A major problem is the difficulty to compare with reference values as no references for very old people (> 75yrs) exist. Therefore this study compares two matched samples of elderly citizen home people with a mean age of 84.1 ± 5.1yrs in female V, 84.3 ± 5.0yrs in female NV, 80.5 ± 7.5yrs in male V, and 80.6 ± 7.3yrs male NV, which is far above the cut-offs of most reference values.
The macronutrient intake was in agreement with the Belgian and Dutch food consumption surveys [17, 18]. Only in male V, protein intake was within the recommendations, while for the other groups protein intake was too high. Mean carbohydrate intake was too low in the male and female NV compared to the RDI. Total fat intake met the RDI in all groups. The saturated fat intake was comparable and, except in the male V, significantly above the RDI in all groups. Mean daily energy intake was lower in V females compared to their NV counterparts and did not reach the recommended value of 7.8 MJ. Most studies found no differences in energy intake between V and NV [1–5]. The difference found between the female groups in this study can be due to several reasons. Female V had a lower protein and saturated fat intake and a higher carbohydrate intake compared to the female NV. The macronutrient profile in V females was closer to the recommendations compared to the NV females. However both groups had protein, saturated fat and carbohydrate intakes not in accordance with the RDI. Although only significant in the females, the V had a lower alcohol intake compared to the NV. The better macronutrient profile and a lower alcohol consumption may be attributed to the often cited "health consciousness" in vegetarians [1, 10]. Mean vitamin and mineral intakes met the RDI in both groups, which indicates a varied choice of micronutrient dense foods in both groups. A study of Brants et al.  indicated a favourable food choice of independently living elderly vegetarians preparing their own meals compared to institutionalized elderly vegetarians. Compared to the institutionalized subjects studied by Brants et al.  our subjects (V and NV) show similar dietary shortcomings. Although the fact that vitamin B12, folic acid, iron, and calcium are often described as critical in a vegetarian diet, the mean blood concentrations of these parameters were normal in both groups. The main explanation can be found in the fact that our subjects were lacto-ovo vegetarians. While estimated zinc intake was sufficient, zinc blood serum was below the reference value in both groups. This indicates that the RDI may be underestimated for elderly as a consequence of their poor absorption due to aging. The mean blood cholesterol concentration was below the 200 mg/dl upper limit in the NV group and above that limit in the V group. These higher blood cholesterol levels are comparable with those found in elderly in Flanders where 86% of the males and 91% of the females between 65 and 69 yrs have blood cholesterol levels above 200 mg/dl . Since the subjects in this study have a mean age over 80 years the values in the vegetarian subjects can be considered as normal. Moreover, several studies suggest that slightly higher cholesterol levels may even protect against infections and atherosclerosis [42–44]. According to these authors, values from 200 to 239 mg/dl are more appropriate in elderly (> 65 yrs) [42–44]. Comparable numbers of subjects not reaching blood reference values were found in both groups. Of note is that vitamin B12 deficiency is not always unique to vegetarians, although it is generally more difficult for vegetarians (especially vegans) to meet vitamin B12 requirements than it is for omnivores. The present results show a mean vitamin B12 status according to the reference values while a comparable number of subjects with atypical values were detected for the V and NV. Therefore a low vitamin B12 status is not likely to be related to neither the vegetarian nor the omnivorous diet. The present results confirm the advice of Elmadfa and Singer  that a regular monitoring of the vitamin B12 status in order to facilitate early detection of low vitamin B12 status and timely treatment before clinical manifestations can develop is required. These concerns should also be taken into account for the other blood parameters.
Mean BMI was around 25 and comparable between V and NV in both genders. This is in contrast with previous studies which showed vegetarians to have a lower BMI [7, 46, 47]. Appleby et al.  found a significant inverse association between dietary fibre and BMI, but this was not confirmed by our results. The lower reported saturated fat intake in female V as compared to the NV did not result in a significantly lower BMI or waist-hip ratio in female V. Except lower subscapular skinfolds in V females no anthropometric differences were found between the V and NV. This can be a consequence of working with small matched samples. The findings concerning the anthropometric characteristics indicate an adequate energy balance, also in the female V who reported an energy intake below the reference value. The waist-hip ratio was comparable between V and NV groups, and was below the upper reference value . Price et al.  and Seidell  showed that for persons aged > 75 yrs, waist-hip ratio should be used instead of BMI because of the positive relationship between mortality and abdominal adiposity. Therefore we can consider that the elderly V and NV in the present study are not at increased risk. Regarding the handgrip strength both groups scored below the reference value and no differences were found between the V and NV. The anthropometric and physical differences compared to the reference values (65 - 75 yrs) may be due to the higher mean age (> 80 yrs) of the studied V and NV groups.
Since in this study V and NV did not substantially differ with respect to dietary intake and nutritional status, one should be careful with the conclusion that only dietary factors of a vegetarian diet cause different morbidity and mortality risks among V and NV.
A limitation of the present study may be the fact that the compared senior citizens homes in this study are located in two different regions (Belgium and the Netherlands). This is mainly due to the fact that, to our knowledge, there is but one exclusively vegetarian senior citizens home in the Netherlands and none in the Dutch speaking region of Belgium. Nonetheless, the national food consumption surveys in both regions showed similar shortcomings typical for European diets regarding the macro- and micronutrient intake. Another drawback is the non-randomized sampling of the two populations studied. More extensive studies have to be undertaken in order to gain more knowledge about the value of vegetarian diets in elderly people.