Protein requirement studies in elderly people: reassessment of previously reviewed data

When comparing results and conclusions of these studies (Cheng et al, 1978, Gersovitz et al, 1982; Uauy et al, 1978; Zanni et al, 1979), it would be helpful if the same N balance formula was used and the same assumptions were applied to the calculations. To this end, Campbell et al (1994), recently reassessed the N balance data from each of these studies using the N balance formula recommended by the 1985 Consultation. This reassessment was possible because individual subjects' N balance data had been published in each of the original articles. The results are shown in Table 2. Recalculation of the data from Cheng et al (1978) increased the mean protein requirement by 21%, to 0.93 ± 0.05 g protein/kg/d, with the increase mainly due to the inclusion of a factor for miscellaneous N losses in the balance formula. Data from the elderly women and men studied by Uauy et al (1978) were combined and a mean protein requirement of 0.81 ± 0.09 g protein/kg/d calculated. The similarity of this value to the original protein requirement estimates by Uauy et al was expected, since an increase in the estimated miscellaneous N losses from 5 mg to 8 mg N/kg/d was the only change in the N balance formula used. Upon recalculation of the data of Zanni et al (1979), the mean protein intake required for N equilibrium increased by 41%, to 0.65 ± 0.19 g protein/kg/d. This dramatic increase was due mainly to the exclusion of data collected during the 17-day protein-free period from the regression. When the recalculated data from all three studies were combined by weighted mean averaging, a mean protein requirement of 0.89 ± 0.05 g protein/kg/d was estimated (Campbell et al, 1994).

Protein requirement studies in elderly people: new data

In addition to the retrospective reassessment of data from these N balance studies of elderly subjects (Cheng et al, 1978; Uauy et al, 1978; Zanni et al, 1979), Campbell et al (1994) also provided N balance data on 12 men and women, aged 56-80 years, who consumed weight-maintenance diets providing either 0.8 or 1.6 g protein/kg/d. A mean protein requirement for N equilibrium of 1.0 g protein/kg/d was estimated based on N balance data collected during days 6-11 of this short term N balance study. This study was limited by the assessment of N balance at only two levels of protein intake, and the lack of opportunity to calculate separate protein requirements for each subject, because each subject consumed only one of the experimental diets. However, it is noteworthy that the lower limit of the 95% confidence interval for the 1.0 g protein/kg/d mean protein requirement is 0.72 g protein/kg/d, a value considerably greater than the 0.6 g protein/kg/d mean protein requirement established by the 1985 Consultation for adults of all ages.

The 1.0 g protein/kg/d mean protein requirement estimated by Campbell et al (1994) is similar to the 0.97 g protein/kg/d mean protein intake that maintained 24 healthy men and women, aged 70-86 years, in N equilibrium (Bunker et al, 1987). Homebound elderly people, aged 7086 years, were reported to be in marked negative N balance at a mean protein intake of 0.67 g protein/kg/d. These data by Bunker et al (1987) actually overestimate N balance because no factor for miscellaneous N losses was used in the balance calculation.

A summary of the new data from Campbell et al (1994), and the recalculation of data from Cheng et al (1978), Uauy et al (1978) and Zanni et al (1979), is provided in Figure 1. A weighted mean protein requirement for elderly people of 0.91 ± 0.043 g protein/kg/d was estimated from data of these four studies combined. As shown, the estimated mean protein requirement in three of the four studies is higher than the safe level of 0.75 g protein/kg/d recommended by the 1985 Consultation for adults of all ages.

After a mean protein requirement is established, an upward adjustment should be made to establish a safe level of intake (recommended allowance) to meet the metabolic needs of virtually all adults. The Consultation (1985) estimated the coefficient of variation of protein requirements of young adults at 12.5%, and suggested that the dietary protein needs of 97.5% of individuals in this population group would be met with a protein intake 25% above the mean protein requirement. Based on this estimate, a safe protein intake for elderly men and women would be 1.0 g protein/kg/d or more high quality protein.

Table 3 Conclusions from protein requirement studies in elderly people

Reference	Original conclusions of the authors	Conclusions based on 1985 FAO/WHO/UNU calculations
Cheng et al (1978)	0.8 g protein/kg/d	0.8 g protein/kg/d
	adequate intake	not adequate intake
Uauy et al (1978)	0.8 g protein/kg d	0.8 g protein/kg/d
	not adequate intake	not adequate intake
Zanni et al (1979)	1973 FAO/WHO safe intake of 0.57 g protein/kg/d adequate intake	Mean protein requirement 41% higher than original estimate. But 0.57 g protein kg/d still considered adequate intake
Gersovitz et al (1982)	0.8 g protein/kg d not adequate intake	Conclusions the same. 0.8 g protein/kg/d not adequate intake
Campbell et al (1994)	0.8 g protein/kg d not adequate intake	Conclusions the same. 1985 Consultation N balance formula used
Castaneda et al (1995)		0.8 g protein/kg d not adequate intake 1985 Consultation N balance formula used

Figure 1 Dietary protein intakes required to achieve nitrogen equilibrium in elderly subjects. Individual subject (open circles) and group mean protein intakes (±2 s.e.m.s; solid bars) at nitrogen equilibrum were calculated from data by Zanni et al (1979), Cheng et al (1978), Uauy et al (1978) and Campbell et al (1994), based on the 1985 FAO/ WHO/UNU nitrogen balance formula (FAO/WHO/UNU, 1985). The dashed line represents the current estimated mean protein requirement for all adults (0.60 g protein/kg/d), and the dotted line represents the current suggested safe level of protein intake for elderly men and women (0.75 g protein/kg/d) (FAO/WO/UNU, 1985). Figure adapted from Campbell et al (1994).

The 1985 report acknowledged that no longer-term N balance data in elderly subjects were available when the Consultation met in October 1981, other than the 30-day balance data of Gersovitz et al (1982). Recently Castenada et al (1995) completed a study assessing longer-term (9 weeks) adaptation to marginal protein intakes in 12 elderly women aged 66-79 years. Subjects consumed weight-maintenance diets providing either 0.45 or 0.92 g protein/kg/d, and N balance was measured for 6-day periods during study weeks 3 and 9. All six women provided 0.45 g protein/kg/d were in negative N balance during week 3 (group mean -15.3 ± 3.3 mg N/kg/d), and the mean negative N balance moved significantly toward N equilibrium at week 9 (group mean - 5.2 ± 1.3 mg N/kg/d). However, this shift in N balance occurred only after these elderly women had accommodated with an 8% decline in body cell mass (assessed by potassium-40 scanning) and an 8% decline in muscle mass (assessed by urinary creatinine). Five of these six women remained in negative balance at week 9. The six women provided 0.92 g protein/kg/d achieved N equilibrium throughout the study, with a mean balance of 7.5 ± 6.3 and 1.7 ± 2.5 mg N/kg/d at weeks 3 and 9, respectively, and maintained body cell mass and muscle mass.

At week 3, a mean protein requirement of 0.78 g protein/kg/d for N equilibrium may be estimated from regression analysis of data from both the 0.45 and 0.92 g protein/kg/d groups. A similar estimated mean protein requirement may be calculated at week 9 (0.82 g protein/ kg/d). Thus, both short- and longer-term N balances support previous reports (Campbell et al, 1994; Gersovitz et al, 1982; Uauy et al, 1978) demonstrating higher protein requirements in healthy elderly subjects.

Short-term N balance studies (1-3 week study periods) have criticized for allowing inadequate time to fully adapt to steady-state at a given protein intake level. Indeed, these data (Castaneda, 1995), and those of Gersowitz et al (1982), show that N balance tends to move toward equilibrium (i.e. zero balance) when extended past 3 weeks in elderly men and women. That the mean protein requirement estimates in elderly women were similar at both 3 and 9 weeks, despite this shift, however, supports the conclusion that short-term balance studies are predictive of protein requirements that would be estimated from longer-term studies (Campbell, 1994). This is fully consistent with the close similarity of mean protein requirement estimates from short- and longer-term balance studies, 0.63 and 0.58 g protein/kg/d, respectively in studies reviewed by the 1985 Consultation in young men. In addition, the data of Castaneda et al demonstrate that N equilibrium shown in longer-term balance studies may be established as a result of significant detrimental changes in body composition. Thus, N balance data generated from longer-term studies do not provide an accurate assessment for protein requirements of subjects in the metabolic state in which they entered the study.

Thus available assessments of protein requirements utilizing both short- and longer-term N balance techniques suggest that the mean protein requirement of elderly men and women may be ³ 0.8 g protein/kg/d. This would indicate a safe level of intake for most elderly people of ³ 1.0 g protein/kg/d.