Health hazards associated with obesity
Large epidemiological studies have demonstrated an association between obesity and mortality. In a subset of individuals aged 50 years who had never smoked, and were followed for 10 years, there was a two- to threefold increase in mortality for those with a BMI > 30 kg/m2.18 Obesity is strongly linked to Type 2 diabetes, hypertension, CVD, some cancers and arthritis, which each contribute to the increase in mortality. The mechanism for this relationship may be related to insulin resistance and hyperinsulinaemia, with subsequent increases in impaired glucose tolerance, increased sympathetic activity, renal sodium retention and vascular tone. In spite of increased use of risk-modifying therapies such as lipid-lowering drugs and antihypertensives, there is no evidence of a reduction in the population risk associated with obesity over time.19
Cardiorespiratory fitness may modify this risk.20–22 A prospective observational study of 25 714 predominantly Caucasian men22 demonstrated that low fitness was common in obese men and an independent predictor of cardiovascular and all-cause mortality and increased the relative risk of mortality to a similar degree as does diabetes. A second important finding in this study was that for each risk factor studied (i.e. diabetes, high cholesterol, hypertension, current smoker and low cardiorespiratory fitness), increasing adiposity was associated with a stepwise increase in mortality risk. That is, for every risk factor examined, the presence of obesity increased the risk.
In the Australian population,23 more than 75% of obese males and 65% of obese females had at least one comorbidity (hypertension, dyslipidaemia or diabetes) and 7–10% had all three. The AusDiab 2005 report demonstrated that compared with those with a normal BMI at baseline, the overweight and obese have a 2- to 4-fold increase in the annual incidence of diabetes and hypertension (see Table 1). For example, the annual incidence of hypertension in obese patients was 5% and for diabetes it was 1.6%. These data are derived from a 5-year follow-up study24 and further information is required to determine the relationship between baseline BMI and the incidence of hypertension and diabetes over time. However, this is of particular relevance to living kidney donors in whom the average age at nephrectomy is 48 years25 and who have a life expectancy of many more decades.
The impact of obesity on risk of diabetes and hypertension is even more pronounced in Aboriginal Australians. Compared with the AusDiab population, the OR (95% CI) for diabetes among normal, overweight and obese (by waist circumference) remote living aboriginal women were 2.6 (06–11.5), 13.1 (6.7–25.7) and 6.1 (4.6–8.0), respectively.8 The risk for diabetes in aboriginal men was 6-fold higher in each of the weight categories. Similar increased prevalence of obesity, diabetes, hypertension and cardiovascular risk were also described in a cohort of urban indigenous people from Perth.26 The adjusted relative risk for the incidence of newly diagnosed diabetes in an 8-year follow-up study was 3- to 4 fold higher for BMI > 25 kg/m2 compared with those with a lean BMI.11 In summary, indigenous Australians have a significantly increased risk of diabetes, hypertension, cardiovascular and kidney disease, which is further magnified even at low levels of adiposity.
In New Zealand, the prevalence of obesity is increased in Maori and Pacific Islander peoples compared with the Caucasian population (BMI ≥ 31 kg/m2 63%, 69% and 26%, respectively).27 Similarly, the prevalence of diabetes is a least 3-fold higher in the Maori and Pacific Islanders and occurs at a younger age (typically between 5 and 10 years younger than Caucasians).28 The relationship between fasting insulin and BMI was independent of ethnicity, suggesting that the high prevalence of diabetes was related to obesity. Hypertension is also increased in the Maori and Pacific Islander population29 and in a large church-based survey, BMI was positively associated with blood pressure (BP), with a 14 mmHg difference in systolic BP between the lowest and highest quartile of BMI in men and 9 mmHg in women.30
At any given level of obesity, the absolute risk of diabetes is consistently higher in Asians, for both men and women. For example, for men with a BMI of 24 kg/m2, the prevalence of diabetes was 5% in Asians versus 2% in Caucasians, and the prevalence of hypertension was 32% in Asians and 17% in Caucasians.31 There is a continuous positive association between baseline BMI and risk of future diabetes, which is stronger in Asians than Caucasian cohorts.32 In the Nurses Health study,33 for each 5-unit increase in BMI, the adjusted relative risk of incident diabetes in Asians was 2.36 (95% CI: 1.83–3.04) and for Caucasians was 1.96 (95% CI: 1.93–2.00). The impact of weight gain from baseline was also a significant factor; in Asians, each 5 kg weight gain was associated with an increase in risk of incident diabetes by 84% (95% CI: 58–114) and 37% (95% CI: 35–38) in Caucasians.
Obesity as a risk factor for CKD
There are several mechanisms by which obesity may be expected to have a detrimental effect on the kidney. Obesity increases single-nephron glomerular filtration rate (GFR), increases activation of the sympathetic nervous and renin-angiotensin systems, promotes salt resorption in the proximal tubule34 and has been associated with specific histological changes including glomerulomegaly and focal segmental sclerosing lesions.35 Obesity is associated with and often precedes multiple factors associated with development of kidney dysfunction – hypertension, diabetes and atherosclerosis but data from longitudinal cohort studies suggest that obesity may also be an independent risk factor for the development of CKD and ESKD36–41 (see Table 2).
Analysis of the Kaiser Permanente cohort40
demonstrated that there is a progressive increase in risk of ESKD associated with obesity, independent of age, gender, race, smoking, previous myocardial infarct, baseline cholesterol, proteinuria and serum creatinine. Compared with normal BMI, the adjusted relative risk for ESKD was 1.87 for overweight and 3.57 for BMI between 30 and 34.9 kg/m2
and 6.12 for BMI between 34 and 39.9 kg/m2
and 7.07 for BMI > 40 kg/m2
. Adjustment for baseline BP and presence of diabetes attenuated the risk slightly but the associations remained strong.
Preliminary data from the AusDiab follow-up study, on the annual incidence of both de novo eGFR < 60 mL/min per 1.73 m2 and albuminuria, demonstrates a stepped increase in risk for each weight category. For the obese, the risk approaches 1% per year and is roughly increased by 50% above the risk of the normal weight participants (personal communication, Steven Chadban). The increase in risk is independent of diabetes.
The Framingham Offspring study39,42
involving 2676 participants with normal baseline kidney function who were followed for a median of 18.5 years, demonstrated an increased risk of stage 3 CKD in obese participants but in a subsequent publication, this was accounted for by the higher prevalence of diabetes and hypertension. There was a 56% increase in risk of proteinuria (>1+ dipstick) in the obese participants, which was independent of hypertension and diabetes.
Pooled data from two community-based longitudinal studies (Atherosclerosis Risk in Communities and the Cardiovascular Health Study with a median follow up of 9.3 years,43
demonstrated that the relationship between obesity and incident CKD was best determined by WHR rather than BMI. Univariate analysis demonstrated a strong relationship between WHR and CKD, but this was no longer significant after adjustment for multiple risk factors. Of note, this study involved an older population (mean baseline age 57 years) in whom most of the comorbidities that associate with obesity, would likely have developed by the time of enrolment. WHR was also associated with an increased risk of the composite end-point of mortality and CKD, whereas increased BMI was protective of this outcome.
The Physicians Health Study38
demonstrated that in healthy men followed for 14 years, higher baseline BMI was associated with an increased risk for CKD following adjustment for potential confounders. In addition, men who had a greater than 10% weight gain, also substantially increased their risk of CKD. This study was flawed by the possible inclusion of men with CKD at baseline.
Analysis of the Hypertension Detection and Follow-up Program37
cohort demonstrated that in the presence of hypertension, overweight and obesity significantly increased the risk of CKD compared with normal BMI. In patients who were not diabetic at baseline, the adjusted OR for CKD at 5 years in the overweight group was 1.22 and for the obese group it was 1.38. This study was conducted in the 1970s and it is not known if lower BP targets and newer antihypertensive agents attenuate the risk. CKD was defined as proteinuria 1+ on dipstick and/or eGFR < 60 mL/min.
A Swedish case control study36
showed that overweight BMI at age 20, was associated with a 3-fold increase in the risk of CKD and was independent of diabetes and hypertension. Patients were stratified for cause of kidney disease and although the influence of overweight at 20 years was greatest for diabetic nephropathy (5-fold increase), overweight patients also had a 3-fold increase in risk of nephrosclerosis and glomerulonephritis.
Other earlier studies have not demonstrated an independent association between obesity and subsequent ESKD. The study by Perry et al.44
involved a high-risk group cohort of hypertensive veterans, of whom 45% died during follow up and the degree to which CKD contributed to their death was not examined. Iseki et al.45
found an association between BMI and ESKD in men that was explained by diabetes.
It is important to note that while there is a fairly consistent increase in relative risk between obesity and kidney disease, the absolute risk of ESKD for an individual is small. Using the Kaiser Permanente population as an example, the adjusted rate of ESKD is 10 per 100 000 person years for normal BMI and 46 per 100 000 person years for BMI 30–34.9 kg/m2. In terms that patients are more likely to comprehend, this equates to a risk of ESKD over 10 years of 1 in every 1000 normal BMI patients, compared with 4.6 in every 1000 obese patients.
The associations between obesity and incident CKD, are to a variable degree dependent on the associated comorbidities of hypertension and diabetes. This is of relevance when assessing donors who have been carefully screened for these risk factors, and the risk associated with obesity in the absence of these is likely to be small. However, these assurances can't be assumed in the young donor in whom they may not yet be manifest.
Obesity may be a greater risk factor for loss of GFR in patients who already have impaired kidney function. This is analogous to the greater impact of hypertension in causing progressive disease in patients with CKD when compared with those with normal kidney function. There are some data (n = 162) to suggest that obesity promotes more rapid loss of renal function in patients with IgA nephropathy.46 Patients who were overweight had heavier proteinuria at time of biopsy, were more likely to be hypertensive, have more severe tubulointerstitial changes on biopsy and to subsequently develop hypertension and renal impairment.
Gestational diabetes: a systematic review47 demonstrated that gestational diabetes is associated with a 17–63% increase in risk of Type 2 diabetes within 5–16 years of pregnancy. The highest risk occurs in the first 5 years after pregnancy and then appears to plateau. BMI > 30 kg/m2 was identified to further increase risk associated with gestational diabetes in most but not all studies.
Renal cell carcinoma (RCC): although RCC only accounts for 2.8% of cancers in Australia (Cancer in Australia, 2001), it is of particular relevance to potential donors. A systematic review48 of 22 small studies demonstrated an increase in the relative risk of RCC of 1.07 (95% CI: 1.05–1.09) per unit increase in BMI and the risk was equivalent in men and women. Therefore, the relative risk for patients with a BMI of 30 kg/m2 is 1.35. Subsequent large cohort studies have been consistent with this finding49,50 although others have failed to find an association between obesity and RCC in men.51,52 There is a biologically plausible link between obesity and RCC as increasing BMI is associated with elevated levels of fasting serum insulin-like growth factor,53 which has been shown to increase cellular proliferation in RCC in animal models.
Kidney stones: analysis of data from the Nurse's Health Study I and II and the Health Professionals Follow-up Study54,55 demonstrated that prevalence and incidence of new stone disease was directly associated with BMI, with a stronger relationship evident in women. The age-adjusted prevalence OR for women with a BMI ≥ 32 kg/m2 compared with 21–22.9 kg/m2 was 1.76 (95% CI: 1.50–2.07), and 1.38 (1.51–2.36) for the same analysis in men. For incident stone formation in women, the OR was 1.89 (1.51–2.36) in women, but not significantly different in men.
The acceptance of obese donors is increasing
Increases in rates of donor obesity have occurred over the past decade and demonstrate regional variation. In a survey of UK transplant centres published in 1999,56 only one centre was identified as accepting patients with a BMI greater than 30 kg/m2 or a weight greater than 20% above ideal. Results of a survey of US centres, published in 1995, reported that only 16% of centres would exclude a donor with moderate obesity.57 The same centres were surveyed in 200558 and the following cut-offs were described – 10% of centres would exclude a donor if BMI > 30 kg/m2, 52% if BMI > 35 kg/m2 and 20% if BMI > 40 kg/m2. Subsequent publications59,60 from the US demonstrate that, in some centres, 20–30% of donors have a BMI > 30 kg/m2 and data from the Organ Procurement and Transplantation Network/United Network for Organ Sharing (OPTN/UNOS) registry suggest that from 7/2004 to 12/2005, 13% of US donors had a BMI > 30 kg/m2.
There are data to suggest that acceptance of obese donors is also increasing in Australia.61 Preliminary data from the ANZ live donor registry presented in 2007 at the ANZSN ASM, suggest that 16% of donors from 2004–2006 had a BMI of between 30 and 35 kg/m2 and 2.3% had a BMI > 35 kg/m2.
Short term risks associated with obesity
Retrospective analysis of a US healthcare registry62 using discharge data for 3074 patients from 28 centres identified comorbidities and complications using ICD-9-CM coding data. Obesity was associated with an increased risk of overall complication rate (OR 1.92, 95% CI 1.06–3.46), however, numbers were too small to assess the impact of obesity on the incidence of major complications, and the study was not able to discriminate between open and laparoscopic nephrectomy. Similar results have been reported from a number of single centre studies, demonstrating an increase in minor complications in obese donors for both open and laparoscopic nephrectomy (see Table 3).59,63,64 Complications are predominantly wound related and include wound infection, seroma and hernias. The rates of wound infection approach 10% in the obese compared with 2% in normal weight donors. Operative time is longer in obese patients – with increases ranging from 10 to 41 min, but no increase in length of hospital stay is reported.59,63,65,66 Nor is there any reported increase in delayed graft function in the recipient. Numbers are small and results relating to conversion from laparoscopic to open procedure are mixed, with some studies reporting no difference59,67 and others66 reporting increased conversion in obese men. They also commented that the perinephric distribution of fat in obese men increased the technical difficulty. There is a consistent pattern of greater blood loss and increased transfusion requirements in obese patients, which is not significant in each of the single centre studies due to small numbers.63,66–69 In addition, laparoscopic donor nephrectomy has been a relatively new technique and there may have been an increased complication rate in the more technically challenging obese patients as part of the learning curve. Rhabdomyolysis is a rare complication of donor nephrectomy. Sporadic case reports of rhabdomyolysis in donors are characterized by the following risk factors – long operative time, laparoscopic procedure and high BMI.68,70–73
Does nephrectomy increase the risk of future renal disease in obese patients? Studies in which baseline BMI has been reported are outlined in Table 4.
There are limited data from which to address this issue. The often quoted follow-up studies of living donors are limited by several significant methodological flaws, studies are retrospective, predominantly Caucasian and the rate of loss to follow up is high. Baseline BMI has not been reported in many of the older studies and obese patients are almost certainly under-represented in the long term follow-up statistics used to educate prospective donors regarding the risks of nephrectomy.
Studies reporting baseline characteristics of obese donors suggest that they are at higher risk of future kidney disease.59,63,74,75 A study from a centre59 with a high use of obese donors, in which 31% of donors had a BMI > 30 kg/m2 gives a detailed analysis of the baseline characteristics of obese donors. Obese donors had a significantly higher pre-nephrectomy BP (137/79 vs 126/73 mmHg), increased history of donor hypertension (14% vs 4%), more adverse lipid profiles, higher fasting glucose levels (although within the normal range) and had a family history of diabetes (47% vs 33%), when compared with donors with a BMI < 25 kg/m2. Data are available at 1 year for approximately 60% of donors in this study, and demonstrates that BP and fasting glucose remained higher, albeit in the acceptable range, and did not incrementally increase post nephrectomy. The post-nephrectomy GFR and rates of microalbuminuria were not different in the obese, within this short timeframe.
Donors who are overweight or obese are more likely to gain weight post donation than those of normal weight.76
There is a probable relationship between BMI and subsequent hypertension.74,76–78 Obese patients are more likely to have higher BP at the time of donation and it is unknown if nephrectomy alters the age of onset or severity of hypertension. A German study of 152 donors, with 93% followed for a mean of 11 years and with pre-nephrectomy BMI of 26 ± 4 kg/m2, demonstrated that baseline BMI was correlated with mean arterial pressure but not change in BP post donation.78
There is no evidence of association between the baseline BMI and development of proteinuria or decline in GFR post donation in predominantly Caucasian populations.78,79 However, the number of donors who were obese at baseline is too small to be able to determine this with any certainty. The study from the Mayo Clinic79 had long-term follow up on 73% of donors with a median follow up of 12 years. Only data on weight are available and is not differentiated for gender. Median weight at donation was 70 kg and weight gain at follow up was 7.5 kg. Baseline weight, change in weight and relative weight (measured/ideal weight) was not a significant predictor of current serum creatinine or change in creatinine. The flaws are use of creatinine rather than GFR and the number of patients who were obese at baseline is unknown.
The most detailed study, examining the impact of obesity is from Pakistan.80 The study was large (736 patients) with a mean follow up of 3 years (range: 6 months to 18 years). At last follow up, 11.5% of patients were obese and obesity was more common in women (17% vs 6%). Obese donors, when compared with the non-obese donors, had significantly higher rates of diabetes (13.5% vs 3%) and hypertension (24% vs 10%). There was a non-significant trend to lower GFR (<60 mL/min) and a higher prevalence of proteinuria in obese donors. This data are concerning and the median follow-up time is short. There is limited detail given in terms of screening donors for diabetes, or presence of family history for diabetes and baseline BMI. There are cultural reasons cited for the high rate of weight gain post donation, and the population studied is one that is ethnically more at risk of developing diabetes.
This study highlights that the safety data drawn from predominantly Caucasian populations, do not necessarily hold true for populations with a greater risk of diabetes and/or kidney disease. A report from the OPTN/UNOS registry81 records 102 individuals as waiting for transplant who have previously been living donors, in which African Americans are over-represented. There is no information on the prevalence of obesity in the group or other identifiable risk factors that may have been present at donation, however, hypertension and diabetes are listed as the cause of ESKD in roughly one third.
The histology of implantation biopsies in obese living donors is subtly different from non-obese donors.82 Increased glomerular planar surface area (GPSA), glomerulomegaly and minor tubular abnormalities are more common in obese donors and there is a trend to increased arterial hyalinosis. There was no difference in the number of segmental sclerotic lesions or degree of interstitial fibrosis. GPSA was correlated with albuminuria, although all donors had 24 h urinary albumins that were within the normal range. Donor follow up was less than 1 year and no difference in serum creatinine was seen between obese and non-obese donors.
A retrospective analysis of 73 patients examined the outcome of unilateral nephrectomy done for clinical indication (i.e. not donors).83 At the time of nephrectomy, patients had normal creatinine and urinalysis, no multisystem disease such as diabetes and no morphological abnormality of the remaining kidney examined by ultrasound. Median follow up was 13.6 years (range: 18 months to 35 years). Twenty of 73 patients developed abnormalities of renal function (proteinuria ± renal insufficiency). Average time to proteinuria was 10 ± 6 years and was slowly progressive in most patients. Thirteen of 73 patients developed renal impairment (serum creatinine > 1.4 mg/dL and creatinine clearance < 70 mL/min per 1.73 m2). Time between development of proteinuria and onset of renal impairment was 4.1 ± 4.3 years. Renal impairment was progressive and two patients were biopsied with histology revealing glomerulomegaly and focal segmental glomerulosclerosis. At time of nephrectomy, BP, age and renal function were similar between those that did and did not develop CKD. There were, however, significant differences in BMI at the time of nephrectomy (BMI 24.9 kg/m2 in normal function group, compared with 33.7 kg/m2 in the abnormal renal function group). BMI was independently associated with proteinuria/renal dysfunction on multivariate analysis (OR 1.34, 95% CI: 1.03–1.76). At 10 years following nephrectomy, the probability of negative proteinuria and normal renal function was 40% and 70%, respectively, in the obese group and 93% and 98%, respectively, for the non-obese patients. It is important not to overinterpret this study, which is retrospective, has small numbers, is subject to ascertainment bias and involved patients who may have had undiagnosed abnormalities of the remaining kidney. However, it does raise some uncertainty about the long-term safety of nephrectomy in obese donors.