A sarcopenia em doentes hospitalizados: estudo do seu impacto clínico e económico e do método antropométrico

Porto, 2015 D 2015 A sarcopenia em doentes hospitalizados: estudo do impacto clínico e económico e do método antropométrico Tese de Doutoramento apresentada à Faculdade de Ciências da Nutrição e Alimentação da Universidade do Porto Ana Sofia Sousa ii iii A sarcopenia em doentes hospitalizados: estudo do impacto clínico e económico e do método antropométrico Sarcopenia among hospitalized patients: clinical and financial impact and anthropometric method Ana Sofia Sousa Porto | 2015 Dissertação de candidatura ao grau de Doutor apresentada à Faculdade de Ciências da Nutrição e Alimentação da Universidade do Porto Orientadora: Prof. Doutora Teresa Maria de Serpa Pinto Freitas do Amaral Professora Associada Faculdade de Ciências da Nutrição e Alimentação da Universidade do Porto iv v Lista de Publicações Fazem parte desta dissertação a seguinte publicação e manuscritos submetidos para publicação: 1. Sousa AS, Guerra RS, Fonseca I, Pichel F, Amaral TF. Sarcopenia among hospitalized patients - A cross-sectional study. Clin Nutr. 2014; 27 (14):00313-6.doi.org/10.1016/j.clnu.2014.12.015. 2. Sousa AS, Guerra RS, Fonseca I, Pichel F, Amaral TF. Sarcopenia and length of hospital stay [submitted for publication]. 3. Sousa AS, Guerra RS, Fonseca I, Ferreira S., Pichel F, Amaral TF. Financial impact of sarcopenia on hospitalization costs [submitted for publication]. 4. Sousa AS, Fonseca I, Pichel F, Amaral TF. Effects of posture and body mass index on body girths assessment [submitted for publication]. 5. Sousa AS, Fonseca I, Pichel F, Amaral TF. Triceps skinfold compressibility in hospitalized patients – an exploratory analysis [submitted for publication]. Colaborei na definição dos objetivos, na recolha, na análise e na interpretação dos dados de todos os manuscritos e do artigo. Fui responsável pela redação da versão inicial de todos os manuscritos e colaborei ativamente na elaboração de todas as versões do artigo. vi vii Agradecimentos À Professora Teresa Amaral. Por todo o encorajamento, preciosa ajuda e orientação ao longo deste percurso. Por me ter sentido sempre apoiada. Pelo exemplo de profissionalismo, dedicação e rigor. Pela disponibilidade com que acolheu e respondeu a todas as minhas dúvidas. Muito obrigada por tudo. Ao Dr. Fernando Pichel, pelo papel fundamental na minha integração na equipa de investigação e no Serviço de Nutrição e Alimentação do Centro Hospitalar do Porto, onde pude implementar e desenvolver o meu trabalho. Por toda a ajuda, de forma pronta e disponível, sempre que necessitei. À Prof. Doutora Isabel Fonseca, pela disponibilidade com que acolheu o meu projeto e pelo seu importante papel no seu desenvolvimento. Por ter participado ativamente na redação de todos os manuscritos e artigo desta tese, pelas críticas construtivas e comentários, que ajudaram sempre a melhorar cada um dos trabalhos. À Rita Guerra, por todo o apoio, essencial no desenvolvimento de todas as etapas do meu trabalho, por toda a colaboração e disponibilidade. viii À Dr.a Susana Ferreira, pela simpatia e disponibilidade com que colaborou no fornecimento dos dados referentes aos custos hospitalares. Ao Eng.º Tiago Andrade, pelo apoio, simpatia e disponibilidade com que orientou os cálculos referentes à compressibilidade das pregas cutâneas. Às colegas do Serviço de Nutrição e Alimentação do Centro Hospitalar do Porto, pela simpatia com que fui acolhida, pela disponibilidade e pela pronta colaboração, sempre que necessitei de ajuda. A todos as pessoas que aceitaram participar neste estudo, pois sem a sua colaboração, nada seria possível. Aos meus pais, pelo vosso apoio, ao longo de toda a minha vida. Porque são um exemplo para mim e porque vos devo tudo o que sou. Por me fazerem acreditar em mim. Muito obrigada por acolherem todas as minhas inseguranças, pela compreensão e palavras de encorajamento. À minha irmã, Francisca, por tudo o que representas para mim. Pela companhia, compreensão e pelas palavras certas, sempre que necessito. Ao João. Obrigada por partilhares a tua vida comigo e me acompanhares em cada etapa. Obrigada pela compreensão em todos os momentos. O teu apoio é fundamental em tudo o que faço. ix Abstract According to the European Working Group on Sarcopenia in Older People (EWGSOP), sarcopenia is a condition defined as the loss of both muscle mass and muscle function (strength or performance). Sarcopenia is a complex phenomenon and there are several factors associated with the development of this condition. Sarcopenia is diagnosed using muscle mass, strength and physical performance data. There is lack of standardization of the diagnostic procedures, but this classification can be improved. Nevertheless, prevalence of sarcopenia varies widely depending on the methodology used for its diagnosis. Moreover, although sarcopenia is mainly considered as a geriatric syndrome, this condition has been also reported as being present in younger adults. However, the problem of sarcopenia among hospitalized younger adults remains to be documented. Sarcopenia has been previously described as being related with poor clinical outcome in hospitalized patients. However, in regards to the association of sarcopenia with length of hospital stay, information is scarce and controversial. Moreover, data on financial burden of sarcopenia in hospital setting is limited to surgical patients. Concerning the assessment of hospitalized patients, anthropometry provides useful information on body composition and is of utmost importance in nutrition risk screening and evaluation. Even though anthropometric measures, such as body circumferences and skinfold thickness, are not considered by the EWGSOP as being suitable for routine use in clinical practice, they still are amongst the most relevant methods for body composition assessment due to their predictive value and their practicability. Therefore, reducing potential sources of error in body xvi cutâneas, não sejam consideradas pelo EWGSOP como adequadas para utilização por rotina na prática clínica no diagnóstico da sarcopenia, estão entre os métodos mais relevantes para a avaliação da composição corporal, devido ao seu valor preditivo e fácil aplicabilidade. Assim, a redução de potenciais fontes de erro na avaliação de perímetros corporais e de pregas cutâneas, poderá ser vantajosa, uma vez que poderá permitir melhorar a validade da antropometria. A presente tese tem como objetivo avaliar a frequência da sarcopenia e estudar o impacto de diferentes critérios de diagnóstico (artigo 1). Este trabalho teve também como objetivo identificar os fatores associados com a sarcopenia e aumentar o conhecimento acerca do seu impacto clínico (artigo 2) e económico (artigo 3) em doentes hospitalizados. Foram ainda objetivos deste trabalho explorar potenciais fontes de erro na avaliação de perímetros corporais (artigo 4) e de pregas cutâneas (artigo 5). Os primeiros três estudos que compõem esta tese foram conduzidos numa amostra consecutiva de, respetivamente, 608, 655 e 656 doentes hospitalizados com idade igual ou superior a 18 anos. Os estudos 4 e 5 foram conduzidos numa amostra de conveniência de 123 e 106 doentes, respetivamente. Os estudos 1,4 e 5 são transversais, enquanto os estudos 2 e 3 são prospetivos. Os resultados do artigo 1 mostram que 25,3% dos doentes avaliados encontravam-se sarcopénicos. No entanto, dependendo da idade e do critério aplicado, a frequência da sarcopenia variou de 5% a 41,1% nos homens e de 4,9% a 38,3% nas mulheres. Encontrou-se uma concordância de 95,7% (k = 0,89) entre os critérios que estimaram a massa muscular através de impedância bioelétrica. xvii De acordo com os critérios do EWGSOP, aproximadamente 20% dos doentes não desnutridos encontravam-se sarcopénicos e 29,5% dos doentes com excesso de peso e 18,7 % dos doentes obesos estavam também sarcopénicos. Além disso, 19,8% dos doentes com idade compreendida entre os 18 e os 64 anos encontravam-se sarcopénicos. Os fatores associados com a presença de sarcopenia (artigo 2) são o sexo masculino, idade igual ou superior a 65 anos, dependência moderada ou grave, desnutrição e admissão hospitalar num serviço médico. Os doentes sarcopénicos apresentam menor probabilidade de ter alta do hospital (Hazard Ratio (HR); Intervalo de Confiança (IC) 95% = 0,71; 0,58-0,86). No entanto, após estratificação por grupos etários, este efeito foi visível apenas nos doentes com idade compreendida entre os 18 e os 64 anos (HR; IC 95% = 0,66; 0,51-0,86). Além disso, doentes com excesso de peso sarcopénico ou obesidade sarcopénica apresentaram maior probabilidade de ter alta (HR; IC 95%= 0,78; 0,61-0,99) do que os doentes sarcopénicos sem excesso de peso ou obesidade (HR; IC 95% = 0,63; 0,48-0,83). O artigo 3 mostra que a sarcopenia aumenta de forma independente os custos de hospitalização em €1240 (IC 95%: €596-1887) nos doentes com idade compreendida entre 18 e 64 anos e em €721 (IC 95%: €13-1429) nos doentes com 65 ou mais anos. O excesso de peso sarcopénico associou-se a um aumento independente de €884 (IC 95%: €295-1476) dos custos de hospitalização. Quanto aos efeitos da postura e do índice de massa corporal (IMC) na avaliação de perímetros corporais (artigo 4), os perímetros corporais obtidos na posição ortostática e em decúbito dorsal foram comparados de acordo com duas xviii categorias de IMC: peso normal e excesso de peso ou obesidade. Encontraram-se diferenças significativas entre 0,6 e 1,1 cm, entre as medições obtidas em posição ortostática e em decúbito dorsal. Os valores dos coeficientes de correlação intraclasse foram ≥0,97 e a concordância variou de 81,3% a 87% (k ponderado ≥0,84). Obtiveram-se resultados semelhantes quando as diferenças foram estratificadas por categorias de IMC. Com o artigo 5 pretendeu-se explorar e descrever, através de uma técnica recentemente desenvolvida, a compressibilidade da prega cutânea tricipital (PCT) e os seus fatores associados numa amostra de doentes hospitalizados. A compressibilidade foi determinada de acordo com uma definição baseada numa medida de tempo (τ) que reflete a resposta dinâmica do tecido adiposo à compressão, e foi também definida de acordo com a diferença entre o valor inicial e o valor final da PCT (diferença de PCT). Resultados provenientes de modelos de regressão linear mostraram que o tempo de compressibilidade (τ) não se encontra significativamente associado com nenhuma das variáveis incluídas. Contudo, a compressibilidade baseada na diferença de valores da PCT encontra-se independentemente associada com a espessura da PCT (coeficiente de regressão (intervalo de confiança a 95%) = 0,38 (0,01-0,05), p=0,002) e com o estado nutricional (coeficiente de regressão (intervalo de confiança a 95%) = 0,23 (0,12-1,23), p=0,018). Os resultados do presente trabalho conduzem às seguintes conclusões: (1) a sarcopenia é frequente em doentes hospitalizados e esta frequência apresenta grande variação, dependendo do critério de diagnóstico aplicado; identificou-se a sarcopenia numa proporção considerável de doentes adultos com idade xix compreendida entre os 18 e os 64 anos e também em doentes não desnutridos, com excesso de peso e em obesos; (2) ser do sexo masculino, com idade igual ou superior a 65 anos, apresentar dependência, estar desnutrido e ser hospitalizado num serviço médico são fatores associados à presença de sarcopenia em doentes adultos hospitalizados; a sarcopenia encontra-se independentemente associada a maior tempo de internamento, embora esta associação seja mais forte em doentes com idade compreendida entre os 18 e os 64 anos; o excesso de peso sarcopénico relaciona-se com maior probabilidade de ter alta do que a sarcopenia não associada com excesso de peso; (3) a sarcopenia associa-se de forma independente com os custos de hospitalização, aumentando-os em 52,7% (58,5% em doentes com idade inferior a 65 anos e 34% em doentes com 65 ou mais anos); (4) a avaliação de perímetros corporais em posição ortostática e em decúbito dorsal em doentes hospitalizados difere mas estas diferenças são pequenas e não dependem das categorias de IMC; (5) numa amostra de doentes hospitalizados, o tempo de compressibilidade (τ) não se encontrava associado com nenhum dos fatores estudados. No entanto, o risco de desnutrição e a espessura da PCT são fatores independentemente associados com um aumento da compressibilidade, definida como a diferença entre o valor inicial e o valor final de PCT. Embora se trate de uma análise exploratória para descrever a compressibilidade e os seus efeitos, estes resultados realçam a necessidade de mais investigação de forma a determinar o método mais preciso para quantificar a compressibilidade, para inferir sobre os fatores associados e controlar o seu efeito. Os resultados da presente tese aumentaram o conhecimento acerca do problema e dos critérios de diagnóstico da sarcopenia numa perspetiva clínica e xx demonstraram que esta condição não é meramente uma síndrome geriátrica em contexto hospitalar. Além disso, o reconhecimento dos pequenos erros associados à postura e à compleição física na avaliação de perímetros corporais e o avanço na possibilidade de quantificar a compressibilidade de pregas cutâneas poderão ser indicações úteis no que concerne à utilização de medições antropométricas. xxi Index Abstract ........................................................................................................... ix Resumo .......................................................................................................... xv Abbreviations .............................................................................................. xxiii Introduction ..................................................................................................... 1 1. Sarcopenia ............................................................................................... 3 1.1. Definition: definition and diagnosis ..................................................... 3 1.2. Prevalence ......................................................................................... 4 1.3. Etiology .............................................................................................. 5 1.3.1. Sarcopenia versus cachexia ......................................................... 7 1.3.2. Sarcopenia versus frailty .............................................................. 7 1.3.3. Sarcopenic obesity ....................................................................... 8 1.3.4. Sarcopenia and associated factors .............................................. 9 1.4. Diagnosis ......................................................................................... 10 1.4.1. Methodology ............................................................................... 10 1.4.2. Definition of cut-offs for diagnostic measurements ..................... 11 1.5. Sarcopenia amongst hospitalized patients ....................................... 12 1.5.1. Clinical impact ............................................................................ 12 1.5.2. Financial impact ......................................................................... 13 2. Anthropometry and body composition assessment ................................ 14 2.1. Assessment of body circumferences: the effect of posture and body complexion .............................................................................. 15 2.2. Skinfold thickness measurement: the effect of compressibility ......... 17 xxii Aims .............................................................................................................. 21 Chapter I ........................................................................................................ 25 Chapter I.a .................................................................................................. 27 Chapter I.b .................................................................................................. 35 Chapter I.c .................................................................................................. 69 Chapter II ....................................................................................................... 95 Chapter II.a ................................................................................................. 97 Chapter II.b ............................................................................................... 121 General discussion ...................................................................................... 143 Future perspectives ..................................................................................... 151 Conclusions ................................................................................................. 155 References .................................................................................................. 159 xxiii Abbreviations BIA, bioelectrical impedance analysis BMI, body mass index EWGSOP, European Working Group on Sarcopenia in Older People HGS, handgrip strength LOS, length of hospital stay MAMC, mid-arm muscle circumference NRS-2002, Nutritional Risk Screening PG-SGA, Patient-Generated Subjective Global Assessment TSF, triceps skinfold WHO, World Health Organization xxiv 1 Introduction 8 Unintended weight loss, exhaustion, weakness, slow gait speed and low physical activity are the features that support a frailty diagnosis (42). Frailty and sarcopenia overlap as most frail older individuals are sarcopenic, and there are older people with sarcopenia who are also frail (3). Notwithstanding this, the concept of frailty goes beyond physical factors and includes also psychological and social dimensions such as cognitive status, social support and other environmental factors (41). 1.3.3. Sarcopenic obesity The co-occurrence of sarcopenia with increased fat mass is defined as sarcopenic obesity, which may carry the cumulative risk derived from each of the two conditions (43, 44). Excess adiposity on its own may generate significant adverse health effects such as hypertension, dyslipidaemia and insulin resistance and there is increasing evidence showing that these risks can be elevated by the addition of low muscle mass (45). Sarcopenic obesity is often observed in malignancy, rheumatoid arthritis and ageing as in this conditions lean body mass is lost while fat mass may be preserved or even increased (46). Sarcopenic obesity has been associated with aggravation on mobility dysfunction, higher dependence in activities of daily living among community dwelling individuals and with higher risk of co-morbidities in hospitalized patients (45). The association between age-related reduction of muscle mass and strength is, in most cases, independent of body mass (47). It had long been described 9 that age-related loss of weight, along with muscle mass, was the main responsible for muscle weakness in older individuals (47). However, it is now clear that muscle composition has influence on muscle quality and function, as the infiltration of fat into muscle, also known as “marbling”, lowers its quality and performance (48). 1.3.4. Sarcopenia and associated factors Several factors have been described to be associated with the presence of sarcopenia among hospitalized and also community dwelling older adults older adults, such as age (49), schooling years (50), smoking (49) and physical activity (11), BMI (11), hormonal factors (50) and hospitalization at a medical ward. Notwithstanding this, according to current knowledge, data on factors associated with sarcopenia among hospitalized younger patients, aged 18 to 64 years, are non-existent and remain to be documented. The research for factors associated with the presence of sarcopenia is of major importance, as it may allow for the recognition of modifiable risk factors. Since modifiable factors are identified, it will be possible to act on prevention. 10 1.4. Diagnosis 1.4.1. Methodology Muscle mass, strength and physical performance are the variables to assess in order to identify sarcopenia according to the EWGSOP (3). It is important to use the most accurate method to evaluate these variables. However, the EWGSOP (3) proposes different possible techniques, considering their characteristics and suitability for use either in research or clinical practice. For muscle mass assessment, imaging techniques, such as computed tomography (CT) scans and magnetic resonance imaging (MRI), are very precise and are able to distinguish fat from other soft tissues, features that make these methods gold standards for estimating muscle mass (3). However, high financial costs and radiation exposure limits the use of these whole-body imaging methods in clinical practice (51). Dual energy X-ray absorptiometry (DXA) is considered the preferred alternative method for research and clinical practice. Nevertheless, lack of portability may be a limitation for clinical use (51). Bioelectrical impedance analysis (BIA) is an economic, easy to use, reproducible technique which used under standard conditions have been found to correlate well with MRI predictions (52). Therefore, this technique is considered a valid alternative to DXA (3). Other muscle mass estimation methods include anthropometric measures such as mid-arm circumference and skinfold thickness. However, changes in adiposity that occur with ageing and the decline of skin elasticity contribute to errors 11 of estimation in older people, making anthropometric measures vulnerable to error (53). The EWGSOP does not recommend anthropometry for routine use in the diagnosis of sarcopenia (3). Considering muscle strength evaluation, handgrip strength is considered the best and most suitable technique. This measurement has been strongly associated with lower extremity muscle power, knee extension, calf cross-sectional muscle area and is also a clinical marker of mobility (54). For physical performance, the Short Physical Performance Battery (SPPB), a combined method for assessing physical performance is referred as a standard measure (3). This method evaluates balance, gait speed, strength and endurance (55). Gait speed, which is a part of the SPPB, can also be applied as a single test (3). Timed get-up-and-go (TGUG) test which measures the time needed to complete a series of functional task and evaluates balance is another possible examination of performance (56, 57). 1.4.2. Definition of cut-offs for diagnostic measurements Cut-off points depend on the selected technique and the lack of reference studies limits value standardization. Therefore, the European consensus proposes various cut-off points for each recommended method or technique based on the information available, considering the assessment of muscle mass using DXA and BIA, muscle strength (by handgrip strength) or physical performance, assessed by SPPB or gait speed (3). 12 Despite the EWGSOP (3) recommendations concerning the most suitable methods and cut-off points, there is lack of standardization of the diagnostic procedures. Thus, prevalence of sarcopenia varies widely depending on the methodology used to diagnose this condition (58, 59). 1.5. Sarcopenia amongst hospitalized patients 1.5.1. Clinical impact Sarcopenia has been previously described as being related with poor clinical outcome in hospitalized patients. This condition has also been associated with higher mortality (7, 60), higher risk of non-elective readmission in a six month period (7) and worst post-operative outcomes (61-65). Length of hospital stay (LOS) is a widely used indicator of the changes that occur during a hospitalization process and can be used as a surrogate marker of health status (66). Thus, prediction of LOS may lead to a maximization of resources (67). Concerning the association of sarcopenia with LOS, information is scarce and controversial. Results from a study conducted among hospitalized patients aged ≥65 years showed that sarcopenic patients with a mean age of 79 years had longer LOS than non-sarcopenic patients (7). In contrast, no differences in LOS between sarcopenic and non-sarcopenic older patients with a mean age of 84.2 years were reported by Cerri et al. (9). Moreover, a study conducted among cancer patients, with sarcopenia defined through computed tomography scans, reported sarcopenic 13 patients to present longer LOS than non-sarcopenic patients (39 days versus 30 days, p<0.001) (61). The establishment of an association between sarcopenia and LOS is, therefore, of utmost importance in order to provide a more effective healthcare plan and reduce adverse consequences. The potential effect of confounding factors on the association between sarcopenia and LOS remains to be described, as well as the quantification of the association of sarcopenia with LOS among a wide-ranging sample of hospitalized patients. Increasing the knowledge on this subject could be a major advantage for clinical setting, in order to provide a more effective healthcare plan and thus reducing the adverse consequences that sarcopenia entails. 1.5.2. Financial impact Considering the previously described impact of sarcopenia on hospitalized individuals, healthcare costs of this condition are expected to be high (68). Notwithstanding this, data on the financial burden of sarcopenia are limited. One study from 2004 (69), which was conducted among representative samples of American adults aged ≥ 60 years, reported that the estimated healthcare cost attributable to sarcopenia defined merely as the loss of muscle mass was $18.5 billion, being $10.8 billion in men and $7.7 billion in women. Nevertheless, although sarcopenia has been previously associated to higher hospitalization costs, this 14 information is still limited to surgical patients (70-72). Thus, the impact of sarcopenia on hospitalization costs among a wider group of patients remains to be documented. Considering the adverse consequences sarcopenia entails among hospitalized patients and the financial constraints that healthcare systems often face, in order to maximize resources and provide a more effective healthcare plan, it is important to recognize and to explore the association of sarcopenia with hospitalization costs. 2. Anthropometry and body composition assessment Anthropometry, from the Greek anthropos, man, and metron, measure, is defined as a measure to study human body dimensions and has been used since ancient times (73, 74). This method was originally used, not for science but for artistic purpose: painters and sculptors needed information about human body constitution so that they could make authentic representations (75). Anthropometry was firstly applied for epidemiological studies, specifically for nutritional assessment and its association with clinical outcomes, in the second half of the twentieth century (76). Anthropometric measures provide useful information on body composition assessment and the techniques used are non-invasive, economic and easy-to-use both for clinical practice and research purposes (77-79). These measures are of utmost importance in nutrition risk screening and assessment, particularly in hospital settings, as undernutrition is an important predictor of poor prognosis and longer length of hospital stay (80, 81). 15 Anthropometric measures are not currently recommended by the EWGSOP as a method for sarcopenia diagnosis. However, as mentioned above, besides being universally used, anthropometry is often the only available method for professionals to evaluate body composition and nutrition status in clinical settings. Therefore, even though anthropometric measures are not considered by this European Consensus as being suitable for routine use in clinical practice, they still are amongst the most relevant methods for body composition assessment. Thus, concerning the identification of sarcopenia, anthropometric measurements can still be applied for screening purposes and, therefore, be used as complimentary methods to the recommended techniques. Body circumferences and skinfold thickness are amongst the most widely used measures, both for research and clinical practice, due to their predictive value and association with a variety of conditions, in addition to their usefulness for predicting other relevant anthropometric measurements, such as weight and height when their measurement is impossible to obtain. 2.1. Assessment of body circumferences: the effect of posture and body complexion For the assessment of body circumferences, protocols currently used (82) indicate supine position as the correct position for performing measurements. However, especially in hospital settings, individuals are often unable to change their 16 body position (83, 84). For instance, to evaluate a critical patient or a bedridden patient, it is necessary to adapt the standard procedure to the body position. Body circumference results, namely arm, waist, hip and calf girths, are frequently used for body composition and nutritional assessment purposes, isolated or as part of undernutrition diagnosis and screening tools (85, 86). The effect of posture on body circumferences assessment could be of major relevance as it can compromise the entire body composition assessment results. Previous reports had shown that body posture influenced anthropometric measurements of the lower limbs in young free-living adults (83) and, more recently, a study conducted among institutionalized and hospitalized older adults aged ≥65 years (87) showed that differences between body circumferences measurements obtained on standing and supine positions did not have clinically relevant impact on nutritional assessment. However, it is not known if age-related differences in body composition can change these results when younger hospitalized individuals are assessed. Nevertheless, information on the effect of posture on body circumferences in hospitalized younger adult patients aged < 65 years, is still lacking. Besides posture, factors such as physical complexion are also susceptible of introducing bias in body circumferences assessment. Anthropometric measurements in obese (or overweight) patients, due to the presence of a larger body size are more susceptible to error, even with a trained anthropometrist. Thus, it not currently known if posture related changes differ when measurements are evaluated in overweight or obese subjects rather than in individuals with normal weight. This possible influence of overweight or obesity in body circumferences 17 assessment needs to be documented, as, in clinical practice, this effect can lead to a misinterpretation of anthropometry and, consequently, a misclassification of nutritional status assessment. 2.2. Skinfold thickness measurement: the effect of compressibility Skinfold thickness data allows for inference on body composition (88). This is a substantially used measurement for monitoring subcutaneous adiposity due to its accessibility and non-invasive nature (89). Moreover, this measurement can be used to predict adiposity and it is integrated, along with mid-arm circumference, in midarm muscle circumference formula, through which it is possible to estimate muscle mass (76). In skinfold thickness measurement using a skinfold calliper, a constant pressure is applied for two seconds (82). Tissue’s dynamic response to this pressure is defined as compressibility and it has been studied using coarse methods based on the comparison between skinfold calliper measurements and subcutaneous fat thickness assessed by imaging methods or cadaver studies (88, 90). Moreover, there are assumptions that underlie the estimation of body fatness, based on skinfold thickness measure: skin thickness is negligible, adipose tissue has constant characteristics and also that proportion of subcutaneous to visceral fat is equivalent in all subjects (88). However, based on the previous studies using empiric comparisons and cadaver studies (89-91), it has been shown that 24 25 Chapter I Sousa AS, Guerra RS, Fonseca I, Pichel F, Amaral TF Sarcopenia among hospitalized patients – A cross-sectional study Clin Nutr. 2014 [published online] Sousa AS, Guerra RS, Fonseca I, Pichel F, Amaral TF Sarcopenia and length of hospital stay Under review Sousa AS, Guerra RS, Fonseca I, Pichel F, Ferreira S, Amaral TF Financial impact of sarcopenia on hospitalization costs Under review 26 27 Chapter I.a Sarcopenia among hospitalized patients – A cross-sectional study 28 29 30 31 32 33 40 Introduction According to the European Working Group on Sarcopenia in Older People (EWGSOP) sarcopenia is defined as a combination of both low muscle mass and low muscle function (1). This condition has been associated with physical disability, low quality of life and higher mortality (1, 2). Sarcopenia is estimated to occur between 5 to 45% of community dwelling older adults (3-5). While this condition is mainly observed in older adults, it can also be present in younger individuals. A study from 2013 by Cherin et al. (6) showed that 9% of the individuals aged between 45 and 54 years and 13.5% of those aged from 55 to 64 years were sarcopenic. Although data concerning sarcopenia in hospitalized patients are scarce, previous studies have described this condition as frequent among hospitalized older patients (7-11), ranging from 10% to 37.3%. Moreover, it has been recently shown that sarcopenia is present in hospitalized patients aged under 65 years, with a frequency equal to 19.8% (10). It has been previously reported that sarcopenia is related with poor clinical outcome in hospitalized older patients, namely higher mortality (7, 9, 12), higher risk of non-elective readmission in a six month period (7) and worst post-operative outcomes (13-16). In a study conducted among hospitalized patients aged ≥65 years (7), sarcopenic patients presenting a mean age of 79 years were reported to have higher length of hospital stay (LOS) than non-sarcopenic patients. In contrast, Cerri et al. (2014) (9) found no differences in LOS between sarcopenic and non-sarcopenic patients, among hospitalized older patients, with a mean age of 84.2 years, ranging 41 from 66 to 100 years. Nevertheless, as far as we are concerned, there are no available data on the impact of sarcopenia on LOS among hospitalized patients aged <65 years. LOS is an indicator of the changes that occur during a hospitalization process and can be used as a surrogate marker of health status (17). Moreover, predicting LOS may lead to a maximization of resources (18). According to our knowledge, data on factors associated with sarcopenia in hospitalized patients are scarce while it is particularly limited among hospitalized younger patients. Moreover, the potential effect of confounding factors on the association between sarcopenia and LOS remains to be described. Identification of sarcopenia and the establishment of an association between this condition and LOS are of utmost importance in order to provide a more effective healthcare plan and thus reducing the adverse consequences this condition entails. This study aims to quantify the association of sarcopenia with LOS, after adjustment for potential confounders and to identify factors associated with sarcopenia among a wide-ranging sample of hospitalized adult patients. 42 Materials and Methods Study sample and design A longitudinal study was conducted in a general, university and 600 beds hospital between July 2011 and December 2014. A consecutive sampling method was applied in medical and surgical wards. Patients were eligible to participate in the study if they were aged 18 years and over, Caucasian, with an expected hospital stay longer than 24 hours, conscious, cooperative and capable of providing written informed consent. Patients unable to perform the handgrip strength (HGS) technique were excluded from the study. This impossibility in carrying out HGS measurement was defined as an inability to understand verbal instructions or having a condition limiting HGS measurement (namely pain). Critically ill patients, i.e., with a life-threatening medical or surgical condition requiring intensive care unit level care, presenting severe organ system dysfunction and needing for active therapeutic support were excluded (19). Pregnancy and patient ward isolation were also defined as exclusion criteria. According to these criteria, patients admitted to neurology, clinical haematology and intensive care unit wards were not recruited whereas participants from the following departments were selected: angiology and vascular surgery, cardiology, digestive surgery, endocrinology, gastroenterology, hepatobiliary surgery, internal medicine, nephrology, non-digestive surgery, orthopaedics, otorhinolaryngology and urology. Therefore, from the daily list of inpatients admitted to each of these wards, those who fulfilled inclusion criteria were invited to participate in the study, until the number of patients had attained the total number of beds of the ward. 43 From 992 patients who fulfilled the inclusion criteria and were invited to participate, 337 (34%) were not included. The reasons were refusals (n=198), cognitive impairment (n=13) and missing data (n=126). All patients were followed up from the time of admission until death, hospital discharge or 30 days after admission. Ethics This research was carried out according to the recommendations established by the Declaration of Helsinki and approved by the Institutional ethics and review boards of Centro Hospitalar do Porto. All study participants provided a written informed consent. Data collection Demographical, clinical data, medical diagnoses and data of hospital admission were retrieved from patient's clinical file at the time of evaluation. Date of hospital discharge, discharge destination (home, another ward, another hospital, continuing care unit and discharge against medical advice or death) and discharge diagnosis were retrieved from hospital records after patient discharge. All other information was obtained by two trained registered nutritionists through a structured questionnaire within 72h of admission to hospital. Education was evaluated by the number of completed school years and the following categories were created: 0-4, 5-12 and more than 12 years. Marital status was categorized as single, married or in a civil partnership, divorced and widowed. Cognitive impairment was evaluated with the Abbreviated Mental Test (AMT) (20). 44 Independence in activities of daily living was assessed with the Katz index (21). Charlson disease severity index (22) was obtained by two previously trained interviewers using medical discharge diagnoses in the patient's clinical record. Patient nutritional status was evaluated with Patient - Generated Subjective Global Assessment (PG-SGA) (23). Standing height (cm) was measured with a metal tape (Rosscraft, Innovations Incorporated, Surrey, Canada) with a 0.1 cm resolution and a headboard. Body weight (kg) was assessed with a calibrated portable beam scale with a 0.5 kg resolution. All anthropometric measurements were performed by two previously trained registered nutritionists using standard methods (24). The intraand interobserver technical error of measurement was calculated for all measurements, respectively, in 17 and 18 individuals. Intra-observer ranged from 0.2% to 0.6%, and inter-observer error ranged from 0 to 1.4%. These values are considered acceptable for trained anthropometrists (25). Body mass index (BMI) was determined through the standard formula [weight (kg) / height2 (m)] and BMI categories were created according to the World Health Organization cut-offs (26). Sarcopenia was defined according to the EWGSOP as the presence of both low muscle mass and low muscle function (1). Whole body resistance (ohms) and reactance (ohms) were assessed through tetrapolar bioelectrical impedance analysis (BIA) using a Biodynamics Model 450 (Seattle, Washington USA) with 0.1 ohm resolution, operating at a single frequency of 50 KHz. Muscle mass was evaluated using the equation of Janssen et al. (2000) (27): [(height2/ resistance × 0.401) + (gender × 3.825) + (age × –0.071)] + 5.102, with 45 height measured in cm; resistance measured in ohms; for gender, men = 1 and women = 0; age measured in years. Muscle mass was adjusted for height. Gender specific cut-off points indicated in the EWGSOP consensus were used (1). Muscle function was evaluated as HGS, using a calibrated Jamar® Hydraulic Hand dynamometer (Sammons Preston, Bolingbrook, IL, USA), with a 0.1 kgf resolution. The Jamar® dynamometer is proposed by the American Society of Hand Therapists as the gold standard for measurements of HGS (28). Each subject undertook three measurements using the non-dominant hand with a one minute interval between measurements and the maximum value was selected (29). Low HGS was classified using the cut-offs proposed in the EWGSOP Consensus (1): less than 30 kgf for men and 20 kgf for women. Statistics According to the normality of variables distribution, evaluated through Kolmogorov-Smirnov test, results were described as mean and standard deviation or as median and interquartile range (IQR) if non-normal distribution. Categorical variables were reported as frequencies. In order to identify variables associated with sarcopenia by bivariable analysis, sarcopenic and non-sarcopenic patients were compared for several demographic and clinical characteristics. Bivariable and multivariable logistic regression models were also conducted. Variables were included in the multivariable logistic regression model considering their potential confounding effect. Length of hospital stay was dichotomized according to a cut-off of 7 days based on the median LOS of the entire sample, and in agreement with the median 46 LOS in Portuguese hospitals (30). Variables associated with longer LOS (≥7 days) were identified comparing patients with and without a long LOS. All the comparisons were computed using Mann-Whitney test, or Student’s t test for independent samples, for continuous variables and Pearson χ2 or Fisher’s exact test for categorical variables. Length of hospital stay was determined from the date of hospital admission and discharge to usual residence (the main event of interest). Patients who were not discharged from the hospital to usual residence within the study period were censored at the time of other events, namely death, transfer (to another hospital ward, to another hospital or to continuing care units) and discharge against medical advice (n=40). Length of hospital stay was censored at 30 days, so patients that remained hospitalized 30 days after hospital admission were also censored (n=16). The Kaplan-Meier method was used to estimate the cumulative probability of being discharge-free over time (i.e. to experience the event of interest, defined as discharge home within the follow-up interval), according to the presence or the absence of sarcopenia. Multivariable Cox proportional hazards regression models were used to estimate adjusted hazard ratios (HR) and corresponding 95% confidence intervals (CI). The following characteristics were considered in the multivariable procedure: presence of sarcopenia (categorical), age (categorical), Charlson index (continuous), nutritional status categories according to PG-SGA (categorical), education (categorical), Katz index (categorical), gender (categorical), marital status (categorical) and AMT (continuous). 47 Statistical significance was set at p < 0.05. All analyses were conducted with the Software Package for Social Sciences (SPSS) for Windows (version 20.0; SPSS, Inc., Chicago, IL). 48 Results Baseline characteristics of the 655 hospitalized patients enrolled in this study, according to sarcopenia status are shown in Table 1. Approximately half of patients were women (46.1%), age ranged between 18 and 90 years old (median (IQR) = 56 (22) years). Frequency of sarcopenia was 24.3%. Within the period the present study was conducted two patients had died. Therefore, mortality rate was 0.3%. Sarcopenic patients were older and presented longer LOS than nonsarcopenic patients (Table 1). Also, they were more likely to be male, to be undernourished and to present higher Charlson index score than non-sarcopenic patients (Table 1). There was a higher proportion of sarcopenic patients in medical wards than in surgical wards. The highest proportion of sarcopenic patients (34.3%) was observed in internal medicine wards. Otorhinolaryngology presented the lowest proportion of sarcopenic patients (1.9%). It is worth noticing that patients aged ≥ 65 years presented lower muscle mass (median (IQR) 24.8 (11.4) kg) than patients aged < 65 years (median (IQR) 26.4 (11.4) kg), p= 0.008. Older patients also presented lower HGS than patients aged<65 years (median (IQR) 22.0 (9.8) kgf versus median (IQR) 24.1 (17.5) kgf), p<0.001. As shown in Table 2, after adjusting for potential confounders, being a male, aged ≥65 years, presenting moderate or severe dependence, being undernourished and being admitted to a medical ward were factors associated with sarcopenia. 49 Table 1 – Participants’ baseline characteristics according to sarcopenia status. Non-sarcopenic (n=496) Sarcopenic (n=159) p Age (years), median (IQR) 54 (24.0) 64 (19.0) <0.001 2 Age categories, n (%) < 65 ≥ 65 367 (74.0) 129 (26.0) 85 (53.5) 74 (46.5) <0.001 1 Gender, n (%) Women Men 244 (49.2) 252 (50.8) 58 (36.5) 101 (63.5) 0.006 1 Education (years), n (%) 0-4 5-12 >12 183 (36.9) 270 (54.4) 43 (8.7) 81 (50.9) 64 (40.3) 14 (8.8) 0.005 1 Marital status, n (%) Single Non-single 91 (18.3) 405 (81.7) 25 (15.7) 134 (84.3) 0.477 1 AMT, median (IQR) 10.0 (1.0) 10.0 (1.0) 0.437 2 Charlson Index, median (IQR) 1.0 (2.0) 2.0 (3.0) 0.002 2 PG-SGA, n (%) Non-undernourished Undernourished 298 (60.1) 198 (39.9) 63 (39.6) 96 (60.4) <0.001 1 Table continued 56 Results from multivariable Cox proportional hazards regression models were displayed for the entire sample and according to age groups (Table 4). The model was adjusted for age, gender, marital status, education, nutritional status, Charlson index, AMT score and Katz index, as these variables could be considered as potential confounders in the association between sarcopenia and LOS. Considering the entire sample and the group of patients aged < 65 years, sarcopenia was consistently associated with lower HR (<1) for being discharged home, meaning that sarcopenic patients presented a lower probability of being discharged home. However, for patients aged ≥ 65 years, sarcopenia was not independently associated with the probability of being discharged home. It is worth noticing that sarcopenic overweight or obese patients presented a higher probability of being discharged home, adjusted HR (95% CI) = 0.78 (0.610.99) than non-overweight sarcopenic patients, adjusted HR (95% CI) = 0.63 (0.480.83). Additionally, LOS had also been stratified according to hospital ward (medical or surgical) and, as expected, there was a higher proportion of patients with a longer LOS (≥7 days) admitted to medical wards (53%) than in surgical wards (47%), p=0.019. Thus, the type of hospital ward was included in an additional multivariable Cox proportional hazards regression model. However, the inclusion of this variable did not modify the results concerning the probability of being discharged home. 57 Table 4 - Hazard ratios (HR) of being discharged home associated with the presence of sarcopenia. All patients (n=655) Age < 65 years (n= 452) Age ≥ 65 years (n= 203) Adjusted HR (95% CI) p Adjusted HR (95% CI) p Adjusted HR (95% CI) p Sarcopenia Non-sarcopenic Sarcopenic 1 0.71 (0.58-0.86) 0.001 1 0.66 (0.51-0.86) 0.002 1 0.80 (0.58-1.10) 0.168 Gender Female Male 1 1.00 (0.85-1.19) 0.969 1 0.97 (0.79-1.18) 0.754 1 1.18 (0.84-1.66) 0.328 Age (years) < 65 ≥ 65 1 0.94 (0.78-1.13) 0.535 - - - - - - - - Education (years) 0-4 4-12 >12 1 0.81 (0.67-0.97) 1.24 (0.92-1.68) 0.023 0.156 1 0.74 (0.59-0.93) 1.23 (0.83-1.33) 0.010 0.293 1 0.82 (0.59-1.14) 0.92 (0.56-1.53) 0.235 0.759 Table continued 58 CI, confidence interval; AMT, Abbreviated Mental Test; PG-SGA, Patient-Generated Subjective Global Assessment. All patients (n=655) Age < 65 years (n= 452) Age ≥ 65 years (n= 203) Adjusted HR (95% CI) P Adjusted HR (95% CI) p Adjusted HR (95% CI) p Marital Status Single Non-single 1 0.87 (0.70-1.07) 0.190 1 0.82 (0.64-1.03) 0.094 1 1.22 (0.66-2.25) 0.531 AMT 1.05 (0.96-1.14) 0.301 1.10 (0.99-1.24) 0.079 0.98 (0.85-1.12) 0.733 Katz index Independent Moderate / Severe dependence 1 0.77 (0.52-1.14) 0.188 1 0.74 (0.45-1.23) 0.250 1 0.71 (0.38-1.31) 0.268 PG-SGA Non-undernourished Moderate / Severe undernutrition 1 0.56 (0.47-0.66) <0.001 1 0.51 (0.41-0.62) <0.001 1 0.70 (0.51-0.97) 0.030 Charlson index 0.94 (0.90-0.98) 0.003 0.95 (0.90-1.00) 0.068 0.90 (0.83-0.98) 0.010 Table 4 continued Table 3 - Hazard ratios (HR) of being discharge-free over time associated with the presence of sarcopenia. 59 Discussion The present study results show that sarcopenic patients presented a lower probability of being discharged from the hospital. Cox analysis revealed that sarcopenia is associated with longer LOS after considering the confounding effect of age, gender, marital status, education, nutritional status, disease severity, cognitive impairment and independence in daily living activities. However, after stratifying this analysis by age groups, this association was only observed for patients aged <65 years. This may be explained by a lower proportion of older patients in the study sample (approximately 31%), which leads to a loss of statistical power, increasing the possibility of occurrence of a type two error or, alternatively, by different clinical characteristics, i.e., the simultaneous presence of several comorbidities in older patients could have diminished the strength of the association of sarcopenia with LOS. This study results increased the knowledge and highlighted the impact of sarcopenia on LOS, specifically among hospitalized younger patients (<65 years). Besides, as far as we are concerned, there were no previous data concerning factors associated with sarcopenia among hospitalized younger patients, with the exception for previous results from a recent study undertaken by our research team (10). Gariballa and Alessa (2013) (7), in a study conducted among hospitalized older patients which defined sarcopenia with muscle mass assessed through midarm muscle circumference and muscle function evaluated by HGS, concluded that LOS was significantly higher in sarcopenic patients compared with non-sarcopenic 60 patients. Otherwise, in a study conducted by Cerri et al. (2014) (9) among hospitalized undernourished older patients, no differences in LOS were found between sarcopenic and non-sarcopenic patients. Using similar methodology, our results for older patients corroborate Cerri et al. (9) findings. Although present results are not in accordance with Gariballa and Alessa, the observed differences between studies may be explained by the use of different methodology, BIA and anthropometry, in the assessment of muscle mass and by different patients’ characteristics. However, our results clearly show a significant association of sarcopenia with prolonged LOS for patients aged under 65 years. The difference observed for < 65 years and ≥ 65 years groups concerning the association of sarcopenia with LOS may be justified by the existence of different characteristics, diagnoses and, even, higher severity of co-morbidities between younger and older adult patients, besides the possible occurrence of a type two error, as hypothesized before in this section. The present study results also showed that sarcopenic overweight or obese patients (BMI ≥ 25 kg/m2) had a higher probability of being discharged home than sarcopenic non-overweight patients. A possible explanation is that overweight (or obese) patients presented significantly higher muscle mass than non-overweight patients. Thus, characteristics related to overweight and higher muscle mass could have introduced a protective effect for being discharged from the hospital. Notwithstanding this, due to the presence of overweight or obesity these patients may not present obvious frailty physical features. This may have influenced caregivers and biased the indication for discharge destination. 61 The present study shows a frequency of sarcopenia among hospitalized older adults of 36.4%, being higher than previous reports, 10% from Gariballa and Alessa (2013) (7), 25.3% from Smoliner et al.(2014) (8), 26% from Rossi et al. (2014) (11) and 21.4% from Cerri et al.(2014) (9). These differences may be due to the use of different methodology and to patients’ characteristics. This study also identified sarcopenia in 18.8% of the hospitalized patients aged under 65 years. However, it is noteworthy that cut-off points used were previously defined for use in older adults, as sarcopenia was considered as a geriatric condition. This situation may have biased present results with a possible under diagnosis of sarcopenia. Patients from intensive care units and other critical patients were excluded from the present study due to their inability to perform the required functional tests to identify sarcopenia. This situation may constitute a study limitation because critical patients due to their clinical condition, would be likely to present muscle mass depletion and reduced function and, therefore, to be sarcopenic. Furthermore, the inclusion of muscle function (physical performance) in the definition and diagnostic criteria of sarcopenia may impair the identification of sarcopenia among critical patients and patients unable to perform functional tests. Moreover, it is important to highlight that HGS of patients unable to stand was measured with individuals on a bed. Although a differential may exist between measurements performed with the individual in a sitting or lying position, care was taken in order to follow strictly HGS measurement protocol (29). Specifically, HGS was obtained from all participants with the unsupported elbow (31). In the present study, muscle mass was estimated through BIA, instead of using Computed Tomography (CT) or Magnetic Resonance Imaging (MRI), the 62 golden standards for quantifying muscle mass, or Dual Energy X-ray Absorptiometry (DXA) the selected alternative for estimating muscle mass in research and clinical use (1). This could be regarded as a study limitation. However, BIA results are readily reproducible and this is an economical, practical and portable method which, used under standard conditions, has been found to be a good alternative to DXA (6). Although BIA may not be reliable in conditions like heart failure, kidney failure, and dehydration, after applying inclusion criteria, not all patients with these conditions were excluded. This may have led to a misclassification of muscle mass and subsequently to a misclassification of sarcopenia. According to hospital discharge records, the proportion of discharged patients aged over 65 years was 38.3% in 2012 and 40% in 2013. Our sample contains less patients aged over 65 years (31%). This may have resulted in a lower representation of an important group of high risk patients, underestimating sarcopenia burden. Nonetheless, the diagnostic criteria of sarcopenia recommended by the European Consensus necessitate the application of functional tests, thus excluding patients who are unable to carry out these tests (9). The lower representation of older patients in this sample may be explained by the need to fulfill the criteria. Several strengths of this study could be highlighted. A large number of hospitalized patients composed this study sample, with a wide age range, 18 to 90 years old. The patients enrolled in the present study were from a multiplicity of hospital surgical and medical wards, which ensured a large variety of diagnoses and different diseases. These characteristics strengthen the generalizability of our results for other hospitalized patients. 63 Survival analysis has the ability of handling data that are censored, which in this study were death, transfer, discharge against medical advice and LOS>30 days. This allows for a better hospital representation, because it permits the inclusion of cases that could not be included with other statistical approaches, namely, with follow-up information unavailable after a certain point, which in our study was 30 days after hospital admission. Nonetheless, only 16 participants (2.4% of the study sample) had a LOS longer than 30 days, thus an extended follow-up period probably would not have changed the results obtained. Although there are some results available concerning mortality and hospital readmission for older patients (7, 9, 12), further research is required in order to assess short-term and long-term consequences of sarcopenia in hospitalized patients. Being male, aged ≥65 years, presenting dependence, being undernourished and being admitted to a medical ward are factors associated with sarcopenia among hospitalized adult patients. Sarcopenia is independently associated with longer LOS, although this association is stronger for patients aged < 65 years. Moreover, sarcopenic overweight is associated with a higher probability of discharge to usual residence than non-overweight sarcopenia. Acknowledgements The authors thank Centro Hospitalar do Porto and all ward directors for facilitating the data collection. R. S. G. received a scholarship from Fundação para a Ciência e a Tecnologia, financing program POPH/FSE, under the project SFRH/BD/61656/2009. 64 Conflicts of interest The authors declare no conflicts of interest. 65 References 1. Cruz-Jentoft AJ, Baeyens JP, Bauer JM, Boirie Y, Cederholm T, Landi F, et al. Sarcopenia: European consensus on definition and diagnosis: Report of the European Working Group on Sarcopenia in Older People. Age Ageing 2010;39(4):412-23. 2. Landi F, Liperoti R, Russo A, Giovannini S, Tosato M, Capoluongo E, et al. Sarcopenia as a risk factor for falls in elderly individuals: results from the ilSIRENTE study. Clin Nutr 2012;31(5):652-8. 3. Janssen I, Heymsfield SB, Ross R. Low relative skeletal muscle mass (sarcopenia) in older persons is associated with functional impairment and physical disability. J Am Geriatr Soc 2002;50(5):889-96. 4. Abellan van Kan G. Epidemiology and consequences of sarcopenia. J Nutr Health Aging 2009;13(8):708-12. 5. Volpato S, Bianchi L, Cherubini A, Landi F, Maggio M, Savino E, et al. Prevalence and clinical correlates of sarcopenia in community-dwelling older people: application of the EWGSOP definition and diagnostic algorithm. J Gerontol A Biol Sci Med Sci 2014;69(4):438-46. 6. Cherin P, Voronska E, Fraoucene N, de Jaeger C. Prevalence of sarcopenia among healthy ambulatory subjects: the sarcopenia begins from 45 years. Aging Clin Exp Res 2014;26(2):137-46. 7. Gariballa S, Alessa A. Sarcopenia: prevalence and prognostic significance in hospitalized patients. Clin Nutr 2013;32(5):772-6. 72 Abstract Background and aims: Data on the association of sarcopenia with costs among hospitalized patients are limited to surgical patients. This study aims to increase knowledge regarding the association of sarcopenia with these costs among a wide-ranging sample of surgical and non-surgical patients. Methods: A prospective study was conducted among hospitalized adult patients. Sarcopenia was identified according to the European Working Group on Sarcopenia in Older People, as low muscle mass, assessed by bioelectrical impedance analysis and low muscle function evaluated by handgrip strength. Hospitalization cost was calculated for each patient based on discharge diagnosis related group codes and determined on the basis of a relative weight value. Costs were defined as the percentage of deviation from the cost of a patient with a relative weight equal to one. Multivariable linear regression models were performed to identify the factors independently associated to hospitalization costs. Results: 656 hospitalized patients aged ≥18 years (24.2% sarcopenic) composed the study sample. Sarcopenia increased hospitalization costs by €1240 (95 % CI: €596-1887) for patients aged <65 years and €721 (95% CI: €13-1429) for patients aged ≥65 years. Sarcopenic overweight was related to an increase in hospitalization costs of €884 (95% CI: €295-1476). Conclusion: Sarcopenia is independently related to hospitalization costs. This condition is estimated to increase hospitalization costs by 58.5% for patients aged <65 years and 34% for patients aged ≥65 years. Key-words: sarcopenia; body composition; handgrip strength; hospital; cost 73 Introduction Sarcopenia is currently defined as a combination of both low muscle mass and low muscle function, according to the European Working Group on Sarcopenia in Older People (EWGSOP) (1). It is estimated that sarcopenia occurs between 5 to 45% of community dwelling older adults (2-4). Although this condition has been mainly described in older adults, it can also be present in younger individuals. Cherin et al. in a study conducted among community dwelling adults (5) showed that 9% of the individuals aged between 45 and 54 years and 13.5% of those aged from 55 to 64 years were sarcopenic. Previous studies have shown that this condition is highly frequent among hospitalized older patients (6-10), ranging from 10% to 37.3% and was identified in circa one fifth of patients aged under 65 years (9). This condition has been associated with physical disability, low quality of life and higher mortality in community dwelling older adults (1, 11). Among hospitalized patients, sarcopenia has been related with poor clinical outcome, namely worst post-operative outcomes (12-15), higher risk of non-elective readmission (6) and higher mortality (6, 8, 16). Considering the impact of sarcopenia on both community dwelling and hospitalized individuals, healthcare costs of this condition are expected to be high (17). However, according to our knowledge, data on the economic burden of sarcopenia are limited. One study from 2004 (18), conducted among representative samples of American adults aged ≥ 60 years, reported that the estimated healthcare cost attributable to sarcopenia defined as the loss of muscle mass was $18.5 billion 74 ($10.8 billion in men, $7.7 billion in women). Recent studies from 2013 (19) and 2015 (20, 21), reported that sarcopenia determined by computed tomography scans, was associated with increased costs in major surgery. Nevertheless, the impact of sarcopenia on hospitalization costs among a wider variety of patients, from surgical and non-surgical wards, remains to be documented. Considering the adverse consequences sarcopenia entails among hospitalized patients and the financial constraints that healthcare systems often face, it is important to recognize and explore the association of sarcopenia with hospitalization costs, in order to maximize resources and provide a more effective healthcare plan. Therefore, the present study aims to increase the knowledge on the association of sarcopenia with costs among a wide-ranging sample of hospitalized patients. 75 Materials and Methods Study sample and design A prospective study was conducted in a general and university hospital between July 2011 and December 2014. A consecutive sampling method was applied in medical and surgical wards. Patients were eligible to participate in the study if they were aged 18 years and over, Caucasian, with an expected hospital stay longer than 24 hours, conscious, cooperative and capable of providing written informed consent. Patients unable to perform the handgrip strength (HGS) technique were excluded from the study. This impossibility in carrying out HGS measurement was defined as an inability to understand verbal instructions or having a condition limiting HGS measurement (namely pain). Critically ill patients, i.e., with a life-threatening medical or surgical condition requiring intensive care unit level care, presenting severe organ system dysfunction and needing for active therapeutic support were excluded (22). Pregnancy and patient ward isolation were also defined as exclusion criteria. According to these criteria, patients admitted to neurology, clinical haematology and intensive care unit wards were not recruited whereas participants from the following departments were selected: angiology and vascular surgery, cardiology, digestive surgery, endocrinology, gastroenterology, hepatobiliary surgery, internal medicine, nephrology, non-digestive surgery, orthopaedics, otorhinolaryngology and urology. Therefore, from the daily list of inpatients admitted to each of these wards, those who fulfilled inclusion criteria were invited to 76 participate in the study, until the number of patients had attained the total number of beds of the ward. From 992 patients who fulfilled the inclusion criteria and were invited to participate, 336 (33.9%) were not included. The reasons were refusals (n=198), cognitive impairment (n=13) and missing data (n=125). Ethics This research was conducted according to the recommendations established by the Declaration of Helsinki and approved by the Institutional ethics and review boards of Centro Hospitalar do Porto. All study participants provided written informed consent. Data collection Demographical, clinical data, medical diagnoses and data of hospital admission were retrieved from patient's clinical file at the time of evaluation. Date of hospital discharge and discharge diagnosis were retrieved from hospital records after the patient had left the hospital. All other information was obtained by two trained registered nutritionists through a structured questionnaire within 72h of admission to the hospital. Education was evaluated by the number of completed school years and the following categories were created: 0-4, 5-12 and more than 12 years. Marital status was categorized as single and not single (married or in a civil partnership, divorced and widowed). Cognitive impairment was evaluated with the Abbreviated Mental Test (AMT) (23). Independence in activities of daily living was assessed with the 77 Katz index (24) and two categories were defined according to the score obtained: ≤ 5 - moderate / severe dependence and 6 - independent. Charlson Disease Severity Index (25) was recorded by two previously trained interviewers using medical discharge diagnoses in the patient's clinical record. Patient nutritional status was evaluated with Patient - Generated Subjective Global Assessment (PG-SGA) (26). Standing height (cm) was measured with a metal tape (Rosscraft, Innovations Incorporated, Surrey, Canada) with a 0.1 cm resolution and a headboard. Body weight (kg) was assessed with a calibrated portable beam scale with a 0.5 kg resolution. All anthropometric measurements were performed by two previously trained registered nutritionists using standard methods (27). The intra-and inter-observer technical error of measurement was calculated for all measurements, respectively, in 17 and 18 individuals. Intraobserver ranged from 0.2% to 0.6%, and inter-observer error ranged from 0 to 1.4%. These values are considered acceptable for trained anthropometrists (28). Sarcopenia was defined according to the EWGSOP as the presence of both low muscle mass and low muscle function (1). Whole body resistance (ohms) and reactance (ohms) were assessed through tetrapolar bioelectrical impedance analysis (BIA) using a Biodynamics Model 450 (Seattle, Washington USA) with 0.1 ohm resolution, operating at a single frequency of 50 KHz. Muscle mass was evaluated using the equation of Janssen et al. (2000) (29): [(height2/ resistance × 0.401) + (gender × 3.825) + (age × –0.071)] + 5.102, with height measured in cm; resistance measured in ohms; for gender, men = 1 and 78 women = 0; age measured in years. Muscle mass was adjusted for height. Gender specific cut-off points indicated in the EWGSOP consensus were used (1). Muscle function was evaluated by HGS, using a calibrated Jamar® Hydraulic Hand dynamometer (Sammons Preston, Bolingbrook, IL, USA), with 0.1 kgf resolution. The Jamar® dynamometer is proposed by the American Society of Hand Therapists as the gold standard for measurements of HGS (30). Each subject undertook three measurements using the non-dominant hand with a one minute interval between measurements and the maximum value was selected (31). Low HGS was classified using the cut-offs proposed in the EWGSOP Consensus (1): less than 30 kgf for men and 20 kgf for women. Body mass index (BMI) was determined through the standard formula [weight (kg) / height2 (m)] and BMI categories were created according to the World Health Organization cut-offs (32). Statistics According to the normality of variables distribution, evaluated through Kolmogorov-Smirnov test, results were described as mean and standard deviation or as median and interquartile range (IQR) if non-normal distribution. Categorical variables were reported as frequencies. Hospitalization cost was calculated for each patient based on discharge diagnosis related group (DRG) codes. The DRG system is used to calculate hospital reimbursements, with the amounts determined on the basis of a relative weight value. This weight value reflects the main diagnosis, surgical interventions, 79 pathologies, complications, clinical procedures, medium length of hospital stay, age, gender and discharge destination. The information about DRG codes and its amounts was obtained from Portuguese Ministerial Directive number 839-A, 31 July 2009 (33) for data obtained between 2011 and 2012, number 163, 24 April 2013 (34) was used for data obtained in 2013 and number 20 from 29 January 2014 (35) was used for data obtained in 2014. The percentage of cost deviation was calculated from the difference between the cost of each patient and the cost of a patient with a relative weight equal to one (€2396 for data obtained between 2011 and 2012; €2142 for data obtained in 2013; €2120 for data obtained in 2014). Percentage of cost deviation was summarized into quartiles using the cutoffs of the sample distribution: ≤-35.3 (24.1%); -35.2,-1.10 (25.8%); -1.09, 88.4 (24.5%); ≥88.5 (25.6%). Length of hospital stay (LOS) was determined from the date of hospital admission and discharge. Length of hospital stay was also dichotomized according to a cut-off of 7 days based on the median LOS of the entire sample, and in agreement with the median LOS in Portuguese hospitals (36). In order to select variables associated with sarcopenia and with percentage of cost deviation, patients were compared for several demographic and clinical characteristics. All the comparisons were computed using Mann-Whitney test, or Student’s t test for independent samples, or Kruskal-Wallis test for continuous variables and Pearson χ2 for categorical variables. Multivariable linear regression models using stepwise method were performed to identify the independent variables associated with percentage of cost deviation. The following variables were included in the model: sarcopenia status 80 (categorical), age (continuous), gender (categorical), marital status (categorical), Katz index (categorical), education (categorical), nutrition status (categorical), hospital ward (categorical), length of hospital stay (categorical), the Abbreviated Mental Test score (continuous) and the Charlson comorbidity index score (continuous). These variables were included, as they were considered potential confounders. Statistical significance was set at p < 0.05. All analyses were conducted with the Software Package for Social Sciences (SPSS) for Windows (version 20.0; SPSS, Inc., Chicago, IL). 81 Results Baseline characteristics of the 656 hospitalized patients enrolled in this study, according to sarcopenia status are shown in Table 1. Approximately half of patients were women (46.1%), age ranged between 18 and 90 years old, median (IQR) = 56 (22) years. Sarcopenia was highly frequent affecting 24.2% of the participants. Sarcopenic patients were older and presented longer LOS than non-sarcopenic patients (Table 1). Also, sarcopenic patients were more likely to be male, to be undernourished and to present higher Charlson index score than non-sarcopenic patients (Table 1). There was a higher proportion of sarcopenic patients in medical wards than in surgical wards (Table 1). The highest proportion of sarcopenic patients (34.3%) was observed in internal medicine wards. Hospitalization costs within the present sample ranged from €387 to €30880, median (IQR) of €2369 (€3094). Patients’ characteristics were stratified according to the percentage of cost deviation quartiles, as shown in Table 2. Sarcopenic patients presented a positive percentage of cost deviation, i.e., these patients present a mean cost higher than the cost of a patient with relative weight equal to one. Otherwise, the percentage of cost deviation was negative for non-sarcopenic patients. Thus, non-sarcopenic patients present a mean cost lower than the cost of a patient with relative weight equal to one. Compared to patients in the upper quartiles of percentage of cost deviation, patients in the lower quartiles had a higher education level, were more likely to be single, were less likely to be dependent and presented better nutrition status and shorter length of hospital stay (Table 2). The highest proportion of sarcopenic patients was found in the highest quartile of percentage of cost deviation distribution. 88 Discussion The present study results show that sarcopenia is associated with a major increase in hospitalization costs, considering the effect of potential confounders. After stratifying the model according to age group, this effect was still visible for both younger and older adults, in spite of being stronger for younger patients. It is worth highlighting that with the exception of sarcopenia, factors associated with hospitalization costs changed across the two different age groups. While age, undernutrition, being on a surgical hospital ward and length of stay were related to higher hospitalization costs for younger patients, in the model carried out for patients aged ≥ 65 years, only length of stay and the ward of hospitalization were associated with the percentage of cost deviation. Moreover, sarcopenic overweight (or obesity) was also a predictor of higher hospitalization costs, even though the association was weaker than the one described for sarcopenic patients. Additionally, it is worth mentioning that the multivariable linear regression model was adjusted for length of hospital stay using a dichotomised variable (< 7 days, ≥7 days), although medium length of hospital stay is included in the weighing value used for calculate hospitalization cost. The inclusion of this variable in the multivariable analyses is justified by the existence of an interval of days of hospitalization for each DRG code. These defined intervals can be wide-ranging. Depending on the interval indicated to each DRG code, the same DRG code can be attributable to a patient with a short LOS (< 7 days) and a patient with a longer LOS (≥ 7 days). The potential confounding effect of different LOS was, therefore, controlled. 89 The present study results increase the knowledge about sarcopenia and hospitalization costs by providing an estimation on this association. Sarcopenia defined by computed tomography scans was related with higher costs among patients who underwent surgery (19, 20, 21). The ability of HGS, as a single parameter in predicting higher hospitalization costs has also been recently described (37). However, as far as we are concerned, there were no previous reports where sarcopenia was defined as low muscle mass and low muscle function and among hospitalized patients with a wide range of diagnoses and age. Therefore, due to differences in methodology, these results are not comparable with previous reports. The impact of sarcopenia in healthcare costs has been described in the United States (18). But, in this report, sarcopenia was defined merely by the loss of muscle mass and the observation was not focused on hospitalized patients. The inclusion of muscle function (physical performance) in the definition and diagnostic criteria of sarcopenia may impair the identification of sarcopenia among critically ill patients and patients unable to perform functional tests. This situation may constitute a study limitation because these patients due to their clinical condition, would be likely to present muscle mass depletion and reduced function and, therefore, to be sarcopenic. In the present study, muscle mass was assessed with BIA, instead of using computed tomography or magnetic resonance imaging which are the golden standards for quantifying muscle mass, or dual energy x-ray absorptiometry, the selected alternative for estimating muscle mass in research and clinical use (1). This could be a study limitation. However, BIA is an economical, practical and reproducible method which, used under standard conditions, has been described as 90 a suitable alternative to dual energy x-ray absorptiometry (1). Although BIA may not be reliable in conditions like heart failure, kidney failure, and dehydration (29) not all patients with these conditions were excluded. This may have caused a misclassification of muscle mass and consequently a misclassification of sarcopenia. According to hospital discharge records, the proportion of discharged older patients (aged ≥ 65 years) was 38.3% in 2012 and 40% in 2013.The proportion of patients aged over 65 years in our sample contains is lower (31%). This situation may have led to a lower representation of an important group of high risk patients, underestimating sarcopenia burden. Nevertheless, the diagnostic criteria of sarcopenia recommended by the European Consensus, requires the application of functional tests, thus excluding patients who are unable to carry out these tests (8). The lower representation of older patients in this sample may be explained by the need to comply with the criteria. The DRG system has been shown to underestimate the real hospitalization costs as it reflects only direct hospitalization costs. Indirect costs, as societal costs, are not taken into account (38). Nonetheless, this methodology was used in the present study as it allows for an assortment of patients with a diversity of diagnoses and procedures. Several strengths of this study could be emphasized. This sample is composed of a large number of hospitalized patients, with a wide age range, from 18 to 90 years old. The patients enrolled in the present study were from a multiplicity of hospital surgical and medical wards, which ensured a variety of diagnoses and 91 diseases. These characteristics strengthen the generalizability of our results for other hospitalized patients. Further investigation is needed in order to explore the extent of the influence of the early identification of sarcopenia in the reduction of adverse outcomes and, therefore, the reduction of inherent hospitalization costs. In conclusion, present research shows that sarcopenia is independently related to hospitalization costs. This condition is estimated to increase hospitalization costs in 52.7% (58.5% for patients aged < 65 years and 34% for patients aged ≥ 65 years). Acknowledgements The authors thank Centro Hospitalar do Porto and all ward directors for facilitating the data collection. R. S. G. received a scholarship from Fundação para a Ciência e a Tecnologia, financing program POPH/FSE, under the project SFRH/BD/61656/2009. The authors declare no conflicts of interest. 92 References 1. Cruz-Jentoft AJ, Baeyens JP, Bauer JM, Boirie Y, Cederholm T, Landi F, et al. Sarcopenia: European consensus on definition and diagnosis: Report of the European Working Group on Sarcopenia in Older People. Age Ageing 2010;39(4):412-23. 2. Janssen I, Heymsfield SB, Ross R. Low relative skeletal muscle mass (sarcopenia) in older persons is associated with functional impairment and physical disability. J Am Geriatr Soc 2002;50(5):889-96. 3. Abellan van Kan G. Epidemiology and consequences of sarcopenia. J Nutr Health Aging 2009;13(8):708-12. 4. Volpato S, Bianchi L, Cherubini A, Landi F, Maggio M, Savino E, et al. Prevalence and clinical correlates of sarcopenia in community-dwelling older people: application of the EWGSOP definition and diagnostic algorithm. J Gerontol A Biol Sci Med Sci 2014;69(4):438-46. 5. Cherin P, Voronska E, Fraoucene N, de Jaeger C. Prevalence of sarcopenia among healthy ambulatory subjects: the sarcopenia begins from 45 years. Aging Clin Exp Res 2014;26(2):137-46. 6. Gariballa S, Alessa A. Sarcopenia: prevalence and prognostic significance in hospitalized patients. Clin Nutr 2013;32(5):772-6. 7. Smoliner C, Sieber CC, Wirth R. Prevalence of sarcopenia in geriatric hospitalized patients. J Am Med Dir Assoc 2014;15(4):267-72. 8. Cerri AP, Bellelli G, Mazzone A, Pittella F, Landi F, Zambon A, et al. Sarcopenia and malnutrition in acutely ill hospitalized elderly: Prevalence and outcomes. Clin Nutr 2014. 9. Sousa AS, Guerra RS, Fonseca I, Pichel F, Amaral TF. Sarcopenia among hospitalized patients - A cross-sectional study. Clin Nutr 2014. 10. Rossi AP, Fantin F, Micciolo R, Bertocchi M, Bertassello P, Zanandrea V, et al. Identifying sarcopenia in acute care setting patients. J Am Med Dir Assoc 2014;15(4):303 e7-12. 11. Landi F, Liperoti R, Russo A, Giovannini S, Tosato M, Capoluongo E, et al. Sarcopenia as a risk factor for falls in elderly individuals: results from the ilSIRENTE study. Clin Nutr 2012;31(5):652-8. 93 12. Otsuji H, Yokoyama Y, Ebata T, Igami T, Sugawara G, Mizuno T, et al. Preoperative Sarcopenia Negatively Impacts Postoperative Outcomes Following Major Hepatectomy with Extrahepatic Bile Duct Resection. World J Surg 2015. 13. Joglekar S, Asghar A, Mott SL, Johnson BE, Button AM, Clark E, et al. Sarcopenia is an independent predictor of complications following pancreatectomy for adenocarcinoma. J Surg Oncol 2014. 14. Du Y, Karvellas CJ, Baracos V, Williams DC, Khadaroo RG. Sarcopenia is a predictor of outcomes in very elderly patients undergoing emergency surgery. Surgery 2014;156(3):521-7. 15. Lieffers JR, Bathe OF, Fassbender K, Winget M, Baracos VE. Sarcopenia is associated with postoperative infection and delayed recovery from colorectal cancer resection surgery. Br J Cancer 2012;107(6):931-6. 16. Vetrano DL, Landi F, Volpato S, Corsonello A, Meloni E, Bernabei R, et al. Association of sarcopenia with shortand long-term mortality in older adults admitted to acute care wards: results from the CRIME study. J Gerontol A Biol Sci Med Sci 2014;69(9):1154-61. 17. Beaudart C, Rizzoli R, Bruyere O, Reginster JY, Biver E. Sarcopenia: burden and challenges for public health. Arch Public Health 2014;72(1):45. 18. Janssen I, Shepard DS, Katzmarzyk PT, Roubenoff R. The healthcare costs of sarcopenia in the United States. J Am Geriatr Soc 2004;52(1):80-5. 19. Sheetz KH, Waits SA, Terjimanian MN, Sullivan J, Campbell DA, Wang SC, et al. Cost of major surgery in the sarcopenic patient. J Am Coll Surg 2013;217(5):813-8. 20. Friedman J, Lussiez A, Sullivan J, Wang S, Englesbe M. Implications of Sarcopenia in Major Surgery. Nutr Clin Pract 2015;30(2):175-9. 21. Kirk PS, Friedman JF, Cron DC, Terjimanian MN, Wang SC, Campbell DA, et al. One-year postoperative resource utilization in sarcopenic patients. J Surg Res 2015;30(15):074. 22. Guidelines for the use of parenteral and enteral nutrition in adult and pediatric patients. JPEN Journal of Parenteral and Enteral Nutrition 2002;26(1 Suppl):1SA138SA. 23. Hodkinson HM. Evaluation of a mental test score for assessment of mental impairment in the elderly. Age Ageing 1972;1(4):233-8. 24. Katz S. Assessing self-maintenance: activities of daily living, mobility, and instrumental activities of daily living. J Am Geriatr Soc 1983;31(12):721-7. 94 25. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 1987;40(5):373-83. 26. Ottery F. Patient-Generated Subjective Global Assessment. In: Polisena PMC, editor. The Clinical Guide to Oncology Nutrition Chicago: The American Dietetic Association; 2000. p. 11-23. 27. Marfell-Jones M OT, Stewart A, Carter L. International standards for anthropometric assessement. Potchefstroom, South Africa: ISAK; 2006. 28. Pederson D, Gore C. Anthropometry Measurement Error. Sydney, Australia University of New South Wales Press; 1996. 29. Janssen I, Heymsfield SB, Baumgartner RN, Ross R. Estimation of skeletal muscle mass by bioelectrical impedance analysis. J Appl Physiol 2000;89(2):46571. 30. Fess E. Grip Strength. 2nd, editor. Chicago: American Society of Hand Therapists; 1992. 31. Vaz M, Thangam S, Prabhu A, Shetty PS. Maximal voluntary contraction as a functional indicator of adult chronic undernutrition. Br J Nutr 1996;76(1):9-15. 32. Physical status: the use and interpretation of anthropometry. Report of a WHO Expert Committee. World Health Organ Tech Rep Ser 1995;854:1-452. 33. Portaria n.º 839-A/2009. Diário da República. Portugal,2009. 34. Portaria n.º 163/2013. Diário da República. Portugal,2013. 35. Portaria nº 20/2014. Diário da República. Portugal,2014. 36. Matos L, Teixeira MA, Henriques A, Tavares MM, Alvares L, Antunes A, et al. [Nutritional status recording in hospitalized patient notes]. Acta Med Port 2007;20(6):503-10. 37. Guerra RS, Amaral TF, Sousa AS, Pichel F, Restivo MT, Ferreira S, et al. Handgrip strength measurement as a predictor of hospitalization costs. Eur J Clin Nutr 2015; 69(2):187-92. 38. Freijer K, Tan SS, Koopmanschap MA, Meijers JM, Halfens RJ, Nuijten MJ. The economic costs of disease related malnutrition. Clin Nutr 2013;32(1):136-41. 95 Chapter II Sousa AS; Fonseca I, Pichel F, Amaral TF Effects of posture and body mass index on body girths assessment Under review Sousa AS; Fonseca I, Pichel F,, Amaral TF Triceps skinfold compressibility in hospitalized patients – an exploratory analysis Under review 96 97 Chapter II.a Effects of posture and body mass index on body girths assessment 104 Patient nutritional status was evaluated with Nutritional Risk Screening 2002 11. Standing height (cm) was measured with a metal tape (Rosscraft, Innovations Incorporated, Surrey, Canada) with a 0.1 cm resolution and a headboard. Body weight (kg) was assessed with a calibrated portable beam scale with 0.5 kg resolution. Measurements of the arm, waist, hip and calf girths were performed by a previously trained registered nutritionist following standard methods 1 and using a metal tape with 1 mm resolution. All measurements were conducted first with the subject in a standing position and after with the subject in a supine position. The intraand interobserver technical error of measurement was calculated for all measurements. Intra-observer ranged from 0.2% to 0.6%, and inter-observer error ranged from 0 to 1.4%. These values are considered acceptable for trained anthropometrists 12,13. BMI was determined through the standard formula [weight (kg) / height2 (m)] 14 and BMI categories were created according to the World Health Organization cut-off values 15. Statistics The normality of variables distribution was tested using Kolmogorov-Smirnov test. Results were described as mean and standard deviation (SD). Categorical variables were reported as frequencies. Differences between proportions were assessed with Pearson χ 2 test. Body girths means obtained with the subject in both standing and supine positions were compared through Student’s t-test for related samples. Differences in body girths means obtained in the two body positions according to BMI categories 105 (normal weight or overweight/obesity) were compared using Student’s t-test for independent samples. Data set was divided into quartiles, according to the distribution of girth measurement values on both standing and supine positions. Association between girth measurements obtained in the two body positions was quantified by Intraclass Correlation Coefficient (ICC) 16. Agreement between quartiles of girth measurements performed in both standing and supine positions was determined using the weighted kappa with the Fleiss classification 17 and also by the graphical method of Bland and Altman 18. Statistical significance was set at p <0.05. All analyses were conducted with the Software Package for Social Sciences (SPSS) for Windows (version 20.0; SPSS, Inc., Chicago, IL). 106 Results Baseline characteristics of the 123 patients enrolled in the present study are displayed in Table 1, for the entire sample and stratified by age groups. Patients’ mean age (SD) was 52.7 (15.7) years and 27.6% patients were aged ≥65 years. Younger patients (aged <65 years) presented higher education level, were more likely to be single and presented better nutrition status than patients aged ≥65 years. The highest proportion of patients was independent according to Katz index. Mean BMI was 26.8 (5.1) kg/m2 and 57.7% patients presented overweight or obesity (BMI ≥25 kg/m2). Body girths mean values obtained in both standing and supine positions are presented in Table 2. Measurements obtained in standing position were systematically higher than measurements obtained in supine position (p ≤0.001). The highest difference (1.1 cm) was observed for hip girth. Nevertheless, all ICC values correspond to a strong correlation (ICC ≥0.97) 16. Moreover, the agreement between measurements is high, ranging from 81.3% to 87% and all kappa values correspond to a good agreement (0.80-1.00) according to Fleiss classification 17 107 Characteristics Entire sample (n=123) < 65 years (n=89) ≥ 65 years (n=34) p Age (years), mean (SD) 52.7 (15.7) 45.5 (11.9) 71.3 (5.3) <0.001a Gender, n (%) Women Men 57 (46.3) 66 (53.7) 45 (50.6) 44 (49.4) 12 (35.3) 22 (64.7) 0.159b Education (years), n (%) 0-4 44 (35.8) 26 (29.2) 18 (53.0) 5-12 >12 64 (52.0) 15 (12.2) 51 (57.3) 12 (13.5) 13 (38.2) 3 (8.8) 0.049b Marital status, n (%) Single Married Widowed Divorced 15 (12.2) 84 (68.3) 15 (12.2) 9 (7.3) 13 (14.6) 64 (71.9) 11 (12.4) 1 (1.1) 2 (5.9) 20 (58.8) 4 (11.8) 8 (23.5) <0.001b Table continued Table 1 – Patients’ baseline characteristics for the entire sample and according to age groups. 108 Characteristics Entire sample (n=123) < 65 years (n=89) ≥ 65 years (n=34) p Katz index, n (%) Independent Moderate and severe dependence 120 (97.6) 3 (2.4) 88 (98.9) 1 (1.1) 32 (94.1) 2 (5.9) 0.185b Nutritional status (NRS-2002), n (%) Normal Risk 114 (92.7) 9 (7.3) 86 (96.6) 3 (3.4) 28 (82.4) 6 (17.6) 0.013b BMI (kg/m2), mean (SD) 26.8 (5.1) 27.1 (5.4) 25.8 (3.9) 0.245a BMI categories (kg/m2), n (%) Normal weight Overweight and obesity 52 (42.3) 71 (57.7) 36 (40.4) 53 (59.6) 16 (47.1) 18 (52.9) 0.545b Table 1 continued SD: standard deviaton; BMI: body mass index; NRS-2002: Nutritional Risk Screening - 2002 a Independent samples t-Test b Pearson Chi-square test 109 Table 2 – Comparison between body girths in standing and supine positions. SD: standard deviation; CI: confidence interval; ICC: intraclass correlation coefficient; a Difference = Girth in standing position - Girth in supine position b Student’s t-test for related samples c Data in quartiles according to sample distribution Girths (cm) Standing, mean (SD) Supine, mean (SD) Differencea mean (95%CI) p b ICC Weighted kappac Agreement (%)c Arm 29.9 (4.3) 29.2 (4.1) 0.7 (0.6-0.9) <0.001 0.97 0.87 83.7 Waist 93.2 (14.2) 92.6 (14.0) 0.6 (0.3-0.9) 0.001 0.99 0.89 86.7 Hip 99.9 (9.9) 98.8 (10.0) 1.1 (0.8-1.4) <0.001 0.99 0.90 87.0 Calf 35.6 (3.5) 35.0 (3.5) 0.6 (0.5-0.8) <0.001 0.97 0.84 81.3 110 Comparison between the differences in body girths obtained in the two body positions across two BMI categories (normal weight and overweight or obesity) are presented in Table 3. Mean differences in body girths measured in the two body positions did not significantly change between patients with normal weight and overweight or obese patients. Agreement, kappa and ICC values are also similar between the two BMI categories. 111 Table 3 – Body girths differences for standing and supine positions, comparison between normal weight (n=52) and overweight or obese (n=71) patients. Girths (cm) Standing / supine difference, normal weight mean (SD) Standing / supine difference, overweight or obesity mean (SD) Differencea, mean (95% CI) p b Agreement (%)c normal weight (kappa)d Agreement (%)c overweight or obesity (kappa)d ICC normal weight ICC overweight or obesity Arm 0.6 (1.0) 0.8 (1.1) -0.2 (-0.6-0.1) 0.147 86.5 (0.82) 83.1 (0.80) 0.92 0.97 Waist 0.2 (0.3) 0.9 (0.2) -0.7 (-1.3-0.04) 0.063 86.5 (0.82) 88.0 (0.89) 0.95 0.99 Hip 1.2 (0.2) 1.1 (0.2) 0.1 (-0.4-0.6) 0.666 80.8 (0.74) 91.6 (0.91) 0.95 0.98 Calf 0.6 (0.1) 0.7 (0.1) -0.1 (-0.4-0.1) 0.332 80.8 (0.72) 81.7 (0.81) 0.95 0.97 SD: standard deviation; CI: confidence interval; ICC: intraclass correlation coefficient a Difference in positions for girths in normal weight - Difference in position for girths in overweight or obesity b Student’s t-test for independent samples c Data in quartiles according to sample distribution d Weighted kappa 112 Bland and Altman graphical representations (Figures 1 to 4) were displayed considering the effect of different BMI categories. For all body girths obtained in the two body positions it is possible to observe that body girth values are consistently higher for overweight or obese patients. There is little dispersion and observed differences are not dependent on the magnitude of the measurements for both normal weight and overweight patients. These representations corroborate graphically the high agreement found for measurements in the two body positions and the similarities between normal weight and overweight patients considering this agreement. Further analysis was carried out in order to assess whether these results were different according to <65 years and ≥65 years age groups. Differences between body girths obtained in the two body positions were not statistically different when stratified by the two age categories, with exception for waist girth in patients aged ≥ 65 years (waist girth mean in standing position = 97.1 cm versus waist girth mean in supine position = 96.5 cm, p = 0.08) that almost reached the statistical significance. Regarding the comparison between normal weight and overweight or obesity, there was no change after stratification by age groups. 113 Figure 2 - Bland and Altman plot to waist girth (cm) obtained in the standing and supine positions and stratified by BMI categories. Figure 1 - Bland and Altman plot to arm girth (cm) obtained in the standing and supine positions and stratified by BMI categories. 120 121 Chapter II.b Triceps skinfold compressibility in hospitalized patients – an exploratory analysis 122 123 Triceps skinfold compressibility in hospitalized patients – an exploratory analysis Ana S. Sousa1, Isabel Fonseca2, Fernando Pichel2, Teresa F. Amaral1,3 1 Faculdade de Ciências da Nutrição e Alimentação, Universidade do Porto, Porto, Portugal 2 Serviço de Nutrição e Alimentação, Centro Hospitalar do Porto, Porto, Portugal 3 UISPA-INEGI, Faculdade de Engenharia, Universidade do Porto, Porto, Portugal Key-words: anthropometry; nutritional assessment; compressibility; skinfold thickness. 124 Abstract Background: Triceps skinfold (TSF) compressibility can introduce error on the measurement and its interpretation. However, it has not been explored yet in a clinical setting. Lipotool® is a digital calliper which acquires 60 measures per second and firstly allows the study of compressibility. Therefore, the present study aims to explore through an innovative technique, TSF compressibility and its associated factors among a sample of hospitalized patients. Methods: A cross-sectional study was conducted among hospitalized adult patients. Evolution of tissue compressibility during two seconds was registered and 120 TSF values were obtained. Compressibility was determined according to time (τ) and according to the difference between the initial value and the final value (TSF difference). Multivariable linear regression models were performed in order to identify factors associated with TSF compressibility. Results: 106 patients (30.2% aged ≥65 years) composed the study sample. Time of compressibility (τ) was no significantly associated with any of the studied variables, but compressibility based on TSF difference was independently associated with TSF thickness (regression coefficient (95% CI) = 0.38 (0.01-0.05), p= 0.002) and nutritional risk (regression coefficient (95% CI) = 0.23 (0.12-1.23), p= 0.018). Conclusion: Among a sample of hospitalized patients, time of compressibility (τ) was not affected by any of the studied factors. However, undernutrition risk and the TSF thickness were factors independently associated with higher compressibility assessed by the difference between the initial and final TSF value. 125 Introduction Skinfold thickness is often used for body composition assessment due the accessibility, the non-invasive features and the ability to measure subcutaneous adiposity (1, 2). In skinfold thickness measurement with a skinfold calliper, a constant pressure is applied for a defined period of time (3, 4). The tissue’s dynamic response to this pressure is defined as compressibility (1, 5). This characteristic has been studied by comparing skinfold calliper measurements and subcutaneous fat thickness assessed by coarse methods such as imaging methods, cadaver studies and empiric comparisons (1, 5). There are underlying suppositions on the estimation from skinfold measurement: skin thickness is negligible, adipose tissue has constant characteristics and also that proportion of subcutaneous to visceral fat is equivalent in all subjects (1). Notwithstanding this, it has been previously shown that compressibility varies according to the sites of measurement and between individuals, influencing the relation between the measurement and the actual adipose tissue thickness, introducing error in the estimation of body fatness (1, 5). Gender (5), age (6), hydration status (6), skin thickness (7), subcutaneous tissue pressure (7) and site of measurement (8) have been previously described as factors associated with compressibility. Nevertheless, over the past few years, knowledge on compressibility has not significantly increased. 126 An integrated system, Lipotool®, was recently developed. This equipment consists of a digital skinfold calliper and a software application (9). The system acquires 60 measurements per second (9). Thus, this novel methodology firstly permits the study of dynamic tissue’s response evolution during the measurement(9). From all skinfold thickness sites, triceps skinfold (TSF) is the most used in clinical practice, as, along with mid-arm circumference, it integrates mid-arm muscle circumference formula, a simple method that allows for the estimation of muscle mass (10). Regarding the wide use of TSF, the minimization of error is of utmost importance in order to provide an adequate use and interpretation for clinical practice. Nonetheless, as far as we are concerned, skinfolds compressibility has not been explored yet in a clinical setting. Therefore, the present study aims to explore through an innovative technique, TSF compressibility and its associated factors among a sample of hospitalized patients. 127 Materials and Methods Study sample and design A cross-sectional study was conducted in a general university hospital. Patients were eligible to participate in the study if they were aged 18 years and over, Caucasian, conscious, cooperative and able to provide written informed consent. Critically ill patients, i.e., with a life-threatening medical or surgical condition requiring intensive care unit level care, presenting severe organ system dysfunction and needing for active therapeutic support were excluded (11). Pregnancy and patient ward isolation were also defined as exclusion criteria. Ethics This research was carried out according to the recommendations established by the Declaration of Helsinki and approved by the institutional ethics and review boards of Centro Hospitalar do Porto. All study participants provided a written informed consent. Data collection Demographical data were obtained by one trained registered nutritionist through a structured questionnaire within 72 hours of admission to hospital. Education was evaluated by the number of completed school years and the following categories were created: 0-4, 5-12 and more than 12 years. Marital status 128 was categorized as single, married or in a civil partnership, divorced and widowed. Independence in activities of daily living was assessed with the Katz index (12). Patients’ nutritional status was evaluated with Nutritional Risk Screening (NRS) 2002 (13). Standing height (cm) was measured with a metal tape (Rosscraft, Innovations Incorporated, Surrey, Canada) with a 0.1 cm resolution and with a headboard. Body weight (kg) was assessed with a calibrated portable beam scale with 0.5 kg resolution. Triceps skinfold thickness (mm) was obtained with Lipotool® digital calliper after performing the measurement during two seconds, as established by the protocol (3). All measurements were performed by the same trained registered nutritionist. The intraand interobserver technical error of measurement was calculated for all measurements. Intra-observer ranged from 0.2% to 0.9%, and inter-observer error ranged from 0 to 6.6%. These values are considered acceptable for trained anthropometrists (14, 15). Body mass index (BMI) was determined through the standard formula [weight (kg) / height2 (m)] (16) and BMI categories were created according to the World Health Organization cut-offs (17). Statistics Results were described as mean and standard deviation (SD) or as median and interquartile range (IQR) according to normality of distribution, assessed with Kolmogorov - Smirnov test. 129 Data on TSF measurements were provided by Lipotool® software and the evolution of tissue compressibility during two seconds was registered, as this method acquires 60 measurements per second. Thus, in the end of the measurement, 120 values were obtained. Therefore, compressibility was determined according to a method based on τ, tau, a measurement of time expressed in seconds, that reflects adipose tissue dynamic response to compression, being an individual characteristic. Thus, lower τ values mean that the skinfold compress faster, and, therefore, presents higher compressibility. τ value was obtained after computing the inverse of the exponent of a regression equation displayed for the 120 measurement sets of each patient. Alternatively, another method was used to define compressibility. This method was based on the difference computed between the initial value and the final value, from the 120 TSF measurements acquired by the digital calliper. Thus, high difference between initial and final TSF value corresponds to high compressibility. Data set was divided into tertiles of TSF, tertiles of τ and tertiles of difference between TSF initial and final value (TSF difference), according to the sample distribution. In order to select variables associated to compressibility, patients’ baseline characteristics were compared across τ tertiles and TSF difference tertiles. Patients’ baseline characteristics were also compared across TSF tertiles. All the comparisons were computed by One-way ANOVA test if distribution was normal, or Kruskal-Wallis test, if non-normal distribution. Categorical variables were reported as frequencies. Differences between proportions were assessed with Pearson χ 2 test.