Anthropometric, hormonal and biochemical characteristics of the study subjects
The study groups were age-matched. Patients with AN had severely decreased BMI and body fat content (Table
1). At baseline, patients with AN had significantly decreased lean body mass and daily resting energy expenditure as compared with control group. Fasting serum concentrations of leptin, glucose, free fatty acids, and total serum protein were significantly reduced in AN group as compared with C group. Fasting LDL cholesterol levels in untreated AN patients did not significantly differ from C group.
Table 1
Characteristics of the study groups.
Age (years) | 22.9 ± 0.50 | 24.9 ± 1.34 | 24.9 ± 1.34 |
Body mass index (kg/m2) | 21.8 ± 0.38 | 15.7 ± 0.30* | 17.5 ± 0.23*+ |
Body fat content (bodystat) (%) | 21.9 ± 1.21 | 9.3 ± 1.65* | 12.8 ± 1.38*+ |
Body fat content (DEXA) (%) | 26.4 ± 1.79 | 13.9 ± 1.44* | 18.0 ± 1.38*+ |
Lean body mass (kg) | 52.5 ± 2.03 | 38.9 ± 1.3* | 41.3 ± 0.96* |
Resting energy expenditure (kcal/day) | 1400.1 ± 35.92 | 1060.0 ± 40.25* | 1176.3 ± 25.05*+ |
Fasting glucose (mmol/l) | 4.4 ± 0.07 | 3.9 ± 0.09* | 4.2 ± 0.08+ |
Total cholesterol (mmol/l) | 4.2 ± 0.14 | 4.6 ± 0.26 | 5.2 ± 0.27*+ |
HDL-cholesterol (mmol/l) | 1.6 ± 0.08 | 1.5 ± 0.10 | 1.6 ± 0.08 |
LDL-cholesterol (mmol/l) | 2.1 ± 0.09 | 2.4 ± 0.18 | 3.0 ± 0.21*+ |
Free fatty acids (mmol/l) | 0.64 ± 0.06 | 0.16 ± 0.02* | Not measured |
Total serum protein (g/l) | 79.1 ± 1.49 | 68.7 ± 2.01* | 75.5 ± 1.04+ |
Albumin (g/l) | 46.4 ± 1.22 | 43.8 ± 1.12 | 47.6 ± 0.72+ |
Total bone mineral density (g/cm2) | 1.09 ± 0.02 | 1.04 ± 0.02* | Not measured |
Lumbar (L1-L4) bone mineral density (g/cm2) | 1.03 ± 0.03 | 0.86 ± 0.03* | Not measured |
Proximal femur bone mineral density (g/cm2) | 0.98 ± 0.02 | 0.78 ± 0.04* | Not measured |
IGF-1 (ng/ml)
|
NA
| 156 ± 18 |
NA
|
MIC-1 (pg/ml)
|
1051 ± 83
|
1872 ± 200*
|
1530 ± 169*
+
|
Leptin (ng/ml) | 11.8 ± 1.78 | 0.97 ± 0.23* | 2.6 ± 0.66*+ |
Fasting serum HDL-cholesterol, total cholesterol and albumin concentrations did not significantly differ between the groups. At baseline, lumbar, proximal femur and total bone mineral density were significantly reduced in AN group relative to C group (Table
1).
The influence of partial realimentation on anthropometric, hormonal, and biochemical characteristics of patients with AN is summarized in Table
1. Patients with AN significantly gained weight and body fat content during realimentation. Fasting serum leptin, total serum cholesterol, glucose, total protein, albumin, LDL-cholesterol concentrations and daily resting energy expenditure significantly increased in patients with AN after partial refeeding as compared with pre-treatment values. Post-treatment fasting serum HDL-cholesterol concentrations were not significantly different as compared with its pre-treatment levels in AN. Post-treatment concentrations of total cholesterol and LDL-cholesterol were significantly increased in patients with AN as compared with controls (Table
1).
Relationship of MIC-1 with other studied parameters
The relationship of serum MIC-1 levels with other parameters was assessed in a combined population of both groups and separately in patients with AN (Tables
2 and
3). In a combined population (Table
2), serum pre-treatment MIC-1 concentrations were correlated with other pre-treatment parameters by Spearman correlation test. The correlations were normalized for BMI. In a combined group of AN and controls pre-treatment MIC-1 positively correlated with age and serum total cholesterol, and was inversely related to BMI, lean body mass, serum glucose, serum albumin, serum leptin, lumbar bone mineral density and resting energy expenditure (Table
2).
Table 2
Relationships of MIC-1 with other parameters in combined population.
MIC1 | r | -0.52 | -0.59 | -0.36 | 0.47 | -0.49 | -0.53 | -0.43 | -0.64 | -0.57 |
| p | 0.000 | 0.001 | 0.02 | 0.002 | 0.008 | 0.003 | 0.04 | 0.000 | 0.000 |
Table 3
Relationships of MIC-1 with other parameters in patients with anorexia nervosa.
Pre-treatment MIC1 | r | -0.60 | - 0.65 | NS | -0.65 | -0.55 | NS | NS | 0.53 |
| p | 0.01 | 0.02 | | 0.007 | 0.03 | | | 0.03 |
Post-treatment MIC1 | r | -0.68 | - 0.60 | -0.62 | -0.52 | -0.70 | -0.55 | 0.53 | 0.51 |
| p | 0.004 | 0.04 | 0.01 | 0.04 | 0.002 | 0.03 | 0.03 | 0.04 |
In patients with AN, univariate correlations were calculated by Spearman correlation test for pre-treatment MIC-1 levels with other pre- and post-treatment parameters (Table
3) and post-treatment MIC-1 levels with other pre- and post-treatment parameters (Table
3), respectively. Both pre- and post-treatment serum MIC-1 levels were inversely related to post-treatment Buzby nutritional risk index, pre-treatment serum insulin-like growth factor concentrations, post-treatment total protein and post-treatment albumin levels. Both pre- and post-treatment MIC-1 levels positively correlated with age. Post-treatment serum MIC-1 levels significantly inversely correlated with post-treatment serum glucose and pre-treatment lumbar bone mineral density and were significantly positively associated with the duration of illness in patients with AN (Table
3). In both groups, we failed to find any significant relationship of MIC-1 with body fat content, serum thyroid hormones (fT3, fT4, TSH), serum free fatty acids, and liver enzymes (aspartate aminotransferase, alanine aminotransferase).
In patients with AN, delta MIC-1 during treatment significantly inversely correlated with delta LBM (r = - 0.70, p = 0.002) during treatment.
Multiple regression analysis was performed in combined population with pre-treatment MIC-1 levels as dependent and pre-treatment BMI, lean body mass, glucose, cholesterol, albumin, lumbar bone mineral density, leptin, resting energy expenditure and age as independent variables. None of the factors included was identified as statistically significant independent predictor of pre-treatment MIC-1 concentrations.
Multiple regression analyses were also performed separately in anorexia nervosa group with pre-treatment MIC-1 and post-treatment MIC-1 levels as dependent parameters, respectively. First multiple regression analysis was performed with pre-treatment MIC-1 levels as dependent and pre- and post-treatment Buzby nutritional index, pre-treatment IGF-1, lumbar bone mineral density, age, duration of illness and post-treatment glucose, total protein and albumin as independent variables. Similarly as in the combined group of AN and control subjects none of the factors included was identified as statistically significant independent predictor of pre-treatment MIC-1 concentrations.
Another multiple regression analysis was performed for post-treatment MIC-levels as dependent and pre- and post-treatment Buzby nutritional index, pre-treatment IGF-1, lumbar bone mineral density, age, duration of illness and post-treatment glucose, total protein and albumin as independent variables. Similarly as in previous cases none of the factors included was identified as statistically significant independent predictor of pre-treatment MIC-1 concentrations.
Finally, to assess a possible role of MIC-1 levels in the regulation bone mineral density two multiple regression analyses were performed. In the first one, pre-treatment bone mineral density was set as dependent variable and age, post-treatment MIC-1, post-treatment blood glucose, post-treatment BMI, post-treatment IGF-1, post-treatment albumin and post-treatment Buzby nutritional index as independent variables. In the second one, pre-treatment bone mineral density was set as dependent variable and age, pre-treatment MIC-1, pre-treatment blood glucose, pre-treatment BMI, pre-treatment IGF-1, pre-treatment albumin and pre-treatment Buzby nutritional index as independent variables. None of the factors included was identified as statistically significant independent predictor of pre-treatment bone mineral density.