A number of socio-economic, biological and lifestyle characteristics change with advancing age and place very old adults (those aged 85 years and over) at increased risk of micronutrient deficiencies. For example, 10–30 % of older adults (aged 65 years and over) have atrophic gastritis and hypochlorhydria( Reference Johnson 1 ), which reduce secretion of acid-pepsin and intrinsic factor, allowing small-bowel bacterial growth and leading to impaired vitamin B12 absorption( Reference Hughes, Ward and Hoey 2 ). Although micronutrient malabsorption is not an inherent consequence of ageing, the absorption of pH-dependent vitamins and minerals, such as folate, vitamin B12, Ca, Fe and β-carotene, might be partially compromised( Reference Johnson 1 , Reference Russell 3 ). Very old adults are also at higher risk of vitamin D deficiency due to reduced skin stores of 7-dehydrocholesterol (provitamin D), renal impairment and reduced renal conversion of its biologically inert to active form (i.e. 25-hydroxyvitamin D to calcitriol), immobility, malnutrition and environmental factors (reviewed in Hill et al. ( Reference Hill, Aspray and Francis 4 )). Micronutrient deficiencies may contribute to disability, frailty and impaired physical function in very old adults( Reference Inzitari, Doets and Bartali 5 ).
In the UK, apart from the Reference Nutrient Intake (RNI) for vitamin D, which sets a dietary reference value (DRV) for people aged 65 years and over, all other DRV for vitamins or minerals apply equally to everyone aged ≥50 years( 6 ). The scarcity of dietary data on very old adults, and lack of evidence on relationships with risk factors and health outcomes, have resulted in DRV based on extrapolations from younger populations( Reference McCormick 7 ).
The 1994/1995 National Diet and Nutrition Survey (NDNS) of people aged 65 years and over identified a significant number of older adults with inadequate micronutrient intakes – namely, vitamin D, Mg and K( Reference Bates, Prentice and Cole 8 ). A review of micronutrient intakes across Europe revealed that inadequacy (assessed against the Nordic Nutrition Recommendations, estimated average intake) was present in more than 20 % of older adults (≥65 years) for vitamin D, folate, Ca and Se( Reference Roman Vinas, Ribas Barba and Ngo 9 ). Similarly, a review of non-institutionalised older adults living in Western countries concluded that at least 30 % were below the estimated average requirement (EAR) for vitamin D, vitamin B2, Ca, Mg and Se( Reference Ter Borg, Verlaan and Hemsworth 10 ).
The aims of this study were to assess daily energy, vitamin and mineral intakes of 85-year-olds participating in the Newcastle 85+ Study, determine their food sources, compare intakes with the current UK DRV and to explore socio-economic and lifestyle determinants of micronutrient intakes.
Newcastle 85+ Study
The Newcastle 85+ Study is a longitudinal, population-based study of health trajectories and outcomes of a cohort of 845 very old people (85 years old at baseline) born in 1921 (for details visit http://research.ncl.ac.uk/85plus)( Reference Collerton, Davies and Jagger 11 – Reference Davies, Kingston and Robinson 13 ). Complete dietary intake data (without protocol violation) were available for 793 participants (302 men and 491 women).
Dietary assessment, micronutrient estimation and supplement use
Dietary intakes were collected using a 24-h multiple-pass recall (24 h-MPR) tool on two distinct occasions (1 week apart and on different days of the week) at baseline (2006/2007) by trained research nurses and in the participant’s usual residence. Food and drink portion sizes were estimated using the Photographic Atlas of Food Portion Sizes ( Reference Nelson, Atkinson and Meyer 14 ). All dietary intake data were independently double entered. Any discrepancies were identified, checked against original records and corrected before data analysis. Energy, vitamin and mineral intakes were estimated using McCance and Widdowson’s the Composition of Food, 6th edition (used as published)( 15 ) together with a purpose-designed in-house Microsoft Office Access database on the nutrient composition of commonly consumed foods( Reference Mendonça, Hill and Granic 16 ). In all, 85 and 90 % of the participants believed that the 24 h-MPR reflected their usual food and drink intakes, respectively. Intakes of energy, vitamin A, β-carotene, vitamin B2, vitamin B6, folate, vitamin B12, vitamin E, vitamin C, vitamin D, Ca, Fe, Mg, K, Na, Se and Zn are reported in this article (excluding supplements). Vitamin and mineral densities per 1 MJ of energy were also calculated.
Supplement use was divided into three categories: no supplements, one supplement and two or more supplements. Information on supplement use was limited to type and brand; therefore, micronutrient-containing supplements were assumed to be taken according to the manufacturer’s specifications. Supplement users were characterised by supplement type: those taking fish and n-3 oil preparations, single mineral/vitamin preparations, multivitamin and/or multimineral preparations, and other supplements. Micronutrient intakes from all sources (including supplements) and the difference (%) between micronutrient intakes from dietary sources only (excluding supplements) were determined, but supplements were not included in the main analysis.
Individual foods were coded and allocated to food groups. In brief, individual foods were allocated to fifteen first-level food groups: cereals and cereal products (CCP), milk and milk products, eggs and egg dishes, oils and fat spreads, meat and meat products, fish and fish dishes, vegetables, potatoes, savoury snacks, nuts and seeds, fruits, sugar, preserves and confectionery, non-alcoholic beverages, alcoholic beverages and miscellaneous( Reference Mendonça, Hill and Granic 16 ). The average contribution of food groups to vitamin and mineral intakes was reported so that ≥90 % of intakes could be explained.
Estimation of misreporting
The proportion of possible misreporters was calculated using a energy intake:BMR cut-off value of 1·05–2·00 (further details can be found in the study by Mendonça et al.( Reference Mendonça, Hill and Granic 16 )). With this cut-off value, 26·3 % were identified as misreporters (21·6 % as under-reporters and 4·7 % as over-reporters). Possible misreporters have not been excluded from the analysis because of the uncertainty surrounding this estimate and the small differences observed between excluding and not excluding misreporters( Reference Mendonça, Hill and Granic 16 ). Further, in 5 % of the participants (n 42), the proxy was the only respondent.
Socio-economic, health and lifestyle factors
Apart from supplement use, details on other socio-economic and lifestyle variables have been previously published( Reference Collerton, Davies and Jagger 11 ) and commented on in the companion paper: ‘Macronutrient intake and food sources in the very old: analysis of the Newcastle 85+ Study’( Reference Mendonça, Hill and Granic 16 ). Participants were classified according to housing: standard, sheltered or institutional housing. Further, participants were characterised as living alone, with spouse or with others, years of full-time education (categorised as <9/10–11/and >12 years) and social class according to the National Statistics Socio-Economic Classification (NS-SEC) three-class scheme( Reference Chandola and Jenkinson 17 ). Participants were also categorised into those with low (scores 0–1), medium (scores 2–6) and high (scores 7–18) physical activity using a validated and purpose-designed physical activity questionnaire( Reference Innerd, Catt and Collerton 18 ).
The Shapiro–Wilk test and quantile-quantile plots were used for normality testing. Normally distributed data are reported as mean values and standard deviations and non-normal data as medians and interquartile ranges (IQR). Baseline characteristics, micronutrient intake and percentage of participants below the Lower Reference Nutrient Intake (LRNI), EAR, RNI and upper level (UL) were calculated using descriptive statistics. If available, LRNI was the preferred DRV to be reported. The LRNI is only supposed to meet the needs of 2·5 % of a given population and intakes below this are likely to be ‘inadequate’. When appropriate, sex differences were assessed with the two-sample t test or the χ 2 test for normally distributed continuous variables and categorical variables, respectively. Most micronutrient intake data were continuous and non-normally distributed; therefore, sex differences were determined by the Mann–Whitney U test. Vitamin and mineral intakes were stratified by housing, living arrangements (with whom participants lived), years of full-time education, social class (coded to the NS-SEC three-class system( Reference Chandola and Jenkinson 17 )) and physical activity groups and compared by multinomial logistic regression. Apart from energy, which was adjusted for sex only, all vitamins and minerals were adjusted for sex and energy. Exploratory and statistical analyses were conducted using the IBM statistical tool SPSS version 22.0. Values of P<0·05 were considered significant.
Men had higher vitamin intakes than women except for vitamin C (Table 2). However, the overall higher vitamin intake by men disappeared when the results were expressed per 1 MJ. Specifically, women’s vitamin A intake was 12 µg retinol equivalents (RE)/MJ or 13 % higher (P=0·008) and vitamin C intake was 20 mg/MJ or 28 % higher (P=0·001) compared with men. Despite 43 % of the participants (n 335) consuming one or more supplements on a regular basis (Table 1), on a population level, vitamin intakes changed only marginally when supplements were included, except for vitamin A and D, which increased by 19·2 % (from 620 to 752 µg-RE) and by 22·5 % (from 2·0 to 2·5 µg), respectively (online Supplementary Table S1). Owing to the modest differences to micronutrient intakes when including supplements, and limitations in supplement frequency data, micronutrient consumption from supplements was not included in the main analysis.
NS-SEC, National Statistics Socio-economic Classification; % en, percentage of energy.
* χ 2 Test for no sex difference unless otherwise stated.
† Excludes participants living in institutions.
‡ Mann–Whitney U test for no sex difference.
§ Independent t test for no sex difference.
Vitamin food sources
Fig. 1 shows the percentage contribution of food groups to vitamin intake for all participants. Meat and meat products contributed to 40% of vitamin A intake – the majority coming from liver and liver products and dishes (94·4 %). Vegetables were the second highest contributors (22·4 %) to vitamin A intake, of which most came from carrots (71·1 %). CCP were the highest contributors (31·5 %) to folate intake, 86·9 % of which came from bread and breakfast cereals. Vegetables were the second highest contributors (15·8 %) to folate intake with 42·4 % coming from cruciferous vegetables. Half (49·6 %) of the vitamin B12 intake from meat and meat products (52·3 %) came from liver and liver products and dishes; one-third (33·8 %) of vitamin D intake came from fish and fish dishes (98·9 % of which was from oily fish) and 23·8 % from CCP (45·2 % of which was from breakfast cereals and 43·3 % from buns, cakes, pastries and fruit pies).
Similar to vitamin intake, men had an overall higher mineral intake than women (24 % higher on average) (Table 2). When expressed per 1 MJ of energy, men still had higher intakes of Fe (P=0·005), Se (P=0·028) and Zn (P<0·001) compared with women but lower Ca intakes (P=0·008). On a population level, supplement contribution to mineral intakes was almost negligible (online Supplementary Table S1). The highest difference between dietary intake with and without supplements was only 2·7 % for Zn (from 7·1 to 7·3 mg).
RE, retinol equivalents.
* Does not include supplements.
† Mann–Whitney U test for no sex difference (median/1 MJ of energy).
‡ Does not include energy from alcohol.
§ Does not include table salt and salt used for cooking.
Mineral food sources
Fig. 1 shows the percentage contribution of food groups to vitamin intakes for all participants. Milk and milk products were the highest contributors (31·3 %) to Ca intake while CCP was second with 27·5 % (36·6 % of which came from bread). Non-alcoholic beverages contributed 18·9 % to Ca intake mainly because tea and coffee (with added milk) were included in this group (95·4 % came from tea, coffee and water). Non-alcoholic beverages accounted for 19 % of K intake (81·5 % of which was from tea, coffee and water). CCP (15·8 %) and potatoes (14·6 %) were the second and third, respectively, highest contributors to K intake. CCP explained 46·7 % of Se intake and 93·2 % of this came from bread. Meat and meat products made a higher contribution to intakes of Fe (19·3 v. 14·2 %), vitamin D (20·3 v. 13·4 %) and vitamin B12 (59·2 v. 47·8 %) for men than for women (data not shown).
The failure of both men and women in the Newcastle 85+ Study to meet several micronutrients’ DRV was widespread (Fig. 2 and online Supplementary Table S2). In all, 20 % of the participants had intakes below the LRNI for Mg, K and Se. The proportion of participants below the LRNI for vitamin A, vitamin B12 and Zn was around 10 %. However, 4·6 % (n 36) of the participants had vitamin A intakes above the UL. The widest disparity between intake and recommendations was seen for vitamin D intake, with >95 % (n 756) of participants having intakes below the RNI for vitamin D of 10 µg/d (EAR or LRNI for vitamin D have not been defined for the UK)( 6 ), and 52·7 % (n 418) of the participants were below the LRNI for Se. In contrast, 82·2 % (n 652) of the participants were above the RNI for Na of 1600 mg/d( 6 ). The 95th percentile of Na intake was 4663 mg/d, and among those with intakes above the RNI the median intake was 2594 mg. Fewer men had intakes below the DRV for vitamin B12, Fe, K and folate than women. The widest difference between men and women not meeting the LRNI was for vitamin B12 (5·0 v. 12·4 %, P<0·001) and Fe (2·3 v. 7·8 %, P<0·001). Meat and meat products were top contributors for both micronutrients.
Micronutrient intake by housing, socio-economic status and physical activity
Table 3 reports the energy, vitamin and mineral intakes in the Newcastle 85+ Study stratified by housing, living arrangements, years of full-time education, social class (past occupation according to NS-SEC) and physical activity. All micronutrient models were adjusted for sex and food energy intake.
NS-SEC, National Statistics Socio-economic Classification; Stand, standard; Institut, institutional housing; Class 1, higher managerial, administrative and professional occupations; Class 2, intermediate occupations; Class 3, routine or manual occupations.
* P<0·05, ** P<0·01, *** P<0·001.
† All models were adjusted for sex and energy intake except for energy intake, which was only adjusted for sex. Standard housing, living alone, ≤9 years of full-time education, class 3 of past occupation and low physical activity were the reference categories.
‡ Does not include supplements.
§ Does not include energy from alcohol.
|| Does not include table salt and salt used for cooking.
Energy and vitamin D intakes were higher in participants who lived in institutional care (nursing or residential) than in standard housing. Conversely, vitamin E, Mg and K intakes were lower in institutional than in standard housing. Participants who lived with their spouses had higher K and Se intakes than those who lived alone. Those with 12 or more years of full-time education had higher intakes of vitamin C, vitamin D, Ca, Mg and K than those with ≤9 years of full-time education. Social class also predicted the intakes of several vitamins and minerals. Participants with previous higher managerial, administrative and professional occupations (class 1) had higher intakes of vitamin B2, folate, Ca, Fe, Mg, K and Zn than those who had routine and manual occupations (class 3). Those with high physical activity had a more nutrient-dense diet including vitamin B6, folate, vitamin E, vitamin C, Fe, Mg, K and Zn than those with lower physical activity.
The median vitamin D, Mg, K and Se intakes were 2·0 (IQR 1·2–6·5) µg/d, 215 (IQR 166–266) mg/d, 2477 (IQR 1890–3023) mg/d and 39·0 (IQR 27·3–55·5) µg/d, respectively. Participants with more full-time years in education, from higher social class and those who were more physically active had more nutrient-dense diets including several vitamins and minerals. The most notable findings are that 20 % or more of the participants in the Newcastle 85+ Study had intakes below the LRNI for Mg, K and Se and that >95 % of the participants were below the RNI of 10 µg/d of vitamin D. Very old adults may be at increased risk of micronutrient deficiencies, which contributes to disability, frailty and loss of physical function( Reference Inzitari, Doets and Bartali 5 ). Therefore, a deeper understanding of the dietary habits of the very old is an important prerequisite for developing evidence-based, age-specific dietary recommendations.
Comparison with other studies
Since the 1994/1995 NDNS of people aged 65 years and over, which included 172 men and 287 women aged 85 years and over (all non-institutionalised), no study has described micronutrient intakes and food sources in a large sample of very old adults in the UK. Most vitamin and mineral intakes were similar between the two studies, except for β-carotene (1141 v. 1516 µg/d), vitamin C (41·4 v. 56·5 mg/d) and Ca (644 v. 731 mg/d), which were higher in the Newcastle 85+ Study participants (intakes from food sources only)( Reference Finch, Doyle and Lowe 19 ). In the 1994/1995 NDNS, less vitamin A (34 v. 40 %) and vitamin B12 (43 v. 53 %) were derived from meat and meat products and less K from non-alcoholic drinks (10 v. 19 %). However, more vitamin B12 (29 v. 13 %), Ca (54 v. 31 %) and K (20 v. 9 %) came from milk and milk products in the 1994/1995 NDNS than in the Newcastle 85+ Study. The food sources of vitamin D were considerably different between the studies with fish and fish dishes making a lower contribution to intake (17 v. 34 %), whereas fat spreads made a higher contribution (23 v. 8 %) in the 1994/1995 NDNS than in our study( Reference Finch, Doyle and Lowe 19 ). The observed differences are unlikely to be due to fortification policies. The Newcastle 85+ Study included 85-year-olds only, whereas the 1994/1995 NDNS included those aged 85 years and over. Other possible reasons include different dietary assessments (4-d weighted diet record v. 2×24 h-MPR) that diverged by more than a decade.
The European Prospective Investigation into Cancer and Nutrition (EPIC)-Oxford third follow-up questionnaire in 2010–2014 included 411 men and 872 women aged 80 years and over( 20 ). Intakes of all vitamins and minerals were at least 20 % higher in the EPIC-Oxford than in the Newcastle 85+ Study participants (Tim Key & Paul Appleby, personal communication). Different descriptive statistics and dietary assessment methods used, different ages (≥80 v. 85-year-olds) and characteristics of the participants (14 % of EPIC-Oxford participants were vegetarians) are potential explanations for the wide differences observed in micronutrient intakes.
The current NDNS rolling programme (from 2008/2009 to 2011/2012 or years 1 to 4) does not yet have enough very old adults for comparison with our study. Nonetheless, it included 428 adults (191 men and 237 women) aged ≥65 years( Reference Bates, Lennox and Prentice 21 ). Although energy intakes were similar between both the studies, vitamin and mineral intakes (without supplements) were slightly higher in the NDNS than in the Newcastle 85+ Study (except for Na where intakes were 1947 and 2383 mg/d, respectively). More than 10 % of the participants had intakes of Mg, K and Se below the LRNI( Reference Bates, Lennox and Prentice 21 ). Similarly, >20 % of the Newcastle 85+ Study participants were also below the LRNI for these minerals.
Public health implications
In the Newcastle 85+ Study, men had higher energy intakes than women; therefore, it was not unexpected that intakes of most micronutrients by men were also higher. However, when vitamin and mineral intakes were expressed per 1 MJ, vitamin A, C and Ca intakes were higher in women than in men. Conversely, men’s diets were more nutrient dense in vitamin B12, Fe and Se compared with women. Higher meat and meat products consumption by men was the main cause for these differences.
Several micronutrient intakes were lower than the current DRV. In all,>20 % of the participants were below the LRNI for Mg, K and Se, whereas 95·3 % of the participants were below the RNI for vitamin D (the Scientific Advisory Committee tentatively set the same RNI as the Committee on Medical Aspects of Food and Nutrition Policy( 22 )). This is of concern because Mg is associated with physical performance( Reference Veronese, Berton and Carraro 23 ), systemic inflammation, endothelial function( Reference Chacko, Song and Nathan 24 ) and bone mineral density in older adults( Reference Orchard, Larson and Alghothani 25 ); inadequate Se has been linked with anaemia( Reference Semba, Ricks and Ferrucci 26 ), cancer and all-cause mortality( Reference Ray, Semba and Walston 27 ); and low vitamin D intake has been consistently associated with musculoskeletal( Reference Hill, Aspray and Francis 4 ) and extra-skeletal outcomes, including cognitive impairment and mortality( Reference Granic, Aspray and Hill 28 , Reference Granic, Hill and Kirkwood 29 ). However, the major ‘inadequacy’ in vitamin D intake may not reflect vitamin D status, as circulating concentrations of 25-hydroxyvitamin D depend largely on sun exposure( Reference Hill, Aspray and Francis 4 ). Higher K intakes are a known protective factor for hypertension( Reference Smith, Klotman and Svetkey 30 ), whereas excessive Na intake is an established risk factor for hypertension in older adults( Reference Alam and Johnson 31 ). In our study, only a fifth of the participants were below the RNI of 1600 mg/d of Na but half of them met the recommendation of <2400 mg/d. Na intake reduction and increased K intake might help reduce the prevalence of stroke and fatal CHD in this population( Reference Strazzullo, D’Elia and Kandala 32 ).
More than 10 % of participants had vitamin A intakes below the LRNI but, interestingly, 5 % had intakes above the UL of 3000 µg-RE/d set by the European Food Safety Authority( 33 ). This classic paradox may not be the result of habitual intake, but the result of consuming high vitamin A-containing foods (e.g. liver and liver dishes) on one or more of the non-consecutive 24-h recalls of the 24 h-MPR( 34 ). In fact, thirty-five out of the thirty-six participants who had vitamin A intakes above the UL of 3000 µg-RE ate liver and liver products at least on one of the 24 h-MPR.
Assessing micronutrient intake inadequacies in this age group has several methodological limitations. Among all, 27 % (n 214) of the participants were classified as cognitively impaired (Standardized Mini-Mental State Examination≤25) (data not shown), which might have played a major part in misreporting (estimated to be 26·3 %). Further, because of a scarcity of nutrition data in this age group, most DRV were extrapolated from younger populations. This leads to uncertainty regarding the health significance of inadequacies in the very old.
In line with previous studies( Reference Darmon and Drewnowski 35 ) and a recent review on socio-economic determinants of micronutrient intakes in older adults( Reference Novakovic, Cavelaars and Geelen 36 ), participants with more education and from a higher social class had overall higher micronutrient intakes. Similarly, perhaps because healthy habits cluster together, those who were more physically active had more nutrient-dense diets. It has been argued that nutrient-dense foods are more expensive than less-healthy foods in the UK and USA( Reference Jones, Conklin and Suhrcke 37 , Reference Aggarwal, Monsivais and Cook 38 ), and this price differential might explain the difference in nutrient density between lower and higher socio-economic status (SES) groups. However, others have challenged the view that healthier foods or dietary patterns are more expensive than unhealthy ones and, for example, price differentials are dependent on the unit of comparison (e.g. per unit of energy, per unit of mass)( Reference Rao, Afshin and Singh 39 , Reference Carlson and Frazão 40 ). Physical proximity to (and/or means to access) fresh-produce stores has been proposed as an explanation for higher micronutrient intakes in high SES groups( Reference Bodor, Rose and Farley 41 ) but this is somewhat debatable in the UK and North-East England( Reference Cummins and Macintyre 42 ). Inaccessibility to fresh produce, higher cost of nutrient-dense foods in the UK and poorer food choices( Reference McKinnon, Giskes and Turrell 43 ) are some of the potential causes that mediate the diet quality gradient between SES groups. In this age group, with more disabilities and lower income, these issues might be exacerbated.
Strengths and weaknesses
The Newcastle 85+ Study was socio-demographically representative of the general UK population. However, all participants were from Newcastle upon Tyne and North Tyneside and of a predominantly white background, which can limit generalisations( Reference Mendonça, Hill and Granic 16 ). We performed 35 % of the 24-h-recalls during summer (June–August), whereas the rest were evenly distributed throughout the other three seasons. Seasonality is known to influence micronutrient intakes, but the slight bias towards summer is unlikely to have changed the results. Although vitamins and minerals are not abundantly present in commonly under-reported foods such as sweets and snacks, the inherent retrospective nature of the 24 h-MPR might have proved challenging for some individuals in this age group. Adamson et al. ( Reference Adamson, Collerton and Davies 44 ) have described in detail the challenges of dietary assessment in this age group and in the pilot study. To reduce patient and interviewer burden, only qualitative data on supplement use were collected. Therefore, the frequency of supplement use had to be estimated on the basis of the manufacturer’s recommendations. Data on Na derived from table salt and salt used in cooking were not recorded, which might have underestimated Na intake in the Newcastle 85+ Study.
Food sources of the selected micronutrients in the Newcastle 85+ Study were diverse; however, as CCP were widely consumed, they were among the top contributors to intakes of several vitamins and minerals. Higher SES and greater physical activity were associated with higher micronutrient intakes. Compared with current DRV, several micronutrient intakes were ‘inadequate’ and Se (52·7 % below the LRNI) and vitamin D (95·3 % below the RNI) showed the greatest disparities. However, the lack of evidence-based, age-specific DRV for micronutrients for the very old means that such information should be interpreted with caution. As energy requirements are dependent on energy expenditure, the decrease in energy needs in later life mirrors the age-dependent fall in physical activity. However, the physiological basis for age-dependent changes in vitamin and mineral requirements (if any) is poorly understood. In the absence of such evidence, it may be appropriate that dietary information for very old people focuses on healthy food choices, on increasing nutrient density and only recommending the use of supplements in specific situations( 45 ).
In summary, this study provides novel insights into micronutrient intakes, their corresponding food sources and the socio-demographic and lifestyle determinants of micronutrient intakes in very old people. Given the dearth of dietary intake data in the very old, the contemporary micronutrient data from our study are likely to be the most reliable for this age group in the UK. These findings need to be confirmed in other cohort studies of the very old.
The authors acknowledge the operational support of the North of England Commissioning Support Unit (formerly NHS North of Tyne) and of the local general practitioners and their staff. The authors thank Professor Tim Key and Dr Paul Appleby for providing dietary data on the EPIC-Oxford for comparison purposes. The authors also thank the research nurses, dietary coders, management and clerical team for outstanding work throughout, as well as many colleagues for their expert advice. Thanks are due especially to the study participants and, where appropriate, their families and carers.
The Newcastle 85+ Study has been funded by the Medical Research Council, Biotechnology and Biological Sciences Research Council and the Dunhill Medical Trust. The research was also supported by the National Institute for Health Research (NIHR), Newcastle Biomedical Research Centre, based at Newcastle upon Tyne Hospitals, NHS Foundation Trust and Newcastle University. A. J. A. is funded by the NIHR as a Professor in translational research.
The authors’ contributions to the study were as follows: N. M., T. R. H., C. J. S. and C. J. designed the study, N. M. analysed the data, performed statistical analyses and wrote the paper, T. B. L. K. is the principal investigator on the Newcastle 85+ Study. All authors contributed to the interpretation of the findings of the study, read, critically reviewed the manuscript, commented and approved the final version of the manuscript.
None of the authors reported any conflicts of interest.
For supplementary material/s referred to in this article, please visit http://dx.doi.org/10.1017/S0007114516002567