Use of adductor pollicis muscle thickness in hospitalized or ambulatory patients: a systematic review

ABSTRACT Objective: to analyze the use of the Adductor Pollicis Muscle Thickness (APMT) as an anthropometric parameter and prognostic indicator in hospitalized or ambulatory patients. Method: systematic review carried out the Web of Science, SCOPUS and Lilacs databases. Results: Twenty-three studies were performed on critical, surgical, oncological, nephropathic and hepatopathic patients, collecting data on bibliographic reference, study site, objectives, number of patients, age group, methodology, main results and conclusion. APMT proved to be a good anthropometric parameter for evaluation of nutritional status in critical patients without edema, and surgical, oncological and nephropathic patients, but presented poor performance for diagnosis of malnutrition in hepatopathic patients. It was a good prognostic indicator for mortality in critical, nephropathic and oncological patients, and also a good predictor of hospitalization in nephropathic patients. There was an association with neurological complications in Hepatic Encephalopathy (HE) in the case of hepatophatic patients, but it was not a predictor of postoperative complications in surgical patients. Conclusion: APTM was considered a good anthropometric parameter in most clinical conditions, except in patients with liver disease and a good prognostic indicator for mortality in critical, oncological and nephropathic patients, and a predictor of neurological complications in HE. Further prognostic investigation, standardization of cutoff points and evaluation of sensitivity and specificity are required.


Introduction
Due to the limited use of sophisticated equipment for analysis of body composition in the clinical practice, for the high costs and the experience required in such procedures, anthropometric and laboratory parameters are still used for the nutritional assessment of hospitalized patients (1) . Thus, new assessement methods are needed within the hospital environment, particularly those that are simple, relatively non-invasive, and that have high sensitivity and preserved specificity (2) .
In this context, a new assessment technique of the muscle compartment called Adductor Pollicis Muscle Thickness (APMT) was introduced in 2004 and it has been used to diagnose muscle loss and consequently, malnutrition (3,4) . The incidence of malnutrition lies between 20 and 69% in hospitalized patients, being higher in critical cases and nutritional levels (4,5) that lead to increased muscle fatigue, loss of contraction force and relaxation rate of the adductor pollicis muscle (APM) (1,2) .
The opposition of the thumb is present in several activities of the daily life of humans. Since the APM is also consumed during catabolism and weakened when in disuse, its trophic condition can reflect the routine of an individual (1) . This muscle can indicate changes in the muscle composition of the whole body, including early changes arising from both malnutrition and recovery of nutritional status (2,4) .  (1) .
The accuracy and reliability of anthropometric measures are influenced by many variables, such as: equipment, technical ability, cooperation of the individual, and variety of reference standards (2) . However, this new technique has many advantages; the APM is a muscle almost devoid of adipose tissue, flat, with a well-defined anatomic point, being the only one that allows the direct measurement of its thickness without need for calculations, and prone to quick, simple, noninvasive, low-cost measurement, easily reproducible by other researchers in both ambulatory and bedridden patients (1,2,4,5) . Thus, the objective of this study is to perform a systematic review on the use of APMT as an anthropometric parameter and prognostic indicator in hospitalized or ambulatory patients, when compared to other methods of assessment of nutritional status.

Methods
The following strategy was used for selection of descriptors and formulation of the guiding question of the systematic review. The group of patients named as "P" included hospitalized or ambulatory patients; the intervention "I" consisted in the application of the APMT; the comparison methods "C" were anthropometric and prognostic; and the expected outcome "O" was a good correlation between the evaluated variables. The guiding question of the research is: "Is the APMT a good anthropometric method and prognostic indicator compared to other methods used to evaluate nutritional status in hospitalized or ambulatory patients?" The following Health Sciences Descriptors were used (DeCS) to search scientific articles: "Hospitalização" OR "Avaliação nutricional" OR "Antropometria" OR "Prognóstico" OR "Desnutrição" OR "Ambulatório Hospitalar" combined with the term "Músculo adutor do polegar", which is not indexed in DeCS, and its English and Spanish translations "Hospitalization", "Hospitalización", "Nutritional Assessment", "Evaluación nutricional", "Anthropometry", "Antropometría", "Prognosis", "Pronosis" "Malnutrition", "Desnutrición", "Servicio Ambulatorio en Hospital"; "Hospital Outpatient Clinics", "Adductor pollicis muscle" and "Músculo aductor del pulgar".

Results
Two hundred and twenty nine articles were initially searched, of which 165 were repeated and 64 were accessed. After the first selection by the criterion what the author considered to be due to the presence of edema in the hands (4) .
In 2015, a study carried out in Porto Alegre consisted of 73.5% of elderly participants, which explains the low values found (3) .
A study carried out in Asia in 2015 observed a significant difference in APMT measurements between races and genders (p < 0.05). No significant correlation was found with mortality at 28 days, hospital outcome, and ICU length of stay (p > 0.05). However, there was a significant and moderate correlation with AC and BMI (p < 0.05), and APMT values were higher when compared to other Brazilian studies (2) . A study conducted in Iran found with mortality (odds ratio 3.8, 95% confidence interval, 1.2 to 5.2, p < 0.01). A significant correlation was also found between the dominant APMT and the SOFA score for organ failure (r = -0.86, p < 0.001) (5) .

Conclusion regarding APMT
To investigate whether APMT is a valid prognostic indicator of morbidity and mortality (4) . APMT can be used as a prognostic indicator in critically ill patients, as a fast, easy and inexpensive method to estimate nutritional status, especially lean body mass. It cannot be considered an independent risk factor for ICU length of stay and days of mechanical ventilation.
To verify the relation of the APMT test with GSA and to correlate it with other anthropometric methods (3) . APMT proved to be a good and efficient method for nutritional risk assessment.
To determine whether APMT can be used as a predictive indicator of mortality (2) .
APMT is not a universally applicable measure and should be used with caution.
To determine if APMT correlates with mortality, morbidity and other anthropometric parameters (5) . APMT was considered a valuable prognostic indicator and a new low cost, reliable and easy tool to assess the nutritional status of critically ill patients.  between APMT and HGS (DH p < 0.001; NDH p < 0.001) and mortality (p < 0.05). However, there was no significant association with length of hospitalization and complications in the PP (data not shown) (7,8) .
In a more recent study (2014) conducted in the same city, APMT did not show a good association with the percentage of weight loss (p = 0.113), AMC (p = 0.806) and AMAc (p = 0.770), but correlated significantly with AC (p = 0.003), TS (p = 0.000) and BMI (p = 0.000).
APMT was also associated with gender and age, as there was a high prevalence of malnutrition in women and the elderly. The values found in the DH were higher, suggesting a more rapid atrophy of this musculature in case of inactivity resulting from malnutrition (9) .

Study conducted in Pelotas (RS) published in 2015
identified an association and a significant linear trend between APMT values and GSA categories (p < 0.001).
It also found a strong association between APMT results for muscle mass depletion and GSA for malnutrition (p < 0.001). In this study, APMT showed low sensitivity (DH 34.9% and NDH 37.7%), and high specificity (greater than 90%) for predicting malnutrition (10) .
In the results of a research carried out in Vitória (ES) in 2016, APMT showed to be correlated with BMI, AMC, AMAc, CC (p < 0.01) and GSA (p = 0.026), with AMC (p = 0.036) as the variable that most influenced APMT values (11) . In a study carried out in the city of Salvador (BA) in the year of 2016, APMT did not present a significant association with the presence of complications in the PP in either hand (DH p = 0.217 and NDH p = 0.148). However, the APMT of the NDH was significantly associated with specific infectious complications (p = 0.030) (12) . In another study conducted in the same year in Recife (PE), APMT was compared to the gold-standard GSA method, but did not present a significant association (DH p = 0.513 and NDH p = 0.842) (13). and NDH p = 0.43) (14) . Two studies carried out in Belo However, no significant association was found with age, dialysis time, TS, hemoglobin, and reactance. Higher risks of hospitalization within six months and mortality were associated with lower APMT values (17) .A study was associated with BMI (p = 0.001), where higher values were more prevalent in overweight individuals and the lowest values in those with low weight (19) . The objectives, number of patients and age group, objective and subjective evaluations used for comparison with APMT and completion of the work with oncological and nephropathic patients are described in the Another study found APMT to be significantly associated only with disease severity (p < 0.05), but also showed a low prevalence of malnutrition, a weak association with GSA (k = 0.222) and no association with the diagnosis of malnutrition and inadequate dietary intake (21) . Two studies conducted in 2015 and 2016 with non-cirrhotic hepatitis C patients and patients before and after hepatic transplantation, respectively, did not identify any individuals with malnutrition and, therefore, APMT presented poor performance (22,23) .
In the year 2016, a study conducted in the city of Botucatu (SP) in cirrhotic patients with Hepatic Encephalopathy (HE) related APMT to disease severity, so that the reduction of 1 mm in the measurement was associated with an increase of 30.7% in the degree of HE (p = 0.0177). The lower values were related to lower states of mental acuity, since measures lower than the cutoff point (6.5mm) were associated with the degrees I and II of HE (p = 0.013) (24) .
Data regarding the objectives, number of patients and age group, objective and subjective evaluations used for comparison with APMT, as well as the conclusion of the studies performed in hepatopathic patients are described in the Figure 5.

Conclusion regarding APMT
To evaluate APMT as a prognostic indicator in patients submitted to valve heart surgery (6) . Complications, mortality, length of hospital length of stay and ICU length of stay † and mechanical ventilation / Mean AM ‡ : 9.54 ± 2.5.
APMT was not considered a superior prognostic indicator, neither the other parameters evaluated.
To determine if APMT measurement is reliable for nutritional evaluation and to check its correlation with other anthropometric, biochemical and clinical parameters (7) . APMT is an easy-to-apply, low cost, reliable and safe tool for nutritional assessment in clinical practice with surgical patients.
To evaluate whether APMT and HGS |||| are reliable prognostic indicators of the postoperative evolution of major surgeries of the gastrointestinal tract (8) .
APMT was considered safe for nutritional assessment in surgical patients and a reliable prognostic indicator in the PP ¶ ¶ .
To estimate the prevalence of malnutrition based on APMT, through cutoff points for surgical patients suggested in the literature (9) . 151 (51.91 ± 15.78) BMI, AC, AMC, TS, AMAc***, percentage of weight loss and clinical diagnosis.
APMT was considered a good method for the diagnosis of both muscular depletion and malnutrition in surgical patients.
To investigate the validity of APMT as a nutritional parameter in surgical patients through comparison with the GSA (10) .
APMT was associated with nutritional status and considered a useful and easily accessible tool to confirm malnutrition.
To evaluate the correlation between APMT and anthropometric measurements, BMI and GSA in the nutritional diagnosis of surgical patients (11) . APMT is a reliable measure able to identify the risk of malnutrition and may be included in the nutritional screening of surgical patients to facilitate nutritional diagnosis.
To investigate the clinical complications in the PP of elective cardiac surgeries (12) .
APMT was associated with infectious complications in the PP.
To evaluate through GSA the nutritional status of surgical patients, comparing it with nutritional screening and objective methods (13) . APMT showed no association with GSA.

Conclusion regarding APMT
To evaluate the preoperative nutritional status of patients with tumor in the upper GIT † and to indicate which methods correlate with hospitalization and mortality (14) . Mortality and length of hospitalization / Mean DH ‡ : 12.9 ± 3.5 and NDH §: 12.0 ± 3.6.
APMT was determined as a reliable predictor of mortality in patients undergoing upper GIT tumor resection surgeries.
To perform nutritional assessment by GSA || and Glasgow score, and to verify its relation with assessment methods (15) .

43
(64.7 ± 12) GSA, Glasgow prognostic score. APMT was considered a useful parameter for classification of nutritional status of cancer patients.
To correlate the nutritional status, the Glasgow score and the complications of anticancer treatment, besides associating the GSA with other methods of nutritional assessment (16) .
APMT showed low agreement with the GSA.
To analyze the correlation between anthropometric, biochemical and EBI ¶ methods and evaluate the relationship between APMT and mortality and hospitalization (17) . APMT was considered a valid method to diagnose malnutrition, predict the risk of hospitalization and mortality in patients with CKD ¶ ¶ undergoing HD***.
To observe the behavior of APMT as a nutritional indicator in patients undergoing HD (18) . APMT assumed HG and was suggested as a indicator of nutritional status in hemodialytic patients.
To analyze the nutritional status of patients undergoing HD and the associated factors (19) .

Conclusion regarding APMT
To compare different nutritional assessment methods for diagnosis of malnutrition in cirrhotic patients (20) .
APMT was poorly associated with GSA and HGS methods, and was not associated with disease severity.
To calculate dietary intake and compare various nutritional assessment methods in chronic, cirrhotic and non-cirrhotic hepatophatic patients (21) .
APMT had a poor performance in the diagnosis of malnutrition.
To carry out an evaluation of the nutritional status of non-cirrhotic hepatitis C patients by comparing several methods and dietary assessment (22) . APMT did not present a good performance in the diagnosis of malnutrition.
To assess the nutritional status of cirrhotic patients before and after liver transplantation during one year of follow-up using various methods and compare the results (23) . APMT presented low sensitivity and was considered inappropriate for the follow-up of pre and post-liver transplantation patients.
APMT was considered a good indicator of cirrhotic complications, as it was associated with the neurological manifestations of HE.

Discussion
APMT showed to be a good anthropometric parameter in critically ill patients without edema (3)(4)(5) , which may be justified by the greater expression of critical and acute conditions in the ICU, protein depletion and muscle loss caused by both the decrease in daily activities and the hypercatabolic status of these patients (25) . APMT was also considered a good prognostic indicator for mortality in this group (4,5) , being efficient even in patients with hand edema in one of the studies (4) , although it was not considered as a good prognosis for ICU length of stay (2)(3)(4) . Perhaps, this may be explained by the fact that APMT evaluates the lean mass and whose preservation is an indication of good evolution in such patients, since malnutrition contributes to increased mortality (25) . No justification was reported as to the and GSA methods (2,5) . The combination of several tools may be more efficient to detect abnormality in the body composition of these patients (3) .
In the group of individuals submitted to surgeries, APMT also proved to be a good method for nutritional evaluation and diagnosis of malnutrition, but it was not considered a good prognostic factor for the prevalence of complications in the PP. Malnutrition in these patients is associated with surgical stress, as this procedure promotes the release of catabolic hormones that cause the degradation of muscle proteins, which is aggravated according to the nutritional deficit of the patient (9) . Thus, the potential detection of malnutrition can be justified by the good specificity of APMT verified in the studies, a characteristic that indicates a low rate of false positive results (7,10) . Furthermore, there is a correlation of APMT with anthropometric measures that happens due to the similar nature of the measures (7) . As for the non-association with PP complications, it is assumed that this result comes from the variability of this condition in function of the underlying disease severity and nutritional status prior to surgery.
In the evaluation of cancer patients, APMT was considered a good parameter for classification of nutritional status, but had low association with GSA (15,16) .
This finding was indicated by the fact that GSA detected early functional alterations, while APMT detected malnutrition and changes in body composition only in later instance (15,16) . Only one study evaluated APMT as a prognostic indicator and found it to be a good parameter of mortality, resulting from its high discriminative power (14) .
With regard to nephropathic patients, APMT was also considered a good method of nutritional assessment compared to anthropometric parameters (17,19) and a good predictor of hospitalization and mortality in the only study that evaluated its prognostic indication (17) , but the number of studies with this group is still small (17,18) . This outcome is based on the peculiarities of CKD patients, namely, reduced food intake and metabolic acidosis combined with uremia, factors that lead to protein catabolism (19) , which along with physical inactivity, further aggravates the disease (17) . Furthermore, malnutrition is per se a risk factor for death in hemodialysis patients (17) .
However, studies evaluating hepatopathic patients revealed poor performance of APMT in the diagnosis of malnutrition, and only one showed an association with neurological complications in HE.
Two studies differed in relation to the severity of the disease. Non-association of APMT with malnutrition can be explained by fluid retention (24) , low sensitivity demonstrated by the method (23) and anatomical muscle alteration that follows functional changes (21) .
An association of APMT with HE complications and disease severity was observed, most likely explained by the marked depletion of arm muscle experienced by cirrhotic patients (24) . The study in which there was no association with disease severity was justified by the low number of patients in the severe stage (20) . speed up the nutritional screening process (10,11,15,25) .
Among the ten studies comparing GSA with APMT, seven found a positive association between them and Soares BGFS, Vicentini AP.
Studies with healthy individuals were excluded from this review because they aimed to establish reference values for this population (26,27) , and the research with  (2,11) .
Thus, it is necessary to determine specific cutoff points for pathologies or groups of individuals and carefully interpret the results. It is also necessary to assess the actual sensitivity and specificity of APMT. In this sense, sensitivity is a parameter that evaluates the probability of the result to be positive, and specificity analyzes the probability of that same result to be negative (29) .

Conclusion
APMT was considered a good anthropometric parameter in the great majority of clinical conditions evaluated, except in patients with liver disease, and presented low sensitivity and high specificity. It was also indicated as a good prognostic indicator for mortality in critical, oncological and renal patients, and as a predictor of neurological complications in hepatic encephalopathy.
However, further investigation of its prognostic usefulness in other clinical conditions, standardization of cutoff points for reference in the measurement classification and assessment of sensitivity and specificity is required.