- Email: [email protected]
Validity of the adductor pollicis muscle as a component of nutritional screening in the hospital setting: A systematic review
Clinical Nutrition ESPEN xxx (2016) 1e7
Contents lists available at ScienceDirect
Clinical Nutrition ESPEN journal homepage: http://www.clinicalnutritionespen.com
Review
Validity of the adductor pollicis muscle as a component of nutritional screening in the hospital setting: A systematic review Charles Chin Han Lew a, b, *, Fangyi Ong c, Michelle Miller a a
Discipline of Nutrition and Dietetics, Flinders University, Australia Dietetics and Nutrition Department, Ng Teng Fong General Hospital Singapore c Dietetics Department, National University Hospital, Singapore b
a r t i c l e i n f o
s u m m a r y
Article history: Received 27 August 2016 Accepted 31 August 2016
Background & aims: The thickness of the adductor pollicis muscle (TAPM) is the only muscle that can be directly measured with a caliper. Recent studies demonstrate that the TAPM could be used as part of nutritional screening, but this has not been systematically reviewed. We aimed to review the validity and reliability of TAPM in identifying malnutrition risk. Methods: PubMed, CINAHL and Scopus were systematically searched. Eligible studies were crosssectional, case-control or cohort studies which recruited adult patients in the hospital, measured the TAPM along with a valid nutritional assessment tool and evaluated the validity of TAPM with univariate, multivariate, discriminative and/or agreement statistics. The Quality Assessment of Diagnostic Accuracy Studies-II was used to evaluate the risk-of-bias. Results: Nine out of 39 studies identified were eligible, and all had some risk-of-bias. Subjective Global Assessment (SGA) was used as the criterion standard in all nine studies. The TAPM amongst malnourished patients were significantly different from their counterparts but this may be over-estimated since all measurements of the SGA and TAPM were non-blinded. Concordance between the TAPM and SGA ranged from poor to good (kappa ranged from 0.04 to 0.25; specificity ranged from 97.8% to 100%), and this may be due to the varied cut-off values used amongst the studies. There were also disparities in the measurement instruments and methods. Reliability was not reported. Conclusion: More studies are needed to establish the reliability of TAPM measurement and cut-off values before the TAPM can be used as a component of nutritional screening. © 2016 European Society for Clinical Nutrition and Metabolism. Published by Elsevier Ltd. All rights reserved.
Keywords: Adductor pollicis Malnutrition Nutritional screening
1. Introduction About 25e40% of patients are malnourished at hospital admission [1e3]. Malnutrition increases short and long-term mortality, length of hospitalization and healthcare costs [4e6]. To reduce the negative impact of malnutrition, nutritional interventions should be timely [7], and this can only be achieved if patients at risk of malnutrition are promptly identified. Therefore, the Joint Commission International stipulates that nutritional screening should be carried out within 24 h of hospital admission [8].
Abbreviations: APM, Adductor pollicis muscle; ROC, Area under the receiver operative characteristics curve; SGA, Subjective Global Assessment; TAPM, Thickness of the adductor pollicis muscle. * Corresponding author. 1 Jurong East Street 21, 609606, Singapore. E-mail address: [email protected] (C.C.H. Lew).
There are more than 30 nutritional screening tools reported in the literature, each using different combinations of variables (e.g. unintentional weight loss and insufficient calorie intake) to determine malnutrition risk [9]. Loss of muscle mass is one of the hallmarks of malnutrition [10,11]. There are several ways to measure loss of muscle mass in the clinical setting, namely bio-impedance analysis, physical examination and anthropometry. The accuracy of bio-impedance analysis is variable especially in the clinical setting as several conditions that are highly prevalent in the hospital (e.g. edema, hypoalbuminemia and some medications) can confound the results [12]. Concerning physical examination and anthropometry where accuracy are both skill dependent, anthropometry may have an advantage over physical examination as it is relatively more objective in measuring muscle loss. In the clinical setting, anthropometry for the measurement of muscle loss is commonly quantified by the mid-arm muscle area
http://dx.doi.org/10.1016/j.clnesp.2016.08.005 2405-4577/© 2016 European Society for Clinical Nutrition and Metabolism. Published by Elsevier Ltd. All rights reserved.
Please cite this article in press as: Lew CCH, et al., Validity of the adductor pollicis muscle as a component of nutritional screening in the hospital setting: A systematic review, Clinical Nutrition ESPEN (2016), http://dx.doi.org/10.1016/j.clnesp.2016.08.005
2
C.C.H. Lew et al. / Clinical Nutrition ESPEN xxx (2016) 1e7
[13]. However, it is an indirect measurement of muscle mass as both the areas of subcutaneous fat and bone in the mid-arm region are adjusted in the calculation of muscle mass. This has been shown to over-estimate muscle mass by up to 25% [14]. In contrast, the thickness of the adductor pollicis muscle (APM) is the only muscle that can be directly measured with a caliper [15]. The APM is located between the thumb and index finger. It is easily accessible and can be directly measured because it is anatomically well-defined, flat in shape and has a minimal amount of subcutaneous fat surrounding it [15]. Recent studies demonstrated that the TAPM is a good surrogate measurement for total muscle mass [13,16]. Therefore, TAPM could potentially be used in combination with other conventional nutritional parameters as part of nutritional screening if it can independently discriminate well- and malnourished patients at the univariate level [17]. We aim to systematically review the validity of the TAPM against valid nutritional assessment tools in identifying malnutrition risk amongst adults in the hospital setting. In addition, we aim to report on the reliability of TAPM measurement. 2. Methods 2.1. Protocol and registration The reporting of this systematic review is in accordance to the PRISMA-statement and the protocol is registered on PROSPERO (Registration number: CRD42015023261). 2.2. Eligibility criteria Since the measurement of TAPM is relatively new, a broad eligibility criteria was developed to maximize the possibility of gathering all relevant articles. All cross-sectional, case-control and cohort studies were eligible for inclusion. In addition, all studies that included the APM and nutritional status as measurement variables were included. No restriction was imposed on publication date and language.
non-pregnant patients in the hospital were included. Articles that used electrical stimulation of the ulnar nerve to evaluate the strength of contraction and maximum rate of slackness of the APM were excluded. This method is not feasible in clinical practice as it requires specialized equipment, and the procedure is painful for the patient. Instead, articles that measured the TAPM along with a valid nutritional assessment tool (i.e. Subjective Global Assessment [SGA] [18] and/or Mini-nutritional Assessment) [19] were included in the review. In addition, articles included used appropriate statistical methods to evaluate the validity of TAPM in differentiating well- and malnourished patients. According to Jones [17], variables that could potentially be included as part of nutritional screening should minimally be able to differentiate well- and malnourished individuals at a univariate level. Therefore, only studies that compared the mean or median of the TAPM between well- and malnourished patients, reported discriminative statistics and/or agreement statistics were included. The PRISMA flow diagram was used to summarize the article selection processes (Fig. 1). 2.5. Data extraction Data were extracted and grouped into four categories as follows: 1) study design, country, patient characteristics (i.e. age and primary diagnosis); 2) timing and assessor the TAPM and nutritional assessment; 3) the prevalence of malnutrition determined by nutritional assessment; and 4) descriptive statistics (i.e. mean, standard deviation, median and inter-quartile range), univariate statistics (i.e. one-way analysis of variance and correlation), discriminative statistics (i.e. sensitivity, specificity and area under the receiver operative characteristics curve [ROC]), multivariate analysis, agreement statistics (i.e. kappa), and reliability statistics (i.e. intra- and inter-assessor error). We contacted five authors [20e24] for further information, and all of them responded
2.3. Information sources and search methods A search strategy was developed with reference to the eligibility criteria and three electronic databases, namely PubMed, CINAHL and Scopus were systematically searched on 2nd May 2015. To maximize the possibility of gathering all relevant studies, both free text terms and broad search terms (i.e. MeSH in PubMed, and CINAHL Headings in CINAHL) were used. Synonyms of ‘malnutrition’ and ‘nutritional status’ were combined with synonyms of ‘screening’ and ‘assessment’ to identify all tools used to measure nutritional status [4]. Results of the latter were combined with all articles that included measurements of the APM. This search strategy was appropriately adapted in all three electronic databases to ensure consistency (Supplementary material). To further ensure all relevant articles were identified, we hand searched the reference lists of the articles that were included in this systematic review. 2.4. Study selection There were a manageable number of articles upon the removal of duplicates. Therefore, instead of screening the titles and abstracts, two reviewers (CHCL and FO) assessed the relevance of the studies independently by evaluating the full-text version of the articles. In all cases of disagreement, consensus was sought through discussion. This review focused on the TAPM measured in a hospital setting. Therefore, only studies that recruited adult (>18 years old) and
Fig. 1. Summary of each stage of the search methods and the quantity of articles retrieved, excluded and included in this systematic review.
Please cite this article in press as: Lew CCH, et al., Validity of the adductor pollicis muscle as a component of nutritional screening in the hospital setting: A systematic review, Clinical Nutrition ESPEN (2016), http://dx.doi.org/10.1016/j.clnesp.2016.08.005
Authors
n
Age (y)
Measurement details
Malnutrition prevalence
Univariate/Multivariate statistics
Discriminative statistics
Bragagnolo et al. [26]
87
53.8 ± 15.4 TAPM and SGA measured within 48 h of admission, by NSP
SGA-A: 11.5% SGA-B: 29.9% SGA-C: 58.6% DTAPM: 62.8% NDTAPM: 65.9%
Mean DTAPM and NDTAPM of SGA-A was significantly higher than SGA-B, p < 0.001 Mean DTAPM and NDTAPM of SGA-B was significantly higher than SGA-C, p < 0.05
SensitivityDTAPM: NA 72.4% NDTAPM: 77.3% SpecificityDTAPM: 100.0% NDTAPM: 100.0% ROCDTAPM: 0.93 (95% CI: 0.86 to 0.99) NDTAPM: 0.92 (95% CI: 0.85 to 0.98)
Bragagnolo et al. [30]
90
53 ± 16
TAPM and SGA measured within 24 h of admission, by NSP
SGA-A: 14.4% SGA-B: 32.2% SGA-C: 53.3% TAPM: NA
Multi-linear regressionCompared to SGA-B and SGA-C, DTAPM and NDTAPM were 4.7 mm thicker in SGA-A, p < 0.001
NA
NA
Caporossi et al. [20]
246 62*
NA
SGA-A: 21.9% SGA-B: 53.7% SGA-C: 24.4% TAPM: NA
TAPM of the right hand (mm)SGA-A: 17.2 ± 5.4 NA SGA-B: 16.8 ± 5.7 SGA-C: 12.9 ± 5.3y TAPM of the left hand (mm)SGA-A: 15.8 ± 4.6 SGA-B: 15.9 ± 5.9 SGA-C: 12.3 ± 5.5y NA NA
NA
Nunes et al. [27]
119 56.3 ± 12.0 TAPM and SGA measured at outpatient visit, by NSP
SGA-A: NA SGA-B and SGA-C: 6.7% TAPM: 14.3%
Agreement statistics
Kappa: 0.25
SGA-A: 30.0% SGA-B: 24.3% SGA-C: 45.7% TAPM: 13.3% SGA-A: 52.0% SGA-B: 46.6% SGA-C: 1.4% TAPM: NA
TAPM (mm)SGA-A: 24.3 ± 4.2‡ SGA-B: 22.5 ± 6.5 SGA-C: 20.0 ± 5.1
NA
Kappa: 0.04, p < 0.05
No difference in the number of patients with TAPM lower or higher than 10 mm in each SGA category, p ¼ 0.55
NA
NA
TAPM and SGA measured before radio-/chemotherapy, by Dietitians (number NA)
SGA-A: 13.9% SGA-B: 13.9% SGA-C: 72.2% TAPM: 44.2%
TAPM (mm)*SGA-A: 21.0 SGA-B: 21.5 SGA-C: 16.5x
NA
Kappa: < 0.20
TAPM and SGA measured within 72 h of admission by two trained Nutritionists
PG-SGA-A: 52.1% PG-SGA-B: 24.1% PG-SGA-C: 23.8% TAPM: NA
Correlation between TAPM and SGA r ¼ 0.194, p < 0.001
NA
NA
SGA-A: 68.4% SGA-B: 24.1% SGA-C: 7.5% TAPM: NA
DTAPM (Male) (mm)*SGA-A: 26.0 (25.0, 28.0)‡ SensitivityDTAPM: 34.9% SGA-B: 19.5 (16.5, 22.0) NDTAPM: 37.7% SGA-C: 18.0 (15.0, 20.0) DTAPM (Female) (mm)*SGA-A: 23.0 (21.0, 25.0)‡
Mauricio et al. [21] 70
60.4 ± 14.3 TAPM and SGA measured before radio-/ chemotherapy, by Dietitians (number NA)
Pereira et al. [22]
73
52.3 ± 17.0 TAPM and SGA measured after haemodialysis, by two trained Nutritionists
Silva et al. [23]
43
NA
Guerra et al. [29]
688 58* (21)
Gonzalez et al. [24]
361 49.6 ± 17.8 TAPM and SGA measured by trained personnel (number and timing NA)
C.C.H. Lew et al. / Clinical Nutrition ESPEN xxx (2016) 1e7
Please cite this article in press as: Lew CCH, et al., Validity of the adductor pollicis muscle as a component of nutritional screening in the hospital setting: A systematic review, Clinical Nutrition ESPEN (2016), http://dx.doi.org/10.1016/j.clnesp.2016.08.005
Table 1 Summary of the validity of the TAPM in differentiating nutritional status.
NA
3
(continued on next page)
SpecificityDTAPM: SGA-B: 18.0 (16.0, 20.0) 98.7% SGA-C: 17.0 (15.0, 18.0) ‡ NDTAPM (Male) (mm)*SGA-A: 25.0 (24.0, 28.0) NDTAPM: 97.8% SGA-B: 18.5 (15.5, 20.0) SGA-C: 16.0 (15.0, 20.0) NDTAPM (Female) (mm)*SGA-A: 22.0 (20.0, 25.0)‡ SGA-B: 18.0 (16.0, 20.0) SGA-C: 16.0 (12.0, 18.0) Correlation with SGADTAPM: r ¼ 0.61, p < 0.05 NDTAPM: r ¼ 0.60, p < 0.05 Multivariate linear regressionDTAPM and NDTAPM of SGA-B and -C were 4.59 mm and 4.56 mm, and 6.51 mm and 6.14 mm thinner than SGA-A respectively, p < 0.001
Discriminative statistics Univariate/Multivariate statistics Malnutrition prevalence Measurement details Age (y)
Gonzalez et al. [24]
n Authors
Table 1 (continued )
Values are means ± standard deviation unless stated, * Median and interquartile range in parenthesis (when available), y p < 0.001 SGA-C compared to SGA-A and SGA-B, ‡ p < 0.05 SGA-A compared to SGA-B and SGA-C, x p < 0.05 SGA-C compared to SGA-B, DTAPM: Dominant thickness of the adductor pollicis muscle, NA: Not available, NDTAPM: Non-dominant thickness of the adductor pollicis muscle, NSP: Non-specific personnel, PG-SGA: PatientGenerated Subjective Global Assessment, ROC: area under the receiver operative characteristics curve, SGA: Subjective Global Assessment, TAPM: Thickness of the adductor pollicis muscle.
C.C.H. Lew et al. / Clinical Nutrition ESPEN xxx (2016) 1e7
Agreement statistics
4
[20e24]. One reviewer (CHCL) extracted the required data and the second reviewer (FO) checked the extracted data. Disagreements were resolved by discussion. If no consensus was reached, a third reviewer's (MM) opinion was sought. 2.6. Study appraisal and synthesis An evidence-based quality assessment tool, i.e. Quality Assessment of Diagnostic Accuracy Studies e II (QUADAS-II) [25] with high construct validity, interrater reliability, and internal consistency was used to evaluate the diagnostic validity of the studies included in this review. Essentially, it is a checklist of seven items that assess the risk of bias in four main domains (i.e. patient selection, index test, reference test and patient flow). The tool also assesses the applicability of the diagnostic studies with reference to the review question. Each domain was scored as “low risk”, “high risk” or “unclear risk” according to the detailed scoring criteria outlined by Whiting et al. [25]. The critical appraisal was independently performed by two reviewers (CHCL and FO), and any disagreement between the two reviewers were resolved by discussion or a third reviewer's (MM) opinion was sought if no consensus was reached. A meta-analysis was not performed as measurements of the TAPM were reported in means and medians. In addition, discriminative statistics were not pooled as the TAPM cut-off points for malnutrition used were different, i.e. Bragagnolo et al. [26] used cut-off values derived in their study whereas Gonzalez et al. [24] used cut-off values derived previously from a group of wellnourished subjects [15]. Similarly, agreement statistics were not pooled as the TAPM cut-off points for malnutrition were different, i.e. Nunes et al. [27] used cut-off values derived by Lameu et al. [28] whereas Mauricio et al. [21] and Silva et al. [23] used cut-off values derived by Gonzalez et al. [15]. 3. Results The literature search generated 73 articles. Upon the removal of duplicates and studies that did not meet the eligibility criteria, a total of nine studies were included in the review. The flow chart of the selection process is provided in Fig. 1. Of the nine studies, eight were carried out in Brazil, and one was carried out in Portugal [29]. All studies were written in English except for two written in Portuguese [26,27]. They were cross-sectional studies except for the studies conducted by Caporossi et al. [20] and Bragagnolo et al. [30] which were prospective cohort studies. Most of the studies recruited medical patients, i.e. those with cancer [21,23], critical illness [20], liver [27] and renal disease [22], and a group of heterogeneous inpatients [29]. The other three studies recruited surgical patients [24,26,30]. The mean or median age of the subjects were between 50 and 62 years old. Malnutrition was diagnosed by the SGA [20e24,26,27,30] or the Patient-Generated Subjective Global Assessment [29], and the prevalence of malnutrition in the inpatient setting ranged from 31.6% [24] to 88.5% [26], and it was 6.7% [27] in the outpatient setting (Table 1). All studies had some risk of bias (Table 2). In the domain of Index Test, both the SGA and the TAPM were carried out in a non-blinded fashion. In the domain of Reference Standard, two studies had partial verification bias because the SGA was not performed on all the patients and the rationale was not reported [23,24]. There were applicability concerns in three studies. Silva et al. [23] and Nunes et al. [27] did not provide details on the methods used to perform the SGA and/or TAPM measurements, and Pereira et al. [22] and Silva et al. [23] did not report the caliper used. The procedures used to measure TAPM were similar in seven studies [21e24,26,29,30]. Measurements were taken while subjects
Please cite this article in press as: Lew CCH, et al., Validity of the adductor pollicis muscle as a component of nutritional screening in the hospital setting: A systematic review, Clinical Nutrition ESPEN (2016), http://dx.doi.org/10.1016/j.clnesp.2016.08.005
C.C.H. Lew et al. / Clinical Nutrition ESPEN xxx (2016) 1e7
5
Table 2 Assessment of methodological risk of bias based on QUADAS-II. Study
Bragagnolo et al. [26] Bragagnolo et al. [30] Caporossi et al. [20] Nunes et al. [27] Mauricio et al. [21] Pereira et al. [22] Silva et al. [23] Gonzalez et al. [24] Guerra et al. [29]
Risk of bias
Applicability concerns
Patient selection
Index test
Reference standard
Flow and timing
Patient selection
Index test
Reference standard
þ þ þ þ þ þ þ þ þ
e e e e e e e e e
þ þ þ þ þ þ e e þ
þ þ þ þ þ þ þ þ þ
þ þ þ þ þ þ þ þ þ
þ þ þ e þ e e þ þ
þ þ þ þ þ þ e þ þ
þ Low risk; e High risk; ? Unclear risk.
were seated, elbows bent at a 90-degree angle and hands resting on the knees. Caporossi et al. [20] however had to modify the procedure according to the limitation of critically ill patients. Measurements were taken while subjects were in a supine position, elbows bent at a 90-degree angle and hands lying on the upper abdomen. Although the procedures were mostly similar, the recorded measurements were derived differently. Most studies used the mean of three consecutive measurements [20,22,24,26,29,30] but other studies used the highest value of three measurements [21,23]. Nunes et al. [27] did not report how the recorded measurements were derived. Calipers used to measure the TAPM were also different amongst the studies. The Cescorf caliper was used in four studies [20,26,27,30], Lange caliper in two studies [21,24], and Harpenden caliper in one study [29]. Pereira et al. [22] and Silva et al. [23] did not report the caliper used. The intra- and inter-assessor reliability of TAPM measurement were not reported in all the studies. Furthermore, the clinicians who carried out the TAPM measurements were not identified in more than half of the studies [20,24,26,27,30]. When specified, measurements were performed by either two nutritionists [22,29] or undefined number of dietitians [21,23]. The difference in the TAPM between well- and malnourished patients was determined mostly by univariate analyses. One-way analysis of variance was used in several studies, and it showed mixed results (Table 1) [20,21,23,24,26]. Only Bragagnolo et al. [26] demonstrated that the TAPM increased significantly with better nutritional status (i.e. SGA-A > SGA-B > SGA-C). This increase in the TAPM with each category of the SGA was not observed in other studies [20,21,23,24]. Gonzalez et al. [24] demonstrated that the APM of well-nourished patients (SGA-A)
were significantly thicker than malnourished patients (SGA-B and SGA-C combined). On the other hand, Mauricio et al. [21], Caporossi et al. [20] and Silva et al. [23] demonstrated that the APM of both well- and mildly-moderately malnourished patients (SGA-A and SGA-B combined) were significantly thicker than severely malnourished patients (SGA-C). Other univariate analyses quantified the correlation and agreement (i.e. kappa) between the TAPM and nutritional status diagnosed by the SGA [21e24,27,29]. Correlation analyses showed mixed results as Gonzalez et al. [24] found moderate negative correlation whereas Pereira et al. [22] and Guerra et al. [29] found either no correlation or weak negative correlation between the SGA and the TAPM. Kappa analyses were performed in three studies, and the agreement between the TAPM and SGA were consistently poor (kappa ranged from 0.04 to 0.25) [21,23,27]. Multilinear regression analysis was used in two studies to determine the difference in the TAPM between well- and malnourished patients. Both studies showed that the TAPM of wellnourished patients were thicker than their counterparts [24,30]. After adjusting for sex, age and/or weight, the APM of the dominant and non-dominant hands in well-nourished patients were at least 4.6 mm thicker than the malnourished patients [24,30]. Only two studies used discriminative statistics [24,26] (i.e. ROC, sensitivity and specificity analyses) and they showed that the TAPM could differentiate well-nourished patients from malnourished patients (combination of mildly-moderately and severely malnourished). Specificity of the TAPM in discriminating wellnourished patients from malnourished patients were high (Table 1) [24,26]. However, the cut-off values used in both studies were different.
Table 3 Thickness of the adductor pollicis muscle cut-off values used to define malnutrition risk. Authors
Method used to determine the TAPM cut-off value to define malnutrition risk
DTAPM (mm)
NDTAPM (mm)
Male
Female
All
Male
Female
All
Bragagnolo et al. [26] Bragagnolo et al. [30] Caporossi et al. [20] Nunes et al. [27]
Receiver Operative Characteristics NA 10th percentile of the subjects' TAPM Referenced Lameu et al. [28] where 421 healthy subjects aged between 18 and 87 years old were recruited. Mean cut-off values were derived from values that were more than one standard deviation away from the mean, and the median.
NA NA NA 9.5y 11.0‡
NA NA NA 8.0y 9.0‡
13.4 NA 9.5 NA NA
NA NA NA 9.5y 11.0‡
NA NA NA 8.0y 9.0‡
13.1 NA 8.3 NA NA
Mauricio et al. [21] Silva et al. [23] Gonzalez et al. [24] Pereira et al. [22] Guerra et al. [29]
All three studies referenced Gonzalez et al. [15] where 300 healthy subjects aged between 18 to 90 years old were recruited. Cut-off values of were derived from the 5th percentile of the subjects' TAPM Median of the subjects' TAPM measurements NA
20.0x 23.0¶ 18.0yy NA NA
16.0x 17.0¶ 14.0yy NA NA
NA NA NA 10‡ NA
19.0x 21.0¶ 16.0yy NA NA
15.0x 16.0¶ 14.0yy NA NA
NA NA NA 10‡ NA
y
mean, ‡ median, x 18e29 years old, ¶ 30e59 years old, yy 60 years old, DTAPM: Dominant thickness of the adductor pollicis muscle, NA: Not available, NDTAPM: Nondominant thickness of the adductor pollicis muscle, TAPM: Thickness of the adductor pollicis muscle.
Please cite this article in press as: Lew CCH, et al., Validity of the adductor pollicis muscle as a component of nutritional screening in the hospital setting: A systematic review, Clinical Nutrition ESPEN (2016), http://dx.doi.org/10.1016/j.clnesp.2016.08.005
6
C.C.H. Lew et al. / Clinical Nutrition ESPEN xxx (2016) 1e7
There were vast variations of cut-off values used to define nutritional status. Nunes et al. [27] used cut-off values provided by Lameu et al. [28] whereas Mauricio et al. [21], Silva et al. [23] and Gonzalez et al. [24] used cut-off values provided by Gonzalez et al. [15]. Other studies self-defined their cut-off values using the ROC analysis with the SGA as the criterion standard [26], or the 10th [20] and 50th [22] percentile of the TAPM measurements of recruited subjects. Table 3 summarizes the cut-off values used in the studies included in the review. 4. Discussion To our knowledge, this is the first systematic review that evaluated the validity and reliability of the TAPM in identifying malnutrition risk. Our results show that the TAPM is not an appropriate component of nutritional screening until further studies evaluate the limitations highlighted in the review. Although univariate and multivariate analyses demonstrated that the TAPM amongst malnourished patients were significantly different from their counterparts, such results may be overestimated because there were some risk-of-bias, i.e. the SGA and TAPM were performed in a non-blinded fashion. Concordance between the TAPM and SGA were mixed. The ROC showed that the TAPM had excellent discrimination (area under the curve: 0.92) [26] and high specificity (ranged from 97.8% to 100%) [24,26] in differentiating well-nourished patients from the malnourished. However, kappa statistics (ranged from 0.04 to 0.25) [21,23,27] demonstrated poor concordance. The disparate findings were likely due to differences in the cut-off values used. Different cut-off values were used to define malnutrition risk (Table 3). Some studies self-derived the cut-off values [20,22,26] whereas other studies [21,23,24,27] used cut-off values derived from healthy populations [15,28]. Lameu et al. [28] and Gonzalez et al. [15] measured the TAPM of well-nourished and healthy individuals and established cut-off values that define malnutrition risk. These cut-off values, however may have limited generalizability as all the subjects were monoethnic, i.e. Brazilians. The applicability of using the TAPM as a component of nutritional screening may be limited by the lack of measurement standardization and reliability measurement. The methods and calipers used to measure the TAPM were variable amongst the studies. The consequence of such differences is demonstrated by the results of Lameu et al. [28] and Gonzalez et al. [15] where cut-off values used to define malnutrition risk were vastly different. Lameu et al. [28] used the Lange caliper and the mean of three measurements whereas Gonzalez et al. [15] used the Crecorf caliper and the highest value of three measurements. These differences could explain the discrepancies in the results of the studies included in the review. None of the studies reported the reliability of measuring the TAPM. Since a clinically useful tool requires both validity and reliability [31], the absence of intra- and inter-reliability data preclude the use of the TAPM in the hospital setting. 5. Future studies Given that body composition can be affected by ethnicity [32], future studies should measure the TAPM amongst healthy individuals from different ethnicities and use similar methods to establish cut-off values that are hand, age, sex and ethnic specific to define malnutrition risk. In addition, future studies should measure the intra- and inter-reliability of measuring the TAPM. This is especially important because such measurement is skilldependent. It is ideal for nutritional screening tools to be parsimonious. Future studies should also determine the significance of including
the TAPM with other conventional nutritional parameters as part of nutritional screening via a multivariate model. Interestingly, the TAPM was demonstrated to have mortality prognostic value in three studies [20,30,33], but such result was not consistently observed [34]. More studies should be carried out to evaluate the prognostic value of the TAPM in the hospital setting. 6. Conclusion This review presented a body of evidence highlighting the limitations of using the TAPM as a component of nutritional screening in the hospital setting. For the TAPM to be considered a valid component of nutritional screening, future studies should determine the reliability of measuring the TAPM as well as establish cut-off values that are hand, age, sex, and ethnic specific to define malnutrition risk. Sources of support Nil. Authors' contribution All authors designed the research, analyzed the studies included in the review, and read and approved the final manuscript; C. C. H. Lew and F. Ong conducted the literature search and prepared the draft manuscript; C. C. H. Lew had primary responsibility for final content. Conflict of interest statement and funding sources None of the authors reported a conflict of interest related to the study. This review did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. Acknowledgments We are grateful to Claudia Canavarro for translating the articles written in Portuguese into English. Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.clnesp.2016.08.005. References [1] Kelly IE, Tessier S, Cahill A, Morris SE, Crumley A, McLaughlin D, et al. Still hungry in hospital: identifying malnutrition in acute hospital admissions. QJM 2000;93:93e8. [2] Kruizenga HM, Van Tulder MW, Seidell JC, Thijs A, Ader HJ, Van Bokhorst-de van der Schueren MA. Effectiveness and cost-effectiveness of early screening and treatment of malnourished patients. Am J Clin Nutr 2005;82: 1082e9. [3] McWhirter JP, Pennington CR. Incidence and recognition of malnutrition in hospital. BMJ 1994;308:945e8. [4] Lew CCH, Yandell R, Fraser RJ, Chua AP, Chong MF, Miller M. Association between malnutrition and clinical outcomes in the intensive care unit: a systematic review. JPEN J Parenter Enter Nutr 2016. http://dx.doi.org/10.1177/ 0148607115625638. [5] Lim SL, Ong KC, Chan YH, Loke WC, Ferguson M, Daniels L. Malnutrition and its impact on cost of hospitalization, length of stay, readmission and 3-year mortality. Clin Nutr 2012;31:345e50. [6] Correia MI, Waitzberg DL. The impact of malnutrition on morbidity, mortality, length of hospital stay and costs evaluated through a multivariate model analysis. Clin Nutr 2003;22:235e9. [7] Rasheed S, Woods RT. Malnutrition and quality of life in older people: a systematic review and meta-analysis. Ageing Res Rev 2013;12:561e6.
Please cite this article in press as: Lew CCH, et al., Validity of the adductor pollicis muscle as a component of nutritional screening in the hospital setting: A systematic review, Clinical Nutrition ESPEN (2016), http://dx.doi.org/10.1016/j.clnesp.2016.08.005
C.C.H. Lew et al. / Clinical Nutrition ESPEN xxx (2016) 1e7 [8] Joint Commission International, Joint Commission on Accreditation of Healthcare Organizations. Joint commission international accreditation standards for hospitals. 5th ed. Joint Commission Resources; 2015. [9] van Bokhorst-de van der Schueren MA, Guaitoli PR, Jansma EP, de Vet HC. Nutrition screening tools: does one size fit all? A systematic review of screening tools for the hospital setting. Clin Nutr 2014;33:39e58. [10] White JV, Guenter P, Jensen G, Malone A, Schofield M. Consensus statement of the Academy of Nutrition and Dietetics/American Society for Parenteral and Enteral Nutrition: characteristics recommended for the identification and documentation of adult malnutrition (undernutrition). J Acad Nutr Diet 2012;112:730e8. [11] Cederholm T, Bosaeus I, Barazzoni R, Bauer J, Van Gossum A, Klek S, et al. Diagnostic criteria for malnutrition e an ESPEN consensus statement. Clin Nutr 2015;34:335e40. [12] Kushner RF, Gudivaka R, Schoeller DA. Clinical characteristics influencing bioelectrical impedance analysis measurements. Am J Clin Nutr 1996;64:423Se7S. [13] Lameu EB, Gerude MF, Campos AC, Luiz RR. The thickness of the adductor pollicis muscle reflects the muscle compartment and may be used as a new anthropometric parameter for nutritional assessment. Curr Opin Clin Nutr Metab Care 2004;7:293e301. [14] Heymsfield SB, McManus C, Smith J, Stevens V, Nixon DW. Anthropometric measurement of muscle mass: revised equations for calculating bone-free arm muscle area. Am J Clin Nutr 1982;36:680e90. [15] Gonzalez MC, Duarte RR, Budziareck MB. Adductor pollicis muscle: reference values of its thickness in a healthy population. Clin Nutr 2010;29:268e71. [16] de Oliveira CM, Kubrusly M, Mota RS, Choukroun G, Neto JB, da Silva CA. Adductor pollicis muscle thickness: a promising anthropometric parameter for patients with chronic renal failure. J Ren Nutr 2012;22:307e16. [17] Jones JM. Development of a nutritional screening or assessment tool using a multivariate technique. Nutrition 2004;20:298e306. [18] Detsky AS, McLaughlin JR, Baker JP, Johnston N, Whittaker S, Mendelson RA, et al. What is subjective global assessment of nutritional status? JPEN J Parenter Enter Nutr 1987;11:8e13. [19] Guigoz Y, Vellas B, Garry P, Vellas B, Albarede J. Mini nutritional assessment: a practical assessment tool for grading the nutritional state of elderly patients. The mini nutritional assessment: MNA Nutrition in the elderly. 1997. p. 15e60. [20] Caporossi FS, Caporossi C, Borges Dock-Nascimento D, de AguilarNascimento JE. Measurement of the thickness of the adductor pollicis muscle as a predictor of outcome in critically ill patients. Nutr Hosp 2012;27:490e5. [21] Mauricio SF, da Silva JB, Bering T, Correia MI. Relationship between nutritional status and the Glasgow Prognostic Score in patients with colorectal cancer. Nutrition 2013;29:625e9.
7
[22] Pereira RA, Caetano AL, Cuppari L, Kamimura MA. Adductor pollicis muscle thickness as a predictor of handgrip strength in hemodialysis patients. J Bras Nefrol 2013;35:177e84. [23] Silva JBd, Maurício SF, Bering T, Correia MITD. The relationship between nutritional status and the Glasgow Prognostic Score in patients with cancer of the esophagus and stomach. Nutr Cancer 2013;65:25e33. [24] Gonzalez MC, Pureza Duarte RR, Orlandi SP, Bielemann RM, Barbosa-Silva TG. Adductor pollicis muscle: a study about its use as a nutritional parameter in surgical patients. Clin Nutr 2014. http://dx.doi.org/10.1016/j.clnu.2014.11.006. [25] Whiting PF, Rutjes AW, Westwood ME, Mallett S, Deeks JJ, Reitsma JB, et al. QUADAS-2: a revised tool for the quality assessment of diagnostic accuracy studies. Ann Intern Med 2011;155:529e36. [26] Bragagnolo R, Caporossi FS, Dock-Nascimento DB, de Aguilar-Nascimento JE. Adductor pollicis muscle thickness: a fast and reliable method for nutritional assessment in surgical patients. Rev Col Bras Cir 2009;36:371e6. [27] Nunes FF, Fernandes SA, Bertolini CM, Rabito EI, Gottschall CBA. Nutritional evaluation of cirrhotic patients: comparison between several methods. Sci Medica 2012;22:12e7. [28] Lameu EB, Gerude MF, Correa RC, Lima KA. Adductor pollicis muscle: a new anthropometric parameter. Rev Hosp Clin Fac Med Sao Paulo 2004;59: 57e62. [29] Guerra RS, Fonseca I, Pichel F, Restivo MT, Amaral TF. Handgrip strength and associated factors in hospitalized patients. JPEN J Parenter Enter Nutr 2015;39:322e30. [30] Bragagnolo R, Caporossi FS, Dock-Nascimento DB, Eduardo de AguilarNascimento J. Handgrip strength and adductor pollicis muscle thickness as predictors of postoperative complications after major operations of the gastrointestinal tract. e-SPEN 2011;6:e21e6. [31] Gavin H. Understanding research methods and statistics in psychology. Sage; 2008. [32] Wulan S, Westerterp K, Plasqui G. Ethnic differences in body composition and the associated metabolic profile: a comparative study between Asians and Caucasians. Maturitas 2010;65:315e9. [33] Ghorabi S, Ardehali H, Amiri Z, Shariatpanahi ZV. Association of the adductor pollicis muscle thickness with clinical outcomes in intensive care unit patients. Nutr Clin Pract 2016. 0884533615621547. [34] Leong Shu-Fen C, Ong V, Kowitlawakul Y, Ling TA, Mukhopadhyay A, Henry J. The adductor pollicis muscle: a poor predictor of clinical outcome in ICU patients. Asia Pac J Clin Nutr 2015;24:605e9.
Please cite this article in press as: Lew CCH, et al., Validity of the adductor pollicis muscle as a component of nutritional screening in the hospital setting: A systematic review, Clinical Nutrition ESPEN (2016), http://dx.doi.org/10.1016/j.clnesp.2016.08.005
Contents lists available at ScienceDirect
Clinical Nutrition ESPEN journal homepage: http://www.clinicalnutritionespen.com
Review
Validity of the adductor pollicis muscle as a component of nutritional screening in the hospital setting: A systematic review Charles Chin Han Lew a, b, *, Fangyi Ong c, Michelle Miller a a
Discipline of Nutrition and Dietetics, Flinders University, Australia Dietetics and Nutrition Department, Ng Teng Fong General Hospital Singapore c Dietetics Department, National University Hospital, Singapore b
a r t i c l e i n f o
s u m m a r y
Article history: Received 27 August 2016 Accepted 31 August 2016
Background & aims: The thickness of the adductor pollicis muscle (TAPM) is the only muscle that can be directly measured with a caliper. Recent studies demonstrate that the TAPM could be used as part of nutritional screening, but this has not been systematically reviewed. We aimed to review the validity and reliability of TAPM in identifying malnutrition risk. Methods: PubMed, CINAHL and Scopus were systematically searched. Eligible studies were crosssectional, case-control or cohort studies which recruited adult patients in the hospital, measured the TAPM along with a valid nutritional assessment tool and evaluated the validity of TAPM with univariate, multivariate, discriminative and/or agreement statistics. The Quality Assessment of Diagnostic Accuracy Studies-II was used to evaluate the risk-of-bias. Results: Nine out of 39 studies identified were eligible, and all had some risk-of-bias. Subjective Global Assessment (SGA) was used as the criterion standard in all nine studies. The TAPM amongst malnourished patients were significantly different from their counterparts but this may be over-estimated since all measurements of the SGA and TAPM were non-blinded. Concordance between the TAPM and SGA ranged from poor to good (kappa ranged from 0.04 to 0.25; specificity ranged from 97.8% to 100%), and this may be due to the varied cut-off values used amongst the studies. There were also disparities in the measurement instruments and methods. Reliability was not reported. Conclusion: More studies are needed to establish the reliability of TAPM measurement and cut-off values before the TAPM can be used as a component of nutritional screening. © 2016 European Society for Clinical Nutrition and Metabolism. Published by Elsevier Ltd. All rights reserved.
Keywords: Adductor pollicis Malnutrition Nutritional screening
1. Introduction About 25e40% of patients are malnourished at hospital admission [1e3]. Malnutrition increases short and long-term mortality, length of hospitalization and healthcare costs [4e6]. To reduce the negative impact of malnutrition, nutritional interventions should be timely [7], and this can only be achieved if patients at risk of malnutrition are promptly identified. Therefore, the Joint Commission International stipulates that nutritional screening should be carried out within 24 h of hospital admission [8].
Abbreviations: APM, Adductor pollicis muscle; ROC, Area under the receiver operative characteristics curve; SGA, Subjective Global Assessment; TAPM, Thickness of the adductor pollicis muscle. * Corresponding author. 1 Jurong East Street 21, 609606, Singapore. E-mail address: [email protected] (C.C.H. Lew).
There are more than 30 nutritional screening tools reported in the literature, each using different combinations of variables (e.g. unintentional weight loss and insufficient calorie intake) to determine malnutrition risk [9]. Loss of muscle mass is one of the hallmarks of malnutrition [10,11]. There are several ways to measure loss of muscle mass in the clinical setting, namely bio-impedance analysis, physical examination and anthropometry. The accuracy of bio-impedance analysis is variable especially in the clinical setting as several conditions that are highly prevalent in the hospital (e.g. edema, hypoalbuminemia and some medications) can confound the results [12]. Concerning physical examination and anthropometry where accuracy are both skill dependent, anthropometry may have an advantage over physical examination as it is relatively more objective in measuring muscle loss. In the clinical setting, anthropometry for the measurement of muscle loss is commonly quantified by the mid-arm muscle area
http://dx.doi.org/10.1016/j.clnesp.2016.08.005 2405-4577/© 2016 European Society for Clinical Nutrition and Metabolism. Published by Elsevier Ltd. All rights reserved.
Please cite this article in press as: Lew CCH, et al., Validity of the adductor pollicis muscle as a component of nutritional screening in the hospital setting: A systematic review, Clinical Nutrition ESPEN (2016), http://dx.doi.org/10.1016/j.clnesp.2016.08.005
2
C.C.H. Lew et al. / Clinical Nutrition ESPEN xxx (2016) 1e7
[13]. However, it is an indirect measurement of muscle mass as both the areas of subcutaneous fat and bone in the mid-arm region are adjusted in the calculation of muscle mass. This has been shown to over-estimate muscle mass by up to 25% [14]. In contrast, the thickness of the adductor pollicis muscle (APM) is the only muscle that can be directly measured with a caliper [15]. The APM is located between the thumb and index finger. It is easily accessible and can be directly measured because it is anatomically well-defined, flat in shape and has a minimal amount of subcutaneous fat surrounding it [15]. Recent studies demonstrated that the TAPM is a good surrogate measurement for total muscle mass [13,16]. Therefore, TAPM could potentially be used in combination with other conventional nutritional parameters as part of nutritional screening if it can independently discriminate well- and malnourished patients at the univariate level [17]. We aim to systematically review the validity of the TAPM against valid nutritional assessment tools in identifying malnutrition risk amongst adults in the hospital setting. In addition, we aim to report on the reliability of TAPM measurement. 2. Methods 2.1. Protocol and registration The reporting of this systematic review is in accordance to the PRISMA-statement and the protocol is registered on PROSPERO (Registration number: CRD42015023261). 2.2. Eligibility criteria Since the measurement of TAPM is relatively new, a broad eligibility criteria was developed to maximize the possibility of gathering all relevant articles. All cross-sectional, case-control and cohort studies were eligible for inclusion. In addition, all studies that included the APM and nutritional status as measurement variables were included. No restriction was imposed on publication date and language.
non-pregnant patients in the hospital were included. Articles that used electrical stimulation of the ulnar nerve to evaluate the strength of contraction and maximum rate of slackness of the APM were excluded. This method is not feasible in clinical practice as it requires specialized equipment, and the procedure is painful for the patient. Instead, articles that measured the TAPM along with a valid nutritional assessment tool (i.e. Subjective Global Assessment [SGA] [18] and/or Mini-nutritional Assessment) [19] were included in the review. In addition, articles included used appropriate statistical methods to evaluate the validity of TAPM in differentiating well- and malnourished patients. According to Jones [17], variables that could potentially be included as part of nutritional screening should minimally be able to differentiate well- and malnourished individuals at a univariate level. Therefore, only studies that compared the mean or median of the TAPM between well- and malnourished patients, reported discriminative statistics and/or agreement statistics were included. The PRISMA flow diagram was used to summarize the article selection processes (Fig. 1). 2.5. Data extraction Data were extracted and grouped into four categories as follows: 1) study design, country, patient characteristics (i.e. age and primary diagnosis); 2) timing and assessor the TAPM and nutritional assessment; 3) the prevalence of malnutrition determined by nutritional assessment; and 4) descriptive statistics (i.e. mean, standard deviation, median and inter-quartile range), univariate statistics (i.e. one-way analysis of variance and correlation), discriminative statistics (i.e. sensitivity, specificity and area under the receiver operative characteristics curve [ROC]), multivariate analysis, agreement statistics (i.e. kappa), and reliability statistics (i.e. intra- and inter-assessor error). We contacted five authors [20e24] for further information, and all of them responded
2.3. Information sources and search methods A search strategy was developed with reference to the eligibility criteria and three electronic databases, namely PubMed, CINAHL and Scopus were systematically searched on 2nd May 2015. To maximize the possibility of gathering all relevant studies, both free text terms and broad search terms (i.e. MeSH in PubMed, and CINAHL Headings in CINAHL) were used. Synonyms of ‘malnutrition’ and ‘nutritional status’ were combined with synonyms of ‘screening’ and ‘assessment’ to identify all tools used to measure nutritional status [4]. Results of the latter were combined with all articles that included measurements of the APM. This search strategy was appropriately adapted in all three electronic databases to ensure consistency (Supplementary material). To further ensure all relevant articles were identified, we hand searched the reference lists of the articles that were included in this systematic review. 2.4. Study selection There were a manageable number of articles upon the removal of duplicates. Therefore, instead of screening the titles and abstracts, two reviewers (CHCL and FO) assessed the relevance of the studies independently by evaluating the full-text version of the articles. In all cases of disagreement, consensus was sought through discussion. This review focused on the TAPM measured in a hospital setting. Therefore, only studies that recruited adult (>18 years old) and
Fig. 1. Summary of each stage of the search methods and the quantity of articles retrieved, excluded and included in this systematic review.
Please cite this article in press as: Lew CCH, et al., Validity of the adductor pollicis muscle as a component of nutritional screening in the hospital setting: A systematic review, Clinical Nutrition ESPEN (2016), http://dx.doi.org/10.1016/j.clnesp.2016.08.005
Authors
n
Age (y)
Measurement details
Malnutrition prevalence
Univariate/Multivariate statistics
Discriminative statistics
Bragagnolo et al. [26]
87
53.8 ± 15.4 TAPM and SGA measured within 48 h of admission, by NSP
SGA-A: 11.5% SGA-B: 29.9% SGA-C: 58.6% DTAPM: 62.8% NDTAPM: 65.9%
Mean DTAPM and NDTAPM of SGA-A was significantly higher than SGA-B, p < 0.001 Mean DTAPM and NDTAPM of SGA-B was significantly higher than SGA-C, p < 0.05
SensitivityDTAPM: NA 72.4% NDTAPM: 77.3% SpecificityDTAPM: 100.0% NDTAPM: 100.0% ROCDTAPM: 0.93 (95% CI: 0.86 to 0.99) NDTAPM: 0.92 (95% CI: 0.85 to 0.98)
Bragagnolo et al. [30]
90
53 ± 16
TAPM and SGA measured within 24 h of admission, by NSP
SGA-A: 14.4% SGA-B: 32.2% SGA-C: 53.3% TAPM: NA
Multi-linear regressionCompared to SGA-B and SGA-C, DTAPM and NDTAPM were 4.7 mm thicker in SGA-A, p < 0.001
NA
NA
Caporossi et al. [20]
246 62*
NA
SGA-A: 21.9% SGA-B: 53.7% SGA-C: 24.4% TAPM: NA
TAPM of the right hand (mm)SGA-A: 17.2 ± 5.4 NA SGA-B: 16.8 ± 5.7 SGA-C: 12.9 ± 5.3y TAPM of the left hand (mm)SGA-A: 15.8 ± 4.6 SGA-B: 15.9 ± 5.9 SGA-C: 12.3 ± 5.5y NA NA
NA
Nunes et al. [27]
119 56.3 ± 12.0 TAPM and SGA measured at outpatient visit, by NSP
SGA-A: NA SGA-B and SGA-C: 6.7% TAPM: 14.3%
Agreement statistics
Kappa: 0.25
SGA-A: 30.0% SGA-B: 24.3% SGA-C: 45.7% TAPM: 13.3% SGA-A: 52.0% SGA-B: 46.6% SGA-C: 1.4% TAPM: NA
TAPM (mm)SGA-A: 24.3 ± 4.2‡ SGA-B: 22.5 ± 6.5 SGA-C: 20.0 ± 5.1
NA
Kappa: 0.04, p < 0.05
No difference in the number of patients with TAPM lower or higher than 10 mm in each SGA category, p ¼ 0.55
NA
NA
TAPM and SGA measured before radio-/chemotherapy, by Dietitians (number NA)
SGA-A: 13.9% SGA-B: 13.9% SGA-C: 72.2% TAPM: 44.2%
TAPM (mm)*SGA-A: 21.0 SGA-B: 21.5 SGA-C: 16.5x
NA
Kappa: < 0.20
TAPM and SGA measured within 72 h of admission by two trained Nutritionists
PG-SGA-A: 52.1% PG-SGA-B: 24.1% PG-SGA-C: 23.8% TAPM: NA
Correlation between TAPM and SGA r ¼ 0.194, p < 0.001
NA
NA
SGA-A: 68.4% SGA-B: 24.1% SGA-C: 7.5% TAPM: NA
DTAPM (Male) (mm)*SGA-A: 26.0 (25.0, 28.0)‡ SensitivityDTAPM: 34.9% SGA-B: 19.5 (16.5, 22.0) NDTAPM: 37.7% SGA-C: 18.0 (15.0, 20.0) DTAPM (Female) (mm)*SGA-A: 23.0 (21.0, 25.0)‡
Mauricio et al. [21] 70
60.4 ± 14.3 TAPM and SGA measured before radio-/ chemotherapy, by Dietitians (number NA)
Pereira et al. [22]
73
52.3 ± 17.0 TAPM and SGA measured after haemodialysis, by two trained Nutritionists
Silva et al. [23]
43
NA
Guerra et al. [29]
688 58* (21)
Gonzalez et al. [24]
361 49.6 ± 17.8 TAPM and SGA measured by trained personnel (number and timing NA)
C.C.H. Lew et al. / Clinical Nutrition ESPEN xxx (2016) 1e7
Please cite this article in press as: Lew CCH, et al., Validity of the adductor pollicis muscle as a component of nutritional screening in the hospital setting: A systematic review, Clinical Nutrition ESPEN (2016), http://dx.doi.org/10.1016/j.clnesp.2016.08.005
Table 1 Summary of the validity of the TAPM in differentiating nutritional status.
NA
3
(continued on next page)
SpecificityDTAPM: SGA-B: 18.0 (16.0, 20.0) 98.7% SGA-C: 17.0 (15.0, 18.0) ‡ NDTAPM (Male) (mm)*SGA-A: 25.0 (24.0, 28.0) NDTAPM: 97.8% SGA-B: 18.5 (15.5, 20.0) SGA-C: 16.0 (15.0, 20.0) NDTAPM (Female) (mm)*SGA-A: 22.0 (20.0, 25.0)‡ SGA-B: 18.0 (16.0, 20.0) SGA-C: 16.0 (12.0, 18.0) Correlation with SGADTAPM: r ¼ 0.61, p < 0.05 NDTAPM: r ¼ 0.60, p < 0.05 Multivariate linear regressionDTAPM and NDTAPM of SGA-B and -C were 4.59 mm and 4.56 mm, and 6.51 mm and 6.14 mm thinner than SGA-A respectively, p < 0.001
Discriminative statistics Univariate/Multivariate statistics Malnutrition prevalence Measurement details Age (y)
Gonzalez et al. [24]
n Authors
Table 1 (continued )
Values are means ± standard deviation unless stated, * Median and interquartile range in parenthesis (when available), y p < 0.001 SGA-C compared to SGA-A and SGA-B, ‡ p < 0.05 SGA-A compared to SGA-B and SGA-C, x p < 0.05 SGA-C compared to SGA-B, DTAPM: Dominant thickness of the adductor pollicis muscle, NA: Not available, NDTAPM: Non-dominant thickness of the adductor pollicis muscle, NSP: Non-specific personnel, PG-SGA: PatientGenerated Subjective Global Assessment, ROC: area under the receiver operative characteristics curve, SGA: Subjective Global Assessment, TAPM: Thickness of the adductor pollicis muscle.
C.C.H. Lew et al. / Clinical Nutrition ESPEN xxx (2016) 1e7
Agreement statistics
4
[20e24]. One reviewer (CHCL) extracted the required data and the second reviewer (FO) checked the extracted data. Disagreements were resolved by discussion. If no consensus was reached, a third reviewer's (MM) opinion was sought. 2.6. Study appraisal and synthesis An evidence-based quality assessment tool, i.e. Quality Assessment of Diagnostic Accuracy Studies e II (QUADAS-II) [25] with high construct validity, interrater reliability, and internal consistency was used to evaluate the diagnostic validity of the studies included in this review. Essentially, it is a checklist of seven items that assess the risk of bias in four main domains (i.e. patient selection, index test, reference test and patient flow). The tool also assesses the applicability of the diagnostic studies with reference to the review question. Each domain was scored as “low risk”, “high risk” or “unclear risk” according to the detailed scoring criteria outlined by Whiting et al. [25]. The critical appraisal was independently performed by two reviewers (CHCL and FO), and any disagreement between the two reviewers were resolved by discussion or a third reviewer's (MM) opinion was sought if no consensus was reached. A meta-analysis was not performed as measurements of the TAPM were reported in means and medians. In addition, discriminative statistics were not pooled as the TAPM cut-off points for malnutrition used were different, i.e. Bragagnolo et al. [26] used cut-off values derived in their study whereas Gonzalez et al. [24] used cut-off values derived previously from a group of wellnourished subjects [15]. Similarly, agreement statistics were not pooled as the TAPM cut-off points for malnutrition were different, i.e. Nunes et al. [27] used cut-off values derived by Lameu et al. [28] whereas Mauricio et al. [21] and Silva et al. [23] used cut-off values derived by Gonzalez et al. [15]. 3. Results The literature search generated 73 articles. Upon the removal of duplicates and studies that did not meet the eligibility criteria, a total of nine studies were included in the review. The flow chart of the selection process is provided in Fig. 1. Of the nine studies, eight were carried out in Brazil, and one was carried out in Portugal [29]. All studies were written in English except for two written in Portuguese [26,27]. They were cross-sectional studies except for the studies conducted by Caporossi et al. [20] and Bragagnolo et al. [30] which were prospective cohort studies. Most of the studies recruited medical patients, i.e. those with cancer [21,23], critical illness [20], liver [27] and renal disease [22], and a group of heterogeneous inpatients [29]. The other three studies recruited surgical patients [24,26,30]. The mean or median age of the subjects were between 50 and 62 years old. Malnutrition was diagnosed by the SGA [20e24,26,27,30] or the Patient-Generated Subjective Global Assessment [29], and the prevalence of malnutrition in the inpatient setting ranged from 31.6% [24] to 88.5% [26], and it was 6.7% [27] in the outpatient setting (Table 1). All studies had some risk of bias (Table 2). In the domain of Index Test, both the SGA and the TAPM were carried out in a non-blinded fashion. In the domain of Reference Standard, two studies had partial verification bias because the SGA was not performed on all the patients and the rationale was not reported [23,24]. There were applicability concerns in three studies. Silva et al. [23] and Nunes et al. [27] did not provide details on the methods used to perform the SGA and/or TAPM measurements, and Pereira et al. [22] and Silva et al. [23] did not report the caliper used. The procedures used to measure TAPM were similar in seven studies [21e24,26,29,30]. Measurements were taken while subjects
Please cite this article in press as: Lew CCH, et al., Validity of the adductor pollicis muscle as a component of nutritional screening in the hospital setting: A systematic review, Clinical Nutrition ESPEN (2016), http://dx.doi.org/10.1016/j.clnesp.2016.08.005
C.C.H. Lew et al. / Clinical Nutrition ESPEN xxx (2016) 1e7
5
Table 2 Assessment of methodological risk of bias based on QUADAS-II. Study
Bragagnolo et al. [26] Bragagnolo et al. [30] Caporossi et al. [20] Nunes et al. [27] Mauricio et al. [21] Pereira et al. [22] Silva et al. [23] Gonzalez et al. [24] Guerra et al. [29]
Risk of bias
Applicability concerns
Patient selection
Index test
Reference standard
Flow and timing
Patient selection
Index test
Reference standard
þ þ þ þ þ þ þ þ þ
e e e e e e e e e
þ þ þ þ þ þ e e þ
þ þ þ þ þ þ þ þ þ
þ þ þ þ þ þ þ þ þ
þ þ þ e þ e e þ þ
þ þ þ þ þ þ e þ þ
þ Low risk; e High risk; ? Unclear risk.
were seated, elbows bent at a 90-degree angle and hands resting on the knees. Caporossi et al. [20] however had to modify the procedure according to the limitation of critically ill patients. Measurements were taken while subjects were in a supine position, elbows bent at a 90-degree angle and hands lying on the upper abdomen. Although the procedures were mostly similar, the recorded measurements were derived differently. Most studies used the mean of three consecutive measurements [20,22,24,26,29,30] but other studies used the highest value of three measurements [21,23]. Nunes et al. [27] did not report how the recorded measurements were derived. Calipers used to measure the TAPM were also different amongst the studies. The Cescorf caliper was used in four studies [20,26,27,30], Lange caliper in two studies [21,24], and Harpenden caliper in one study [29]. Pereira et al. [22] and Silva et al. [23] did not report the caliper used. The intra- and inter-assessor reliability of TAPM measurement were not reported in all the studies. Furthermore, the clinicians who carried out the TAPM measurements were not identified in more than half of the studies [20,24,26,27,30]. When specified, measurements were performed by either two nutritionists [22,29] or undefined number of dietitians [21,23]. The difference in the TAPM between well- and malnourished patients was determined mostly by univariate analyses. One-way analysis of variance was used in several studies, and it showed mixed results (Table 1) [20,21,23,24,26]. Only Bragagnolo et al. [26] demonstrated that the TAPM increased significantly with better nutritional status (i.e. SGA-A > SGA-B > SGA-C). This increase in the TAPM with each category of the SGA was not observed in other studies [20,21,23,24]. Gonzalez et al. [24] demonstrated that the APM of well-nourished patients (SGA-A)
were significantly thicker than malnourished patients (SGA-B and SGA-C combined). On the other hand, Mauricio et al. [21], Caporossi et al. [20] and Silva et al. [23] demonstrated that the APM of both well- and mildly-moderately malnourished patients (SGA-A and SGA-B combined) were significantly thicker than severely malnourished patients (SGA-C). Other univariate analyses quantified the correlation and agreement (i.e. kappa) between the TAPM and nutritional status diagnosed by the SGA [21e24,27,29]. Correlation analyses showed mixed results as Gonzalez et al. [24] found moderate negative correlation whereas Pereira et al. [22] and Guerra et al. [29] found either no correlation or weak negative correlation between the SGA and the TAPM. Kappa analyses were performed in three studies, and the agreement between the TAPM and SGA were consistently poor (kappa ranged from 0.04 to 0.25) [21,23,27]. Multilinear regression analysis was used in two studies to determine the difference in the TAPM between well- and malnourished patients. Both studies showed that the TAPM of wellnourished patients were thicker than their counterparts [24,30]. After adjusting for sex, age and/or weight, the APM of the dominant and non-dominant hands in well-nourished patients were at least 4.6 mm thicker than the malnourished patients [24,30]. Only two studies used discriminative statistics [24,26] (i.e. ROC, sensitivity and specificity analyses) and they showed that the TAPM could differentiate well-nourished patients from malnourished patients (combination of mildly-moderately and severely malnourished). Specificity of the TAPM in discriminating wellnourished patients from malnourished patients were high (Table 1) [24,26]. However, the cut-off values used in both studies were different.
Table 3 Thickness of the adductor pollicis muscle cut-off values used to define malnutrition risk. Authors
Method used to determine the TAPM cut-off value to define malnutrition risk
DTAPM (mm)
NDTAPM (mm)
Male
Female
All
Male
Female
All
Bragagnolo et al. [26] Bragagnolo et al. [30] Caporossi et al. [20] Nunes et al. [27]
Receiver Operative Characteristics NA 10th percentile of the subjects' TAPM Referenced Lameu et al. [28] where 421 healthy subjects aged between 18 and 87 years old were recruited. Mean cut-off values were derived from values that were more than one standard deviation away from the mean, and the median.
NA NA NA 9.5y 11.0‡
NA NA NA 8.0y 9.0‡
13.4 NA 9.5 NA NA
NA NA NA 9.5y 11.0‡
NA NA NA 8.0y 9.0‡
13.1 NA 8.3 NA NA
Mauricio et al. [21] Silva et al. [23] Gonzalez et al. [24] Pereira et al. [22] Guerra et al. [29]
All three studies referenced Gonzalez et al. [15] where 300 healthy subjects aged between 18 to 90 years old were recruited. Cut-off values of were derived from the 5th percentile of the subjects' TAPM Median of the subjects' TAPM measurements NA
20.0x 23.0¶ 18.0yy NA NA
16.0x 17.0¶ 14.0yy NA NA
NA NA NA 10‡ NA
19.0x 21.0¶ 16.0yy NA NA
15.0x 16.0¶ 14.0yy NA NA
NA NA NA 10‡ NA
y
mean, ‡ median, x 18e29 years old, ¶ 30e59 years old, yy 60 years old, DTAPM: Dominant thickness of the adductor pollicis muscle, NA: Not available, NDTAPM: Nondominant thickness of the adductor pollicis muscle, TAPM: Thickness of the adductor pollicis muscle.
Please cite this article in press as: Lew CCH, et al., Validity of the adductor pollicis muscle as a component of nutritional screening in the hospital setting: A systematic review, Clinical Nutrition ESPEN (2016), http://dx.doi.org/10.1016/j.clnesp.2016.08.005
6
C.C.H. Lew et al. / Clinical Nutrition ESPEN xxx (2016) 1e7
There were vast variations of cut-off values used to define nutritional status. Nunes et al. [27] used cut-off values provided by Lameu et al. [28] whereas Mauricio et al. [21], Silva et al. [23] and Gonzalez et al. [24] used cut-off values provided by Gonzalez et al. [15]. Other studies self-defined their cut-off values using the ROC analysis with the SGA as the criterion standard [26], or the 10th [20] and 50th [22] percentile of the TAPM measurements of recruited subjects. Table 3 summarizes the cut-off values used in the studies included in the review. 4. Discussion To our knowledge, this is the first systematic review that evaluated the validity and reliability of the TAPM in identifying malnutrition risk. Our results show that the TAPM is not an appropriate component of nutritional screening until further studies evaluate the limitations highlighted in the review. Although univariate and multivariate analyses demonstrated that the TAPM amongst malnourished patients were significantly different from their counterparts, such results may be overestimated because there were some risk-of-bias, i.e. the SGA and TAPM were performed in a non-blinded fashion. Concordance between the TAPM and SGA were mixed. The ROC showed that the TAPM had excellent discrimination (area under the curve: 0.92) [26] and high specificity (ranged from 97.8% to 100%) [24,26] in differentiating well-nourished patients from the malnourished. However, kappa statistics (ranged from 0.04 to 0.25) [21,23,27] demonstrated poor concordance. The disparate findings were likely due to differences in the cut-off values used. Different cut-off values were used to define malnutrition risk (Table 3). Some studies self-derived the cut-off values [20,22,26] whereas other studies [21,23,24,27] used cut-off values derived from healthy populations [15,28]. Lameu et al. [28] and Gonzalez et al. [15] measured the TAPM of well-nourished and healthy individuals and established cut-off values that define malnutrition risk. These cut-off values, however may have limited generalizability as all the subjects were monoethnic, i.e. Brazilians. The applicability of using the TAPM as a component of nutritional screening may be limited by the lack of measurement standardization and reliability measurement. The methods and calipers used to measure the TAPM were variable amongst the studies. The consequence of such differences is demonstrated by the results of Lameu et al. [28] and Gonzalez et al. [15] where cut-off values used to define malnutrition risk were vastly different. Lameu et al. [28] used the Lange caliper and the mean of three measurements whereas Gonzalez et al. [15] used the Crecorf caliper and the highest value of three measurements. These differences could explain the discrepancies in the results of the studies included in the review. None of the studies reported the reliability of measuring the TAPM. Since a clinically useful tool requires both validity and reliability [31], the absence of intra- and inter-reliability data preclude the use of the TAPM in the hospital setting. 5. Future studies Given that body composition can be affected by ethnicity [32], future studies should measure the TAPM amongst healthy individuals from different ethnicities and use similar methods to establish cut-off values that are hand, age, sex and ethnic specific to define malnutrition risk. In addition, future studies should measure the intra- and inter-reliability of measuring the TAPM. This is especially important because such measurement is skilldependent. It is ideal for nutritional screening tools to be parsimonious. Future studies should also determine the significance of including
the TAPM with other conventional nutritional parameters as part of nutritional screening via a multivariate model. Interestingly, the TAPM was demonstrated to have mortality prognostic value in three studies [20,30,33], but such result was not consistently observed [34]. More studies should be carried out to evaluate the prognostic value of the TAPM in the hospital setting. 6. Conclusion This review presented a body of evidence highlighting the limitations of using the TAPM as a component of nutritional screening in the hospital setting. For the TAPM to be considered a valid component of nutritional screening, future studies should determine the reliability of measuring the TAPM as well as establish cut-off values that are hand, age, sex, and ethnic specific to define malnutrition risk. Sources of support Nil. Authors' contribution All authors designed the research, analyzed the studies included in the review, and read and approved the final manuscript; C. C. H. Lew and F. Ong conducted the literature search and prepared the draft manuscript; C. C. H. Lew had primary responsibility for final content. Conflict of interest statement and funding sources None of the authors reported a conflict of interest related to the study. This review did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. Acknowledgments We are grateful to Claudia Canavarro for translating the articles written in Portuguese into English. Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.clnesp.2016.08.005. References [1] Kelly IE, Tessier S, Cahill A, Morris SE, Crumley A, McLaughlin D, et al. Still hungry in hospital: identifying malnutrition in acute hospital admissions. QJM 2000;93:93e8. [2] Kruizenga HM, Van Tulder MW, Seidell JC, Thijs A, Ader HJ, Van Bokhorst-de van der Schueren MA. Effectiveness and cost-effectiveness of early screening and treatment of malnourished patients. Am J Clin Nutr 2005;82: 1082e9. [3] McWhirter JP, Pennington CR. Incidence and recognition of malnutrition in hospital. BMJ 1994;308:945e8. [4] Lew CCH, Yandell R, Fraser RJ, Chua AP, Chong MF, Miller M. Association between malnutrition and clinical outcomes in the intensive care unit: a systematic review. JPEN J Parenter Enter Nutr 2016. http://dx.doi.org/10.1177/ 0148607115625638. [5] Lim SL, Ong KC, Chan YH, Loke WC, Ferguson M, Daniels L. Malnutrition and its impact on cost of hospitalization, length of stay, readmission and 3-year mortality. Clin Nutr 2012;31:345e50. [6] Correia MI, Waitzberg DL. The impact of malnutrition on morbidity, mortality, length of hospital stay and costs evaluated through a multivariate model analysis. Clin Nutr 2003;22:235e9. [7] Rasheed S, Woods RT. Malnutrition and quality of life in older people: a systematic review and meta-analysis. Ageing Res Rev 2013;12:561e6.
Please cite this article in press as: Lew CCH, et al., Validity of the adductor pollicis muscle as a component of nutritional screening in the hospital setting: A systematic review, Clinical Nutrition ESPEN (2016), http://dx.doi.org/10.1016/j.clnesp.2016.08.005
C.C.H. Lew et al. / Clinical Nutrition ESPEN xxx (2016) 1e7 [8] Joint Commission International, Joint Commission on Accreditation of Healthcare Organizations. Joint commission international accreditation standards for hospitals. 5th ed. Joint Commission Resources; 2015. [9] van Bokhorst-de van der Schueren MA, Guaitoli PR, Jansma EP, de Vet HC. Nutrition screening tools: does one size fit all? A systematic review of screening tools for the hospital setting. Clin Nutr 2014;33:39e58. [10] White JV, Guenter P, Jensen G, Malone A, Schofield M. Consensus statement of the Academy of Nutrition and Dietetics/American Society for Parenteral and Enteral Nutrition: characteristics recommended for the identification and documentation of adult malnutrition (undernutrition). J Acad Nutr Diet 2012;112:730e8. [11] Cederholm T, Bosaeus I, Barazzoni R, Bauer J, Van Gossum A, Klek S, et al. Diagnostic criteria for malnutrition e an ESPEN consensus statement. Clin Nutr 2015;34:335e40. [12] Kushner RF, Gudivaka R, Schoeller DA. Clinical characteristics influencing bioelectrical impedance analysis measurements. Am J Clin Nutr 1996;64:423Se7S. [13] Lameu EB, Gerude MF, Campos AC, Luiz RR. The thickness of the adductor pollicis muscle reflects the muscle compartment and may be used as a new anthropometric parameter for nutritional assessment. Curr Opin Clin Nutr Metab Care 2004;7:293e301. [14] Heymsfield SB, McManus C, Smith J, Stevens V, Nixon DW. Anthropometric measurement of muscle mass: revised equations for calculating bone-free arm muscle area. Am J Clin Nutr 1982;36:680e90. [15] Gonzalez MC, Duarte RR, Budziareck MB. Adductor pollicis muscle: reference values of its thickness in a healthy population. Clin Nutr 2010;29:268e71. [16] de Oliveira CM, Kubrusly M, Mota RS, Choukroun G, Neto JB, da Silva CA. Adductor pollicis muscle thickness: a promising anthropometric parameter for patients with chronic renal failure. J Ren Nutr 2012;22:307e16. [17] Jones JM. Development of a nutritional screening or assessment tool using a multivariate technique. Nutrition 2004;20:298e306. [18] Detsky AS, McLaughlin JR, Baker JP, Johnston N, Whittaker S, Mendelson RA, et al. What is subjective global assessment of nutritional status? JPEN J Parenter Enter Nutr 1987;11:8e13. [19] Guigoz Y, Vellas B, Garry P, Vellas B, Albarede J. Mini nutritional assessment: a practical assessment tool for grading the nutritional state of elderly patients. The mini nutritional assessment: MNA Nutrition in the elderly. 1997. p. 15e60. [20] Caporossi FS, Caporossi C, Borges Dock-Nascimento D, de AguilarNascimento JE. Measurement of the thickness of the adductor pollicis muscle as a predictor of outcome in critically ill patients. Nutr Hosp 2012;27:490e5. [21] Mauricio SF, da Silva JB, Bering T, Correia MI. Relationship between nutritional status and the Glasgow Prognostic Score in patients with colorectal cancer. Nutrition 2013;29:625e9.
7
[22] Pereira RA, Caetano AL, Cuppari L, Kamimura MA. Adductor pollicis muscle thickness as a predictor of handgrip strength in hemodialysis patients. J Bras Nefrol 2013;35:177e84. [23] Silva JBd, Maurício SF, Bering T, Correia MITD. The relationship between nutritional status and the Glasgow Prognostic Score in patients with cancer of the esophagus and stomach. Nutr Cancer 2013;65:25e33. [24] Gonzalez MC, Pureza Duarte RR, Orlandi SP, Bielemann RM, Barbosa-Silva TG. Adductor pollicis muscle: a study about its use as a nutritional parameter in surgical patients. Clin Nutr 2014. http://dx.doi.org/10.1016/j.clnu.2014.11.006. [25] Whiting PF, Rutjes AW, Westwood ME, Mallett S, Deeks JJ, Reitsma JB, et al. QUADAS-2: a revised tool for the quality assessment of diagnostic accuracy studies. Ann Intern Med 2011;155:529e36. [26] Bragagnolo R, Caporossi FS, Dock-Nascimento DB, de Aguilar-Nascimento JE. Adductor pollicis muscle thickness: a fast and reliable method for nutritional assessment in surgical patients. Rev Col Bras Cir 2009;36:371e6. [27] Nunes FF, Fernandes SA, Bertolini CM, Rabito EI, Gottschall CBA. Nutritional evaluation of cirrhotic patients: comparison between several methods. Sci Medica 2012;22:12e7. [28] Lameu EB, Gerude MF, Correa RC, Lima KA. Adductor pollicis muscle: a new anthropometric parameter. Rev Hosp Clin Fac Med Sao Paulo 2004;59: 57e62. [29] Guerra RS, Fonseca I, Pichel F, Restivo MT, Amaral TF. Handgrip strength and associated factors in hospitalized patients. JPEN J Parenter Enter Nutr 2015;39:322e30. [30] Bragagnolo R, Caporossi FS, Dock-Nascimento DB, Eduardo de AguilarNascimento J. Handgrip strength and adductor pollicis muscle thickness as predictors of postoperative complications after major operations of the gastrointestinal tract. e-SPEN 2011;6:e21e6. [31] Gavin H. Understanding research methods and statistics in psychology. Sage; 2008. [32] Wulan S, Westerterp K, Plasqui G. Ethnic differences in body composition and the associated metabolic profile: a comparative study between Asians and Caucasians. Maturitas 2010;65:315e9. [33] Ghorabi S, Ardehali H, Amiri Z, Shariatpanahi ZV. Association of the adductor pollicis muscle thickness with clinical outcomes in intensive care unit patients. Nutr Clin Pract 2016. 0884533615621547. [34] Leong Shu-Fen C, Ong V, Kowitlawakul Y, Ling TA, Mukhopadhyay A, Henry J. The adductor pollicis muscle: a poor predictor of clinical outcome in ICU patients. Asia Pac J Clin Nutr 2015;24:605e9.
Please cite this article in press as: Lew CCH, et al., Validity of the adductor pollicis muscle as a component of nutritional screening in the hospital setting: A systematic review, Clinical Nutrition ESPEN (2016), http://dx.doi.org/10.1016/j.clnesp.2016.08.005