The Factors Effective on Bone Mineral Density in Peritoneal Dialysis Patients

Bone mineral metabolism deteriorates gradually beginning from the early stages of chronic kidney disease (CKD). But, the KDIGO-2009 guideline could not provide high quality data regarding the biochemical tests related with bone mineral disorder in CKD. The aim of our study was to analyse the relationship between bone mineral densitometry and clinical and biochemical parameters in chronic peritoneal dialysis (PD) patients. Besides the demographic parameters, routine hematological and biochemical analysis results of PD patients followed up in our clinic were recorded. Bone mineral density (BMD) was measured at lumbar vertebrae and femur neck using DEXA machine. The mean lumbar Tand femur T-score were -1.03±1.20 (minimum:-3.73; maximum:+1.75) and -2.49±1.20 (minimum:-4.63; maximum:-0.51), respectively. Lumbar T score was significantly higher than femur T score (p<0.0001). Eight patients had BMD within normal limits; there was osteopenia in 16 and osteoporosis in 29 patients. While there was a negative correlation between femur T-score and age (r=-0.36, p=0.026), no correlation was detected between lumbar T-score and age (r=-0.17, p=0.21). With multivariate analysis of the factors related with femur T-score; age and body mass index (BMI) were the independent determinants while gender, parathyroid hormone levels and use of active vitamin D were not effective. Age was related negatively while BMI was related positively with BMD: BMD measurement at femur is more accurate than that at lumbar vertebrae in PD patients. BMD is low in most of the PD patients; and age and BMI are the major determina-


INTRODUCTION
Chronic kidney disease (CKD) is an important health problem for both the patients and the general population due to its high morbidity and mortality.The early and proper diagnosis of pathologies is important in those patients.These pathologies increase significantly when dialysis initiated.As it is well known; defects in bone structure and mineral metabolism occur in CKD beginning from the early stages.Measurement of bone mineral density (BMD) is a reliable predictor of future fracture in postmenopausal and senile osteoporosis in the general population.The decrease in bone mass results in increased friability.It is not possible to consider patients with stage 5 CKD similar to the general population.Since osteoporosis is only a part of a very complex metabolic bone disorders including secondary hyperparathyroidism, osteomalacia and adynamic bone disease.
According to KDIGO 2009 guidelines; routine measurement of BMD in patients with stage 3-5 CKD is not recommended; because it is not useful in determining fracture risk and identifying the type of bone mineral disorder.But the grade and quality of the recommendations are not strong enough (1).All of the studies leading to this comment are cross sectional with varying and indefinite results.Similarly, the biochemical tests for diagnosis of CKD related bone mineral disease are also not strong.Our aim was to study the relationship of BMD findings with clinical and biochemical parameters among peritoneal dialysis (PD) patients followed up in our unit.

MATERIAL AND METHODS
Chronic PD patients who gave informed consent and who did not have exclusion criteria among 69 patients followed up in our PD unit were included into this cross sectional study.Patients aged less than 18 and more than 80 years, those with dialysis duration less than three months and patients with malignancy were excluded.The demographic data including age, gender, weight, height, body mass index (BMI) were recorded as well as primary kidney disease, duration of CKD and dialysis, the modality of PD treatment, the drugs used (angiotensin converting enzyme inhibitors-ACEi-, angiotensin receptor blockers-ARB-, other antihypertensive drugs, statins, active vitamin D) were recorded.

Comorbidities
Hypertension, hyperlipidemia, diabetes mellitus and ischemic heart disease were recorded as comorbidities.Hypertension was accepted to be present in the presence of history of hypertension; or measurement of blood pressure more than 140/90 mmHg at two separate times.Diabetes mellitus was diagnosed in in the presence of fasting blood glucose level above 126 mg/dl; or blood glucose level above 200 mg/dl at any time or at the second hour of oral glucose tolerance test; and was recorded if the patient had history of diabetes mellitus.The diagnosis of hyperlipidemia was put according to the criteria of National Cholesterol Education Program Adult Treatment Panel (NCEP ATP) III; age and the other risk factors of the patients.Ischemic heart disease was diagnosed if the patient had history of acute coronary syndrome, coronary artery by-pass surgery or coronary interventions; typical symptoms of coronary artery disease (angina pectoris, angina equivalent), and positive findings on electrocardiography, echocardiography, stress tests or coronary angiography.

Measurement of BMD
The measurement was performed at L2-L4 anteroposterior lumbar vertebrae and femoral neck using Norland Cooper Surgical Inc.-USA dual-energy X-ray absorptiometry (DEXA) machine.Results were expressed as T scores.Patients were divided into three groups de-scribed by World Health Organization according to the T scores measured (2): Normal (Femur T score>-1); osteopenia (-2.5<T score<-1) and osteoporosis (T score<-2.5).

Statistical Analysis
The statistical analysis was carried out with Statistical Package for Social Sciences for Windows ver.15.0.All numerical variables were given as mean ± standard deviation.Two groups were compared with paired Student's t-test or Mann Whitney U tests when necessary.Chi-square test with Yates correction and Fisher's exact test were used for 2X2 contingency tables when appropriate for non-numerical data.Correlations between T scores and other numerical parameters were analyzed with Spearman's rho correlation test.Groups were compared with Student's t-test.Comparisons in the BMD groups and PTH groups were made by Kruskal Wallis-H analysis of variance because the distribution was abnormal.P values less than 0.05 were accepted as significant.Multivariate analyses of the parameters related to femur T score was carried out with linear regression model with "enter" method.

RESULTS
Fifty three patients (female/male: 30/23) were included in the study.The mean age in female and male patients was 50±15 years and 56±16 years, respectively.The demographic data and the basal laboratory findings are presented in  ance, calcium, phosphorus, ALP and PTH levels were similar in patients with normal BMD, osteopenia and osteoporosis.Patients with normal BMD were all female (p=0.011);and the groups with osteopenia and osteoporosis were similar regarding gender (p= 0.17) (Table 3).Femoral T scores were significantly higher in women (Table 4, Figure 2).
There was no correlation of BMD with the presence of DM, hypertension, hyperlipidemia and ischemic heart disease.Patients grouped according to their PTH levels as low (<150pg/ml, n: 6), normal (150-300pg/ml, n: 6) and high (>300pg/ml, n: 41) had similar lumbar (p=0.18) and femur (p=0.37)T scores.With multivariate analysis conducted with parameters related with femur T score; age and BMI were found to be the independent variables determining femur T score, while gender, PTH level and active vitamin D treatment did not affect T score (Table 5).BMD was negatively correlated with age and positively with BMI.

DISCUSSION
We detected in our study a statistically significant difference between lumbar and femur T scores in our patients (p<0.001).BMD measured by DEXA at lumbar vertebrae may be overestimating in both uremic and nonuremic population.This is thought to be due to the presence of osteophytes, aortic calcification and scoliosis (3,4).T scores at the femoral neck have been measured as lower than in the lumbar vertebrae in other studies also (5)(6)(7)(8)(9).Arici et al (10) and Pongchaiyakul et al (11) found difference between T scores measured at hip in uremic patients compared with the control group; while corresponding measurements at spine were similar.
When T score at femur was considered; eight of the patients (15%) had normal T scores while

Figure 2. The femur and lumbar T scores according to the BMD groups
phosphorus, ALP and PTH levels.Different osteopenia and osteoporosis prevalence rates have been reported in dialysis patients in the literature (Table 6).The higher osteoporosis prevalence detected in our study could not be explained by the mean age, dialysis duration or BMI; but the most striking difference was about PTH levels which was higher than the above mentioned studies (560±429 pg/ml vs. 214.4±306.6 and 298±301 pg/ml).
Although PTH levels were relatively high, there was no correlation between PTH levels and T scores.Subgroups with low, normal and high PTH levels had similar T scores at femur and lumbar vertebrae.Grzegorzewska et al (14) reported that elevated PTH level is important in prediction of BMD with the lowest values seen in patients with the highest PTH levels.But there are studies showing no relation (9,19,20) or negative correlation also (7,(15)(16)(17)(18).Beyond the speculations related to the geographical or racial differances, these variations may be associated to PTH level.Because it is well known that the effects of PTH on bone structure and turnover are long term effects.So it is difficult to make connection between PTH level and BMD in a cross sectional study.Moreover, PTH levels change continuously due to many factors including dialysis dose, medications, measurement methods, etc. Supporting this idea, PTH levels reported in the above mentioned studies have high level of standard deviation.
The strong correlation between age and BMD seen in the general population may be weakened in the dialysis population due to many factors affecting on BMD other than age.We detected a statistically significant negative correlation between age and T score measured at femur (r= -0.36, p= 0.026), while age was not correlated with the corresponding measurement at lumbar vertebrae.
Ersoy et al ( 9) reported a strong negative correlation with age; while Taal et al (7) found this relation only in females and Grzegorzewska et al ( 8) only with measurement at femur, similar to our study.The prevalence of osteopenia and osteoporosis were similar in male and female patients (p= 0.17); while patients with normal BMD value were all female (p=0.011).Female patients were found to have higher T scores at femur compared with male patients.But with multivariate analysis, the relation with gender disappeared, leaving age and BMI as the major determinants of T score measures at femur.This may be explained by lower mean age higher mean BMI detected in females.Ersoy et al (9) reported no correlation between gender and BMD, while Taal et al (7) detected negative correlation between female gender and BMD at hip in hemodialysis (HD) patients.Orlic et al (14) found BMD values at both femur and lumbar vertebrae lower in female HD patients.Negri et al (5) reported that the mean femur and lumbar T scores were similar in both gender in PD patients; but total bone mineral content was significantly lower in female patients.In a recent study conducted in HD patients, it was reported that the decrease in bone volumetric density detected by high resolution peripheric quantitative computerized tomography was less in males compared with females (21).Grzegorzewska et al (8) found no relation between BMD (both femur and lumbar vertebrae) and gender in patients with dialysis duration more than 20 months.We detected no correlation between BMD and weight, BMI, serum albumin levels.But with multivariate analysis; BMI together with age were the major independent variable for T core measured at femur.This finding is consistent with the literature showing positive correlation between BMD and BMI (7,9,22,23).

225
In conclusion, BMD measurements performed at femur are more reliable compared with measurements at lumbar vertebrae in PD patients.The majority of PD patients have low T scores; with age and BMI being the major determinants.

Table 1 .
The demographic and clinical data of the patients.

Table 2 .
Biochemical data and BMD values.

Table 4 .
The relationship between gender and BMD.

Table 3 .
The relationship between gender and BMD.

Table 5 .
Multivariate analysis results of factors affecting on T cores measure at femoral neck.