There has been a rapid increase in the number of elderly people in Pakistan 1 hence maintaining health and well-being of this age group is becoming even more important. Beside so many other health risks associated with old age, this population is potentially the most vulnerable group for malnutrition 2 . Poor dentition, neuropsychological problems and immobility in older age directly affect their nutritional status 3 .

The prevalence of overweight and obesity is increasing 4 , particularly in the elderly 5 , where it is associated with increased mortality and a number of metabolic and cardiac disorders 6 . Overweight and obesity also contributes to functional decline and disability in the elderly 7 . At the same time, quite significant numbers of old individuals are reported to suffer from underweight and are at higher risk for acute illness and death 8 . They also have significantly higher risk of dying within the first year of hospitalization than those with adequate nutrition 9 . Weight loss has been shown to be associated with a higher risk of disability 10 . Decreased body Mass Index (BMI) is an indicator of chronic energy deficiency and malnutrition, and is associated with compromised immune function, increased susceptibility to infectious illnesses, and reduced survival in the elderly 6 .

Similar to other developing countries, Pakistan can be expected to experience the impact of an increasingly ageing population over the next few decades 1 , with a steady rise in the average life expectancy from 59.1 years in 1991 to 65 years in 2002. This quite sudden demographic shift can be very challenging in terms of health and nutritional care. Essential information about individuals' food intake and habits, activity, cultural influences, and the economic and social situation provide a database for nutritional assessment. Developed countries have established dedicated health care systems in order to meet the special needs of the elderly. However, such programs are lacking in developing countries like Pakistan. To the best of our knowledge, so far no separate study has been undertaken to document the nutritional status of the elderly in Pakistan and this type of important information thus remains fragmentary or absent. Those nutritional surveys that have been conducted in the past, however, do show very marginal nutritional status and high nutrient deficiencies in the general population (not specifically the aged) 1 . In this context of higher prevalence of malnutrition in general population in Pakistan, it can be assumed that the elderly might have an even more impaired nutritional status. The present study, therefore, aimed to investigate the nutritional status and nutrient intake of Pakistani elderly. The results are expected to help in designing policies and making plans regarding health care provision for the elderly in Pakistan. Nutritional status is particularly worrisome in the context of the ageing population, which is becoming a serious demographic problem. Hence, elucidating the nutritional status of the elderly is of prime importance for formulating preventive strategies to lower morbidity rates, improve quality of life and reduce health care costs.

The present study included only apparently healthy individuals with no recent past or present smoking or any other drug addiction history. Table 1 shows general and socio-demographic characteristics of the study subjects. Table 1 also shows % number of elderly in four BMI categories and mean (SD) % BF of elderly in these BMI categories. As evident, more than half (51%) of study subjects were illiterate and relatively a high number (82%) were living with their families. Based on BMI, there were 13.1, 3.1, and 10.8% obese, overweight and underweight elderly, respectively. The mean (SD) % BF ranged from 15.5 (6.41) to 38.4(7.21), respectively in the underweight and obese elderly.

Table 2 shows % number of overweight and obese elderly defined by BMI, WC and WHR. Most of the overweight and/or obese elderly defined by any of these three criteria were in the age group of 60.1 - 70 yr. Based on BMI, WC and WHR, 8.6, 4.9, and 29.2% elderly were either overweight or obese in this age category; the highest as compared to other age categories. The other age category with the second highest percent prevalence of obesity and/or overweight was 70.1-80 yr. The prevalence of WHR-defined obesity was the highest (23.2%) in the age group 60.1 - 70 yr. Furthermore, in all age groups WHR gave the highest prevalence of obesity followed by BMI- and WC-defined obesity. These results show that either BMI or WC alone may underestimate the prevalence of obesity in elderly and, therefore, WHR may be a stronger and more sensitive indicator for estimation of obesity and/or overweight in epidemiological studies. These results further show that in elderly central or abdominal obesity (assessed by WC or WHR) may be more prevalent than general obesity (assessed by BMI).

Table 3 presents the mean daily intake of selected nutrients by elderly stratified by BMI groups. There were large differences in nutrient intake comparing all the three groups (i.e. obese, overweight and underweight) to the normal weight group. Obese and overweight elderly seemed to be consuming significantly (p < 0.0001) more energy than people of normal weight but significantly less protein, calcium, iron, vitamins A and C. Further, the results show that underweight elderly had significantly lower mean intake of all nutrients studied as compared to the normal weight elderly (p value ranged from 0.0001 - 0.0006).

The % number of elderly with adequate nutrient intake in each BMI category is depicted in Figure 1. Overall, very few elderly had adequate energy and protein intake. In obese and overweight categories, 100 and 84% of the elderly had adequate energy intake, while very few people in those two categories had adequate protein intake. Similarly, in the normal weight and underweight BMI categories, adequate energy and protein intake were reported for 64 and 22, and 47 and 17%, respectively. Similarly, for minerals and vitamins, even lesser than 45% of the elderly in obese, overweight and underweight categories had an adequate intake of Ca, Fe, Zn, vitamin A and vitamin C. As expected, the percentage of normal weight elderly with adequate intake for these nutrients was higher than either of the other BMI categories.

One encouraging fact was that the participation rate in this study was fairly high (73.6%). Because subjects in poor health are often not able and also not willing to participate, selectivity in favor of subjects in better health can hardly be avoided in studies involving the elderly. The same holds true for poorly-educated persons 19 .

The nutritional assessment of free-living elderly in Pakistan in the present study has demonstrated the need to promote a healthy lifestyle in this population. BMI, WC, WHR, and % BF measurements showed that most of the elderly people had abnormal nutritional status with very high energy intake in the obese category and inadequately lower energy intake in the rest of the BMI categories. The need for the elderly to improve their nutritional status and balance their dietary intake has been a long-standing topic of discussion among nutritionists. Many studies have associated higher energy intake with obesity and overweight and lower energy intake with body decomposition, which may result in a decreased DNA repair capability, lower plasma glucose levels, diminished insulin sensitivity and overall unhealthy lifespan 6 19 .

In current study, all the anthropometric variables were included on the basis of their association with food habits, health and well-being in the elderly 20 . Weight reflects the recent and present balance between energy utilization 21 . Height/stature reflects genetic potential and nutritional status during growth and is also related to fat-free or lean body mass, which is a good index of protein stores 22 . BMI calculated from weight and height 23 is related to percentage of body fat and to fat-free mass, while WC and HC are useful indices of adipose tissue and central obesity 24 .

The present study highlights an alarmingly high prevalence of overweight, obese and underweight even in relatively healthy and wealthy Pakistani elderly men, measured either by BMI, WC or WHR. In particular, very high numbers (43.6%) of elderly were found to be either overweight or obese assessed by WHR (Table 2), which is especially important in view of the fact that Asian adults have higher cardiovascular risk factors already at lower BMI and WC than Western populations 16 . These arguments may support the fact that alone BMI is not enough to determine the risk of developing obesity-related conditions. Excess abdominal fat, regardless of overall body fat, will predispose to obesity-related disease. This highlights the importance of measuring WHR. It is possible that two persons with very similar BMI may vary substantially in the proportion of abdominal fat. Accordingly, a person with a BMI in the "normal" weight range may exceed the safe range of abdominal fat. In aged individuals with a decline in lean muscle mass, their BMI may not change or may even decrease, but fat levels could increase with the accompanying redistribution of body fat. WHR and WC are useful and reliable measures of abdominal obesity but both of them have their individual strengths and weaknesses and both are usually measured in a clinical evaluation.

In addition, BMI has also been criticized for its poor discrimination between fat and muscle mass. Thus, those individuals who are overweight not because of an increased amount of body fat, may have a high BMI value, but should not be considered obese. There are data indicating that even though BMI is a reliable measure of fatness in children and young individuals 25 , an adolescent's percentage of fat can change by as much as -3 to +7% without any difference in BMI. For an individual adult, the same BMI can correspond to changes in fat of ±5% 26 . Additionally, BMI seems to have a reduced applicability to the elderly 27 . For this very reason, WC and WHR are used for better discrimination of obesity, particularly the central or abdominal obesity 24 26 28 . However, all these anthropometric measurements have certain limitations 29 and therefore, cannot be used in isolation to predict results.

Data on nutritional status of elderly is also very fragmentary in Pakistan. Other studies documenting the prevalence of obesity and overweight in the elderly seem essentially absent. There has been no nationwide study to document the prevalence of obesity in the other population groups either. Some small-scale local studies, however, reported variable rates of overweight and obesity in Pakistan 30 . Higher prevalence of obesity and/or overweight in Pakistani population with increasing age has also been reported previously 30 31 . The results of these studies are in close agreement with ours, finding the highest mean measurements of BMI, WC and WHR in the elderly age group of 60.1-70 yr. The difference in prevalence as reported by the current and the previous studies might be mainly due to difference of age of the sample, sample size and sample characteristics.

In current study, we found fewer elderly had adequate nutrient intakes (Figure 1). Energy intake seemed to be adequate (66.7-100% of the recommended intake) in 100, 84 and 64%, respectively of obese, overweight and normal weight elderly, but only in 22% of the underweight elderly. The overall number of elderly individuals with adequate energy intake was 67.5%, which means more than 33% were energy-deficient and had inadequate (<66.7% of the recommended intake) energy intake.

The prevalence of energy deficiency in Pakistan is not unexpected 32 , particularly in the elderly 33 . If BMI < 18.5 kg/m² is used as an indicator of chronic energy deficiency in the elderly 34 , prevalence of chronic energy deficiency as high as 13.1% is reported in the current study. Low BMI values in relation to low energy intake in Asian elderly populations have also been reported in the IUNS Study 35 . Even in developed countries, data show a high prevalence of energy deficiency in the elderly 36 . Lower energy intake causes body decomposition 18 . On the other hand, due to problems with mastication and poor dentition 33 37 , elderly prefer caloric-dense foods with proportionally limited amounts of other necessary nutrients, which might be a contributing factor to age-related obesity and deficient intake of other important nutrients.

In current study, protein intake in all four BMI categories seemed to be inadequate (Table 2). Only very few elderly had adequate (66.7-100% of the recommendation) protein intake in the four BMI categories (Figure 1A): 25, 21, 47, and 17% of the obese, overweight, normal weight and underweight elderly, respectively, with an overall of 27.5%, had adequate intake. This implies that a large proportion (72.5%) of the elderly had inadequate (<66.7% of the recommendation) protein intake. Requirements for protein in the elderly are still under debate 31 ; but it is quite safe to say that there was a high risk of protein deficiency in our study group of the elderly.

The % number of elderly in the four BMI categories with adequate Ca, Fe, Zn (Figure 1B) and vitamin A and vitamin C (Figure 1C) intake ranged from 21 - 58% for Ca; 31 - 61% for Fe; 25 - 69% for Zn; 13 - 59% for vitamin A and 28 - 82% for vitamin C. However, the overall numbers of elderly with adequate intake of these nutrients were only 37, 43, 41, 30, and 47%, respectively. To the best of our knowledge, there have been no separate data on the intake of these nutrients by Pakistani elderly. However it has been reported that mean intake of Ca, Fe and Zn by adults in the general Pakistani population is much lower than the recommendations 38 . Mean calcium, iron, and zinc intake in the present study seemed well within the intake range of most countries 39 . However, the % number of subjects with adequate intake of these nutrients was very low.

It is also noteworthy that most nutrients consumed by the elderly in the present study were derived from plant sources (data not shown). This intake pattern is similar to that in many other developing countries 40 , which may be one of the reasons for deficiencies in certain nutrients in this age group. For example, phytates present in whole-grain breads, cereals, legumes and other plant foods bind zinc and inhibit its absorption 41 . Factors found mainly in plant foods including phosphorus, flavonoids, oxalates and soy protein can also inhibit iron absorption and decrease its bioavailability 42 .

The correlation analyses (Figure 2) show that with increasing age there was a significant decrease in BMI (p = 0.0028; r = -0.1304). Energy (p = 0.0564; r = -0.1236) and protein intake (p = 0.0776; r = -0.0771) tended to decrease with age but not significantly, while a non-significant increase in WC (p = 0.3124; r = 0.0422) and significant increase in % BF (p = <0.0001; r = 0.3655) with age were noted. Unlike WC, WHR decreased with age. However, this decrease was not significant statistically (p = 0.1220; r = -0.0675). Studies show a decrease in BMI with age, particularly after 60 yr 43 44 , an increase in fat mass 45 and a decrease in energy intake 36 . However, these changes are very variable 43 44 45 . Nevertheless, all these associations of selected anthropometric measurements and nutrients with age are important from the aging and nutrition point of view as an understanding of the underlying factors affecting body composition may facilitate correction by simple nutritional interventions. An increase in body fat with aging may be partly attributed to a loss in muscle mass, even in independently-living healthy subjects 27 . Furthermore, skeletal muscle mass loss in men is masked by weight stability, resulting from a corresponding increase in total body fat mass. Progression of sarcopenia, particularly in men, may therefore be clinically silent and comparable to the loss of bone mineral density in osteoporosis 27 .

In conclusion, there is a high prevalence of underweight, overweight and obesity in elderly Pakistani men. We report a limitation of prediction made either by BMI, WC or WHR alone as a measure of overweight and obesity, based on our results and the published literature. The nutritional data demonstrated that majority of subjects had a suboptimal nutrient intake. We propose that the current BMI-based categories be reviewed for the Pakistani population, particularly for the elderly. Furthermore, we suggest that BMI, WC and WHR should be used in combination to define nutritional status. In addition, we suggest that attention should also be paid to the problem of underweight in old age.

The authors declare that they have no competing interests.

IA and GP designed research; IA, and PIP conducted research and collected the data; IA and AL analyzed the data; IA wrote the manuscript; Critical revision of the manuscript for important intellectual content was the responsibility of IA, AL and GP. IA had full access to all the data in the study and takes full responsibility for the integrity of the data and the accuracy of the analysis. All authors read and approved the final manuscript.