Disease-related malnutrition is one of the most, or even the most, frequent comorbidity in any healthcare setting and has relevant negative consequences on the patient's outcome(Reference Norman, Pichard and Lochs1–Reference Cereda3). Although the prevalence of overweight/obesity has been reported to be about twice as high as in the general population, nutritional risk and malnutrition may also occur in Parkinson's disease (PD), particularly in the advanced stages of the disease(Reference Barichella, Cereda and Pezzoli4, Reference Barichella, Villa and Massarotto5). The pathogenesis of weight loss (WL) and/or low body weight in PD has not been elucidated, but the main causes appear to be reduced energy intake, due to several neurological, social and psychological factors, and increased energy expenditure secondary to dyskinesias(Reference Barichella, Cereda and Pezzoli4, Reference Cereda, Pezzoli and Barichella6, Reference Bachmann and Trenkwalder7). Despite being mainly a neurodegenerative motor disorder, PD is frequently characterised by a number of gastrointestinal dysautonomia symptoms of nutritional interest, such as dysphagia, sialorrhoea and constipation, which in turn could result in an impairment of nutritional status through the modification of eating behaviours and energy balance. A Movement Disorders Society-commissioned task force has recently suggested that clinical rating scales should be used to screen for these complications(Reference Evatt, Chaudhuri and Chou8). An overview of prevalence data on malnutrition in PD has recently been provided(Reference Sheard, Ash and Silburn9), but multivariate evaluations of dysautonomia symptoms and other features of PD, with an impact on nutritional risk, are limited(Reference Wang, Wan and Cheng10). With this background, we designed the present study to investigate the prevalence of nutritional risk in PD and to assess the relationship between nutritional risk and PD features, including dysautonomia symptoms, as assessed by the available clinical rating scales.
Materials and methods
Study design
The study was conducted according to the guidelines laid down in the Declaration of Helsinki, and all the procedures involving human subjects were approved by the ethics committee of our institute (Istituti Clinici di Perfezionamento, Milan). Written informed consent was obtained from all patients.
We designed a prospective single-centre cohort study of all PD patients hospitalised (mean length of hospital stay: 4 d) on a scheduled basis from January 2011 to October 2011 in the ward of the Parkinson Institute (Istituti Clinici di Perfezionamento Hospital, Milano). These included new patients seen at the outpatients' clinic and admitted for diagnosis confirmation and setting of drug treatment, as well as those for whom PD diagnosis had been already made. In the latter case, according to our follow-up protocols, PD patients are periodically hospitalised (every 2·5–3 years) for the re-evaluation of pharmacological treatment and the eventual execution of neuroimaging procedures.
A cross-sectional analysis of baseline data was the objective of the present study.
Subject inclusion and assessment
All patients admitted to the Institute were potentially eligible for inclusion in the study, unless they had not given their informed consent to participate in writing.
All the subjects were assessed for demographic information, PD, anthropometry, biochemistry, nutritional risk and gastrointestional dysautonomia symptons, as described.
Demographic information
Age and sex were noted.
Information on Parkinson's disease
Anti-Parkinson medications, duration of PD and its clinical rating by the Hoehn and Yahr scale (a commonly used system for describing how the symptoms of PD progress)(Reference Hoehn and Yahr11) and the unified Parkinson's disease rating scale(Reference Fahn, Elton, UPDRS Program Members, Fahn, Marsden, Goldstein and Calne12) were used. The unified Parkinson's disease rating scale is made up of several questions (answered by direct interviewing and clinical observation) grouped in four sections and addressing the following areas: part I, mentation, behaviour and mood; part II, self-evaluation of the activities of daily life (including speech, swallowing, handwriting, dressing, hygiene, falling, salivation, turning in bed, walking, cutting food, freezing when walking, tremor and sensory complaints); part III, clinician-scored motor evaluation; and part IV, complications of therapy. In regard to this scale, attention was focused on the scores of parts II and III.
Anthropometry
Weight (to the nearest 0·1 kg), standing height (to the nearest cm), BMI (calculated as the ratio between weight (kg) and height (m) squared), mid upper-arm circumference (cm), triceps skinfold (mm) and arm muscle area were measured. In the case of abnormal spinal curvature or evident camptocormia, the use of standing height was replaced by that of estimated height from knee–heel length. All measurements were performed according to standard procedures(13–Reference Cereda, Bertoli and Vanotti16). The mean of three measurements was considered in the analyses. Moreover, self-reported information on usual body weight was collected in order to assess the history of WL in the last 3–6 months.
Biochemistry
Venous blood samples were drawn after 8–12 h of fasting for the determination of serum albumin and transferrin. The following threshold values were considered indicative of an impairment in nutritional status: albumin < 35 g/l and transferrin < 2000 mg/l(Reference Omran and Morley17).
Nutritional risk
Nutritional screening was conducted through the Malnutrition Universal Screening Tool (MUST)(Reference Stratton, Hackston and Longmore18). We used this tool because, after being initially developed for use in the community, it appears to be the most popular for hospitalised patients worldwide and is recommended by the European Society of Clinical Nutrition and Metabolism due to its high degree of reliability (low inter-observer variation) and association with outcome(Reference Kondrup, Allison and Elia19, Reference Schindler, Pernicka and Laviano20). The MUST is based on three clinical parameters (BMI, WL and metabolic stress due to acute disease) that have been associated with poor outcome. Each parameter is rated as 0, 1 or 2 as follows: BMI>20 kg/m2= 0, 18·5–20·0 kg/m2= 1 and < 18·5 kg/m2= 2; WL < 5 % = 0, WL 5–10 % = 1 and WL>10 % = 2; acute disease: absent = 0 and present = 2. Accordingly, the total score (sum of the sub-scores) enables evaluation of the overall risk of malnutrition as follows: 0 = low, 1 = medium and 2 = high.
Gastrointestinal dysautonomia symptoms of nutritional interest
The presence of dysphagia was assessed through the administration of the swallowing disturbance questionnaire(Reference Evatt, Chaudhuri and Chou8, Reference Manor, Giladi and Cohen21). The scale includes fourteen items, each with a score ranging from 0 to 3, except one question with a score ranging from 0·5 to 2·5. It results in a total score ranging from 0·5 to 44·5. A cut-off score ≥ 11 has been suggested for the diagnosis of dysphagia needing treatment. However, a score >0·5 already means that the patients experience some degree of difficulty in swallowing. One of the questions of the scale regards dysphagia to liquids, with a score ranging from 0 (absence of dysphagia to liquids) to 3 (severe dysphagia to liquids); a score ≥ 2 has been arbitrarily considered as threatening. The sialorrhoea clinical scale for PD(Reference Evatt, Chaudhuri and Chou8, Reference Perez Lloret, Pirán Arce and Rossi22) was used for the evaluation of sialorrhoea. The scale consists of seven questions, each with a score ranging from 0 to 3, resulting in a total score of 0–21. A cut-off score >11 has been suggested for the diagnosis of sialorrhoea needing treatment; however, a score higher than 0 already means that the patient is not able to control salivation normally. Finally, Rome III criteria (constipation module) were used for the assessment of the presence of chronic constipation. According to these criteria, patients are considered to have functional constipation if they score ≥ 2 points answering the questionnaire(23). At present, no officially recommended assessment scales exist for PD patients, but this tool is commonly used in gastroenterology and has recently been used in this patient population(Reference Kaye, Gage and Kimber24).
Statistical analysis
The data of the patients were used in an anonymous format with a progressive number and collected in a password-protected database. All statistical analyses were carried out using the software MEDCALC® for Windows, version 11.3.0.0 (MedCalc Software). The prevalence of malnutrition risk was computed together with its exact 95 % binomial CI (95 % CI). Continuous variables were reported as mean and standard deviation or median and interquartile range (25th–75th percentile) and categorical variables as counts and percentages. Group comparisons were performed using Fisher's exact test (categorical variables) and Student's t test or the Mann–Whitney U test (continuous variables) when appropriate, according to normal distribution. Afterwards, multiple logistic regression analysis of non-collinear variables was performed to investigate the parameters independently associated with nutritional risk according to the MUST (score ≥ 1). In these models, we decided to include only variables having a P < 0·25 at univariate analyses. Accordingly, data were reported as OR and their respective 95 % CI. All statistical analyses were performed by setting the level of significance at a two-tailed P < 0·05.
Results
General features
From January 2011 to October 2011, 208 PD patients (32·2 % females) were evaluated consecutively. Most of them were taking levodopa (89·3 %), either alone (33·1 %) or in combination with other therapies. According to WHO criteria(13), the distribution of patients among nutritional status categories was as follows: underweight (BMI < 18·5 kg/m2), 1 %; normal weight (BMI 18·5–24·9 kg/m2), 33·0 %; overweight (BMI 25–29·9 kg/m2), 46·4 %; and obesity (BMI ≥ 30 kg/m2), 19·6 %.
Nutritional risk by Malnutrition Universal Screening Tool and related features
Nutritional risk (MUST score ≥ 1) was observed in thirty-six patients, with a prevalence of 17·2 (95 % CI 12·1, 24·0) %. Severe nutritional risk (MUST score ≥ 2) was recorded in ten patients (5 %) and relied mainly on WL (>10 % in seven patients). The clinical features of the population by nutritional risk are summarised in Table 1. Patients at nutritional risk presented with lower BMI, reported WL more frequently, but none reported the presence of acute disease. Moreover, they were more likely to have more severe symptoms, as assessed by the Hoehn–Yahr staging system, to experience sialorrhoea and to require higher levodopa dosages than those with a normal nutritional status.
Table 1 Demographic and clinical features of the study population by nutritional risk (Malnutrition Universal Screening Tool Score ≥1) (Mean values and standard deviations; number of counts and percentages)
UPDRS, unified Parkinson's disease rating scale (part II, activity of daily living; part III, motor examination); SCS-PD, sialorrhoea clinical scale for Parkinson's disease; SQD score, swallowing disturbance questionnaire.
* Percentages are calculated within single groups.
† Variables were compared between groups with Student's t test, the Mann–Whitney U test or Fisher's exact test as appropriate.
Gastrointestinal dysautonomia symptoms and related features
In the overall population, the frequency of sialorrhoea, dysphagia, dysphagia to liquids and constipation was 10·6, 11·0, 14·4 and 59·6 %, respectively. However, the proportion of patients reporting at least one symptom indicating difficulties in swallowing and some degree of sialorrhoea was 58·2 and 55·8 %, respectively. The clinical features of the population are presented by complication in Table 2. With the exception of constipation, dysautonomia symptoms were significantly associated with each other and with more severe symptoms, either assessed by the unified Parkinson's disease rating scale or the Hoehn–Yahr staging system. Patients with sialorrhoea were also characterised by longer disease duration. Higher levodopa dosages were also observed in patients with sialorrhoea and dysphagia.
Table 2 Clinical features of the study population by presence of gastrointestinal dysautonomia symptoms (Mean values and standard deviations; counts and percentages)
SCS-PD, sialorrhoea clinical scale for Parkinson's disease; SQD score, swallowing disturbance questionnaire; UPDRS, unified Parkinson's disease rating scale (part II, activity of daily living; part III, motor examination); MUST, malnutrition universal screening tool.
* Continuous and categorical variables were compared between groups with Student's t test, the Mann–Whitney U test or Fisher's exact test as appropriate
† Percentages are calculated within single groups.
‡ P< 0·02.
§ P< 0·001.
∥ P< 0·05.
¶ P< 0·01.
Predictors of nutritional risk
Finally, we evaluated the independent predictors of nutritional risk by logistic regression analyses (Table 3). In multivariable models, nutritional risk was significantly associated with more severe disease, as assessed according to Hoehn–Yahr stage, levodopa dosage and the number of non-motor complications. No single dysautonomic symptom was independently associated with nutritional risk.
Table 3 Associations with nutritional risk (Malnutrition Universal Screening Tool score ≥1): logistic regression analyses of non-collinear variables (Odds ratios and 95 % confidence intervals)
UPDRS, unified Parkinson's disease rating scale (part II, activity of daily living; part III, motor examination); SCS-PD, sialorrhoea clinical scale for Parkinson's disease; SQD score, swallowing disturbance questionnaire.
* Variables arbitrarily included were those having a P< 0·25 at univariate analyses.
† OR per additional symptom (SCS-PD score >11, SQD total score >11, SQD score for liquids ≥ 2 and presence of constipation).
Discussion
The present study showed that the prevalence of nutritional risk in PD patients hospitalised on a scheduled basis is about half of that of the general hospital population. The present data are in agreement with those of previous studies investigating the prevalence of nutritional derangements by means of multidimensional screening tools(Reference Barichella, Villa and Massarotto5, Reference Wang, Wan and Cheng10, Reference Jaafar, Gray and Porter25). In respect to this issue, the present study emphasises the importance of using these assessment tools, because the large variation in prevalence across previously published studies appears to be due to differences in nutritional assessment methods and in the definition of malnutrition(Reference Sheard, Ash and Silburn9). Several reports, including the present, have reported that the prevalence of overweight and obesity is about twice as high as in the general Italian population(Reference Barichella, Cereda and Pezzoli4, Reference Cereda, Cassani and Barichella26). This has been explained as a consequence of the changes in lifestyle (e.g. reduction in physical activity) and the important improvements in the therapeutic management (pharmacological and/or surgical) of the disease achieved in the last decades(Reference Barichella, Cereda and Pezzoli4). Accordingly, using only BMI in the nutritional screening or assessment of the PD patient is a limitation, which could lead to a significant underestimation of nutritional derangements. Also body weight excess deserves attention from a nutritional standpoint due to its association with morbidity and mortality(13). However, PD patients appear apparently protected from CVD due to a more favourable metabolic profile, which seems to be independent of nutritional status and total and abdominal adiposity(Reference Barichella, Cereda and Pezzoli4, Reference Cereda, Cassani and Barichella26, Reference Cereda, Cassani and Barichella27).
The key feature of nutritional risk in the present study was WL, a finding in agreement with a previous report(Reference Barichella, Cereda and Pezzoli4, Reference Chen, Zhang and Hernán28). The present prevalence data have important implications for clinical practice, because the risk of malnutrition in elderly patients has been related to the setting and the level of assistance, with significantly lower rates being reported in outpatient populations(Reference Cereda3). However, the lower prevalence of nutritional risk, compared with the general hospital population(Reference Norman, Pichard and Lochs1, Reference Cereda, Lucchin and Pedrolli2, Reference Stratton, Hackston and Longmore18–Reference Schindler, Pernicka and Laviano20), reasonably reflects the pathophysiology of nutritional risk. The present study population included outpatients who were not hospitalised for acute conditions frequently characterised by an inflammatory background. This is supported also by other data because biochemical protein markers of nutritional status associated with inflammation(Reference Omran and Morley17), particularly albumin, were unrelated to MUST score. Along with this, it has been shown that the reduction in body weight in PD patients occurs mainly at the expense of fat mass rather than lean body mass(Reference Barichella, Cereda and Pezzoli4, Reference Chen, Zhang and Hernán28), a body compartment that is generally affected by inflammatory diseases(Reference Norman, Pichard and Lochs1). In respect with the setting of provenience, the present study agrees with that by Jaafar et al. (Reference Jaafar, Gray and Porter25). On the other hand, although the MUST was primarily developed for the community setting and its use in hospitals has been supported by several studies(Reference Norman, Pichard and Lochs1, Reference Stratton, Hackston and Longmore18–Reference Schindler, Pernicka and Laviano20), the present observations suggest raising questions about the appropriateness of this tool in PD outpatients. Given the prevalence of overweight/obesity in PD patients and the prevalent absence of acute disease, it is possible that the MUST may have not correctly identified all those at risk. In hospital inpatients, prevalence of nutritional risk by MUST is reported to be up to 60 % and in outpatients up to 30 %(Reference Stratton, Hackston and Longmore18). All these aspects are of importance because screening represents the first step in nutritional assessment and further investigation is warranted when risk is detected.
Previous studies have attempted to explain the causes of WL in PD patients, and available evidence suggests that both changes in energy intake and expenditure contribute to this process(Reference Barichella, Cereda and Pezzoli4). Increased energy requirements have been associated with motor function deterioration, particularly with muscle rigidity and levodopa-induced dyskinesia(Reference Barichella, Cereda and Pezzoli4, Reference Cereda, Pezzoli and Barichella6). Moreover, a stimulating effect of levodopa on energy metabolism has been proposed(Reference Kashihara29). The present results are consistent with previous findings, as levodopa doses and severity of disease were independently associated with nutritional risk, thus supporting the belief that malnutrition is intrinsically related to the course of the disease and levodopa-related complications(Reference Barichella, Cereda and Pezzoli4, Reference Cereda, Pezzoli and Barichella6). Akinetic state, as reflected by more severe disease, could also result in impaired food intake due to poor hand–mouth coordination. It has been suggested that the reduction in energy intake may depend on other factors, such as reduced food interest secondary to hyposmia, psychosocial factors (e.g. mood disorders or difficulties in preparing meals secondary to disability), as well as gastrointestinal dysfunction (e.g. swallowing disturbances)(Reference Barichella, Cereda and Pezzoli4, Reference Bachmann and Trenkwalder7, Reference Kashihara29). Unfortunately, in the present study, we did not assess energy intake due to technical difficulties. Patients admitted to our Institute came from the whole Italian country and quantitative dietary assessments by means of food diaries could not be planned, while the use of the 24-h dietary recall may result in over-or underestimation of daily energy intake. However, our research group has previously demonstrated that nutritional status deterioration by the mini nutritional assessment was mainly related to those questions focusing on dietary habits(Reference Barichella, Villa and Massarotto5). On the other hand, we found that swallowing disturbances, which are generally associated with reduced food intake, were unrelated to nutritional risk. Protein redistribution diets could contribute to WL in PD patients(Reference Cereda, Barichella and Pedrolli30). However, although neurologists should recommend the adherence to this dietetic regimen on an individualised basis to those experiencing motor fluctuations, many PD patients have a dietary pattern similar to that of the general population(Reference Marczewska, De Notaris and Sieri31).
In contrast with previous evidence collected in patients with other diseases, in the present survey, nutritional risk was not associated with reduced activities of daily living(Reference Cereda3). Moreover, only the number of dysautonomia symptoms was significantly correlated with nutritional risk, whilst no independent effect for single symptoms was demonstrated. Although it can be argued that difficulty in swallowing solid food due to oropharyngeal dysfunction may increase WL, mainly in the advanced rather than in early stages of PD(Reference Barichella, Cereda and Pezzoli4, Reference Bachmann and Trenkwalder7), in the present study, neither swallowing difficulties nor nutritional risk were related to duration of PD.
There is only one report, as far as we are aware, investigating the association between malnutrition and gastrointestinal dysautonomia symptoms other than dysphagia. Wang et al. (Reference Wang, Wan and Cheng10) found that constipation predicted nutritional risk. Someone could argue that the lack of association with single non-motor complications depends on the method used for the assessment. Indeed, clinical rating scales are simple instruments to be used for screening purposes, leaving the clinician the possibility to continue the diagnostic work-up with more complex and expensive assessments, should complications be suspected. Nonetheless, the prevalence of dysautonomic symptoms detected in the present study using clinical rating scales was similar to the figures previously reported in the literature (dysphagia, 30–80 %; sialorrhoea, 70–75 %; and constipation, 60 %)(Reference Barichella, Cereda and Pezzoli4, Reference Kaye, Gage and Kimber24–Reference Adler32) and symptoms were correlated to each other. Moreover, taken together, these symptoms were associated with nutritional risk, to which they probably contribute to different extents. Sialorrhoea consists of the alteration of salivation control associated with unpleasant drooling. In PD patients, it is more likely to be secondary to swallowing abnormalities than to excessive salivary production(Reference Evatt, Chaudhuri and Chou8, Reference Adler32), and the association found with other symptoms of autonomic nervous system dysfunction supports this hypothesis. Sialorrhoea may be responsible for discomfort or social embarrassment, thus inducing the patient to change dietary habits, which, in turn, could be also negatively affected by swallowing difficulties or abdominal distension due to constipation.
A potential inclusion bias should be taken into account. Patients with severe disability and/or cognitive impairment were more likely to be excluded. Accordingly, the prevalence of nutritional risk, as well as the contribution of these conditions in the overall population of PD patients, may have been partly underestimated. However, our data are likely to be representative of the PD outpatient population, because patients with severe disability and/or cognitive impairment are frequently admitted to long-term care facilities.
Finally, depression was not included among the assessments. This limitation is also recognised because Wang et al. (Reference Wang, Wan and Cheng10) have recently demonstrated a significant association with nutritional risk.
In conclusion, the present paper shows that, in PD patients, the nutritional domain should be object of periodic evaluations during the course of the disease. In respect of this, attention should be focused on several factors and through a multidisciplinary approach because the heterogeneous changes potentially occurring seem to be of multifactorial origin. This approach could allow the introduction of effective dietary treatment promptly. Nonetheless, the use of other screening tools along with the implications of nutritional risk on patient outcome should be addressed by future studies.
Acknowledgements
The authors wish to thank Jennifer S. Hartwig, MD, for editorial assistance. All the authors certify that there are no affiliations with or involvement in any organisation or entity with a direct financial interest in the subject matter or materials discussed in the manuscript. E. Cereda has received consultancy honoraria and investigator grants from Nutricia Italia and the ‘Fondazione Grigioni per il Morbo di Parkinson’. All authors significantly contributed to the work, and read and approved the final version of the manuscript. All authors, external and internal, had full access to all of the data (including statistical reports and tables) in the study and can take responsibility for the integrity of the data and the accuracy of the data analysis. In particular, contributions were as follows: M. B. contributed to the study design, data interpretation, review and critique. E. Cereda performed the data analysis, data interpretation and drafting of the manuscript. C. M., L. I. and C. P. carried out data collection. R. Cancello was involved in data collection and data interpretation. R. Caccialanza carried out data interpretation. G. P. was involved in the supervision, review and critique. Erica Cassani aided in study design, data collection, and drafting of the manuscript. The present study was supported by the ‘Fondazione Grigioni per il Morbo di Parkinson’.
