**4. Discussion**

This study sought to determine the effect of fruit and vegetable consumption on all-cause mortality in a representative sample of Spanish community-dwelling older adults who were followed up for a period of approximately 6 years. Our results show that consuming five or more servings per day increases the probability of surviving in the general older population with two chronic conditions by 27%, compared to those who consume three or fewer servings per day. However, this beneficial effect of fruit and vegetable consumption is not found among participants with none or one chronic condition, or three or more.

Most participants (65.7%) in the study did not adhere to the WHO recommendation of consuming a minimum of five servings of fruit and vegetables per day, with the median of servings per day being four. In a very large Spanish sample of university graduates, the mean consumption of fruits and vegetables was 343 g/day and 525 g/day [36], respectively, with an equivalence of approximately four and six portions of 80 g. However, few epidemiological studies have described the patterns of fruit and

vegetable consumption among the older Spanish population. For example, in the Seniors-ENRICA study, a population-based cohort of Spanish older adults aged 60+, a total of 22.5% participants reported having five or more portions of fruit and vegetables a day [23], slightly inferior to the 34.3% found in our study. These discrepancies could be explained by the different tools used to assess fruit and vegetable consumption.

Previous population-based studies have repeatedly reported an inverse association between fruit and vegetable consumption, and all-cause and disease-specific mortality in the older population [37–39], although some studies have reported inconsistent findings about whether this greater risk of all-cause mortality might be mainly due to CVD-related or non-cardiovascular-related deaths, such as cancer [40]. Our findings contribute to this evidence by showing that this effect is exerted through a protective effect in the presence of multimorbidity. Additionally, the consumption of both fruit and vegetables rather than the consumption of only fruit or vegetables seems to be especially beneficial for reducing the risk of CVD [39] and non-cardiovascular diseases [3]. Our findings also support the recommendation of a minimum of five servings per day of fruit and vegetables whereas there were no differences in terms of increased risk of mortality among older adults with low (equal to or less than three servings/d) or medium (four servings/d) consumption. This is in line with previous studies that investigated the risk of all-cause mortality associated with a dose-response of fruit and vegetable consumption in a large population-based cohort aged 45–83 [8]. The authors found that consuming fewer than five servings a day was associated with progressively shorter survival and higher mortality rate, whereas consuming more than five servings did not add any benefits with respect to survival.

A systematic review conducted by Nunes et al [41] showed an overall positive association between multimorbidity (defined as the presence of two or more chronic diseases) and mortality (HR = 1.44, 95%CI = 134–1.55). In the unadjusted model, we found that only three or more chronic conditions were related to a 62% higher probability of having a shorter survival and dying, compared to those with none or one chronic condition. We found that the beneficial effect of consuming five or more fruit and vegetable servings per day is exerted in those having two CCs, but not three or more. Additionally, this protective effect seems to be beyond the confounding effects of other risk factors, such as obesity, physical activity, smoking, gender, or educational level. Participants suffering from three or more chronic conditions might represent complex patients, who might be in need of intensive care. The presence of multiple diseases is related to interactions between morbidities, inadequate use of medication, polypharmacy [42], and frailty [43]. Thus, the protective effect of high intake of fruit and vegetables might not be sufficient to reduce the risk of death in people with three or more CCs. It is also possible that older adults with three or more CCs have been given a prescription of a balanced diet, or have been advised to quit or reduce smoking and alcohol intake [44], which might in turn explain the lack of association between fruit and vegetable intake and time to death in this particular subgroup. Despite this, the beneficial effect of consuming five or more servings per day of fruit and vegetables could be huge. Taking into account that an important proportion of Spanish older adults do not reach the recommended five servings per day of fruit and vegetables along with the high prevalence of multimorbidity in this population, interventions promoting fruit and vegetable consumption among older adults might have a positive impact on reducing the risk of death and increasing their quality of life. Future research is needed to learn whether fruit and vegetable intake is particularly beneficial in reducing the risk of death for a particular pair of diseases.

There are several mechanisms by which fruit and vegetable consumption can reduce the risk of mortality in older adults. Fruit and vegetables contain a variety of nutrients and phytochemicals (i.e., fibre, vitamin C, carotenoids, antioxidants, potassium, and flavonoids) that act through several biological mechanisms to reduce the risk of chronic conditions and premature mortality [3]. Greater intake of fruit and vegetables has also been linked to a greater adherence to the Mediterranean diet in older adults (characterized by abundant consumption of olive oil, minimally processed, locally grown vegetables, fruits, nuts, legumes, and cereals, and proteins coming mainly from fish and shellfish) [45] and to reduced consumption of sweet foods [46] which in turn might also prevent CVD [47], several types of cancer [48,49], cognitive decline, and dementia [50,51], while increasing longevity [52]. Our study did not include data on adherence to the Mediterranean diet or other potential dietary risk factors for non-communicable diseases and risk of mortality, such as consumption of red and processed meat [53] or ultraprocessed food [54]. More studies are needed to determine whether the beneficial e ffect of fruit and vegetable intake on the probability of survival among people with multimorbidity is maintained or attenuated by the presence of these diet-related risk factors. Additionally, the way in which fruits and vegetables are consumed (e.g., raw or cooked) might also play an important role in the potential protective factor among older adults with chronic conditions. Another mechanism by which fruit and vegetable consumption might impact the risk of mortality among older adults is the presence of unhealthy lifestyles among those who consume less fruit and vegetables. Previous research has indicated an inverse association between fruit and vegetable intake and smoking [55], alcohol consumption [56], obesity [57], and sedentarism [58]. The beneficial e ffects of consuming fruit and vegetables, such as lower systemic inflammation [59], reduced oxidative stress [60], and decreased platelet aggregation [61], may partially reduce the e ffects of smoking and alcohol intake [55,56]. However, we did not find significant interactions between fruit and vegetable consumption and smoking status, alcohol consumption, obesity, or low levels of physical activity. Future research is needed to replicate these results.

Our study had some limitations. First, health variables, such as fruit and vegetable consumption, tobacco and physical activity, were self-reported, thus potentially leading to measurement errors or misclassification. Additionally, recall bias might also be present. Second, it was assumed that the fruit and vegetable intake pattern was unchanged during the follow-up period. Third, it is possible that the beneficial e ffect of fruit and vegetable consumption is not observed among participants with none or one chronic condition because they are more likely to survive during the follow-up period. Thus, longer periods of follow-up might be needed. Fourth, measuring fruit and vegetable consumption might be problematic. For example, the study did not extensively measure the dietary habits of the sample through a 24-hour dietary recall or a frequency questionnaire; thus, some measurement bias might have been introduced. Questions concerning the number of fruit and vegetable servings were asked once, ye<sup>t</sup> they may be prone to seasonable bias as well. Additionally, these questions were aggregated, and the e ffect of this variable could be due to specific sorts of fruits and vegetables. Fourth, residual confounding might explain our findings. For example, consuming vitamin supplements or specific diet patterns such as the Mediterranean diet could be related to both fruit and vegetable consumption and mortality. However, findings were adjusted for several potential confounders, such as smoking status, alcohol consumption, physical activity, and obesity.

In sum, the finding that a high level of fruit and vegetable consumption (reaching the threshold of five or more servings per day) significantly reduces the risk of mortality among older adults with two chronic conditions has several implications. As has been shown in the present study, fruit and vegetable intake in the general population of older adults does not approach recommended levels. Interventions to increase fruit and vegetable intake in older adults should take into account their unique nutritional needs and barriers, as well as several characteristics that might influence their fruit and vegetable intake, such as appetite loss, tooth loss and oral problems, changes in perception of hunger, taste acuity and sense of smell (sometimes associated with drugs' side e ffects), and mobility di fficulties in shopping [18]. These factors should be taken into account when designing interventions to promote fruit and vegetable consumption geared to the older population.

**Author Contributions:** Conceptualization, B.O., C.A.E., J.L.A.-M., and J.M.H.; Methodology, B.O., M.V.M., E.L., M.M., N.M.M., D.M.-A., J.L.A.-M., and J.M.H.; Formal analysis, B.O. and M.V.M.; Investigation, B.O., M.V.M., E.L., M.M., N.M.M., D.M.-A., J.L.A.-M., and J.M.H.; Resources, B.O., M.V.M., E.L., M.M., N.M.M., D.M.-A., J.L.A.-M., and J.M.H.; Data curation, B.O., M.V.M., E.L., N.M.M., D.M.-A.; Writing–Original draft preparation, B.O., C.A.E. and A.A.; Writing–Review & Editing, all authors; Supervision, C.A.E., J.L.A.-M., and J.M.H.; Project administration, B.O., E.L., M.M., N.M.M., D.M.-A., J.L.A.-M., and J.M.H., Funding acquisition, M.M., J.L.A.-M., and J.M.H.

**Funding:** This work was supported by the Seventh Framework Programme [grant number 223071-COURAGE Study]; the Instituto de Salud Carlos III-FIS [grant numbers PS09/00295, PS09/01845, PI12/01490, PI13/00059, PI16/00218,

and PI16/01073]; the European Regional Development Fund (ERDF) "A Way to Build Europe" [grant numbers PI12/01490 and PI13/00059]; Horizon 2020 Research and Innovation Programme [grant number 635316-ATHLOS]; and the Centro de Investigación Biomédica en Red de Salud Mental. B.O.'s work is supported by the PERIS program 2016-2020 "Ajuts per a la Incorporació de Científics i Tecnòlegs" [grant number SLT006/17/00066], with the support of the Health Department of the Generalitat de Catalunya. E.L.'s work is supported by the Sara Borrell postdoctoral programme (CD18/00099) of the Instituto de Salud Carlos III (Spain) and co-funded by European Union (ERDF/ESF, "Investing in your future"). Funding sources did not have any role in study design, data collection, analysis and interpretation of data, writing of the article, or the decision to submit it for publication.

**Conflicts of Interest:** The authors declare no conflict of interest.
