Quick Links

Recommendations Summary

CKD: Nutrition Assessment: Body Mass Index (BMI) as a Predictor of Mortality (2020)

Click here to see the explanation of recommendation ratings (Strong, Fair, Weak, Consensus, Insufficient Evidence) and labels (Imperative or Conditional). To see more detail on the evidence from which the following recommendations were drawn, use the hyperlinks in the Supporting Evidence Section below.

  • Recommendation(s)

    CKD: Body Mass Index (BMI) as a Predictor of Mortality, Maintenance Hemodialysis (MHD)

    In adults with CKD 5D on MHD, we suggest that overweight/obese status (based on BMI) can be used as a predictor of lower mortality, whereas, underweight status and morbid obesity (based on BMI) can be used as a predictor of higher mortality (2B).

    Rating: Fair

    CKD: Body Mass Index (BMI) as a Predictor of Mortality, Peritoneal Dialysis (PD)

    In adults with CKD 5D on PD, we suggest that underweight status (based on BMI) can be used as a predictor of higher mortality (2C).

    Rating: Weak

    CKD: Body Mass Index (BMI) as a Predictor of Mortality, Non-Dialyzed

    In adults with CKD 1-5, it is reasonable to consider using underweight status (based on BMI) as a predictor of higher mortality, though the mortality risk associated with overweight or obesity status (based on BMI) is not clear (OPINION).

    Rating: Consensus

    CKD: Body Mass Index (BMI) as a Predictor of Mortality, Post-Transplant

    In posttransplantation adults,  it is reasonable to consider using underweight and overweight/obesity status (based on BMI) as a predictor of higher mortality (OPINION).

    Rating: Consensus

    • Risks/Harms of Implementing This Recommendation

      There are no obvious risks or harms associated with these recommendations.

    • Conditions of Application

      • BMI is not an ideal marker of obesity, since it cannot differentiate between higher weights due to increased adiposity vs. muscularity and it cannot identify visceral adiposity, which has negative metabolic effects. 
      • To ensure accuracy of BMI, height should be measured periodically.
      • The standard weight status categories associated with BMI ranges for adults should be used in the CKD population; these include <18.5 kg/m2 for underweight; 18.5 to 24.9 kg/m2 for normal weight; 25.0 to 29.9 kg/m2 for overweight; and ≥30 kg/m2 for obese.
      • Limited evidence suggested that obesity (BMI ≥30 kg/m2) may be a risk factor for higher mortality in individuals who are on dialysis and under the age of 65. Therefore, practitioners should consider patient age when determining mortality risk according to BMI.
      • In patients on dialysis, weight to calculate BMI should be measured following dialysis treatment to improve accuracy.

    • Potential Costs Associated with Application

      There are no obvious costs associated with these recommendations. 

    • Recommendation Narrative

      Methods of assessing body composition, including anthropometric measurements, are components of the nutrition assessment in CKD.  Anthropometric measurements are practical, inexpensive and non-invasive techniques that describe body mass, size, shape, and levels of fatness and leanness; they are the most basic and indirect methods of assessing body composition. Regardless of the method selected to assess body composition, none are perfect, and the errors surrounding them should not be ignored. Errors may have clinical relevance, especially if the individual is treated and observed over time (Chumlea et al 2008). Moreover, the results of the measures are only as useful as the availability of suitable reference data from a group of persons of at least the same age, race, gender and disease status.

      Twenty-four studies reported on the use of BMI to assess nutritional status, including 17 prediction studies (Molnar et al 2001, de Roij van Zuijewijn et al 2015, Araujo et al 2006, Kim et al 2014, Kalantar-Zadeh et al 2005, Yen et al 2010, Chatzot et al 2009, Wiesholzer et al 2003, Leavey et al 2001, Mathew et al 2015, Badve et al 2014, Hoogeveen et al 2012 and Lievense et al 2012, Leinig et al 2001 and McDonald et al 2003) and nine correlation studies (Atif et al 2013, Beberashvile et al 2009, Bross et al 2010, Kadiri et al 2011, Kahraman et l 2007, Leinig et al 2008, Nakao et al 2007, Steiber et al 2007, Visser et al 1999). There were no studies examining validity or reliability of using BMI in this population to classify nutritional status.

      Maintenance Hemodialysis Patients 
      Eight studies examined MHD patients only. Seven studies examined mortality risk according to BMI category. In three studies (Chatzot et al., Wiesholzer et al and Yen et al),  the authors examined mortality risk according to traditional weight categories (underweight, normal weight, overweight and obese), although in a study with Taiwanese participants (Yen et al),  these categories were defined differently. In five additional studies, the authors examined risk according to 5 to 11 BMI categories (Kalantar-Zadeh et al, Leavey et al. and Badve et al, Ricks et al 2011, Doshi et al 2016).

      In one study that only compared two groups (< or >25 kg/m2), the authors found no association between BMI and mortality at 10 years (Araujo e al 2006).  However, in the remaining studies in which BMI was examined according to traditional weight status groups or by 5 to 11 categories, there was consistently a higher risk of mortality for participants who were underweight, and lower risk for participants who were overweight or obese (Kalantar-Zadeh et al 2005, Yen et al 2010, Chatzot et al 2009, Wiesholzer et al 2003, Leavey et al 2001 and Badve et al 2014, Ricks et al 2011, Doshi et al 2016). Length of follow-up for these studies ranged from 1.34 to 10 years. There was an inverse relationship with mortality when BMI was measured as a continuous variable in three studies (Molnar et al 2010, Wiesholzer et al 2003 and Leavey et al 2001),  but Harell’s C statistic was not significant in de Roij van Zuijewijn, et al.

      Findings from correlation studies indicated that BMI was positively associated with albumin levels, fat and lean body mass measured by a variety of methods in HD patients. Beberashvili et al showed that serum albumin was significantly and positively correlated with BMI and FM in MHD patients (Beberashvili et al 2009). The higher BMI group had greater LBM (p=0.001) and FM (p=0.0001), and higher phase angle (PA) and ECM/BCM (p<0.05). MHD patients with elevated BMI demonstrate better nutritional status compared to normal BMI or overweight patients. Severity of inflammation was not related to BMI in MHD patients.

      Bross et al indicated that BMI had a strong linear correlation with total body fat percentage measured by NIR and BIA (Segal) (r ≥ 0.85) in MHD patient. Fat tissue index, as estimated by BIA, was significantly correlated with BMI in the study by Aatif et al. In another study, Kadiri et al showed that BMI was positively correlated with FM (r=0.493, p=0.002), serum albumin (r=0.340, p=0.04), and anemia in MHD patients. BMI was negatively correlated with CRP (r=-0.065, p=0.702) but had no correlation with LBM (r=0.278, p=0.085). Kahraman et al 2007 studied the relationship between CRP and BMI status and found that CRP levels were significantly higher in obese and underweight MHD patients compared with normal and overweight patients (p<0.05).

      Steiber, et al. found that mean BMI was significantly different across the 5 categories of SGA (p<0.05) in MHD patients. Visser et al demonstrated that there was a strong correlation between the 7-point SGA scale and BMI in MHD patients (r=0.79, p<0.001), % fat (r=0.77, p<0.001).

      Maintenance Hemodialysis and Peritoneal Dialysis Patients 
      Three studies reported on the relationship between BMI and mortality in a combination of MHD and PD patients (Badve, et al. reported results for MHD and PD patients separately). In Mathew et al 2015, participants who survived had higher baseline BMIs compared to the group that did not survive, but BMI category was not a significant predictor. Hoogeveen et al 2012 demonstrated that underweight and obesity were risk factors in a combination of MHD/PD patients less than 65 years of age, but for those who were at least 65, there was no relationship between BMI and mortality. Lievense et al demonstrated that PD patients had lower mortality risk compared to MHD patients.

      Leinig et al showed that there was a positive correlation between BMI and FM in predialysis (r=0.67, p=0.0002), in MHD (r= 0.67, p=0.0002), and peritoneal dialysis (r=0.79, p<0.0001) patients. Nakao et al indicated that BMI was significantly correlated with BPI in MHD and PD patients (r values ranging from 0.778 to 0.886, p<0.0001). Hoogeveen et al 2012 followed dialysis patients < or ≥65 years of age for seven years. In the multivariable adjusted model, compared to those with “normal” weight status, those who were categorized as underweight (2.00 (1.30-3.07) and obese (1.57 (1.08-2.28) had a significantly higher hazard of mortality for those who were <65 years, but there was no significant relationship between weight status and mortality for those ≥65 years of age. 

      Peritoneal Dialysis Patients 
      Four studies reported on the relationship between BMI and mortality in PD patients. Badve et al. found that underweight increased mortality risk at 2.3 years, but results regarding higher BMI categories were not consistent. Leinig et al. found no difference in mortality risk according to whether PD patients had a BMI < or >23 kg/m2 at 2 years. McDonald et al. found that, in adjusted analysis, PD patients who were obese had higher risk of mortality (up to 10 years) compared to patients with normal weight status. In the study by Kim et al., the group with the lowest quartile of BMI had the highest mortality risk at 2 years, but there were no other significant associations. In a systematic review performed by Ahmadi et al 2016, authors confirmed an increased risk of 1 year mortality for people with CKD who were underweight, but this relationship did not persist for 2, 3 and 5 year mortality. Conversely, Ahmadi et al 2016 found that overweight/obesity status decreased mortality risk at 1, but not 2, 3 or 5 years (Ahmadi et al 2016).  

      Non-dialyzed Patients 
      Finally, two studies examined the relationship between BMI and mortality in non-dialyzed CKD patients. Madero, et al. examined risk according to BMI quartile and found no relationship. Hanks, et al. took a different approach and examined risk not only according to traditional BMI categories, but also according to whether participants were metabolically healthy. Of those who were metabolically healthy, there was decreased risk for overweight/obese participants compared to those with a normal BMI. However, there was no difference in mortality risk according to weight status in those who were metabolically unhealthy. These findings were consistent with a systematic reveiw by Ahmadi et al 2015.

      Post-transplant Patients
      A systematic review by Ahmadi et al 2014 examined the relationship between BMI and mortality in 150, 000+ adults with CKD with kidney transplant. Authors conclude that, compared to participants with “normal” weight status at baseline, those who were underweight [HR (95% CI): 1.09 (1.02, 1.20)] or overweight/obese [1.20 (1.14, 1.23)] were at increased hazard of mortality.

    • Recommendation Strength Rationale

      The evidence supporting these recommendations is based on Grade II/Grade B, Grade III/Grade C evidence as well as Consensus/expert opinion. 

    • Minority Opinions

      Consensus reached.