In our study, we found that Blacks have a lower mortality than non-Hispanic Whites for higher BMI groups while Hispanics have a consistently lower mortality than Non-Hispanic Whites for almost every BMI groups.
There may be another potential explanation for the stronger survival advantage of higher BMI in Blacks compared to other races. BMI per se may not be an appropriate measure to characterize nutritional status, body composition, obesity or muscle mass in dialysis patients.(18
) Blacks at any given BMI may have a greater muscle mass than other races/ethnicities. Indeed density of lean body mass is higher in Blacks than in Whites;(70
) and more muscle mass appears to have an association with better survival.(23
) Moreover, in Blacks without CKD higher BMI (>25 kg/m2
) tends to associate with only slight increase in risks of death.(71
) Hence, in Black maintenance HD patients, higher muscles mass could be responsible for the observed association between BMI and survival.
Our study is limited by its observational nature. Like all observational studies, this analysis cannot distinguish direct causation from confounding associations (here, association of BMI with other factors related to survival). It is also important to note the limitation of BMI as a measure of “obesity per se”, and the potential that higher BMI may be a marker of better nutrition/higher muscle mass/edema as well as higher visceral adiposity. Another potential limitation is that we excluded more than 17000 patients with outlier BMI; however, this is unlikely to lead to substantial bias, because excluded patients had similar demographic and laboratory values compared to individuals in the study (see Table S1
). Additionally, race and ethnicity categories were based on self-identified data. Our study did not have detailed serum markers of inflammation, oxidative stress, or oxidized lipids, which could have provided additional evidence to reinforce proposed mechanistic pathways for our findings of powerful associations. A notable strength of our study is its large sample size and the availability of laboratory values, which allowed us to account for important covariates in the multivariate analyses and to examine specific associations, thus providing points of focus in future studies. To the best of our knowledge our study is the first to compare the association between BMI and survival among Blacks, Non-Hispanic Whites and Hispanics in maintenance HD patients.
In this study we found that obesity, as measured by BMI, was associated with greater survival in maintenance HD patients of all races and ethnicity, but different patterns were observed across the three mutually exclusive racial and ethnic groups. The survival advantage of obesity was most prominent among Blacks. Our analyses suggest a protective but differential role for higher BMI across racial and ethnic groups of maintenance HD patients.