Portal de Periódicos da CAPES

You’re Not Big—You’re Just Tall, That’s All!

2015; American Society of Nephrology; Volume: 27; Issue: 2 Linguagem: Inglês

10.1681/asn.2015070816

1533-3450

John T. Daugirdas,

Cardiovascular Function and Risk Factors

The title of this editorial comes from a blues song written by Jimmy Reed, titled "Big Boss Man," the refrain being "Big boss man… you ain't so big—you're just tall, that's all." I'm 6'3" (190 cm) tall, so I was happy to read the recent New England Journal of Medicine article by Nelson et al. identifying a lower coronary artery disease risk in tall people.1 The case for tallness being beneficial in that report was strengthened by the methodologic approach used to investigate the role of height; height marked by genotype was the independent variable, limiting interference from well known confounders of tallness such as nutrition, race, age, and socioeconomic status. However, somewhat dismaying for tall people, two recent studies have been published showing that tallness in dialysis patients is associated with a negative effect on survival. A study by Elsayed et al. in this issue of JASN looked at US patients who began dialysis between 1995 and 2008.2 After patients with incomplete data were excluded, slightly more than 1 million patients made up the final study cohort. Height, measured at the time of beginning dialysis treatments, was divided into sex-specific quintiles. Death information was obtained from the US Centers for Medicare and Medicaid Services death notification form and the Social Security Administration Death Master File. Two-year mortality rates were then calculated for each race and sex subgroup, and the usual adjustments and interactions were examined. What did the results show? When adjusted for age, tallness was associated with adverse 2-year survival in men, such that the death rate in the maximum height quintile was about 4%–5% greater than death in the lowest quintile (272 versus 260 deaths per 1000 person-years, respectively). In women, there was no relationship when the entire patient population was considered. There also seemed to be a race-related effect, because there were opposite effects in black versus white patients. In whites, tallness was a risk factor for death, and the risk increased somewhat monotonically across quintiles 2–5 in both men and women. In blacks, increased height was associated with lower death risk, and the increased death risk was somewhat more pronounced in the lowest height quintiles of both sexes. When further multiple adjustments were made, the risk of increased height seemed to be similar for both men and women, with the association being absent in blacks and pronounced in whites and Asians. In both sexes, the percentage of patients with identified Hispanic ethnicity was markedly concentrated in the lower quintiles. As expected, weight increased monotonically with height quintile in both sexes, but body mass index (BMI; calculated as kilograms divided by height in meters squared) was relatively similar across height quintiles 2–5 for both men and women, being somewhat increased in the lowest height quintiles. Overall, the relationship between height and outcome was similar in patients being treated with hemodialysis and peritoneal dialysis, and the association was not diminished by adjusting for a variety of treatment-related factors or vascular access. No information was presented with regard to cause-specific mortality. In a recent CJASN article that also examined height-associated mortality risk, Shapiro et al. reported a prospective analysis of patients receiving hemodialysis at one large dialysis organization (DaVita) from mid-2001 through mid-2006.3 In that study, cause-specific mortality was available, and height-associated mortality risk was determined after adjusting for weight only, for weight plus age, and then additionally for case mix and then also for laboratory values. The sample size was 117,000, height was expressed in terms of race-adjusted deciles (separately for men and women), and Hispanic identification was treated as a race; thus, race was defined as white, Asian, black, Hispanic, or other. There was little effect of height decile on mortality in analyses adjusted only for weight, but a positive association appeared when an age adjustment was added. Adjusting the analysis for race and sex seemed to make little difference overall, but the association between height and mortality seemed to be slightly magnified by adjusting for case mix; addition of laboratory values to the adjustment had little additional effect. The investigators also looked at patient weight and found the well known, inverse association between weight and mortality; surprisingly, they found that the abilities (receiver-operating characteristic [ROC]) to predict mortality of case mix–adjusted height and weight were similar. In contrast with the results of Elsayed et al. (which reported on a data set that presumably included many, if not most, of the DaVita patients), the association between height and mortality was similar in black and white patients in the study by Shapiro et al. The authors of the two studies compare their results with results in non-CKD populations, and they point out that associations between height and mortality in this broader population also depend to an extent on the statistical adjustments made; a positive effect of height on survival found in some early studies was no longer seen or was even reversed after adjustment for age and social status.2,3 What do these results mean, and do they provide us any useful information in either better treating dialysis patients or in understanding body size–related adjustments commonly performed when looking at either nutrition-related outcomes or dialysis adequacy? One clinical question regularly encountered in the treatment of dialysis patients concerns what to do with a patient who is moderately to markedly obese. In patients with normal kidney function, obesity has been associated with the metabolic syndrome, risk of diabetes, and increased risk of cardiovascular disease and even CKD.4 In assessing obesity risk, BMI is commonly used, with >30 kg/m2 being indicative of increased risk (BMI cutoffs associated with increased risk can be lower in Asians). In dialysis patients, it has been much more difficult to show a risk associated with increased BMI, and in fact, obese dialysis patients are at a relatively low risk of death, while mortality risk increases exponentially in the lowest BMI subgroups.5 The association between higher BMI and good outcomes in patients on dialysis (or with other chronic illness) has been explained primarily in terms of weight. This situation is not unusual when considering patients with chronic illness, because a high BMI implies a certain degree of resistance to that illness; in such patients, body mass is maintained by way of good nutrition in the face of the disease process. By contrast, a low BMI in patients with chronic illness is usually associated with poor appetite linked to poor health status. Attention has been paid to fine-tuning the associations between BMI and outcome, because the weight term of BMI can be high as a result of either increased muscle mass (associated with a good outcome in all populations) or increased body fat. In the normal population, the adverse risk of high BMI is thought to be attributable primarily to increased body fat, and especially central obesity, and measurements to more specifically measure this risk include use of the waist-to-hip circumference ratio or more sophisticated imaging methods to determine body fat, especially visceral fat.6 In dialysis patients, the survival benefit of high BMI has been linked to higher muscle mass but high fat mass appears to be protective as well.6 In analyses of BMI versus outcome in dialysis patients, up until now, there has been little thought given to the denominator of BMI, height squared. Because the height term is in the denominator, increased height will mathematically reduce BMI for any given level of body weight. To what extent height may explain associations between BMI and outcome in dialysis patients has not been examined. Anthropometrically derived estimates of total body water, a surrogate for lean body mass, are positively related to both height and weight, and anthropometric estimates of body water are strongly positively associated with survival.7 For this reason, it would not be expected that the inverse association between height and survival shown in the studies by Elsayed, Shapiro, and their colleagues could alone explain the well known positive association between BMI and survival, and it would seem that the weight term of BMI, a surrogate for increased muscle mass (and possibly, increased fat mass) would be primary. However, detailed analysis of the effect of the height and weight terms of various measures of body size has yet to be done. The ROC data presented by Shapiro et al. suggest that height also needs to be taken into consideration. Another issue where height and weight come into play is in terms of scaling of the dialysis dose. In the study by Shapiro et al., average Kt/V was calculated in each height decile for both men and women. There was a slight inverse association between height and delivered hemodialysis Kt/V (1.62 in the lowest decile, 1.43 in the highest decile), leading the authors to speculate that perhaps some of this height association with mortality might have been driven by dialysis adequacy differences and/or by an intrinsic need for more dialysis in taller patients as a result of increased susceptibility to uremic toxins. In the study by Elsayed et al.,2 Kt/V data were not available, but dialysis time was. Data presented in the authors' supplemental tables showed that the height–mortality association was present in subgroups that were being dialyzed either for 3 hours per session or 4 hours per session; the adverse association of height and mortality appeared to be magnified in men dialyzed 3 hours per session versus men dialyzed for 4 hours.2 To better clarify whether differences in dialysis dose scaled either to volume or body surface area might potentially explain the observed relationship between height and mortality, it would have been useful to have dialysis adequacy data in terms of either Kt/V or Kt normalized to body surface area in the study by Elsayed et al. The persistence of the association between height and mortality in patients treated by peritoneal dialysis dampens somewhat the enthusiasm for such a potential mechanistic explanation. We cannot change the height of a given patient, and if taller patients are, for some reason, at risk, it is not clear how to mitigate this, because risks in the study by Shapiro et al. seemed to be increased to similar degrees for cardiovascular, infectious, and cancer-associated mortality. These two articles are intriguing, in that they remind us that both height and weight need to be considered when analyzing outcomes by measures of body size that include both of these terms such as BMI, total body water, or body surface area. At this point, it is not at all clear what the physiology of increased mortality risk associated with taller stature might be, nor how this new knowledge might affect clinical practice. Disclosures None.