AUTHOR COPY. Predictors of long-term recovery in complex activities of daily living before discharge from the stroke unit

Transcription

1 NeuroRehabilitation 33 (2013) DOI: /NRE IOS Press 217 Predictors of long-term recovery in complex activities of daily living before discharge from the stroke unit David Cioncoloni a,b, Giuseppe Martini c, Pietro Piu d, Sabrina Taddei a, Maurizio Acampa c, Francesca Guideri c, Rossana Tassi c and Riccardo Mazzocchio e, a U.O.P. Professioni della Riabilitazione, Azienda Ospedaliera Universitaria Senese, Università di Siena, Siena, Italy b Scuola di Dottorato in Scienze Cognitive, Azienda Ospedaliera Universitaria Senese, Università di Siena, Siena, Italy c Area Stroke Unit, Azienda Ospedaliera Universitaria Senese, Università di Siena, Siena, Italy d Scuola di Dottorato in Scienze Neurologiche Applicate, Azienda Ospedaliera Universitaria Senese, Università di Siena, Siena, Italy e S.C. Neurologia e Neurofisiologia Clinica, Azienda Ospedaliera Universitaria Senese, Università di Siena, Siena, Italy Abstract. BACKGROUND AND PURPOSE: There is a need for individuating those post-stroke patients who may benefit from an optimal and customised rehabilitation plan aiming at early reintegration in community life participation. This study investigated whether the gain of independence in complex Activities of Daily Living (ADL) may be predicted before the discharge from the stroke unit using simple bedside determinants. METHODS: In 104 first-ever stroke patients with no previous disability, ten determinants at 10 days after stroke were selected. Multivariable logistic regression analysis was applied to identify the prognostic determinants able to predict independence in complex ADL, as measured by modified Rankin Scale grade 2. RESULTS: The model shows that having a Barthel Index 9, a Motricity Index- Upper Limb 75, an age 70 and being a male resulted in 100% probability of achieving independence in complex ADL. If three of the four determinants were present, the probability was more than 90%. With the presence of two of the four determinants, the probability ranged from 87% to 28%. With the presence of only one determinant, the probability was 13%. CONCLUSIONS: Accurate prediction of independence in complex ADL can be made before the discharge from the stroke unit. The strength of the paretic upper limb, age, gender, and the ability of performing basic ADL are the significant variables. The probability of favorable prognosis depends on the presence and on the robustness of each single determinant. Keywords: Stroke, prognosis, recovery of function 1. Introduction Stroke is the leading cause of chronic disability in the United States and Europe [1, 2]. Approximately Address for correspondence: Dr. Riccardo Mazzocchio, Azienda Ospedaliera Universitaria Senese, Siena, Italy. Tel.: ; Fax: ; mazzocchio@unisi.it. 50 million of survivors worldwide reported significant deficits in physical, cognitive, and emotional functions, and 25% to 74% of these require some assistance or are fully dependent on caregivers for ADL [3]. In addition, it is estimated that only 50% of these is able to regain the ability to walk in the community [4] and that 60% of young stroke survivors, who are independent /13/$ IOS Press and the authors. All rights reserved

2 218 D. Cioncoloni et al. / Predictors of long-term recovery in complex activities of daily living in ADL, experience impairments in participation in the community life [5]. Functional prognosis is paramount in rehabilitation. Simple models with good clinical performance [6], using clinical factors measured at fixed moments [7], may contribute to an accurate and early prediction of functional outcome optimizing post-stroke management. Robust measures are required for assessment and measuring outcome [8] considering that functional capacity measured when the patient is in a standard and known setting might not reflect real-world patient s functional performance [4]. There is evidence that clinical measurements are time-dependent and not linearly related with stroke recovery [7]. To develop an accurate prognostic model, they should be administered at the time points when their prediction properties are optimal [6, 9, 10]. In addition, if assessment scales are not clinimetrically fit for purpose, the risk is to produce meaningless data on outcome evaluation [8]. Therefore, for prognostic models to be accurate [6] and generally applicable in clinical practice [11], they should be set on variables showing good prognostic properties and developed using the most frequent measurements adopted in stroke trials at their optimal moment of prediction capability. However, it is not clear what factors, clinical determinants and measurements are critical for predicting functional recovery in complex ADL. The aim of this study was to verify whether the probability to regain independence in complex ADL at six months post-stroke, including the mainly outdoor activities in the local community, could be prognosticated at the discharge from the stroke unit using a simple model based on the presence or absence and on the robustness of a choice of significant predictive variables. 2. Methods 2.1. Design This prospective cohort study included 123 patients affected by hemispheric ischemic stroke and spontaneous, non-traumatic intracerebral hemorrhage who were admitted to the Stroke Unit of Siena University Hospital during a period of 14 months. Measurements were taken within 48 hours after stroke and on day 10; final outcome was assessed at six months. Rehabilitation intervention started at two or three days post-stroke in Stroke Unit Area. After the discharge from the Stroke Unit, patients followed an ad hoc Rehabilitation Project till functional ability could be improved. The procedures were approved by the review body of the University Hospital and conformed to the Declaration of Helsinki Subjects Subjects were recruited according to the following inclusion criteria: (1) First-ever stroke in one hemisphere and no previous functional alteration (Barthel Index = 20; modified Rankin Scale = 0); (2) monoparesis or -plegia or hemi-paresis or -plegia within 48 hours after stroke; (3) aged 18 years or older; (4) able to follow the instructions. Critical subjects with no functional recovery prognosis at 48 hours were not eligible. Nineteen subjects were excluded because dead or not available at the follow up. In conclusion, 104 patients were eligible for the study Dependent variable The modified Rankin Scale (mrs) [8] was chosen as outcome variable. Unlike scales that focus solely on basic ADL, mrs is an ordinal, hierarchical and responsive grading of global disabilities which describes the ability to perform outdoor activities, complex ADL and need of assistance relative to the patient s previous lifestyle. Properly conducted, the mrs scores patients based on perception of functioning within the context of their own lives and, as such, has the potential to offer a meaningful assessment for evaluating global post-stroke functional recovery, whereby one grade is capable of describing patients status. The mrs quantifies disability using a grading from zero (no symptoms) to 5 (severe disability) with some adding a category of mrs 6 (death). The mrs manifests the maximum sensitivity in differentiating functional recovery at six months post-stroke [10]. At this time, there is no overlap in the frequency distribution between grade 2 and 3, making a cut-off between grades 0 2 and grades 3 5 possible. For this study, patients with mrs 2 were considered independent in complex ADL and those with mrs 3 not independent [12]. The cut-off grade 2 describes patients with slight disability, but who can live alone for a period of a week or more without concern, so that they are able to bathe, use toilet, prepare or get meals as well as to perform in independence the mainly outdoor activities in local community.

3 D. Cioncoloni et al. / Predictors of long-term recovery in complex activities of daily living Independent variables Eleven independent variables were selected at 10 days post-stroke. They were: (1) gender; (2) age; (3) hemisphere of stroke (left; right); (4) level of the global neurological status (National Institute of Health Stroke Scale [NIHSS] total score); presence or absence (yes; no) of (5) hemianopsia, (6) sensory loss, (7) extinction or inattention, (NIHSS, items 3, 8, 11); muscle strength of (8) upper and (9) lower limb, (total score of the Motricity Index [MI] upper and lower limb respectively); (10) presence or absence (yes; no) of sitting balance (Trunk Control Tests [TCT], item 3); (11) ability to perform basic ADL (Barthel Index [BI], total score). These variables have been shown to have good clinimetrical and predictive properties. Indeed, the NIHSS is used in most stroke trials for evaluating the patients neurological status within the acute phase post-stroke, and it is highly associated with the functional outcome on ADL three months [13]. The BI is one of the most frequently administered scales for evaluating basic ADL in the stroke unit setting. It shows good discriminative properties for predicting outcome in basic ADL at six month, if assessed after five days after stroke [14]. Concerning the post-stroke neurological deficits, the most prominent impairment is considered muscle weakness and its distribution in the lower and in the upper limb [15], so that its role in functional prognosis has been frequently tested [6, 16]. Gender [17] and age [18] have also been implicated in functional prognosis as well as the presence of extinction and inattention [19], hemianopsia [20] and sensory loss [20, 21] Data analysis The clinical factors were dichotomized on the basis of their presence or absence. For determining the optimal cut off of each measurement total score and age, the Receiver-Operating Characteristic (ROC) curve analysis was applied by using sensitivity/1-specificity. To determine the bivariable association between the candidate determinants at 10 days and the recovery of independence in complex ADL at six months, classified by dichotomized mrs, binary logistic regression was performed to calculate significances, odds ratios (OR) and the 95% confidence intervals (95% CI) for each individual variable, respectively. Only the variables which presented a liberal significant level of p 0.1 were included in subsequent multivariable logistic regression analysis. The eligible variables for the multivariable logistic regression were tested using the Spearman s rho (r s ) to prevent spurious relationship. If the correlation coefficient (r s )was 0.75, only one variable was selected for further use in the multivariable analysis based on clinical significance. A stepwise, forward LR multivariable logistic regression analysis was performed with the remaining determinants. These analyses were tested 2-tailed using critical p-values for entry and removal of 0.01 and 0.05, respectively. Using the constants and regression coefficients of the included determinants in the equation p = 1/(1 + (e [ ( 0 + 1X1 + 2X2 + 3X3 + + nxn)] )), the probability to achieve independence in complex ADL were calculated. Finally, 2-way contingency tables were used to calculate sensitivity, specificity, Positive Predicted Value (PPV) and Negative Predicted Value (NPV), including their 95% CIs, for the derived model. Statistics was performed using SPSS, version Results Table 1 shows the main demographic and clinical characteristics of the 104 patients included in the analysis. Table 2 reports the ORs and their 95% CIs of the binary logistic regression between the dependent and each independent variable respectively. Ten of the 11 variables presented a p-value < 0.1, so that they could be included in the multivariable logistic regression analysis. Indeed, no high associations were found between the selected variables (r s < 0.75). The significant determinants in the simultaneous multivariable regression analysis for the prediction model were: (1) MI-upper limb, (2) age, (3) gender and (4) BI. Table 3 shows the significant determinants, the cut off points and the probabilities to regain independence in complex ADL at six months, based on the equation model, using the constant and regression coefficient of the significant determinants. At 10 days after stroke, patients who were males and presented with a MI-upper limb score 75, age 70 years and, a BI score 9 had 100% probability to regain independence in complex ADL. If three of the four criteria were present, the probability was more than 90%. With the presence of two of the four positive predictive determinants, the probability ranged from 87% (MI-upper limb score 75, age 70 years) to 28% (BI 9, male gender). With the presence of only one determinant, the higher value was 13%. Finally, the sensitivity of the model was 0.89 (95% CI ), the specificity 0.87 (95% CI ), the

4 220 D. Cioncoloni et al. / Predictors of long-term recovery in complex activities of daily living Table 1 Patients characteristics at assessment time N 104 Gender, male/female 49/55 Age, years (total, male, female) a (12.34) (10.87) (13.54) Hemisphere of stroke, left/right 40/64 Stroke, ischemic/hemorrhagic 85/19 Type of stroke (Bamford) LACI 3 PACI 41 TACI 39 POCI 21 Days spent in hospital a (6.36) rtpa, yes/no (ischemic patients) 15/70 NIHSS, item 3: vision (hemianopsia), pres./abs. (10 days) 27/77 NIHSS, item 8: sensory (sensory loss), pres./abs. (10 days) 54/50 NIHSS, item 11: extinction and inattention, pres./abs. (10 days) 28/76 NIHSS total score (48 hours, 10 days) b 6.00 (6.75) 4.00 (6.00) MI-upper limb, total score (48 hours, 10 days) b (58.75) (58.00) MI-lower limb, total score (48 hours, 10 days) b (52.25) (54.00) TCT, balance in sitting position, pres./abs. (10 days) 74/30 TCT, total score (10 days) b (52.00) BI, total score (10 days, 6 months) b 9.00 (11.00) (10.00) mrs, total score (6 months) b 2.00 (3.00) Abbreviations: LACI, lacunar infarct; PACI, partial anterior circulation infarct; TACI, total anterior circulation infarct; POCI, posterior circulation infarct; rtpa = medical treatment with recombinant Tissue Plasminogen Activator. NIHSS = National Institute of Health Stroke Scale; MI, Motricity Index; TCT, Trunk Control Test; BI, Barthel Index; mrs = modified Rankin Scale. a Mean and Standard Deviation. b Median and Interquartile Ranges. Table 2 Variables associated with dichotomized mrs at 6 months in the binary logistic regression Determinants OR 95% CI p-value Gender (male/female) a Age (1 70; 0 > 70) a Hemisphere of stroke (left/right) a NIHSS 3, vision (hemianopsia) (0 > 0; 1 = 0) a NIHSS 8, sensory (sensory loss) (0 > 0; 1 = 0) a NIHSS 11, extinction - inattention (0 > 0; 1 = 0) a NIHSS, total score (1 4;0>4) b MI arm, total score (0 < 75; 1 75) b MI leg, total score (0 < 75; 1 75) b TCT, balance in sitting position (0 < 25; 1 = 25) a BI, total score (0 < 9; 1 9) b Abbreviations: OR = Odds Ratio; 95% CI = 95% Confidence Interval; NIHSS = National Institute of Health Stroke Scale; MI arm = Motricity Index upper limb; Mi leg = Motricity Index lower limb; TCT = Trunk Control Test; BI = Barthel Index. a Cut-off points are based on clinical grounds. b Cut-off points are based on analysis of receiver-operating characteristic curves. PPV 0.89 (95% CI ) and the NPV 0.87 (95% CI ). 4. Discussion The main result of this study was that age and gender, together with upper limb strength and BI measured at ten days post-stroke are the significant determinants for prediction of functional recovery of complex ADL at six months, as measured by the mrs. The model shows that having a BI 9, a MI upper limb 75, an age 70 and being a male resulted in 100% probability of achieving independence in complex ADL. The probability was high and approximately the same in patients who presented three of the four variables with little change in probability whichever of the four was absent. Inspection of the regression coefficients of the multiple regression analysis (Table 3) shows a difference roughly of 60% in the probability of favourable prognosis depending on the robustness of any of the two

5 D. Cioncoloni et al. / Predictors of long-term recovery in complex activities of daily living 221 Table 3 Probability to be independent in complex ADL at six months post-stroke (N. 104) mrs 2 at six months post-stroke BI MI Arm Age Gender True negatives, N False negatives, N False positives, N True positives, N p Cut-off p =1/(1+(exp ( BI MI arm age gender) )) Abbreviations: mrs = modified Rankin Scale; BI = Barthel Index; MI Arm = Motricity Index upper limb; = male; p = probability. variables with which the equation was performed. The higher probability (87%) of predicting good outcome occurred when the regression coefficients of MI-upper limb and age were taken on, similarly to the probability related to the presence of three determinants. These data indicate that the gain of independence in complex ADL depends mainly on the age and on the residual severity of the upper limb paresis after the acute phase poststroke. More specifically, a MI-upper limb total score 75 is more frequently achieved when patients are able to perform two specific movements (shoulder abduction - elbow flexion) against resistance and, thirdly, when a functional pinch grip is present. This is in line with the evidence that final recovery of upper limb strength and independence in basic ADL depend mainly on the initial severity of motor impairment or function [22, 23] and residual integrity of the corticospinal tract [24]. Thus, the robustness of the variable MI-upper limb mirrors the anatomical and physiological integrity of the corticospinal tract post-stroke, which is crucial for regaining independence in complex ADL at six months [24]. Regarding age, our finding is in line with current evidence predicting post-stroke outcome [6, 18]. It confirms that vascular compensation and a more pronounced neural plasticity in the lesser old population during the process of neurological recovery are presumably relevant contributing factors for a more favorable outcome, as experimental stroke models in rats have shown [25]. Moreover, comorbidities [16] and the rapidly increasing frequency of post-stroke cognitive deficits and dementia in the elderly [26], may be considered additional factors which contribute to a worse outcome. The third model s significant variable was gender. Indeed, males had a better functional outcome than females. The predictive value of this determinant is in line with some observational studies [27, 28] which show a worse outcome in females. Nevertheless, biological explanations remain lacking and the poorer outcome in female seems to be associated with other covariates, such as more advanced age, poorer health before stroke, tendency to sustain more severe stroke and different social status [27]. In our model, male with age 70 years presented a higher probability to regain independence than female. Finally, considering a cut-off value of 9, the BI showed significant discriminative properties in predicting final outcome in complex ADL. It is worth underlining that, the cut-off of 9 was two points higher than the score used for predicting independence in basic ADL (BI 19) at six months [14], at time moment evaluations similar to those used in our study. Based on these findings, the long-term prognostic value of simpler measures such as the Stroke Levity Scale [29] may be also explored to facilitate easier adoption by clinicians. In view of the recent definition of handicap as restriction of participation and involvement in general life situations [30], the results of the present study may help optimize rehabilitation plans and treatments focused on an earlier recovery in participation. In patients with good or quite good prognosis, an earlier restoration of

6 222 D. Cioncoloni et al. / Predictors of long-term recovery in complex activities of daily living independence in complex ADL could decrease poststroke consequences related to social isolation and the risk of suboptimal Health-Related Quality of Life on final outcome, on account of depression [31]. This study presents some limits such as the inclusion of hemorrhage and ischemic stroke together in the cohort. In addition, a cross-validation of the model with an independent sample of stroke unit patients, having the same inclusion criteria, is needed. In conclusion, in stroke patients, strength of the paretic upper limb, age, gender, and ability to perform basic ADL before the discharge from the stroke unit are the significant predictors at 10 days post-stroke of independence in complex ADL at six months. Acknowledgments We wish to thank all the patients who participated in the study, the nurse staff of the Stroke Unit of Siena University Hospital, the rehabilitation staff of Siena University Hospital and ASL 7, and Drs. Marotta, D Andrea and Lo Giudice for helping with data collection. References [1] Roger, V. L., Go, A. S., Lloyd-Jones, D. M., Benjamin, E. J., Berry, J. D., Borden, W. B., et al. (2012). Heart disease and stroke statistics 2012 update: A report from the american heart association. Circulation, 125, [2] Zhang, Y., Chapman, A. M., Plested, M., Jackson, D., & Purroy, F. (2012). The Incidence, Prevalence, and Mortality of Stroke in France, Germany, Italy, Spain, the UK, and the US: A Literature Review. Stroke Res Treat, 2012, [3] Miller, E. L., Murray, L., Richards, L., Zorowitz, R. D., Bakas, T., Clark, P., et al. (2010). Comprehensive overview of nursing and interdisciplinary rehabilitation care of the stroke patient: A scientific statement from the American Heart Association. Stroke, 41, [4] Lord, S. E., McPherson, K., McNaughton, H. K., Rochester, L., & Weatherall, M. (2004). Community ambulation after stroke: How important and obtainable is it and what measures appear predictive? Archives of Physical Medical. Rehabilitation, 85, [5] Röding, J., Glader, E. L., Malm, J., Eriksson, M., & Lindström, B. (2009). Perceived impaired physical and cognitive functions after stroke in men and women between 18 and 55 years of age a national survey. Disabil Rehabil, 31, [6] Veerbeek, J. M., Kwakkel, G., van Wegen, E. E. H., Johannes, C. F., Ket, J. C. F., & Heymans, M. W. (2011). Early Prediction of Outcome of Activities of Daily Living After Stroke: A Systematic Review. Stroke, 42, [7] Kwakkel, G., Kollen, B., & Twisk, J. (2006). Impact of time on improvement of outcome after stroke. Stroke, 37, [8] Quinn, T. J., Dawson, J., Walters, M. R., & Lees, K. R. (2009). Reliability of the Modified Rankin Scale: A Systematic Review. Stroke, 40, [9] Woldag, H., Gerhold, L. L., de Groot, M., Wohlfart, K., Wagner, A., & Hummelsheim, H. (2006). Early prediction of functional outcome after stroke. Brain Inj, 20, [10] Cioncoloni, D., Piu, P., Tassi, R., Acampa, M., Guideri, F., Taddei, S., et al. (2012). Relationship between the modified Rankin Scale and the Barthel Index in the process of functional recovery after stroke. NeuroRehabilitation, 30, [11] Moons, K., Altman, D., Vergouwe, Y., & Royston, P. (2009). Prognosis and prognostic research: Application and impact of prognostic models in clinical practice. BMJ, 338, b606. [12] Weisscher, N., Vermeulen, M., Roos, Y. B., & de Haan, R. J. (2008). What should be defined as good outcome in stroke trials; a modified Rankin score of 0-1 or 0-2? J Neurol, 255, [13] Kwakkel, G., Veerbeek, J. M., van Wegen, E. E., Nijland, R., Harmeling-van der Wel, B. C., Dippel, D. W., et al. (2010). Predictive value of the NIHSS for ADL outcome after ischemic hemispheric stroke: Does timing of early assessment matter? J Neurol Sci, 15(294), [14] Kwakkel, G., Veerbeek, J. M., Harmeling-van der Wel, B. C., van Wegen, E., & Kollen, B. J., (2011). Early Prediction of functional Outcome after Stroke (EPOS) Investigators. Diagnostic accuracy of the Barthel Index for measuring activities of daily living outcome after ischemic hemispheric stroke: Does early poststroke timing of assessment matter? Stroke, 42, [15] Tyson, S. F., Chillala, J., Hanley, M., Selley, A. B., & Tallis, R. C. (2006). Distribution of weakness in the upper and lower limbs post-stroke. Disabil Rehabil, 28, [16] Suzuki, M., Omori, Y., Sugimura, S., Miyamoto, M., Sugimura, Y., Kirimoto, H., & Yamada, S. (2011). Predicting recovery of bilateral upper extremity muscle strength after stroke. J Rehabil Med, 43, [17] Gall, S. L., Tran, P. L., Martin, K., Blizzard, L., & Srikanth, V. (2012). Sex differences in long-term outcomes after stroke: Functional outcomes, handicap, and quality of life. Stroke, 43, [18] Knoflach, M., Matosevic, B., Rücker, M., Furtner, M., Mair, A., Wille, G., et al. (2012). Austrian Stroke Unit Registry Collaborators. Functional recovery after ischemic stroke a matter of age: Data from the Austrian Stroke Unit Registry. Neurology, 78, [19] Goedert, K. M., Chen, P., Botticello, A., Masmela, J. R., Adler, U., & Barrett, A. M. (2012). Psychometric evaluation of neglect assessment reveals motor-exploratory predictor of functional disability in acute-stage spatial neglect. Arch Phys Med Rehabil, 93, [20] Patel, A., Duncan, P., Lai, S., & Studenski, S. (2000). The relation between impairments and functional outcomes poststroke. Arch Phys Med Rehabil, 81, [21] Tyson, S. F., Hanley, M., Chillala, J., Selley, A. B., & Tallis, R. C. (2008). Sensory loss in hospital-admitted people with stroke: Characteristics, associated factors, and relationship with function. Neurorehabil Neural Repair, 22, [22] Coupar, F., Pollock, A., Rowe, P., Weir, C., & Langhorne, P. (2012). Predictors of upper limb recovery after stroke: A systematic review and meta-analysis. Clin Rehabil, 26, [23] Harris, J. E., & Eng, J. J. (2007). Paretic upper-limb strength best explains arm activity in people with stroke. Phys Ther, 87,

7 D. Cioncoloni et al. / Predictors of long-term recovery in complex activities of daily living 223 [24] Stinear, C. M., Barber, P. A., Smale, P. R., Coxon, J. P., Fleming, M. K., & Byblow, W. D. (2007). Functional potential in chronic stroke patients depends on corticospinal tract integrity. Brain, 130, [25] Petcu, E. B., Sfredel, V., Platt, D., Herndon, J. G., Kessler, C., & Popa-Wagner, A. (2008). Cellular and molecular events underlying the dysregulated response of the aged brain to stroke: A mini-review. Gerontology, 54, [26] Pendlebury, S. T., & Rothwell, P. M. (2009). Prevalence, incidence, and factors associated with pre-stroke and post-stroke dementia: A systematic review and meta-analysis. Lancet Neurol, 8, [27] Gall, S. L., Donnan, G., Dewey, H. M., Macdonell, R., Sturm, J., Gilligan, A., et al. (2010). Sex differences in presentation, severity, and management of stroke in a population-based study. Neurology, 74, [28] Appelros, P., Stegmayr, B., & Terent, A. (2010). A review on sex differences in stroke treatment and outcome. Acta Neurol Scand, 121, [29] Owolabi, M. O., & Platz, T. (2008). Proposing the Stroke Levity Scale: A valid, reliable, simple, and time-saving measure of stroke severity. Eur J Neurol, 15, [30] World Health Organization: International Classification of Functioning, Disability and Health: ICF. Geneva, Switzerland, WHO, [31] Haley, W. E., Roth, D. L., Kissela, B., Perkins, M., & Howard, G. (2011). Quality of life after stroke: A prospective longitudinal study. Qual Life Res, 20,